Crypt-xxHash-0.09000755001750001750 015146654120 15433 5ustar00dchernenkodchernenko000000000000Crypt-xxHash-0.09/Build.PL000444001750001750 360615146654120 17071 0ustar00dchernenkodchernenko000000000000package main; use utf8; use warnings; use strict; use Module::Build (); use Config (); use File::Which qw/ which /; use IPC::Run qw/ run timeout /; my $has_cpp_compiler = scalar(which("c++")); my $arch; if (my $which =(which("uname"))[0]) { my ($out, $err); my $is_success = run( [$which, '-m'], \undef, \$out, \$err, timeout(1) ); chomp($out); $arch = $out if $out; } my %not_sse_archs = map { $_ => 1 } qw/ armv7l aarch64 /; my %Config = %Config::Config; my @cc_flags = split(/ /, $Config{'ccflags'}); push(@cc_flags, "-O3", "-msse2", "-msse4.2") if $arch && !$not_sse_archs{$arch}; push(@cc_flags, "-DCPP=1", "-std=c++17") if $has_cpp_compiler; push(@cc_flags, "-std=c17") if !$has_cpp_compiler; my $build = Module::Build->new( module_name => 'Crypt::xxHash', license => 'bsd', create_readme => 1, configure_requires => { 'Module::Build' => undef, # xs }, config => { $has_cpp_compiler ? (cc => 'c++', ld => 'c++') : (), }, build_requires => { 'File::Which' => undef, 'IPC::Run' => undef, 'ExtUtils::CBuilder' => undef, 'Test::More' => undef, 'Math::Int64' => undef # 32bit perl without int64 }, requires => { }, needs_compiler => 1, extra_compiler_flags => [@cc_flags], extra_linker_flags => [@cc_flags], include_dirs => ['.', 'src', 'ext/xxHash'], # ext/xxHash/tests/collisions/allcodecs c_source => ['.', 'src'], xs_files => { './xs/xxHash.xs' => './lib/Crypt/xxHash.xs' }, meta_merge => { 'resources' => { 'bugtracker' => 'https://github.com/KalessinD/perl_crypt_xxhash/issues', 'repository' => 'https://github.com/KalessinD/perl_crypt_xxhash', }, } ); $build->create_build_script; 1; __END__ Crypt-xxHash-0.09/Changes000444001750001750 114715146654120 17066 0ustar00dchernenkodchernenko000000000000Changes ======= 0.09 - 2026 Feb 22 - updating xxHash C++ library up to v0.8.3 - @CindyLinz added 64 strem methods 0.08 0 2025 Nov 20 - ignore versions of required packages 0.07 - 2022 Sep 27 - do not use SSE instructions for ARM architecture 0.06 - 2021 Dec 02 - fixes in MANIFEST files 0.05 - 2021 Nov 28 - using v.0.8.1 of xxHash 0.04 - 2021 Apr 28 - Update xxHash up to v0.8.0 0.03 - 2021 March 26 - Fix possible issues with compilation under Perl with thread support 0.02 - 2021 March 16 - Fix dependencies 0.01 - 2021 Feb 19 - Initial version (xxHash v0.7.4). Crypt-xxHash-0.09/LICENSE000444001750001750 306715146654120 16603 0ustar00dchernenkodchernenko000000000000This software is Copyright (c) 2013 by Chernenko Dmitriy. This is free software, licensed under: The (three-clause) BSD License The BSD License Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. * Neither the name of Chernenko Dmitriy nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. Crypt-xxHash-0.09/MANIFEST000444001750001750 147715146654120 16732 0ustar00dchernenkodchernenko000000000000Build.PL Changes ext/xxHash/.travis.yml ext/xxHash/appveyor.yml ext/xxHash/CHANGELOG ext/xxHash/clib.json ext/xxHash/cmake_unofficial/.gitignore ext/xxHash/cmake_unofficial/CMakeLists.txt ext/xxHash/cmake_unofficial/JoinPaths.cmake ext/xxHash/cmake_unofficial/README.md ext/xxHash/cmake_unofficial/xxHashConfig.cmake.in ext/xxHash/doc/README.md ext/xxHash/doc/xxhash.cry ext/xxHash/doc/xxhash_spec.md ext/xxHash/Doxyfile ext/xxHash/Doxyfile-internal ext/xxHash/fuzz/fuzzer.c ext/xxHash/libxxhash.pc.in ext/xxHash/LICENSE ext/xxHash/Makefile ext/xxHash/README.md ext/xxHash/SECURITY.md ext/xxHash/xxh3.h ext/xxHash/xxh_x86dispatch.c ext/xxHash/xxh_x86dispatch.h ext/xxHash/xxhash.c ext/xxHash/xxhash.h lib/Crypt/xxHash.pm LICENSE MANIFEST MANIFEST.SKIP META.json META.yml README README.md src/ppport.h t/Crypt-xxHash.t xs/xxHash.xs Crypt-xxHash-0.09/MANIFEST.SKIP000444001750001750 17315146654120 17447 0ustar00dchernenkodchernenko000000000000#!install_default ^MYMETA\.yml$ ^MYMETA\.json$ ^\.git \.vstags ^_build ^Build$ ^blib \.bak$ \.old$ xs/\.c$ xs/\.h$ xs/\.o$ Crypt-xxHash-0.09/META.json000444001750001750 236015146654120 17212 0ustar00dchernenkodchernenko000000000000{ "abstract" : "xxHash implementation for Perl", "author" : [ "Chernenko Dmitiry cdn@cpan.org" ], "dynamic_config" : 1, "generated_by" : "Module::Build version 0.4234", "license" : [ "bsd" ], "meta-spec" : { "url" : "http://search.cpan.org/perldoc?CPAN::Meta::Spec", "version" : 2 }, "name" : "Crypt-xxHash", "prereqs" : { "build" : { "requires" : { "ExtUtils::CBuilder" : "0", "File::Which" : "0", "IPC::Run" : "0", "Math::Int64" : "0", "Test::More" : "0" } }, "configure" : { "requires" : { "Module::Build" : "0" } } }, "provides" : { "Crypt::xxHash" : { "file" : "lib/Crypt/xxHash.pm", "version" : "0.09" } }, "release_status" : "stable", "resources" : { "bugtracker" : { "web" : "https://github.com/KalessinD/perl_crypt_xxhash/issues" }, "license" : [ "http://opensource.org/licenses/BSD-3-Clause" ], "repository" : { "url" : "https://github.com/KalessinD/perl_crypt_xxhash" } }, "version" : "0.09", "x_serialization_backend" : "JSON::PP version 4.16" } Crypt-xxHash-0.09/META.yml000444001750001750 146315146654120 17045 0ustar00dchernenkodchernenko000000000000--- abstract: 'xxHash implementation for Perl' author: - 'Chernenko Dmitiry cdn@cpan.org' build_requires: ExtUtils::CBuilder: '0' File::Which: '0' IPC::Run: '0' Math::Int64: '0' Test::More: '0' configure_requires: Module::Build: '0' dynamic_config: 1 generated_by: 'Module::Build version 0.4234, CPAN::Meta::Converter version 2.150010' license: bsd meta-spec: url: http://module-build.sourceforge.net/META-spec-v1.4.html version: '1.4' name: Crypt-xxHash provides: Crypt::xxHash: file: lib/Crypt/xxHash.pm version: '0.09' resources: bugtracker: https://github.com/KalessinD/perl_crypt_xxhash/issues license: http://opensource.org/licenses/BSD-3-Clause repository: https://github.com/KalessinD/perl_crypt_xxhash version: '0.09' x_serialization_backend: 'CPAN::Meta::YAML version 0.018' Crypt-xxHash-0.09/README000444001750001750 765715146654120 16467 0ustar00dchernenkodchernenko000000000000NAME Crypt::xxHash - xxHash implementation for Perl SYNOPSIS use Crypt::xxHash qw/ xxhash32 xxhash32_hex xxhash64 xxhash64_hex xxhash3_64bits xxhash3_64bits_hex xxhash3_128bits_hex /; my $hash = xxhash32( $data, $seed ); my $hex = xxhash32_hex( $data, $seed ); my $hash_64 = xxhash64( $data, $seed ); my $hex_64 = xxhash64_hex( $data, $seed ); my $hash_64 = xxhash3_64bits( $data, $seed ); my $hex_64 = xxhash3_64bits_hex( $data, $seed ); my $hex_128 = xxhash3_128bits_hex( $data, $seed ); DESCRIPTION xxHash is a super fast algorithm that claims to work at speeds close to RAM limits. This package provides 32- and 64-bit hash functions. As bonus it provides 128-bit hex method. This package was inspired by Digest::xxHash, but it includes the fresh C code and has some optimisations. Thus all hex methods are implemented in Perl XS. Also this module doesn't use Math::Int64 in favor of native 64 bit digits. The used version of xxHash is v0.8.0 $h = xxhash32( $data, $seed ) Returns a 32 bit hash. $h = xxhash32_hex( $data, $seed ) Returns a 32 bit hash converted into hex string. $h = xxhash64( $data, $seed ) Returns a 64 bit hash. $h = xxhash64_hex( $data, $seed ) Returns a 64 bit hash converted into hex string $h = xxhash3_64bits( $data, $seed ) Returns a 64 bit hash which calculated by using xxHash3 algorithm. $h = xxhash3_64bits_hex( $data, $seed ) Returns a 64 bit hash which calculated by using xxHash3 algorithm. This hash is converted into hex string. $h = xxhash3_128its_hex( $data, $seed ) Returns a 128 bit hash which calculated by using xxHash3 algorithm. This hash is converted into hex string. $stream = xxhash3_64bits_stream($seed) =head2 xxhash3_64bits_stream_update($stream, $more_data) =head2 $h = xxhash3_64bits_stream_digest($stream) =head2 $h = xxhash3_64bits_stream_digest_hex($stream) Get a 64 bit hash from segmented data by calling xxhash3_64bits_stream_update multiple times. sub hash_a_file { my($fh) = @_; my $stream = xxhash3_64bits_stream(12345); my $buf; while( read $fh, $buf, 1024 ) { xxhash3_64bits_stream_update($stream, $buf); } return xxhash3_64bits_stream_digest($stream); # return xxhash3_64bits_stream_digest_hex($stream); # hex version } # the resources $stream occupied will be released here. SPEED There are some official benchmark results can be found on the project web-site https://github.com/Cyan4973/xxHash BENCHMARKS Below you can find some benchmarks in comparison with Digest::xxHash. The "xxhash32" methods in those two packages have the same realisation. so they have the same throughput capacity. But other methods have some differences: Rate Digest::xxHash::xxhash32_hex Crypt::xxHash::xxhash32_hex Digest::xxHash::xxhash32_hex 2577320/s -- -54% Crypt::xxHash::xxhash32_hex 5543237/s 115% -- Rate Digest::xxHash::xxhash64 Crypt::xxHash::xxhash64 Digest::xxHash::xxhash64 201729/s -- -99% Crypt::xxHash::xxhash64 14893617/s 7283% -- Rate Digest::xxHash::xxhash64_hex Crypt::xxHash::xxhash64_hex Digest::xxHash::xxhash64_hex 185048/s -- -96% Crypt::xxHash::xxhash64_hex 4926108/s 2562% -- LICENSE xxHash is covered by the BSD license. AUTHOR Chernenko Dmitiry cdn@cpan.org xxHash by Yann Collet. Crypt-xxHash-0.09/README.md000444001750001750 737415146654120 17062 0ustar00dchernenkodchernenko000000000000# perl_crypt_xxhash ### NAME Crypt::xxHash - xxHash implementation for Perl ### SYNOPSIS ```perl use Crypt::xxHash qw/ xxhash32 xxhash32_hex xxhash64 xxhash64_hex xxhash3_64bits xxhash3_64bits_hex xxhash3_128bits_hex /; my $hash = xxhash32( $data, $seed ); my $hex = xxhash32_hex( $data, $seed ); my $hash_64 = xxhash64( $data, $seed ); my $hex_64 = xxhash64_hex( $data, $seed ); my $hash_64 = xxhash3_64bits( $data, $seed ); my $hex_64 = xxhash3_64bits_hex( $data, $seed ); my $hex_128 = xxhash3_128bits_hex( $data, $seed ); ``` ### DESCRIPTION xxHash is a super fast algorithm that claims to work at speeds close to RAM limits. This package provides 32- and 64-bit hash functions. As bonus it provides 128-bit hex method. This package was inspired by "Digest::xxHash", but it includes the fresh C code and has some optimisations. Thus all hex methods are implemented in Perl XS. Also this module doesn't use "Math::Int64" in favor of native 64 bit digits. The used version of xxHash is v0.8.0 ##### $h = xxhash32( $data, $seed ) Returns a 32 bit hash. ##### $h = xxhash32_hex( $data, $seed ) Returns a 32 bit hash converted into hex string. ##### $h = xxhash64( $data, $seed ) Returns a 64 bit hash. =head2 $h = xxhash64_hex( $data, $seed ) Returns a 64 bit hash converted into hex string ##### $h = xxhash3_64bits( $data, $seed ) Returns a 64 bit hash which calculated by using xxHash3 algorithm. ##### $h = xxhash3_64bits_hex( $data, $seed ) Returns a 64 bit hash which calculated by using xxHash3 algorithm. This hash is converted into hex string. ##### $h = xxhash3_128its_hex( $data, $seed ) Returns a 128 bit hash which calculated by using xxHash3 algorithm. This hash is converted into hex string. ##### $stream = xxhash3_64bits_stream($seed) ##### xxhash3_64bits_stream_update($stream, $more_data) ##### $h = xxhash3_64bits_stream_digest($stream) ##### $h = xxhash3_64bits_stream_digest_hex($stream) Get a 64 bit hash from segmented data by calling xxhash3_64bits_stream_update multiple times. ```perl sub hash_a_file { my($fh) = @_; my $stream = xxhash3_64bits_stream(12345); my $buf; while( read $fh, $buf, 1024 ) { xxhash3_64bits_stream_update($stream, $buf); } return xxhash3_64bits_stream_digest($stream); # return xxhash3_64bits_stream_digest_hex($stream); # hex version } # the resources $stream occupied will be released here. ``` ### SPEED There are some official benchmark results can be found on the project web-site L ### Benchmarks Below you can find some benchmarks in comparison with C. The "xxhash32" methods in those two packages have the same realisation. so they have the same throughput capacity. But other methods have some differences: ``` Rate Digest::xxHash::xxhash32_hex Crypt::xxHash::xxhash32_hex Digest::xxHash::xxhash32_hex 2577320/s -- -54% Crypt::xxHash::xxhash32_hex 5543237/s 115% -- Rate Digest::xxHash::xxhash64 Crypt::xxHash::xxhash64 Digest::xxHash::xxhash64 201729/s -- -99% Crypt::xxHash::xxhash64 14893617/s 7283% -- Rate Digest::xxHash::xxhash64_hex Crypt::xxHash::xxhash64_hex Digest::xxHash::xxhash64_hex 185048/s -- -96% Crypt::xxHash::xxhash64_hex 4926108/s 2562% -- ``` ### LICENSE xxHash is covered by the BSD license. ### AUTHOR Chernenko Dmitiry cdn@cpan.org xxHash by Yann Collet. Crypt-xxHash-0.09/ext000755001750001750 015146654120 16233 5ustar00dchernenkodchernenko000000000000Crypt-xxHash-0.09/ext/xxHash000755001750001750 015146654120 17476 5ustar00dchernenkodchernenko000000000000Crypt-xxHash-0.09/ext/xxHash/.travis.yml000444001750001750 1223415146654120 21766 0ustar00dchernenkodchernenko000000000000language: c # Dump CPU info before start before_install: - cat /proc/cpuinfo || echo /proc/cpuinfo is not present matrix: fast_finish: true include: - name: General linux x64 tests arch: amd64 addons: apt: packages: - g++-multilib - gcc-multilib - cppcheck script: - make -B test-all - make clean - CFLAGS="-Werror" MOREFLAGS="-Wno-sign-conversion" make dispatch # removing sign conversion warnings due to a bug in gcc-5's definition of some AVX512 intrinsics - make clean - CC=g++ CFLAGS="-O1 -mavx512f -Werror" make - make clean - CC=g++ CFLAGS="-Wall -Wextra -Werror" make DISPATCH=1 - name: Check results consistency on x64 arch: amd64 script: - CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR make check # Scalar code path - make clean - CPPFLAGS=-DXXH_VECTOR=XXH_SSE2 make check # SSE2 code path - make clean - CPPFLAGS="-mavx2 -DXXH_VECTOR=XXH_AVX2" make check # AVX2 code path - make clean - CPPFLAGS="-mavx512f -DXXH_VECTOR=XXH_AVX512" make check # AVX512 code path - make clean - CPPFLAGS=-DXXH_REROLL=1 make check # reroll code path (#240) - make -C tests/bench - name: macOS General Test os: osx compiler: clang script: - CFLAGS="-Werror" make # test library build - make clean - make test MOREFLAGS='-Werror' | tee # test scenario where `stdout` is not the console - name: ARM compilation and consistency checks (Qemu) dist: xenial arch: amd64 addons: apt: packages: - qemu-system-arm - qemu-user-static - gcc-arm-linux-gnueabi - libc6-dev-armel-cross script: # arm (32-bit) - CC=arm-linux-gnueabi-gcc CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR LDFLAGS=-static RUN_ENV=qemu-arm-static make check # Scalar code path - make clean # NEON (32-bit) - CC=arm-linux-gnueabi-gcc CPPFLAGS=-DXXH_VECTOR=XXH_NEON CFLAGS="-O3 -march=armv7-a -fPIC -mfloat-abi=softfp -mfpu=neon-vfpv4" LDFLAGS=-static RUN_ENV=qemu-arm-static make check # NEON code path - name: aarch64 compilation and consistency checks dist: xenial arch: arm64 script: # aarch64 - CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR make check # Scalar code path # NEON (64-bit) - make clean - CPPFLAGS=-DXXH_VECTOR=XXH_NEON make check # NEON code path # clang - make clean - CC=clang CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR make check # Scalar code path # clang + NEON - make clean - CC=clang CPPFLAGS=-DXXH_VECTOR=XXH_NEON make check # NEON code path # We need Bionic here because the QEMU versions shipped in the older repos # do not support POWER8 emulation, and compiling QEMU from source is a pain. - name: PowerPC + PPC64 compilation and consistency checks (Qemu on Bionic) dist: bionic arch: amd64 addons: apt: packages: - qemu-system-ppc - qemu-user-static - gcc-powerpc-linux-gnu - gcc-powerpc64-linux-gnu - libc6-dev-powerpc-cross - libc6-dev-ppc64-cross script: - CC=powerpc-linux-gnu-gcc RUN_ENV=qemu-ppc-static LDFLAGS=-static make check # Scalar code path - make clean - CC=powerpc64-linux-gnu-gcc RUN_ENV=qemu-ppc64-static CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR CFLAGS="-O3" LDFLAGS="-static -m64" make check # Scalar code path # VSX code - make clean - CC=powerpc64-linux-gnu-gcc RUN_ENV="qemu-ppc64-static -cpu power8" CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-O3 -maltivec -mvsx -mcpu=power8 -mpower8-vector" LDFLAGS="-static -m64" make check # VSX code path # altivec.h redefinition issue #426 - make clean - CC=powerpc64-linux-gnu-gcc CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-maltivec -mvsx -mcpu=power8 -mpower8-vector" make -C tests test_ppc_redefine - name: PPC64LE compilation and consistency checks dist: xenial arch: ppc64le script: # Scalar (universal) code path - CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR LDFLAGS=-static make check # VSX code path (64-bit) - make clean - CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-O3 -maltivec -mvsx -mpower8-vector -mcpu=power8" LDFLAGS="-static" make check # altivec.h redefinition issue #426 - make clean - CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-maltivec -mvsx -mcpu=power8 -mpower8-vector" make -C tests test_ppc_redefine - name: IBM s390x compilation and consistency checks dist: bionic arch: s390x script: # Scalar (universal) code path - CPPFLAGS=-DXXH_VECTOR=XXH_SCALAR LDFLAGS=-static make check # s390x code path (64-bit) - make clean - CPPFLAGS=-DXXH_VECTOR=XXH_VSX CFLAGS="-O3 -march=arch11 -mzvector" LDFLAGS="-static" make check - name: cmake build test script: - cd cmake_unofficial - mkdir build - cd build - cmake .. - CFLAGS=-Werror make Crypt-xxHash-0.09/ext/xxHash/CHANGELOG000444001750001750 1423015146654120 21065 0ustar00dchernenkodchernenko000000000000v0.8.3 - fix : variant `XXH3_128bits_withSecretandSeed()` could produce an invalid result in some specific set of conditions, #894 by @hltj - cli : vector extension detected at runtime on x86/x64, enabled by default - cli : new commands `--filelist` and `--files-from`, by @Ian-Clowes - cli : XXH3 64-bits GNU format can now be generated and checked (command `-H3`) - portability: LoongArch SX SIMD extension, by @lrzlin - portability: can build on AIX, suggested by @likema - portability: validated for SPARC cpus v0.8.2 - fix : XXH3 S390x vector implementation (@hzhuang1) - fix : PowerPC vector compilation with IBM XL compiler (@MaxiBoether) - perf : improved WASM speed by x2/x3 using SIMD128 (@easyaspi314) - perf : improved speed (+20%) for XXH3 on ARM NEON (@easyaspi314) - cli : Fix filename contain /LF character (@t-mat) - cli : Support # comment lines in --check files (@t-mat) - cli : Support commands --binary and --ignore-missing (@t-mat) - build: fix -Og compilation (@easyaspi314, @t-mat) - build: fix pkgconfig generation with cmake (@ilya-fedin) - build: fix icc compilation - build: fix cmake install directories - build: new build options XXH_NO_XXH3, XXH_SIZE_OPT and XXH_NO_STREAM to reduce binary size (@easyaspi314) - build: dedicated install targets (@ffontaine) - build: support DISPATCH mode in cmake (@hzhuang1) - portability: fix x86dispatch when building with Visual + clang-cl (@t-mat) - portability: SVE vector implementation of XXH3 (@hzhuang1) - portability: compatibility with freestanding environments, using XXH_NO_STDLIB - portability: can build on Haiku (@Begasus) - portability: validated on m68k and risc-v - doc : XXH3 specification (@Adrien1018) - doc : improved doxygen documentation (@easyaspi314, @t-mat) - misc : dedicated sanity test binary (@t-mat) v0.8.1 - perf : much improved performance for XXH3 streaming variants, notably on gcc and msvc - perf : improved XXH64 speed and latency on small inputs - perf : small XXH32 speed and latency improvement on small inputs of random size - perf : minor stack usage improvement for XXH32 and XXH64 - api : new experimental variants XXH3_*_withSecretandSeed() - api : update XXH3_generateSecret(), can no generate secret of any size (>= XXH3_SECRET_SIZE_MIN) - cli : xxhsum can now generate and check XXH3 checksums, using command `-H3` - build: can build xxhash without XXH3, with new build macro XXH_NO_XXH3 - build: fix xxh_x86dispatch build with MSVC, by @apankrat - build: XXH_INLINE_ALL can always be used safely, even after XXH_NAMESPACE or a previous XXH_INLINE_ALL - build: improved PPC64LE vector support, by @mpe - install: fix pkgconfig, by @ellert - install: compatibility with Haiku, by @Begasus - doc : code comments made compatible with doxygen, by @easyaspi314 - misc : XXH_ACCEPT_NULL_INPUT_POINTER is no longer necessary, all functions can accept NULL input pointers, as long as size == 0 - misc : complete refactor of CI tests on Github Actions, offering much larger coverage, by @t-mat - misc : xxhsum code base split into multiple specialized units, within directory cli/, by @easyaspi314 v0.8.0 - api : stabilize XXH3 - cli : xxhsum can parse BSD-style --check lines, by @WayneD - cli : `xxhsum -` accepts console input, requested by @jaki - cli : xxhsum accepts -- separator, by @jaki - cli : fix : print correct default algo for symlinked helpers, by @martinetd - install: improved pkgconfig script, allowing custom install locations, requested by @ellert v0.7.4 - perf: automatic vector detection and selection at runtime (`xxh_x86dispatch.h`), initiated by @easyaspi314 - perf: added AVX512 support, by @gzm55 - api : new: secret generator `XXH_generateSecret()`, suggested by @koraa - api : fix: XXH3_state_t is movable, identified by @koraa - api : fix: state is correctly aligned in AVX mode (unlike `malloc()`), by @easyaspi314 - api : fix: streaming generated wrong values in some combination of random ingestion lengths, reported by @WayneD - cli : fix unicode print on Windows, by @easyaspi314 - cli : can `-c` check file generated by sfv - build: `make DISPATCH=1` generates `xxhsum` and `libxxhash` with runtime vector detection (x86/x64 only) - install: cygwin installation support - doc : Cryptol specification of XXH32 and XXH64, by @weaversa v0.7.3 - perf: improved speed for large inputs (~+20%) - perf: improved latency for small inputs (~10%) - perf: s390x Vectorial code, by @easyaspi314 - cli: improved support for Unicode filenames on Windows, thanks to @easyaspi314 and @t-mat - api: `xxhash.h` can now be included in any order, with and without `XXH_STATIC_LINKING_ONLY` and `XXH_INLINE_ALL` - build: xxHash's implementation transferred into `xxhash.h`. No more need to have `xxhash.c` in the `/include` directory for `XXH_INLINE_ALL` to work - install: created pkg-config file, by @bket - install: VCpkg installation instructions, by @LilyWangL - doc: Highly improved code documentation, by @easyaspi314 - misc: New test tool in `/tests/collisions`: brute force collision tester for 64-bit hashes v0.7.2 - Fixed collision ratio of `XXH128` for some specific input lengths, reported by @svpv - Improved `VSX` and `NEON` variants, by @easyaspi314 - Improved performance of scalar code path (`XXH_VECTOR=0`), by @easyaspi314 - `xxhsum`: can generate 128-bit hashes with the `-H2` option (note: for experimental purposes only! `XXH128` is not yet frozen) - `xxhsum`: option `-q` removes status notifications v0.7.1 - Secret first: the algorithm computation can be altered by providing a "secret", which is any blob of bytes, of size >= `XXH3_SECRET_SIZE_MIN`. - `seed` is still available, and acts as a secret generator - updated `ARM NEON` variant by @easyaspi314 - Streaming implementation is available - Improve compatibility and performance with Visual Studio, with help from @aras-p - Better integration when using `XXH_INLINE_ALL`: do not pollute host namespace, use its own macros, such as `XXH_ASSERT()`, `XXH_ALIGN`, etc. - 128-bit variant provides helper functions for comparison of hashes. - Better `clang` generation of `rotl` instruction, thanks to @easyaspi314 - `XXH_REROLL` build macro to reduce binary size, by @easyaspi314 - Improved `cmake` script, by @Mezozoysky - Full benchmark program provided in `/tests/bench` Crypt-xxHash-0.09/ext/xxHash/Doxyfile000444001750001750 410715146654120 21343 0ustar00dchernenkodchernenko000000000000# Doxygen config for xxHash DOXYFILE_ENCODING = UTF-8 PROJECT_NAME = "xxHash" PROJECT_NUMBER = "0.8.2" PROJECT_BRIEF = "Extremely fast non-cryptographic hash function" OUTPUT_DIRECTORY = doxygen OUTPUT_LANGUAGE = English # We already separate the internal docs. INTERNAL_DOCS = NO # Consistency SORT_MEMBER_DOCS = NO BRIEF_MEMBER_DESC = YES REPEAT_BRIEF = YES # Warnings QUIET = YES # Until we document everything WARN_IF_UNDOCUMENTED = NO # TODO: Add the other files. It is just xxhash.h for now. FILE_PATTERNS = xxhash.h xxh_x86dispatch.c # Note: xxHash's source files are technically ASCII only. INPUT_ENCODING = UTF-8 TAB_SIZE = 4 MARKDOWN_SUPPORT = YES # xxHash is a C library OPTIMIZE_OUTPUT_FOR_C = YES # We hide private part from public document EXTRACT_STATIC = NO # We hide private part from public document EXTRACT_PRIVATE = NO # Document the macros MACRO_EXPANSION = YES EXPAND_ONLY_PREDEF = YES # Predefine some macros to clean up the output. PREDEFINED = "XXH_DOXYGEN=" \ "XXH_PUBLIC_API=" \ "XXH_NOESCAPE=" \ "XXH_FORCE_INLINE=static inline" \ "XXH_NO_INLINE=static" \ "XXH_RESTRICT=restrict" \ "XSUM_API=" \ "XXH_STATIC_LINKING_ONLY" \ "XXH_IMPLEMENTATION" \ "XXH_PUREF=[[gnu::pure]]" \ "XXH_CONSTF=[[gnu::const]]" \ "XXH_MALLOCF=[[gnu::malloc]]" \ "XXH_ALIGN(N)=alignas(N)" \ "XXH_ALIGN_MEMBER(align,type)=alignas(align) type" # We want HTML docs GENERATE_HTML = YES HTML_OUTPUT = html HTML_FILE_EXTENSION = .html # Tweak the colors a bit HTML_COLORSTYLE_HUE = 220 HTML_COLORSTYLE_GAMMA = 100 HTML_COLORSTYLE_SAT = 100 # We don't want LaTeX. GENERATE_LATEX = NO Crypt-xxHash-0.09/ext/xxHash/Doxyfile-internal000444001750001750 410515146654120 23153 0ustar00dchernenkodchernenko000000000000# Doxygen config for xxHash DOXYFILE_ENCODING = UTF-8 PROJECT_NAME = "xxHash" PROJECT_NUMBER = "0.8.2" PROJECT_BRIEF = "Extremely fast non-cryptographic hash function" OUTPUT_DIRECTORY = doxygen OUTPUT_LANGUAGE = English # We already separate the internal docs. INTERNAL_DOCS = YES # Consistency SORT_MEMBER_DOCS = NO BRIEF_MEMBER_DESC = YES REPEAT_BRIEF = YES # Warnings QUIET = YES # Until we document everything WARN_IF_UNDOCUMENTED = NO # TODO: Add the other files. It is just xxhash.h for now. FILE_PATTERNS = xxhash.h xxh_x86dispatch.c # Note: xxHash's source files are technically ASCII only. INPUT_ENCODING = UTF-8 TAB_SIZE = 4 MARKDOWN_SUPPORT = YES # xxHash is a C library OPTIMIZE_OUTPUT_FOR_C = YES # So we can document the internals EXTRACT_STATIC = YES # We show private part in the internal document EXTRACT_PRIVATE = YES # Document the macros MACRO_EXPANSION = YES EXPAND_ONLY_PREDEF = YES # Predefine some macros to clean up the output. PREDEFINED = "XXH_DOXYGEN=" \ "XXH_PUBLIC_API=" \ "XXH_NOESCAPE=" \ "XXH_FORCE_INLINE=static inline" \ "XXH_NO_INLINE=static" \ "XXH_RESTRICT=restrict" \ "XSUM_API=" \ "XXH_STATIC_LINKING_ONLY" \ "XXH_IMPLEMENTATION" \ "XXH_PUREF=[[gnu::pure]]" \ "XXH_CONSTF=[[gnu::const]]" \ "XXH_MALLOCF=[[gnu::malloc]]" \ "XXH_ALIGN(N)=alignas(N)" \ "XXH_ALIGN_MEMBER(align,type)=alignas(align) type" # We want HTML docs GENERATE_HTML = YES HTML_OUTPUT = html HTML_FILE_EXTENSION = .html # Tweak the colors a bit HTML_COLORSTYLE_HUE = 220 HTML_COLORSTYLE_GAMMA = 100 HTML_COLORSTYLE_SAT = 100 # We don't want LaTeX. GENERATE_LATEX = NO Crypt-xxHash-0.09/ext/xxHash/LICENSE000444001750001750 255515146654120 20647 0ustar00dchernenkodchernenko000000000000xxHash Library Copyright (c) 2012-2021 Yann Collet All rights reserved. BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) Redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met: * Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer. * Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution. THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT HOLDER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. Crypt-xxHash-0.09/ext/xxHash/Makefile000444001750001750 6305315146654120 21322 0ustar00dchernenkodchernenko000000000000# ################################################################ # xxHash Makefile # Copyright (C) 2012-2024 Yann Collet # # GPL v2 License # # This program is free software; you can redistribute it and/or modify # it under the terms of the GNU General Public License as published by # the Free Software Foundation; either version 2 of the License, or # (at your option) any later version. # # This program is distributed in the hope that it will be useful, # but WITHOUT ANY WARRANTY; without even the implied warranty of # MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the # GNU General Public License for more details. # # You should have received a copy of the GNU General Public License along # with this program; if not, write to the Free Software Foundation, Inc., # 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA. # # You can contact the author at: # - xxHash homepage: https://www.xxhash.com # - xxHash source repository: https://github.com/Cyan4973/xxHash # ################################################################ # xxhsum: provides 32/64 bits hash of one or multiple files, or stdin # ################################################################ Q = $(if $(filter 1,$(V) $(VERBOSE)),,@) # Version numbers SED ?= sed SED_ERE_OPT ?= -E LIBVER_MAJOR_SCRIPT:=`$(SED) -n '/define XXH_VERSION_MAJOR/s/.*[[:blank:]]\([0-9][0-9]*\).*/\1/p' < xxhash.h` LIBVER_MINOR_SCRIPT:=`$(SED) -n '/define XXH_VERSION_MINOR/s/.*[[:blank:]]\([0-9][0-9]*\).*/\1/p' < xxhash.h` LIBVER_PATCH_SCRIPT:=`$(SED) -n '/define XXH_VERSION_RELEASE/s/.*[[:blank:]]\([0-9][0-9]*\).*/\1/p' < xxhash.h` LIBVER_MAJOR := $(shell echo $(LIBVER_MAJOR_SCRIPT)) LIBVER_MINOR := $(shell echo $(LIBVER_MINOR_SCRIPT)) LIBVER_PATCH := $(shell echo $(LIBVER_PATCH_SCRIPT)) LIBVER := $(LIBVER_MAJOR).$(LIBVER_MINOR).$(LIBVER_PATCH) CFLAGS ?= -O3 DEBUGFLAGS+=-Wall -Wextra -Wconversion -Wcast-qual -Wcast-align -Wshadow \ -Wstrict-aliasing=1 -Wswitch-enum -Wdeclaration-after-statement \ -Wstrict-prototypes -Wundef -Wpointer-arith -Wformat-security \ -Wvla -Wformat=2 -Winit-self -Wfloat-equal -Wwrite-strings \ -Wredundant-decls -Wstrict-overflow=2 CFLAGS += $(DEBUGFLAGS) $(MOREFLAGS) FLAGS = $(CFLAGS) $(CPPFLAGS) XXHSUM_VERSION = $(LIBVER) UNAME := $(shell uname) # Define *.exe as extension for Windows systems ifneq (,$(filter Windows%,$(OS))) EXT =.exe else EXT = endif # automatically enable runtime vector dispatch on x86/64 targets detect_x86_arch = $(shell $(CC) -dumpmachine | grep -E 'i[3-6]86|x86_64') ifneq ($(strip $(call detect_x86_arch)),) #note: can be overridden at compile time, by setting DISPATCH=0 DISPATCH ?= 1 endif ifeq ($(NODE_JS),1) # Link in unrestricted filesystem support LDFLAGS += -sNODERAWFS # Set flag to fix isatty() support CPPFLAGS += -DXSUM_NODE_JS=1 endif # OS X linker doesn't support -soname, and use different extension # see: https://developer.apple.com/library/mac/documentation/DeveloperTools/Conceptual/DynamicLibraries/100-Articles/DynamicLibraryDesignGuidelines.html ifeq ($(UNAME), Darwin) SHARED_EXT = dylib SHARED_EXT_MAJOR = $(LIBVER_MAJOR).$(SHARED_EXT) SHARED_EXT_VER = $(LIBVER).$(SHARED_EXT) SONAME_FLAGS = -install_name $(LIBDIR)/libxxhash.$(SHARED_EXT_MAJOR) -compatibility_version $(LIBVER_MAJOR) -current_version $(LIBVER) else SONAME_FLAGS = -Wl,-soname=libxxhash.$(SHARED_EXT).$(LIBVER_MAJOR) SHARED_EXT = so SHARED_EXT_MAJOR = $(SHARED_EXT).$(LIBVER_MAJOR) SHARED_EXT_VER = $(SHARED_EXT).$(LIBVER) endif LIBXXH = libxxhash.$(SHARED_EXT_VER) XXHSUM_SRC_DIR = cli XXHSUM_SPLIT_SRCS = $(XXHSUM_SRC_DIR)/xxhsum.c \ $(XXHSUM_SRC_DIR)/xsum_os_specific.c \ $(XXHSUM_SRC_DIR)/xsum_arch.c \ $(XXHSUM_SRC_DIR)/xsum_output.c \ $(XXHSUM_SRC_DIR)/xsum_sanity_check.c \ $(XXHSUM_SRC_DIR)/xsum_bench.c XXHSUM_SPLIT_OBJS = $(XXHSUM_SPLIT_SRCS:.c=.o) XXHSUM_HEADERS = $(XXHSUM_SRC_DIR)/xsum_config.h \ $(XXHSUM_SRC_DIR)/xsum_arch.h \ $(XXHSUM_SRC_DIR)/xsum_os_specific.h \ $(XXHSUM_SRC_DIR)/xsum_output.h \ $(XXHSUM_SRC_DIR)/xsum_sanity_check.h \ $(XXHSUM_SRC_DIR)/xsum_bench.h ## generate CLI and libraries in release mode (default for `make`) .PHONY: default default: DEBUGFLAGS= default: lib xxhsum_and_links .PHONY: all all: lib xxhsum xxhsum_inlinedXXH ## xxhsum is the command line interface (CLI) ifeq ($(DISPATCH),1) xxhsum: CPPFLAGS += -DXXHSUM_DISPATCH=1 xxhsum: xxh_x86dispatch.o endif xxhsum: xxhash.o $(XXHSUM_SPLIT_OBJS) $(CC) $(FLAGS) $^ $(LDFLAGS) -o $@$(EXT) xxhsum32: CFLAGS += -m32 ## generate CLI in 32-bits mode xxhsum32: xxhash.c $(XXHSUM_SPLIT_SRCS) ## do not generate object (avoid mixing different ABI) $(CC) $(FLAGS) $^ $(LDFLAGS) -o $@$(EXT) ## dispatch only works for x86/x64 systems dispatch: CPPFLAGS += -DXXHSUM_DISPATCH=1 dispatch: xxhash.o xxh_x86dispatch.o $(XXHSUM_SPLIT_SRCS) $(CC) $(FLAGS) $^ $(LDFLAGS) -o $@$(EXT) xxhash.o: xxhash.c xxhash.h xxhsum.o: $(XXHSUM_SRC_DIR)/xxhsum.c $(XXHSUM_HEADERS) \ xxhash.h xxh_x86dispatch.h xxh_x86dispatch.o: xxh_x86dispatch.c xxh_x86dispatch.h xxhash.h .PHONY: xxhsum_and_links xxhsum_and_links: xxhsum xxh32sum xxh64sum xxh128sum xxh3sum xxh32sum xxh64sum xxh128sum xxh3sum: xxhsum ln -sf $<$(EXT) $@$(EXT) xxhsum_inlinedXXH: CPPFLAGS += -DXXH_INLINE_ALL xxhsum_inlinedXXH: $(XXHSUM_SPLIT_SRCS) $(CC) $(FLAGS) $< -o $@$(EXT) # library libxxhash.a: ARFLAGS = rcs libxxhash.a: xxhash.o $(AR) $(ARFLAGS) $@ $^ $(LIBXXH): LDFLAGS += -shared ifeq (,$(filter Windows%,$(OS))) $(LIBXXH): CFLAGS += -fPIC endif ifeq ($(LIBXXH_DISPATCH),1) $(LIBXXH): xxh_x86dispatch.c endif $(LIBXXH): xxhash.c $(CC) $(FLAGS) $^ $(LDFLAGS) $(SONAME_FLAGS) -o $@ ln -sf $@ libxxhash.$(SHARED_EXT_MAJOR) ln -sf $@ libxxhash.$(SHARED_EXT) .PHONY: libxxhash libxxhash: ## generate dynamic xxhash library libxxhash: $(LIBXXH) .PHONY: lib lib: ## generate static and dynamic xxhash libraries lib: libxxhash.a libxxhash # helper targets AWK = awk GREP = grep SORT = sort NM = nm .PHONY: list list: ## list all Makefile targets $(Q)$(MAKE) -pRrq -f $(lastword $(MAKEFILE_LIST)) : 2>/dev/null | $(AWK) -v RS= -F: '/^# File/,/^# Finished Make data base/ {if ($$1 !~ "^[#.]") {print $$1}}' | $(SORT) | egrep -v -e '^[^[:alnum:]]' -e '^$@$$' | xargs .PHONY: help help: ## list documented targets $(Q)$(GREP) -E '^[0-9a-zA-Z_-]+:.*?## .*$$' $(MAKEFILE_LIST) | \ $(SORT) | \ $(AWK) 'BEGIN {FS = ":.*?## "}; {printf "\033[36m%-30s\033[0m %s\n", $$1, $$2}' .PHONY: clean clean: ## remove all build artifacts $(Q)$(RM) -r *.dSYM # Mac OS-X specific $(Q)$(RM) core *.o *.obj *.$(SHARED_EXT) *.$(SHARED_EXT).* *.a libxxhash.pc $(Q)$(RM) xxhsum$(EXT) xxhsum32$(EXT) xxhsum_inlinedXXH$(EXT) dispatch$(EXT) $(Q)$(RM) xxhsum.wasm xxhsum.js xxhsum.html $(Q)$(RM) xxh32sum$(EXT) xxh64sum$(EXT) xxh128sum$(EXT) xxh3sum$(EXT) $(Q)$(RM) fuzzer $(Q)$(RM) $(XXHSUM_SRC_DIR)/*.o $(XXHSUM_SRC_DIR)/*.obj $(MAKE) -C tests clean $(MAKE) -C tests/bench clean $(MAKE) -C tests/collisions clean @echo cleaning completed # ================================================= # tests # ================================================= # make check can be run with cross-compiled binaries on emulated environments (qemu user mode) # by setting $(RUN_ENV) to the target emulation environment .PHONY: check check: xxhsum test_sanity ## basic tests for xxhsum CLI, set RUN_ENV for emulated environments # stdin # If you get "Wrong parameters" on Emscripten+Node.js, recompile with `NODE_JS=1` $(RUN_ENV) ./xxhsum$(EXT) < xxhash.c # multiple files $(RUN_ENV) ./xxhsum$(EXT) xxhash.* # internal bench $(RUN_ENV) ./xxhsum$(EXT) -bi0 # long bench command $(RUN_ENV) ./xxhsum$(EXT) --benchmark-all -i0 # bench multiple variants $(RUN_ENV) ./xxhsum$(EXT) -b1,2,3 -i0 # file bench $(RUN_ENV) ./xxhsum$(EXT) -bi0 xxhash.c # 32-bit $(RUN_ENV) ./xxhsum$(EXT) -H0 xxhash.c # 128-bit $(RUN_ENV) ./xxhsum$(EXT) -H2 xxhash.c # XXH3 (enforce BSD style) $(RUN_ENV) ./xxhsum$(EXT) -H3 xxhash.c | grep "XXH3" # request incorrect variant $(RUN_ENV) ./xxhsum$(EXT) -H9 xxhash.c ; test $$? -eq 1 @printf "\n ....... checks completed successfully ....... \n" .PHONY: test-unicode test-unicode: $(MAKE) -C tests test_unicode .PHONY: test_sanity test_sanity: $(MAKE) -C tests test_sanity .PHONY: test-mem VALGRIND = valgrind --leak-check=yes --error-exitcode=1 test-mem: RUN_ENV = $(VALGRIND) test-mem: xxhsum check .PHONY: test32 test32: xxhsum32 @echo ---- test 32-bit ---- ./xxhsum32 -bi0 xxhash.c TEST_FILES = xxhsum$(EXT) xxhash.c xxhash.h .PHONY: test-xxhsum-c test-xxhsum-c: xxhsum # xxhsum to/from pipe ./xxhsum $(TEST_FILES) | ./xxhsum -c - ./xxhsum -H0 $(TEST_FILES) | ./xxhsum -c - # xxhsum -c is unable to verify checksum of file from STDIN (#470) ./xxhsum < README.md > .test.README.md.xxh ./xxhsum -c .test.README.md.xxh < README.md # xxhsum -q does not display "Loading" message into stderr (#251) ! ./xxhsum -q $(TEST_FILES) 2>&1 | grep Loading # xxhsum does not display "Loading" message into stderr either ! ./xxhsum $(TEST_FILES) 2>&1 | grep Loading # Check that xxhsum do display filename that it failed to open. LC_ALL=C ./xxhsum nonexistent 2>&1 | grep "Error: Could not open 'nonexistent'" # xxhsum to/from file, shell redirection ./xxhsum $(TEST_FILES) > .test.xxh64 ./xxhsum --tag $(TEST_FILES) > .test.xxh64_tag ./xxhsum --little-endian $(TEST_FILES) > .test.le_xxh64 ./xxhsum --tag --little-endian $(TEST_FILES) > .test.le_xxh64_tag ./xxhsum -H0 $(TEST_FILES) > .test.xxh32 ./xxhsum -H0 --tag $(TEST_FILES) > .test.xxh32_tag ./xxhsum -H0 --little-endian $(TEST_FILES) > .test.le_xxh32 ./xxhsum -H0 --tag --little-endian $(TEST_FILES) > .test.le_xxh32_tag ./xxhsum -H2 $(TEST_FILES) > .test.xxh128 ./xxhsum -H2 --tag $(TEST_FILES) > .test.xxh128_tag ./xxhsum -H2 --little-endian $(TEST_FILES) > .test.le_xxh128 ./xxhsum -H2 --tag --little-endian $(TEST_FILES) > .test.le_xxh128_tag ./xxhsum -H3 $(TEST_FILES) > .test.xxh3 ./xxhsum -H3 --tag $(TEST_FILES) > .test.xxh3_tag ./xxhsum -H3 --little-endian $(TEST_FILES) > .test.le_xxh3 ./xxhsum -H3 --tag --little-endian $(TEST_FILES) > .test.le_xxh3_tag ./xxhsum -c .test.xxh* ./xxhsum -c --little-endian .test.le_xxh* ./xxhsum -c .test.*_tag # read list of files from stdin ./xxhsum -c < .test.xxh32 ./xxhsum -c < .test.xxh64 ./xxhsum -c < .test.xxh128 ./xxhsum -c < .test.xxh3 cat .test.xxh* | ./xxhsum -c - # check variant with '*' marker as second separator $(SED) 's/ / \*/' .test.xxh32 | ./xxhsum -c # bsd-style output ./xxhsum --tag xxhsum* | $(GREP) XXH64 ./xxhsum --tag -H0 xxhsum* | $(GREP) XXH32 ./xxhsum --tag -H1 xxhsum* | $(GREP) XXH64 ./xxhsum --tag -H2 xxhsum* | $(GREP) XXH128 ./xxhsum --tag -H3 xxhsum* | $(GREP) XXH3 ./xxhsum -H3 xxhsum* | $(GREP) XXH3_ # prefix for GNU format ./xxhsum --tag -H32 xxhsum* | $(GREP) XXH32 ./xxhsum --tag -H64 xxhsum* | $(GREP) XXH64 ./xxhsum --tag -H128 xxhsum* | $(GREP) XXH128 ./xxhsum --tag -H0 --little-endian xxhsum* | $(GREP) XXH32_LE ./xxhsum --tag -H1 --little-endian xxhsum* | $(GREP) XXH64_LE ./xxhsum --tag -H2 --little-endian xxhsum* | $(GREP) XXH128_LE ./xxhsum --tag -H3 --little-endian xxhsum* | $(GREP) XXH3_LE ./xxhsum --tag -H32 --little-endian xxhsum* | $(GREP) XXH32_LE ./xxhsum --tag -H64 --little-endian xxhsum* | $(GREP) XXH64_LE ./xxhsum --tag -H128 --little-endian xxhsum* | $(GREP) XXH128_LE # check bsd-style ./xxhsum --tag xxhsum* | ./xxhsum -c ./xxhsum --tag -H32 --little-endian xxhsum* | ./xxhsum -c # xxhsum -c warns improperly format lines. echo '12345678 ' >>.test.xxh32 ./xxhsum -c .test.xxh32 | $(GREP) improperly echo '123456789 file' >>.test.xxh64 ./xxhsum -c .test.xxh64 | $(GREP) improperly # Expects "FAILED" echo "0000000000000000 LICENSE" | ./xxhsum -c -; test $$? -eq 1 echo "00000000 LICENSE" | ./xxhsum -c -; test $$? -eq 1 # Expects "FAILED open or read" echo "0000000000000000 test-expects-file-not-found" | ./xxhsum -c -; test $$? -eq 1 echo "00000000 test-expects-file-not-found" | ./xxhsum -c -; test $$? -eq 1 # --filelist echo xxhash.c > .test.filenames $(RUN_ENV) ./xxhsum$(EXT) --filelist .test.filenames # --filelist from stdin cat .test.filenames | $(RUN_ENV) ./xxhsum$(EXT) --filelist @$(RM) .test.* LIB_FUZZING_ENGINE?="-fsanitize=fuzzer" CC_VERSION := $(shell $(CC) --version) ifneq (,$(findstring clang,$(CC_VERSION))) fuzzer: libxxhash.a fuzz/fuzzer.c $(CC) $(CFLAGS) $(LIB_FUZZING_ENGINE) -I. -o fuzzer fuzz/fuzzer.c -L. -Wl,-Bstatic -lxxhash -Wl,-Bdynamic endif .PHONY: test-filename-escape test-filename-escape: $(MAKE) -C tests test_filename_escape .PHONY: test-cli-comment-line test-cli-comment-line: $(MAKE) -C tests test_cli_comment_line .PHONY: test-cli-ignore-missing test-cli-ignore-missing: $(MAKE) -C tests test_cli_ignore_missing .PHONY: armtest armtest: clean @echo ---- test ARM compilation ---- CC=arm-linux-gnueabi-gcc MOREFLAGS="-Werror -static" $(MAKE) xxhsum .PHONY: clangtest clangtest: clean @echo ---- test clang compilation ---- CC=clang MOREFLAGS="-Werror -Wconversion -Wno-sign-conversion" $(MAKE) all .PHONY: gcc-og-test gcc-og-test: clean @echo ---- test gcc -Og compilation ---- CFLAGS="-Og -Wall -Wextra -Wundef -Wshadow -Wcast-align -Werror -fPIC" CPPFLAGS="-DXXH_NO_INLINE_HINTS" MOREFLAGS="-Werror" $(MAKE) all .PHONY: cxxtest cxxtest: clean @echo ---- test C++ compilation ---- CC="$(CXX) -Wno-deprecated" $(MAKE) all CFLAGS="-O3 -Wall -Wextra -Wundef -Wshadow -Wcast-align -Werror -fPIC" # In strict C90 mode, there is no `long long` type support, # consequently, only XXH32 can be compiled. .PHONY: c90test ifeq ($(NO_C90_TEST),true) c90test: @echo no c90 compatibility test else c90test: CPPFLAGS += -DXXH_NO_LONG_LONG c90test: CFLAGS += -std=c90 -Werror -pedantic c90test: xxhash.c @echo ---- test strict C90 compilation [xxh32 only] ---- $(RM) xxhash.o $(CC) $(FLAGS) $^ -c $(NM) xxhash.o | $(GREP) XXH64 ; test $$? -eq 1 $(RM) xxhash.o endif .PHONY: noxxh3test noxxh3test: CPPFLAGS += -DXXH_NO_XXH3 noxxh3test: CFLAGS += -Werror -pedantic -Wno-long-long # XXH64 requires long long support noxxh3test: OFILE = xxh_noxxh3.o noxxh3test: xxhash.c @echo ---- test compilation without XXH3 ---- $(CC) $(FLAGS) -c $^ -o $(OFILE) $(NM) $(OFILE) | $(GREP) XXH3_ ; test $$? -eq 1 $(RM) $(OFILE) .PHONY: nostreamtest nostreamtest: CPPFLAGS += -DXXH_NO_STREAM nostreamtest: CFLAGS += -Werror -pedantic -Wno-long-long # XXH64 requires long long support nostreamtest: OFILE = xxh_nostream.o nostreamtest: xxhash.c @echo ---- test compilation without streaming ---- $(CC) $(FLAGS) -c $^ -o $(OFILE) $(NM) $(OFILE) | $(GREP) update ; test $$? -eq 1 $(RM) $(OFILE) .PHONY: nostdlibtest nostdlibtest: CPPFLAGS += -DXXH_NO_STDLIB nostdlibtest: CFLAGS += -Werror -pedantic -Wno-long-long # XXH64 requires long long support nostdlibtest: OFILE = xxh_nostdlib.o nostdlibtest: xxhash.c @echo ---- test compilation without \ ---- $(CC) $(FLAGS) -c $^ -o $(OFILE) $(NM) $(OFILE) | $(GREP) "U _free\|U free" ; test $$? -eq 1 $(RM) $(OFILE) .PHONY: usan usan: CC=clang usan: CXX=clang++ usan: ## check CLI runtime for undefined behavior, using clang's sanitizer @echo ---- check undefined behavior - sanitize ---- $(MAKE) clean $(MAKE) test CC=$(CC) CXX=$(CXX) MOREFLAGS="-g -fsanitize=undefined -fno-sanitize-recover=all" .PHONY: staticAnalyze SCANBUILD ?= scan-build staticAnalyze: clean ## check C source files using $(SCANBUILD) static analyzer @echo ---- static analyzer - $(SCANBUILD) ---- CFLAGS="-g -Werror" $(SCANBUILD) --status-bugs -v $(MAKE) all CPPCHECK ?= cppcheck .PHONY: cppcheck cppcheck: ## check C source files using $(CPPCHECK) static analyzer @echo ---- static analyzer - $(CPPCHECK) ---- $(CPPCHECK) . --force --enable=warning,portability,performance,style --error-exitcode=1 > /dev/null .PHONY: namespaceTest namespaceTest: ## ensure XXH_NAMESPACE redefines all public symbols $(CC) -c xxhash.c $(CC) -DXXH_NAMESPACE=TEST_ -c xxhash.c -o xxhash2.o $(CC) xxhash.o xxhash2.o $(XXHSUM_SPLIT_SRCS) -o xxhsum2 # will fail if one namespace missing (symbol collision) $(RM) *.o xxhsum2 # clean MAN = $(XXHSUM_SRC_DIR)/xxhsum.1 MD2ROFF ?= ronn MD2ROFF_FLAGS ?= --roff --warnings --manual="User Commands" --organization="xxhsum $(XXHSUM_VERSION)" $(MAN): $(XXHSUM_SRC_DIR)/xxhsum.1.md xxhash.h cat $< | $(MD2ROFF) $(MD2ROFF_FLAGS) | $(SED) -n '/^\.\\\".*/!p' > $@ .PHONY: man man: $(MAN) ## generate man page from markdown source .PHONY: clean-man clean-man: $(RM) xxhsum.1 .PHONY: preview-man preview-man: man man ./xxhsum.1 .PHONY: test test: DEBUGFLAGS += -DXXH_DEBUGLEVEL=1 test: all namespaceTest check test-xxhsum-c c90test test-tools noxxh3test nostdlibtest # this test checks that including "xxhash.h" multiple times and with different directives still compiles properly .PHONY: test-multiInclude test-multiInclude: $(MAKE) -C tests test_multiInclude .PHONY: test-inline-notexposed test-inline-notexposed: xxhsum_inlinedXXH $(NM) xxhsum_inlinedXXH | $(GREP) "t _XXH32_" ; test $$? -eq 1 # no XXH32 symbol should be left $(NM) xxhsum_inlinedXXH | $(GREP) "t _XXH64_" ; test $$? -eq 1 # no XXH64 symbol should be left .PHONY: test-inline test-inline: test-inline-notexposed test-multiInclude .PHONY: test-all test-all: CFLAGS += -Werror test-all: test test32 test-unicode clangtest gcc-og-test cxxtest usan test-inline listL120 trailingWhitespace test-xxh-nnn-sums .PHONY: test-tools test-tools: CFLAGS=-Werror $(MAKE) -C tests/bench CFLAGS=-Werror $(MAKE) -C tests/collisions check .PHONY: test-xxh-nnn-sums test-xxh-nnn-sums: xxhsum_and_links ./xxhsum README.md > tmp.xxhsum.out # xxhsum outputs xxh64 ./xxh32sum README.md > tmp.xxh32sum.out ./xxh64sum README.md > tmp.xxh64sum.out ./xxh128sum README.md > tmp.xxh128sum.out ./xxh3sum README.md > tmp.xxh3sum.out cat tmp.xxhsum.out cat tmp.xxh32sum.out cat tmp.xxh64sum.out cat tmp.xxh128sum.out cat tmp.xxh3sum.out ./xxhsum -c tmp.xxhsum.out ./xxhsum -c tmp.xxh32sum.out ./xxhsum -c tmp.xxh64sum.out ./xxhsum -c tmp.xxh128sum.out ./xxhsum -c tmp.xxh3sum.out ./xxh32sum -c tmp.xxhsum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh32sum -c tmp.xxh32sum.out ./xxh32sum -c tmp.xxh64sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh32sum -c tmp.xxh128sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh32sum -c tmp.xxh3sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh64sum -c tmp.xxhsum.out ./xxh64sum -c tmp.xxh32sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh64sum -c tmp.xxh64sum.out ./xxh64sum -c tmp.xxh128sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh64sum -c tmp.xxh3sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh128sum -c tmp.xxhsum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh128sum -c tmp.xxh32sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh128sum -c tmp.xxh64sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh128sum -c tmp.xxh128sum.out ./xxh128sum -c tmp.xxh3sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh3sum -c tmp.xxhsum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh3sum -c tmp.xxh32sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh3sum -c tmp.xxh64sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh3sum -c tmp.xxh128sum.out ; test $$? -eq 1 # expects "no properly formatted" ./xxh3sum -c tmp.xxh3sum.out .PHONY: listL120 listL120: # extract lines >= 120 characters in *.{c,h}, by Takayuki Matsuoka (note: $$, for Makefile compatibility) find . -type f -name '*.c' -o -name '*.h' | while read -r filename; do awk 'length > 120 {print FILENAME "(" FNR "): " $$0}' $$filename; done .PHONY: trailingWhitespace trailingWhitespace: ! $(GREP) -E "`printf '[ \\t]$$'`" cli/*.c cli/*.h cli/*.1 *.c *.h LICENSE Makefile cmake_unofficial/CMakeLists.txt .PHONY: lint-unicode lint-unicode: ./tests/unicode_lint.sh # ========================================================= # make install is validated only for the following targets # ========================================================= ifneq (,$(filter Linux Darwin GNU/kFreeBSD GNU Haiku OpenBSD FreeBSD NetBSD DragonFly SunOS CYGWIN% , $(UNAME))) DESTDIR ?= # directory variables: GNU conventions prefer lowercase # see https://www.gnu.org/prep/standards/html_node/Makefile-Conventions.html # support both lower and uppercase (BSD), use uppercase in script prefix ?= /usr/local PREFIX ?= $(prefix) exec_prefix ?= $(PREFIX) EXEC_PREFIX ?= $(exec_prefix) libdir ?= $(EXEC_PREFIX)/lib LIBDIR ?= $(libdir) includedir ?= $(PREFIX)/include INCLUDEDIR ?= $(includedir) bindir ?= $(EXEC_PREFIX)/bin BINDIR ?= $(bindir) datarootdir ?= $(PREFIX)/share mandir ?= $(datarootdir)/man man1dir ?= $(mandir)/man1 ifneq (,$(filter $(UNAME),FreeBSD NetBSD DragonFly)) PKGCONFIGDIR ?= $(PREFIX)/libdata/pkgconfig else PKGCONFIGDIR ?= $(LIBDIR)/pkgconfig endif ifneq (,$(filter $(UNAME),OpenBSD FreeBSD NetBSD DragonFly SunOS)) MANDIR ?= $(PREFIX)/man/man1 else MANDIR ?= $(man1dir) endif ifneq (,$(filter $(UNAME),SunOS)) INSTALL ?= ginstall else INSTALL ?= install endif INSTALL_PROGRAM ?= $(INSTALL) INSTALL_DATA ?= $(INSTALL) -m 644 MAKE_DIR ?= $(INSTALL) -d -m 755 # Escape special symbols by putting each character into its separate class EXEC_PREFIX_REGEX ?= $(shell echo "$(EXEC_PREFIX)" | $(SED) $(SED_ERE_OPT) -e "s/([^^])/[\1]/g" -e "s/\\^/\\\\^/g") PREFIX_REGEX ?= $(shell echo "$(PREFIX)" | $(SED) $(SED_ERE_OPT) -e "s/([^^])/[\1]/g" -e "s/\\^/\\\\^/g") PCLIBDIR ?= $(shell echo "$(LIBDIR)" | $(SED) -n $(SED_ERE_OPT) -e "s@^$(EXEC_PREFIX_REGEX)(/|$$)@@p") PCINCDIR ?= $(shell echo "$(INCLUDEDIR)" | $(SED) -n $(SED_ERE_OPT) -e "s@^$(PREFIX_REGEX)(/|$$)@@p") PCEXECDIR?= $(if $(filter $(PREFIX),$(EXEC_PREFIX)),$$\{prefix\},$(EXEC_PREFIX)) ifeq (,$(PCLIBDIR)) # Additional prefix check is required, since the empty string is technically a # valid PCLIBDIR ifeq (,$(shell echo "$(LIBDIR)" | $(SED) -n $(SED_ERE_OPT) -e "\\@^$(EXEC_PREFIX_REGEX)(/|$$)@ p")) $(error configured libdir ($(LIBDIR)) is outside of exec_prefix ($(EXEC_PREFIX)), can't generate pkg-config file) endif endif ifeq (,$(PCINCDIR)) # Additional prefix check is required, since the empty string is technically a # valid PCINCDIR ifeq (,$(shell echo "$(INCLUDEDIR)" | $(SED) -n $(SED_ERE_OPT) -e "\\@^$(PREFIX_REGEX)(/|$$)@ p")) $(error configured includedir ($(INCLUDEDIR)) is outside of prefix ($(PREFIX)), can't generate pkg-config file) endif endif libxxhash.pc: libxxhash.pc.in @echo creating pkgconfig $(Q)$(SED) $(SED_ERE_OPT) -e 's|@PREFIX@|$(PREFIX)|' \ -e 's|@EXECPREFIX@|$(PCEXECDIR)|' \ -e 's|@LIBDIR@|$$\{exec_prefix\}/$(PCLIBDIR)|' \ -e 's|@INCLUDEDIR@|$$\{prefix\}/$(PCINCDIR)|' \ -e 's|@VERSION@|$(LIBVER)|' \ $< > $@ install_libxxhash.a: libxxhash.a @echo Installing libxxhash.a $(Q)$(MAKE_DIR) $(DESTDIR)$(LIBDIR) $(Q)$(INSTALL_DATA) libxxhash.a $(DESTDIR)$(LIBDIR) install_libxxhash: libxxhash @echo Installing libxxhash $(Q)$(MAKE_DIR) $(DESTDIR)$(LIBDIR) $(Q)$(INSTALL_PROGRAM) $(LIBXXH) $(DESTDIR)$(LIBDIR) $(Q)ln -sf $(LIBXXH) $(DESTDIR)$(LIBDIR)/libxxhash.$(SHARED_EXT_MAJOR) $(Q)ln -sf $(LIBXXH) $(DESTDIR)$(LIBDIR)/libxxhash.$(SHARED_EXT) install_libxxhash.includes: $(Q)$(INSTALL) -d -m 755 $(DESTDIR)$(INCLUDEDIR) # includes $(Q)$(INSTALL_DATA) xxhash.h $(DESTDIR)$(INCLUDEDIR) $(Q)$(INSTALL_DATA) xxh3.h $(DESTDIR)$(INCLUDEDIR) # for compatibility, will be removed in v0.9.0 ifeq ($(LIBXXH_DISPATCH),1) $(Q)$(INSTALL_DATA) xxh_x86dispatch.h $(DESTDIR)$(INCLUDEDIR) endif install_libxxhash.pc: libxxhash.pc @echo Installing pkgconfig $(Q)$(MAKE_DIR) $(DESTDIR)$(PKGCONFIGDIR)/ $(Q)$(INSTALL_DATA) libxxhash.pc $(DESTDIR)$(PKGCONFIGDIR)/ install_xxhsum: xxhsum @echo Installing xxhsum $(Q)$(MAKE_DIR) $(DESTDIR)$(BINDIR)/ $(Q)$(INSTALL_PROGRAM) xxhsum $(DESTDIR)$(BINDIR)/xxhsum $(Q)ln -sf xxhsum $(DESTDIR)$(BINDIR)/xxh32sum $(Q)ln -sf xxhsum $(DESTDIR)$(BINDIR)/xxh64sum $(Q)ln -sf xxhsum $(DESTDIR)$(BINDIR)/xxh128sum $(Q)ln -sf xxhsum $(DESTDIR)$(BINDIR)/xxh3sum install_man: @echo Installing man pages $(Q)$(MAKE_DIR) $(DESTDIR)$(MANDIR)/ $(Q)$(INSTALL_DATA) $(MAN) $(DESTDIR)$(MANDIR)/xxhsum.1 $(Q)ln -sf xxhsum.1 $(DESTDIR)$(MANDIR)/xxh32sum.1 $(Q)ln -sf xxhsum.1 $(DESTDIR)$(MANDIR)/xxh64sum.1 $(Q)ln -sf xxhsum.1 $(DESTDIR)$(MANDIR)/xxh128sum.1 $(Q)ln -sf xxhsum.1 $(DESTDIR)$(MANDIR)/xxh3sum.1 .PHONY: install ## install libraries, CLI, links and man pages install: install_libxxhash.a install_libxxhash install_libxxhash.includes install_libxxhash.pc install_xxhsum install_man @echo xxhash installation completed .PHONY: uninstall uninstall: ## uninstall libraries, CLI, links and man page $(Q)$(RM) $(DESTDIR)$(LIBDIR)/libxxhash.a $(Q)$(RM) $(DESTDIR)$(LIBDIR)/libxxhash.$(SHARED_EXT) $(Q)$(RM) $(DESTDIR)$(LIBDIR)/libxxhash.$(SHARED_EXT_MAJOR) $(Q)$(RM) $(DESTDIR)$(LIBDIR)/$(LIBXXH) $(Q)$(RM) $(DESTDIR)$(INCLUDEDIR)/xxhash.h $(Q)$(RM) $(DESTDIR)$(INCLUDEDIR)/xxh3.h $(Q)$(RM) $(DESTDIR)$(INCLUDEDIR)/xxh_x86dispatch.h $(Q)$(RM) $(DESTDIR)$(PKGCONFIGDIR)/libxxhash.pc $(Q)$(RM) $(DESTDIR)$(BINDIR)/xxh32sum $(Q)$(RM) $(DESTDIR)$(BINDIR)/xxh64sum $(Q)$(RM) $(DESTDIR)$(BINDIR)/xxh128sum $(Q)$(RM) $(DESTDIR)$(BINDIR)/xxh3sum $(Q)$(RM) $(DESTDIR)$(BINDIR)/xxhsum $(Q)$(RM) $(DESTDIR)$(MANDIR)/xxh32sum.1 $(Q)$(RM) $(DESTDIR)$(MANDIR)/xxh64sum.1 $(Q)$(RM) $(DESTDIR)$(MANDIR)/xxh128sum.1 $(Q)$(RM) $(DESTDIR)$(MANDIR)/xxh3sum.1 $(Q)$(RM) $(DESTDIR)$(MANDIR)/xxhsum.1 @echo xxhsum successfully uninstalled endif Crypt-xxHash-0.09/ext/xxHash/README.md000444001750001750 4162315146654120 21140 0ustar00dchernenkodchernenko000000000000 xxHash - Extremely fast hash algorithm ====================================== xxHash is an Extremely fast Hash algorithm, processing at RAM speed limits. Code is highly portable, and produces hashes identical across all platforms (little / big endian). The library includes the following algorithms : - XXH32 : generates 32-bit hashes, using 32-bit arithmetic - XXH64 : generates 64-bit hashes, using 64-bit arithmetic - XXH3 (since `v0.8.0`): generates 64 or 128-bit hashes, using vectorized arithmetic. The 128-bit variant is called XXH128. All variants successfully complete the [SMHasher](https://code.google.com/p/smhasher/wiki/SMHasher) test suite which evaluates the quality of hash functions (collision, dispersion and randomness). Additional tests, which evaluate more thoroughly speed and collision properties of 64-bit hashes, [are also provided](https://github.com/Cyan4973/xxHash/tree/dev/tests). |Branch |Status | |------------|---------| |release | [![Build Status](https://github.com/Cyan4973/xxHash/actions/workflows/ci.yml/badge.svg?branch=release)](https://github.com/Cyan4973/xxHash/actions?query=branch%3Arelease+) | |dev | [![Build Status](https://github.com/Cyan4973/xxHash/actions/workflows/ci.yml/badge.svg?branch=dev)](https://github.com/Cyan4973/xxHash/actions?query=branch%3Adev+) | Benchmarks ------------------------- The benchmarked reference system uses an Intel i7-9700K cpu, and runs Ubuntu x64 20.04. The [open source benchmark program] is compiled with `clang` v10.0 using `-O3` flag. | Hash Name | Width | Bandwidth (GB/s) | Small Data Velocity | Quality | Comment | | --------- | ----- | ---------------- | ----- | --- | --- | | __XXH3__ (SSE2) | 64 | 31.5 GB/s | 133.1 | 10 | __XXH128__ (SSE2) | 128 | 29.6 GB/s | 118.1 | 10 | _RAM sequential read_ | N/A | 28.0 GB/s | N/A | N/A | _for reference_ | City64 | 64 | 22.0 GB/s | 76.6 | 10 | T1ha2 | 64 | 22.0 GB/s | 99.0 | 9 | Slightly worse [collisions] | City128 | 128 | 21.7 GB/s | 57.7 | 10 | __XXH64__ | 64 | 19.4 GB/s | 71.0 | 10 | SpookyHash | 64 | 19.3 GB/s | 53.2 | 10 | Mum | 64 | 18.0 GB/s | 67.0 | 9 | Slightly worse [collisions] | __XXH32__ | 32 | 9.7 GB/s | 71.9 | 10 | City32 | 32 | 9.1 GB/s | 66.0 | 10 | Murmur3 | 32 | 3.9 GB/s | 56.1 | 10 | SipHash | 64 | 3.0 GB/s | 43.2 | 10 | FNV64 | 64 | 1.2 GB/s | 62.7 | 5 | Poor avalanche properties | Blake2 | 256 | 1.1 GB/s | 5.1 | 10 | Cryptographic | SHA1 | 160 | 0.8 GB/s | 5.6 | 10 | Cryptographic but broken | MD5 | 128 | 0.6 GB/s | 7.8 | 10 | Cryptographic but broken [open source benchmark program]: https://github.com/Cyan4973/xxHash/tree/release/tests/bench [collisions]: https://github.com/Cyan4973/xxHash/wiki/Collision-ratio-comparison#collision-study note 1: Small data velocity is a _rough_ evaluation of algorithm's efficiency on small data. For more detailed analysis, please refer to next paragraph. note 2: some algorithms feature _faster than RAM_ speed. In which case, they can only reach their full speed potential when input is already in CPU cache (L3 or better). Otherwise, they max out on RAM speed limit. ### Small data Performance on large data is only one part of the picture. Hashing is also very useful in constructions like hash tables and bloom filters. In these use cases, it's frequent to hash a lot of small data (starting at a few bytes). Algorithm's performance can be very different for such scenarios, since parts of the algorithm, such as initialization or finalization, become fixed cost. The impact of branch mis-prediction also becomes much more present. XXH3 has been designed for excellent performance on both long and small inputs, which can be observed in the following graph: ![XXH3, latency, random size](https://user-images.githubusercontent.com/750081/61976089-aedeab00-af9f-11e9-9239-e5375d6c080f.png) For a more detailed analysis, please visit the wiki : https://github.com/Cyan4973/xxHash/wiki/Performance-comparison#benchmarks-concentrating-on-small-data- Quality ------------------------- Speed is not the only property that matters. Produced hash values must respect excellent dispersion and randomness properties, so that any sub-section of it can be used to maximally spread out a table or index, as well as reduce the amount of collisions to the minimal theoretical level, following the [birthday paradox]. `xxHash` has been tested with Austin Appleby's excellent SMHasher test suite, and passes all tests, ensuring reasonable quality levels. It also passes extended tests from [newer forks of SMHasher], featuring additional scenarios and conditions. Finally, xxHash provides its own [massive collision tester](https://github.com/Cyan4973/xxHash/tree/dev/tests/collisions), able to generate and compare billions of hashes to test the limits of 64-bit hash algorithms. On this front too, xxHash features good results, in line with the [birthday paradox]. A more detailed analysis is documented [in the wiki](https://github.com/Cyan4973/xxHash/wiki/Collision-ratio-comparison). [birthday paradox]: https://en.wikipedia.org/wiki/Birthday_problem [newer forks of SMHasher]: https://github.com/rurban/smhasher ### Build modifiers The following macros can be set at compilation time to modify `libxxhash`'s behavior. They are generally disabled by default. - `XXH_INLINE_ALL`: Make all functions `inline`, implementation is directly included within `xxhash.h`. Inlining functions is beneficial for speed, notably for small keys. It's _extremely effective_ when key's length is expressed as _a compile time constant_, with performance improvements observed in the +200% range . See [this article](https://fastcompression.blogspot.com/2018/03/xxhash-for-small-keys-impressive-power.html) for details. - `XXH_PRIVATE_API`: same outcome as `XXH_INLINE_ALL`. Still available for legacy support. The name underlines that `XXH_*` symbol names will not be exported. - `XXH_STATIC_LINKING_ONLY`: gives access to internal state declaration, required for static allocation. Incompatible with dynamic linking, due to risks of ABI changes. - `XXH_NAMESPACE`: Prefixes all symbols with the value of `XXH_NAMESPACE`. This macro can only use compilable character set. Useful to evade symbol naming collisions, in case of multiple inclusions of xxHash's source code. Client applications still use the regular function names, as symbols are automatically translated through `xxhash.h`. - `XXH_FORCE_ALIGN_CHECK`: Use a faster direct read path when input is aligned. This option can result in dramatic performance improvement on architectures unable to load memory from unaligned addresses when input to hash happens to be aligned on 32 or 64-bit boundaries. It is (slightly) detrimental on platform with good unaligned memory access performance (same instruction for both aligned and unaligned accesses). This option is automatically disabled on `x86`, `x64` and `aarch64`, and enabled on all other platforms. - `XXH_FORCE_MEMORY_ACCESS`: The default method `0` uses a portable `memcpy()` notation. Method `1` uses a gcc-specific `packed` attribute, which can provide better performance for some targets. Method `2` forces unaligned reads, which is not standard compliant, but might sometimes be the only way to extract better read performance. Method `3` uses a byteshift operation, which is best for old compilers which don't inline `memcpy()` or big-endian systems without a byteswap instruction. - `XXH_CPU_LITTLE_ENDIAN`: By default, endianness is determined by a runtime test resolved at compile time. If, for some reason, the compiler cannot simplify the runtime test, it can cost performance. It's possible to skip auto-detection and simply state that the architecture is little-endian by setting this macro to 1. Setting it to 0 states big-endian. - `XXH_ENABLE_AUTOVECTORIZE`: Auto-vectorization may be triggered for XXH32 and XXH64, depending on cpu vector capabilities and compiler version. Note: auto-vectorization tends to be triggered more easily with recent versions of `clang`. For XXH32, SSE4.1 or equivalent (NEON) is enough, while XXH64 requires AVX512. Unfortunately, auto-vectorization is generally detrimental to XXH performance. For this reason, the xxhash source code tries to prevent auto-vectorization by default. That being said, systems evolve, and this conclusion is not forthcoming. For example, it has been reported that recent Zen4 cpus are more likely to improve performance with vectorization. Therefore, should you prefer or want to test vectorized code, you can enable this flag: it will remove the no-vectorization protection code, thus making it more likely for XXH32 and XXH64 to be auto-vectorized. - `XXH32_ENDJMP`: Switch multi-branch finalization stage of XXH32 by a single jump. This is generally undesirable for performance, especially when hashing inputs of random sizes. But depending on exact architecture and compiler, a jump might provide slightly better performance on small inputs. Disabled by default. - `XXH_IMPORT`: MSVC specific: should only be defined for dynamic linking, as it prevents linkage errors. - `XXH_NO_STDLIB`: Disable invocation of `` functions, notably `malloc()` and `free()`. `libxxhash`'s `XXH*_createState()` will always fail and return `NULL`. But one-shot hashing (like `XXH32()`) or streaming using statically allocated states still work as expected. This build flag is useful for embedded environments without dynamic allocation. - `XXH_DEBUGLEVEL` : When set to any value >= 1, enables `assert()` statements. This (slightly) slows down execution, but may help finding bugs during debugging sessions. #### Binary size control - `XXH_NO_XXH3` : removes symbols related to `XXH3` (both 64 & 128 bits) from generated binary. `XXH3` is by far the largest contributor to `libxxhash` size, so it's useful to reduce binary size for applications which do not employ `XXH3`. - `XXH_NO_LONG_LONG`: removes compilation of algorithms relying on 64-bit `long long` types which include `XXH3` and `XXH64`. Only `XXH32` will be compiled. Useful for targets (architectures and compilers) without 64-bit support. - `XXH_NO_STREAM`: Disables the streaming API, limiting the library to single shot variants only. - `XXH_NO_INLINE_HINTS`: By default, xxHash uses `__attribute__((always_inline))` and `__forceinline` to improve performance at the cost of code size. Defining this macro to 1 will mark all internal functions as `static`, allowing the compiler to decide whether to inline a function or not. This is very useful when optimizing for smallest binary size, and is automatically defined when compiling with `-O0`, `-Os`, `-Oz`, or `-fno-inline` on GCC and Clang. It may also be required to successfully compile using `-Og`, depending on compiler version. - `XXH_SIZE_OPT`: `0`: default, optimize for speed `1`: default for `-Os` and `-Oz`: disables some speed hacks for size optimization `2`: makes code as small as possible, performance may cry #### Build modifiers specific for XXH3 - `XXH_VECTOR` : manually select a vector instruction set (default: auto-selected at compilation time). Available instruction sets are `XXH_SCALAR`, `XXH_SSE2`, `XXH_AVX2`, `XXH_AVX512`, `XXH_NEON` and `XXH_VSX`. Compiler may require additional flags to ensure proper support (for example, `gcc` on x86_64 requires `-mavx2` for `AVX2`, or `-mavx512f` for `AVX512`). - `XXH_PREFETCH_DIST` : select prefetching distance. For close-to-metal adaptation to specific hardware platforms. XXH3 only. - `XXH_NO_PREFETCH` : disable prefetching. Some platforms or situations may perform better without prefetching. XXH3 only. #### Makefile variables When compiling the Command Line Interface `xxhsum` using `make`, the following environment variables can also be set : - `DISPATCH=1` : use `xxh_x86dispatch.c`, select at runtime between `scalar`, `sse2`, `avx2` or `avx512` instruction set. This option is only valid for `x86`/`x64` systems. It is enabled by default when target `x86`/`x64` is detected. It can be forcefully turned off using `DISPATCH=0`. - `LIBXXH_DISPATCH=1` : same idea, implemented a runtime vector extension detector, but within `libxxhash`. This parameter is disabled by default. When enabled (only valid for `x86`/`x64` systems), new symbols published in `xxh_x86dispatch.h` become accessible. At the time of this writing, it's required to include `xxh_x86dispatch.h` in order to access the symbols with runtime vector extension detection. - `XXH_1ST_SPEED_TARGET` : select an initial speed target, expressed in MB/s, for the first speed test in benchmark mode. Benchmark will adjust the target at subsequent iterations, but the first test is made "blindly" by targeting this speed. Currently conservatively set to 10 MB/s, to support very slow (emulated) platforms. - `NODE_JS=1` : When compiling `xxhsum` for Node.js with Emscripten, this links the `NODERAWFS` library for unrestricted filesystem access and patches `isatty` to make the command line utility correctly detect the terminal. This does make the binary specific to Node.js. ### Building xxHash - Using vcpkg You can download and install xxHash using the [vcpkg](https://github.com/Microsoft/vcpkg) dependency manager: git clone https://github.com/Microsoft/vcpkg.git cd vcpkg ./bootstrap-vcpkg.sh ./vcpkg integrate install ./vcpkg install xxhash The xxHash port in vcpkg is kept up to date by Microsoft team members and community contributors. If the version is out of date, please [create an issue or pull request](https://github.com/Microsoft/vcpkg) on the vcpkg repository. ### Example The simplest example calls xxhash 64-bit variant as a one-shot function generating a hash value from a single buffer, and invoked from a C/C++ program: ```C #include "xxhash.h" (...) XXH64_hash_t hash = XXH64(buffer, size, seed); } ``` Streaming variant is more involved, but makes it possible to provide data incrementally: ```C #include "stdlib.h" /* abort() */ #include "xxhash.h" XXH64_hash_t calcul_hash_streaming(FileHandler fh) { /* create a hash state */ XXH64_state_t* const state = XXH64_createState(); if (state==NULL) abort(); size_t const bufferSize = SOME_SIZE; void* const buffer = malloc(bufferSize); if (buffer==NULL) abort(); /* Initialize state with selected seed */ XXH64_hash_t const seed = 0; /* or any other value */ if (XXH64_reset(state, seed) == XXH_ERROR) abort(); /* Feed the state with input data, any size, any number of times */ (...) while ( /* some data left */ ) { size_t const length = get_more_data(buffer, bufferSize, fh); if (XXH64_update(state, buffer, length) == XXH_ERROR) abort(); (...) } (...) /* Produce the final hash value */ XXH64_hash_t const hash = XXH64_digest(state); /* State could be re-used; but in this example, it is simply freed */ free(buffer); XXH64_freeState(state); return hash; } ``` ### License The library files `xxhash.c` and `xxhash.h` are BSD licensed. The utility `xxhsum` is GPL licensed. ### Other programming languages Beyond the C reference version, xxHash is also available from many different programming languages, thanks to great contributors. They are [listed here](http://www.xxhash.com/#other-languages). ### Packaging status Many distributions bundle a package manager which allows easy xxhash installation as both a `libxxhash` library and `xxhsum` command line interface. [![Packaging status](https://repology.org/badge/vertical-allrepos/xxhash.svg)](https://repology.org/project/xxhash/versions) ### Special Thanks - Takayuki Matsuoka, aka @t-mat, for creating `xxhsum -c` and great support during early xxh releases - Mathias Westerdahl, aka @JCash, for introducing the first version of `XXH64` - Devin Hussey, aka @easyaspi314, for incredible low-level optimizations on `XXH3` and `XXH128` Crypt-xxHash-0.09/ext/xxHash/SECURITY.md000444001750001750 126015146654120 21423 0ustar00dchernenkodchernenko000000000000# Security Policy ## Supported Versions Security updates are applied only to the latest release. ## Reporting a Vulnerability If you have discovered a security vulnerability in this project, please report it privately. **Do not disclose it as a public issue.** This gives us time to work with you to fix the issue before public exposure, reducing the chance that the exploit will be used before a patch is released. Please disclose it at [security advisory](https://github.com/Cyan4973/xxHash/security/advisories/new). This project is maintained by a team of volunteers on a reasonable-effort basis. As such, please give us at least 90 days to work on a fix before public exposure. Crypt-xxHash-0.09/ext/xxHash/appveyor.yml000444001750001750 714215146654120 22227 0ustar00dchernenkodchernenko000000000000#---------------------------------# # general configuration # #---------------------------------# version: 1.0.{build} max_jobs: 2 #---------------------------------# # environment configuration # #---------------------------------# clone_depth: 2 environment: matrix: - COMPILER: "visual" ARCH: "x64" TEST_XXHSUM: "true" - COMPILER: "visual" ARCH: "Win32" TEST_XXHSUM: "true" - COMPILER: "visual" ARCH: "Win32" APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2013 TEST_XXHSUM: "true" - COMPILER: "visual" ARCH: "ARM" # Below tests are now disabled due to redundancy. # Their equivalent already runs correctly on Github Actions. # - COMPILER: "visual" # ARCH: "x64" # APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2017 # TEST_XXHSUM: "true" # - COMPILER: "visual" # ARCH: "ARM64" # APPVEYOR_BUILD_WORKER_IMAGE: Visual Studio 2017 # # note: ARM64 is not available with Visual Studio 14 2015, which is default for Appveyor # The following tests were also flacky on Appveyor, for various reasons. # - COMPILER: "gcc" # PLATFORM: "mingw64" # - COMPILER: "gcc" # PLATFORM: "mingw32" # - COMPILER: "gcc" # PLATFORM: "clang" install: - ECHO Installing %COMPILER% %PLATFORM% %ARCH% - MKDIR bin - if [%COMPILER%]==[gcc] SET PATH_ORIGINAL=%PATH% - if [%COMPILER%]==[gcc] ( SET "PATH_MINGW32=c:\MinGW\bin;c:\MinGW\usr\bin" && SET "PATH_MINGW64=c:\msys64\mingw64\bin;c:\msys64\usr\bin" && COPY C:\MinGW\bin\mingw32-make.exe C:\MinGW\bin\make.exe && COPY C:\MinGW\bin\gcc.exe C:\MinGW\bin\cc.exe ) #---------------------------------# # build configuration # #---------------------------------# build_script: - if [%PLATFORM%]==[mingw32] SET PATH=%PATH_MINGW32%;%PATH_ORIGINAL% - if [%PLATFORM%]==[mingw64] SET PATH=%PATH_MINGW64%;%PATH_ORIGINAL% - if [%PLATFORM%]==[clang] SET PATH=%PATH_MINGW64%;%PATH_ORIGINAL% - ECHO *** - ECHO Building %COMPILER% %PLATFORM% %ARCH% - ECHO *** - if [%COMPILER%]==[gcc] ( if [%PLATFORM%]==[clang] ( clang -v ) ELSE ( gcc -v ) ) - if [%COMPILER%]==[gcc] ( echo ----- && make -v && echo ----- && if not [%PLATFORM%]==[clang] ( if [%PLATFORM%]==[mingw32] ( SET CPPFLAGS=-DPOOL_MT=0 ) && make -B clean test MOREFLAGS=-Werror ) ELSE ( SET CXXFLAGS=--std=c++14 && make -B clean test CC=clang CXX=clang++ MOREFLAGS="--target=x86_64-w64-mingw32 -Werror -Wno-pass-failed" NO_C90_TEST=true ) && make -C tests/bench ) # note 1: strict c90 tests with clang fail, due to (erroneous) presence on `inline` keyword in some included system file # note 2: multi-threading code doesn't work with mingw32, disabled through POOL_MT=0 # note 3: clang requires C++14 to compile sort because its own code contains c++14-only code - if [%COMPILER%]==[visual] ( cd cmake_unofficial && cmake . -DCMAKE_BUILD_TYPE=Release -A %ARCH% -DXXHASH_C_FLAGS="/WX" && cmake --build . --config Release ) #---------------------------------# # tests configuration # #---------------------------------# test_script: # note: can only run x86 and x64 binaries on Appveyor # note: if %COMPILER%==gcc, xxhsum was already tested within `make test` - if [%TEST_XXHSUM%]==[true] ( ECHO *** && ECHO Testing %COMPILER% %PLATFORM% %ARCH% && ECHO *** && cd Release && xxhsum.exe -bi1 && ECHO ------- xxhsum tested ------- ) #---------------------------------# # artifacts configuration # #---------------------------------# # none yet Crypt-xxHash-0.09/ext/xxHash/clib.json000444001750001750 37715146654120 21426 0ustar00dchernenkodchernenko000000000000{ "name": "xxhash", "version": "0.8.2", "repo": "Cyan4973/xxhash", "description": "Extremely fast non-cryptographic hash algorithm", "keywords": ["xxhash", "hashing"], "license": "BSD-2-Clause", "src": [ "xxhash.c", "xxhash.h" ] } Crypt-xxHash-0.09/ext/xxHash/libxxhash.pc.in000444001750001750 63215146654120 22537 0ustar00dchernenkodchernenko000000000000# xxHash - Extremely fast hash algorithm # Copyright (C) 2012-2021, Yann Collet, Facebook # BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) prefix=@PREFIX@ exec_prefix=@EXECPREFIX@ includedir=@INCLUDEDIR@ libdir=@LIBDIR@ Name: xxhash Description: extremely fast hash algorithm URL: http://www.xxhash.com/ Version: @VERSION@ Libs: -L${libdir} -lxxhash Cflags: -I${includedir} Crypt-xxHash-0.09/ext/xxHash/xxh3.h000444001750001750 453215146654120 20702 0ustar00dchernenkodchernenko000000000000/* * xxHash - Extremely Fast Hash algorithm * Development source file for `xxh3` * Copyright (C) 2019-2021 Yann Collet * * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions are * met: * * * Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * Redistributions in binary form must reproduce the above * copyright notice, this list of conditions and the following disclaimer * in the documentation and/or other materials provided with the * distribution. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * You can contact the author at: * - xxHash homepage: https://www.xxhash.com * - xxHash source repository: https://github.com/Cyan4973/xxHash */ /* * Note: This file used to host the source code of XXH3_* variants. * during the development period. * The source code is now properly integrated within xxhash.h. * * xxh3.h is no longer useful, * but it is still provided for compatibility with source code * which used to include it directly. * * Programs are now highly discouraged to include xxh3.h. * Include `xxhash.h` instead, which is the officially supported interface. * * In the future, xxh3.h will start to generate warnings, then errors, * then it will be removed from source package and from include directory. */ /* Simulate the same impact as including the old xxh3.h source file */ #define XXH_INLINE_ALL #include "xxhash.h" Crypt-xxHash-0.09/ext/xxHash/xxh_x86dispatch.c000444001750001750 7657115146654120 23073 0ustar00dchernenkodchernenko000000000000/* * xxHash - Extremely Fast Hash algorithm * Copyright (C) 2020-2021 Yann Collet * * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions are * met: * * * Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * Redistributions in binary form must reproduce the above * copyright notice, this list of conditions and the following disclaimer * in the documentation and/or other materials provided with the * distribution. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * You can contact the author at: * - xxHash homepage: https://www.xxhash.com * - xxHash source repository: https://github.com/Cyan4973/xxHash */ /*! * @file xxh_x86dispatch.c * * Automatic dispatcher code for the @ref XXH3_family on x86-based targets. * * Optional add-on. * * **Compile this file with the default flags for your target.** * Note that compiling with flags like `-mavx*`, `-march=native`, or `/arch:AVX*` * will make the resulting binary incompatible with cpus not supporting the requested instruction set. * * @defgroup dispatch x86 Dispatcher * @{ */ #if defined (__cplusplus) extern "C" { #endif #if !(defined(__x86_64__) || defined(__i386__) || defined(_M_IX86) || defined(_M_X64)) # error "Dispatching is currently only supported on x86 and x86_64." #endif /*! @cond Doxygen ignores this part */ #ifndef XXH_HAS_INCLUDE # ifdef __has_include /* * Not defined as XXH_HAS_INCLUDE(x) (function-like) because * this causes segfaults in Apple Clang 4.2 (on Mac OS X 10.7 Lion) */ # define XXH_HAS_INCLUDE __has_include # else # define XXH_HAS_INCLUDE(x) 0 # endif #endif /*! @endcond */ /*! * @def XXH_DISPATCH_SCALAR * @brief Enables/dispatching the scalar code path. * * If this is defined to 0, SSE2 support is assumed. This reduces code size * when the scalar path is not needed. * * This is automatically defined to 0 when... * - SSE2 support is enabled in the compiler * - Targeting x86_64 * - Targeting Android x86 * - Targeting macOS */ #ifndef XXH_DISPATCH_SCALAR # if defined(__SSE2__) || (defined(_M_IX86_FP) && _M_IX86_FP >= 2) /* SSE2 on by default */ \ || defined(__x86_64__) || defined(_M_X64) /* x86_64 */ \ || defined(__ANDROID__) || defined(__APPLE__) /* Android or macOS */ # define XXH_DISPATCH_SCALAR 0 /* disable */ # else # define XXH_DISPATCH_SCALAR 1 # endif #endif /*! * @def XXH_DISPATCH_AVX2 * @brief Enables/disables dispatching for AVX2. * * This is automatically detected if it is not defined. * - GCC 4.7 and later are known to support AVX2, but >4.9 is required for * to get the AVX2 intrinsics and typedefs without -mavx -mavx2. * - Visual Studio 2013 Update 2 and later are known to support AVX2. * - The GCC/Clang internal header `` is detected. While this is * not allowed to be included directly, it still appears in the builtin * include path and is detectable with `__has_include`. * * @see XXH_AVX2 */ #ifndef XXH_DISPATCH_AVX2 # if (defined(__GNUC__) && (__GNUC__ > 4)) /* GCC 5.0+ */ \ || (defined(_MSC_VER) && _MSC_VER >= 1900) /* VS 2015+ */ \ || (defined(_MSC_FULL_VER) && _MSC_FULL_VER >= 180030501) /* VS 2013 Update 2 */ \ || XXH_HAS_INCLUDE() /* GCC/Clang internal header */ # define XXH_DISPATCH_AVX2 1 /* enable dispatch towards AVX2 */ # else # define XXH_DISPATCH_AVX2 0 # endif #endif /* XXH_DISPATCH_AVX2 */ /*! * @def XXH_DISPATCH_AVX512 * @brief Enables/disables dispatching for AVX512. * * Automatically detected if one of the following conditions is met: * - GCC 4.9 and later are known to support AVX512. * - Visual Studio 2017 and later are known to support AVX2. * - The GCC/Clang internal header `` is detected. While this * is not allowed to be included directly, it still appears in the builtin * include path and is detectable with `__has_include`. * * @see XXH_AVX512 */ #ifndef XXH_DISPATCH_AVX512 # if (defined(__GNUC__) \ && (__GNUC__ > 4 || (__GNUC__ == 4 && __GNUC_MINOR__ >= 9))) /* GCC 4.9+ */ \ || (defined(_MSC_VER) && _MSC_VER >= 1910) /* VS 2017+ */ \ || XXH_HAS_INCLUDE() /* GCC/Clang internal header */ # define XXH_DISPATCH_AVX512 1 /* enable dispatch towards AVX512 */ # else # define XXH_DISPATCH_AVX512 0 # endif #endif /* XXH_DISPATCH_AVX512 */ /*! * @def XXH_TARGET_SSE2 * @brief Allows a function to be compiled with SSE2 intrinsics. * * Uses `__attribute__((__target__("sse2")))` on GCC to allow SSE2 to be used * even with `-mno-sse2`. * * @def XXH_TARGET_AVX2 * @brief Like @ref XXH_TARGET_SSE2, but for AVX2. * * @def XXH_TARGET_AVX512 * @brief Like @ref XXH_TARGET_SSE2, but for AVX512. * */ #if defined(__GNUC__) # include /* SSE2 */ # if XXH_DISPATCH_AVX2 || XXH_DISPATCH_AVX512 # include /* AVX2, AVX512F */ # endif # define XXH_TARGET_SSE2 __attribute__((__target__("sse2"))) # define XXH_TARGET_AVX2 __attribute__((__target__("avx2"))) # define XXH_TARGET_AVX512 __attribute__((__target__("avx512f"))) #elif defined(__clang__) && defined(_MSC_VER) /* clang-cl.exe */ # include /* SSE2 */ # if XXH_DISPATCH_AVX2 || XXH_DISPATCH_AVX512 # include /* AVX2, AVX512F */ # include # include # include # include # endif # define XXH_TARGET_SSE2 __attribute__((__target__("sse2"))) # define XXH_TARGET_AVX2 __attribute__((__target__("avx2"))) # define XXH_TARGET_AVX512 __attribute__((__target__("avx512f"))) #elif defined(_MSC_VER) # include # define XXH_TARGET_SSE2 # define XXH_TARGET_AVX2 # define XXH_TARGET_AVX512 #else # error "Dispatching is currently not supported for your compiler." #endif /*! @cond Doxygen ignores this part */ #ifdef XXH_DISPATCH_DEBUG /* debug logging */ # include # define XXH_debugPrint(str) { fprintf(stderr, "DEBUG: xxHash dispatch: %s \n", str); fflush(NULL); } #else # define XXH_debugPrint(str) ((void)0) # undef NDEBUG /* avoid redefinition */ # define NDEBUG #endif /*! @endcond */ #include #ifndef XXH_DOXYGEN #define XXH_INLINE_ALL #define XXH_X86DISPATCH #include "xxhash.h" #endif /*! @cond Doxygen ignores this part */ #ifndef XXH_HAS_ATTRIBUTE # ifdef __has_attribute # define XXH_HAS_ATTRIBUTE(...) __has_attribute(__VA_ARGS__) # else # define XXH_HAS_ATTRIBUTE(...) 0 # endif #endif /*! @endcond */ /*! @cond Doxygen ignores this part */ #if XXH_HAS_ATTRIBUTE(constructor) # define XXH_CONSTRUCTOR __attribute__((constructor)) # define XXH_DISPATCH_MAYBE_NULL 0 #else # define XXH_CONSTRUCTOR # define XXH_DISPATCH_MAYBE_NULL 1 #endif /*! @endcond */ /*! @cond Doxygen ignores this part */ /* * Support both AT&T and Intel dialects * * GCC doesn't convert AT&T syntax to Intel syntax, and will error out if * compiled with -masm=intel. Instead, it supports dialect switching with * curly braces: { AT&T syntax | Intel syntax } * * Clang's integrated assembler automatically converts AT&T syntax to Intel if * needed, making the dialect switching useless (it isn't even supported). * * Note: Comments are written in the inline assembly itself. */ #ifdef __clang__ # define XXH_I_ATT(intel, att) att "\n\t" #else # define XXH_I_ATT(intel, att) "{" att "|" intel "}\n\t" #endif /*! @endcond */ /*! * @private * @brief Runs CPUID. * * @param eax , ecx The parameters to pass to CPUID, %eax and %ecx respectively. * @param abcd The array to store the result in, `{ eax, ebx, ecx, edx }` */ static void XXH_cpuid(xxh_u32 eax, xxh_u32 ecx, xxh_u32* abcd) { #if defined(_MSC_VER) __cpuidex((int*)abcd, eax, ecx); #else xxh_u32 ebx, edx; # if defined(__i386__) && defined(__PIC__) __asm__( "# Call CPUID\n\t" "#\n\t" "# On 32-bit x86 with PIC enabled, we are not allowed to overwrite\n\t" "# EBX, so we use EDI instead.\n\t" XXH_I_ATT("mov edi, ebx", "movl %%ebx, %%edi") XXH_I_ATT("cpuid", "cpuid" ) XXH_I_ATT("xchg edi, ebx", "xchgl %%ebx, %%edi") : "=D" (ebx), # else __asm__( "# Call CPUID\n\t" XXH_I_ATT("cpuid", "cpuid") : "=b" (ebx), # endif "+a" (eax), "+c" (ecx), "=d" (edx)); abcd[0] = eax; abcd[1] = ebx; abcd[2] = ecx; abcd[3] = edx; #endif } /* * Modified version of Intel's guide * https://software.intel.com/en-us/articles/how-to-detect-new-instruction-support-in-the-4th-generation-intel-core-processor-family */ #if XXH_DISPATCH_AVX2 || XXH_DISPATCH_AVX512 /*! * @private * @brief Runs `XGETBV`. * * While the CPU may support AVX2, the operating system might not properly save * the full YMM/ZMM registers. * * xgetbv is used for detecting this: Any compliant operating system will define * a set of flags in the xcr0 register indicating how it saves the AVX registers. * * You can manually disable this flag on Windows by running, as admin: * * bcdedit.exe /set xsavedisable 1 * * and rebooting. Run the same command with 0 to re-enable it. */ static xxh_u64 XXH_xgetbv(void) { #if defined(_MSC_VER) return _xgetbv(0); /* min VS2010 SP1 compiler is required */ #else xxh_u32 xcr0_lo, xcr0_hi; __asm__( "# Call XGETBV\n\t" "#\n\t" "# Older assemblers (e.g. macOS's ancient GAS version) don't support\n\t" "# the XGETBV opcode, so we encode it by hand instead.\n\t" "# See for details.\n\t" ".byte 0x0f, 0x01, 0xd0\n\t" : "=a" (xcr0_lo), "=d" (xcr0_hi) : "c" (0)); return xcr0_lo | ((xxh_u64)xcr0_hi << 32); #endif } #endif /*! @cond Doxygen ignores this part */ #define XXH_SSE2_CPUID_MASK (1 << 26) #define XXH_OSXSAVE_CPUID_MASK ((1 << 26) | (1 << 27)) #define XXH_AVX2_CPUID_MASK (1 << 5) #define XXH_AVX2_XGETBV_MASK ((1 << 2) | (1 << 1)) #define XXH_AVX512F_CPUID_MASK (1 << 16) #define XXH_AVX512F_XGETBV_MASK ((7 << 5) | (1 << 2) | (1 << 1)) /*! @endcond */ /*! * @private * @brief Returns the best XXH3 implementation. * * Runs various CPUID/XGETBV tests to try and determine the best implementation. * * @return The best @ref XXH_VECTOR implementation. * @see XXH_VECTOR_TYPES */ int XXH_featureTest(void) { xxh_u32 abcd[4]; xxh_u32 max_leaves; int best = XXH_SCALAR; #if XXH_DISPATCH_AVX2 || XXH_DISPATCH_AVX512 xxh_u64 xgetbv_val; #endif #if defined(__GNUC__) && defined(__i386__) xxh_u32 cpuid_supported; __asm__( "# For the sake of ruthless backwards compatibility, check if CPUID\n\t" "# is supported in the EFLAGS on i386.\n\t" "# This is not necessary on x86_64 - CPUID is mandatory.\n\t" "# The ID flag (bit 21) in the EFLAGS register indicates support\n\t" "# for the CPUID instruction. If a software procedure can set and\n\t" "# clear this flag, the processor executing the procedure supports\n\t" "# the CPUID instruction.\n\t" "# \n\t" "#\n\t" "# Routine is from .\n\t" "# Save EFLAGS\n\t" XXH_I_ATT("pushfd", "pushfl" ) "# Store EFLAGS\n\t" XXH_I_ATT("pushfd", "pushfl" ) "# Invert the ID bit in stored EFLAGS\n\t" XXH_I_ATT("xor dword ptr[esp], 0x200000", "xorl $0x200000, (%%esp)") "# Load stored EFLAGS (with ID bit inverted)\n\t" XXH_I_ATT("popfd", "popfl" ) "# Store EFLAGS again (ID bit may or not be inverted)\n\t" XXH_I_ATT("pushfd", "pushfl" ) "# eax = modified EFLAGS (ID bit may or may not be inverted)\n\t" XXH_I_ATT("pop eax", "popl %%eax" ) "# eax = whichever bits were changed\n\t" XXH_I_ATT("xor eax, dword ptr[esp]", "xorl (%%esp), %%eax" ) "# Restore original EFLAGS\n\t" XXH_I_ATT("popfd", "popfl" ) "# eax = zero if ID bit can't be changed, else non-zero\n\t" XXH_I_ATT("and eax, 0x200000", "andl $0x200000, %%eax" ) : "=a" (cpuid_supported) :: "cc"); if (XXH_unlikely(!cpuid_supported)) { XXH_debugPrint("CPUID support is not detected!"); return best; } #endif /* Check how many CPUID pages we have */ XXH_cpuid(0, 0, abcd); max_leaves = abcd[0]; /* Shouldn't happen on hardware, but happens on some QEMU configs. */ if (XXH_unlikely(max_leaves == 0)) { XXH_debugPrint("Max CPUID leaves == 0!"); return best; } /* Check for SSE2, OSXSAVE and xgetbv */ XXH_cpuid(1, 0, abcd); /* * Test for SSE2. The check is redundant on x86_64, but it doesn't hurt. */ if (XXH_unlikely((abcd[3] & XXH_SSE2_CPUID_MASK) != XXH_SSE2_CPUID_MASK)) return best; XXH_debugPrint("SSE2 support detected."); best = XXH_SSE2; #if XXH_DISPATCH_AVX2 || XXH_DISPATCH_AVX512 /* Make sure we have enough leaves */ if (XXH_unlikely(max_leaves < 7)) return best; /* Test for OSXSAVE and XGETBV */ if ((abcd[2] & XXH_OSXSAVE_CPUID_MASK) != XXH_OSXSAVE_CPUID_MASK) return best; /* CPUID check for AVX features */ XXH_cpuid(7, 0, abcd); xgetbv_val = XXH_xgetbv(); #if XXH_DISPATCH_AVX2 /* Validate that AVX2 is supported by the CPU */ if ((abcd[1] & XXH_AVX2_CPUID_MASK) != XXH_AVX2_CPUID_MASK) return best; /* Validate that the OS supports YMM registers */ if ((xgetbv_val & XXH_AVX2_XGETBV_MASK) != XXH_AVX2_XGETBV_MASK) { XXH_debugPrint("AVX2 supported by the CPU, but not the OS."); return best; } /* AVX2 supported */ XXH_debugPrint("AVX2 support detected."); best = XXH_AVX2; #endif #if XXH_DISPATCH_AVX512 /* Check if AVX512F is supported by the CPU */ if ((abcd[1] & XXH_AVX512F_CPUID_MASK) != XXH_AVX512F_CPUID_MASK) { XXH_debugPrint("AVX512F not supported by CPU"); return best; } /* Validate that the OS supports ZMM registers */ if ((xgetbv_val & XXH_AVX512F_XGETBV_MASK) != XXH_AVX512F_XGETBV_MASK) { XXH_debugPrint("AVX512F supported by the CPU, but not the OS."); return best; } /* AVX512F supported */ XXH_debugPrint("AVX512F support detected."); best = XXH_AVX512; #endif #endif return best; } /* === Vector implementations === */ /*! @cond PRIVATE */ /*! * @private * @brief Defines the various dispatch functions. * * TODO: Consolidate? * * @param suffix The suffix for the functions, e.g. sse2 or scalar * @param target XXH_TARGET_* or empty. */ #define XXH_DEFINE_DISPATCH_FUNCS(suffix, target) \ \ /* === XXH3, default variants === */ \ \ XXH_NO_INLINE target XXH64_hash_t \ XXHL64_default_##suffix(XXH_NOESCAPE const void* XXH_RESTRICT input, \ size_t len) \ { \ return XXH3_hashLong_64b_internal( \ input, len, XXH3_kSecret, sizeof(XXH3_kSecret), \ XXH3_accumulate_##suffix, XXH3_scrambleAcc_##suffix \ ); \ } \ \ /* === XXH3, Seeded variants === */ \ \ XXH_NO_INLINE target XXH64_hash_t \ XXHL64_seed_##suffix(XXH_NOESCAPE const void* XXH_RESTRICT input, size_t len, \ XXH64_hash_t seed) \ { \ return XXH3_hashLong_64b_withSeed_internal( \ input, len, seed, XXH3_accumulate_##suffix, \ XXH3_scrambleAcc_##suffix, XXH3_initCustomSecret_##suffix \ ); \ } \ \ /* === XXH3, Secret variants === */ \ \ XXH_NO_INLINE target XXH64_hash_t \ XXHL64_secret_##suffix(XXH_NOESCAPE const void* XXH_RESTRICT input, \ size_t len, XXH_NOESCAPE const void* secret, \ size_t secretLen) \ { \ return XXH3_hashLong_64b_internal( \ input, len, secret, secretLen, \ XXH3_accumulate_##suffix, XXH3_scrambleAcc_##suffix \ ); \ } \ \ /* === XXH3 update variants === */ \ \ XXH_NO_INLINE target XXH_errorcode \ XXH3_update_##suffix(XXH_NOESCAPE XXH3_state_t* state, \ XXH_NOESCAPE const void* input, size_t len) \ { \ return XXH3_update(state, (const xxh_u8*)input, len, \ XXH3_accumulate_##suffix, XXH3_scrambleAcc_##suffix); \ } \ \ /* === XXH128 default variants === */ \ \ XXH_NO_INLINE target XXH128_hash_t \ XXHL128_default_##suffix(XXH_NOESCAPE const void* XXH_RESTRICT input, \ size_t len) \ { \ return XXH3_hashLong_128b_internal( \ input, len, XXH3_kSecret, sizeof(XXH3_kSecret), \ XXH3_accumulate_##suffix, XXH3_scrambleAcc_##suffix \ ); \ } \ \ /* === XXH128 Secret variants === */ \ \ XXH_NO_INLINE target XXH128_hash_t \ XXHL128_secret_##suffix(XXH_NOESCAPE const void* XXH_RESTRICT input, \ size_t len, \ XXH_NOESCAPE const void* XXH_RESTRICT secret, \ size_t secretLen) \ { \ return XXH3_hashLong_128b_internal( \ input, len, (const xxh_u8*)secret, secretLen, \ XXH3_accumulate_##suffix, XXH3_scrambleAcc_##suffix); \ } \ \ /* === XXH128 Seeded variants === */ \ \ XXH_NO_INLINE target XXH128_hash_t \ XXHL128_seed_##suffix(XXH_NOESCAPE const void* XXH_RESTRICT input, size_t len,\ XXH64_hash_t seed) \ { \ return XXH3_hashLong_128b_withSeed_internal(input, len, seed, \ XXH3_accumulate_##suffix, XXH3_scrambleAcc_##suffix, \ XXH3_initCustomSecret_##suffix); \ } /*! @endcond */ /* End XXH_DEFINE_DISPATCH_FUNCS */ /*! @cond Doxygen ignores this part */ #if XXH_DISPATCH_SCALAR XXH_DEFINE_DISPATCH_FUNCS(scalar, /* nothing */) #endif XXH_DEFINE_DISPATCH_FUNCS(sse2, XXH_TARGET_SSE2) #if XXH_DISPATCH_AVX2 XXH_DEFINE_DISPATCH_FUNCS(avx2, XXH_TARGET_AVX2) #endif #if XXH_DISPATCH_AVX512 XXH_DEFINE_DISPATCH_FUNCS(avx512, XXH_TARGET_AVX512) #endif #undef XXH_DEFINE_DISPATCH_FUNCS /*! @endcond */ /* ==== Dispatchers ==== */ /*! @cond Doxygen ignores this part */ typedef XXH64_hash_t (*XXH3_dispatchx86_hashLong64_default)(XXH_NOESCAPE const void* XXH_RESTRICT, size_t); typedef XXH64_hash_t (*XXH3_dispatchx86_hashLong64_withSeed)(XXH_NOESCAPE const void* XXH_RESTRICT, size_t, XXH64_hash_t); typedef XXH64_hash_t (*XXH3_dispatchx86_hashLong64_withSecret)(XXH_NOESCAPE const void* XXH_RESTRICT, size_t, XXH_NOESCAPE const void* XXH_RESTRICT, size_t); typedef XXH_errorcode (*XXH3_dispatchx86_update)(XXH_NOESCAPE XXH3_state_t*, XXH_NOESCAPE const void*, size_t); typedef struct { XXH3_dispatchx86_hashLong64_default hashLong64_default; XXH3_dispatchx86_hashLong64_withSeed hashLong64_seed; XXH3_dispatchx86_hashLong64_withSecret hashLong64_secret; XXH3_dispatchx86_update update; } XXH_dispatchFunctions_s; #define XXH_NB_DISPATCHES 4 /*! @endcond */ /*! * @private * @brief Table of dispatchers for @ref XXH3_64bits(). * * @pre The indices must match @ref XXH_VECTOR_TYPE. */ static const XXH_dispatchFunctions_s XXH_kDispatch[XXH_NB_DISPATCHES] = { #if XXH_DISPATCH_SCALAR /* Scalar */ { XXHL64_default_scalar, XXHL64_seed_scalar, XXHL64_secret_scalar, XXH3_update_scalar }, #else /* Scalar */ { NULL, NULL, NULL, NULL }, #endif /* SSE2 */ { XXHL64_default_sse2, XXHL64_seed_sse2, XXHL64_secret_sse2, XXH3_update_sse2 }, #if XXH_DISPATCH_AVX2 /* AVX2 */ { XXHL64_default_avx2, XXHL64_seed_avx2, XXHL64_secret_avx2, XXH3_update_avx2 }, #else /* AVX2 */ { NULL, NULL, NULL, NULL }, #endif #if XXH_DISPATCH_AVX512 /* AVX512 */ { XXHL64_default_avx512, XXHL64_seed_avx512, XXHL64_secret_avx512, XXH3_update_avx512 } #else /* AVX512 */ { NULL, NULL, NULL, NULL } #endif }; /*! * @private * @brief The selected dispatch table for @ref XXH3_64bits(). */ static XXH_dispatchFunctions_s XXH_g_dispatch = { NULL, NULL, NULL, NULL }; /*! @cond Doxygen ignores this part */ typedef XXH128_hash_t (*XXH3_dispatchx86_hashLong128_default)(XXH_NOESCAPE const void* XXH_RESTRICT, size_t); typedef XXH128_hash_t (*XXH3_dispatchx86_hashLong128_withSeed)(XXH_NOESCAPE const void* XXH_RESTRICT, size_t, XXH64_hash_t); typedef XXH128_hash_t (*XXH3_dispatchx86_hashLong128_withSecret)(XXH_NOESCAPE const void* XXH_RESTRICT, size_t, const void* XXH_RESTRICT, size_t); typedef struct { XXH3_dispatchx86_hashLong128_default hashLong128_default; XXH3_dispatchx86_hashLong128_withSeed hashLong128_seed; XXH3_dispatchx86_hashLong128_withSecret hashLong128_secret; XXH3_dispatchx86_update update; } XXH_dispatch128Functions_s; /*! @endcond */ /*! * @private * @brief Table of dispatchers for @ref XXH3_128bits(). * * @pre The indices must match @ref XXH_VECTOR_TYPE. */ static const XXH_dispatch128Functions_s XXH_kDispatch128[XXH_NB_DISPATCHES] = { #if XXH_DISPATCH_SCALAR /* Scalar */ { XXHL128_default_scalar, XXHL128_seed_scalar, XXHL128_secret_scalar, XXH3_update_scalar }, #else /* Scalar */ { NULL, NULL, NULL, NULL }, #endif /* SSE2 */ { XXHL128_default_sse2, XXHL128_seed_sse2, XXHL128_secret_sse2, XXH3_update_sse2 }, #if XXH_DISPATCH_AVX2 /* AVX2 */ { XXHL128_default_avx2, XXHL128_seed_avx2, XXHL128_secret_avx2, XXH3_update_avx2 }, #else /* AVX2 */ { NULL, NULL, NULL, NULL }, #endif #if XXH_DISPATCH_AVX512 /* AVX512 */ { XXHL128_default_avx512, XXHL128_seed_avx512, XXHL128_secret_avx512, XXH3_update_avx512 } #else /* AVX512 */ { NULL, NULL, NULL, NULL } #endif }; /*! * @private * @brief The selected dispatch table for @ref XXH3_64bits(). */ static XXH_dispatch128Functions_s XXH_g_dispatch128 = { NULL, NULL, NULL, NULL }; /*! * @private * @brief Runs a CPUID check and sets the correct dispatch tables. */ static XXH_CONSTRUCTOR void XXH_setDispatch(void) { int vecID = XXH_featureTest(); XXH_STATIC_ASSERT(XXH_AVX512 == XXH_NB_DISPATCHES-1); assert(XXH_SCALAR <= vecID && vecID <= XXH_AVX512); #if !XXH_DISPATCH_SCALAR assert(vecID != XXH_SCALAR); #endif #if !XXH_DISPATCH_AVX512 assert(vecID != XXH_AVX512); #endif #if !XXH_DISPATCH_AVX2 assert(vecID != XXH_AVX2); #endif XXH_g_dispatch = XXH_kDispatch[vecID]; XXH_g_dispatch128 = XXH_kDispatch128[vecID]; } /* ==== XXH3 public functions ==== */ /*! @cond Doxygen ignores this part */ static XXH64_hash_t XXH3_hashLong_64b_defaultSecret_selection(const void* XXH_RESTRICT input, size_t len, XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) { (void)seed64; (void)secret; (void)secretLen; if (XXH_DISPATCH_MAYBE_NULL && XXH_g_dispatch.hashLong64_default == NULL) XXH_setDispatch(); return XXH_g_dispatch.hashLong64_default(input, len); } XXH64_hash_t XXH3_64bits_dispatch(XXH_NOESCAPE const void* input, size_t len) { return XXH3_64bits_internal(input, len, 0, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_64b_defaultSecret_selection); } static XXH64_hash_t XXH3_hashLong_64b_withSeed_selection(const void* XXH_RESTRICT input, size_t len, XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) { (void)secret; (void)secretLen; if (XXH_DISPATCH_MAYBE_NULL && XXH_g_dispatch.hashLong64_seed == NULL) XXH_setDispatch(); return XXH_g_dispatch.hashLong64_seed(input, len, seed64); } XXH64_hash_t XXH3_64bits_withSeed_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) { return XXH3_64bits_internal(input, len, seed, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_64b_withSeed_selection); } static XXH64_hash_t XXH3_hashLong_64b_withSecret_selection(const void* XXH_RESTRICT input, size_t len, XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) { (void)seed64; if (XXH_DISPATCH_MAYBE_NULL && XXH_g_dispatch.hashLong64_secret == NULL) XXH_setDispatch(); return XXH_g_dispatch.hashLong64_secret(input, len, secret, secretLen); } XXH64_hash_t XXH3_64bits_withSecret_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretLen) { return XXH3_64bits_internal(input, len, 0, secret, secretLen, XXH3_hashLong_64b_withSecret_selection); } XXH_errorcode XXH3_64bits_update_dispatch(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len) { if (XXH_DISPATCH_MAYBE_NULL && XXH_g_dispatch.update == NULL) XXH_setDispatch(); return XXH_g_dispatch.update(state, (const xxh_u8*)input, len); } /*! @endcond */ /* ==== XXH128 public functions ==== */ /*! @cond Doxygen ignores this part */ static XXH128_hash_t XXH3_hashLong_128b_defaultSecret_selection(const void* input, size_t len, XXH64_hash_t seed64, const void* secret, size_t secretLen) { (void)seed64; (void)secret; (void)secretLen; if (XXH_DISPATCH_MAYBE_NULL && XXH_g_dispatch128.hashLong128_default == NULL) XXH_setDispatch(); return XXH_g_dispatch128.hashLong128_default(input, len); } XXH128_hash_t XXH3_128bits_dispatch(XXH_NOESCAPE const void* input, size_t len) { return XXH3_128bits_internal(input, len, 0, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_128b_defaultSecret_selection); } static XXH128_hash_t XXH3_hashLong_128b_withSeed_selection(const void* input, size_t len, XXH64_hash_t seed64, const void* secret, size_t secretLen) { (void)secret; (void)secretLen; if (XXH_DISPATCH_MAYBE_NULL && XXH_g_dispatch128.hashLong128_seed == NULL) XXH_setDispatch(); return XXH_g_dispatch128.hashLong128_seed(input, len, seed64); } XXH128_hash_t XXH3_128bits_withSeed_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) { return XXH3_128bits_internal(input, len, seed, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_128b_withSeed_selection); } static XXH128_hash_t XXH3_hashLong_128b_withSecret_selection(const void* input, size_t len, XXH64_hash_t seed64, const void* secret, size_t secretLen) { (void)seed64; if (XXH_DISPATCH_MAYBE_NULL && XXH_g_dispatch128.hashLong128_secret == NULL) XXH_setDispatch(); return XXH_g_dispatch128.hashLong128_secret(input, len, secret, secretLen); } XXH128_hash_t XXH3_128bits_withSecret_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretLen) { return XXH3_128bits_internal(input, len, 0, secret, secretLen, XXH3_hashLong_128b_withSecret_selection); } XXH_errorcode XXH3_128bits_update_dispatch(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len) { if (XXH_DISPATCH_MAYBE_NULL && XXH_g_dispatch128.update == NULL) XXH_setDispatch(); return XXH_g_dispatch128.update(state, (const xxh_u8*)input, len); } /*! @endcond */ #if defined (__cplusplus) } #endif /*! @} */ Crypt-xxHash-0.09/ext/xxHash/xxh_x86dispatch.h000444001750001750 773215146654120 23051 0ustar00dchernenkodchernenko000000000000/* * xxHash - XXH3 Dispatcher for x86-based targets * Copyright (C) 2020-2024 Yann Collet * * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions are * met: * * * Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * Redistributions in binary form must reproduce the above * copyright notice, this list of conditions and the following disclaimer * in the documentation and/or other materials provided with the * distribution. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * You can contact the author at: * - xxHash homepage: https://www.xxhash.com * - xxHash source repository: https://github.com/Cyan4973/xxHash */ #ifndef XXH_X86DISPATCH_H_13563687684 #define XXH_X86DISPATCH_H_13563687684 #include "xxhash.h" /* XXH64_hash_t, XXH3_state_t */ #if defined (__cplusplus) extern "C" { #endif /*! * @brief Returns the best XXH3 implementation for x86 * * @return The best @ref XXH_VECTOR implementation. * @see XXH_VECTOR_TYPES */ XXH_PUBLIC_API int XXH_featureTest(void); XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_dispatch(XXH_NOESCAPE const void* input, size_t len); XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_withSeed_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed); XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_withSecret_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretLen); XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update_dispatch(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len); XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_dispatch(XXH_NOESCAPE const void* input, size_t len); XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_withSeed_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed); XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_withSecret_dispatch(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretLen); XXH_PUBLIC_API XXH_errorcode XXH3_128bits_update_dispatch(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len); #if defined (__cplusplus) } #endif /* automatic replacement of XXH3 functions. * can be disabled by setting XXH_DISPATCH_DISABLE_REPLACE */ #ifndef XXH_DISPATCH_DISABLE_REPLACE # undef XXH3_64bits # define XXH3_64bits XXH3_64bits_dispatch # undef XXH3_64bits_withSeed # define XXH3_64bits_withSeed XXH3_64bits_withSeed_dispatch # undef XXH3_64bits_withSecret # define XXH3_64bits_withSecret XXH3_64bits_withSecret_dispatch # undef XXH3_64bits_update # define XXH3_64bits_update XXH3_64bits_update_dispatch # undef XXH128 # define XXH128 XXH3_128bits_withSeed_dispatch # undef XXH3_128bits # define XXH3_128bits XXH3_128bits_dispatch # undef XXH3_128bits_withSeed # define XXH3_128bits_withSeed XXH3_128bits_withSeed_dispatch # undef XXH3_128bits_withSecret # define XXH3_128bits_withSecret XXH3_128bits_withSecret_dispatch # undef XXH3_128bits_update # define XXH3_128bits_update XXH3_128bits_update_dispatch #endif /* XXH_DISPATCH_DISABLE_REPLACE */ #endif /* XXH_X86DISPATCH_H_13563687684 */ Crypt-xxHash-0.09/ext/xxHash/xxhash.c000444001750001750 347715146654120 21315 0ustar00dchernenkodchernenko000000000000/* * xxHash - Extremely Fast Hash algorithm * Copyright (C) 2012-2023 Yann Collet * * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions are * met: * * * Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * Redistributions in binary form must reproduce the above * copyright notice, this list of conditions and the following disclaimer * in the documentation and/or other materials provided with the * distribution. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * You can contact the author at: * - xxHash homepage: https://www.xxhash.com * - xxHash source repository: https://github.com/Cyan4973/xxHash */ /* * xxhash.c instantiates functions defined in xxhash.h */ #define XXH_STATIC_LINKING_ONLY /* access advanced declarations */ #define XXH_IMPLEMENTATION /* access definitions */ #include "xxhash.h" Crypt-xxHash-0.09/ext/xxHash/xxhash.h000444001750001750 101403715146654120 21375 0ustar00dchernenkodchernenko000000000000/* * xxHash - Extremely Fast Hash algorithm * Header File * Copyright (C) 2012-2023 Yann Collet * * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) * * Redistribution and use in source and binary forms, with or without * modification, are permitted provided that the following conditions are * met: * * * Redistributions of source code must retain the above copyright * notice, this list of conditions and the following disclaimer. * * Redistributions in binary form must reproduce the above * copyright notice, this list of conditions and the following disclaimer * in the documentation and/or other materials provided with the * distribution. * * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. * * You can contact the author at: * - xxHash homepage: https://www.xxhash.com * - xxHash source repository: https://github.com/Cyan4973/xxHash */ /*! * @mainpage xxHash * * xxHash is an extremely fast non-cryptographic hash algorithm, working at RAM speed * limits. * * It is proposed in four flavors, in three families: * 1. @ref XXH32_family * - Classic 32-bit hash function. Simple, compact, and runs on almost all * 32-bit and 64-bit systems. * 2. @ref XXH64_family * - Classic 64-bit adaptation of XXH32. Just as simple, and runs well on most * 64-bit systems (but _not_ 32-bit systems). * 3. @ref XXH3_family * - Modern 64-bit and 128-bit hash function family which features improved * strength and performance across the board, especially on smaller data. * It benefits greatly from SIMD and 64-bit without requiring it. * * Benchmarks * --- * The reference system uses an Intel i7-9700K CPU, and runs Ubuntu x64 20.04. * The open source benchmark program is compiled with clang v10.0 using -O3 flag. * * | Hash Name | ISA ext | Width | Large Data Speed | Small Data Velocity | * | -------------------- | ------- | ----: | ---------------: | ------------------: | * | XXH3_64bits() | @b AVX2 | 64 | 59.4 GB/s | 133.1 | * | MeowHash | AES-NI | 128 | 58.2 GB/s | 52.5 | * | XXH3_128bits() | @b AVX2 | 128 | 57.9 GB/s | 118.1 | * | CLHash | PCLMUL | 64 | 37.1 GB/s | 58.1 | * | XXH3_64bits() | @b SSE2 | 64 | 31.5 GB/s | 133.1 | * | XXH3_128bits() | @b SSE2 | 128 | 29.6 GB/s | 118.1 | * | RAM sequential read | | N/A | 28.0 GB/s | N/A | * | ahash | AES-NI | 64 | 22.5 GB/s | 107.2 | * | City64 | | 64 | 22.0 GB/s | 76.6 | * | T1ha2 | | 64 | 22.0 GB/s | 99.0 | * | City128 | | 128 | 21.7 GB/s | 57.7 | * | FarmHash | AES-NI | 64 | 21.3 GB/s | 71.9 | * | XXH64() | | 64 | 19.4 GB/s | 71.0 | * | SpookyHash | | 64 | 19.3 GB/s | 53.2 | * | Mum | | 64 | 18.0 GB/s | 67.0 | * | CRC32C | SSE4.2 | 32 | 13.0 GB/s | 57.9 | * | XXH32() | | 32 | 9.7 GB/s | 71.9 | * | City32 | | 32 | 9.1 GB/s | 66.0 | * | Blake3* | @b AVX2 | 256 | 4.4 GB/s | 8.1 | * | Murmur3 | | 32 | 3.9 GB/s | 56.1 | * | SipHash* | | 64 | 3.0 GB/s | 43.2 | * | Blake3* | @b SSE2 | 256 | 2.4 GB/s | 8.1 | * | HighwayHash | | 64 | 1.4 GB/s | 6.0 | * | FNV64 | | 64 | 1.2 GB/s | 62.7 | * | Blake2* | | 256 | 1.1 GB/s | 5.1 | * | SHA1* | | 160 | 0.8 GB/s | 5.6 | * | MD5* | | 128 | 0.6 GB/s | 7.8 | * @note * - Hashes which require a specific ISA extension are noted. SSE2 is also noted, * even though it is mandatory on x64. * - Hashes with an asterisk are cryptographic. Note that MD5 is non-cryptographic * by modern standards. * - Small data velocity is a rough average of algorithm's efficiency for small * data. For more accurate information, see the wiki. * - More benchmarks and strength tests are found on the wiki: * https://github.com/Cyan4973/xxHash/wiki * * Usage * ------ * All xxHash variants use a similar API. Changing the algorithm is a trivial * substitution. * * @pre * For functions which take an input and length parameter, the following * requirements are assumed: * - The range from [`input`, `input + length`) is valid, readable memory. * - The only exception is if the `length` is `0`, `input` may be `NULL`. * - For C++, the objects must have the *TriviallyCopyable* property, as the * functions access bytes directly as if it was an array of `unsigned char`. * * @anchor single_shot_example * **Single Shot** * * These functions are stateless functions which hash a contiguous block of memory, * immediately returning the result. They are the easiest and usually the fastest * option. * * XXH32(), XXH64(), XXH3_64bits(), XXH3_128bits() * * @code{.c} * #include * #include "xxhash.h" * * // Example for a function which hashes a null terminated string with XXH32(). * XXH32_hash_t hash_string(const char* string, XXH32_hash_t seed) * { * // NULL pointers are only valid if the length is zero * size_t length = (string == NULL) ? 0 : strlen(string); * return XXH32(string, length, seed); * } * @endcode * * * @anchor streaming_example * **Streaming** * * These groups of functions allow incremental hashing of unknown size, even * more than what would fit in a size_t. * * XXH32_reset(), XXH64_reset(), XXH3_64bits_reset(), XXH3_128bits_reset() * * @code{.c} * #include * #include * #include "xxhash.h" * // Example for a function which hashes a FILE incrementally with XXH3_64bits(). * XXH64_hash_t hashFile(FILE* f) * { * // Allocate a state struct. Do not just use malloc() or new. * XXH3_state_t* state = XXH3_createState(); * assert(state != NULL && "Out of memory!"); * // Reset the state to start a new hashing session. * XXH3_64bits_reset(state); * char buffer[4096]; * size_t count; * // Read the file in chunks * while ((count = fread(buffer, 1, sizeof(buffer), f)) != 0) { * // Run update() as many times as necessary to process the data * XXH3_64bits_update(state, buffer, count); * } * // Retrieve the finalized hash. This will not change the state. * XXH64_hash_t result = XXH3_64bits_digest(state); * // Free the state. Do not use free(). * XXH3_freeState(state); * return result; * } * @endcode * * Streaming functions generate the xxHash value from an incremental input. * This method is slower than single-call functions, due to state management. * For small inputs, prefer `XXH32()` and `XXH64()`, which are better optimized. * * An XXH state must first be allocated using `XXH*_createState()`. * * Start a new hash by initializing the state with a seed using `XXH*_reset()`. * * Then, feed the hash state by calling `XXH*_update()` as many times as necessary. * * The function returns an error code, with 0 meaning OK, and any other value * meaning there is an error. * * Finally, a hash value can be produced anytime, by using `XXH*_digest()`. * This function returns the nn-bits hash as an int or long long. * * It's still possible to continue inserting input into the hash state after a * digest, and generate new hash values later on by invoking `XXH*_digest()`. * * When done, release the state using `XXH*_freeState()`. * * * @anchor canonical_representation_example * **Canonical Representation** * * The default return values from XXH functions are unsigned 32, 64 and 128 bit * integers. * This the simplest and fastest format for further post-processing. * * However, this leaves open the question of what is the order on the byte level, * since little and big endian conventions will store the same number differently. * * The canonical representation settles this issue by mandating big-endian * convention, the same convention as human-readable numbers (large digits first). * * When writing hash values to storage, sending them over a network, or printing * them, it's highly recommended to use the canonical representation to ensure * portability across a wider range of systems, present and future. * * The following functions allow transformation of hash values to and from * canonical format. * * XXH32_canonicalFromHash(), XXH32_hashFromCanonical(), * XXH64_canonicalFromHash(), XXH64_hashFromCanonical(), * XXH128_canonicalFromHash(), XXH128_hashFromCanonical(), * * @code{.c} * #include * #include "xxhash.h" * * // Example for a function which prints XXH32_hash_t in human readable format * void printXxh32(XXH32_hash_t hash) * { * XXH32_canonical_t cano; * XXH32_canonicalFromHash(&cano, hash); * size_t i; * for(i = 0; i < sizeof(cano.digest); ++i) { * printf("%02x", cano.digest[i]); * } * printf("\n"); * } * * // Example for a function which converts XXH32_canonical_t to XXH32_hash_t * XXH32_hash_t convertCanonicalToXxh32(XXH32_canonical_t cano) * { * XXH32_hash_t hash = XXH32_hashFromCanonical(&cano); * return hash; * } * @endcode * * * @file xxhash.h * xxHash prototypes and implementation */ #if defined (__cplusplus) extern "C" { #endif /* **************************** * INLINE mode ******************************/ /*! * @defgroup public Public API * Contains details on the public xxHash functions. * @{ */ #ifdef XXH_DOXYGEN /*! * @brief Gives access to internal state declaration, required for static allocation. * * Incompatible with dynamic linking, due to risks of ABI changes. * * Usage: * @code{.c} * #define XXH_STATIC_LINKING_ONLY * #include "xxhash.h" * @endcode */ # define XXH_STATIC_LINKING_ONLY /* Do not undef XXH_STATIC_LINKING_ONLY for Doxygen */ /*! * @brief Gives access to internal definitions. * * Usage: * @code{.c} * #define XXH_STATIC_LINKING_ONLY * #define XXH_IMPLEMENTATION * #include "xxhash.h" * @endcode */ # define XXH_IMPLEMENTATION /* Do not undef XXH_IMPLEMENTATION for Doxygen */ /*! * @brief Exposes the implementation and marks all functions as `inline`. * * Use these build macros to inline xxhash into the target unit. * Inlining improves performance on small inputs, especially when the length is * expressed as a compile-time constant: * * https://fastcompression.blogspot.com/2018/03/xxhash-for-small-keys-impressive-power.html * * It also keeps xxHash symbols private to the unit, so they are not exported. * * Usage: * @code{.c} * #define XXH_INLINE_ALL * #include "xxhash.h" * @endcode * Do not compile and link xxhash.o as a separate object, as it is not useful. */ # define XXH_INLINE_ALL # undef XXH_INLINE_ALL /*! * @brief Exposes the implementation without marking functions as inline. */ # define XXH_PRIVATE_API # undef XXH_PRIVATE_API /*! * @brief Emulate a namespace by transparently prefixing all symbols. * * If you want to include _and expose_ xxHash functions from within your own * library, but also want to avoid symbol collisions with other libraries which * may also include xxHash, you can use @ref XXH_NAMESPACE to automatically prefix * any public symbol from xxhash library with the value of @ref XXH_NAMESPACE * (therefore, avoid empty or numeric values). * * Note that no change is required within the calling program as long as it * includes `xxhash.h`: Regular symbol names will be automatically translated * by this header. */ # define XXH_NAMESPACE /* YOUR NAME HERE */ # undef XXH_NAMESPACE #endif #if (defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API)) \ && !defined(XXH_INLINE_ALL_31684351384) /* this section should be traversed only once */ # define XXH_INLINE_ALL_31684351384 /* give access to the advanced API, required to compile implementations */ # undef XXH_STATIC_LINKING_ONLY /* avoid macro redef */ # define XXH_STATIC_LINKING_ONLY /* make all functions private */ # undef XXH_PUBLIC_API # if defined(__GNUC__) # define XXH_PUBLIC_API static __inline __attribute__((__unused__)) # elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) # define XXH_PUBLIC_API static inline # elif defined(_MSC_VER) # define XXH_PUBLIC_API static __inline # else /* note: this version may generate warnings for unused static functions */ # define XXH_PUBLIC_API static # endif /* * This part deals with the special case where a unit wants to inline xxHash, * but "xxhash.h" has previously been included without XXH_INLINE_ALL, * such as part of some previously included *.h header file. * Without further action, the new include would just be ignored, * and functions would effectively _not_ be inlined (silent failure). * The following macros solve this situation by prefixing all inlined names, * avoiding naming collision with previous inclusions. */ /* Before that, we unconditionally #undef all symbols, * in case they were already defined with XXH_NAMESPACE. * They will then be redefined for XXH_INLINE_ALL */ # undef XXH_versionNumber /* XXH32 */ # undef XXH32 # undef XXH32_createState # undef XXH32_freeState # undef XXH32_reset # undef XXH32_update # undef XXH32_digest # undef XXH32_copyState # undef XXH32_canonicalFromHash # undef XXH32_hashFromCanonical /* XXH64 */ # undef XXH64 # undef XXH64_createState # undef XXH64_freeState # undef XXH64_reset # undef XXH64_update # undef XXH64_digest # undef XXH64_copyState # undef XXH64_canonicalFromHash # undef XXH64_hashFromCanonical /* XXH3_64bits */ # undef XXH3_64bits # undef XXH3_64bits_withSecret # undef XXH3_64bits_withSeed # undef XXH3_64bits_withSecretandSeed # undef XXH3_createState # undef XXH3_freeState # undef XXH3_copyState # undef XXH3_64bits_reset # undef XXH3_64bits_reset_withSeed # undef XXH3_64bits_reset_withSecret # undef XXH3_64bits_update # undef XXH3_64bits_digest # undef XXH3_generateSecret /* XXH3_128bits */ # undef XXH128 # undef XXH3_128bits # undef XXH3_128bits_withSeed # undef XXH3_128bits_withSecret # undef XXH3_128bits_reset # undef XXH3_128bits_reset_withSeed # undef XXH3_128bits_reset_withSecret # undef XXH3_128bits_reset_withSecretandSeed # undef XXH3_128bits_update # undef XXH3_128bits_digest # undef XXH128_isEqual # undef XXH128_cmp # undef XXH128_canonicalFromHash # undef XXH128_hashFromCanonical /* Finally, free the namespace itself */ # undef XXH_NAMESPACE /* employ the namespace for XXH_INLINE_ALL */ # define XXH_NAMESPACE XXH_INLINE_ /* * Some identifiers (enums, type names) are not symbols, * but they must nonetheless be renamed to avoid redeclaration. * Alternative solution: do not redeclare them. * However, this requires some #ifdefs, and has a more dispersed impact. * Meanwhile, renaming can be achieved in a single place. */ # define XXH_IPREF(Id) XXH_NAMESPACE ## Id # define XXH_OK XXH_IPREF(XXH_OK) # define XXH_ERROR XXH_IPREF(XXH_ERROR) # define XXH_errorcode XXH_IPREF(XXH_errorcode) # define XXH32_canonical_t XXH_IPREF(XXH32_canonical_t) # define XXH64_canonical_t XXH_IPREF(XXH64_canonical_t) # define XXH128_canonical_t XXH_IPREF(XXH128_canonical_t) # define XXH32_state_s XXH_IPREF(XXH32_state_s) # define XXH32_state_t XXH_IPREF(XXH32_state_t) # define XXH64_state_s XXH_IPREF(XXH64_state_s) # define XXH64_state_t XXH_IPREF(XXH64_state_t) # define XXH3_state_s XXH_IPREF(XXH3_state_s) # define XXH3_state_t XXH_IPREF(XXH3_state_t) # define XXH128_hash_t XXH_IPREF(XXH128_hash_t) /* Ensure the header is parsed again, even if it was previously included */ # undef XXHASH_H_5627135585666179 # undef XXHASH_H_STATIC_13879238742 #endif /* XXH_INLINE_ALL || XXH_PRIVATE_API */ /* **************************************************************** * Stable API *****************************************************************/ #ifndef XXHASH_H_5627135585666179 #define XXHASH_H_5627135585666179 1 /*! @brief Marks a global symbol. */ #if !defined(XXH_INLINE_ALL) && !defined(XXH_PRIVATE_API) # if defined(_WIN32) && defined(_MSC_VER) && (defined(XXH_IMPORT) || defined(XXH_EXPORT)) # ifdef XXH_EXPORT # define XXH_PUBLIC_API __declspec(dllexport) # elif XXH_IMPORT # define XXH_PUBLIC_API __declspec(dllimport) # endif # else # define XXH_PUBLIC_API /* do nothing */ # endif #endif #ifdef XXH_NAMESPACE # define XXH_CAT(A,B) A##B # define XXH_NAME2(A,B) XXH_CAT(A,B) # define XXH_versionNumber XXH_NAME2(XXH_NAMESPACE, XXH_versionNumber) /* XXH32 */ # define XXH32 XXH_NAME2(XXH_NAMESPACE, XXH32) # define XXH32_createState XXH_NAME2(XXH_NAMESPACE, XXH32_createState) # define XXH32_freeState XXH_NAME2(XXH_NAMESPACE, XXH32_freeState) # define XXH32_reset XXH_NAME2(XXH_NAMESPACE, XXH32_reset) # define XXH32_update XXH_NAME2(XXH_NAMESPACE, XXH32_update) # define XXH32_digest XXH_NAME2(XXH_NAMESPACE, XXH32_digest) # define XXH32_copyState XXH_NAME2(XXH_NAMESPACE, XXH32_copyState) # define XXH32_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH32_canonicalFromHash) # define XXH32_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH32_hashFromCanonical) /* XXH64 */ # define XXH64 XXH_NAME2(XXH_NAMESPACE, XXH64) # define XXH64_createState XXH_NAME2(XXH_NAMESPACE, XXH64_createState) # define XXH64_freeState XXH_NAME2(XXH_NAMESPACE, XXH64_freeState) # define XXH64_reset XXH_NAME2(XXH_NAMESPACE, XXH64_reset) # define XXH64_update XXH_NAME2(XXH_NAMESPACE, XXH64_update) # define XXH64_digest XXH_NAME2(XXH_NAMESPACE, XXH64_digest) # define XXH64_copyState XXH_NAME2(XXH_NAMESPACE, XXH64_copyState) # define XXH64_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH64_canonicalFromHash) # define XXH64_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH64_hashFromCanonical) /* XXH3_64bits */ # define XXH3_64bits XXH_NAME2(XXH_NAMESPACE, XXH3_64bits) # define XXH3_64bits_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSecret) # define XXH3_64bits_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSeed) # define XXH3_64bits_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSecretandSeed) # define XXH3_createState XXH_NAME2(XXH_NAMESPACE, XXH3_createState) # define XXH3_freeState XXH_NAME2(XXH_NAMESPACE, XXH3_freeState) # define XXH3_copyState XXH_NAME2(XXH_NAMESPACE, XXH3_copyState) # define XXH3_64bits_reset XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset) # define XXH3_64bits_reset_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSeed) # define XXH3_64bits_reset_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSecret) # define XXH3_64bits_reset_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSecretandSeed) # define XXH3_64bits_update XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_update) # define XXH3_64bits_digest XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_digest) # define XXH3_generateSecret XXH_NAME2(XXH_NAMESPACE, XXH3_generateSecret) # define XXH3_generateSecret_fromSeed XXH_NAME2(XXH_NAMESPACE, XXH3_generateSecret_fromSeed) /* XXH3_128bits */ # define XXH128 XXH_NAME2(XXH_NAMESPACE, XXH128) # define XXH3_128bits XXH_NAME2(XXH_NAMESPACE, XXH3_128bits) # define XXH3_128bits_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSeed) # define XXH3_128bits_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSecret) # define XXH3_128bits_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSecretandSeed) # define XXH3_128bits_reset XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset) # define XXH3_128bits_reset_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSeed) # define XXH3_128bits_reset_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSecret) # define XXH3_128bits_reset_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSecretandSeed) # define XXH3_128bits_update XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_update) # define XXH3_128bits_digest XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_digest) # define XXH128_isEqual XXH_NAME2(XXH_NAMESPACE, XXH128_isEqual) # define XXH128_cmp XXH_NAME2(XXH_NAMESPACE, XXH128_cmp) # define XXH128_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH128_canonicalFromHash) # define XXH128_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH128_hashFromCanonical) #endif /* ************************************* * Compiler specifics ***************************************/ /* specific declaration modes for Windows */ #if !defined(XXH_INLINE_ALL) && !defined(XXH_PRIVATE_API) # if defined(_WIN32) && defined(_MSC_VER) && (defined(XXH_IMPORT) || defined(XXH_EXPORT)) # ifdef XXH_EXPORT # define XXH_PUBLIC_API __declspec(dllexport) # elif XXH_IMPORT # define XXH_PUBLIC_API __declspec(dllimport) # endif # else # define XXH_PUBLIC_API /* do nothing */ # endif #endif #if defined (__GNUC__) # define XXH_CONSTF __attribute__((__const__)) # define XXH_PUREF __attribute__((__pure__)) # define XXH_MALLOCF __attribute__((__malloc__)) #else # define XXH_CONSTF /* disable */ # define XXH_PUREF # define XXH_MALLOCF #endif /* ************************************* * Version ***************************************/ #define XXH_VERSION_MAJOR 0 #define XXH_VERSION_MINOR 8 #define XXH_VERSION_RELEASE 3 /*! @brief Version number, encoded as two digits each */ #define XXH_VERSION_NUMBER (XXH_VERSION_MAJOR *100*100 + XXH_VERSION_MINOR *100 + XXH_VERSION_RELEASE) /*! * @brief Obtains the xxHash version. * * This is mostly useful when xxHash is compiled as a shared library, * since the returned value comes from the library, as opposed to header file. * * @return @ref XXH_VERSION_NUMBER of the invoked library. */ XXH_PUBLIC_API XXH_CONSTF unsigned XXH_versionNumber (void); /* **************************** * Common basic types ******************************/ #include /* size_t */ /*! * @brief Exit code for the streaming API. */ typedef enum { XXH_OK = 0, /*!< OK */ XXH_ERROR /*!< Error */ } XXH_errorcode; /*-********************************************************************** * 32-bit hash ************************************************************************/ #if defined(XXH_DOXYGEN) /* Don't show include */ /*! * @brief An unsigned 32-bit integer. * * Not necessarily defined to `uint32_t` but functionally equivalent. */ typedef uint32_t XXH32_hash_t; #elif !defined (__VMS) \ && (defined (__cplusplus) \ || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) # ifdef _AIX # include # else # include # endif typedef uint32_t XXH32_hash_t; #else # include # if UINT_MAX == 0xFFFFFFFFUL typedef unsigned int XXH32_hash_t; # elif ULONG_MAX == 0xFFFFFFFFUL typedef unsigned long XXH32_hash_t; # else # error "unsupported platform: need a 32-bit type" # endif #endif /*! * @} * * @defgroup XXH32_family XXH32 family * @ingroup public * Contains functions used in the classic 32-bit xxHash algorithm. * * @note * XXH32 is useful for older platforms, with no or poor 64-bit performance. * Note that the @ref XXH3_family provides competitive speed for both 32-bit * and 64-bit systems, and offers true 64/128 bit hash results. * * @see @ref XXH64_family, @ref XXH3_family : Other xxHash families * @see @ref XXH32_impl for implementation details * @{ */ /*! * @brief Calculates the 32-bit hash of @p input using xxHash32. * * @param input The block of data to be hashed, at least @p length bytes in size. * @param length The length of @p input, in bytes. * @param seed The 32-bit seed to alter the hash's output predictably. * * @pre * The memory between @p input and @p input + @p length must be valid, * readable, contiguous memory. However, if @p length is `0`, @p input may be * `NULL`. In C++, this also must be *TriviallyCopyable*. * * @return The calculated 32-bit xxHash32 value. * * @see @ref single_shot_example "Single Shot Example" for an example. */ XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32 (const void* input, size_t length, XXH32_hash_t seed); #ifndef XXH_NO_STREAM /*! * @typedef struct XXH32_state_s XXH32_state_t * @brief The opaque state struct for the XXH32 streaming API. * * @see XXH32_state_s for details. * @see @ref streaming_example "Streaming Example" */ typedef struct XXH32_state_s XXH32_state_t; /*! * @brief Allocates an @ref XXH32_state_t. * * @return An allocated pointer of @ref XXH32_state_t on success. * @return `NULL` on failure. * * @note Must be freed with XXH32_freeState(). * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_MALLOCF XXH32_state_t* XXH32_createState(void); /*! * @brief Frees an @ref XXH32_state_t. * * @param statePtr A pointer to an @ref XXH32_state_t allocated with @ref XXH32_createState(). * * @return @ref XXH_OK. * * @note @p statePtr must be allocated with XXH32_createState(). * * @see @ref streaming_example "Streaming Example" * */ XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr); /*! * @brief Copies one @ref XXH32_state_t to another. * * @param dst_state The state to copy to. * @param src_state The state to copy from. * @pre * @p dst_state and @p src_state must not be `NULL` and must not overlap. */ XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dst_state, const XXH32_state_t* src_state); /*! * @brief Resets an @ref XXH32_state_t to begin a new hash. * * @param statePtr The state struct to reset. * @param seed The 32-bit seed to alter the hash result predictably. * * @pre * @p statePtr must not be `NULL`. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note This function resets and seeds a state. Call it before @ref XXH32_update(). * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH32_reset (XXH32_state_t* statePtr, XXH32_hash_t seed); /*! * @brief Consumes a block of @p input to an @ref XXH32_state_t. * * @param statePtr The state struct to update. * @param input The block of data to be hashed, at least @p length bytes in size. * @param length The length of @p input, in bytes. * * @pre * @p statePtr must not be `NULL`. * @pre * The memory between @p input and @p input + @p length must be valid, * readable, contiguous memory. However, if @p length is `0`, @p input may be * `NULL`. In C++, this also must be *TriviallyCopyable*. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note Call this to incrementally consume blocks of data. * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* statePtr, const void* input, size_t length); /*! * @brief Returns the calculated hash value from an @ref XXH32_state_t. * * @param statePtr The state struct to calculate the hash from. * * @pre * @p statePtr must not be `NULL`. * * @return The calculated 32-bit xxHash32 value from that state. * * @note * Calling XXH32_digest() will not affect @p statePtr, so you can update, * digest, and update again. * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32_digest (const XXH32_state_t* statePtr); #endif /* !XXH_NO_STREAM */ /******* Canonical representation *******/ /*! * @brief Canonical (big endian) representation of @ref XXH32_hash_t. */ typedef struct { unsigned char digest[4]; /*!< Hash bytes, big endian */ } XXH32_canonical_t; /*! * @brief Converts an @ref XXH32_hash_t to a big endian @ref XXH32_canonical_t. * * @param dst The @ref XXH32_canonical_t pointer to be stored to. * @param hash The @ref XXH32_hash_t to be converted. * * @pre * @p dst must not be `NULL`. * * @see @ref canonical_representation_example "Canonical Representation Example" */ XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash); /*! * @brief Converts an @ref XXH32_canonical_t to a native @ref XXH32_hash_t. * * @param src The @ref XXH32_canonical_t to convert. * * @pre * @p src must not be `NULL`. * * @return The converted hash. * * @see @ref canonical_representation_example "Canonical Representation Example" */ XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src); /*! @cond Doxygen ignores this part */ #ifdef __has_attribute # define XXH_HAS_ATTRIBUTE(x) __has_attribute(x) #else # define XXH_HAS_ATTRIBUTE(x) 0 #endif /*! @endcond */ /*! @cond Doxygen ignores this part */ /* * C23 __STDC_VERSION__ number hasn't been specified yet. For now * leave as `201711L` (C17 + 1). * TODO: Update to correct value when its been specified. */ #define XXH_C23_VN 201711L /*! @endcond */ /*! @cond Doxygen ignores this part */ /* C-language Attributes are added in C23. */ #if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= XXH_C23_VN) && defined(__has_c_attribute) # define XXH_HAS_C_ATTRIBUTE(x) __has_c_attribute(x) #else # define XXH_HAS_C_ATTRIBUTE(x) 0 #endif /*! @endcond */ /*! @cond Doxygen ignores this part */ #if defined(__cplusplus) && defined(__has_cpp_attribute) # define XXH_HAS_CPP_ATTRIBUTE(x) __has_cpp_attribute(x) #else # define XXH_HAS_CPP_ATTRIBUTE(x) 0 #endif /*! @endcond */ /*! @cond Doxygen ignores this part */ /* * Define XXH_FALLTHROUGH macro for annotating switch case with the 'fallthrough' attribute * introduced in CPP17 and C23. * CPP17 : https://en.cppreference.com/w/cpp/language/attributes/fallthrough * C23 : https://en.cppreference.com/w/c/language/attributes/fallthrough */ #if XXH_HAS_C_ATTRIBUTE(fallthrough) || XXH_HAS_CPP_ATTRIBUTE(fallthrough) # define XXH_FALLTHROUGH [[fallthrough]] #elif XXH_HAS_ATTRIBUTE(__fallthrough__) # define XXH_FALLTHROUGH __attribute__ ((__fallthrough__)) #else # define XXH_FALLTHROUGH /* fallthrough */ #endif /*! @endcond */ /*! @cond Doxygen ignores this part */ /* * Define XXH_NOESCAPE for annotated pointers in public API. * https://clang.llvm.org/docs/AttributeReference.html#noescape * As of writing this, only supported by clang. */ #if XXH_HAS_ATTRIBUTE(noescape) # define XXH_NOESCAPE __attribute__((__noescape__)) #else # define XXH_NOESCAPE #endif /*! @endcond */ /*! * @} * @ingroup public * @{ */ #ifndef XXH_NO_LONG_LONG /*-********************************************************************** * 64-bit hash ************************************************************************/ #if defined(XXH_DOXYGEN) /* don't include */ /*! * @brief An unsigned 64-bit integer. * * Not necessarily defined to `uint64_t` but functionally equivalent. */ typedef uint64_t XXH64_hash_t; #elif !defined (__VMS) \ && (defined (__cplusplus) \ || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) # ifdef _AIX # include # else # include # endif typedef uint64_t XXH64_hash_t; #else # include # if defined(__LP64__) && ULONG_MAX == 0xFFFFFFFFFFFFFFFFULL /* LP64 ABI says uint64_t is unsigned long */ typedef unsigned long XXH64_hash_t; # else /* the following type must have a width of 64-bit */ typedef unsigned long long XXH64_hash_t; # endif #endif /*! * @} * * @defgroup XXH64_family XXH64 family * @ingroup public * @{ * Contains functions used in the classic 64-bit xxHash algorithm. * * @note * XXH3 provides competitive speed for both 32-bit and 64-bit systems, * and offers true 64/128 bit hash results. * It provides better speed for systems with vector processing capabilities. */ /*! * @brief Calculates the 64-bit hash of @p input using xxHash64. * * @param input The block of data to be hashed, at least @p length bytes in size. * @param length The length of @p input, in bytes. * @param seed The 64-bit seed to alter the hash's output predictably. * * @pre * The memory between @p input and @p input + @p length must be valid, * readable, contiguous memory. However, if @p length is `0`, @p input may be * `NULL`. In C++, this also must be *TriviallyCopyable*. * * @return The calculated 64-bit xxHash64 value. * * @see @ref single_shot_example "Single Shot Example" for an example. */ XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed); /******* Streaming *******/ #ifndef XXH_NO_STREAM /*! * @brief The opaque state struct for the XXH64 streaming API. * * @see XXH64_state_s for details. * @see @ref streaming_example "Streaming Example" */ typedef struct XXH64_state_s XXH64_state_t; /* incomplete type */ /*! * @brief Allocates an @ref XXH64_state_t. * * @return An allocated pointer of @ref XXH64_state_t on success. * @return `NULL` on failure. * * @note Must be freed with XXH64_freeState(). * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_MALLOCF XXH64_state_t* XXH64_createState(void); /*! * @brief Frees an @ref XXH64_state_t. * * @param statePtr A pointer to an @ref XXH64_state_t allocated with @ref XXH64_createState(). * * @return @ref XXH_OK. * * @note @p statePtr must be allocated with XXH64_createState(). * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr); /*! * @brief Copies one @ref XXH64_state_t to another. * * @param dst_state The state to copy to. * @param src_state The state to copy from. * @pre * @p dst_state and @p src_state must not be `NULL` and must not overlap. */ XXH_PUBLIC_API void XXH64_copyState(XXH_NOESCAPE XXH64_state_t* dst_state, const XXH64_state_t* src_state); /*! * @brief Resets an @ref XXH64_state_t to begin a new hash. * * @param statePtr The state struct to reset. * @param seed The 64-bit seed to alter the hash result predictably. * * @pre * @p statePtr must not be `NULL`. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note This function resets and seeds a state. Call it before @ref XXH64_update(). * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH64_reset (XXH_NOESCAPE XXH64_state_t* statePtr, XXH64_hash_t seed); /*! * @brief Consumes a block of @p input to an @ref XXH64_state_t. * * @param statePtr The state struct to update. * @param input The block of data to be hashed, at least @p length bytes in size. * @param length The length of @p input, in bytes. * * @pre * @p statePtr must not be `NULL`. * @pre * The memory between @p input and @p input + @p length must be valid, * readable, contiguous memory. However, if @p length is `0`, @p input may be * `NULL`. In C++, this also must be *TriviallyCopyable*. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note Call this to incrementally consume blocks of data. * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH_NOESCAPE XXH64_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); /*! * @brief Returns the calculated hash value from an @ref XXH64_state_t. * * @param statePtr The state struct to calculate the hash from. * * @pre * @p statePtr must not be `NULL`. * * @return The calculated 64-bit xxHash64 value from that state. * * @note * Calling XXH64_digest() will not affect @p statePtr, so you can update, * digest, and update again. * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64_digest (XXH_NOESCAPE const XXH64_state_t* statePtr); #endif /* !XXH_NO_STREAM */ /******* Canonical representation *******/ /*! * @brief Canonical (big endian) representation of @ref XXH64_hash_t. */ typedef struct { unsigned char digest[sizeof(XXH64_hash_t)]; } XXH64_canonical_t; /*! * @brief Converts an @ref XXH64_hash_t to a big endian @ref XXH64_canonical_t. * * @param dst The @ref XXH64_canonical_t pointer to be stored to. * @param hash The @ref XXH64_hash_t to be converted. * * @pre * @p dst must not be `NULL`. * * @see @ref canonical_representation_example "Canonical Representation Example" */ XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH_NOESCAPE XXH64_canonical_t* dst, XXH64_hash_t hash); /*! * @brief Converts an @ref XXH64_canonical_t to a native @ref XXH64_hash_t. * * @param src The @ref XXH64_canonical_t to convert. * * @pre * @p src must not be `NULL`. * * @return The converted hash. * * @see @ref canonical_representation_example "Canonical Representation Example" */ XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64_hashFromCanonical(XXH_NOESCAPE const XXH64_canonical_t* src); #ifndef XXH_NO_XXH3 /*! * @} * ************************************************************************ * @defgroup XXH3_family XXH3 family * @ingroup public * @{ * * XXH3 is a more recent hash algorithm featuring: * - Improved speed for both small and large inputs * - True 64-bit and 128-bit outputs * - SIMD acceleration * - Improved 32-bit viability * * Speed analysis methodology is explained here: * * https://fastcompression.blogspot.com/2019/03/presenting-xxh3.html * * Compared to XXH64, expect XXH3 to run approximately * ~2x faster on large inputs and >3x faster on small ones, * exact differences vary depending on platform. * * XXH3's speed benefits greatly from SIMD and 64-bit arithmetic, * but does not require it. * Most 32-bit and 64-bit targets that can run XXH32 smoothly can run XXH3 * at competitive speeds, even without vector support. Further details are * explained in the implementation. * * XXH3 has a fast scalar implementation, but it also includes accelerated SIMD * implementations for many common platforms: * - AVX512 * - AVX2 * - SSE2 * - ARM NEON * - WebAssembly SIMD128 * - POWER8 VSX * - s390x ZVector * This can be controlled via the @ref XXH_VECTOR macro, but it automatically * selects the best version according to predefined macros. For the x86 family, an * automatic runtime dispatcher is included separately in @ref xxh_x86dispatch.c. * * XXH3 implementation is portable: * it has a generic C90 formulation that can be compiled on any platform, * all implementations generate exactly the same hash value on all platforms. * Starting from v0.8.0, it's also labelled "stable", meaning that * any future version will also generate the same hash value. * * XXH3 offers 2 variants, _64bits and _128bits. * * When only 64 bits are needed, prefer invoking the _64bits variant, as it * reduces the amount of mixing, resulting in faster speed on small inputs. * It's also generally simpler to manipulate a scalar return type than a struct. * * The API supports one-shot hashing, streaming mode, and custom secrets. */ /*! * @ingroup tuning * @brief Possible values for @ref XXH_VECTOR. * * Unless set explicitly, determined automatically. */ # define XXH_SCALAR 0 /*!< Portable scalar version */ # define XXH_SSE2 1 /*!< SSE2 for Pentium 4, Opteron, all x86_64. */ # define XXH_AVX2 2 /*!< AVX2 for Haswell and Bulldozer */ # define XXH_AVX512 3 /*!< AVX512 for Skylake and Icelake */ # define XXH_NEON 4 /*!< NEON for most ARMv7-A, all AArch64, and WASM SIMD128 */ # define XXH_VSX 5 /*!< VSX and ZVector for POWER8/z13 (64-bit) */ # define XXH_SVE 6 /*!< SVE for some ARMv8-A and ARMv9-A */ # define XXH_LSX 7 /*!< LSX (128-bit SIMD) for LoongArch64 */ /*-********************************************************************** * XXH3 64-bit variant ************************************************************************/ /*! * @brief Calculates 64-bit unseeded variant of XXH3 hash of @p input. * * @param input The block of data to be hashed, at least @p length bytes in size. * @param length The length of @p input, in bytes. * * @pre * The memory between @p input and @p input + @p length must be valid, * readable, contiguous memory. However, if @p length is `0`, @p input may be * `NULL`. In C++, this also must be *TriviallyCopyable*. * * @return The calculated 64-bit XXH3 hash value. * * @note * This is equivalent to @ref XXH3_64bits_withSeed() with a seed of `0`, however * it may have slightly better performance due to constant propagation of the * defaults. * * @see * XXH3_64bits_withSeed(), XXH3_64bits_withSecret(): other seeding variants * @see @ref single_shot_example "Single Shot Example" for an example. */ XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits(XXH_NOESCAPE const void* input, size_t length); /*! * @brief Calculates 64-bit seeded variant of XXH3 hash of @p input. * * @param input The block of data to be hashed, at least @p length bytes in size. * @param length The length of @p input, in bytes. * @param seed The 64-bit seed to alter the hash result predictably. * * @pre * The memory between @p input and @p input + @p length must be valid, * readable, contiguous memory. However, if @p length is `0`, @p input may be * `NULL`. In C++, this also must be *TriviallyCopyable*. * * @return The calculated 64-bit XXH3 hash value. * * @note * seed == 0 produces the same results as @ref XXH3_64bits(). * * This variant generates a custom secret on the fly based on default secret * altered using the @p seed value. * * While this operation is decently fast, note that it's not completely free. * * @see @ref single_shot_example "Single Shot Example" for an example. */ XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_withSeed(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed); /*! * The bare minimum size for a custom secret. * * @see * XXH3_64bits_withSecret(), XXH3_64bits_reset_withSecret(), * XXH3_128bits_withSecret(), XXH3_128bits_reset_withSecret(). */ #define XXH3_SECRET_SIZE_MIN 136 /*! * @brief Calculates 64-bit variant of XXH3 with a custom "secret". * * @param data The block of data to be hashed, at least @p len bytes in size. * @param len The length of @p data, in bytes. * @param secret The secret data. * @param secretSize The length of @p secret, in bytes. * * @return The calculated 64-bit XXH3 hash value. * * @pre * The memory between @p data and @p data + @p len must be valid, * readable, contiguous memory. However, if @p length is `0`, @p data may be * `NULL`. In C++, this also must be *TriviallyCopyable*. * * It's possible to provide any blob of bytes as a "secret" to generate the hash. * This makes it more difficult for an external actor to prepare an intentional collision. * The main condition is that @p secretSize *must* be large enough (>= @ref XXH3_SECRET_SIZE_MIN). * However, the quality of the secret impacts the dispersion of the hash algorithm. * Therefore, the secret _must_ look like a bunch of random bytes. * Avoid "trivial" or structured data such as repeated sequences or a text document. * Whenever in doubt about the "randomness" of the blob of bytes, * consider employing @ref XXH3_generateSecret() instead (see below). * It will generate a proper high entropy secret derived from the blob of bytes. * Another advantage of using XXH3_generateSecret() is that * it guarantees that all bits within the initial blob of bytes * will impact every bit of the output. * This is not necessarily the case when using the blob of bytes directly * because, when hashing _small_ inputs, only a portion of the secret is employed. * * @see @ref single_shot_example "Single Shot Example" for an example. */ XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_withSecret(XXH_NOESCAPE const void* data, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize); /******* Streaming *******/ #ifndef XXH_NO_STREAM /* * Streaming requires state maintenance. * This operation costs memory and CPU. * As a consequence, streaming is slower than one-shot hashing. * For better performance, prefer one-shot functions whenever applicable. */ /*! * @brief The opaque state struct for the XXH3 streaming API. * * @see XXH3_state_s for details. * @see @ref streaming_example "Streaming Example" */ typedef struct XXH3_state_s XXH3_state_t; XXH_PUBLIC_API XXH_MALLOCF XXH3_state_t* XXH3_createState(void); XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr); /*! * @brief Copies one @ref XXH3_state_t to another. * * @param dst_state The state to copy to. * @param src_state The state to copy from. * @pre * @p dst_state and @p src_state must not be `NULL` and must not overlap. */ XXH_PUBLIC_API void XXH3_copyState(XXH_NOESCAPE XXH3_state_t* dst_state, XXH_NOESCAPE const XXH3_state_t* src_state); /*! * @brief Resets an @ref XXH3_state_t to begin a new hash. * * @param statePtr The state struct to reset. * * @pre * @p statePtr must not be `NULL`. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note * - This function resets `statePtr` and generate a secret with default parameters. * - Call this function before @ref XXH3_64bits_update(). * - Digest will be equivalent to `XXH3_64bits()`. * * @see @ref streaming_example "Streaming Example" * */ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr); /*! * @brief Resets an @ref XXH3_state_t with 64-bit seed to begin a new hash. * * @param statePtr The state struct to reset. * @param seed The 64-bit seed to alter the hash result predictably. * * @pre * @p statePtr must not be `NULL`. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note * - This function resets `statePtr` and generate a secret from `seed`. * - Call this function before @ref XXH3_64bits_update(). * - Digest will be equivalent to `XXH3_64bits_withSeed()`. * * @see @ref streaming_example "Streaming Example" * */ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed); /*! * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. * * @param statePtr The state struct to reset. * @param secret The secret data. * @param secretSize The length of @p secret, in bytes. * * @pre * @p statePtr must not be `NULL`. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note * `secret` is referenced, it _must outlive_ the hash streaming session. * * Similar to one-shot API, `secretSize` must be >= @ref XXH3_SECRET_SIZE_MIN, * and the quality of produced hash values depends on secret's entropy * (secret's content should look like a bunch of random bytes). * When in doubt about the randomness of a candidate `secret`, * consider employing `XXH3_generateSecret()` instead (see below). * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize); /*! * @brief Consumes a block of @p input to an @ref XXH3_state_t. * * @param statePtr The state struct to update. * @param input The block of data to be hashed, at least @p length bytes in size. * @param length The length of @p input, in bytes. * * @pre * @p statePtr must not be `NULL`. * @pre * The memory between @p input and @p input + @p length must be valid, * readable, contiguous memory. However, if @p length is `0`, @p input may be * `NULL`. In C++, this also must be *TriviallyCopyable*. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note Call this to incrementally consume blocks of data. * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update (XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); /*! * @brief Returns the calculated XXH3 64-bit hash value from an @ref XXH3_state_t. * * @param statePtr The state struct to calculate the hash from. * * @pre * @p statePtr must not be `NULL`. * * @return The calculated XXH3 64-bit hash value from that state. * * @note * Calling XXH3_64bits_digest() will not affect @p statePtr, so you can update, * digest, and update again. * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_digest (XXH_NOESCAPE const XXH3_state_t* statePtr); #endif /* !XXH_NO_STREAM */ /* note : canonical representation of XXH3 is the same as XXH64 * since they both produce XXH64_hash_t values */ /*-********************************************************************** * XXH3 128-bit variant ************************************************************************/ /*! * @brief The return value from 128-bit hashes. * * Stored in little endian order, although the fields themselves are in native * endianness. */ typedef struct { XXH64_hash_t low64; /*!< `value & 0xFFFFFFFFFFFFFFFF` */ XXH64_hash_t high64; /*!< `value >> 64` */ } XXH128_hash_t; /*! * @brief Calculates 128-bit unseeded variant of XXH3 of @p data. * * @param data The block of data to be hashed, at least @p length bytes in size. * @param len The length of @p data, in bytes. * * @return The calculated 128-bit variant of XXH3 value. * * The 128-bit variant of XXH3 has more strength, but it has a bit of overhead * for shorter inputs. * * This is equivalent to @ref XXH3_128bits_withSeed() with a seed of `0`, however * it may have slightly better performance due to constant propagation of the * defaults. * * @see XXH3_128bits_withSeed(), XXH3_128bits_withSecret(): other seeding variants * @see @ref single_shot_example "Single Shot Example" for an example. */ XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits(XXH_NOESCAPE const void* data, size_t len); /*! @brief Calculates 128-bit seeded variant of XXH3 hash of @p data. * * @param data The block of data to be hashed, at least @p length bytes in size. * @param len The length of @p data, in bytes. * @param seed The 64-bit seed to alter the hash result predictably. * * @return The calculated 128-bit variant of XXH3 value. * * @note * seed == 0 produces the same results as @ref XXH3_64bits(). * * This variant generates a custom secret on the fly based on default secret * altered using the @p seed value. * * While this operation is decently fast, note that it's not completely free. * * @see XXH3_128bits(), XXH3_128bits_withSecret(): other seeding variants * @see @ref single_shot_example "Single Shot Example" for an example. */ XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_withSeed(XXH_NOESCAPE const void* data, size_t len, XXH64_hash_t seed); /*! * @brief Calculates 128-bit variant of XXH3 with a custom "secret". * * @param data The block of data to be hashed, at least @p len bytes in size. * @param len The length of @p data, in bytes. * @param secret The secret data. * @param secretSize The length of @p secret, in bytes. * * @return The calculated 128-bit variant of XXH3 value. * * It's possible to provide any blob of bytes as a "secret" to generate the hash. * This makes it more difficult for an external actor to prepare an intentional collision. * The main condition is that @p secretSize *must* be large enough (>= @ref XXH3_SECRET_SIZE_MIN). * However, the quality of the secret impacts the dispersion of the hash algorithm. * Therefore, the secret _must_ look like a bunch of random bytes. * Avoid "trivial" or structured data such as repeated sequences or a text document. * Whenever in doubt about the "randomness" of the blob of bytes, * consider employing @ref XXH3_generateSecret() instead (see below). * It will generate a proper high entropy secret derived from the blob of bytes. * Another advantage of using XXH3_generateSecret() is that * it guarantees that all bits within the initial blob of bytes * will impact every bit of the output. * This is not necessarily the case when using the blob of bytes directly * because, when hashing _small_ inputs, only a portion of the secret is employed. * * @see @ref single_shot_example "Single Shot Example" for an example. */ XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_withSecret(XXH_NOESCAPE const void* data, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize); /******* Streaming *******/ #ifndef XXH_NO_STREAM /* * Streaming requires state maintenance. * This operation costs memory and CPU. * As a consequence, streaming is slower than one-shot hashing. * For better performance, prefer one-shot functions whenever applicable. * * XXH3_128bits uses the same XXH3_state_t as XXH3_64bits(). * Use already declared XXH3_createState() and XXH3_freeState(). * * All reset and streaming functions have same meaning as their 64-bit counterpart. */ /*! * @brief Resets an @ref XXH3_state_t to begin a new hash. * * @param statePtr The state struct to reset. * * @pre * @p statePtr must not be `NULL`. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note * - This function resets `statePtr` and generate a secret with default parameters. * - Call it before @ref XXH3_128bits_update(). * - Digest will be equivalent to `XXH3_128bits()`. * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr); /*! * @brief Resets an @ref XXH3_state_t with 64-bit seed to begin a new hash. * * @param statePtr The state struct to reset. * @param seed The 64-bit seed to alter the hash result predictably. * * @pre * @p statePtr must not be `NULL`. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note * - This function resets `statePtr` and generate a secret from `seed`. * - Call it before @ref XXH3_128bits_update(). * - Digest will be equivalent to `XXH3_128bits_withSeed()`. * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed); /*! * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. * * @param statePtr The state struct to reset. * @param secret The secret data. * @param secretSize The length of @p secret, in bytes. * * @pre * @p statePtr must not be `NULL`. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * `secret` is referenced, it _must outlive_ the hash streaming session. * Similar to one-shot API, `secretSize` must be >= @ref XXH3_SECRET_SIZE_MIN, * and the quality of produced hash values depends on secret's entropy * (secret's content should look like a bunch of random bytes). * When in doubt about the randomness of a candidate `secret`, * consider employing `XXH3_generateSecret()` instead (see below). * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize); /*! * @brief Consumes a block of @p input to an @ref XXH3_state_t. * * Call this to incrementally consume blocks of data. * * @param statePtr The state struct to update. * @param input The block of data to be hashed, at least @p length bytes in size. * @param length The length of @p input, in bytes. * * @pre * @p statePtr must not be `NULL`. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @note * The memory between @p input and @p input + @p length must be valid, * readable, contiguous memory. However, if @p length is `0`, @p input may be * `NULL`. In C++, this also must be *TriviallyCopyable*. * */ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_update (XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); /*! * @brief Returns the calculated XXH3 128-bit hash value from an @ref XXH3_state_t. * * @param statePtr The state struct to calculate the hash from. * * @pre * @p statePtr must not be `NULL`. * * @return The calculated XXH3 128-bit hash value from that state. * * @note * Calling XXH3_128bits_digest() will not affect @p statePtr, so you can update, * digest, and update again. * */ XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_digest (XXH_NOESCAPE const XXH3_state_t* statePtr); #endif /* !XXH_NO_STREAM */ /* Following helper functions make it possible to compare XXH128_hast_t values. * Since XXH128_hash_t is a structure, this capability is not offered by the language. * Note: For better performance, these functions can be inlined using XXH_INLINE_ALL */ /*! * @brief Check equality of two XXH128_hash_t values * * @param h1 The 128-bit hash value. * @param h2 Another 128-bit hash value. * * @return `1` if `h1` and `h2` are equal. * @return `0` if they are not. */ XXH_PUBLIC_API XXH_PUREF int XXH128_isEqual(XXH128_hash_t h1, XXH128_hash_t h2); /*! * @brief Compares two @ref XXH128_hash_t * * This comparator is compatible with stdlib's `qsort()`/`bsearch()`. * * @param h128_1 Left-hand side value * @param h128_2 Right-hand side value * * @return >0 if @p h128_1 > @p h128_2 * @return =0 if @p h128_1 == @p h128_2 * @return <0 if @p h128_1 < @p h128_2 */ XXH_PUBLIC_API XXH_PUREF int XXH128_cmp(XXH_NOESCAPE const void* h128_1, XXH_NOESCAPE const void* h128_2); /******* Canonical representation *******/ typedef struct { unsigned char digest[sizeof(XXH128_hash_t)]; } XXH128_canonical_t; /*! * @brief Converts an @ref XXH128_hash_t to a big endian @ref XXH128_canonical_t. * * @param dst The @ref XXH128_canonical_t pointer to be stored to. * @param hash The @ref XXH128_hash_t to be converted. * * @pre * @p dst must not be `NULL`. * @see @ref canonical_representation_example "Canonical Representation Example" */ XXH_PUBLIC_API void XXH128_canonicalFromHash(XXH_NOESCAPE XXH128_canonical_t* dst, XXH128_hash_t hash); /*! * @brief Converts an @ref XXH128_canonical_t to a native @ref XXH128_hash_t. * * @param src The @ref XXH128_canonical_t to convert. * * @pre * @p src must not be `NULL`. * * @return The converted hash. * @see @ref canonical_representation_example "Canonical Representation Example" */ XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH128_hashFromCanonical(XXH_NOESCAPE const XXH128_canonical_t* src); #endif /* !XXH_NO_XXH3 */ #endif /* XXH_NO_LONG_LONG */ /*! * @} */ #endif /* XXHASH_H_5627135585666179 */ #if defined(XXH_STATIC_LINKING_ONLY) && !defined(XXHASH_H_STATIC_13879238742) #define XXHASH_H_STATIC_13879238742 /* **************************************************************************** * This section contains declarations which are not guaranteed to remain stable. * They may change in future versions, becoming incompatible with a different * version of the library. * These declarations should only be used with static linking. * Never use them in association with dynamic linking! ***************************************************************************** */ /* * These definitions are only present to allow static allocation * of XXH states, on stack or in a struct, for example. * Never **ever** access their members directly. */ /*! * @internal * @brief Structure for XXH32 streaming API. * * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. Otherwise it is * an opaque type. This allows fields to safely be changed. * * Typedef'd to @ref XXH32_state_t. * Do not access the members of this struct directly. * @see XXH64_state_s, XXH3_state_s */ struct XXH32_state_s { XXH32_hash_t total_len_32; /*!< Total length hashed, modulo 2^32 */ XXH32_hash_t large_len; /*!< Whether the hash is >= 16 (handles @ref total_len_32 overflow) */ XXH32_hash_t acc[4]; /*!< Accumulator lanes */ unsigned char buffer[16]; /*!< Internal buffer for partial reads. */ XXH32_hash_t bufferedSize; /*!< Amount of data in @ref buffer */ XXH32_hash_t reserved; /*!< Reserved field. Do not read nor write to it. */ }; /* typedef'd to XXH32_state_t */ #ifndef XXH_NO_LONG_LONG /* defined when there is no 64-bit support */ /*! * @internal * @brief Structure for XXH64 streaming API. * * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. Otherwise it is * an opaque type. This allows fields to safely be changed. * * Typedef'd to @ref XXH64_state_t. * Do not access the members of this struct directly. * @see XXH32_state_s, XXH3_state_s */ struct XXH64_state_s { XXH64_hash_t total_len; /*!< Total length hashed. This is always 64-bit. */ XXH64_hash_t acc[4]; /*!< Accumulator lanes */ unsigned char buffer[32]; /*!< Internal buffer for partial reads.. */ XXH32_hash_t bufferedSize; /*!< Amount of data in @ref buffer */ XXH32_hash_t reserved32; /*!< Reserved field, needed for padding anyways*/ XXH64_hash_t reserved64; /*!< Reserved field. Do not read or write to it. */ }; /* typedef'd to XXH64_state_t */ #ifndef XXH_NO_XXH3 #if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* >= C11 */ # define XXH_ALIGN(n) _Alignas(n) #elif defined(__cplusplus) && (__cplusplus >= 201103L) /* >= C++11 */ /* In C++ alignas() is a keyword */ # define XXH_ALIGN(n) alignas(n) #elif defined(__GNUC__) # define XXH_ALIGN(n) __attribute__ ((aligned(n))) #elif defined(_MSC_VER) # define XXH_ALIGN(n) __declspec(align(n)) #else # define XXH_ALIGN(n) /* disabled */ #endif /* Old GCC versions only accept the attribute after the type in structures. */ #if !(defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L)) /* C11+ */ \ && ! (defined(__cplusplus) && (__cplusplus >= 201103L)) /* >= C++11 */ \ && defined(__GNUC__) # define XXH_ALIGN_MEMBER(align, type) type XXH_ALIGN(align) #else # define XXH_ALIGN_MEMBER(align, type) XXH_ALIGN(align) type #endif /*! * @brief The size of the internal XXH3 buffer. * * This is the optimal update size for incremental hashing. * * @see XXH3_64b_update(), XXH3_128b_update(). */ #define XXH3_INTERNALBUFFER_SIZE 256 /*! * @internal * @brief Default size of the secret buffer (and @ref XXH3_kSecret). * * This is the size used in @ref XXH3_kSecret and the seeded functions. * * Not to be confused with @ref XXH3_SECRET_SIZE_MIN. */ #define XXH3_SECRET_DEFAULT_SIZE 192 /*! * @internal * @brief Structure for XXH3 streaming API. * * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. * Otherwise it is an opaque type. * Never use this definition in combination with dynamic library. * This allows fields to safely be changed in the future. * * @note ** This structure has a strict alignment requirement of 64 bytes!! ** * Do not allocate this with `malloc()` or `new`, * it will not be sufficiently aligned. * Use @ref XXH3_createState() and @ref XXH3_freeState(), or stack allocation. * * Typedef'd to @ref XXH3_state_t. * Do never access the members of this struct directly. * * @see XXH3_INITSTATE() for stack initialization. * @see XXH3_createState(), XXH3_freeState(). * @see XXH32_state_s, XXH64_state_s */ struct XXH3_state_s { XXH_ALIGN_MEMBER(64, XXH64_hash_t acc[8]); /*!< The 8 accumulators. See @ref XXH32_state_s::v and @ref XXH64_state_s::v */ XXH_ALIGN_MEMBER(64, unsigned char customSecret[XXH3_SECRET_DEFAULT_SIZE]); /*!< Used to store a custom secret generated from a seed. */ XXH_ALIGN_MEMBER(64, unsigned char buffer[XXH3_INTERNALBUFFER_SIZE]); /*!< The internal buffer. @see XXH32_state_s::mem32 */ XXH32_hash_t bufferedSize; /*!< The amount of memory in @ref buffer, @see XXH32_state_s::memsize */ XXH32_hash_t useSeed; /*!< Reserved field. Needed for padding on 64-bit. */ size_t nbStripesSoFar; /*!< Number or stripes processed. */ XXH64_hash_t totalLen; /*!< Total length hashed. 64-bit even on 32-bit targets. */ size_t nbStripesPerBlock; /*!< Number of stripes per block. */ size_t secretLimit; /*!< Size of @ref customSecret or @ref extSecret */ XXH64_hash_t seed; /*!< Seed for _withSeed variants. Must be zero otherwise, @see XXH3_INITSTATE() */ XXH64_hash_t reserved64; /*!< Reserved field. */ const unsigned char* extSecret; /*!< Reference to an external secret for the _withSecret variants, NULL * for other variants. */ /* note: there may be some padding at the end due to alignment on 64 bytes */ }; /* typedef'd to XXH3_state_t */ #undef XXH_ALIGN_MEMBER /*! * @brief Initializes a stack-allocated `XXH3_state_s`. * * When the @ref XXH3_state_t structure is merely emplaced on stack, * it should be initialized with XXH3_INITSTATE() or a memset() * in case its first reset uses XXH3_NNbits_reset_withSeed(). * This init can be omitted if the first reset uses default or _withSecret mode. * This operation isn't necessary when the state is created with XXH3_createState(). * Note that this doesn't prepare the state for a streaming operation, * it's still necessary to use XXH3_NNbits_reset*() afterwards. */ #define XXH3_INITSTATE(XXH3_state_ptr) \ do { \ XXH3_state_t* tmp_xxh3_state_ptr = (XXH3_state_ptr); \ tmp_xxh3_state_ptr->seed = 0; \ tmp_xxh3_state_ptr->extSecret = NULL; \ } while(0) /*! * @brief Calculates the 128-bit hash of @p data using XXH3. * * @param data The block of data to be hashed, at least @p len bytes in size. * @param len The length of @p data, in bytes. * @param seed The 64-bit seed to alter the hash's output predictably. * * @pre * The memory between @p data and @p data + @p len must be valid, * readable, contiguous memory. However, if @p len is `0`, @p data may be * `NULL`. In C++, this also must be *TriviallyCopyable*. * * @return The calculated 128-bit XXH3 value. * * @see @ref single_shot_example "Single Shot Example" for an example. */ XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH128(XXH_NOESCAPE const void* data, size_t len, XXH64_hash_t seed); /* === Experimental API === */ /* Symbols defined below must be considered tied to a specific library version. */ /*! * @brief Derive a high-entropy secret from any user-defined content, named customSeed. * * @param secretBuffer A writable buffer for derived high-entropy secret data. * @param secretSize Size of secretBuffer, in bytes. Must be >= XXH3_SECRET_SIZE_MIN. * @param customSeed A user-defined content. * @param customSeedSize Size of customSeed, in bytes. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * The generated secret can be used in combination with `*_withSecret()` functions. * The `_withSecret()` variants are useful to provide a higher level of protection * than 64-bit seed, as it becomes much more difficult for an external actor to * guess how to impact the calculation logic. * * The function accepts as input a custom seed of any length and any content, * and derives from it a high-entropy secret of length @p secretSize into an * already allocated buffer @p secretBuffer. * * The generated secret can then be used with any `*_withSecret()` variant. * The functions @ref XXH3_128bits_withSecret(), @ref XXH3_64bits_withSecret(), * @ref XXH3_128bits_reset_withSecret() and @ref XXH3_64bits_reset_withSecret() * are part of this list. They all accept a `secret` parameter * which must be large enough for implementation reasons (>= @ref XXH3_SECRET_SIZE_MIN) * _and_ feature very high entropy (consist of random-looking bytes). * These conditions can be a high bar to meet, so @ref XXH3_generateSecret() can * be employed to ensure proper quality. * * @p customSeed can be anything. It can have any size, even small ones, * and its content can be anything, even "poor entropy" sources such as a bunch * of zeroes. The resulting `secret` will nonetheless provide all required qualities. * * @pre * - @p secretSize must be >= @ref XXH3_SECRET_SIZE_MIN * - When @p customSeedSize > 0, supplying NULL as customSeed is undefined behavior. * * Example code: * @code{.c} * #include * #include * #include * #define XXH_STATIC_LINKING_ONLY // expose unstable API * #include "xxhash.h" * // Hashes argv[2] using the entropy from argv[1]. * int main(int argc, char* argv[]) * { * char secret[XXH3_SECRET_SIZE_MIN]; * if (argv != 3) { return 1; } * XXH3_generateSecret(secret, sizeof(secret), argv[1], strlen(argv[1])); * XXH64_hash_t h = XXH3_64bits_withSecret( * argv[2], strlen(argv[2]), * secret, sizeof(secret) * ); * printf("%016llx\n", (unsigned long long) h); * } * @endcode */ XXH_PUBLIC_API XXH_errorcode XXH3_generateSecret(XXH_NOESCAPE void* secretBuffer, size_t secretSize, XXH_NOESCAPE const void* customSeed, size_t customSeedSize); /*! * @brief Generate the same secret as the _withSeed() variants. * * @param secretBuffer A writable buffer of @ref XXH3_SECRET_DEFAULT_SIZE bytes * @param seed The 64-bit seed to alter the hash result predictably. * * The generated secret can be used in combination with *`*_withSecret()` and `_withSecretandSeed()` variants. * * Example C++ `std::string` hash class: * @code{.cpp} * #include * #define XXH_STATIC_LINKING_ONLY // expose unstable API * #include "xxhash.h" * // Slow, seeds each time * class HashSlow { * XXH64_hash_t seed; * public: * HashSlow(XXH64_hash_t s) : seed{s} {} * size_t operator()(const std::string& x) const { * return size_t{XXH3_64bits_withSeed(x.c_str(), x.length(), seed)}; * } * }; * // Fast, caches the seeded secret for future uses. * class HashFast { * unsigned char secret[XXH3_SECRET_DEFAULT_SIZE]; * public: * HashFast(XXH64_hash_t s) { * XXH3_generateSecret_fromSeed(secret, seed); * } * size_t operator()(const std::string& x) const { * return size_t{ * XXH3_64bits_withSecret(x.c_str(), x.length(), secret, sizeof(secret)) * }; * } * }; * @endcode */ XXH_PUBLIC_API void XXH3_generateSecret_fromSeed(XXH_NOESCAPE void* secretBuffer, XXH64_hash_t seed); /*! * @brief Maximum size of "short" key in bytes. */ #define XXH3_MIDSIZE_MAX 240 /*! * @brief Calculates 64/128-bit seeded variant of XXH3 hash of @p data. * * @param data The block of data to be hashed, at least @p len bytes in size. * @param len The length of @p data, in bytes. * @param secret The secret data. * @param secretSize The length of @p secret, in bytes. * @param seed The 64-bit seed to alter the hash result predictably. * * These variants generate hash values using either: * - @p seed for "short" keys (< @ref XXH3_MIDSIZE_MAX = 240 bytes) * - @p secret for "large" keys (>= @ref XXH3_MIDSIZE_MAX). * * This generally benefits speed, compared to `_withSeed()` or `_withSecret()`. * `_withSeed()` has to generate the secret on the fly for "large" keys. * It's fast, but can be perceptible for "not so large" keys (< 1 KB). * `_withSecret()` has to generate the masks on the fly for "small" keys, * which requires more instructions than _withSeed() variants. * Therefore, _withSecretandSeed variant combines the best of both worlds. * * When @p secret has been generated by XXH3_generateSecret_fromSeed(), * this variant produces *exactly* the same results as `_withSeed()` variant, * hence offering only a pure speed benefit on "large" input, * by skipping the need to regenerate the secret for every large input. * * Another usage scenario is to hash the secret to a 64-bit hash value, * for example with XXH3_64bits(), which then becomes the seed, * and then employ both the seed and the secret in _withSecretandSeed(). * On top of speed, an added benefit is that each bit in the secret * has a 50% chance to swap each bit in the output, via its impact to the seed. * * This is not guaranteed when using the secret directly in "small data" scenarios, * because only portions of the secret are employed for small data. */ XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_withSecretandSeed(XXH_NOESCAPE const void* data, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed); /*! * @brief Calculates 128-bit seeded variant of XXH3 hash of @p data. * * @param data The memory segment to be hashed, at least @p len bytes in size. * @param length The length of @p data, in bytes. * @param secret The secret used to alter hash result predictably. * @param secretSize The length of @p secret, in bytes (must be >= XXH3_SECRET_SIZE_MIN) * @param seed64 The 64-bit seed to alter the hash result predictably. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @see XXH3_64bits_withSecretandSeed(): contract is the same. */ XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t length, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed64); #ifndef XXH_NO_STREAM /*! * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. * * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). * @param secret The secret data. * @param secretSize The length of @p secret, in bytes. * @param seed64 The 64-bit seed to alter the hash result predictably. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @see XXH3_64bits_withSecretandSeed(). Contract is identical. */ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed64); /*! * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. * * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). * @param secret The secret data. * @param secretSize The length of @p secret, in bytes. * @param seed64 The 64-bit seed to alter the hash result predictably. * * @return @ref XXH_OK on success. * @return @ref XXH_ERROR on failure. * * @see XXH3_64bits_withSecretandSeed(). Contract is identical. * * Note: there was a bug in an earlier version of this function (<= v0.8.2) * that would make it generate an incorrect hash value * when @p seed == 0 and @p length < XXH3_MIDSIZE_MAX * and @p secret is different from XXH3_generateSecret_fromSeed(). * As stated in the contract, the correct hash result must be * the same as XXH3_128bits_withSeed() when @p length <= XXH3_MIDSIZE_MAX. * Results generated by this older version are wrong, hence not comparable. */ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed64); #endif /* !XXH_NO_STREAM */ #endif /* !XXH_NO_XXH3 */ #endif /* XXH_NO_LONG_LONG */ #if defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API) # define XXH_IMPLEMENTATION #endif #endif /* defined(XXH_STATIC_LINKING_ONLY) && !defined(XXHASH_H_STATIC_13879238742) */ /* ======================================================================== */ /* ======================================================================== */ /* ======================================================================== */ /*-********************************************************************** * xxHash implementation *-********************************************************************** * xxHash's implementation used to be hosted inside xxhash.c. * * However, inlining requires implementation to be visible to the compiler, * hence be included alongside the header. * Previously, implementation was hosted inside xxhash.c, * which was then #included when inlining was activated. * This construction created issues with a few build and install systems, * as it required xxhash.c to be stored in /include directory. * * xxHash implementation is now directly integrated within xxhash.h. * As a consequence, xxhash.c is no longer needed in /include. * * xxhash.c is still available and is still useful. * In a "normal" setup, when xxhash is not inlined, * xxhash.h only exposes the prototypes and public symbols, * while xxhash.c can be built into an object file xxhash.o * which can then be linked into the final binary. ************************************************************************/ #if ( defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API) \ || defined(XXH_IMPLEMENTATION) ) && !defined(XXH_IMPLEM_13a8737387) # define XXH_IMPLEM_13a8737387 /* ************************************* * Tuning parameters ***************************************/ /*! * @defgroup tuning Tuning parameters * @{ * * Various macros to control xxHash's behavior. */ #ifdef XXH_DOXYGEN /*! * @brief Define this to disable 64-bit code. * * Useful if only using the @ref XXH32_family and you have a strict C90 compiler. */ # define XXH_NO_LONG_LONG # undef XXH_NO_LONG_LONG /* don't actually */ /*! * @brief Controls how unaligned memory is accessed. * * By default, access to unaligned memory is controlled by `memcpy()`, which is * safe and portable. * * Unfortunately, on some target/compiler combinations, the generated assembly * is sub-optimal. * * The below switch allow selection of a different access method * in the search for improved performance. * * @par Possible options: * * - `XXH_FORCE_MEMORY_ACCESS=0` (default): `memcpy` * @par * Use `memcpy()`. Safe and portable. Note that most modern compilers will * eliminate the function call and treat it as an unaligned access. * * - `XXH_FORCE_MEMORY_ACCESS=1`: `__attribute__((aligned(1)))` * @par * Depends on compiler extensions and is therefore not portable. * This method is safe _if_ your compiler supports it, * and *generally* as fast or faster than `memcpy`. * * - `XXH_FORCE_MEMORY_ACCESS=2`: Direct cast * @par * Casts directly and dereferences. This method doesn't depend on the * compiler, but it violates the C standard as it directly dereferences an * unaligned pointer. It can generate buggy code on targets which do not * support unaligned memory accesses, but in some circumstances, it's the * only known way to get the most performance. * * - `XXH_FORCE_MEMORY_ACCESS=3`: Byteshift * @par * Also portable. This can generate the best code on old compilers which don't * inline small `memcpy()` calls, and it might also be faster on big-endian * systems which lack a native byteswap instruction. However, some compilers * will emit literal byteshifts even if the target supports unaligned access. * * * @warning * Methods 1 and 2 rely on implementation-defined behavior. Use these with * care, as what works on one compiler/platform/optimization level may cause * another to read garbage data or even crash. * * See https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html for details. * * Prefer these methods in priority order (0 > 3 > 1 > 2) */ # define XXH_FORCE_MEMORY_ACCESS 0 /*! * @def XXH_SIZE_OPT * @brief Controls how much xxHash optimizes for size. * * xxHash, when compiled, tends to result in a rather large binary size. This * is mostly due to heavy usage to forced inlining and constant folding of the * @ref XXH3_family to increase performance. * * However, some developers prefer size over speed. This option can * significantly reduce the size of the generated code. When using the `-Os` * or `-Oz` options on GCC or Clang, this is defined to 1 by default, * otherwise it is defined to 0. * * Most of these size optimizations can be controlled manually. * * This is a number from 0-2. * - `XXH_SIZE_OPT` == 0: Default. xxHash makes no size optimizations. Speed * comes first. * - `XXH_SIZE_OPT` == 1: Default for `-Os` and `-Oz`. xxHash is more * conservative and disables hacks that increase code size. It implies the * options @ref XXH_NO_INLINE_HINTS == 1, @ref XXH_FORCE_ALIGN_CHECK == 0, * and @ref XXH3_NEON_LANES == 8 if they are not already defined. * - `XXH_SIZE_OPT` == 2: xxHash tries to make itself as small as possible. * Performance may cry. For example, the single shot functions just use the * streaming API. */ # define XXH_SIZE_OPT 0 /*! * @def XXH_FORCE_ALIGN_CHECK * @brief If defined to non-zero, adds a special path for aligned inputs (XXH32() * and XXH64() only). * * This is an important performance trick for architectures without decent * unaligned memory access performance. * * It checks for input alignment, and when conditions are met, uses a "fast * path" employing direct 32-bit/64-bit reads, resulting in _dramatically * faster_ read speed. * * The check costs one initial branch per hash, which is generally negligible, * but not zero. * * Moreover, it's not useful to generate an additional code path if memory * access uses the same instruction for both aligned and unaligned * addresses (e.g. x86 and aarch64). * * In these cases, the alignment check can be removed by setting this macro to 0. * Then the code will always use unaligned memory access. * Align check is automatically disabled on x86, x64, ARM64, and some ARM chips * which are platforms known to offer good unaligned memory accesses performance. * * It is also disabled by default when @ref XXH_SIZE_OPT >= 1. * * This option does not affect XXH3 (only XXH32 and XXH64). */ # define XXH_FORCE_ALIGN_CHECK 0 /*! * @def XXH_NO_INLINE_HINTS * @brief When non-zero, sets all functions to `static`. * * By default, xxHash tries to force the compiler to inline almost all internal * functions. * * This can usually improve performance due to reduced jumping and improved * constant folding, but significantly increases the size of the binary which * might not be favorable. * * Additionally, sometimes the forced inlining can be detrimental to performance, * depending on the architecture. * * XXH_NO_INLINE_HINTS marks all internal functions as static, giving the * compiler full control on whether to inline or not. * * When not optimizing (-O0), using `-fno-inline` with GCC or Clang, or if * @ref XXH_SIZE_OPT >= 1, this will automatically be defined. */ # define XXH_NO_INLINE_HINTS 0 /*! * @def XXH3_INLINE_SECRET * @brief Determines whether to inline the XXH3 withSecret code. * * When the secret size is known, the compiler can improve the performance * of XXH3_64bits_withSecret() and XXH3_128bits_withSecret(). * * However, if the secret size is not known, it doesn't have any benefit. This * happens when xxHash is compiled into a global symbol. Therefore, if * @ref XXH_INLINE_ALL is *not* defined, this will be defined to 0. * * Additionally, this defaults to 0 on GCC 12+, which has an issue with function pointers * that are *sometimes* force inline on -Og, and it is impossible to automatically * detect this optimization level. */ # define XXH3_INLINE_SECRET 0 /*! * @def XXH32_ENDJMP * @brief Whether to use a jump for `XXH32_finalize`. * * For performance, `XXH32_finalize` uses multiple branches in the finalizer. * This is generally preferable for performance, * but depending on exact architecture, a jmp may be preferable. * * This setting is only possibly making a difference for very small inputs. */ # define XXH32_ENDJMP 0 /*! * @internal * @brief Redefines old internal names. * * For compatibility with code that uses xxHash's internals before the names * were changed to improve namespacing. There is no other reason to use this. */ # define XXH_OLD_NAMES # undef XXH_OLD_NAMES /* don't actually use, it is ugly. */ /*! * @def XXH_NO_STREAM * @brief Disables the streaming API. * * When xxHash is not inlined and the streaming functions are not used, disabling * the streaming functions can improve code size significantly, especially with * the @ref XXH3_family which tends to make constant folded copies of itself. */ # define XXH_NO_STREAM # undef XXH_NO_STREAM /* don't actually */ #endif /* XXH_DOXYGEN */ /*! * @} */ #ifndef XXH_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ /* prefer __packed__ structures (method 1) for GCC * < ARMv7 with unaligned access (e.g. Raspbian armhf) still uses byte shifting, so we use memcpy * which for some reason does unaligned loads. */ # if defined(__GNUC__) && !(defined(__ARM_ARCH) && __ARM_ARCH < 7 && defined(__ARM_FEATURE_UNALIGNED)) # define XXH_FORCE_MEMORY_ACCESS 1 # endif #endif #ifndef XXH_SIZE_OPT /* default to 1 for -Os or -Oz */ # if (defined(__GNUC__) || defined(__clang__)) && defined(__OPTIMIZE_SIZE__) # define XXH_SIZE_OPT 1 # else # define XXH_SIZE_OPT 0 # endif #endif #ifndef XXH_FORCE_ALIGN_CHECK /* can be defined externally */ /* don't check on sizeopt, x86, aarch64, or arm when unaligned access is available */ # if XXH_SIZE_OPT >= 1 || \ defined(__i386) || defined(__x86_64__) || defined(__aarch64__) || defined(__ARM_FEATURE_UNALIGNED) \ || defined(_M_IX86) || defined(_M_X64) || defined(_M_ARM64) || defined(_M_ARM) /* visual */ # define XXH_FORCE_ALIGN_CHECK 0 # else # define XXH_FORCE_ALIGN_CHECK 1 # endif #endif #ifndef XXH_NO_INLINE_HINTS # if XXH_SIZE_OPT >= 1 || defined(__NO_INLINE__) /* -O0, -fno-inline */ # define XXH_NO_INLINE_HINTS 1 # else # define XXH_NO_INLINE_HINTS 0 # endif #endif #ifndef XXH3_INLINE_SECRET # if (defined(__GNUC__) && !defined(__clang__) && __GNUC__ >= 12) \ || !defined(XXH_INLINE_ALL) # define XXH3_INLINE_SECRET 0 # else # define XXH3_INLINE_SECRET 1 # endif #endif #ifndef XXH32_ENDJMP /* generally preferable for performance */ # define XXH32_ENDJMP 0 #endif /*! * @defgroup impl Implementation * @{ */ /* ************************************* * Includes & Memory related functions ***************************************/ #if defined(XXH_NO_STREAM) /* nothing */ #elif defined(XXH_NO_STDLIB) /* When requesting to disable any mention of stdlib, * the library loses the ability to invoked malloc / free. * In practice, it means that functions like `XXH*_createState()` * will always fail, and return NULL. * This flag is useful in situations where * xxhash.h is integrated into some kernel, embedded or limited environment * without access to dynamic allocation. */ static XXH_CONSTF void* XXH_malloc(size_t s) { (void)s; return NULL; } static void XXH_free(void* p) { (void)p; } #else /* * Modify the local functions below should you wish to use * different memory routines for malloc() and free() */ #include /*! * @internal * @brief Modify this function to use a different routine than malloc(). */ static XXH_MALLOCF void* XXH_malloc(size_t s) { return malloc(s); } /*! * @internal * @brief Modify this function to use a different routine than free(). */ static void XXH_free(void* p) { free(p); } #endif /* XXH_NO_STDLIB */ #include /*! * @internal * @brief Modify this function to use a different routine than memcpy(). */ static void* XXH_memcpy(void* dest, const void* src, size_t size) { return memcpy(dest,src,size); } #include /* ULLONG_MAX */ /* ************************************* * Compiler Specific Options ***************************************/ #ifdef _MSC_VER /* Visual Studio warning fix */ # pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ #endif #if XXH_NO_INLINE_HINTS /* disable inlining hints */ # if defined(__GNUC__) || defined(__clang__) # define XXH_FORCE_INLINE static __attribute__((__unused__)) # else # define XXH_FORCE_INLINE static # endif # define XXH_NO_INLINE static /* enable inlining hints */ #elif defined(__GNUC__) || defined(__clang__) # define XXH_FORCE_INLINE static __inline__ __attribute__((__always_inline__, __unused__)) # define XXH_NO_INLINE static __attribute__((__noinline__)) #elif defined(_MSC_VER) /* Visual Studio */ # define XXH_FORCE_INLINE static __forceinline # define XXH_NO_INLINE static __declspec(noinline) #elif defined (__cplusplus) \ || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L)) /* C99 */ # define XXH_FORCE_INLINE static inline # define XXH_NO_INLINE static #else # define XXH_FORCE_INLINE static # define XXH_NO_INLINE static #endif #if defined(XXH_INLINE_ALL) # define XXH_STATIC XXH_FORCE_INLINE #else # define XXH_STATIC static #endif #if XXH3_INLINE_SECRET # define XXH3_WITH_SECRET_INLINE XXH_FORCE_INLINE #else # define XXH3_WITH_SECRET_INLINE XXH_NO_INLINE #endif #if ((defined(sun) || defined(__sun)) && __cplusplus) /* Solaris includes __STDC_VERSION__ with C++. Tested with GCC 5.5 */ # define XXH_RESTRICT /* disable */ #elif defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* >= C99 */ # define XXH_RESTRICT restrict #elif (defined (__GNUC__) && ((__GNUC__ > 3) || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1))) \ || (defined (__clang__)) \ || (defined (_MSC_VER) && (_MSC_VER >= 1400)) \ || (defined (__INTEL_COMPILER) && (__INTEL_COMPILER >= 1300)) /* * There are a LOT more compilers that recognize __restrict but this * covers the major ones. */ # define XXH_RESTRICT __restrict #else # define XXH_RESTRICT /* disable */ #endif /* ************************************* * Debug ***************************************/ /*! * @ingroup tuning * @def XXH_DEBUGLEVEL * @brief Sets the debugging level. * * XXH_DEBUGLEVEL is expected to be defined externally, typically via the * compiler's command line options. The value must be a number. */ #ifndef XXH_DEBUGLEVEL # ifdef DEBUGLEVEL /* backwards compat */ # define XXH_DEBUGLEVEL DEBUGLEVEL # else # define XXH_DEBUGLEVEL 0 # endif #endif #if (XXH_DEBUGLEVEL>=1) # include /* note: can still be disabled with NDEBUG */ # define XXH_ASSERT(c) assert(c) #else # if defined(__INTEL_COMPILER) # define XXH_ASSERT(c) XXH_ASSUME((unsigned char) (c)) # else # define XXH_ASSERT(c) XXH_ASSUME(c) # endif #endif /* note: use after variable declarations */ #ifndef XXH_STATIC_ASSERT # if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* C11 */ # define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { _Static_assert((c),m); } while(0) # elif defined(__cplusplus) && (__cplusplus >= 201103L) /* C++11 */ # define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { static_assert((c),m); } while(0) # else # define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { struct xxh_sa { char x[(c) ? 1 : -1]; }; } while(0) # endif # define XXH_STATIC_ASSERT(c) XXH_STATIC_ASSERT_WITH_MESSAGE((c),#c) #endif /*! * @internal * @def XXH_COMPILER_GUARD(var) * @brief Used to prevent unwanted optimizations for @p var. * * It uses an empty GCC inline assembly statement with a register constraint * which forces @p var into a general purpose register (eg eax, ebx, ecx * on x86) and marks it as modified. * * This is used in a few places to avoid unwanted autovectorization (e.g. * XXH32_round()). All vectorization we want is explicit via intrinsics, * and _usually_ isn't wanted elsewhere. * * We also use it to prevent unwanted constant folding for AArch64 in * XXH3_initCustomSecret_scalar(). */ #if defined(__GNUC__) || defined(__clang__) # define XXH_COMPILER_GUARD(var) __asm__("" : "+r" (var)) #else # define XXH_COMPILER_GUARD(var) ((void)0) #endif /* Specifically for NEON vectors which use the "w" constraint, on * Clang. */ #if defined(__clang__) && defined(__ARM_ARCH) && !defined(__wasm__) # define XXH_COMPILER_GUARD_CLANG_NEON(var) __asm__("" : "+w" (var)) #else # define XXH_COMPILER_GUARD_CLANG_NEON(var) ((void)0) #endif /* ************************************* * Basic Types ***************************************/ #if !defined (__VMS) \ && (defined (__cplusplus) \ || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) # ifdef _AIX # include # else # include # endif typedef uint8_t xxh_u8; #else typedef unsigned char xxh_u8; #endif typedef XXH32_hash_t xxh_u32; #ifdef XXH_OLD_NAMES # warning "XXH_OLD_NAMES is planned to be removed starting v0.9. If the program depends on it, consider moving away from it by employing newer type names directly" # define BYTE xxh_u8 # define U8 xxh_u8 # define U32 xxh_u32 #endif /* *** Memory access *** */ /*! * @internal * @fn xxh_u32 XXH_read32(const void* ptr) * @brief Reads an unaligned 32-bit integer from @p ptr in native endianness. * * Affected by @ref XXH_FORCE_MEMORY_ACCESS. * * @param ptr The pointer to read from. * @return The 32-bit native endian integer from the bytes at @p ptr. */ /*! * @internal * @fn xxh_u32 XXH_readLE32(const void* ptr) * @brief Reads an unaligned 32-bit little endian integer from @p ptr. * * Affected by @ref XXH_FORCE_MEMORY_ACCESS. * * @param ptr The pointer to read from. * @return The 32-bit little endian integer from the bytes at @p ptr. */ /*! * @internal * @fn xxh_u32 XXH_readBE32(const void* ptr) * @brief Reads an unaligned 32-bit big endian integer from @p ptr. * * Affected by @ref XXH_FORCE_MEMORY_ACCESS. * * @param ptr The pointer to read from. * @return The 32-bit big endian integer from the bytes at @p ptr. */ /*! * @internal * @fn xxh_u32 XXH_readLE32_align(const void* ptr, XXH_alignment align) * @brief Like @ref XXH_readLE32(), but has an option for aligned reads. * * Affected by @ref XXH_FORCE_MEMORY_ACCESS. * Note that when @ref XXH_FORCE_ALIGN_CHECK == 0, the @p align parameter is * always @ref XXH_alignment::XXH_unaligned. * * @param ptr The pointer to read from. * @param align Whether @p ptr is aligned. * @pre * If @p align == @ref XXH_alignment::XXH_aligned, @p ptr must be 4 byte * aligned. * @return The 32-bit little endian integer from the bytes at @p ptr. */ #if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) /* * Manual byteshift. Best for old compilers which don't inline memcpy. * We actually directly use XXH_readLE32 and XXH_readBE32. */ #elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2)) /* * Force direct memory access. Only works on CPU which support unaligned memory * access in hardware. */ static xxh_u32 XXH_read32(const void* memPtr) { return *(const xxh_u32*) memPtr; } #elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1)) /* * __attribute__((aligned(1))) is supported by gcc and clang. Originally the * documentation claimed that it only increased the alignment, but actually it * can decrease it on gcc, clang, and icc: * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69502, * https://gcc.godbolt.org/z/xYez1j67Y. */ #ifdef XXH_OLD_NAMES typedef union { xxh_u32 u32; } __attribute__((__packed__)) unalign; #endif static xxh_u32 XXH_read32(const void* ptr) { typedef __attribute__((__aligned__(1))) xxh_u32 xxh_unalign32; return *((const xxh_unalign32*)ptr); } #else /* * Portable and safe solution. Generally efficient. * see: https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html */ static xxh_u32 XXH_read32(const void* memPtr) { xxh_u32 val; XXH_memcpy(&val, memPtr, sizeof(val)); return val; } #endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */ /* *** Endianness *** */ /*! * @ingroup tuning * @def XXH_CPU_LITTLE_ENDIAN * @brief Whether the target is little endian. * * Defined to 1 if the target is little endian, or 0 if it is big endian. * It can be defined externally, for example on the compiler command line. * * If it is not defined, * a runtime check (which is usually constant folded) is used instead. * * @note * This is not necessarily defined to an integer constant. * * @see XXH_isLittleEndian() for the runtime check. */ #ifndef XXH_CPU_LITTLE_ENDIAN /* * Try to detect endianness automatically, to avoid the nonstandard behavior * in `XXH_isLittleEndian()` */ # if defined(_WIN32) /* Windows is always little endian */ \ || defined(__LITTLE_ENDIAN__) \ || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) # define XXH_CPU_LITTLE_ENDIAN 1 # elif defined(__BIG_ENDIAN__) \ || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) # define XXH_CPU_LITTLE_ENDIAN 0 # else /*! * @internal * @brief Runtime check for @ref XXH_CPU_LITTLE_ENDIAN. * * Most compilers will constant fold this. */ static int XXH_isLittleEndian(void) { /* * Portable and well-defined behavior. * Don't use static: it is detrimental to performance. */ const union { xxh_u32 u; xxh_u8 c[4]; } one = { 1 }; return one.c[0]; } # define XXH_CPU_LITTLE_ENDIAN XXH_isLittleEndian() # endif #endif /* **************************************** * Compiler-specific Functions and Macros ******************************************/ #define XXH_GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) #ifdef __has_builtin # define XXH_HAS_BUILTIN(x) __has_builtin(x) #else # define XXH_HAS_BUILTIN(x) 0 #endif /* * C23 and future versions have standard "unreachable()". * Once it has been implemented reliably we can add it as an * additional case: * * ``` * #if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= XXH_C23_VN) * # include * # ifdef unreachable * # define XXH_UNREACHABLE() unreachable() * # endif * #endif * ``` * * Note C++23 also has std::unreachable() which can be detected * as follows: * ``` * #if defined(__cpp_lib_unreachable) && (__cpp_lib_unreachable >= 202202L) * # include * # define XXH_UNREACHABLE() std::unreachable() * #endif * ``` * NB: `__cpp_lib_unreachable` is defined in the `` header. * We don't use that as including `` in `extern "C"` blocks * doesn't work on GCC12 */ #if XXH_HAS_BUILTIN(__builtin_unreachable) # define XXH_UNREACHABLE() __builtin_unreachable() #elif defined(_MSC_VER) # define XXH_UNREACHABLE() __assume(0) #else # define XXH_UNREACHABLE() #endif #if XXH_HAS_BUILTIN(__builtin_assume) # define XXH_ASSUME(c) __builtin_assume(c) #else # define XXH_ASSUME(c) if (!(c)) { XXH_UNREACHABLE(); } #endif /*! * @internal * @def XXH_rotl32(x,r) * @brief 32-bit rotate left. * * @param x The 32-bit integer to be rotated. * @param r The number of bits to rotate. * @pre * @p r > 0 && @p r < 32 * @note * @p x and @p r may be evaluated multiple times. * @return The rotated result. */ #if !defined(NO_CLANG_BUILTIN) && XXH_HAS_BUILTIN(__builtin_rotateleft32) \ && XXH_HAS_BUILTIN(__builtin_rotateleft64) # define XXH_rotl32 __builtin_rotateleft32 # define XXH_rotl64 __builtin_rotateleft64 #elif XXH_HAS_BUILTIN(__builtin_stdc_rotate_left) # define XXH_rotl32 __builtin_stdc_rotate_left # define XXH_rotl64 __builtin_stdc_rotate_left /* Note: although _rotl exists for minGW (GCC under windows), performance seems poor */ #elif defined(_MSC_VER) # define XXH_rotl32(x,r) _rotl(x,r) # define XXH_rotl64(x,r) _rotl64(x,r) #else # define XXH_rotl32(x,r) (((x) << (r)) | ((x) >> (32 - (r)))) # define XXH_rotl64(x,r) (((x) << (r)) | ((x) >> (64 - (r)))) #endif /*! * @internal * @fn xxh_u32 XXH_swap32(xxh_u32 x) * @brief A 32-bit byteswap. * * @param x The 32-bit integer to byteswap. * @return @p x, byteswapped. */ #if defined(_MSC_VER) /* Visual Studio */ # define XXH_swap32 _byteswap_ulong #elif XXH_GCC_VERSION >= 403 # define XXH_swap32 __builtin_bswap32 #else static xxh_u32 XXH_swap32 (xxh_u32 x) { return ((x << 24) & 0xff000000 ) | ((x << 8) & 0x00ff0000 ) | ((x >> 8) & 0x0000ff00 ) | ((x >> 24) & 0x000000ff ); } #endif /* *************************** * Memory reads *****************************/ /*! * @internal * @brief Enum to indicate whether a pointer is aligned. */ typedef enum { XXH_aligned, /*!< Aligned */ XXH_unaligned /*!< Possibly unaligned */ } XXH_alignment; /* * XXH_FORCE_MEMORY_ACCESS==3 is an endian-independent byteshift load. * * This is ideal for older compilers which don't inline memcpy. */ #if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) XXH_FORCE_INLINE xxh_u32 XXH_readLE32(const void* memPtr) { const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; return bytePtr[0] | ((xxh_u32)bytePtr[1] << 8) | ((xxh_u32)bytePtr[2] << 16) | ((xxh_u32)bytePtr[3] << 24); } XXH_FORCE_INLINE xxh_u32 XXH_readBE32(const void* memPtr) { const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; return bytePtr[3] | ((xxh_u32)bytePtr[2] << 8) | ((xxh_u32)bytePtr[1] << 16) | ((xxh_u32)bytePtr[0] << 24); } #else XXH_FORCE_INLINE xxh_u32 XXH_readLE32(const void* ptr) { return XXH_CPU_LITTLE_ENDIAN ? XXH_read32(ptr) : XXH_swap32(XXH_read32(ptr)); } static xxh_u32 XXH_readBE32(const void* ptr) { return XXH_CPU_LITTLE_ENDIAN ? XXH_swap32(XXH_read32(ptr)) : XXH_read32(ptr); } #endif XXH_FORCE_INLINE xxh_u32 XXH_readLE32_align(const void* ptr, XXH_alignment align) { if (align==XXH_unaligned) { return XXH_readLE32(ptr); } else { return XXH_CPU_LITTLE_ENDIAN ? *(const xxh_u32*)ptr : XXH_swap32(*(const xxh_u32*)ptr); } } /* ************************************* * Misc ***************************************/ /*! @ingroup public */ XXH_PUBLIC_API unsigned XXH_versionNumber (void) { return XXH_VERSION_NUMBER; } /* ******************************************************************* * 32-bit hash functions *********************************************************************/ /*! * @} * @defgroup XXH32_impl XXH32 implementation * @ingroup impl * * Details on the XXH32 implementation. * @{ */ /* #define instead of static const, to be used as initializers */ #define XXH_PRIME32_1 0x9E3779B1U /*!< 0b10011110001101110111100110110001 */ #define XXH_PRIME32_2 0x85EBCA77U /*!< 0b10000101111010111100101001110111 */ #define XXH_PRIME32_3 0xC2B2AE3DU /*!< 0b11000010101100101010111000111101 */ #define XXH_PRIME32_4 0x27D4EB2FU /*!< 0b00100111110101001110101100101111 */ #define XXH_PRIME32_5 0x165667B1U /*!< 0b00010110010101100110011110110001 */ #ifdef XXH_OLD_NAMES # define PRIME32_1 XXH_PRIME32_1 # define PRIME32_2 XXH_PRIME32_2 # define PRIME32_3 XXH_PRIME32_3 # define PRIME32_4 XXH_PRIME32_4 # define PRIME32_5 XXH_PRIME32_5 #endif /*! * @internal * @brief Normal stripe processing routine. * * This shuffles the bits so that any bit from @p input impacts several bits in * @p acc. * * @param acc The accumulator lane. * @param input The stripe of input to mix. * @return The mixed accumulator lane. */ static xxh_u32 XXH32_round(xxh_u32 acc, xxh_u32 input) { acc += input * XXH_PRIME32_2; acc = XXH_rotl32(acc, 13); acc *= XXH_PRIME32_1; #if (defined(__SSE4_1__) || defined(__aarch64__) || defined(__wasm_simd128__)) && !defined(XXH_ENABLE_AUTOVECTORIZE) /* * UGLY HACK: * A compiler fence is used to prevent GCC and Clang from * autovectorizing the XXH32 loop (pragmas and attributes don't work for some * reason) without globally disabling SSE4.1. * * The reason we want to avoid vectorization is because despite working on * 4 integers at a time, there are multiple factors slowing XXH32 down on * SSE4: * - There's a ridiculous amount of lag from pmulld (10 cycles of latency on * newer chips!) making it slightly slower to multiply four integers at * once compared to four integers independently. Even when pmulld was * fastest, Sandy/Ivy Bridge, it is still not worth it to go into SSE * just to multiply unless doing a long operation. * * - Four instructions are required to rotate, * movqda tmp, v // not required with VEX encoding * pslld tmp, 13 // tmp <<= 13 * psrld v, 19 // x >>= 19 * por v, tmp // x |= tmp * compared to one for scalar: * roll v, 13 // reliably fast across the board * shldl v, v, 13 // Sandy Bridge and later prefer this for some reason * * - Instruction level parallelism is actually more beneficial here because * the SIMD actually serializes this operation: While v1 is rotating, v2 * can load data, while v3 can multiply. SSE forces them to operate * together. * * This is also enabled on AArch64, as Clang is *very aggressive* in vectorizing * the loop. NEON is only faster on the A53, and with the newer cores, it is less * than half the speed. * * Additionally, this is used on WASM SIMD128 because it JITs to the same * SIMD instructions and has the same issue. */ XXH_COMPILER_GUARD(acc); #endif return acc; } /*! * @internal * @brief Mixes all bits to finalize the hash. * * The final mix ensures that all input bits have a chance to impact any bit in * the output digest, resulting in an unbiased distribution. * * @param hash The hash to avalanche. * @return The avalanched hash. */ static xxh_u32 XXH32_avalanche(xxh_u32 hash) { hash ^= hash >> 15; hash *= XXH_PRIME32_2; hash ^= hash >> 13; hash *= XXH_PRIME32_3; hash ^= hash >> 16; return hash; } #define XXH_get32bits(p) XXH_readLE32_align(p, align) /*! * @internal * @brief Sets up the initial accumulator state for XXH32(). */ XXH_FORCE_INLINE void XXH32_initAccs(xxh_u32 *acc, xxh_u32 seed) { XXH_ASSERT(acc != NULL); acc[0] = seed + XXH_PRIME32_1 + XXH_PRIME32_2; acc[1] = seed + XXH_PRIME32_2; acc[2] = seed + 0; acc[3] = seed - XXH_PRIME32_1; } /*! * @internal * @brief Consumes a block of data for XXH32(). * * @return the end input pointer. */ XXH_FORCE_INLINE const xxh_u8 * XXH32_consumeLong( xxh_u32 *XXH_RESTRICT acc, xxh_u8 const *XXH_RESTRICT input, size_t len, XXH_alignment align ) { const xxh_u8* const bEnd = input + len; const xxh_u8* const limit = bEnd - 15; XXH_ASSERT(acc != NULL); XXH_ASSERT(input != NULL); XXH_ASSERT(len >= 16); do { acc[0] = XXH32_round(acc[0], XXH_get32bits(input)); input += 4; acc[1] = XXH32_round(acc[1], XXH_get32bits(input)); input += 4; acc[2] = XXH32_round(acc[2], XXH_get32bits(input)); input += 4; acc[3] = XXH32_round(acc[3], XXH_get32bits(input)); input += 4; } while (input < limit); return input; } /*! * @internal * @brief Merges the accumulator lanes together for XXH32() */ XXH_FORCE_INLINE XXH_PUREF xxh_u32 XXH32_mergeAccs(const xxh_u32 *acc) { XXH_ASSERT(acc != NULL); return XXH_rotl32(acc[0], 1) + XXH_rotl32(acc[1], 7) + XXH_rotl32(acc[2], 12) + XXH_rotl32(acc[3], 18); } /*! * @internal * @brief Processes the last 0-15 bytes of @p ptr. * * There may be up to 15 bytes remaining to consume from the input. * This final stage will digest them to ensure that all input bytes are present * in the final mix. * * @param hash The hash to finalize. * @param ptr The pointer to the remaining input. * @param len The remaining length, modulo 16. * @param align Whether @p ptr is aligned. * @return The finalized hash. * @see XXH64_finalize(). */ static XXH_PUREF xxh_u32 XXH32_finalize(xxh_u32 hash, const xxh_u8* ptr, size_t len, XXH_alignment align) { #define XXH_PROCESS1 do { \ hash += (*ptr++) * XXH_PRIME32_5; \ hash = XXH_rotl32(hash, 11) * XXH_PRIME32_1; \ } while (0) #define XXH_PROCESS4 do { \ hash += XXH_get32bits(ptr) * XXH_PRIME32_3; \ ptr += 4; \ hash = XXH_rotl32(hash, 17) * XXH_PRIME32_4; \ } while (0) if (ptr==NULL) XXH_ASSERT(len == 0); /* Compact rerolled version; generally faster */ if (!XXH32_ENDJMP) { len &= 15; while (len >= 4) { XXH_PROCESS4; len -= 4; } while (len > 0) { XXH_PROCESS1; --len; } return XXH32_avalanche(hash); } else { switch(len&15) /* or switch(bEnd - p) */ { case 12: XXH_PROCESS4; XXH_FALLTHROUGH; /* fallthrough */ case 8: XXH_PROCESS4; XXH_FALLTHROUGH; /* fallthrough */ case 4: XXH_PROCESS4; return XXH32_avalanche(hash); case 13: XXH_PROCESS4; XXH_FALLTHROUGH; /* fallthrough */ case 9: XXH_PROCESS4; XXH_FALLTHROUGH; /* fallthrough */ case 5: XXH_PROCESS4; XXH_PROCESS1; return XXH32_avalanche(hash); case 14: XXH_PROCESS4; XXH_FALLTHROUGH; /* fallthrough */ case 10: XXH_PROCESS4; XXH_FALLTHROUGH; /* fallthrough */ case 6: XXH_PROCESS4; XXH_PROCESS1; XXH_PROCESS1; return XXH32_avalanche(hash); case 15: XXH_PROCESS4; XXH_FALLTHROUGH; /* fallthrough */ case 11: XXH_PROCESS4; XXH_FALLTHROUGH; /* fallthrough */ case 7: XXH_PROCESS4; XXH_FALLTHROUGH; /* fallthrough */ case 3: XXH_PROCESS1; XXH_FALLTHROUGH; /* fallthrough */ case 2: XXH_PROCESS1; XXH_FALLTHROUGH; /* fallthrough */ case 1: XXH_PROCESS1; XXH_FALLTHROUGH; /* fallthrough */ case 0: return XXH32_avalanche(hash); } XXH_ASSERT(0); return hash; /* reaching this point is deemed impossible */ } } #ifdef XXH_OLD_NAMES # define PROCESS1 XXH_PROCESS1 # define PROCESS4 XXH_PROCESS4 #else # undef XXH_PROCESS1 # undef XXH_PROCESS4 #endif /*! * @internal * @brief The implementation for @ref XXH32(). * * @param input , len , seed Directly passed from @ref XXH32(). * @param align Whether @p input is aligned. * @return The calculated hash. */ XXH_FORCE_INLINE XXH_PUREF xxh_u32 XXH32_endian_align(const xxh_u8* input, size_t len, xxh_u32 seed, XXH_alignment align) { xxh_u32 h32; if (input==NULL) XXH_ASSERT(len == 0); if (len>=16) { xxh_u32 acc[4]; XXH32_initAccs(acc, seed); input = XXH32_consumeLong(acc, input, len, align); h32 = XXH32_mergeAccs(acc); } else { h32 = seed + XXH_PRIME32_5; } h32 += (xxh_u32)len; return XXH32_finalize(h32, input, len&15, align); } /*! @ingroup XXH32_family */ XXH_PUBLIC_API XXH32_hash_t XXH32 (const void* input, size_t len, XXH32_hash_t seed) { #if !defined(XXH_NO_STREAM) && XXH_SIZE_OPT >= 2 /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ XXH32_state_t state; XXH32_reset(&state, seed); XXH32_update(&state, (const xxh_u8*)input, len); return XXH32_digest(&state); #else if (XXH_FORCE_ALIGN_CHECK) { if ((((size_t)input) & 3) == 0) { /* Input is 4-bytes aligned, leverage the speed benefit */ return XXH32_endian_align((const xxh_u8*)input, len, seed, XXH_aligned); } } return XXH32_endian_align((const xxh_u8*)input, len, seed, XXH_unaligned); #endif } /******* Hash streaming *******/ #ifndef XXH_NO_STREAM /*! @ingroup XXH32_family */ XXH_PUBLIC_API XXH32_state_t* XXH32_createState(void) { return (XXH32_state_t*)XXH_malloc(sizeof(XXH32_state_t)); } /*! @ingroup XXH32_family */ XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr) { XXH_free(statePtr); return XXH_OK; } /*! @ingroup XXH32_family */ XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dstState, const XXH32_state_t* srcState) { XXH_memcpy(dstState, srcState, sizeof(*dstState)); } /*! @ingroup XXH32_family */ XXH_PUBLIC_API XXH_errorcode XXH32_reset(XXH32_state_t* statePtr, XXH32_hash_t seed) { XXH_ASSERT(statePtr != NULL); memset(statePtr, 0, sizeof(*statePtr)); XXH32_initAccs(statePtr->acc, seed); return XXH_OK; } /*! @ingroup XXH32_family */ XXH_PUBLIC_API XXH_errorcode XXH32_update(XXH32_state_t* state, const void* input, size_t len) { if (input==NULL) { XXH_ASSERT(len == 0); return XXH_OK; } state->total_len_32 += (XXH32_hash_t)len; state->large_len |= (XXH32_hash_t)((len>=16) | (state->total_len_32>=16)); XXH_ASSERT(state->bufferedSize < sizeof(state->buffer)); if (len < sizeof(state->buffer) - state->bufferedSize) { /* fill in tmp buffer */ XXH_memcpy(state->buffer + state->bufferedSize, input, len); state->bufferedSize += (XXH32_hash_t)len; return XXH_OK; } { const xxh_u8* xinput = (const xxh_u8*)input; const xxh_u8* const bEnd = xinput + len; if (state->bufferedSize) { /* non-empty buffer: complete first */ XXH_memcpy(state->buffer + state->bufferedSize, xinput, sizeof(state->buffer) - state->bufferedSize); xinput += sizeof(state->buffer) - state->bufferedSize; /* then process one round */ (void)XXH32_consumeLong(state->acc, state->buffer, sizeof(state->buffer), XXH_aligned); state->bufferedSize = 0; } XXH_ASSERT(xinput <= bEnd); if ((size_t)(bEnd - xinput) >= sizeof(state->buffer)) { /* Process the remaining data */ xinput = XXH32_consumeLong(state->acc, xinput, (size_t)(bEnd - xinput), XXH_unaligned); } if (xinput < bEnd) { /* Copy the leftover to the tmp buffer */ XXH_memcpy(state->buffer, xinput, (size_t)(bEnd-xinput)); state->bufferedSize = (unsigned)(bEnd-xinput); } } return XXH_OK; } /*! @ingroup XXH32_family */ XXH_PUBLIC_API XXH32_hash_t XXH32_digest(const XXH32_state_t* state) { xxh_u32 h32; if (state->large_len) { h32 = XXH32_mergeAccs(state->acc); } else { h32 = state->acc[2] /* == seed */ + XXH_PRIME32_5; } h32 += state->total_len_32; return XXH32_finalize(h32, state->buffer, state->bufferedSize, XXH_aligned); } #endif /* !XXH_NO_STREAM */ /******* Canonical representation *******/ /*! @ingroup XXH32_family */ XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash) { XXH_STATIC_ASSERT(sizeof(XXH32_canonical_t) == sizeof(XXH32_hash_t)); if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap32(hash); XXH_memcpy(dst, &hash, sizeof(*dst)); } /*! @ingroup XXH32_family */ XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src) { return XXH_readBE32(src); } #ifndef XXH_NO_LONG_LONG /* ******************************************************************* * 64-bit hash functions *********************************************************************/ /*! * @} * @ingroup impl * @{ */ /******* Memory access *******/ typedef XXH64_hash_t xxh_u64; #ifdef XXH_OLD_NAMES # define U64 xxh_u64 #endif #if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) /* * Manual byteshift. Best for old compilers which don't inline memcpy. * We actually directly use XXH_readLE64 and XXH_readBE64. */ #elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2)) /* Force direct memory access. Only works on CPU which support unaligned memory access in hardware */ static xxh_u64 XXH_read64(const void* memPtr) { return *(const xxh_u64*) memPtr; } #elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1)) /* * __attribute__((aligned(1))) is supported by gcc and clang. Originally the * documentation claimed that it only increased the alignment, but actually it * can decrease it on gcc, clang, and icc: * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69502, * https://gcc.godbolt.org/z/xYez1j67Y. */ #ifdef XXH_OLD_NAMES typedef union { xxh_u32 u32; xxh_u64 u64; } __attribute__((__packed__)) unalign64; #endif static xxh_u64 XXH_read64(const void* ptr) { typedef __attribute__((__aligned__(1))) xxh_u64 xxh_unalign64; return *((const xxh_unalign64*)ptr); } #else /* * Portable and safe solution. Generally efficient. * see: https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html */ static xxh_u64 XXH_read64(const void* memPtr) { xxh_u64 val; XXH_memcpy(&val, memPtr, sizeof(val)); return val; } #endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */ #if defined(_MSC_VER) /* Visual Studio */ # define XXH_swap64 _byteswap_uint64 #elif XXH_GCC_VERSION >= 403 # define XXH_swap64 __builtin_bswap64 #else static xxh_u64 XXH_swap64(xxh_u64 x) { return ((x << 56) & 0xff00000000000000ULL) | ((x << 40) & 0x00ff000000000000ULL) | ((x << 24) & 0x0000ff0000000000ULL) | ((x << 8) & 0x000000ff00000000ULL) | ((x >> 8) & 0x00000000ff000000ULL) | ((x >> 24) & 0x0000000000ff0000ULL) | ((x >> 40) & 0x000000000000ff00ULL) | ((x >> 56) & 0x00000000000000ffULL); } #endif /* XXH_FORCE_MEMORY_ACCESS==3 is an endian-independent byteshift load. */ #if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) XXH_FORCE_INLINE xxh_u64 XXH_readLE64(const void* memPtr) { const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; return bytePtr[0] | ((xxh_u64)bytePtr[1] << 8) | ((xxh_u64)bytePtr[2] << 16) | ((xxh_u64)bytePtr[3] << 24) | ((xxh_u64)bytePtr[4] << 32) | ((xxh_u64)bytePtr[5] << 40) | ((xxh_u64)bytePtr[6] << 48) | ((xxh_u64)bytePtr[7] << 56); } XXH_FORCE_INLINE xxh_u64 XXH_readBE64(const void* memPtr) { const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; return bytePtr[7] | ((xxh_u64)bytePtr[6] << 8) | ((xxh_u64)bytePtr[5] << 16) | ((xxh_u64)bytePtr[4] << 24) | ((xxh_u64)bytePtr[3] << 32) | ((xxh_u64)bytePtr[2] << 40) | ((xxh_u64)bytePtr[1] << 48) | ((xxh_u64)bytePtr[0] << 56); } #else XXH_FORCE_INLINE xxh_u64 XXH_readLE64(const void* ptr) { return XXH_CPU_LITTLE_ENDIAN ? XXH_read64(ptr) : XXH_swap64(XXH_read64(ptr)); } static xxh_u64 XXH_readBE64(const void* ptr) { return XXH_CPU_LITTLE_ENDIAN ? XXH_swap64(XXH_read64(ptr)) : XXH_read64(ptr); } #endif XXH_FORCE_INLINE xxh_u64 XXH_readLE64_align(const void* ptr, XXH_alignment align) { if (align==XXH_unaligned) return XXH_readLE64(ptr); else return XXH_CPU_LITTLE_ENDIAN ? *(const xxh_u64*)ptr : XXH_swap64(*(const xxh_u64*)ptr); } /******* xxh64 *******/ /*! * @} * @defgroup XXH64_impl XXH64 implementation * @ingroup impl * * Details on the XXH64 implementation. * @{ */ /* #define rather that static const, to be used as initializers */ #define XXH_PRIME64_1 0x9E3779B185EBCA87ULL /*!< 0b1001111000110111011110011011000110000101111010111100101010000111 */ #define XXH_PRIME64_2 0xC2B2AE3D27D4EB4FULL /*!< 0b1100001010110010101011100011110100100111110101001110101101001111 */ #define XXH_PRIME64_3 0x165667B19E3779F9ULL /*!< 0b0001011001010110011001111011000110011110001101110111100111111001 */ #define XXH_PRIME64_4 0x85EBCA77C2B2AE63ULL /*!< 0b1000010111101011110010100111011111000010101100101010111001100011 */ #define XXH_PRIME64_5 0x27D4EB2F165667C5ULL /*!< 0b0010011111010100111010110010111100010110010101100110011111000101 */ #ifdef XXH_OLD_NAMES # define PRIME64_1 XXH_PRIME64_1 # define PRIME64_2 XXH_PRIME64_2 # define PRIME64_3 XXH_PRIME64_3 # define PRIME64_4 XXH_PRIME64_4 # define PRIME64_5 XXH_PRIME64_5 #endif /*! @copydoc XXH32_round */ static xxh_u64 XXH64_round(xxh_u64 acc, xxh_u64 input) { acc += input * XXH_PRIME64_2; acc = XXH_rotl64(acc, 31); acc *= XXH_PRIME64_1; #if (defined(__AVX512F__)) && !defined(XXH_ENABLE_AUTOVECTORIZE) /* * DISABLE AUTOVECTORIZATION: * A compiler fence is used to prevent GCC and Clang from * autovectorizing the XXH64 loop (pragmas and attributes don't work for some * reason) without globally disabling AVX512. * * Autovectorization of XXH64 tends to be detrimental, * though the exact outcome may change depending on exact cpu and compiler version. * For information, it has been reported as detrimental for Skylake-X, * but possibly beneficial for Zen4. * * The default is to disable auto-vectorization, * but you can select to enable it instead using `XXH_ENABLE_AUTOVECTORIZE` build variable. */ XXH_COMPILER_GUARD(acc); #endif return acc; } static xxh_u64 XXH64_mergeRound(xxh_u64 acc, xxh_u64 val) { val = XXH64_round(0, val); acc ^= val; acc = acc * XXH_PRIME64_1 + XXH_PRIME64_4; return acc; } /*! @copydoc XXH32_avalanche */ static xxh_u64 XXH64_avalanche(xxh_u64 hash) { hash ^= hash >> 33; hash *= XXH_PRIME64_2; hash ^= hash >> 29; hash *= XXH_PRIME64_3; hash ^= hash >> 32; return hash; } #define XXH_get64bits(p) XXH_readLE64_align(p, align) /*! * @internal * @brief Sets up the initial accumulator state for XXH64(). */ XXH_FORCE_INLINE void XXH64_initAccs(xxh_u64 *acc, xxh_u64 seed) { XXH_ASSERT(acc != NULL); acc[0] = seed + XXH_PRIME64_1 + XXH_PRIME64_2; acc[1] = seed + XXH_PRIME64_2; acc[2] = seed + 0; acc[3] = seed - XXH_PRIME64_1; } /*! * @internal * @brief Consumes a block of data for XXH64(). * * @return the end input pointer. */ XXH_FORCE_INLINE const xxh_u8 * XXH64_consumeLong( xxh_u64 *XXH_RESTRICT acc, xxh_u8 const *XXH_RESTRICT input, size_t len, XXH_alignment align ) { const xxh_u8* const bEnd = input + len; const xxh_u8* const limit = bEnd - 31; XXH_ASSERT(acc != NULL); XXH_ASSERT(input != NULL); XXH_ASSERT(len >= 32); do { /* reroll on 32-bit */ if (sizeof(void *) < sizeof(xxh_u64)) { size_t i; for (i = 0; i < 4; i++) { acc[i] = XXH64_round(acc[i], XXH_get64bits(input)); input += 8; } } else { acc[0] = XXH64_round(acc[0], XXH_get64bits(input)); input += 8; acc[1] = XXH64_round(acc[1], XXH_get64bits(input)); input += 8; acc[2] = XXH64_round(acc[2], XXH_get64bits(input)); input += 8; acc[3] = XXH64_round(acc[3], XXH_get64bits(input)); input += 8; } } while (input < limit); return input; } /*! * @internal * @brief Merges the accumulator lanes together for XXH64() */ XXH_FORCE_INLINE XXH_PUREF xxh_u64 XXH64_mergeAccs(const xxh_u64 *acc) { XXH_ASSERT(acc != NULL); { xxh_u64 h64 = XXH_rotl64(acc[0], 1) + XXH_rotl64(acc[1], 7) + XXH_rotl64(acc[2], 12) + XXH_rotl64(acc[3], 18); /* reroll on 32-bit */ if (sizeof(void *) < sizeof(xxh_u64)) { size_t i; for (i = 0; i < 4; i++) { h64 = XXH64_mergeRound(h64, acc[i]); } } else { h64 = XXH64_mergeRound(h64, acc[0]); h64 = XXH64_mergeRound(h64, acc[1]); h64 = XXH64_mergeRound(h64, acc[2]); h64 = XXH64_mergeRound(h64, acc[3]); } return h64; } } /*! * @internal * @brief Processes the last 0-31 bytes of @p ptr. * * There may be up to 31 bytes remaining to consume from the input. * This final stage will digest them to ensure that all input bytes are present * in the final mix. * * @param hash The hash to finalize. * @param ptr The pointer to the remaining input. * @param len The remaining length, modulo 32. * @param align Whether @p ptr is aligned. * @return The finalized hash * @see XXH32_finalize(). */ XXH_STATIC XXH_PUREF xxh_u64 XXH64_finalize(xxh_u64 hash, const xxh_u8* ptr, size_t len, XXH_alignment align) { if (ptr==NULL) XXH_ASSERT(len == 0); len &= 31; while (len >= 8) { xxh_u64 const k1 = XXH64_round(0, XXH_get64bits(ptr)); ptr += 8; hash ^= k1; hash = XXH_rotl64(hash,27) * XXH_PRIME64_1 + XXH_PRIME64_4; len -= 8; } if (len >= 4) { hash ^= (xxh_u64)(XXH_get32bits(ptr)) * XXH_PRIME64_1; ptr += 4; hash = XXH_rotl64(hash, 23) * XXH_PRIME64_2 + XXH_PRIME64_3; len -= 4; } while (len > 0) { hash ^= (*ptr++) * XXH_PRIME64_5; hash = XXH_rotl64(hash, 11) * XXH_PRIME64_1; --len; } return XXH64_avalanche(hash); } #ifdef XXH_OLD_NAMES # define PROCESS1_64 XXH_PROCESS1_64 # define PROCESS4_64 XXH_PROCESS4_64 # define PROCESS8_64 XXH_PROCESS8_64 #else # undef XXH_PROCESS1_64 # undef XXH_PROCESS4_64 # undef XXH_PROCESS8_64 #endif /*! * @internal * @brief The implementation for @ref XXH64(). * * @param input , len , seed Directly passed from @ref XXH64(). * @param align Whether @p input is aligned. * @return The calculated hash. */ XXH_FORCE_INLINE XXH_PUREF xxh_u64 XXH64_endian_align(const xxh_u8* input, size_t len, xxh_u64 seed, XXH_alignment align) { xxh_u64 h64; if (input==NULL) XXH_ASSERT(len == 0); if (len>=32) { /* Process a large block of data */ xxh_u64 acc[4]; XXH64_initAccs(acc, seed); input = XXH64_consumeLong(acc, input, len, align); h64 = XXH64_mergeAccs(acc); } else { h64 = seed + XXH_PRIME64_5; } h64 += (xxh_u64) len; return XXH64_finalize(h64, input, len, align); } /*! @ingroup XXH64_family */ XXH_PUBLIC_API XXH64_hash_t XXH64 (XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) { #if !defined(XXH_NO_STREAM) && XXH_SIZE_OPT >= 2 /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ XXH64_state_t state; XXH64_reset(&state, seed); XXH64_update(&state, (const xxh_u8*)input, len); return XXH64_digest(&state); #else if (XXH_FORCE_ALIGN_CHECK) { if ((((size_t)input) & 7)==0) { /* Input is aligned, let's leverage the speed advantage */ return XXH64_endian_align((const xxh_u8*)input, len, seed, XXH_aligned); } } return XXH64_endian_align((const xxh_u8*)input, len, seed, XXH_unaligned); #endif } /******* Hash Streaming *******/ #ifndef XXH_NO_STREAM /*! @ingroup XXH64_family*/ XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void) { return (XXH64_state_t*)XXH_malloc(sizeof(XXH64_state_t)); } /*! @ingroup XXH64_family */ XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr) { XXH_free(statePtr); return XXH_OK; } /*! @ingroup XXH64_family */ XXH_PUBLIC_API void XXH64_copyState(XXH_NOESCAPE XXH64_state_t* dstState, const XXH64_state_t* srcState) { XXH_memcpy(dstState, srcState, sizeof(*dstState)); } /*! @ingroup XXH64_family */ XXH_PUBLIC_API XXH_errorcode XXH64_reset(XXH_NOESCAPE XXH64_state_t* statePtr, XXH64_hash_t seed) { XXH_ASSERT(statePtr != NULL); memset(statePtr, 0, sizeof(*statePtr)); XXH64_initAccs(statePtr->acc, seed); return XXH_OK; } /*! @ingroup XXH64_family */ XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH_NOESCAPE XXH64_state_t* state, XXH_NOESCAPE const void* input, size_t len) { if (input==NULL) { XXH_ASSERT(len == 0); return XXH_OK; } state->total_len += len; XXH_ASSERT(state->bufferedSize <= sizeof(state->buffer)); if (len < sizeof(state->buffer) - state->bufferedSize) { /* fill in tmp buffer */ XXH_memcpy(state->buffer + state->bufferedSize, input, len); state->bufferedSize += (XXH32_hash_t)len; return XXH_OK; } { const xxh_u8* xinput = (const xxh_u8*)input; const xxh_u8* const bEnd = xinput + len; if (state->bufferedSize) { /* non-empty buffer => complete first */ XXH_memcpy(state->buffer + state->bufferedSize, xinput, sizeof(state->buffer) - state->bufferedSize); xinput += sizeof(state->buffer) - state->bufferedSize; /* and process one round */ (void)XXH64_consumeLong(state->acc, state->buffer, sizeof(state->buffer), XXH_aligned); state->bufferedSize = 0; } XXH_ASSERT(xinput <= bEnd); if ((size_t)(bEnd - xinput) >= sizeof(state->buffer)) { /* Process the remaining data */ xinput = XXH64_consumeLong(state->acc, xinput, (size_t)(bEnd - xinput), XXH_unaligned); } if (xinput < bEnd) { /* Copy the leftover to the tmp buffer */ XXH_memcpy(state->buffer, xinput, (size_t)(bEnd-xinput)); state->bufferedSize = (unsigned)(bEnd-xinput); } } return XXH_OK; } /*! @ingroup XXH64_family */ XXH_PUBLIC_API XXH64_hash_t XXH64_digest(XXH_NOESCAPE const XXH64_state_t* state) { xxh_u64 h64; if (state->total_len >= 32) { h64 = XXH64_mergeAccs(state->acc); } else { h64 = state->acc[2] /*seed*/ + XXH_PRIME64_5; } h64 += (xxh_u64) state->total_len; return XXH64_finalize(h64, state->buffer, (size_t)state->total_len, XXH_aligned); } #endif /* !XXH_NO_STREAM */ /******* Canonical representation *******/ /*! @ingroup XXH64_family */ XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH_NOESCAPE XXH64_canonical_t* dst, XXH64_hash_t hash) { XXH_STATIC_ASSERT(sizeof(XXH64_canonical_t) == sizeof(XXH64_hash_t)); if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap64(hash); XXH_memcpy(dst, &hash, sizeof(*dst)); } /*! @ingroup XXH64_family */ XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(XXH_NOESCAPE const XXH64_canonical_t* src) { return XXH_readBE64(src); } #ifndef XXH_NO_XXH3 /* ********************************************************************* * XXH3 * New generation hash designed for speed on small keys and vectorization ************************************************************************ */ /*! * @} * @defgroup XXH3_impl XXH3 implementation * @ingroup impl * @{ */ /* === Compiler specifics === */ #if (defined(__GNUC__) && (__GNUC__ >= 3)) \ || (defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 800)) \ || defined(__clang__) # define XXH_likely(x) __builtin_expect(x, 1) # define XXH_unlikely(x) __builtin_expect(x, 0) #else # define XXH_likely(x) (x) # define XXH_unlikely(x) (x) #endif #ifndef XXH_HAS_INCLUDE # ifdef __has_include /* * Not defined as XXH_HAS_INCLUDE(x) (function-like) because * this causes segfaults in Apple Clang 4.2 (on Mac OS X 10.7 Lion) */ # define XXH_HAS_INCLUDE __has_include # else # define XXH_HAS_INCLUDE(x) 0 # endif #endif #if defined(__GNUC__) || defined(__clang__) # if defined(__ARM_FEATURE_SVE) # include # endif # if defined(__ARM_NEON__) || defined(__ARM_NEON) \ || (defined(_M_ARM) && _M_ARM >= 7) \ || defined(_M_ARM64) || defined(_M_ARM64EC) \ || (defined(__wasm_simd128__) && XXH_HAS_INCLUDE()) /* WASM SIMD128 via SIMDe */ # define inline __inline__ /* circumvent a clang bug */ # include # undef inline # elif defined(__AVX2__) # include # elif defined(__SSE2__) # include # elif defined(__loongarch_sx) # include # endif #endif #if defined(_MSC_VER) # include #endif /* * One goal of XXH3 is to make it fast on both 32-bit and 64-bit, while * remaining a true 64-bit/128-bit hash function. * * This is done by prioritizing a subset of 64-bit operations that can be * emulated without too many steps on the average 32-bit machine. * * For example, these two lines seem similar, and run equally fast on 64-bit: * * xxh_u64 x; * x ^= (x >> 47); // good * x ^= (x >> 13); // bad * * However, to a 32-bit machine, there is a major difference. * * x ^= (x >> 47) looks like this: * * x.lo ^= (x.hi >> (47 - 32)); * * while x ^= (x >> 13) looks like this: * * // note: funnel shifts are not usually cheap. * x.lo ^= (x.lo >> 13) | (x.hi << (32 - 13)); * x.hi ^= (x.hi >> 13); * * The first one is significantly faster than the second, simply because the * shift is larger than 32. This means: * - All the bits we need are in the upper 32 bits, so we can ignore the lower * 32 bits in the shift. * - The shift result will always fit in the lower 32 bits, and therefore, * we can ignore the upper 32 bits in the xor. * * Thanks to this optimization, XXH3 only requires these features to be efficient: * * - Usable unaligned access * - A 32-bit or 64-bit ALU * - If 32-bit, a decent ADC instruction * - A 32 or 64-bit multiply with a 64-bit result * - For the 128-bit variant, a decent byteswap helps short inputs. * * The first two are already required by XXH32, and almost all 32-bit and 64-bit * platforms which can run XXH32 can run XXH3 efficiently. * * Thumb-1, the classic 16-bit only subset of ARM's instruction set, is one * notable exception. * * First of all, Thumb-1 lacks support for the UMULL instruction which * performs the important long multiply. This means numerous __aeabi_lmul * calls. * * Second of all, the 8 functional registers are just not enough. * Setup for __aeabi_lmul, byteshift loads, pointers, and all arithmetic need * Lo registers, and this shuffling results in thousands more MOVs than A32. * * A32 and T32 don't have this limitation. They can access all 14 registers, * do a 32->64 multiply with UMULL, and the flexible operand allowing free * shifts is helpful, too. * * Therefore, we do a quick sanity check. * * If compiling Thumb-1 for a target which supports ARM instructions, we will * emit a warning, as it is not a "sane" platform to compile for. * * Usually, if this happens, it is because of an accident and you probably need * to specify -march, as you likely meant to compile for a newer architecture. * * Credit: large sections of the vectorial and asm source code paths * have been contributed by @easyaspi314 */ #if defined(__thumb__) && !defined(__thumb2__) && defined(__ARM_ARCH_ISA_ARM) # warning "XXH3 is highly inefficient without ARM or Thumb-2." #endif /* ========================================== * Vectorization detection * ========================================== */ #ifdef XXH_DOXYGEN /*! * @ingroup tuning * @brief Overrides the vectorization implementation chosen for XXH3. * * Can be defined to 0 to disable SIMD or any of the values mentioned in * @ref XXH_VECTOR_TYPE. * * If this is not defined, it uses predefined macros to determine the best * implementation. */ # define XXH_VECTOR XXH_SCALAR /*! * @ingroup tuning * @brief Selects the minimum alignment for XXH3's accumulators. * * When using SIMD, this should match the alignment required for said vector * type, so, for example, 32 for AVX2. * * Default: Auto detected. */ # define XXH_ACC_ALIGN 8 #endif /* Actual definition */ #ifndef XXH_DOXYGEN #endif #ifndef XXH_VECTOR /* can be defined on command line */ # if defined(__ARM_FEATURE_SVE) # define XXH_VECTOR XXH_SVE # elif ( \ defined(__ARM_NEON__) || defined(__ARM_NEON) /* gcc */ \ || defined(_M_ARM) || defined(_M_ARM64) || defined(_M_ARM64EC) /* msvc */ \ || (defined(__wasm_simd128__) && XXH_HAS_INCLUDE()) /* wasm simd128 via SIMDe */ \ ) && ( \ defined(_WIN32) || defined(__LITTLE_ENDIAN__) /* little endian only */ \ || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) \ ) # define XXH_VECTOR XXH_NEON # elif defined(__AVX512F__) # define XXH_VECTOR XXH_AVX512 # elif defined(__AVX2__) # define XXH_VECTOR XXH_AVX2 # elif defined(__SSE2__) || defined(_M_AMD64) || defined(_M_X64) || (defined(_M_IX86_FP) && (_M_IX86_FP == 2)) # define XXH_VECTOR XXH_SSE2 # elif (defined(__PPC64__) && defined(__POWER8_VECTOR__)) \ || (defined(__s390x__) && defined(__VEC__)) \ && defined(__GNUC__) /* TODO: IBM XL */ # define XXH_VECTOR XXH_VSX # elif defined(__loongarch_sx) # define XXH_VECTOR XXH_LSX # else # define XXH_VECTOR XXH_SCALAR # endif #endif /* __ARM_FEATURE_SVE is only supported by GCC & Clang. */ #if (XXH_VECTOR == XXH_SVE) && !defined(__ARM_FEATURE_SVE) # ifdef _MSC_VER # pragma warning(once : 4606) # else # warning "__ARM_FEATURE_SVE isn't supported. Use SCALAR instead." # endif # undef XXH_VECTOR # define XXH_VECTOR XXH_SCALAR #endif /* * Controls the alignment of the accumulator, * for compatibility with aligned vector loads, which are usually faster. */ #ifndef XXH_ACC_ALIGN # if defined(XXH_X86DISPATCH) # define XXH_ACC_ALIGN 64 /* for compatibility with avx512 */ # elif XXH_VECTOR == XXH_SCALAR /* scalar */ # define XXH_ACC_ALIGN 8 # elif XXH_VECTOR == XXH_SSE2 /* sse2 */ # define XXH_ACC_ALIGN 16 # elif XXH_VECTOR == XXH_AVX2 /* avx2 */ # define XXH_ACC_ALIGN 32 # elif XXH_VECTOR == XXH_NEON /* neon */ # define XXH_ACC_ALIGN 16 # elif XXH_VECTOR == XXH_VSX /* vsx */ # define XXH_ACC_ALIGN 16 # elif XXH_VECTOR == XXH_AVX512 /* avx512 */ # define XXH_ACC_ALIGN 64 # elif XXH_VECTOR == XXH_SVE /* sve */ # define XXH_ACC_ALIGN 64 # elif XXH_VECTOR == XXH_LSX /* lsx */ # define XXH_ACC_ALIGN 64 # endif #endif #if defined(XXH_X86DISPATCH) || XXH_VECTOR == XXH_SSE2 \ || XXH_VECTOR == XXH_AVX2 || XXH_VECTOR == XXH_AVX512 # define XXH_SEC_ALIGN XXH_ACC_ALIGN #elif XXH_VECTOR == XXH_SVE # define XXH_SEC_ALIGN XXH_ACC_ALIGN #else # define XXH_SEC_ALIGN 8 #endif #if defined(__GNUC__) || defined(__clang__) # define XXH_ALIASING __attribute__((__may_alias__)) #else # define XXH_ALIASING /* nothing */ #endif /* * UGLY HACK: * GCC usually generates the best code with -O3 for xxHash. * * However, when targeting AVX2, it is overzealous in its unrolling resulting * in code roughly 3/4 the speed of Clang. * * There are other issues, such as GCC splitting _mm256_loadu_si256 into * _mm_loadu_si128 + _mm256_inserti128_si256. This is an optimization which * only applies to Sandy and Ivy Bridge... which don't even support AVX2. * * That is why when compiling the AVX2 version, it is recommended to use either * -O2 -mavx2 -march=haswell * or * -O2 -mavx2 -mno-avx256-split-unaligned-load * for decent performance, or to use Clang instead. * * Fortunately, we can control the first one with a pragma that forces GCC into * -O2, but the other one we can't control without "failed to inline always * inline function due to target mismatch" warnings. */ #if XXH_VECTOR == XXH_AVX2 /* AVX2 */ \ && defined(__GNUC__) && !defined(__clang__) /* GCC, not Clang */ \ && defined(__OPTIMIZE__) && XXH_SIZE_OPT <= 0 /* respect -O0 and -Os */ # pragma GCC push_options # pragma GCC optimize("-O2") #endif #if XXH_VECTOR == XXH_NEON /* * UGLY HACK: While AArch64 GCC on Linux does not seem to care, on macOS, GCC -O3 * optimizes out the entire hashLong loop because of the aliasing violation. * * However, GCC is also inefficient at load-store optimization with vld1q/vst1q, * so the only option is to mark it as aliasing. */ typedef uint64x2_t xxh_aliasing_uint64x2_t XXH_ALIASING; /*! * @internal * @brief `vld1q_u64` but faster and alignment-safe. * * On AArch64, unaligned access is always safe, but on ARMv7-a, it is only * *conditionally* safe (`vld1` has an alignment bit like `movdq[ua]` in x86). * * GCC for AArch64 sees `vld1q_u8` as an intrinsic instead of a load, so it * prohibits load-store optimizations. Therefore, a direct dereference is used. * * Otherwise, `vld1q_u8` is used with `vreinterpretq_u8_u64` to do a safe * unaligned load. */ #if defined(__aarch64__) && defined(__GNUC__) && !defined(__clang__) XXH_FORCE_INLINE uint64x2_t XXH_vld1q_u64(void const* ptr) /* silence -Wcast-align */ { return *(xxh_aliasing_uint64x2_t const *)ptr; } #else XXH_FORCE_INLINE uint64x2_t XXH_vld1q_u64(void const* ptr) { return vreinterpretq_u64_u8(vld1q_u8((uint8_t const*)ptr)); } #endif /*! * @internal * @brief `vmlal_u32` on low and high halves of a vector. * * This is a workaround for AArch64 GCC < 11 which implemented arm_neon.h with * inline assembly and were therefore incapable of merging the `vget_{low, high}_u32` * with `vmlal_u32`. */ #if defined(__aarch64__) && defined(__GNUC__) && !defined(__clang__) && __GNUC__ < 11 XXH_FORCE_INLINE uint64x2_t XXH_vmlal_low_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) { /* Inline assembly is the only way */ __asm__("umlal %0.2d, %1.2s, %2.2s" : "+w" (acc) : "w" (lhs), "w" (rhs)); return acc; } XXH_FORCE_INLINE uint64x2_t XXH_vmlal_high_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) { /* This intrinsic works as expected */ return vmlal_high_u32(acc, lhs, rhs); } #else /* Portable intrinsic versions */ XXH_FORCE_INLINE uint64x2_t XXH_vmlal_low_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) { return vmlal_u32(acc, vget_low_u32(lhs), vget_low_u32(rhs)); } /*! @copydoc XXH_vmlal_low_u32 * Assume the compiler converts this to vmlal_high_u32 on aarch64 */ XXH_FORCE_INLINE uint64x2_t XXH_vmlal_high_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) { return vmlal_u32(acc, vget_high_u32(lhs), vget_high_u32(rhs)); } #endif /*! * @ingroup tuning * @brief Controls the NEON to scalar ratio for XXH3 * * This can be set to 2, 4, 6, or 8. * * ARM Cortex CPUs are _very_ sensitive to how their pipelines are used. * * For example, the Cortex-A73 can dispatch 3 micro-ops per cycle, but only 2 of those * can be NEON. If you are only using NEON instructions, you are only using 2/3 of the CPU * bandwidth. * * This is even more noticeable on the more advanced cores like the Cortex-A76 which * can dispatch 8 micro-ops per cycle, but still only 2 NEON micro-ops at once. * * Therefore, to make the most out of the pipeline, it is beneficial to run 6 NEON lanes * and 2 scalar lanes, which is chosen by default. * * This does not apply to Apple processors or 32-bit processors, which run better with * full NEON. These will default to 8. Additionally, size-optimized builds run 8 lanes. * * This change benefits CPUs with large micro-op buffers without negatively affecting * most other CPUs: * * | Chipset | Dispatch type | NEON only | 6:2 hybrid | Diff. | * |:----------------------|:--------------------|----------:|-----------:|------:| * | Snapdragon 730 (A76) | 2 NEON/8 micro-ops | 8.8 GB/s | 10.1 GB/s | ~16% | * | Snapdragon 835 (A73) | 2 NEON/3 micro-ops | 5.1 GB/s | 5.3 GB/s | ~5% | * | Marvell PXA1928 (A53) | In-order dual-issue | 1.9 GB/s | 1.9 GB/s | 0% | * | Apple M1 | 4 NEON/8 micro-ops | 37.3 GB/s | 36.1 GB/s | ~-3% | * * It also seems to fix some bad codegen on GCC, making it almost as fast as clang. * * When using WASM SIMD128, if this is 2 or 6, SIMDe will scalarize 2 of the lanes meaning * it effectively becomes worse 4. * * @see XXH3_accumulate_512_neon() */ # ifndef XXH3_NEON_LANES # if (defined(__aarch64__) || defined(__arm64__) || defined(_M_ARM64) || defined(_M_ARM64EC)) \ && !defined(__APPLE__) && XXH_SIZE_OPT <= 0 # define XXH3_NEON_LANES 6 # else # define XXH3_NEON_LANES XXH_ACC_NB # endif # endif #endif /* XXH_VECTOR == XXH_NEON */ /* * VSX and Z Vector helpers. * * This is very messy, and any pull requests to clean this up are welcome. * * There are a lot of problems with supporting VSX and s390x, due to * inconsistent intrinsics, spotty coverage, and multiple endiannesses. */ #if XXH_VECTOR == XXH_VSX /* Annoyingly, these headers _may_ define three macros: `bool`, `vector`, * and `pixel`. This is a problem for obvious reasons. * * These keywords are unnecessary; the spec literally says they are * equivalent to `__bool`, `__vector`, and `__pixel` and may be undef'd * after including the header. * * We use pragma push_macro/pop_macro to keep the namespace clean. */ # pragma push_macro("bool") # pragma push_macro("vector") # pragma push_macro("pixel") /* silence potential macro redefined warnings */ # undef bool # undef vector # undef pixel # if defined(__s390x__) # include # else # include # endif /* Restore the original macro values, if applicable. */ # pragma pop_macro("pixel") # pragma pop_macro("vector") # pragma pop_macro("bool") typedef __vector unsigned long long xxh_u64x2; typedef __vector unsigned char xxh_u8x16; typedef __vector unsigned xxh_u32x4; /* * UGLY HACK: Similar to aarch64 macOS GCC, s390x GCC has the same aliasing issue. */ typedef xxh_u64x2 xxh_aliasing_u64x2 XXH_ALIASING; # ifndef XXH_VSX_BE # if defined(__BIG_ENDIAN__) \ || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) # define XXH_VSX_BE 1 # elif defined(__VEC_ELEMENT_REG_ORDER__) && __VEC_ELEMENT_REG_ORDER__ == __ORDER_BIG_ENDIAN__ # warning "-maltivec=be is not recommended. Please use native endianness." # define XXH_VSX_BE 1 # else # define XXH_VSX_BE 0 # endif # endif /* !defined(XXH_VSX_BE) */ # if XXH_VSX_BE # if defined(__POWER9_VECTOR__) || (defined(__clang__) && defined(__s390x__)) # define XXH_vec_revb vec_revb # else /*! * A polyfill for POWER9's vec_revb(). */ XXH_FORCE_INLINE xxh_u64x2 XXH_vec_revb(xxh_u64x2 val) { xxh_u8x16 const vByteSwap = { 0x07, 0x06, 0x05, 0x04, 0x03, 0x02, 0x01, 0x00, 0x0F, 0x0E, 0x0D, 0x0C, 0x0B, 0x0A, 0x09, 0x08 }; return vec_perm(val, val, vByteSwap); } # endif # endif /* XXH_VSX_BE */ /*! * Performs an unaligned vector load and byte swaps it on big endian. */ XXH_FORCE_INLINE xxh_u64x2 XXH_vec_loadu(const void *ptr) { xxh_u64x2 ret; XXH_memcpy(&ret, ptr, sizeof(xxh_u64x2)); # if XXH_VSX_BE ret = XXH_vec_revb(ret); # endif return ret; } /* * vec_mulo and vec_mule are very problematic intrinsics on PowerPC * * These intrinsics weren't added until GCC 8, despite existing for a while, * and they are endian dependent. Also, their meaning swap depending on version. * */ # if defined(__s390x__) /* s390x is always big endian, no issue on this platform */ # define XXH_vec_mulo vec_mulo # define XXH_vec_mule vec_mule # elif defined(__clang__) && XXH_HAS_BUILTIN(__builtin_altivec_vmuleuw) && !defined(__ibmxl__) /* Clang has a better way to control this, we can just use the builtin which doesn't swap. */ /* The IBM XL Compiler (which defined __clang__) only implements the vec_* operations */ # define XXH_vec_mulo __builtin_altivec_vmulouw # define XXH_vec_mule __builtin_altivec_vmuleuw # else /* gcc needs inline assembly */ /* Adapted from https://github.com/google/highwayhash/blob/master/highwayhash/hh_vsx.h. */ XXH_FORCE_INLINE xxh_u64x2 XXH_vec_mulo(xxh_u32x4 a, xxh_u32x4 b) { xxh_u64x2 result; __asm__("vmulouw %0, %1, %2" : "=v" (result) : "v" (a), "v" (b)); return result; } XXH_FORCE_INLINE xxh_u64x2 XXH_vec_mule(xxh_u32x4 a, xxh_u32x4 b) { xxh_u64x2 result; __asm__("vmuleuw %0, %1, %2" : "=v" (result) : "v" (a), "v" (b)); return result; } # endif /* XXH_vec_mulo, XXH_vec_mule */ #endif /* XXH_VECTOR == XXH_VSX */ #if XXH_VECTOR == XXH_SVE #define ACCRND(acc, offset) \ do { \ svuint64_t input_vec = svld1_u64(mask, xinput + offset); \ svuint64_t secret_vec = svld1_u64(mask, xsecret + offset); \ svuint64_t mixed = sveor_u64_x(mask, secret_vec, input_vec); \ svuint64_t swapped = svtbl_u64(input_vec, kSwap); \ svuint64_t mixed_lo = svextw_u64_x(mask, mixed); \ svuint64_t mixed_hi = svlsr_n_u64_x(mask, mixed, 32); \ svuint64_t mul = svmad_u64_x(mask, mixed_lo, mixed_hi, swapped); \ acc = svadd_u64_x(mask, acc, mul); \ } while (0) #endif /* XXH_VECTOR == XXH_SVE */ /* prefetch * can be disabled, by declaring XXH_NO_PREFETCH build macro */ #if defined(XXH_NO_PREFETCH) # define XXH_PREFETCH(ptr) (void)(ptr) /* disabled */ #else # if XXH_SIZE_OPT >= 1 # define XXH_PREFETCH(ptr) (void)(ptr) # elif defined(_MSC_VER) && (defined(_M_X64) || defined(_M_IX86)) /* _mm_prefetch() not defined outside of x86/x64 */ # include /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */ # define XXH_PREFETCH(ptr) _mm_prefetch((const char*)(ptr), _MM_HINT_T0) # elif defined(__GNUC__) && ( (__GNUC__ >= 4) || ( (__GNUC__ == 3) && (__GNUC_MINOR__ >= 1) ) ) # define XXH_PREFETCH(ptr) __builtin_prefetch((ptr), 0 /* rw==read */, 3 /* locality */) # else # define XXH_PREFETCH(ptr) (void)(ptr) /* disabled */ # endif #endif /* XXH_NO_PREFETCH */ /* ========================================== * XXH3 default settings * ========================================== */ #define XXH_SECRET_DEFAULT_SIZE 192 /* minimum XXH3_SECRET_SIZE_MIN */ #if (XXH_SECRET_DEFAULT_SIZE < XXH3_SECRET_SIZE_MIN) # error "default keyset is not large enough" #endif /*! Pseudorandom secret taken directly from FARSH. */ XXH_ALIGN(64) static const xxh_u8 XXH3_kSecret[XXH_SECRET_DEFAULT_SIZE] = { 0xb8, 0xfe, 0x6c, 0x39, 0x23, 0xa4, 0x4b, 0xbe, 0x7c, 0x01, 0x81, 0x2c, 0xf7, 0x21, 0xad, 0x1c, 0xde, 0xd4, 0x6d, 0xe9, 0x83, 0x90, 0x97, 0xdb, 0x72, 0x40, 0xa4, 0xa4, 0xb7, 0xb3, 0x67, 0x1f, 0xcb, 0x79, 0xe6, 0x4e, 0xcc, 0xc0, 0xe5, 0x78, 0x82, 0x5a, 0xd0, 0x7d, 0xcc, 0xff, 0x72, 0x21, 0xb8, 0x08, 0x46, 0x74, 0xf7, 0x43, 0x24, 0x8e, 0xe0, 0x35, 0x90, 0xe6, 0x81, 0x3a, 0x26, 0x4c, 0x3c, 0x28, 0x52, 0xbb, 0x91, 0xc3, 0x00, 0xcb, 0x88, 0xd0, 0x65, 0x8b, 0x1b, 0x53, 0x2e, 0xa3, 0x71, 0x64, 0x48, 0x97, 0xa2, 0x0d, 0xf9, 0x4e, 0x38, 0x19, 0xef, 0x46, 0xa9, 0xde, 0xac, 0xd8, 0xa8, 0xfa, 0x76, 0x3f, 0xe3, 0x9c, 0x34, 0x3f, 0xf9, 0xdc, 0xbb, 0xc7, 0xc7, 0x0b, 0x4f, 0x1d, 0x8a, 0x51, 0xe0, 0x4b, 0xcd, 0xb4, 0x59, 0x31, 0xc8, 0x9f, 0x7e, 0xc9, 0xd9, 0x78, 0x73, 0x64, 0xea, 0xc5, 0xac, 0x83, 0x34, 0xd3, 0xeb, 0xc3, 0xc5, 0x81, 0xa0, 0xff, 0xfa, 0x13, 0x63, 0xeb, 0x17, 0x0d, 0xdd, 0x51, 0xb7, 0xf0, 0xda, 0x49, 0xd3, 0x16, 0x55, 0x26, 0x29, 0xd4, 0x68, 0x9e, 0x2b, 0x16, 0xbe, 0x58, 0x7d, 0x47, 0xa1, 0xfc, 0x8f, 0xf8, 0xb8, 0xd1, 0x7a, 0xd0, 0x31, 0xce, 0x45, 0xcb, 0x3a, 0x8f, 0x95, 0x16, 0x04, 0x28, 0xaf, 0xd7, 0xfb, 0xca, 0xbb, 0x4b, 0x40, 0x7e, }; static const xxh_u64 PRIME_MX1 = 0x165667919E3779F9ULL; /*!< 0b0001011001010110011001111001000110011110001101110111100111111001 */ static const xxh_u64 PRIME_MX2 = 0x9FB21C651E98DF25ULL; /*!< 0b1001111110110010000111000110010100011110100110001101111100100101 */ #ifdef XXH_OLD_NAMES # define kSecret XXH3_kSecret #endif #ifdef XXH_DOXYGEN /*! * @brief Calculates a 32-bit to 64-bit long multiply. * * Implemented as a macro. * * Wraps `__emulu` on MSVC x86 because it tends to call `__allmul` when it doesn't * need to (but it shouldn't need to anyways, it is about 7 instructions to do * a 64x64 multiply...). Since we know that this will _always_ emit `MULL`, we * use that instead of the normal method. * * If you are compiling for platforms like Thumb-1 and don't have a better option, * you may also want to write your own long multiply routine here. * * @param x, y Numbers to be multiplied * @return 64-bit product of the low 32 bits of @p x and @p y. */ XXH_FORCE_INLINE xxh_u64 XXH_mult32to64(xxh_u64 x, xxh_u64 y) { return (x & 0xFFFFFFFF) * (y & 0xFFFFFFFF); } #elif defined(_MSC_VER) && defined(_M_IX86) # define XXH_mult32to64(x, y) __emulu((unsigned)(x), (unsigned)(y)) #else /* * Downcast + upcast is usually better than masking on older compilers like * GCC 4.2 (especially 32-bit ones), all without affecting newer compilers. * * The other method, (x & 0xFFFFFFFF) * (y & 0xFFFFFFFF), will AND both operands * and perform a full 64x64 multiply -- entirely redundant on 32-bit. */ # define XXH_mult32to64(x, y) ((xxh_u64)(xxh_u32)(x) * (xxh_u64)(xxh_u32)(y)) #endif /*! * @brief Calculates a 64->128-bit long multiply. * * Uses `__uint128_t` and `_umul128` if available, otherwise uses a scalar * version. * * @param lhs , rhs The 64-bit integers to be multiplied * @return The 128-bit result represented in an @ref XXH128_hash_t. */ static XXH128_hash_t XXH_mult64to128(xxh_u64 lhs, xxh_u64 rhs) { /* * GCC/Clang __uint128_t method. * * On most 64-bit targets, GCC and Clang define a __uint128_t type. * This is usually the best way as it usually uses a native long 64-bit * multiply, such as MULQ on x86_64 or MUL + UMULH on aarch64. * * Usually. * * Despite being a 32-bit platform, Clang (and emscripten) define this type * despite not having the arithmetic for it. This results in a laggy * compiler builtin call which calculates a full 128-bit multiply. * In that case it is best to use the portable one. * https://github.com/Cyan4973/xxHash/issues/211#issuecomment-515575677 */ #if (defined(__GNUC__) || defined(__clang__)) && !defined(__wasm__) \ && defined(__SIZEOF_INT128__) \ || (defined(_INTEGRAL_MAX_BITS) && _INTEGRAL_MAX_BITS >= 128) __uint128_t const product = (__uint128_t)lhs * (__uint128_t)rhs; XXH128_hash_t r128; r128.low64 = (xxh_u64)(product); r128.high64 = (xxh_u64)(product >> 64); return r128; /* * MSVC for x64's _umul128 method. * * xxh_u64 _umul128(xxh_u64 Multiplier, xxh_u64 Multiplicand, xxh_u64 *HighProduct); * * This compiles to single operand MUL on x64. */ #elif (defined(_M_X64) || defined(_M_IA64)) && !defined(_M_ARM64EC) #ifndef _MSC_VER # pragma intrinsic(_umul128) #endif xxh_u64 product_high; xxh_u64 const product_low = _umul128(lhs, rhs, &product_high); XXH128_hash_t r128; r128.low64 = product_low; r128.high64 = product_high; return r128; /* * MSVC for ARM64's __umulh method. * * This compiles to the same MUL + UMULH as GCC/Clang's __uint128_t method. */ #elif defined(_M_ARM64) || defined(_M_ARM64EC) #ifndef _MSC_VER # pragma intrinsic(__umulh) #endif XXH128_hash_t r128; r128.low64 = lhs * rhs; r128.high64 = __umulh(lhs, rhs); return r128; #else /* * Portable scalar method. Optimized for 32-bit and 64-bit ALUs. * * This is a fast and simple grade school multiply, which is shown below * with base 10 arithmetic instead of base 0x100000000. * * 9 3 // D2 lhs = 93 * x 7 5 // D2 rhs = 75 * ---------- * 1 5 // D2 lo_lo = (93 % 10) * (75 % 10) = 15 * 4 5 | // D2 hi_lo = (93 / 10) * (75 % 10) = 45 * 2 1 | // D2 lo_hi = (93 % 10) * (75 / 10) = 21 * + 6 3 | | // D2 hi_hi = (93 / 10) * (75 / 10) = 63 * --------- * 2 7 | // D2 cross = (15 / 10) + (45 % 10) + 21 = 27 * + 6 7 | | // D2 upper = (27 / 10) + (45 / 10) + 63 = 67 * --------- * 6 9 7 5 // D4 res = (27 * 10) + (15 % 10) + (67 * 100) = 6975 * * The reasons for adding the products like this are: * 1. It avoids manual carry tracking. Just like how * (9 * 9) + 9 + 9 = 99, the same applies with this for UINT64_MAX. * This avoids a lot of complexity. * * 2. It hints for, and on Clang, compiles to, the powerful UMAAL * instruction available in ARM's Digital Signal Processing extension * in 32-bit ARMv6 and later, which is shown below: * * void UMAAL(xxh_u32 *RdLo, xxh_u32 *RdHi, xxh_u32 Rn, xxh_u32 Rm) * { * xxh_u64 product = (xxh_u64)*RdLo * (xxh_u64)*RdHi + Rn + Rm; * *RdLo = (xxh_u32)(product & 0xFFFFFFFF); * *RdHi = (xxh_u32)(product >> 32); * } * * This instruction was designed for efficient long multiplication, and * allows this to be calculated in only 4 instructions at speeds * comparable to some 64-bit ALUs. * * 3. It isn't terrible on other platforms. Usually this will be a couple * of 32-bit ADD/ADCs. */ /* First calculate all of the cross products. */ xxh_u64 const lo_lo = XXH_mult32to64(lhs & 0xFFFFFFFF, rhs & 0xFFFFFFFF); xxh_u64 const hi_lo = XXH_mult32to64(lhs >> 32, rhs & 0xFFFFFFFF); xxh_u64 const lo_hi = XXH_mult32to64(lhs & 0xFFFFFFFF, rhs >> 32); xxh_u64 const hi_hi = XXH_mult32to64(lhs >> 32, rhs >> 32); /* Now add the products together. These will never overflow. */ xxh_u64 const cross = (lo_lo >> 32) + (hi_lo & 0xFFFFFFFF) + lo_hi; xxh_u64 const upper = (hi_lo >> 32) + (cross >> 32) + hi_hi; xxh_u64 const lower = (cross << 32) | (lo_lo & 0xFFFFFFFF); XXH128_hash_t r128; r128.low64 = lower; r128.high64 = upper; return r128; #endif } /*! * @brief Calculates a 64-bit to 128-bit multiply, then XOR folds it. * * The reason for the separate function is to prevent passing too many structs * around by value. This will hopefully inline the multiply, but we don't force it. * * @param lhs , rhs The 64-bit integers to multiply * @return The low 64 bits of the product XOR'd by the high 64 bits. * @see XXH_mult64to128() */ static xxh_u64 XXH3_mul128_fold64(xxh_u64 lhs, xxh_u64 rhs) { XXH128_hash_t product = XXH_mult64to128(lhs, rhs); return product.low64 ^ product.high64; } /*! Seems to produce slightly better code on GCC for some reason. */ XXH_FORCE_INLINE XXH_CONSTF xxh_u64 XXH_xorshift64(xxh_u64 v64, int shift) { XXH_ASSERT(0 <= shift && shift < 64); return v64 ^ (v64 >> shift); } /* * This is a fast avalanche stage, * suitable when input bits are already partially mixed */ static XXH64_hash_t XXH3_avalanche(xxh_u64 h64) { h64 = XXH_xorshift64(h64, 37); h64 *= PRIME_MX1; h64 = XXH_xorshift64(h64, 32); return h64; } /* * This is a stronger avalanche, * inspired by Pelle Evensen's rrmxmx * preferable when input has not been previously mixed */ static XXH64_hash_t XXH3_rrmxmx(xxh_u64 h64, xxh_u64 len) { /* this mix is inspired by Pelle Evensen's rrmxmx */ h64 ^= XXH_rotl64(h64, 49) ^ XXH_rotl64(h64, 24); h64 *= PRIME_MX2; h64 ^= (h64 >> 35) + len ; h64 *= PRIME_MX2; return XXH_xorshift64(h64, 28); } /* ========================================== * Short keys * ========================================== * One of the shortcomings of XXH32 and XXH64 was that their performance was * sub-optimal on short lengths. It used an iterative algorithm which strongly * favored lengths that were a multiple of 4 or 8. * * Instead of iterating over individual inputs, we use a set of single shot * functions which piece together a range of lengths and operate in constant time. * * Additionally, the number of multiplies has been significantly reduced. This * reduces latency, especially when emulating 64-bit multiplies on 32-bit. * * Depending on the platform, this may or may not be faster than XXH32, but it * is almost guaranteed to be faster than XXH64. */ /* * At very short lengths, there isn't enough input to fully hide secrets, or use * the entire secret. * * There is also only a limited amount of mixing we can do before significantly * impacting performance. * * Therefore, we use different sections of the secret and always mix two secret * samples with an XOR. This should have no effect on performance on the * seedless or withSeed variants because everything _should_ be constant folded * by modern compilers. * * The XOR mixing hides individual parts of the secret and increases entropy. * * This adds an extra layer of strength for custom secrets. */ XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t XXH3_len_1to3_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) { XXH_ASSERT(input != NULL); XXH_ASSERT(1 <= len && len <= 3); XXH_ASSERT(secret != NULL); /* * len = 1: combined = { input[0], 0x01, input[0], input[0] } * len = 2: combined = { input[1], 0x02, input[0], input[1] } * len = 3: combined = { input[2], 0x03, input[0], input[1] } */ { xxh_u8 const c1 = input[0]; xxh_u8 const c2 = input[len >> 1]; xxh_u8 const c3 = input[len - 1]; xxh_u32 const combined = ((xxh_u32)c1 << 16) | ((xxh_u32)c2 << 24) | ((xxh_u32)c3 << 0) | ((xxh_u32)len << 8); xxh_u64 const bitflip = (XXH_readLE32(secret) ^ XXH_readLE32(secret+4)) + seed; xxh_u64 const keyed = (xxh_u64)combined ^ bitflip; return XXH64_avalanche(keyed); } } XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t XXH3_len_4to8_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) { XXH_ASSERT(input != NULL); XXH_ASSERT(secret != NULL); XXH_ASSERT(4 <= len && len <= 8); seed ^= (xxh_u64)XXH_swap32((xxh_u32)seed) << 32; { xxh_u32 const input1 = XXH_readLE32(input); xxh_u32 const input2 = XXH_readLE32(input + len - 4); xxh_u64 const bitflip = (XXH_readLE64(secret+8) ^ XXH_readLE64(secret+16)) - seed; xxh_u64 const input64 = input2 + (((xxh_u64)input1) << 32); xxh_u64 const keyed = input64 ^ bitflip; return XXH3_rrmxmx(keyed, len); } } XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t XXH3_len_9to16_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) { XXH_ASSERT(input != NULL); XXH_ASSERT(secret != NULL); XXH_ASSERT(9 <= len && len <= 16); { xxh_u64 const bitflip1 = (XXH_readLE64(secret+24) ^ XXH_readLE64(secret+32)) + seed; xxh_u64 const bitflip2 = (XXH_readLE64(secret+40) ^ XXH_readLE64(secret+48)) - seed; xxh_u64 const input_lo = XXH_readLE64(input) ^ bitflip1; xxh_u64 const input_hi = XXH_readLE64(input + len - 8) ^ bitflip2; xxh_u64 const acc = len + XXH_swap64(input_lo) + input_hi + XXH3_mul128_fold64(input_lo, input_hi); return XXH3_avalanche(acc); } } XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t XXH3_len_0to16_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) { XXH_ASSERT(len <= 16); { if (XXH_likely(len > 8)) return XXH3_len_9to16_64b(input, len, secret, seed); if (XXH_likely(len >= 4)) return XXH3_len_4to8_64b(input, len, secret, seed); if (len) return XXH3_len_1to3_64b(input, len, secret, seed); return XXH64_avalanche(seed ^ (XXH_readLE64(secret+56) ^ XXH_readLE64(secret+64))); } } /* * DISCLAIMER: There are known *seed-dependent* multicollisions here due to * multiplication by zero, affecting hashes of lengths 17 to 240. * * However, they are very unlikely. * * Keep this in mind when using the unseeded XXH3_64bits() variant: As with all * unseeded non-cryptographic hashes, it does not attempt to defend itself * against specially crafted inputs, only random inputs. * * Compared to classic UMAC where a 1 in 2^31 chance of 4 consecutive bytes * cancelling out the secret is taken an arbitrary number of times (addressed * in XXH3_accumulate_512), this collision is very unlikely with random inputs * and/or proper seeding: * * This only has a 1 in 2^63 chance of 8 consecutive bytes cancelling out, in a * function that is only called up to 16 times per hash with up to 240 bytes of * input. * * This is not too bad for a non-cryptographic hash function, especially with * only 64 bit outputs. * * The 128-bit variant (which trades some speed for strength) is NOT affected * by this, although it is always a good idea to use a proper seed if you care * about strength. */ XXH_FORCE_INLINE xxh_u64 XXH3_mix16B(const xxh_u8* XXH_RESTRICT input, const xxh_u8* XXH_RESTRICT secret, xxh_u64 seed64) { #if defined(__GNUC__) && !defined(__clang__) /* GCC, not Clang */ \ && defined(__i386__) && defined(__SSE2__) /* x86 + SSE2 */ \ && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable like XXH32 hack */ /* * UGLY HACK: * GCC for x86 tends to autovectorize the 128-bit multiply, resulting in * slower code. * * By forcing seed64 into a register, we disrupt the cost model and * cause it to scalarize. See `XXH32_round()` * * FIXME: Clang's output is still _much_ faster -- On an AMD Ryzen 3600, * XXH3_64bits @ len=240 runs at 4.6 GB/s with Clang 9, but 3.3 GB/s on * GCC 9.2, despite both emitting scalar code. * * GCC generates much better scalar code than Clang for the rest of XXH3, * which is why finding a more optimal codepath is an interest. */ XXH_COMPILER_GUARD(seed64); #endif { xxh_u64 const input_lo = XXH_readLE64(input); xxh_u64 const input_hi = XXH_readLE64(input+8); return XXH3_mul128_fold64( input_lo ^ (XXH_readLE64(secret) + seed64), input_hi ^ (XXH_readLE64(secret+8) - seed64) ); } } /* For mid range keys, XXH3 uses a Mum-hash variant. */ XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t XXH3_len_17to128_64b(const xxh_u8* XXH_RESTRICT input, size_t len, const xxh_u8* XXH_RESTRICT secret, size_t secretSize, XXH64_hash_t seed) { XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; XXH_ASSERT(16 < len && len <= 128); { xxh_u64 acc = len * XXH_PRIME64_1; #if XXH_SIZE_OPT >= 1 /* Smaller and cleaner, but slightly slower. */ unsigned int i = (unsigned int)(len - 1) / 32; do { acc += XXH3_mix16B(input+16 * i, secret+32*i, seed); acc += XXH3_mix16B(input+len-16*(i+1), secret+32*i+16, seed); } while (i-- != 0); #else if (len > 32) { if (len > 64) { if (len > 96) { acc += XXH3_mix16B(input+48, secret+96, seed); acc += XXH3_mix16B(input+len-64, secret+112, seed); } acc += XXH3_mix16B(input+32, secret+64, seed); acc += XXH3_mix16B(input+len-48, secret+80, seed); } acc += XXH3_mix16B(input+16, secret+32, seed); acc += XXH3_mix16B(input+len-32, secret+48, seed); } acc += XXH3_mix16B(input+0, secret+0, seed); acc += XXH3_mix16B(input+len-16, secret+16, seed); #endif return XXH3_avalanche(acc); } } XXH_NO_INLINE XXH_PUREF XXH64_hash_t XXH3_len_129to240_64b(const xxh_u8* XXH_RESTRICT input, size_t len, const xxh_u8* XXH_RESTRICT secret, size_t secretSize, XXH64_hash_t seed) { XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); #define XXH3_MIDSIZE_STARTOFFSET 3 #define XXH3_MIDSIZE_LASTOFFSET 17 { xxh_u64 acc = len * XXH_PRIME64_1; xxh_u64 acc_end; unsigned int const nbRounds = (unsigned int)len / 16; unsigned int i; XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); for (i=0; i<8; i++) { acc += XXH3_mix16B(input+(16*i), secret+(16*i), seed); } /* last bytes */ acc_end = XXH3_mix16B(input + len - 16, secret + XXH3_SECRET_SIZE_MIN - XXH3_MIDSIZE_LASTOFFSET, seed); XXH_ASSERT(nbRounds >= 8); acc = XXH3_avalanche(acc); #if defined(__clang__) /* Clang */ \ && (defined(__ARM_NEON) || defined(__ARM_NEON__)) /* NEON */ \ && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable */ /* * UGLY HACK: * Clang for ARMv7-A tries to vectorize this loop, similar to GCC x86. * In everywhere else, it uses scalar code. * * For 64->128-bit multiplies, even if the NEON was 100% optimal, it * would still be slower than UMAAL (see XXH_mult64to128). * * Unfortunately, Clang doesn't handle the long multiplies properly and * converts them to the nonexistent "vmulq_u64" intrinsic, which is then * scalarized into an ugly mess of VMOV.32 instructions. * * This mess is difficult to avoid without turning autovectorization * off completely, but they are usually relatively minor and/or not * worth it to fix. * * This loop is the easiest to fix, as unlike XXH32, this pragma * _actually works_ because it is a loop vectorization instead of an * SLP vectorization. */ #pragma clang loop vectorize(disable) #endif for (i=8 ; i < nbRounds; i++) { /* * Prevents clang for unrolling the acc loop and interleaving with this one. */ XXH_COMPILER_GUARD(acc); acc_end += XXH3_mix16B(input+(16*i), secret+(16*(i-8)) + XXH3_MIDSIZE_STARTOFFSET, seed); } return XXH3_avalanche(acc + acc_end); } } /* ======= Long Keys ======= */ #define XXH_STRIPE_LEN 64 #define XXH_SECRET_CONSUME_RATE 8 /* nb of secret bytes consumed at each accumulation */ #define XXH_ACC_NB (XXH_STRIPE_LEN / sizeof(xxh_u64)) #ifdef XXH_OLD_NAMES # define STRIPE_LEN XXH_STRIPE_LEN # define ACC_NB XXH_ACC_NB #endif #ifndef XXH_PREFETCH_DIST # ifdef __clang__ # define XXH_PREFETCH_DIST 320 # else # if (XXH_VECTOR == XXH_AVX512) # define XXH_PREFETCH_DIST 512 # else # define XXH_PREFETCH_DIST 384 # endif # endif /* __clang__ */ #endif /* XXH_PREFETCH_DIST */ /* * These macros are to generate an XXH3_accumulate() function. * The two arguments select the name suffix and target attribute. * * The name of this symbol is XXH3_accumulate_() and it calls * XXH3_accumulate_512_(). * * It may be useful to hand implement this function if the compiler fails to * optimize the inline function. */ #define XXH3_ACCUMULATE_TEMPLATE(name) \ void \ XXH3_accumulate_##name(xxh_u64* XXH_RESTRICT acc, \ const xxh_u8* XXH_RESTRICT input, \ const xxh_u8* XXH_RESTRICT secret, \ size_t nbStripes) \ { \ size_t n; \ for (n = 0; n < nbStripes; n++ ) { \ const xxh_u8* const in = input + n*XXH_STRIPE_LEN; \ XXH_PREFETCH(in + XXH_PREFETCH_DIST); \ XXH3_accumulate_512_##name( \ acc, \ in, \ secret + n*XXH_SECRET_CONSUME_RATE); \ } \ } XXH_FORCE_INLINE void XXH_writeLE64(void* dst, xxh_u64 v64) { if (!XXH_CPU_LITTLE_ENDIAN) v64 = XXH_swap64(v64); XXH_memcpy(dst, &v64, sizeof(v64)); } /* Several intrinsic functions below are supposed to accept __int64 as argument, * as documented in https://software.intel.com/sites/landingpage/IntrinsicsGuide/ . * However, several environments do not define __int64 type, * requiring a workaround. */ #if !defined (__VMS) \ && (defined (__cplusplus) \ || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) typedef int64_t xxh_i64; #else /* the following type must have a width of 64-bit */ typedef long long xxh_i64; #endif /* * XXH3_accumulate_512 is the tightest loop for long inputs, and it is the most optimized. * * It is a hardened version of UMAC, based off of FARSH's implementation. * * This was chosen because it adapts quite well to 32-bit, 64-bit, and SIMD * implementations, and it is ridiculously fast. * * We harden it by mixing the original input to the accumulators as well as the product. * * This means that in the (relatively likely) case of a multiply by zero, the * original input is preserved. * * On 128-bit inputs, we swap 64-bit pairs when we add the input to improve * cross-pollination, as otherwise the upper and lower halves would be * essentially independent. * * This doesn't matter on 64-bit hashes since they all get merged together in * the end, so we skip the extra step. * * Both XXH3_64bits and XXH3_128bits use this subroutine. */ #if (XXH_VECTOR == XXH_AVX512) \ || (defined(XXH_DISPATCH_AVX512) && XXH_DISPATCH_AVX512 != 0) #ifndef XXH_TARGET_AVX512 # define XXH_TARGET_AVX512 /* disable attribute target */ #endif XXH_FORCE_INLINE XXH_TARGET_AVX512 void XXH3_accumulate_512_avx512(void* XXH_RESTRICT acc, const void* XXH_RESTRICT input, const void* XXH_RESTRICT secret) { __m512i* const xacc = (__m512i *) acc; XXH_ASSERT((((size_t)acc) & 63) == 0); XXH_STATIC_ASSERT(XXH_STRIPE_LEN == sizeof(__m512i)); { /* data_vec = input[0]; */ __m512i const data_vec = _mm512_loadu_si512 (input); /* key_vec = secret[0]; */ __m512i const key_vec = _mm512_loadu_si512 (secret); /* data_key = data_vec ^ key_vec; */ __m512i const data_key = _mm512_xor_si512 (data_vec, key_vec); /* data_key_lo = data_key >> 32; */ __m512i const data_key_lo = _mm512_srli_epi64 (data_key, 32); /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ __m512i const product = _mm512_mul_epu32 (data_key, data_key_lo); /* xacc[0] += swap(data_vec); */ __m512i const data_swap = _mm512_shuffle_epi32(data_vec, (_MM_PERM_ENUM)_MM_SHUFFLE(1, 0, 3, 2)); __m512i const sum = _mm512_add_epi64(*xacc, data_swap); /* xacc[0] += product; */ *xacc = _mm512_add_epi64(product, sum); } } XXH_FORCE_INLINE XXH_TARGET_AVX512 XXH3_ACCUMULATE_TEMPLATE(avx512) /* * XXH3_scrambleAcc: Scrambles the accumulators to improve mixing. * * Multiplication isn't perfect, as explained by Google in HighwayHash: * * // Multiplication mixes/scrambles bytes 0-7 of the 64-bit result to * // varying degrees. In descending order of goodness, bytes * // 3 4 2 5 1 6 0 7 have quality 228 224 164 160 100 96 36 32. * // As expected, the upper and lower bytes are much worse. * * Source: https://github.com/google/highwayhash/blob/0aaf66b/highwayhash/hh_avx2.h#L291 * * Since our algorithm uses a pseudorandom secret to add some variance into the * mix, we don't need to (or want to) mix as often or as much as HighwayHash does. * * This isn't as tight as XXH3_accumulate, but still written in SIMD to avoid * extraction. * * Both XXH3_64bits and XXH3_128bits use this subroutine. */ XXH_FORCE_INLINE XXH_TARGET_AVX512 void XXH3_scrambleAcc_avx512(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) { XXH_ASSERT((((size_t)acc) & 63) == 0); XXH_STATIC_ASSERT(XXH_STRIPE_LEN == sizeof(__m512i)); { __m512i* const xacc = (__m512i*) acc; const __m512i prime32 = _mm512_set1_epi32((int)XXH_PRIME32_1); /* xacc[0] ^= (xacc[0] >> 47) */ __m512i const acc_vec = *xacc; __m512i const shifted = _mm512_srli_epi64 (acc_vec, 47); /* xacc[0] ^= secret; */ __m512i const key_vec = _mm512_loadu_si512 (secret); __m512i const data_key = _mm512_ternarylogic_epi32(key_vec, acc_vec, shifted, 0x96 /* key_vec ^ acc_vec ^ shifted */); /* xacc[0] *= XXH_PRIME32_1; */ __m512i const data_key_hi = _mm512_srli_epi64 (data_key, 32); __m512i const prod_lo = _mm512_mul_epu32 (data_key, prime32); __m512i const prod_hi = _mm512_mul_epu32 (data_key_hi, prime32); *xacc = _mm512_add_epi64(prod_lo, _mm512_slli_epi64(prod_hi, 32)); } } XXH_FORCE_INLINE XXH_TARGET_AVX512 void XXH3_initCustomSecret_avx512(void* XXH_RESTRICT customSecret, xxh_u64 seed64) { XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 63) == 0); XXH_STATIC_ASSERT(XXH_SEC_ALIGN == 64); XXH_ASSERT(((size_t)customSecret & 63) == 0); (void)(&XXH_writeLE64); { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / sizeof(__m512i); __m512i const seed_pos = _mm512_set1_epi64((xxh_i64)seed64); __m512i const seed = _mm512_mask_sub_epi64(seed_pos, 0xAA, _mm512_set1_epi8(0), seed_pos); const __m512i* const src = (const __m512i*) ((const void*) XXH3_kSecret); __m512i* const dest = ( __m512i*) customSecret; int i; XXH_ASSERT(((size_t)src & 63) == 0); /* control alignment */ XXH_ASSERT(((size_t)dest & 63) == 0); for (i=0; i < nbRounds; ++i) { dest[i] = _mm512_add_epi64(_mm512_load_si512(src + i), seed); } } } #endif #if (XXH_VECTOR == XXH_AVX2) \ || (defined(XXH_DISPATCH_AVX2) && XXH_DISPATCH_AVX2 != 0) #ifndef XXH_TARGET_AVX2 # define XXH_TARGET_AVX2 /* disable attribute target */ #endif XXH_FORCE_INLINE XXH_TARGET_AVX2 void XXH3_accumulate_512_avx2( void* XXH_RESTRICT acc, const void* XXH_RESTRICT input, const void* XXH_RESTRICT secret) { XXH_ASSERT((((size_t)acc) & 31) == 0); { __m256i* const xacc = (__m256i *) acc; /* Unaligned. This is mainly for pointer arithmetic, and because * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ const __m256i* const xinput = (const __m256i *) input; /* Unaligned. This is mainly for pointer arithmetic, and because * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ const __m256i* const xsecret = (const __m256i *) secret; size_t i; for (i=0; i < XXH_STRIPE_LEN/sizeof(__m256i); i++) { /* data_vec = xinput[i]; */ __m256i const data_vec = _mm256_loadu_si256 (xinput+i); /* key_vec = xsecret[i]; */ __m256i const key_vec = _mm256_loadu_si256 (xsecret+i); /* data_key = data_vec ^ key_vec; */ __m256i const data_key = _mm256_xor_si256 (data_vec, key_vec); /* data_key_lo = data_key >> 32; */ __m256i const data_key_lo = _mm256_srli_epi64 (data_key, 32); /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ __m256i const product = _mm256_mul_epu32 (data_key, data_key_lo); /* xacc[i] += swap(data_vec); */ __m256i const data_swap = _mm256_shuffle_epi32(data_vec, _MM_SHUFFLE(1, 0, 3, 2)); __m256i const sum = _mm256_add_epi64(xacc[i], data_swap); /* xacc[i] += product; */ xacc[i] = _mm256_add_epi64(product, sum); } } } XXH_FORCE_INLINE XXH_TARGET_AVX2 XXH3_ACCUMULATE_TEMPLATE(avx2) XXH_FORCE_INLINE XXH_TARGET_AVX2 void XXH3_scrambleAcc_avx2(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) { XXH_ASSERT((((size_t)acc) & 31) == 0); { __m256i* const xacc = (__m256i*) acc; /* Unaligned. This is mainly for pointer arithmetic, and because * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ const __m256i* const xsecret = (const __m256i *) secret; const __m256i prime32 = _mm256_set1_epi32((int)XXH_PRIME32_1); size_t i; for (i=0; i < XXH_STRIPE_LEN/sizeof(__m256i); i++) { /* xacc[i] ^= (xacc[i] >> 47) */ __m256i const acc_vec = xacc[i]; __m256i const shifted = _mm256_srli_epi64 (acc_vec, 47); __m256i const data_vec = _mm256_xor_si256 (acc_vec, shifted); /* xacc[i] ^= xsecret; */ __m256i const key_vec = _mm256_loadu_si256 (xsecret+i); __m256i const data_key = _mm256_xor_si256 (data_vec, key_vec); /* xacc[i] *= XXH_PRIME32_1; */ __m256i const data_key_hi = _mm256_srli_epi64 (data_key, 32); __m256i const prod_lo = _mm256_mul_epu32 (data_key, prime32); __m256i const prod_hi = _mm256_mul_epu32 (data_key_hi, prime32); xacc[i] = _mm256_add_epi64(prod_lo, _mm256_slli_epi64(prod_hi, 32)); } } } XXH_FORCE_INLINE XXH_TARGET_AVX2 void XXH3_initCustomSecret_avx2(void* XXH_RESTRICT customSecret, xxh_u64 seed64) { XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 31) == 0); XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE / sizeof(__m256i)) == 6); XXH_STATIC_ASSERT(XXH_SEC_ALIGN <= 64); (void)(&XXH_writeLE64); XXH_PREFETCH(customSecret); { __m256i const seed = _mm256_set_epi64x((xxh_i64)(0U - seed64), (xxh_i64)seed64, (xxh_i64)(0U - seed64), (xxh_i64)seed64); const __m256i* const src = (const __m256i*) ((const void*) XXH3_kSecret); __m256i* dest = ( __m256i*) customSecret; # if defined(__GNUC__) || defined(__clang__) /* * On GCC & Clang, marking 'dest' as modified will cause the compiler: * - do not extract the secret from sse registers in the internal loop * - use less common registers, and avoid pushing these reg into stack */ XXH_COMPILER_GUARD(dest); # endif XXH_ASSERT(((size_t)src & 31) == 0); /* control alignment */ XXH_ASSERT(((size_t)dest & 31) == 0); /* GCC -O2 need unroll loop manually */ dest[0] = _mm256_add_epi64(_mm256_load_si256(src+0), seed); dest[1] = _mm256_add_epi64(_mm256_load_si256(src+1), seed); dest[2] = _mm256_add_epi64(_mm256_load_si256(src+2), seed); dest[3] = _mm256_add_epi64(_mm256_load_si256(src+3), seed); dest[4] = _mm256_add_epi64(_mm256_load_si256(src+4), seed); dest[5] = _mm256_add_epi64(_mm256_load_si256(src+5), seed); } } #endif /* x86dispatch always generates SSE2 */ #if (XXH_VECTOR == XXH_SSE2) || defined(XXH_X86DISPATCH) #ifndef XXH_TARGET_SSE2 # define XXH_TARGET_SSE2 /* disable attribute target */ #endif XXH_FORCE_INLINE XXH_TARGET_SSE2 void XXH3_accumulate_512_sse2( void* XXH_RESTRICT acc, const void* XXH_RESTRICT input, const void* XXH_RESTRICT secret) { /* SSE2 is just a half-scale version of the AVX2 version. */ XXH_ASSERT((((size_t)acc) & 15) == 0); { __m128i* const xacc = (__m128i *) acc; /* Unaligned. This is mainly for pointer arithmetic, and because * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ const __m128i* const xinput = (const __m128i *) input; /* Unaligned. This is mainly for pointer arithmetic, and because * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ const __m128i* const xsecret = (const __m128i *) secret; size_t i; for (i=0; i < XXH_STRIPE_LEN/sizeof(__m128i); i++) { /* data_vec = xinput[i]; */ __m128i const data_vec = _mm_loadu_si128 (xinput+i); /* key_vec = xsecret[i]; */ __m128i const key_vec = _mm_loadu_si128 (xsecret+i); /* data_key = data_vec ^ key_vec; */ __m128i const data_key = _mm_xor_si128 (data_vec, key_vec); /* data_key_lo = data_key >> 32; */ __m128i const data_key_lo = _mm_shuffle_epi32 (data_key, _MM_SHUFFLE(0, 3, 0, 1)); /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ __m128i const product = _mm_mul_epu32 (data_key, data_key_lo); /* xacc[i] += swap(data_vec); */ __m128i const data_swap = _mm_shuffle_epi32(data_vec, _MM_SHUFFLE(1,0,3,2)); __m128i const sum = _mm_add_epi64(xacc[i], data_swap); /* xacc[i] += product; */ xacc[i] = _mm_add_epi64(product, sum); } } } XXH_FORCE_INLINE XXH_TARGET_SSE2 XXH3_ACCUMULATE_TEMPLATE(sse2) XXH_FORCE_INLINE XXH_TARGET_SSE2 void XXH3_scrambleAcc_sse2(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) { XXH_ASSERT((((size_t)acc) & 15) == 0); { __m128i* const xacc = (__m128i*) acc; /* Unaligned. This is mainly for pointer arithmetic, and because * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ const __m128i* const xsecret = (const __m128i *) secret; const __m128i prime32 = _mm_set1_epi32((int)XXH_PRIME32_1); size_t i; for (i=0; i < XXH_STRIPE_LEN/sizeof(__m128i); i++) { /* xacc[i] ^= (xacc[i] >> 47) */ __m128i const acc_vec = xacc[i]; __m128i const shifted = _mm_srli_epi64 (acc_vec, 47); __m128i const data_vec = _mm_xor_si128 (acc_vec, shifted); /* xacc[i] ^= xsecret[i]; */ __m128i const key_vec = _mm_loadu_si128 (xsecret+i); __m128i const data_key = _mm_xor_si128 (data_vec, key_vec); /* xacc[i] *= XXH_PRIME32_1; */ __m128i const data_key_hi = _mm_shuffle_epi32 (data_key, _MM_SHUFFLE(0, 3, 0, 1)); __m128i const prod_lo = _mm_mul_epu32 (data_key, prime32); __m128i const prod_hi = _mm_mul_epu32 (data_key_hi, prime32); xacc[i] = _mm_add_epi64(prod_lo, _mm_slli_epi64(prod_hi, 32)); } } } XXH_FORCE_INLINE XXH_TARGET_SSE2 void XXH3_initCustomSecret_sse2(void* XXH_RESTRICT customSecret, xxh_u64 seed64) { XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 15) == 0); (void)(&XXH_writeLE64); { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / sizeof(__m128i); # if defined(_MSC_VER) && defined(_M_IX86) && _MSC_VER < 1900 /* MSVC 32bit mode does not support _mm_set_epi64x before 2015 */ XXH_ALIGN(16) const xxh_i64 seed64x2[2] = { (xxh_i64)seed64, (xxh_i64)(0U - seed64) }; __m128i const seed = _mm_load_si128((__m128i const*)seed64x2); # else __m128i const seed = _mm_set_epi64x((xxh_i64)(0U - seed64), (xxh_i64)seed64); # endif int i; const void* const src16 = XXH3_kSecret; __m128i* dst16 = (__m128i*) customSecret; # if defined(__GNUC__) || defined(__clang__) /* * On GCC & Clang, marking 'dest' as modified will cause the compiler: * - do not extract the secret from sse registers in the internal loop * - use less common registers, and avoid pushing these reg into stack */ XXH_COMPILER_GUARD(dst16); # endif XXH_ASSERT(((size_t)src16 & 15) == 0); /* control alignment */ XXH_ASSERT(((size_t)dst16 & 15) == 0); for (i=0; i < nbRounds; ++i) { dst16[i] = _mm_add_epi64(_mm_load_si128((const __m128i *)src16+i), seed); } } } #endif #if (XXH_VECTOR == XXH_NEON) /* forward declarations for the scalar routines */ XXH_FORCE_INLINE void XXH3_scalarRound(void* XXH_RESTRICT acc, void const* XXH_RESTRICT input, void const* XXH_RESTRICT secret, size_t lane); XXH_FORCE_INLINE void XXH3_scalarScrambleRound(void* XXH_RESTRICT acc, void const* XXH_RESTRICT secret, size_t lane); /*! * @internal * @brief The bulk processing loop for NEON and WASM SIMD128. * * The NEON code path is actually partially scalar when running on AArch64. This * is to optimize the pipelining and can have up to 15% speedup depending on the * CPU, and it also mitigates some GCC codegen issues. * * @see XXH3_NEON_LANES for configuring this and details about this optimization. * * NEON's 32-bit to 64-bit long multiply takes a half vector of 32-bit * integers instead of the other platforms which mask full 64-bit vectors, * so the setup is more complicated than just shifting right. * * Additionally, there is an optimization for 4 lanes at once noted below. * * Since, as stated, the most optimal amount of lanes for Cortexes is 6, * there needs to be *three* versions of the accumulate operation used * for the remaining 2 lanes. * * WASM's SIMD128 uses SIMDe's arm_neon.h polyfill because the intrinsics overlap * nearly perfectly. */ XXH_FORCE_INLINE void XXH3_accumulate_512_neon( void* XXH_RESTRICT acc, const void* XXH_RESTRICT input, const void* XXH_RESTRICT secret) { XXH_ASSERT((((size_t)acc) & 15) == 0); XXH_STATIC_ASSERT(XXH3_NEON_LANES > 0 && XXH3_NEON_LANES <= XXH_ACC_NB && XXH3_NEON_LANES % 2 == 0); { /* GCC for darwin arm64 does not like aliasing here */ xxh_aliasing_uint64x2_t* const xacc = (xxh_aliasing_uint64x2_t*) acc; /* We don't use a uint32x4_t pointer because it causes bus errors on ARMv7. */ uint8_t const* xinput = (const uint8_t *) input; uint8_t const* xsecret = (const uint8_t *) secret; size_t i; #ifdef __wasm_simd128__ /* * On WASM SIMD128, Clang emits direct address loads when XXH3_kSecret * is constant propagated, which results in it converting it to this * inside the loop: * * a = v128.load(XXH3_kSecret + 0 + $secret_offset, offset = 0) * b = v128.load(XXH3_kSecret + 16 + $secret_offset, offset = 0) * ... * * This requires a full 32-bit address immediate (and therefore a 6 byte * instruction) as well as an add for each offset. * * Putting an asm guard prevents it from folding (at the cost of losing * the alignment hint), and uses the free offset in `v128.load` instead * of adding secret_offset each time which overall reduces code size by * about a kilobyte and improves performance. */ XXH_COMPILER_GUARD(xsecret); #endif /* Scalar lanes use the normal scalarRound routine */ for (i = XXH3_NEON_LANES; i < XXH_ACC_NB; i++) { XXH3_scalarRound(acc, input, secret, i); } i = 0; /* 4 NEON lanes at a time. */ for (; i+1 < XXH3_NEON_LANES / 2; i+=2) { /* data_vec = xinput[i]; */ uint64x2_t data_vec_1 = XXH_vld1q_u64(xinput + (i * 16)); uint64x2_t data_vec_2 = XXH_vld1q_u64(xinput + ((i+1) * 16)); /* key_vec = xsecret[i]; */ uint64x2_t key_vec_1 = XXH_vld1q_u64(xsecret + (i * 16)); uint64x2_t key_vec_2 = XXH_vld1q_u64(xsecret + ((i+1) * 16)); /* data_swap = swap(data_vec) */ uint64x2_t data_swap_1 = vextq_u64(data_vec_1, data_vec_1, 1); uint64x2_t data_swap_2 = vextq_u64(data_vec_2, data_vec_2, 1); /* data_key = data_vec ^ key_vec; */ uint64x2_t data_key_1 = veorq_u64(data_vec_1, key_vec_1); uint64x2_t data_key_2 = veorq_u64(data_vec_2, key_vec_2); /* * If we reinterpret the 64x2 vectors as 32x4 vectors, we can use a * de-interleave operation for 4 lanes in 1 step with `vuzpq_u32` to * get one vector with the low 32 bits of each lane, and one vector * with the high 32 bits of each lane. * * The intrinsic returns a double vector because the original ARMv7-a * instruction modified both arguments in place. AArch64 and SIMD128 emit * two instructions from this intrinsic. * * [ dk11L | dk11H | dk12L | dk12H ] -> [ dk11L | dk12L | dk21L | dk22L ] * [ dk21L | dk21H | dk22L | dk22H ] -> [ dk11H | dk12H | dk21H | dk22H ] */ uint32x4x2_t unzipped = vuzpq_u32( vreinterpretq_u32_u64(data_key_1), vreinterpretq_u32_u64(data_key_2) ); /* data_key_lo = data_key & 0xFFFFFFFF */ uint32x4_t data_key_lo = unzipped.val[0]; /* data_key_hi = data_key >> 32 */ uint32x4_t data_key_hi = unzipped.val[1]; /* * Then, we can split the vectors horizontally and multiply which, as for most * widening intrinsics, have a variant that works on both high half vectors * for free on AArch64. A similar instruction is available on SIMD128. * * sum = data_swap + (u64x2) data_key_lo * (u64x2) data_key_hi */ uint64x2_t sum_1 = XXH_vmlal_low_u32(data_swap_1, data_key_lo, data_key_hi); uint64x2_t sum_2 = XXH_vmlal_high_u32(data_swap_2, data_key_lo, data_key_hi); /* * Clang reorders * a += b * c; // umlal swap.2d, dkl.2s, dkh.2s * c += a; // add acc.2d, acc.2d, swap.2d * to * c += a; // add acc.2d, acc.2d, swap.2d * c += b * c; // umlal acc.2d, dkl.2s, dkh.2s * * While it would make sense in theory since the addition is faster, * for reasons likely related to umlal being limited to certain NEON * pipelines, this is worse. A compiler guard fixes this. */ XXH_COMPILER_GUARD_CLANG_NEON(sum_1); XXH_COMPILER_GUARD_CLANG_NEON(sum_2); /* xacc[i] = acc_vec + sum; */ xacc[i] = vaddq_u64(xacc[i], sum_1); xacc[i+1] = vaddq_u64(xacc[i+1], sum_2); } /* Operate on the remaining NEON lanes 2 at a time. */ for (; i < XXH3_NEON_LANES / 2; i++) { /* data_vec = xinput[i]; */ uint64x2_t data_vec = XXH_vld1q_u64(xinput + (i * 16)); /* key_vec = xsecret[i]; */ uint64x2_t key_vec = XXH_vld1q_u64(xsecret + (i * 16)); /* acc_vec_2 = swap(data_vec) */ uint64x2_t data_swap = vextq_u64(data_vec, data_vec, 1); /* data_key = data_vec ^ key_vec; */ uint64x2_t data_key = veorq_u64(data_vec, key_vec); /* For two lanes, just use VMOVN and VSHRN. */ /* data_key_lo = data_key & 0xFFFFFFFF; */ uint32x2_t data_key_lo = vmovn_u64(data_key); /* data_key_hi = data_key >> 32; */ uint32x2_t data_key_hi = vshrn_n_u64(data_key, 32); /* sum = data_swap + (u64x2) data_key_lo * (u64x2) data_key_hi; */ uint64x2_t sum = vmlal_u32(data_swap, data_key_lo, data_key_hi); /* Same Clang workaround as before */ XXH_COMPILER_GUARD_CLANG_NEON(sum); /* xacc[i] = acc_vec + sum; */ xacc[i] = vaddq_u64 (xacc[i], sum); } } } XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(neon) XXH_FORCE_INLINE void XXH3_scrambleAcc_neon(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) { XXH_ASSERT((((size_t)acc) & 15) == 0); { xxh_aliasing_uint64x2_t* xacc = (xxh_aliasing_uint64x2_t*) acc; uint8_t const* xsecret = (uint8_t const*) secret; size_t i; /* WASM uses operator overloads and doesn't need these. */ #ifndef __wasm_simd128__ /* { prime32_1, prime32_1 } */ uint32x2_t const kPrimeLo = vdup_n_u32(XXH_PRIME32_1); /* { 0, prime32_1, 0, prime32_1 } */ uint32x4_t const kPrimeHi = vreinterpretq_u32_u64(vdupq_n_u64((xxh_u64)XXH_PRIME32_1 << 32)); #endif /* AArch64 uses both scalar and neon at the same time */ for (i = XXH3_NEON_LANES; i < XXH_ACC_NB; i++) { XXH3_scalarScrambleRound(acc, secret, i); } for (i=0; i < XXH3_NEON_LANES / 2; i++) { /* xacc[i] ^= (xacc[i] >> 47); */ uint64x2_t acc_vec = xacc[i]; uint64x2_t shifted = vshrq_n_u64(acc_vec, 47); uint64x2_t data_vec = veorq_u64(acc_vec, shifted); /* xacc[i] ^= xsecret[i]; */ uint64x2_t key_vec = XXH_vld1q_u64(xsecret + (i * 16)); uint64x2_t data_key = veorq_u64(data_vec, key_vec); /* xacc[i] *= XXH_PRIME32_1 */ #ifdef __wasm_simd128__ /* SIMD128 has multiply by u64x2, use it instead of expanding and scalarizing */ xacc[i] = data_key * XXH_PRIME32_1; #else /* * Expanded version with portable NEON intrinsics * * lo(x) * lo(y) + (hi(x) * lo(y) << 32) * * prod_hi = hi(data_key) * lo(prime) << 32 * * Since we only need 32 bits of this multiply a trick can be used, reinterpreting the vector * as a uint32x4_t and multiplying by { 0, prime, 0, prime } to cancel out the unwanted bits * and avoid the shift. */ uint32x4_t prod_hi = vmulq_u32 (vreinterpretq_u32_u64(data_key), kPrimeHi); /* Extract low bits for vmlal_u32 */ uint32x2_t data_key_lo = vmovn_u64(data_key); /* xacc[i] = prod_hi + lo(data_key) * XXH_PRIME32_1; */ xacc[i] = vmlal_u32(vreinterpretq_u64_u32(prod_hi), data_key_lo, kPrimeLo); #endif } } } #endif #if (XXH_VECTOR == XXH_VSX) XXH_FORCE_INLINE void XXH3_accumulate_512_vsx( void* XXH_RESTRICT acc, const void* XXH_RESTRICT input, const void* XXH_RESTRICT secret) { /* presumed aligned */ xxh_aliasing_u64x2* const xacc = (xxh_aliasing_u64x2*) acc; xxh_u8 const* const xinput = (xxh_u8 const*) input; /* no alignment restriction */ xxh_u8 const* const xsecret = (xxh_u8 const*) secret; /* no alignment restriction */ xxh_u64x2 const v32 = { 32, 32 }; size_t i; for (i = 0; i < XXH_STRIPE_LEN / sizeof(xxh_u64x2); i++) { /* data_vec = xinput[i]; */ xxh_u64x2 const data_vec = XXH_vec_loadu(xinput + 16*i); /* key_vec = xsecret[i]; */ xxh_u64x2 const key_vec = XXH_vec_loadu(xsecret + 16*i); xxh_u64x2 const data_key = data_vec ^ key_vec; /* shuffled = (data_key << 32) | (data_key >> 32); */ xxh_u32x4 const shuffled = (xxh_u32x4)vec_rl(data_key, v32); /* product = ((xxh_u64x2)data_key & 0xFFFFFFFF) * ((xxh_u64x2)shuffled & 0xFFFFFFFF); */ xxh_u64x2 const product = XXH_vec_mulo((xxh_u32x4)data_key, shuffled); /* acc_vec = xacc[i]; */ xxh_u64x2 acc_vec = xacc[i]; acc_vec += product; /* swap high and low halves */ #ifdef __s390x__ acc_vec += vec_permi(data_vec, data_vec, 2); #else acc_vec += vec_xxpermdi(data_vec, data_vec, 2); #endif xacc[i] = acc_vec; } } XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(vsx) XXH_FORCE_INLINE void XXH3_scrambleAcc_vsx(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) { XXH_ASSERT((((size_t)acc) & 15) == 0); { xxh_aliasing_u64x2* const xacc = (xxh_aliasing_u64x2*) acc; const xxh_u8* const xsecret = (const xxh_u8*) secret; /* constants */ xxh_u64x2 const v32 = { 32, 32 }; xxh_u64x2 const v47 = { 47, 47 }; xxh_u32x4 const prime = { XXH_PRIME32_1, XXH_PRIME32_1, XXH_PRIME32_1, XXH_PRIME32_1 }; size_t i; for (i = 0; i < XXH_STRIPE_LEN / sizeof(xxh_u64x2); i++) { /* xacc[i] ^= (xacc[i] >> 47); */ xxh_u64x2 const acc_vec = xacc[i]; xxh_u64x2 const data_vec = acc_vec ^ (acc_vec >> v47); /* xacc[i] ^= xsecret[i]; */ xxh_u64x2 const key_vec = XXH_vec_loadu(xsecret + 16*i); xxh_u64x2 const data_key = data_vec ^ key_vec; /* xacc[i] *= XXH_PRIME32_1 */ /* prod_lo = ((xxh_u64x2)data_key & 0xFFFFFFFF) * ((xxh_u64x2)prime & 0xFFFFFFFF); */ xxh_u64x2 const prod_even = XXH_vec_mule((xxh_u32x4)data_key, prime); /* prod_hi = ((xxh_u64x2)data_key >> 32) * ((xxh_u64x2)prime >> 32); */ xxh_u64x2 const prod_odd = XXH_vec_mulo((xxh_u32x4)data_key, prime); xacc[i] = prod_odd + (prod_even << v32); } } } #endif #if (XXH_VECTOR == XXH_SVE) XXH_FORCE_INLINE void XXH3_accumulate_512_sve( void* XXH_RESTRICT acc, const void* XXH_RESTRICT input, const void* XXH_RESTRICT secret) { uint64_t *xacc = (uint64_t *)acc; const uint64_t *xinput = (const uint64_t *)(const void *)input; const uint64_t *xsecret = (const uint64_t *)(const void *)secret; svuint64_t kSwap = sveor_n_u64_z(svptrue_b64(), svindex_u64(0, 1), 1); uint64_t element_count = svcntd(); if (element_count >= 8) { svbool_t mask = svptrue_pat_b64(SV_VL8); svuint64_t vacc = svld1_u64(mask, xacc); ACCRND(vacc, 0); svst1_u64(mask, xacc, vacc); } else if (element_count == 2) { /* sve128 */ svbool_t mask = svptrue_pat_b64(SV_VL2); svuint64_t acc0 = svld1_u64(mask, xacc + 0); svuint64_t acc1 = svld1_u64(mask, xacc + 2); svuint64_t acc2 = svld1_u64(mask, xacc + 4); svuint64_t acc3 = svld1_u64(mask, xacc + 6); ACCRND(acc0, 0); ACCRND(acc1, 2); ACCRND(acc2, 4); ACCRND(acc3, 6); svst1_u64(mask, xacc + 0, acc0); svst1_u64(mask, xacc + 2, acc1); svst1_u64(mask, xacc + 4, acc2); svst1_u64(mask, xacc + 6, acc3); } else { svbool_t mask = svptrue_pat_b64(SV_VL4); svuint64_t acc0 = svld1_u64(mask, xacc + 0); svuint64_t acc1 = svld1_u64(mask, xacc + 4); ACCRND(acc0, 0); ACCRND(acc1, 4); svst1_u64(mask, xacc + 0, acc0); svst1_u64(mask, xacc + 4, acc1); } } XXH_FORCE_INLINE void XXH3_accumulate_sve(xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT input, const xxh_u8* XXH_RESTRICT secret, size_t nbStripes) { if (nbStripes != 0) { uint64_t *xacc = (uint64_t *)acc; const uint64_t *xinput = (const uint64_t *)(const void *)input; const uint64_t *xsecret = (const uint64_t *)(const void *)secret; svuint64_t kSwap = sveor_n_u64_z(svptrue_b64(), svindex_u64(0, 1), 1); uint64_t element_count = svcntd(); if (element_count >= 8) { svbool_t mask = svptrue_pat_b64(SV_VL8); svuint64_t vacc = svld1_u64(mask, xacc + 0); do { /* svprfd(svbool_t, void *, enum svfprop); */ svprfd(mask, xinput + 128, SV_PLDL1STRM); ACCRND(vacc, 0); xinput += 8; xsecret += 1; nbStripes--; } while (nbStripes != 0); svst1_u64(mask, xacc + 0, vacc); } else if (element_count == 2) { /* sve128 */ svbool_t mask = svptrue_pat_b64(SV_VL2); svuint64_t acc0 = svld1_u64(mask, xacc + 0); svuint64_t acc1 = svld1_u64(mask, xacc + 2); svuint64_t acc2 = svld1_u64(mask, xacc + 4); svuint64_t acc3 = svld1_u64(mask, xacc + 6); do { svprfd(mask, xinput + 128, SV_PLDL1STRM); ACCRND(acc0, 0); ACCRND(acc1, 2); ACCRND(acc2, 4); ACCRND(acc3, 6); xinput += 8; xsecret += 1; nbStripes--; } while (nbStripes != 0); svst1_u64(mask, xacc + 0, acc0); svst1_u64(mask, xacc + 2, acc1); svst1_u64(mask, xacc + 4, acc2); svst1_u64(mask, xacc + 6, acc3); } else { svbool_t mask = svptrue_pat_b64(SV_VL4); svuint64_t acc0 = svld1_u64(mask, xacc + 0); svuint64_t acc1 = svld1_u64(mask, xacc + 4); do { svprfd(mask, xinput + 128, SV_PLDL1STRM); ACCRND(acc0, 0); ACCRND(acc1, 4); xinput += 8; xsecret += 1; nbStripes--; } while (nbStripes != 0); svst1_u64(mask, xacc + 0, acc0); svst1_u64(mask, xacc + 4, acc1); } } } #endif #if (XXH_VECTOR == XXH_LSX) #define _LSX_SHUFFLE(z, y, x, w) (((z) << 6) | ((y) << 4) | ((x) << 2) | (w)) XXH_FORCE_INLINE void XXH3_accumulate_512_lsx( void* XXH_RESTRICT acc, const void* XXH_RESTRICT input, const void* XXH_RESTRICT secret) { XXH_ASSERT((((size_t)acc) & 15) == 0); { __m128i* const xacc = (__m128i *) acc; const __m128i* const xinput = (const __m128i *) input; const __m128i* const xsecret = (const __m128i *) secret; for (size_t i = 0; i < XXH_STRIPE_LEN / sizeof(__m128i); i++) { /* data_vec = xinput[i]; */ __m128i const data_vec = __lsx_vld(xinput + i, 0); /* key_vec = xsecret[i]; */ __m128i const key_vec = __lsx_vld(xsecret + i, 0); /* data_key = data_vec ^ key_vec; */ __m128i const data_key = __lsx_vxor_v(data_vec, key_vec); /* data_key_lo = data_key >> 32; */ __m128i const data_key_lo = __lsx_vsrli_d(data_key, 32); // __m128i const data_key_lo = __lsx_vsrli_d(data_key, 32); /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ __m128i const product = __lsx_vmulwev_d_wu(data_key, data_key_lo); /* xacc[i] += swap(data_vec); */ __m128i const data_swap = __lsx_vshuf4i_w(data_vec, _LSX_SHUFFLE(1, 0, 3, 2)); __m128i const sum = __lsx_vadd_d(xacc[i], data_swap); /* xacc[i] += product; */ xacc[i] = __lsx_vadd_d(product, sum); } } } XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(lsx) XXH_FORCE_INLINE void XXH3_scrambleAcc_lsx(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) { XXH_ASSERT((((size_t)acc) & 15) == 0); { __m128i* const xacc = (__m128i*) acc; const __m128i* const xsecret = (const __m128i *) secret; const __m128i prime32 = __lsx_vreplgr2vr_w((int)XXH_PRIME32_1); for (size_t i = 0; i < XXH_STRIPE_LEN / sizeof(__m128i); i++) { /* xacc[i] ^= (xacc[i] >> 47) */ __m128i const acc_vec = xacc[i]; __m128i const shifted = __lsx_vsrli_d(acc_vec, 47); __m128i const data_vec = __lsx_vxor_v(acc_vec, shifted); /* xacc[i] ^= xsecret[i]; */ __m128i const key_vec = __lsx_vld(xsecret + i, 0); __m128i const data_key = __lsx_vxor_v(data_vec, key_vec); /* xacc[i] *= XXH_PRIME32_1; */ __m128i const data_key_hi = __lsx_vsrli_d(data_key, 32); __m128i const prod_lo = __lsx_vmulwev_d_wu(data_key, prime32); __m128i const prod_hi = __lsx_vmulwev_d_wu(data_key_hi, prime32); xacc[i] = __lsx_vadd_d(prod_lo, __lsx_vslli_d(prod_hi, 32)); } } } #endif /* scalar variants - universal */ #if defined(__aarch64__) && (defined(__GNUC__) || defined(__clang__)) /* * In XXH3_scalarRound(), GCC and Clang have a similar codegen issue, where they * emit an excess mask and a full 64-bit multiply-add (MADD X-form). * * While this might not seem like much, as AArch64 is a 64-bit architecture, only * big Cortex designs have a full 64-bit multiplier. * * On the little cores, the smaller 32-bit multiplier is used, and full 64-bit * multiplies expand to 2-3 multiplies in microcode. This has a major penalty * of up to 4 latency cycles and 2 stall cycles in the multiply pipeline. * * Thankfully, AArch64 still provides the 32-bit long multiply-add (UMADDL) which does * not have this penalty and does the mask automatically. */ XXH_FORCE_INLINE xxh_u64 XXH_mult32to64_add64(xxh_u64 lhs, xxh_u64 rhs, xxh_u64 acc) { xxh_u64 ret; /* note: %x = 64-bit register, %w = 32-bit register */ __asm__("umaddl %x0, %w1, %w2, %x3" : "=r" (ret) : "r" (lhs), "r" (rhs), "r" (acc)); return ret; } #else XXH_FORCE_INLINE xxh_u64 XXH_mult32to64_add64(xxh_u64 lhs, xxh_u64 rhs, xxh_u64 acc) { return XXH_mult32to64((xxh_u32)lhs, (xxh_u32)rhs) + acc; } #endif /*! * @internal * @brief Scalar round for @ref XXH3_accumulate_512_scalar(). * * This is extracted to its own function because the NEON path uses a combination * of NEON and scalar. */ XXH_FORCE_INLINE void XXH3_scalarRound(void* XXH_RESTRICT acc, void const* XXH_RESTRICT input, void const* XXH_RESTRICT secret, size_t lane) { xxh_u64* xacc = (xxh_u64*) acc; xxh_u8 const* xinput = (xxh_u8 const*) input; xxh_u8 const* xsecret = (xxh_u8 const*) secret; XXH_ASSERT(lane < XXH_ACC_NB); XXH_ASSERT(((size_t)acc & (XXH_ACC_ALIGN-1)) == 0); { xxh_u64 const data_val = XXH_readLE64(xinput + lane * 8); xxh_u64 const data_key = data_val ^ XXH_readLE64(xsecret + lane * 8); xacc[lane ^ 1] += data_val; /* swap adjacent lanes */ xacc[lane] = XXH_mult32to64_add64(data_key /* & 0xFFFFFFFF */, data_key >> 32, xacc[lane]); } } /*! * @internal * @brief Processes a 64 byte block of data using the scalar path. */ XXH_FORCE_INLINE void XXH3_accumulate_512_scalar(void* XXH_RESTRICT acc, const void* XXH_RESTRICT input, const void* XXH_RESTRICT secret) { size_t i; /* ARM GCC refuses to unroll this loop, resulting in a 24% slowdown on ARMv6. */ #if defined(__GNUC__) && !defined(__clang__) \ && (defined(__arm__) || defined(__thumb2__)) \ && defined(__ARM_FEATURE_UNALIGNED) /* no unaligned access just wastes bytes */ \ && XXH_SIZE_OPT <= 0 # pragma GCC unroll 8 #endif for (i=0; i < XXH_ACC_NB; i++) { XXH3_scalarRound(acc, input, secret, i); } } XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(scalar) /*! * @internal * @brief Scalar scramble step for @ref XXH3_scrambleAcc_scalar(). * * This is extracted to its own function because the NEON path uses a combination * of NEON and scalar. */ XXH_FORCE_INLINE void XXH3_scalarScrambleRound(void* XXH_RESTRICT acc, void const* XXH_RESTRICT secret, size_t lane) { xxh_u64* const xacc = (xxh_u64*) acc; /* presumed aligned */ const xxh_u8* const xsecret = (const xxh_u8*) secret; /* no alignment restriction */ XXH_ASSERT((((size_t)acc) & (XXH_ACC_ALIGN-1)) == 0); XXH_ASSERT(lane < XXH_ACC_NB); { xxh_u64 const key64 = XXH_readLE64(xsecret + lane * 8); xxh_u64 acc64 = xacc[lane]; acc64 = XXH_xorshift64(acc64, 47); acc64 ^= key64; acc64 *= XXH_PRIME32_1; xacc[lane] = acc64; } } /*! * @internal * @brief Scrambles the accumulators after a large chunk has been read */ XXH_FORCE_INLINE void XXH3_scrambleAcc_scalar(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) { size_t i; for (i=0; i < XXH_ACC_NB; i++) { XXH3_scalarScrambleRound(acc, secret, i); } } XXH_FORCE_INLINE void XXH3_initCustomSecret_scalar(void* XXH_RESTRICT customSecret, xxh_u64 seed64) { /* * We need a separate pointer for the hack below, * which requires a non-const pointer. * Any decent compiler will optimize this out otherwise. */ const xxh_u8* kSecretPtr = XXH3_kSecret; XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 15) == 0); #if defined(__GNUC__) && defined(__aarch64__) /* * UGLY HACK: * GCC and Clang generate a bunch of MOV/MOVK pairs for aarch64, and they are * placed sequentially, in order, at the top of the unrolled loop. * * While MOVK is great for generating constants (2 cycles for a 64-bit * constant compared to 4 cycles for LDR), it fights for bandwidth with * the arithmetic instructions. * * I L S * MOVK * MOVK * MOVK * MOVK * ADD * SUB STR * STR * By forcing loads from memory (as the asm line causes the compiler to assume * that XXH3_kSecretPtr has been changed), the pipelines are used more * efficiently: * I L S * LDR * ADD LDR * SUB STR * STR * * See XXH3_NEON_LANES for details on the pipsline. * * XXH3_64bits_withSeed, len == 256, Snapdragon 835 * without hack: 2654.4 MB/s * with hack: 3202.9 MB/s */ XXH_COMPILER_GUARD(kSecretPtr); #endif { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / 16; int i; for (i=0; i < nbRounds; i++) { /* * The asm hack causes the compiler to assume that kSecretPtr aliases with * customSecret, and on aarch64, this prevented LDP from merging two * loads together for free. Putting the loads together before the stores * properly generates LDP. */ xxh_u64 lo = XXH_readLE64(kSecretPtr + 16*i) + seed64; xxh_u64 hi = XXH_readLE64(kSecretPtr + 16*i + 8) - seed64; XXH_writeLE64((xxh_u8*)customSecret + 16*i, lo); XXH_writeLE64((xxh_u8*)customSecret + 16*i + 8, hi); } } } typedef void (*XXH3_f_accumulate)(xxh_u64* XXH_RESTRICT, const xxh_u8* XXH_RESTRICT, const xxh_u8* XXH_RESTRICT, size_t); typedef void (*XXH3_f_scrambleAcc)(void* XXH_RESTRICT, const void*); typedef void (*XXH3_f_initCustomSecret)(void* XXH_RESTRICT, xxh_u64); #if (XXH_VECTOR == XXH_AVX512) #define XXH3_accumulate_512 XXH3_accumulate_512_avx512 #define XXH3_accumulate XXH3_accumulate_avx512 #define XXH3_scrambleAcc XXH3_scrambleAcc_avx512 #define XXH3_initCustomSecret XXH3_initCustomSecret_avx512 #elif (XXH_VECTOR == XXH_AVX2) #define XXH3_accumulate_512 XXH3_accumulate_512_avx2 #define XXH3_accumulate XXH3_accumulate_avx2 #define XXH3_scrambleAcc XXH3_scrambleAcc_avx2 #define XXH3_initCustomSecret XXH3_initCustomSecret_avx2 #elif (XXH_VECTOR == XXH_SSE2) #define XXH3_accumulate_512 XXH3_accumulate_512_sse2 #define XXH3_accumulate XXH3_accumulate_sse2 #define XXH3_scrambleAcc XXH3_scrambleAcc_sse2 #define XXH3_initCustomSecret XXH3_initCustomSecret_sse2 #elif (XXH_VECTOR == XXH_NEON) #define XXH3_accumulate_512 XXH3_accumulate_512_neon #define XXH3_accumulate XXH3_accumulate_neon #define XXH3_scrambleAcc XXH3_scrambleAcc_neon #define XXH3_initCustomSecret XXH3_initCustomSecret_scalar #elif (XXH_VECTOR == XXH_VSX) #define XXH3_accumulate_512 XXH3_accumulate_512_vsx #define XXH3_accumulate XXH3_accumulate_vsx #define XXH3_scrambleAcc XXH3_scrambleAcc_vsx #define XXH3_initCustomSecret XXH3_initCustomSecret_scalar #elif (XXH_VECTOR == XXH_SVE) #define XXH3_accumulate_512 XXH3_accumulate_512_sve #define XXH3_accumulate XXH3_accumulate_sve #define XXH3_scrambleAcc XXH3_scrambleAcc_scalar #define XXH3_initCustomSecret XXH3_initCustomSecret_scalar #elif (XXH_VECTOR == XXH_LSX) #define XXH3_accumulate_512 XXH3_accumulate_512_lsx #define XXH3_accumulate XXH3_accumulate_lsx #define XXH3_scrambleAcc XXH3_scrambleAcc_lsx #define XXH3_initCustomSecret XXH3_initCustomSecret_scalar #else /* scalar */ #define XXH3_accumulate_512 XXH3_accumulate_512_scalar #define XXH3_accumulate XXH3_accumulate_scalar #define XXH3_scrambleAcc XXH3_scrambleAcc_scalar #define XXH3_initCustomSecret XXH3_initCustomSecret_scalar #endif #if XXH_SIZE_OPT >= 1 /* don't do SIMD for initialization */ # undef XXH3_initCustomSecret # define XXH3_initCustomSecret XXH3_initCustomSecret_scalar #endif XXH_FORCE_INLINE void XXH3_hashLong_internal_loop(xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT input, size_t len, const xxh_u8* XXH_RESTRICT secret, size_t secretSize, XXH3_f_accumulate f_acc, XXH3_f_scrambleAcc f_scramble) { size_t const nbStripesPerBlock = (secretSize - XXH_STRIPE_LEN) / XXH_SECRET_CONSUME_RATE; size_t const block_len = XXH_STRIPE_LEN * nbStripesPerBlock; size_t const nb_blocks = (len - 1) / block_len; size_t n; XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); for (n = 0; n < nb_blocks; n++) { f_acc(acc, input + n*block_len, secret, nbStripesPerBlock); f_scramble(acc, secret + secretSize - XXH_STRIPE_LEN); } /* last partial block */ XXH_ASSERT(len > XXH_STRIPE_LEN); { size_t const nbStripes = ((len - 1) - (block_len * nb_blocks)) / XXH_STRIPE_LEN; XXH_ASSERT(nbStripes <= (secretSize / XXH_SECRET_CONSUME_RATE)); f_acc(acc, input + nb_blocks*block_len, secret, nbStripes); /* last stripe */ { const xxh_u8* const p = input + len - XXH_STRIPE_LEN; #define XXH_SECRET_LASTACC_START 7 /* not aligned on 8, last secret is different from acc & scrambler */ XXH3_accumulate_512(acc, p, secret + secretSize - XXH_STRIPE_LEN - XXH_SECRET_LASTACC_START); } } } XXH_FORCE_INLINE xxh_u64 XXH3_mix2Accs(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret) { return XXH3_mul128_fold64( acc[0] ^ XXH_readLE64(secret), acc[1] ^ XXH_readLE64(secret+8) ); } static XXH_PUREF XXH64_hash_t XXH3_mergeAccs(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret, xxh_u64 start) { xxh_u64 result64 = start; size_t i = 0; for (i = 0; i < 4; i++) { result64 += XXH3_mix2Accs(acc+2*i, secret + 16*i); #if defined(__clang__) /* Clang */ \ && (defined(__arm__) || defined(__thumb__)) /* ARMv7 */ \ && (defined(__ARM_NEON) || defined(__ARM_NEON__)) /* NEON */ \ && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable */ /* * UGLY HACK: * Prevent autovectorization on Clang ARMv7-a. Exact same problem as * the one in XXH3_len_129to240_64b. Speeds up shorter keys > 240b. * XXH3_64bits, len == 256, Snapdragon 835: * without hack: 2063.7 MB/s * with hack: 2560.7 MB/s */ XXH_COMPILER_GUARD(result64); #endif } return XXH3_avalanche(result64); } /* do not align on 8, so that the secret is different from the accumulator */ #define XXH_SECRET_MERGEACCS_START 11 static XXH_PUREF XXH64_hash_t XXH3_finalizeLong_64b(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret, xxh_u64 len) { return XXH3_mergeAccs(acc, secret + XXH_SECRET_MERGEACCS_START, len * XXH_PRIME64_1); } #define XXH3_INIT_ACC { XXH_PRIME32_3, XXH_PRIME64_1, XXH_PRIME64_2, XXH_PRIME64_3, \ XXH_PRIME64_4, XXH_PRIME32_2, XXH_PRIME64_5, XXH_PRIME32_1 } XXH_FORCE_INLINE XXH64_hash_t XXH3_hashLong_64b_internal(const void* XXH_RESTRICT input, size_t len, const void* XXH_RESTRICT secret, size_t secretSize, XXH3_f_accumulate f_acc, XXH3_f_scrambleAcc f_scramble) { XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[XXH_ACC_NB] = XXH3_INIT_ACC; XXH3_hashLong_internal_loop(acc, (const xxh_u8*)input, len, (const xxh_u8*)secret, secretSize, f_acc, f_scramble); /* converge into final hash */ XXH_STATIC_ASSERT(sizeof(acc) == 64); XXH_ASSERT(secretSize >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); return XXH3_finalizeLong_64b(acc, (const xxh_u8*)secret, (xxh_u64)len); } /* * It's important for performance to transmit secret's size (when it's static) * so that the compiler can properly optimize the vectorized loop. * This makes a big performance difference for "medium" keys (<1 KB) when using AVX instruction set. * When the secret size is unknown, or on GCC 12 where the mix of NO_INLINE and FORCE_INLINE * breaks -Og, this is XXH_NO_INLINE. */ XXH3_WITH_SECRET_INLINE XXH64_hash_t XXH3_hashLong_64b_withSecret(const void* XXH_RESTRICT input, size_t len, XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) { (void)seed64; return XXH3_hashLong_64b_internal(input, len, secret, secretLen, XXH3_accumulate, XXH3_scrambleAcc); } /* * It's preferable for performance that XXH3_hashLong is not inlined, * as it results in a smaller function for small data, easier to the instruction cache. * Note that inside this no_inline function, we do inline the internal loop, * and provide a statically defined secret size to allow optimization of vector loop. */ XXH_NO_INLINE XXH_PUREF XXH64_hash_t XXH3_hashLong_64b_default(const void* XXH_RESTRICT input, size_t len, XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) { (void)seed64; (void)secret; (void)secretLen; return XXH3_hashLong_64b_internal(input, len, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_accumulate, XXH3_scrambleAcc); } /* * XXH3_hashLong_64b_withSeed(): * Generate a custom key based on alteration of default XXH3_kSecret with the seed, * and then use this key for long mode hashing. * * This operation is decently fast but nonetheless costs a little bit of time. * Try to avoid it whenever possible (typically when seed==0). * * It's important for performance that XXH3_hashLong is not inlined. Not sure * why (uop cache maybe?), but the difference is large and easily measurable. */ XXH_FORCE_INLINE XXH64_hash_t XXH3_hashLong_64b_withSeed_internal(const void* input, size_t len, XXH64_hash_t seed, XXH3_f_accumulate f_acc, XXH3_f_scrambleAcc f_scramble, XXH3_f_initCustomSecret f_initSec) { #if XXH_SIZE_OPT <= 0 if (seed == 0) return XXH3_hashLong_64b_internal(input, len, XXH3_kSecret, sizeof(XXH3_kSecret), f_acc, f_scramble); #endif { XXH_ALIGN(XXH_SEC_ALIGN) xxh_u8 secret[XXH_SECRET_DEFAULT_SIZE]; f_initSec(secret, seed); return XXH3_hashLong_64b_internal(input, len, secret, sizeof(secret), f_acc, f_scramble); } } /* * It's important for performance that XXH3_hashLong is not inlined. */ XXH_NO_INLINE XXH64_hash_t XXH3_hashLong_64b_withSeed(const void* XXH_RESTRICT input, size_t len, XXH64_hash_t seed, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) { (void)secret; (void)secretLen; return XXH3_hashLong_64b_withSeed_internal(input, len, seed, XXH3_accumulate, XXH3_scrambleAcc, XXH3_initCustomSecret); } typedef XXH64_hash_t (*XXH3_hashLong64_f)(const void* XXH_RESTRICT, size_t, XXH64_hash_t, const xxh_u8* XXH_RESTRICT, size_t); XXH_FORCE_INLINE XXH64_hash_t XXH3_64bits_internal(const void* XXH_RESTRICT input, size_t len, XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen, XXH3_hashLong64_f f_hashLong) { XXH_ASSERT(secretLen >= XXH3_SECRET_SIZE_MIN); /* * If an action is to be taken if `secretLen` condition is not respected, * it should be done here. * For now, it's a contract pre-condition. * Adding a check and a branch here would cost performance at every hash. * Also, note that function signature doesn't offer room to return an error. */ if (len <= 16) return XXH3_len_0to16_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, seed64); if (len <= 128) return XXH3_len_17to128_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); if (len <= XXH3_MIDSIZE_MAX) return XXH3_len_129to240_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); return f_hashLong(input, len, seed64, (const xxh_u8*)secret, secretLen); } /* === Public entry point === */ /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH64_hash_t XXH3_64bits(XXH_NOESCAPE const void* input, size_t length) { return XXH3_64bits_internal(input, length, 0, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_64b_default); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_withSecret(XXH_NOESCAPE const void* input, size_t length, XXH_NOESCAPE const void* secret, size_t secretSize) { return XXH3_64bits_internal(input, length, 0, secret, secretSize, XXH3_hashLong_64b_withSecret); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_withSeed(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed) { return XXH3_64bits_internal(input, length, seed, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_64b_withSeed); } XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t length, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) { if (length <= XXH3_MIDSIZE_MAX) return XXH3_64bits_internal(input, length, seed, XXH3_kSecret, sizeof(XXH3_kSecret), NULL); return XXH3_hashLong_64b_withSecret(input, length, seed, (const xxh_u8*)secret, secretSize); } /* === XXH3 streaming === */ #ifndef XXH_NO_STREAM /* * Malloc's a pointer that is always aligned to @align. * * This must be freed with `XXH_alignedFree()`. * * malloc typically guarantees 16 byte alignment on 64-bit systems and 8 byte * alignment on 32-bit. This isn't enough for the 32 byte aligned loads in AVX2 * or on 32-bit, the 16 byte aligned loads in SSE2 and NEON. * * This underalignment previously caused a rather obvious crash which went * completely unnoticed due to XXH3_createState() not actually being tested. * Credit to RedSpah for noticing this bug. * * The alignment is done manually: Functions like posix_memalign or _mm_malloc * are avoided: To maintain portability, we would have to write a fallback * like this anyways, and besides, testing for the existence of library * functions without relying on external build tools is impossible. * * The method is simple: Overallocate, manually align, and store the offset * to the original behind the returned pointer. * * Align must be a power of 2 and 8 <= align <= 128. */ static XXH_MALLOCF void* XXH_alignedMalloc(size_t s, size_t align) { XXH_ASSERT(align <= 128 && align >= 8); /* range check */ XXH_ASSERT((align & (align-1)) == 0); /* power of 2 */ XXH_ASSERT(s != 0 && s < (s + align)); /* empty/overflow */ { /* Overallocate to make room for manual realignment and an offset byte */ xxh_u8* base = (xxh_u8*)XXH_malloc(s + align); if (base != NULL) { /* * Get the offset needed to align this pointer. * * Even if the returned pointer is aligned, there will always be * at least one byte to store the offset to the original pointer. */ size_t offset = align - ((size_t)base & (align - 1)); /* base % align */ /* Add the offset for the now-aligned pointer */ xxh_u8* ptr = base + offset; XXH_ASSERT((size_t)ptr % align == 0); /* Store the offset immediately before the returned pointer. */ ptr[-1] = (xxh_u8)offset; return ptr; } return NULL; } } /* * Frees an aligned pointer allocated by XXH_alignedMalloc(). Don't pass * normal malloc'd pointers, XXH_alignedMalloc has a specific data layout. */ static void XXH_alignedFree(void* p) { if (p != NULL) { xxh_u8* ptr = (xxh_u8*)p; /* Get the offset byte we added in XXH_malloc. */ xxh_u8 offset = ptr[-1]; /* Free the original malloc'd pointer */ xxh_u8* base = ptr - offset; XXH_free(base); } } /*! @ingroup XXH3_family */ /*! * @brief Allocate an @ref XXH3_state_t. * * @return An allocated pointer of @ref XXH3_state_t on success. * @return `NULL` on failure. * * @note Must be freed with XXH3_freeState(). * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH3_state_t* XXH3_createState(void) { XXH3_state_t* const state = (XXH3_state_t*)XXH_alignedMalloc(sizeof(XXH3_state_t), 64); if (state==NULL) return NULL; XXH3_INITSTATE(state); return state; } /*! @ingroup XXH3_family */ /*! * @brief Frees an @ref XXH3_state_t. * * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). * * @return @ref XXH_OK. * * @note Must be allocated with XXH3_createState(). * * @see @ref streaming_example "Streaming Example" */ XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr) { XXH_alignedFree(statePtr); return XXH_OK; } /*! @ingroup XXH3_family */ XXH_PUBLIC_API void XXH3_copyState(XXH_NOESCAPE XXH3_state_t* dst_state, XXH_NOESCAPE const XXH3_state_t* src_state) { XXH_memcpy(dst_state, src_state, sizeof(*dst_state)); } static void XXH3_reset_internal(XXH3_state_t* statePtr, XXH64_hash_t seed, const void* secret, size_t secretSize) { size_t const initStart = offsetof(XXH3_state_t, bufferedSize); size_t const initLength = offsetof(XXH3_state_t, nbStripesPerBlock) - initStart; XXH_ASSERT(offsetof(XXH3_state_t, nbStripesPerBlock) > initStart); XXH_ASSERT(statePtr != NULL); /* set members from bufferedSize to nbStripesPerBlock (excluded) to 0 */ memset((char*)statePtr + initStart, 0, initLength); statePtr->acc[0] = XXH_PRIME32_3; statePtr->acc[1] = XXH_PRIME64_1; statePtr->acc[2] = XXH_PRIME64_2; statePtr->acc[3] = XXH_PRIME64_3; statePtr->acc[4] = XXH_PRIME64_4; statePtr->acc[5] = XXH_PRIME32_2; statePtr->acc[6] = XXH_PRIME64_5; statePtr->acc[7] = XXH_PRIME32_1; statePtr->seed = seed; statePtr->useSeed = (seed != 0); statePtr->extSecret = (const unsigned char*)secret; XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); statePtr->secretLimit = secretSize - XXH_STRIPE_LEN; statePtr->nbStripesPerBlock = statePtr->secretLimit / XXH_SECRET_CONSUME_RATE; } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr) { if (statePtr == NULL) return XXH_ERROR; XXH3_reset_internal(statePtr, 0, XXH3_kSecret, XXH_SECRET_DEFAULT_SIZE); return XXH_OK; } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize) { if (statePtr == NULL) return XXH_ERROR; XXH3_reset_internal(statePtr, 0, secret, secretSize); if (secret == NULL) return XXH_ERROR; if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; return XXH_OK; } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed) { if (statePtr == NULL) return XXH_ERROR; if (seed==0) return XXH3_64bits_reset(statePtr); if ((seed != statePtr->seed) || (statePtr->extSecret != NULL)) XXH3_initCustomSecret(statePtr->customSecret, seed); XXH3_reset_internal(statePtr, seed, NULL, XXH_SECRET_DEFAULT_SIZE); return XXH_OK; } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed64) { if (statePtr == NULL) return XXH_ERROR; if (secret == NULL) return XXH_ERROR; if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; XXH3_reset_internal(statePtr, seed64, secret, secretSize); statePtr->useSeed = 1; /* always, even if seed64==0 */ return XXH_OK; } /*! * @internal * @brief Processes a large input for XXH3_update() and XXH3_digest_long(). * * Unlike XXH3_hashLong_internal_loop(), this can process data that overlaps a block. * * @param acc Pointer to the 8 accumulator lanes * @param nbStripesSoFarPtr In/out pointer to the number of leftover stripes in the block* * @param nbStripesPerBlock Number of stripes in a block * @param input Input pointer * @param nbStripes Number of stripes to process * @param secret Secret pointer * @param secretLimit Offset of the last block in @p secret * @param f_acc Pointer to an XXH3_accumulate implementation * @param f_scramble Pointer to an XXH3_scrambleAcc implementation * @return Pointer past the end of @p input after processing */ XXH_FORCE_INLINE const xxh_u8 * XXH3_consumeStripes(xxh_u64* XXH_RESTRICT acc, size_t* XXH_RESTRICT nbStripesSoFarPtr, size_t nbStripesPerBlock, const xxh_u8* XXH_RESTRICT input, size_t nbStripes, const xxh_u8* XXH_RESTRICT secret, size_t secretLimit, XXH3_f_accumulate f_acc, XXH3_f_scrambleAcc f_scramble) { const xxh_u8* initialSecret = secret + *nbStripesSoFarPtr * XXH_SECRET_CONSUME_RATE; /* Process full blocks */ if (nbStripes >= (nbStripesPerBlock - *nbStripesSoFarPtr)) { /* Process the initial partial block... */ size_t nbStripesThisIter = nbStripesPerBlock - *nbStripesSoFarPtr; do { /* Accumulate and scramble */ f_acc(acc, input, initialSecret, nbStripesThisIter); f_scramble(acc, secret + secretLimit); input += nbStripesThisIter * XXH_STRIPE_LEN; nbStripes -= nbStripesThisIter; /* Then continue the loop with the full block size */ nbStripesThisIter = nbStripesPerBlock; initialSecret = secret; } while (nbStripes >= nbStripesPerBlock); *nbStripesSoFarPtr = 0; } /* Process a partial block */ if (nbStripes > 0) { f_acc(acc, input, initialSecret, nbStripes); input += nbStripes * XXH_STRIPE_LEN; *nbStripesSoFarPtr += nbStripes; } /* Return end pointer */ return input; } #ifndef XXH3_STREAM_USE_STACK # if XXH_SIZE_OPT <= 0 && !defined(__clang__) /* clang doesn't need additional stack space */ # define XXH3_STREAM_USE_STACK 1 # endif #endif /* * Both XXH3_64bits_update and XXH3_128bits_update use this routine. */ XXH_FORCE_INLINE XXH_errorcode XXH3_update(XXH3_state_t* XXH_RESTRICT const state, const xxh_u8* XXH_RESTRICT input, size_t len, XXH3_f_accumulate f_acc, XXH3_f_scrambleAcc f_scramble) { if (input==NULL) { XXH_ASSERT(len == 0); return XXH_OK; } XXH_ASSERT(state != NULL); { const xxh_u8* const bEnd = input + len; const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; #if defined(XXH3_STREAM_USE_STACK) && XXH3_STREAM_USE_STACK >= 1 /* For some reason, gcc and MSVC seem to suffer greatly * when operating accumulators directly into state. * Operating into stack space seems to enable proper optimization. * clang, on the other hand, doesn't seem to need this trick */ XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[8]; XXH_memcpy(acc, state->acc, sizeof(acc)); #else xxh_u64* XXH_RESTRICT const acc = state->acc; #endif state->totalLen += len; XXH_ASSERT(state->bufferedSize <= XXH3_INTERNALBUFFER_SIZE); /* small input : just fill in tmp buffer */ if (len <= XXH3_INTERNALBUFFER_SIZE - state->bufferedSize) { XXH_memcpy(state->buffer + state->bufferedSize, input, len); state->bufferedSize += (XXH32_hash_t)len; return XXH_OK; } /* total input is now > XXH3_INTERNALBUFFER_SIZE */ #define XXH3_INTERNALBUFFER_STRIPES (XXH3_INTERNALBUFFER_SIZE / XXH_STRIPE_LEN) XXH_STATIC_ASSERT(XXH3_INTERNALBUFFER_SIZE % XXH_STRIPE_LEN == 0); /* clean multiple */ /* * Internal buffer is partially filled (always, except at beginning) * Complete it, then consume it. */ if (state->bufferedSize) { size_t const loadSize = XXH3_INTERNALBUFFER_SIZE - state->bufferedSize; XXH_memcpy(state->buffer + state->bufferedSize, input, loadSize); input += loadSize; XXH3_consumeStripes(acc, &state->nbStripesSoFar, state->nbStripesPerBlock, state->buffer, XXH3_INTERNALBUFFER_STRIPES, secret, state->secretLimit, f_acc, f_scramble); state->bufferedSize = 0; } XXH_ASSERT(input < bEnd); if (bEnd - input > XXH3_INTERNALBUFFER_SIZE) { size_t nbStripes = (size_t)(bEnd - 1 - input) / XXH_STRIPE_LEN; input = XXH3_consumeStripes(acc, &state->nbStripesSoFar, state->nbStripesPerBlock, input, nbStripes, secret, state->secretLimit, f_acc, f_scramble); XXH_memcpy(state->buffer + sizeof(state->buffer) - XXH_STRIPE_LEN, input - XXH_STRIPE_LEN, XXH_STRIPE_LEN); } /* Some remaining input (always) : buffer it */ XXH_ASSERT(input < bEnd); XXH_ASSERT(bEnd - input <= XXH3_INTERNALBUFFER_SIZE); XXH_ASSERT(state->bufferedSize == 0); XXH_memcpy(state->buffer, input, (size_t)(bEnd-input)); state->bufferedSize = (XXH32_hash_t)(bEnd-input); #if defined(XXH3_STREAM_USE_STACK) && XXH3_STREAM_USE_STACK >= 1 /* save stack accumulators into state */ XXH_memcpy(state->acc, acc, sizeof(acc)); #endif } return XXH_OK; } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len) { return XXH3_update(state, (const xxh_u8*)input, len, XXH3_accumulate, XXH3_scrambleAcc); } XXH_FORCE_INLINE void XXH3_digest_long (XXH64_hash_t* acc, const XXH3_state_t* state, const unsigned char* secret) { xxh_u8 lastStripe[XXH_STRIPE_LEN]; const xxh_u8* lastStripePtr; /* * Digest on a local copy. This way, the state remains unaltered, and it can * continue ingesting more input afterwards. */ XXH_memcpy(acc, state->acc, sizeof(state->acc)); if (state->bufferedSize >= XXH_STRIPE_LEN) { /* Consume remaining stripes then point to remaining data in buffer */ size_t const nbStripes = (state->bufferedSize - 1) / XXH_STRIPE_LEN; size_t nbStripesSoFar = state->nbStripesSoFar; XXH3_consumeStripes(acc, &nbStripesSoFar, state->nbStripesPerBlock, state->buffer, nbStripes, secret, state->secretLimit, XXH3_accumulate, XXH3_scrambleAcc); lastStripePtr = state->buffer + state->bufferedSize - XXH_STRIPE_LEN; } else { /* bufferedSize < XXH_STRIPE_LEN */ /* Copy to temp buffer */ size_t const catchupSize = XXH_STRIPE_LEN - state->bufferedSize; XXH_ASSERT(state->bufferedSize > 0); /* there is always some input buffered */ XXH_memcpy(lastStripe, state->buffer + sizeof(state->buffer) - catchupSize, catchupSize); XXH_memcpy(lastStripe + catchupSize, state->buffer, state->bufferedSize); lastStripePtr = lastStripe; } /* Last stripe */ XXH3_accumulate_512(acc, lastStripePtr, secret + state->secretLimit - XXH_SECRET_LASTACC_START); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_digest (XXH_NOESCAPE const XXH3_state_t* state) { const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; if (state->totalLen > XXH3_MIDSIZE_MAX) { XXH_ALIGN(XXH_ACC_ALIGN) XXH64_hash_t acc[XXH_ACC_NB]; XXH3_digest_long(acc, state, secret); return XXH3_finalizeLong_64b(acc, secret, (xxh_u64)state->totalLen); } /* totalLen <= XXH3_MIDSIZE_MAX: digesting a short input */ if (state->useSeed) return XXH3_64bits_withSeed(state->buffer, (size_t)state->totalLen, state->seed); return XXH3_64bits_withSecret(state->buffer, (size_t)(state->totalLen), secret, state->secretLimit + XXH_STRIPE_LEN); } #endif /* !XXH_NO_STREAM */ /* ========================================== * XXH3 128 bits (a.k.a XXH128) * ========================================== * XXH3's 128-bit variant has better mixing and strength than the 64-bit variant, * even without counting the significantly larger output size. * * For example, extra steps are taken to avoid the seed-dependent collisions * in 17-240 byte inputs (See XXH3_mix16B and XXH128_mix32B). * * This strength naturally comes at the cost of some speed, especially on short * lengths. Note that longer hashes are about as fast as the 64-bit version * due to it using only a slight modification of the 64-bit loop. * * XXH128 is also more oriented towards 64-bit machines. It is still extremely * fast for a _128-bit_ hash on 32-bit (it usually clears XXH64). */ XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t XXH3_len_1to3_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) { /* A doubled version of 1to3_64b with different constants. */ XXH_ASSERT(input != NULL); XXH_ASSERT(1 <= len && len <= 3); XXH_ASSERT(secret != NULL); /* * len = 1: combinedl = { input[0], 0x01, input[0], input[0] } * len = 2: combinedl = { input[1], 0x02, input[0], input[1] } * len = 3: combinedl = { input[2], 0x03, input[0], input[1] } */ { xxh_u8 const c1 = input[0]; xxh_u8 const c2 = input[len >> 1]; xxh_u8 const c3 = input[len - 1]; xxh_u32 const combinedl = ((xxh_u32)c1 <<16) | ((xxh_u32)c2 << 24) | ((xxh_u32)c3 << 0) | ((xxh_u32)len << 8); xxh_u32 const combinedh = XXH_rotl32(XXH_swap32(combinedl), 13); xxh_u64 const bitflipl = (XXH_readLE32(secret) ^ XXH_readLE32(secret+4)) + seed; xxh_u64 const bitfliph = (XXH_readLE32(secret+8) ^ XXH_readLE32(secret+12)) - seed; xxh_u64 const keyed_lo = (xxh_u64)combinedl ^ bitflipl; xxh_u64 const keyed_hi = (xxh_u64)combinedh ^ bitfliph; XXH128_hash_t h128; h128.low64 = XXH64_avalanche(keyed_lo); h128.high64 = XXH64_avalanche(keyed_hi); return h128; } } XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t XXH3_len_4to8_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) { XXH_ASSERT(input != NULL); XXH_ASSERT(secret != NULL); XXH_ASSERT(4 <= len && len <= 8); seed ^= (xxh_u64)XXH_swap32((xxh_u32)seed) << 32; { xxh_u32 const input_lo = XXH_readLE32(input); xxh_u32 const input_hi = XXH_readLE32(input + len - 4); xxh_u64 const input_64 = input_lo + ((xxh_u64)input_hi << 32); xxh_u64 const bitflip = (XXH_readLE64(secret+16) ^ XXH_readLE64(secret+24)) + seed; xxh_u64 const keyed = input_64 ^ bitflip; /* Shift len to the left to ensure it is even, this avoids even multiplies. */ XXH128_hash_t m128 = XXH_mult64to128(keyed, XXH_PRIME64_1 + (len << 2)); m128.high64 += (m128.low64 << 1); m128.low64 ^= (m128.high64 >> 3); m128.low64 = XXH_xorshift64(m128.low64, 35); m128.low64 *= PRIME_MX2; m128.low64 = XXH_xorshift64(m128.low64, 28); m128.high64 = XXH3_avalanche(m128.high64); return m128; } } XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t XXH3_len_9to16_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) { XXH_ASSERT(input != NULL); XXH_ASSERT(secret != NULL); XXH_ASSERT(9 <= len && len <= 16); { xxh_u64 const bitflipl = (XXH_readLE64(secret+32) ^ XXH_readLE64(secret+40)) - seed; xxh_u64 const bitfliph = (XXH_readLE64(secret+48) ^ XXH_readLE64(secret+56)) + seed; xxh_u64 const input_lo = XXH_readLE64(input); xxh_u64 input_hi = XXH_readLE64(input + len - 8); XXH128_hash_t m128 = XXH_mult64to128(input_lo ^ input_hi ^ bitflipl, XXH_PRIME64_1); /* * Put len in the middle of m128 to ensure that the length gets mixed to * both the low and high bits in the 128x64 multiply below. */ m128.low64 += (xxh_u64)(len - 1) << 54; input_hi ^= bitfliph; /* * Add the high 32 bits of input_hi to the high 32 bits of m128, then * add the long product of the low 32 bits of input_hi and XXH_PRIME32_2 to * the high 64 bits of m128. * * The best approach to this operation is different on 32-bit and 64-bit. */ if (sizeof(void *) < sizeof(xxh_u64)) { /* 32-bit */ /* * 32-bit optimized version, which is more readable. * * On 32-bit, it removes an ADC and delays a dependency between the two * halves of m128.high64, but it generates an extra mask on 64-bit. */ m128.high64 += (input_hi & 0xFFFFFFFF00000000ULL) + XXH_mult32to64((xxh_u32)input_hi, XXH_PRIME32_2); } else { /* * 64-bit optimized (albeit more confusing) version. * * Uses some properties of addition and multiplication to remove the mask: * * Let: * a = input_hi.lo = (input_hi & 0x00000000FFFFFFFF) * b = input_hi.hi = (input_hi & 0xFFFFFFFF00000000) * c = XXH_PRIME32_2 * * a + (b * c) * Inverse Property: x + y - x == y * a + (b * (1 + c - 1)) * Distributive Property: x * (y + z) == (x * y) + (x * z) * a + (b * 1) + (b * (c - 1)) * Identity Property: x * 1 == x * a + b + (b * (c - 1)) * * Substitute a, b, and c: * input_hi.hi + input_hi.lo + ((xxh_u64)input_hi.lo * (XXH_PRIME32_2 - 1)) * * Since input_hi.hi + input_hi.lo == input_hi, we get this: * input_hi + ((xxh_u64)input_hi.lo * (XXH_PRIME32_2 - 1)) */ m128.high64 += input_hi + XXH_mult32to64((xxh_u32)input_hi, XXH_PRIME32_2 - 1); } /* m128 ^= XXH_swap64(m128 >> 64); */ m128.low64 ^= XXH_swap64(m128.high64); { /* 128x64 multiply: h128 = m128 * XXH_PRIME64_2; */ XXH128_hash_t h128 = XXH_mult64to128(m128.low64, XXH_PRIME64_2); h128.high64 += m128.high64 * XXH_PRIME64_2; h128.low64 = XXH3_avalanche(h128.low64); h128.high64 = XXH3_avalanche(h128.high64); return h128; } } } /* * Assumption: `secret` size is >= XXH3_SECRET_SIZE_MIN */ XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t XXH3_len_0to16_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) { XXH_ASSERT(len <= 16); { if (len > 8) return XXH3_len_9to16_128b(input, len, secret, seed); if (len >= 4) return XXH3_len_4to8_128b(input, len, secret, seed); if (len) return XXH3_len_1to3_128b(input, len, secret, seed); { XXH128_hash_t h128; xxh_u64 const bitflipl = XXH_readLE64(secret+64) ^ XXH_readLE64(secret+72); xxh_u64 const bitfliph = XXH_readLE64(secret+80) ^ XXH_readLE64(secret+88); h128.low64 = XXH64_avalanche(seed ^ bitflipl); h128.high64 = XXH64_avalanche( seed ^ bitfliph); return h128; } } } /* * A bit slower than XXH3_mix16B, but handles multiply by zero better. */ XXH_FORCE_INLINE XXH128_hash_t XXH128_mix32B(XXH128_hash_t acc, const xxh_u8* input_1, const xxh_u8* input_2, const xxh_u8* secret, XXH64_hash_t seed) { acc.low64 += XXH3_mix16B (input_1, secret+0, seed); acc.low64 ^= XXH_readLE64(input_2) + XXH_readLE64(input_2 + 8); acc.high64 += XXH3_mix16B (input_2, secret+16, seed); acc.high64 ^= XXH_readLE64(input_1) + XXH_readLE64(input_1 + 8); return acc; } XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t XXH3_len_17to128_128b(const xxh_u8* XXH_RESTRICT input, size_t len, const xxh_u8* XXH_RESTRICT secret, size_t secretSize, XXH64_hash_t seed) { XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; XXH_ASSERT(16 < len && len <= 128); { XXH128_hash_t acc; acc.low64 = len * XXH_PRIME64_1; acc.high64 = 0; #if XXH_SIZE_OPT >= 1 { /* Smaller, but slightly slower. */ unsigned int i = (unsigned int)(len - 1) / 32; do { acc = XXH128_mix32B(acc, input+16*i, input+len-16*(i+1), secret+32*i, seed); } while (i-- != 0); } #else if (len > 32) { if (len > 64) { if (len > 96) { acc = XXH128_mix32B(acc, input+48, input+len-64, secret+96, seed); } acc = XXH128_mix32B(acc, input+32, input+len-48, secret+64, seed); } acc = XXH128_mix32B(acc, input+16, input+len-32, secret+32, seed); } acc = XXH128_mix32B(acc, input, input+len-16, secret, seed); #endif { XXH128_hash_t h128; h128.low64 = acc.low64 + acc.high64; h128.high64 = (acc.low64 * XXH_PRIME64_1) + (acc.high64 * XXH_PRIME64_4) + ((len - seed) * XXH_PRIME64_2); h128.low64 = XXH3_avalanche(h128.low64); h128.high64 = (XXH64_hash_t)0 - XXH3_avalanche(h128.high64); return h128; } } } XXH_NO_INLINE XXH_PUREF XXH128_hash_t XXH3_len_129to240_128b(const xxh_u8* XXH_RESTRICT input, size_t len, const xxh_u8* XXH_RESTRICT secret, size_t secretSize, XXH64_hash_t seed) { XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); { XXH128_hash_t acc; unsigned i; acc.low64 = len * XXH_PRIME64_1; acc.high64 = 0; /* * We set as `i` as offset + 32. We do this so that unchanged * `len` can be used as upper bound. This reaches a sweet spot * where both x86 and aarch64 get simple agen and good codegen * for the loop. */ for (i = 32; i < 160; i += 32) { acc = XXH128_mix32B(acc, input + i - 32, input + i - 16, secret + i - 32, seed); } acc.low64 = XXH3_avalanche(acc.low64); acc.high64 = XXH3_avalanche(acc.high64); /* * NB: `i <= len` will duplicate the last 32-bytes if * len % 32 was zero. This is an unfortunate necessity to keep * the hash result stable. */ for (i=160; i <= len; i += 32) { acc = XXH128_mix32B(acc, input + i - 32, input + i - 16, secret + XXH3_MIDSIZE_STARTOFFSET + i - 160, seed); } /* last bytes */ acc = XXH128_mix32B(acc, input + len - 16, input + len - 32, secret + XXH3_SECRET_SIZE_MIN - XXH3_MIDSIZE_LASTOFFSET - 16, (XXH64_hash_t)0 - seed); { XXH128_hash_t h128; h128.low64 = acc.low64 + acc.high64; h128.high64 = (acc.low64 * XXH_PRIME64_1) + (acc.high64 * XXH_PRIME64_4) + ((len - seed) * XXH_PRIME64_2); h128.low64 = XXH3_avalanche(h128.low64); h128.high64 = (XXH64_hash_t)0 - XXH3_avalanche(h128.high64); return h128; } } } static XXH_PUREF XXH128_hash_t XXH3_finalizeLong_128b(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret, size_t secretSize, xxh_u64 len) { XXH128_hash_t h128; h128.low64 = XXH3_finalizeLong_64b(acc, secret, len); h128.high64 = XXH3_mergeAccs(acc, secret + secretSize - XXH_STRIPE_LEN - XXH_SECRET_MERGEACCS_START, ~(len * XXH_PRIME64_2)); return h128; } XXH_FORCE_INLINE XXH128_hash_t XXH3_hashLong_128b_internal(const void* XXH_RESTRICT input, size_t len, const xxh_u8* XXH_RESTRICT secret, size_t secretSize, XXH3_f_accumulate f_acc, XXH3_f_scrambleAcc f_scramble) { XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[XXH_ACC_NB] = XXH3_INIT_ACC; XXH3_hashLong_internal_loop(acc, (const xxh_u8*)input, len, secret, secretSize, f_acc, f_scramble); /* converge into final hash */ XXH_STATIC_ASSERT(sizeof(acc) == 64); XXH_ASSERT(secretSize >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); return XXH3_finalizeLong_128b(acc, secret, secretSize, (xxh_u64)len); } /* * It's important for performance that XXH3_hashLong() is not inlined. */ XXH_NO_INLINE XXH_PUREF XXH128_hash_t XXH3_hashLong_128b_default(const void* XXH_RESTRICT input, size_t len, XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen) { (void)seed64; (void)secret; (void)secretLen; return XXH3_hashLong_128b_internal(input, len, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_accumulate, XXH3_scrambleAcc); } /* * It's important for performance to pass @p secretLen (when it's static) * to the compiler, so that it can properly optimize the vectorized loop. * * When the secret size is unknown, or on GCC 12 where the mix of NO_INLINE and FORCE_INLINE * breaks -Og, this is XXH_NO_INLINE. */ XXH3_WITH_SECRET_INLINE XXH128_hash_t XXH3_hashLong_128b_withSecret(const void* XXH_RESTRICT input, size_t len, XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen) { (void)seed64; return XXH3_hashLong_128b_internal(input, len, (const xxh_u8*)secret, secretLen, XXH3_accumulate, XXH3_scrambleAcc); } XXH_FORCE_INLINE XXH128_hash_t XXH3_hashLong_128b_withSeed_internal(const void* XXH_RESTRICT input, size_t len, XXH64_hash_t seed64, XXH3_f_accumulate f_acc, XXH3_f_scrambleAcc f_scramble, XXH3_f_initCustomSecret f_initSec) { if (seed64 == 0) return XXH3_hashLong_128b_internal(input, len, XXH3_kSecret, sizeof(XXH3_kSecret), f_acc, f_scramble); { XXH_ALIGN(XXH_SEC_ALIGN) xxh_u8 secret[XXH_SECRET_DEFAULT_SIZE]; f_initSec(secret, seed64); return XXH3_hashLong_128b_internal(input, len, (const xxh_u8*)secret, sizeof(secret), f_acc, f_scramble); } } /* * It's important for performance that XXH3_hashLong is not inlined. */ XXH_NO_INLINE XXH128_hash_t XXH3_hashLong_128b_withSeed(const void* input, size_t len, XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen) { (void)secret; (void)secretLen; return XXH3_hashLong_128b_withSeed_internal(input, len, seed64, XXH3_accumulate, XXH3_scrambleAcc, XXH3_initCustomSecret); } typedef XXH128_hash_t (*XXH3_hashLong128_f)(const void* XXH_RESTRICT, size_t, XXH64_hash_t, const void* XXH_RESTRICT, size_t); XXH_FORCE_INLINE XXH128_hash_t XXH3_128bits_internal(const void* input, size_t len, XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen, XXH3_hashLong128_f f_hl128) { XXH_ASSERT(secretLen >= XXH3_SECRET_SIZE_MIN); /* * If an action is to be taken if `secret` conditions are not respected, * it should be done here. * For now, it's a contract pre-condition. * Adding a check and a branch here would cost performance at every hash. */ if (len <= 16) return XXH3_len_0to16_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, seed64); if (len <= 128) return XXH3_len_17to128_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); if (len <= XXH3_MIDSIZE_MAX) return XXH3_len_129to240_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); return f_hl128(input, len, seed64, secret, secretLen); } /* === Public XXH128 API === */ /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH128_hash_t XXH3_128bits(XXH_NOESCAPE const void* input, size_t len) { return XXH3_128bits_internal(input, len, 0, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_128b_default); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_withSecret(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize) { return XXH3_128bits_internal(input, len, 0, (const xxh_u8*)secret, secretSize, XXH3_hashLong_128b_withSecret); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_withSeed(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) { return XXH3_128bits_internal(input, len, seed, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_128b_withSeed); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) { if (len <= XXH3_MIDSIZE_MAX) return XXH3_128bits_internal(input, len, seed, XXH3_kSecret, sizeof(XXH3_kSecret), NULL); return XXH3_hashLong_128b_withSecret(input, len, seed, secret, secretSize); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH128_hash_t XXH128(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) { return XXH3_128bits_withSeed(input, len, seed); } /* === XXH3 128-bit streaming === */ #ifndef XXH_NO_STREAM /* * All initialization and update functions are identical to 64-bit streaming variant. * The only difference is the finalization routine. */ /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr) { return XXH3_64bits_reset(statePtr); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize) { return XXH3_64bits_reset_withSecret(statePtr, secret, secretSize); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed) { return XXH3_64bits_reset_withSeed(statePtr, seed); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) { return XXH3_64bits_reset_withSecretandSeed(statePtr, secret, secretSize, seed); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_128bits_update(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len) { return XXH3_64bits_update(state, input, len); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_digest (XXH_NOESCAPE const XXH3_state_t* state) { const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; if (state->totalLen > XXH3_MIDSIZE_MAX) { XXH_ALIGN(XXH_ACC_ALIGN) XXH64_hash_t acc[XXH_ACC_NB]; XXH3_digest_long(acc, state, secret); XXH_ASSERT(state->secretLimit + XXH_STRIPE_LEN >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); return XXH3_finalizeLong_128b(acc, secret, state->secretLimit + XXH_STRIPE_LEN, (xxh_u64)state->totalLen); } /* len <= XXH3_MIDSIZE_MAX : short code */ if (state->useSeed) return XXH3_128bits_withSeed(state->buffer, (size_t)state->totalLen, state->seed); return XXH3_128bits_withSecret(state->buffer, (size_t)(state->totalLen), secret, state->secretLimit + XXH_STRIPE_LEN); } #endif /* !XXH_NO_STREAM */ /* 128-bit utility functions */ #include /* memcmp, memcpy */ /* return : 1 is equal, 0 if different */ /*! @ingroup XXH3_family */ XXH_PUBLIC_API int XXH128_isEqual(XXH128_hash_t h1, XXH128_hash_t h2) { /* note : XXH128_hash_t is compact, it has no padding byte */ return !(memcmp(&h1, &h2, sizeof(h1))); } /* This prototype is compatible with stdlib's qsort(). * @return : >0 if *h128_1 > *h128_2 * <0 if *h128_1 < *h128_2 * =0 if *h128_1 == *h128_2 */ /*! @ingroup XXH3_family */ XXH_PUBLIC_API int XXH128_cmp(XXH_NOESCAPE const void* h128_1, XXH_NOESCAPE const void* h128_2) { XXH128_hash_t const h1 = *(const XXH128_hash_t*)h128_1; XXH128_hash_t const h2 = *(const XXH128_hash_t*)h128_2; int const hcmp = (h1.high64 > h2.high64) - (h2.high64 > h1.high64); /* note : bets that, in most cases, hash values are different */ if (hcmp) return hcmp; return (h1.low64 > h2.low64) - (h2.low64 > h1.low64); } /*====== Canonical representation ======*/ /*! @ingroup XXH3_family */ XXH_PUBLIC_API void XXH128_canonicalFromHash(XXH_NOESCAPE XXH128_canonical_t* dst, XXH128_hash_t hash) { XXH_STATIC_ASSERT(sizeof(XXH128_canonical_t) == sizeof(XXH128_hash_t)); if (XXH_CPU_LITTLE_ENDIAN) { hash.high64 = XXH_swap64(hash.high64); hash.low64 = XXH_swap64(hash.low64); } XXH_memcpy(dst, &hash.high64, sizeof(hash.high64)); XXH_memcpy((char*)dst + sizeof(hash.high64), &hash.low64, sizeof(hash.low64)); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH128_hash_t XXH128_hashFromCanonical(XXH_NOESCAPE const XXH128_canonical_t* src) { XXH128_hash_t h; h.high64 = XXH_readBE64(src); h.low64 = XXH_readBE64(src->digest + 8); return h; } /* ========================================== * Secret generators * ========================================== */ #define XXH_MIN(x, y) (((x) > (y)) ? (y) : (x)) XXH_FORCE_INLINE void XXH3_combine16(void* dst, XXH128_hash_t h128) { XXH_writeLE64( dst, XXH_readLE64(dst) ^ h128.low64 ); XXH_writeLE64( (char*)dst+8, XXH_readLE64((char*)dst+8) ^ h128.high64 ); } /*! @ingroup XXH3_family */ XXH_PUBLIC_API XXH_errorcode XXH3_generateSecret(XXH_NOESCAPE void* secretBuffer, size_t secretSize, XXH_NOESCAPE const void* customSeed, size_t customSeedSize) { #if (XXH_DEBUGLEVEL >= 1) XXH_ASSERT(secretBuffer != NULL); XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); #else /* production mode, assert() are disabled */ if (secretBuffer == NULL) return XXH_ERROR; if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; #endif if (customSeedSize == 0) { customSeed = XXH3_kSecret; customSeedSize = XXH_SECRET_DEFAULT_SIZE; } #if (XXH_DEBUGLEVEL >= 1) XXH_ASSERT(customSeed != NULL); #else if (customSeed == NULL) return XXH_ERROR; #endif /* Fill secretBuffer with a copy of customSeed - repeat as needed */ { size_t pos = 0; while (pos < secretSize) { size_t const toCopy = XXH_MIN((secretSize - pos), customSeedSize); memcpy((char*)secretBuffer + pos, customSeed, toCopy); pos += toCopy; } } { size_t const nbSeg16 = secretSize / 16; size_t n; XXH128_canonical_t scrambler; XXH128_canonicalFromHash(&scrambler, XXH128(customSeed, customSeedSize, 0)); for (n=0; n. cmake_minimum_required (VERSION 3.10 FATAL_ERROR) set(XXHASH_DIR "${CMAKE_CURRENT_SOURCE_DIR}/..") file(STRINGS "${XXHASH_DIR}/xxhash.h" XXHASH_VERSION_MAJOR REGEX "^#define XXH_VERSION_MAJOR +([0-9]+) *$") string(REGEX REPLACE "^#define XXH_VERSION_MAJOR +([0-9]+) *$" "\\1" XXHASH_VERSION_MAJOR "${XXHASH_VERSION_MAJOR}") file(STRINGS "${XXHASH_DIR}/xxhash.h" XXHASH_VERSION_MINOR REGEX "^#define XXH_VERSION_MINOR +([0-9]+) *$") string(REGEX REPLACE "^#define XXH_VERSION_MINOR +([0-9]+) *$" "\\1" XXHASH_VERSION_MINOR "${XXHASH_VERSION_MINOR}") file(STRINGS "${XXHASH_DIR}/xxhash.h" XXHASH_VERSION_RELEASE REGEX "^#define XXH_VERSION_RELEASE +([0-9]+) *$") string(REGEX REPLACE "^#define XXH_VERSION_RELEASE +([0-9]+) *$" "\\1" XXHASH_VERSION_RELEASE "${XXHASH_VERSION_RELEASE}") set(XXHASH_VERSION_STRING "${XXHASH_VERSION_MAJOR}.${XXHASH_VERSION_MINOR}.${XXHASH_VERSION_RELEASE}") set(XXHASH_LIB_VERSION ${XXHASH_VERSION_STRING}) set(XXHASH_LIB_SOVERSION "${XXHASH_VERSION_MAJOR}") mark_as_advanced(XXHASH_VERSION_MAJOR XXHASH_VERSION_MINOR XXHASH_VERSION_RELEASE XXHASH_VERSION_STRING XXHASH_LIB_VERSION XXHASH_LIB_SOVERSION) if("${CMAKE_VERSION}" VERSION_LESS "3.13") #message(WARNING "CMake ${CMAKE_VERSION} has no CMP0077 policy: options will erase uncached/untyped normal vars!") else() cmake_policy (SET CMP0077 NEW) endif() cmake_policy (SET CMP0048 NEW) project(xxHash VERSION ${XXHASH_VERSION_STRING} LANGUAGES C) if(NOT CMAKE_BUILD_TYPE AND NOT CMAKE_CONFIGURATION_TYPES) set(CMAKE_BUILD_TYPE "Release" CACHE STRING "Project build type" FORCE) set_property(CACHE CMAKE_BUILD_TYPE PROPERTY STRINGS "Debug" "Release" "RelWithDebInfo" "MinSizeRel") endif() if(NOT CMAKE_CONFIGURATION_TYPES) message(STATUS "xxHash build type: ${CMAKE_BUILD_TYPE}") endif() # Enable assert() statements in debug builds if("${CMAKE_BUILD_TYPE}" STREQUAL "Debug") if("${CMAKE_VERSION}" VERSION_LESS "3.12") # add_compile_definitions is not available for older cmake => do nothing else() add_compile_definitions(XXH_DEBUGLEVEL=1) endif() endif() option(BUILD_SHARED_LIBS "Build shared library" ON) option(XXHASH_BUILD_XXHSUM "Build the xxhsum binary" ON) # If XXHASH is being bundled in another project, we don't want to # install anything. However, we want to let people override this, so # we'll use the XXHASH_BUNDLED_MODE variable to let them do that; just # set it to OFF in your project before you add_subdirectory(xxhash/cmake_unofficial). if(NOT DEFINED XXHASH_BUNDLED_MODE) if("${PROJECT_SOURCE_DIR}" STREQUAL "${CMAKE_SOURCE_DIR}") set(XXHASH_BUNDLED_MODE OFF) else() set(XXHASH_BUNDLED_MODE ON) endif() endif() set(XXHASH_BUNDLED_MODE ${XXHASH_BUNDLED_MODE} CACHE BOOL "" FORCE) mark_as_advanced(XXHASH_BUNDLED_MODE) # Allow people to choose whether to build shared or static libraries # via the BUILD_SHARED_LIBS option unless we are in bundled mode, in # which case we always use static libraries. include(CMakeDependentOption) CMAKE_DEPENDENT_OPTION(BUILD_SHARED_LIBS "Build shared libraries" ON "NOT XXHASH_BUNDLED_MODE" OFF) # detect architecture for DISPATCH mode CMAKE_HOST_SYSTEM_INFORMATION(RESULT PLATFORM QUERY OS_PLATFORM) message(STATUS "Architecture: ${PLATFORM}") # libxxhash if((DEFINED DISPATCH) AND (DEFINED PLATFORM)) # Only support DISPATCH option on x86_64. if(("${PLATFORM}" STREQUAL "x86_64") OR ("${PLATFORM}" STREQUAL "AMD64")) set(XXHSUM_DISPATCH ON) message(STATUS "Enable xxHash dispatch mode") add_library(xxhash "${XXHASH_DIR}/xxh_x86dispatch.c" "${XXHASH_DIR}/xxhash.c" ) set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} -DXXHSUM_DISPATCH=1") else() add_library(xxhash "${XXHASH_DIR}/xxhash.c") endif() else() add_library(xxhash "${XXHASH_DIR}/xxhash.c") endif() add_library(${PROJECT_NAME}::xxhash ALIAS xxhash) target_include_directories(xxhash PUBLIC $ $) if (BUILD_SHARED_LIBS) target_compile_definitions(xxhash PUBLIC XXH_EXPORT) endif () set_target_properties(xxhash PROPERTIES SOVERSION "${XXHASH_LIB_SOVERSION}" VERSION "${XXHASH_VERSION_STRING}") if(XXHASH_BUILD_XXHSUM) set(XXHSUM_DIR "${XXHASH_DIR}/cli") # xxhsum set(XXHSUM_SOURCES) if (XXHSUM_DISPATCH) list(APPEND XXHSUM_SOURCES "${XXHASH_DIR}/xxh_x86dispatch.c") endif() list(APPEND XXHSUM_SOURCES "${XXHSUM_DIR}/xxhsum.c" "${XXHSUM_DIR}/xsum_arch.c" "${XXHSUM_DIR}/xsum_os_specific.c" "${XXHSUM_DIR}/xsum_output.c" "${XXHSUM_DIR}/xsum_sanity_check.c" "${XXHSUM_DIR}/xsum_bench.c" ) add_executable(xxhsum ${XXHSUM_SOURCES}) add_executable(${PROJECT_NAME}::xxhsum ALIAS xxhsum) target_link_libraries(xxhsum PRIVATE xxhash) target_include_directories(xxhsum PRIVATE "${XXHASH_DIR}") endif(XXHASH_BUILD_XXHSUM) # Extra warning flags include (CheckCCompilerFlag) if (XXHASH_C_FLAGS) set(CMAKE_C_FLAGS "${CMAKE_C_FLAGS} ${XXHASH_C_FLAGS}") endif() if(NOT XXHASH_BUNDLED_MODE) include(GNUInstallDirs) install(TARGETS xxhash EXPORT xxHashTargets RUNTIME DESTINATION "${CMAKE_INSTALL_BINDIR}" LIBRARY DESTINATION "${CMAKE_INSTALL_LIBDIR}" ARCHIVE DESTINATION "${CMAKE_INSTALL_LIBDIR}") install(FILES "${XXHASH_DIR}/xxhash.h" DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}") install(FILES "${XXHASH_DIR}/xxh3.h" DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}") if(DISPATCH) install(FILES "${XXHASH_DIR}/xxh_x86dispatch.h" DESTINATION "${CMAKE_INSTALL_INCLUDEDIR}") endif() if(XXHASH_BUILD_XXHSUM) install(TARGETS xxhsum EXPORT xxHashTargets RUNTIME DESTINATION "${CMAKE_INSTALL_BINDIR}") install(FILES "${XXHSUM_DIR}/xxhsum.1" DESTINATION "${CMAKE_INSTALL_MANDIR}/man1") endif(XXHASH_BUILD_XXHSUM) include(CMakePackageConfigHelpers) set(xxHash_VERSION_CONFIG "${PROJECT_BINARY_DIR}/xxHashConfigVersion.cmake") set(xxHash_PROJECT_CONFIG "${PROJECT_BINARY_DIR}/xxHashConfig.cmake") set(xxHash_TARGETS_CONFIG "${PROJECT_BINARY_DIR}/xxHashTargets.cmake") set(xxHash_CONFIG_INSTALL_DIR "${CMAKE_INSTALL_LIBDIR}/cmake/xxHash/") write_basic_package_version_file(${xxHash_VERSION_CONFIG} VERSION ${XXHASH_VERSION_STRING} COMPATIBILITY AnyNewerVersion) configure_package_config_file( ${PROJECT_SOURCE_DIR}/xxHashConfig.cmake.in ${xxHash_PROJECT_CONFIG} INSTALL_DESTINATION ${xxHash_CONFIG_INSTALL_DIR}) export(EXPORT xxHashTargets FILE ${xxHash_TARGETS_CONFIG} NAMESPACE ${PROJECT_NAME}::) install(FILES ${xxHash_PROJECT_CONFIG} ${xxHash_VERSION_CONFIG} DESTINATION ${xxHash_CONFIG_INSTALL_DIR}) install(EXPORT xxHashTargets DESTINATION ${xxHash_CONFIG_INSTALL_DIR} NAMESPACE ${PROJECT_NAME}::) # configure and install pkg-config include(JoinPaths.cmake) set(PREFIX ${CMAKE_INSTALL_PREFIX}) set(EXECPREFIX "\${prefix}") join_paths(INCLUDEDIR "\${prefix}" "${CMAKE_INSTALL_INCLUDEDIR}") join_paths(LIBDIR "\${prefix}" "${CMAKE_INSTALL_LIBDIR}") set(VERSION "${XXHASH_VERSION_STRING}") configure_file(${XXHASH_DIR}/libxxhash.pc.in ${CMAKE_BINARY_DIR}/libxxhash.pc @ONLY) install(FILES ${CMAKE_BINARY_DIR}/libxxhash.pc DESTINATION ${CMAKE_INSTALL_LIBDIR}/pkgconfig) endif(NOT XXHASH_BUNDLED_MODE) include(CPack) Crypt-xxHash-0.09/ext/xxHash/cmake_unofficial/JoinPaths.cmake000444001750001750 146115146654120 26015 0ustar00dchernenkodchernenko000000000000# This module provides function for joining paths # known from most languages # # SPDX-License-Identifier: (MIT OR CC0-1.0) # Copyright 2020 Jan Tojnar # https://github.com/jtojnar/cmake-snips # # Modelled after Python’s os.path.join # https://docs.python.org/3.7/library/os.path.html#os.path.join # Windows not supported function(join_paths joined_path first_path_segment) set(temp_path "${first_path_segment}") foreach(current_segment IN LISTS ARGN) if(NOT ("${current_segment}" STREQUAL "")) if(IS_ABSOLUTE "${current_segment}") set(temp_path "${current_segment}") else() set(temp_path "${temp_path}/${current_segment}") endif() endif() endforeach() set(${joined_path} "${temp_path}" PARENT_SCOPE) endfunction() Crypt-xxHash-0.09/ext/xxHash/cmake_unofficial/README.md000444001750001750 226715146654120 24400 0ustar00dchernenkodchernenko000000000000 ## Usage ### Way 1: import targets Build xxHash targets: cd mkdir build cd build cmake ../cmake_unofficial [options] cmake --build . cmake --build . --target install #optional Where possible options are: - `-DXXHASH_BUILD_ENABLE_INLINE_API=`: adds xxhash.c for the `-DXXH_INLINE_ALL` api. ON by default. - `-DXXHASH_BUILD_XXHSUM=`: build the command line binary. ON by default - `-DBUILD_SHARED_LIBS=`: build dynamic library. ON by default. - `-DCMAKE_INSTALL_PREFIX=`: use custom install prefix path. - `-DDISPATCH=`: enable dispatch mode. OFF by default. Add lines into downstream CMakeLists.txt: find_package(xxHash 0.7 CONFIG REQUIRED) ... target_link_libraries(MyTarget PRIVATE xxHash::xxhash) ### Way 2: Add subdirectory Add lines into downstream CMakeLists.txt: option(BUILD_SHARED_LIBS "Build shared libs" OFF) #optional ... set(XXHASH_BUILD_ENABLE_INLINE_API OFF) #optional set(XXHASH_BUILD_XXHSUM OFF) #optional add_subdirectory( EXCLUDE_FROM_ALL) ... target_link_libraries(MyTarget PRIVATE xxHash::xxhash) Crypt-xxHash-0.09/ext/xxHash/cmake_unofficial/xxHashConfig.cmake.in000444001750001750 11015146654120 27062 0ustar00dchernenkodchernenko000000000000@PACKAGE_INIT@ include(${CMAKE_CURRENT_LIST_DIR}/xxHashTargets.cmake) Crypt-xxHash-0.09/ext/xxHash/doc000755001750001750 015146654120 20243 5ustar00dchernenkodchernenko000000000000Crypt-xxHash-0.09/ext/xxHash/doc/README.md000444001750001750 63515146654120 21643 0ustar00dchernenkodchernenko000000000000xxHash Specification ======================= This directory contains material defining the xxHash algorithm. It's described in [this specification document](xxhash_spec.md). The algorithm is also be illustrated by a [simple educational library](https://github.com/easyaspi314/xxhash-clean), written by @easyaspi314 and designed for readability (as opposed to the reference library which is designed for speed). Crypt-xxHash-0.09/ext/xxHash/doc/xxhash.cry000444001750001750 1602615146654120 22447 0ustar00dchernenkodchernenko000000000000module xxhash where /** * The 32-bit variant of xxHash. The first argument is the sequence * of L bytes to hash. The second argument is a seed value. */ XXH32 : {L} (fin L) => [L][8] -> [32] -> [32] XXH32 input seed = XXH32_avalanche acc1 where (stripes16 # stripes4 # stripes1) = input accR = foldl XXH32_rounds (XXH32_init seed) (split stripes16 : [L/16][16][8]) accL = `(L % 2^^32) + if (`L:Integer) < 16 then seed + PRIME32_5 else XXH32_converge accR acc4 = foldl XXH32_digest4 accL (split stripes4 : [(L%16)/4][4][8]) acc1 = foldl XXH32_digest1 acc4 (stripes1 : [L%4][8]) /** * The 64-bit variant of xxHash. The first argument is the sequence * of L bytes to hash. The second argument is a seed value. */ XXH64 : {L} (fin L) => [L][8] -> [64] -> [64] XXH64 input seed = XXH64_avalanche acc1 where (stripes32 # stripes8 # stripes4 # stripes1) = input accR = foldl XXH64_rounds (XXH64_init seed) (split stripes32 : [L/32][32][8]) accL = `(L % 2^^64) + if (`L:Integer) < 32 then seed + PRIME64_5 else XXH64_converge accR acc8 = foldl XXH64_digest8 accL (split stripes8 : [(L%32)/8][8][8]) acc4 = foldl XXH64_digest4 acc8 (split stripes4 : [(L%8)/4][4][8]) acc1 = foldl XXH64_digest1 acc4 (stripes1 : [L%4][8]) private //Utility functions /** * Combines a sequence of bytes into a word using the little-endian * convention. */ toLE bytes = join (reverse bytes) //32-bit xxHash helper functions //32-bit prime number constants PRIME32_1 = 0x9E3779B1 : [32] PRIME32_2 = 0x85EBCA77 : [32] PRIME32_3 = 0xC2B2AE3D : [32] PRIME32_4 = 0x27D4EB2F : [32] PRIME32_5 = 0x165667B1 : [32] /** * The property shows that the hexadecimal representation of the * PRIME32 constants is the same as the binary representation. */ property PRIME32s_as_bits_correct = (PRIME32_1 == 0b10011110001101110111100110110001) /\ (PRIME32_2 == 0b10000101111010111100101001110111) /\ (PRIME32_3 == 0b11000010101100101010111000111101) /\ (PRIME32_4 == 0b00100111110101001110101100101111) /\ (PRIME32_5 == 0b00010110010101100110011110110001) /** * This function initializes the four internal accumulators of XXH32. */ XXH32_init : [32] -> [4][32] XXH32_init seed = [acc1, acc2, acc3, acc4] where acc1 = seed + PRIME32_1 + PRIME32_2 acc2 = seed + PRIME32_2 acc3 = seed + 0 acc4 = seed - PRIME32_1 /** * This processes a single lane of the main round function of XXH32. */ XXH32_round : [32] -> [32] -> [32] XXH32_round accN laneN = ((accN + laneN * PRIME32_2) <<< 13) * PRIME32_1 /** * This is the main round function of XXH32 and processes a stripe, * i.e. 4 lanes with 4 bytes each. */ XXH32_rounds : [4][32] -> [16][8] -> [4][32] XXH32_rounds accs stripe = [ XXH32_round accN (toLE laneN) | accN <- accs | laneN <- split stripe ] /** * This function combines the four lane accumulators into a single * 32-bit value. */ XXH32_converge : [4][32] -> [32] XXH32_converge [acc1, acc2, acc3, acc4] = (acc1 <<< 1) + (acc2 <<< 7) + (acc3 <<< 12) + (acc4 <<< 18) /** * This function digests a four byte lane */ XXH32_digest4 : [32] -> [4][8] -> [32] XXH32_digest4 acc lane = ((acc + toLE lane * PRIME32_3) <<< 17) * PRIME32_4 /** * This function digests a single byte lane */ XXH32_digest1 : [32] -> [8] -> [32] XXH32_digest1 acc lane = ((acc + (0 # lane) * PRIME32_5) <<< 11) * PRIME32_1 /** * This function ensures that all input bits have a chance to impact * any bit in the output digest, resulting in an unbiased * distribution. */ XXH32_avalanche : [32] -> [32] XXH32_avalanche acc0 = acc5 where acc1 = acc0 ^ (acc0 >> 15) acc2 = acc1 * PRIME32_2 acc3 = acc2 ^ (acc2 >> 13) acc4 = acc3 * PRIME32_3 acc5 = acc4 ^ (acc4 >> 16) //64-bit xxHash helper functions //64-bit prime number constants PRIME64_1 = 0x9E3779B185EBCA87 : [64] PRIME64_2 = 0xC2B2AE3D27D4EB4F : [64] PRIME64_3 = 0x165667B19E3779F9 : [64] PRIME64_4 = 0x85EBCA77C2B2AE63 : [64] PRIME64_5 = 0x27D4EB2F165667C5 : [64] /** * The property shows that the hexadecimal representation of the * PRIME64 constants is the same as the binary representation. */ property PRIME64s_as_bits_correct = (PRIME64_1 == 0b1001111000110111011110011011000110000101111010111100101010000111) /\ (PRIME64_2 == 0b1100001010110010101011100011110100100111110101001110101101001111) /\ (PRIME64_3 == 0b0001011001010110011001111011000110011110001101110111100111111001) /\ (PRIME64_4 == 0b1000010111101011110010100111011111000010101100101010111001100011) /\ (PRIME64_5 == 0b0010011111010100111010110010111100010110010101100110011111000101) /** * This function initializes the four internal accumulators of XXH64. */ XXH64_init : [64] -> [4][64] XXH64_init seed = [acc1, acc2, acc3, acc4] where acc1 = seed + PRIME64_1 + PRIME64_2 acc2 = seed + PRIME64_2 acc3 = seed + 0 acc4 = seed - PRIME64_1 /** * This processes a single lane of the main round function of XXH64. */ XXH64_round : [64] -> [64] -> [64] XXH64_round accN laneN = ((accN + laneN * PRIME64_2) <<< 31) * PRIME64_1 /** * This is the main round function of XXH64 and processes a stripe, * i.e. 4 lanes with 8 bytes each. */ XXH64_rounds : [4][64] -> [32][8] -> [4][64] XXH64_rounds accs stripe = [ XXH64_round accN (toLE laneN) | accN <- accs | laneN <- split stripe ] /** * This is a helper function, used to merge the four lane accumulators. */ mergeAccumulator : [64] -> [64] -> [64] mergeAccumulator acc accN = (acc ^ XXH64_round 0 accN) * PRIME64_1 + PRIME64_4 /** * This function combines the four lane accumulators into a single * 64-bit value. */ XXH64_converge : [4][64] -> [64] XXH64_converge [acc1, acc2, acc3, acc4] = foldl mergeAccumulator ((acc1 <<< 1) + (acc2 <<< 7) + (acc3 <<< 12) + (acc4 <<< 18)) [acc1, acc2, acc3, acc4] /** * This function digests an eight byte lane */ XXH64_digest8 : [64] -> [8][8] -> [64] XXH64_digest8 acc lane = ((acc ^ XXH64_round 0 (toLE lane)) <<< 27) * PRIME64_1 + PRIME64_4 /** * This function digests a four byte lane */ XXH64_digest4 : [64] -> [4][8] -> [64] XXH64_digest4 acc lane = ((acc ^ (0 # toLE lane) * PRIME64_1) <<< 23) * PRIME64_2 + PRIME64_3 /** * This function digests a single byte lane */ XXH64_digest1 : [64] -> [8] -> [64] XXH64_digest1 acc lane = ((acc ^ (0 # lane) * PRIME64_5) <<< 11) * PRIME64_1 /** * This function ensures that all input bits have a chance to impact * any bit in the output digest, resulting in an unbiased * distribution. */ XXH64_avalanche : [64] -> [64] XXH64_avalanche acc0 = acc5 where acc1 = acc0 ^ (acc0 >> 33) acc2 = acc1 * PRIME64_2 acc3 = acc2 ^ (acc2 >> 29) acc4 = acc3 * PRIME64_3 acc5 = acc4 ^ (acc4 >> 32) Crypt-xxHash-0.09/ext/xxHash/doc/xxhash_spec.md000444001750001750 10465015146654120 23305 0ustar00dchernenkodchernenko000000000000xxHash fast digest algorithm ====================== ### Notices Copyright (c) Yann Collet Permission is granted to copy and distribute this document for any purpose and without charge, including translations into other languages and incorporation into compilations, provided that the copyright notice and this notice are preserved, and that any substantive changes or deletions from the original are clearly marked. Distribution of this document is unlimited. ### Version 0.2.0 (29/06/23) Table of Contents --------------------- - [Introduction](#introduction) - [XXH32 algorithm description](#xxh32-algorithm-description) - [XXH64 algorithm description](#xxh64-algorithm-description) - [XXH3 algorithm description](#xxh3-algorithm-overview) - [Small inputs](#xxh3-algorithm-description-for-small-inputs) - [Medium inputs](#xxh3-algorithm-description-for-medium-inputs) - [Large inputs](#xxh3-algorithm-description-for-large-inputs) - [Performance considerations](#performance-considerations) - [Reference Implementation](#reference-implementation) Introduction ---------------- This document describes the xxHash digest algorithm for both 32-bit and 64-bit variants, named `XXH32` and `XXH64`. The algorithm takes an input a message of arbitrary length and an optional seed value, then produces an output of 32 or 64-bit as "fingerprint" or "digest". xxHash is primarily designed for speed. It is labeled non-cryptographic, and is not meant to avoid intentional collisions (same digest for 2 different messages), or to prevent producing a message with a predefined digest. XXH32 is designed to be fast on 32-bit machines. XXH64 is designed to be fast on 64-bit machines. Both variants produce different output. However, a given variant shall produce exactly the same output, irrespective of the cpu / os used. In particular, the result remains identical whatever the endianness and width of the cpu is. ### Operation notations All operations are performed modulo {32,64} bits. Arithmetic overflows are expected. `XXH32` uses 32-bit modular operations. `XXH64` and `XXH3` use 64-bit modular operations. When an operation ingests input or secret as multi-bytes values, it reads it using little-endian convention. - `+`: denotes modular addition - `-`: denotes modular subtraction - `*`: denotes modular multiplication - **Exception:** In `XXH3`, if it is in the form `(u128)x * (u128)y`, it denotes 64-bit by 64-bit normal multiplication into a full 128-bit result. - `X <<< s`: denotes the value obtained by circularly shifting (rotating) `X` left by `s` bit positions. - `X >> s`: denotes the value obtained by shifting `X` right by s bit positions. Upper `s` bits become `0`. - `X << s`: denotes the value obtained by shifting `X` left by s bit positions. Lower `s` bits become `0`. - `X xor Y`: denotes the bit-wise XOR of `X` and `Y` (same width). - `X | Y`: denotes the bit-wise OR of `X` and `Y` (same width). - `~X`: denotes the bit-wise negation of `X`. XXH32 Algorithm Description ------------------------------------- ### Overview We begin by supposing that we have a message of any length `L` as input, and that we wish to find its digest. Here `L` is an arbitrary nonnegative integer; `L` may be zero. The following steps are performed to compute the digest of the message. The algorithm collect and transform input in _stripes_ of 16 bytes. The transforms are stored inside 4 "accumulators", each one storing an unsigned 32-bit value. Each accumulator can be processed independently in parallel, speeding up processing for cpu with multiple execution units. The algorithm uses 32-bits addition, multiplication, rotate, shift and xor operations. Many operations require some 32-bits prime number constants, all defined below: ```c static const u32 PRIME32_1 = 0x9E3779B1U; // 0b10011110001101110111100110110001 static const u32 PRIME32_2 = 0x85EBCA77U; // 0b10000101111010111100101001110111 static const u32 PRIME32_3 = 0xC2B2AE3DU; // 0b11000010101100101010111000111101 static const u32 PRIME32_4 = 0x27D4EB2FU; // 0b00100111110101001110101100101111 static const u32 PRIME32_5 = 0x165667B1U; // 0b00010110010101100110011110110001 ``` These constants are prime numbers, and feature a good mix of bits 1 and 0, neither too regular, nor too dissymmetric. These properties help dispersion capabilities. ### Step 1. Initialize internal accumulators Each accumulator gets an initial value based on optional `seed` input. Since the `seed` is optional, it can be `0`. ```c u32 acc1 = seed + PRIME32_1 + PRIME32_2; u32 acc2 = seed + PRIME32_2; u32 acc3 = seed + 0; u32 acc4 = seed - PRIME32_1; ``` #### Special case: input is less than 16 bytes When the input is too small (< 16 bytes), the algorithm will not process any stripes. Consequently, it will not make use of parallel accumulators. In this case, a simplified initialization is performed, using a single accumulator: ```c u32 acc = seed + PRIME32_5; ``` The algorithm then proceeds directly to step 4. ### Step 2. Process stripes A stripe is a contiguous segment of 16 bytes. It is evenly divided into 4 _lanes_, of 4 bytes each. The first lane is used to update accumulator 1, the second lane is used to update accumulator 2, and so on. Each lane read its associated 32-bit value using __little-endian__ convention. For each {lane, accumulator}, the update process is called a _round_, and applies the following formula: ```c accN = accN + (laneN * PRIME32_2); accN = accN <<< 13; accN = accN * PRIME32_1; ``` This shuffles the bits so that any bit from input _lane_ impacts several bits in output _accumulator_. All operations are performed modulo 2^32. Input is consumed one full stripe at a time. Step 2 is looped as many times as necessary to consume the whole input, except for the last remaining bytes which cannot form a stripe (< 16 bytes). When that happens, move to step 3. ### Step 3. Accumulator convergence All 4 lane accumulators from the previous steps are merged to produce a single remaining accumulator of the same width (32-bit). The associated formula is as follows: ```c acc = (acc1 <<< 1) + (acc2 <<< 7) + (acc3 <<< 12) + (acc4 <<< 18); ``` ### Step 4. Add input length The input total length is presumed known at this stage. This step is just about adding the length to accumulator, so that it participates to final mixing. ```c acc = acc + (u32)inputLength; ``` Note that, if input length is so large that it requires more than 32-bits, only the lower 32-bits are added to the accumulator. ### Step 5. Consume remaining input There may be up to 15 bytes remaining to consume from the input. The final stage will digest them according to following pseudo-code: ```c while (remainingLength >= 4) { lane = read_32bit_little_endian(input_ptr); acc = acc + lane * PRIME32_3; acc = (acc <<< 17) * PRIME32_4; input_ptr += 4; remainingLength -= 4; } while (remainingLength >= 1) { lane = read_byte(input_ptr); acc = acc + lane * PRIME32_5; acc = (acc <<< 11) * PRIME32_1; input_ptr += 1; remainingLength -= 1; } ``` This process ensures that all input bytes are present in the final mix. ### Step 6. Final mix (avalanche) The final mix ensures that all input bits have a chance to impact any bit in the output digest, resulting in an unbiased distribution. This is also called avalanche effect. ```c acc = acc xor (acc >> 15); acc = acc * PRIME32_2; acc = acc xor (acc >> 13); acc = acc * PRIME32_3; acc = acc xor (acc >> 16); ``` ### Step 7. Output The `XXH32()` function produces an unsigned 32-bit value as output. For systems which require to store and/or display the result in binary or hexadecimal format, the canonical format is defined to reproduce the same value as the natural decimal format, hence follows __big-endian__ convention (most significant byte first). XXH64 Algorithm Description ------------------------------------- ### Overview `XXH64`'s algorithm structure is very similar to `XXH32` one. The major difference is that `XXH64` uses 64-bit arithmetic, speeding up memory transfer for 64-bit compliant systems, but also relying on cpu capability to efficiently perform 64-bit operations. The algorithm collects and transforms input in _stripes_ of 32 bytes. The transforms are stored inside 4 "accumulators", each one storing an unsigned 64-bit value. Each accumulator can be processed independently in parallel, speeding up processing for cpu with multiple execution units. The algorithm uses 64-bit addition, multiplication, rotate, shift and xor operations. Many operations require some 64-bit prime number constants, all defined below: ```c static const u64 PRIME64_1 = 0x9E3779B185EBCA87ULL; // 0b1001111000110111011110011011000110000101111010111100101010000111 static const u64 PRIME64_2 = 0xC2B2AE3D27D4EB4FULL; // 0b1100001010110010101011100011110100100111110101001110101101001111 static const u64 PRIME64_3 = 0x165667B19E3779F9ULL; // 0b0001011001010110011001111011000110011110001101110111100111111001 static const u64 PRIME64_4 = 0x85EBCA77C2B2AE63ULL; // 0b1000010111101011110010100111011111000010101100101010111001100011 static const u64 PRIME64_5 = 0x27D4EB2F165667C5ULL; // 0b0010011111010100111010110010111100010110010101100110011111000101 ``` These constants are prime numbers, and feature a good mix of bits 1 and 0, neither too regular, nor too dissymmetric. These properties help dispersion capabilities. ### Step 1. Initialize internal accumulators Each accumulator gets an initial value based on optional `seed` input. Since the `seed` is optional, it can be `0`. ```c u64 acc1 = seed + PRIME64_1 + PRIME64_2; u64 acc2 = seed + PRIME64_2; u64 acc3 = seed + 0; u64 acc4 = seed - PRIME64_1; ``` #### Special case: input is less than 32 bytes When the input is too small (< 32 bytes), the algorithm will not process any stripes. Consequently, it will not make use of parallel accumulators. In this case, a simplified initialization is performed, using a single accumulator: ```c u64 acc = seed + PRIME64_5; ``` The algorithm then proceeds directly to step 4. ### Step 2. Process stripes A stripe is a contiguous segment of 32 bytes. It is evenly divided into 4 _lanes_, of 8 bytes each. The first lane is used to update accumulator 1, the second lane is used to update accumulator 2, and so on. Each lane read its associated 64-bit value using __little-endian__ convention. For each {lane, accumulator}, the update process is called a _round_, and applies the following formula: ```c round(accN,laneN): accN = accN + (laneN * PRIME64_2); accN = accN <<< 31; return accN * PRIME64_1; ``` This shuffles the bits so that any bit from input _lane_ impacts several bits in output _accumulator_. All operations are performed modulo 2^64. Input is consumed one full stripe at a time. Step 2 is looped as many times as necessary to consume the whole input, except for the last remaining bytes which cannot form a stripe (< 32 bytes). When that happens, move to step 3. ### Step 3. Accumulator convergence All 4 lane accumulators from previous steps are merged to produce a single remaining accumulator of same width (64-bit). The associated formula is as follows. Note that accumulator convergence is more complex than 32-bit variant, and requires to define another function called _mergeAccumulator()_: ```c mergeAccumulator(acc,accN): acc = acc xor round(0, accN); acc = acc * PRIME64_1; return acc + PRIME64_4; ``` which is then used in the convergence formula: ```c acc = (acc1 <<< 1) + (acc2 <<< 7) + (acc3 <<< 12) + (acc4 <<< 18); acc = mergeAccumulator(acc, acc1); acc = mergeAccumulator(acc, acc2); acc = mergeAccumulator(acc, acc3); acc = mergeAccumulator(acc, acc4); ``` ### Step 4. Add input length The input total length is presumed known at this stage. This step is just about adding the length to accumulator, so that it participates to final mixing. ```c acc = acc + inputLength; ``` ### Step 5. Consume remaining input There may be up to 31 bytes remaining to consume from the input. The final stage will digest them according to following pseudo-code: ```c while (remainingLength >= 8) { lane = read_64bit_little_endian(input_ptr); acc = acc xor round(0, lane); acc = (acc <<< 27) * PRIME64_1; acc = acc + PRIME64_4; input_ptr += 8; remainingLength -= 8; } if (remainingLength >= 4) { lane = read_32bit_little_endian(input_ptr); acc = acc xor (lane * PRIME64_1); acc = (acc <<< 23) * PRIME64_2; acc = acc + PRIME64_3; input_ptr += 4; remainingLength -= 4; } while (remainingLength >= 1) { lane = read_byte(input_ptr); acc = acc xor (lane * PRIME64_5); acc = (acc <<< 11) * PRIME64_1; input_ptr += 1; remainingLength -= 1; } ``` This process ensures that all input bytes are present in the final mix. ### Step 6. Final mix (avalanche) The final mix ensures that all input bits have a chance to impact any bit in the output digest, resulting in an unbiased distribution. This is also called avalanche effect. ```c acc = acc xor (acc >> 33); acc = acc * PRIME64_2; acc = acc xor (acc >> 29); acc = acc * PRIME64_3; acc = acc xor (acc >> 32); ``` ### Step 7. Output The `XXH64()` function produces an unsigned 64-bit value as output. For systems which require to store and/or display the result in binary or hexadecimal format, the canonical format is defined to reproduce the same value as the natural decimal format, hence follows __big-endian__ convention (most significant byte first). XXH3 Algorithm Overview ------------------------------------- XXH3 comes in two different versions: XXH3-64 and XXH3-128 (or XXH128), producing 64 and 128 bits of output, respectively. XXH3 uses different algorithms for small (0-16 bytes), medium (17-240 bytes), and large (241+ bytes) inputs. The algorithms for small and medium inputs are optimized for performance. The three algorithms are described in the following sections. Many operations require some 64-bit prime number constants, which are mostly the same constants used in XXH32 and XXH64, all defined below: ```c static const u64 PRIME32_1 = 0x9E3779B1U; // 0b10011110001101110111100110110001 static const u64 PRIME32_2 = 0x85EBCA77U; // 0b10000101111010111100101001110111 static const u64 PRIME32_3 = 0xC2B2AE3DU; // 0b11000010101100101010111000111101 static const u64 PRIME64_1 = 0x9E3779B185EBCA87ULL; // 0b1001111000110111011110011011000110000101111010111100101010000111 static const u64 PRIME64_2 = 0xC2B2AE3D27D4EB4FULL; // 0b1100001010110010101011100011110100100111110101001110101101001111 static const u64 PRIME64_3 = 0x165667B19E3779F9ULL; // 0b0001011001010110011001111011000110011110001101110111100111111001 static const u64 PRIME64_4 = 0x85EBCA77C2B2AE63ULL; // 0b1000010111101011110010100111011111000010101100101010111001100011 static const u64 PRIME64_5 = 0x27D4EB2F165667C5ULL; // 0b0010011111010100111010110010111100010110010101100110011111000101 static const u64 PRIME_MX1 = 0x165667919E3779F9ULL; // 0b0001011001010110011001111001000110011110001101110111100111111001 static const u64 PRIME_MX2 = 0x9FB21C651E98DF25ULL; // 0b1001111110110010000111000110010100011110100110001101111100100101 ``` The `XXH3_64bits()` function produces an unsigned 64-bit value. The `XXH3_128bits()` function produces a `XXH128_hash_t` struct containing `low64` and `high64` - the lower and higher 64-bit half values of the result, respectively. For systems requiring storing and/or displaying the result in binary or hexadecimal format, the canonical format is defined to reproduce the same value as the natural decimal format, hence following **big-endian** convention (most significant byte first). ### Seed and Secret XXH3 provides seeded hashing by introducing two configurable constants used in the hashing process: the seed and the secret. The seed is an unsigned 64-bit value, and the secret is an array of bytes that is at least 136 bytes in size. The default seed is 0, and the default secret is the following 192-byte value: ```c static const u8 defaultSecret[192] = { 0xb8, 0xfe, 0x6c, 0x39, 0x23, 0xa4, 0x4b, 0xbe, 0x7c, 0x01, 0x81, 0x2c, 0xf7, 0x21, 0xad, 0x1c, 0xde, 0xd4, 0x6d, 0xe9, 0x83, 0x90, 0x97, 0xdb, 0x72, 0x40, 0xa4, 0xa4, 0xb7, 0xb3, 0x67, 0x1f, 0xcb, 0x79, 0xe6, 0x4e, 0xcc, 0xc0, 0xe5, 0x78, 0x82, 0x5a, 0xd0, 0x7d, 0xcc, 0xff, 0x72, 0x21, 0xb8, 0x08, 0x46, 0x74, 0xf7, 0x43, 0x24, 0x8e, 0xe0, 0x35, 0x90, 0xe6, 0x81, 0x3a, 0x26, 0x4c, 0x3c, 0x28, 0x52, 0xbb, 0x91, 0xc3, 0x00, 0xcb, 0x88, 0xd0, 0x65, 0x8b, 0x1b, 0x53, 0x2e, 0xa3, 0x71, 0x64, 0x48, 0x97, 0xa2, 0x0d, 0xf9, 0x4e, 0x38, 0x19, 0xef, 0x46, 0xa9, 0xde, 0xac, 0xd8, 0xa8, 0xfa, 0x76, 0x3f, 0xe3, 0x9c, 0x34, 0x3f, 0xf9, 0xdc, 0xbb, 0xc7, 0xc7, 0x0b, 0x4f, 0x1d, 0x8a, 0x51, 0xe0, 0x4b, 0xcd, 0xb4, 0x59, 0x31, 0xc8, 0x9f, 0x7e, 0xc9, 0xd9, 0x78, 0x73, 0x64, 0xea, 0xc5, 0xac, 0x83, 0x34, 0xd3, 0xeb, 0xc3, 0xc5, 0x81, 0xa0, 0xff, 0xfa, 0x13, 0x63, 0xeb, 0x17, 0x0d, 0xdd, 0x51, 0xb7, 0xf0, 0xda, 0x49, 0xd3, 0x16, 0x55, 0x26, 0x29, 0xd4, 0x68, 0x9e, 0x2b, 0x16, 0xbe, 0x58, 0x7d, 0x47, 0xa1, 0xfc, 0x8f, 0xf8, 0xb8, 0xd1, 0x7a, 0xd0, 0x31, 0xce, 0x45, 0xcb, 0x3a, 0x8f, 0x95, 0x16, 0x04, 0x28, 0xaf, 0xd7, 0xfb, 0xca, 0xbb, 0x4b, 0x40, 0x7e, }; ``` The seed and the secret can be optionally specified using the `*_withSecret` and `*_withSeed` versions of the hash function. The seed and the secret cannot be specified simultaneously (`*_withSecretAndSeed` is actually `*_withSeed` for short and medium inputs <= 240 bytes, and `*_withSecret` for large inputs). When one is specified, the other one uses the default value. There is one exception though: when input is large (> 240 bytes) and a seed is given, a secret is derived from the seed value and the default secret using the following procedure: ```c deriveSecret(u64 seed): u64 derivedSecret[24] = defaultSecret[0:192]; for (i = 0; i < 12; i++) { derivedSecret[i*2] += seed; derivedSecret[i*2+1] -= seed; } return derivedSecret; // convert to u8[192] (little-endian) ``` The derivation treats the secrets as 24 64-bit values. In XXH3 algorithms, the secret is always read similarly by treating a contiguous segment of the array as one or more 32-bit or 64-bit values. **The secret values are always read using little-endian convention**. ### Final Mixing Step (avalanche) To make sure that all input bits have a chance to impact any bit in the output digest (avalanche effect), the final step of the XXH3 algorithm is usually one of the two fixed operations that mix the bits in a 64-bit value. These operations are denoted `avalanche()` and `avalanche_XXH64()` in the following XXH3 description. ```c avalanche(u64 x): x = x xor (x >> 37); x = x * PRIME_MX1; x = x xor (x >> 32); return x; avalanche_XXH64(u64 x): x = x xor (x >> 33); x = x * PRIME64_2; x = x xor (x >> 29); x = x * PRIME64_3; x = x xor (x >> 32); return x; ``` XXH3 Algorithm Description (for small inputs) ------------------------------------- The algorithm for small inputs (0-16 bytes of input) is further divided into 4 cases: empty, 1-3 bytes, 4-8 bytes, and 9-16 bytes of input. The algorithm uses byte-swap operations. The byte-swap operation reverses the byte order in a 32-bit or 64-bit value. It is denoted `bswap32` and `bswap64` for its 32-bit and 64-bit versions, respectively. ### Empty input The hash of empty input is calculated from the seed and a segment of the secret: ```c XXH3_64_empty(): u64 secretWords[2] = secret[56:72]; return avalanche_XXH64(seed xor secretWords[0] xor secretWords[1]); XXH3_128_empty(): u64 secretWords[4] = secret[64:96]; return {avalanche_XXH64(seed xor secretWords[0] xor secretWords[1]), // lower half avalanche_XXH64(seed xor secretWords[2] xor secretWords[3])}; // higher half ``` ### 1-3 bytes of input The algorithm starts from a single 32-bit value combining the input bytes and its length: ```c u32 combined = (u32)input[inputLength-1] | ((u32)inputLength << 8) | ((u32)input[0] << 16) | ((u32)input[inputLength>>1] << 24); // LSB 8 16 24 MSB // | last byte | length | first byte | middle-or-last byte | ``` Then the final output is calculated from the value and the first 8 bytes (XXH3-64) or 16 bytes (XXH3-128) of the secret to produce the final result. The secret here is read as 32-bit values instead of the usual 64-bit values. ```c XXH3_64_1to3(): u32 secretWords[2] = secret[0:8]; u64 value = ((u64)(secretWords[0] xor secretWords[1]) + seed) xor (u64)combined; return avalanche_XXH64(value); XXH3_128_1to3(): u32 secretWords[4] = secret[0:16]; u64 low = ((u64)(secretWords[0] xor secretWords[1]) + seed) xor (u64)combined; u64 high = ((u64)(secretWords[2] xor secretWords[3]) - seed) xor (u64)(bswap32(combined) <<< 13); // note that the bswap32(combined) <<< 13 above is 32-bit rotate return {avalanche_XXH64(low), // lower half avalanche_XXH64(high)}; // higher half ``` Note that the XXH3-64 result is the lower half of XXH3-128 result. ### 4-8 bytes of input The algorithm starts from reading the first and last 4 bytes of the input as little-endian 32-bit values, and a modified seed: ```c u32 inputFirst = input[0:4]; u32 inputLast = input[inputLength-4:inputLength]; u64 modifiedSeed = seed xor ((u64)bswap32((u32)lowerHalf(seed)) << 32); ``` Again, these values are combined with a segment of the secret to produce the final value. ```c XXH3_64_4to8(): u64 secretWords[2] = secret[8:24]; u64 combined = (u64)inputLast | ((u64)inputFirst << 32); u64 value = ((secretWords[0] xor secretWords[1]) - modifiedSeed) xor combined; value = value xor (value <<< 49) xor (value <<< 24); value = value * PRIME_MX2; value = value xor ((value >> 35) + inputLength); value = value * PRIME_MX2; value = value xor (value >> 28); return value; XXH3_128_4to8(): u64 secretWords[2] = secret[16:32]; u64 combined = (u64)inputFirst | ((u64)inputLast << 32); u64 value = ((secretWords[0] xor secretWords[1]) + modifiedSeed) xor combined; u128 mulResult = (u128)value * (u128)(PRIME64_1 + (inputLength << 2)); u64 high = higherHalf(mulResult); // mulResult >> 64 u64 low = lowerHalf(mulResult); // mulResult & 0xFFFFFFFFFFFFFFFF high = high + (low << 1); low = low xor (high >> 3); low = low xor (low >> 35); low = low * PRIME_MX2; low = low xor (low >> 28); high = avalanche(high); return {low, high}; ``` ### 9-16 bytes of input The algorithm starts from reading the first and last 8 bytes of the input as little-endian 64-bit values: ```c u64 inputFirst = input[0:8]; u64 inputLast = input[inputLength-8:inputLength]; ``` Once again, these values are combined with a segment of the secret to produce the final value. ```c XXH3_64_9to16(): u64 secretWords[4] = secret[24:56]; u64 low = ((secretWords[0] xor secretWords[1]) + seed) xor inputFirst; u64 high = ((secretWords[2] xor secretWords[3]) - seed) xor inputLast; u128 mulResult = (u128)low * (u128)high; u64 value = inputLength + bswap64(low) + high + (u64)(lowerHalf(mulResult) xor higherHalf(mulResult)); return avalanche(value); XXH3_128_9to16(): u64 secretWords[4] = secret[32:64]; u64 val1 = ((secretWords[0] xor secretWords[1]) - seed) xor inputFirst xor inputLast; u64 val2 = ((secretWords[2] xor secretWords[3]) + seed) xor inputLast; u128 mulResult = (u128)val1 * (u128)PRIME64_1; u64 low = lowerHalf(mulResult) + ((u64)(inputLength - 1) << 54); u64 high = higherHalf(mulResult) + ((u64)higherHalf(val2) << 32) + (u64)lowerHalf(val2) * PRIME32_2; // the above line can also be simplified to higherHalf(mulResult) + val2 + (u64)lowerHalf(val2) * (PRIME32_2 - 1); low = low xor bswap64(high); // the following three lines are in fact a 128x64 -> 128 multiplication ({low,high} = (u128){low,high} * PRIME64_2) u128 mulResult2 = (u128)low * (u128)PRIME64_2; low = lowerHalf(mulResult2); high = higherHalf(mulResult2) + high * PRIME64_2; return {avalanche(low), // lower half avalanche(high)}; // higher half ``` XXH3 Algorithm Description (for medium inputs) ------------------------------------- This algorithm is used for medium inputs (17-240 bytes of input). Its internal hash state is stored inside 1 (XXH3-64) or 2 (XXH3-128) "accumulators", each storing an unsigned 64-bit value. ### Step 1. Initialize internal accumulators The accumulator(s) are initialized based on the input length. ```c // For XXH3-64 u64 acc = inputLength * PRIME64_1; // For XXH3-128 u64 acc[2] = {inputLength * PRIME64_1, 0}; ``` ### Step 2. Process the input This step is further divided into two cases: one for 17-128 bytes of input, and one for 129-240 bytes of input. #### Mixing operation This step uses a mixing operation that mixes a 16-byte segment of data, a 16-byte segment of secret and the seed into a 64-bit value as a building block. This operation treat the segment of data and secret as little-endian 64-bit values. ```c mixStep(u8 data[16], size secretOffset, u64 seed): u64 dataWords[2] = data[0:16]; u64 secretWords[2] = secret[secretOffset:secretOffset+16]; u128 mulResult = (u128)(dataWords[0] xor (secretWords[0] + seed)) * (u128)(dataWords[1] xor (secretWords[1] - seed)); return lowerHalf(mulResult) xor higherHalf(mulResult); ``` The mixing operation is always invoked in groups of two in XXH3-128, where two 16-byte segments of data are mixed with a 32-byte segment of secret, and the accumulators are updated accordingly. ```c mixTwoChunks(u8 data1[16], u8 data2[16], size secretOffset, u64 seed): u64 dataWords1[2] = data1[0:16]; // again, little-endian conversion u64 dataWords2[2] = data2[0:16]; acc[0] = acc[0] + mixStep(data1, secretOffset, seed); acc[1] = acc[1] + mixStep(data2, secretOffset + 16, seed); acc[0] = acc[0] xor (dataWords2[0] + dataWords2[1]); acc[1] = acc[1] xor (dataWords1[0] + dataWords1[1]); ``` The input is split into several 16-byte chunks and mixed, and the result is added to the accumulator(s). #### 17-128 bytes of input The input is read as *N* 16-byte chunks starting from the beginning and *N* chunks starting from the end, where *N* is the smallest number that these 2*N* chunks cover the whole input. These chunks are paired up and mixed, and the results are accumulated to the accumulator(s). ```c // the loop variable `i` should be signed to avoid underflow in implementation processInput_XXH3_64_17to128(): u64 numRounds = ((inputLength - 1) >> 5) + 1; for (i = numRounds - 1; i >= 0; i--) { size offsetStart = i*16; size offsetEnd = inputLength - i*16 - 16; acc += mixStep(input[offsetStart:offsetStart+16], i*32, seed); acc += mixStep(input[offsetEnd:offsetEnd+16], i*32+16, seed); } processInput_XXH3_128_17to128(): u64 numRounds = ((inputLength - 1) >> 5) + 1; for (i = numRounds - 1; i >= 0; i--) { size offsetStart = i*16; size offsetEnd = inputLength - i*16 - 16; mixTwoChunks(input[offsetStart:offsetStart+16], input[offsetEnd:offsetEnd+16], i*32, seed); } ``` #### 129-240 bytes of input The input is split into 16-byte (XXH3-64) or 32-byte (XXH3-128) chunks. The first 128 bytes are first mixed chunk by chunk, followed by an intermediate avalanche operation. Then the remaining full chunks are processed, and finally the last 16/32 bytes are treated as a chunk to process. ```c processInput_XXH3_64_129to240(): u64 numChunks = inputLength >> 4; for (i = 0; i < 8; i++) { acc += mixStep(input[i*16:i*16+16], i*16, seed); } acc = avalanche(acc); for (i = 8; i < numChunks; i++) { acc += mixStep(input[i*16:i*16+16], (i-8)*16 + 3, seed); } acc += mixStep(input[inputLength-16:inputLength], 119, seed); processInput_XXH3_128_129to240(): u64 numChunks = inputLength >> 5; for (i = 0; i < 4; i++) { mixTwoChunks(input[i*32:i*32+16], input[i*32+16:i*32+32], i*32, seed); } acc[0] = avalanche(acc[0]); acc[1] = avalanche(acc[1]); for (i = 4; i < numChunks; i++) { mixTwoChunks(input[i*32:i*32+16], input[i*32+16:i*32+32], (i-4)*32 + 3, seed); } // note that the half-chunk order and the seed is different here mixTwoChunks(input[inputLength-16:inputLength], input[inputLength-32:inputLength-16], 103, (u64)0 - seed); ``` ### Step 3. Finalization The final result is extracted from the accumulator(s). ```c XXH3_64_17to240(): return avalanche(acc); XXH3_128_17to240(): u64 low = acc[0] + acc[1]; u64 high = (acc[0] * PRIME64_1) + (acc[1] * PRIME64_4) + (((u64)inputLength - seed) * PRIME64_2); return {avalanche(low), // lower half (u64)0 - avalanche(high)}; // higher half ``` XXH3 Algorithm Description (for large inputs) ------------------------------------- This algorithm is used for inputs larger than 240 bytes. The internal hash state is stored inside 8 "accumulators", each one storing an unsigned 64-bit value. ### Step 1. Initialize internal accumulators The accumulators are initialized to fixed constants: ```c u64 acc[8] = { PRIME32_3, PRIME64_1, PRIME64_2, PRIME64_3, PRIME64_4, PRIME32_2, PRIME64_5, PRIME32_1}; ``` ### Step 2. Process blocks The input is consumed and processed one full block at a time. The size of the block depends on the length of the secret. Specifically, a block consists of several 64-byte stripes. The number of stripes per block is `floor((secretLength-64)/8)` . For the default 192-byte secret, there are 16 stripes in a block, and thus the block size is 1024 bytes. ```c secretLength = lengthInBytes(secret); // default 192; at least 136 stripesPerBlock = (secretLength-64) / 8; // default 16; at least 9 blockSize = 64 * stripesPerBlock; // default 1024; at least 576 ``` The process of processing a full block is called a *round*. It consists of the following two sub-steps: #### Step 2-1. Process stripes in the block A stripe is evenly divided into 8 lanes, of 8 bytes each. In an accumulation step, one stripe and a 64-byte contiguous segment of the secret are used to update the accumulators. Each lane reads its associated 64-bit value using little-endian convention. The accumulation step applies the following procedure: ```c accumulate(u64 stripe[8], size secretOffset): u64 secretWords[8] = secret[secretOffset:secretOffset+64]; for (i = 0; i < 8; i++) { u64 value = stripe[i] xor secretWords[i]; acc[i xor 1] = acc[i xor 1] + stripe[i]; acc[i] = acc[i] + (u64)lowerHalf(value) * (u64)higherHalf(value); // (value and 0xFFFFFFFF) * (value >> 32) } ``` The accumulation step is repeated for all stripes in a block, using different segments of the secret, starting from the first 64 bytes for the first stripe, and offset by 8 bytes for each following round: ```c round_accumulate(u8 block[blockSize]): for (n = 0; n < stripesPerBlock; n++) { u64 stripe[8] = block[n*64:n*64+64]; // 64 bytes = 8 u64s accumulate(stripe, n*8); } ``` #### Step 2-2. Scramble accumulators After the accumulation steps are finished for all stripes in the block, the accumulators are scrambled using the last 64 bytes of the secret. ```c round_scramble(): u64 secretWords[8] = secret[secretLength-64:secretLength]; for (i = 0; i < 8; i++) { acc[i] = acc[i] xor (acc[i] >> 47); acc[i] = acc[i] xor secretWords[i]; acc[i] = acc[i] * PRIME32_1; } ``` A round is thus a `round_accumulate` followed by a `round_scramble`: ```c round(u8 block[blockSize]): round_accumulate(block); round_scramble(); ``` Step 2 is looped to consume the input until there are less than or equal to `blockSize` bytes of input left. Note that we leave the last block to the next step even if it is a full block. ### Step 3. Process the last block and the last 64 bytes Accumulation steps are run for the stripes in the last block, except for the last stripe (whether it is full or not). After that, run a final accumulation step by treating the last 64 bytes as a stripe. Note that the last 64 bytes might overlap with the second-to-last block. ```c // len is the size of the last block (1 <= len <= blockSize) lastRound(u8 block[], size len, u64 lastStripe[8]): size nFullStripes = (len-1)/64; for (n = 0; n < nFullStripes; n++) { u64 stripe[8] = block[n*64:n*64+64]; accumulate(stripe, n * 8); } accumulate(lastStripe, secretLength - 71); ``` ### Step 4. Finalization In the finalization step, a merging procedure is used to extract a single 64-bit value from the accumulators, using an initial seed value and a 64-byte segment of the secret. ```c finalMerge(u64 initValue, size secretOffset): u64 secretWords[8] = secret[secretOffset:secretOffset+64]; u64 result = initValue; for (i = 0; i < 4; i++) { // 64-bit by 64-bit multiplication to 128-bit full result u128 mulResult = (u128)(acc[i*2] xor secretWords[i*2]) * (u128)(acc[i*2+1] xor secretWords[i*2+1]); result = result + (lowerHalf(mulResult) xor higherHalf(mulResult)); // (mulResult and 0xFFFFFFFFFFFFFFFF) xor (mulResult >> 64) } return avalanche(result); ``` XXH3-128 runs the merging procedure twice for the two halves of the result, using different secret segments and different initial values derived from the total input length. The XXH3-64 result is just the lower half of the XXH3-128 result. ```c XXH3_64_large(): return finalMerge((u64)inputLength * PRIME64_1, 11); XXH3_128_large(): return {finalMerge((u64)inputLength * PRIME64_1, 11), // lower half finalMerge(~((u64)inputLength * PRIME64_2), secretLength - 75)}; // higher half ``` Performance considerations ---------------------------------- The xxHash algorithms are simple and compact to implement. They provide a system independent "fingerprint" or digest of a message of arbitrary length. The algorithm allows input to be streamed and processed in multiple steps. In such case, an internal buffer is needed to ensure data is presented to the algorithm in full stripes. On 64-bit systems, the 64-bit variant `XXH64` is generally faster to compute, so it is a recommended variant, even when only 32-bit are needed. On 32-bit systems though, positions are reversed: `XXH64` performance is reduced, due to its usage of 64-bit arithmetic. `XXH32` becomes a faster variant. Finally, when vector operations are possible, `XXH3` is likely the faster variant. Reference Implementation ---------------------------------------- A reference library written in C is available at https://www.xxhash.com. The web page also links to multiple other implementations written in many different languages. It links to the [github project page](https://github.com/Cyan4973/xxHash) where an [issue board](https://github.com/Cyan4973/xxHash/issues) can be used for further public discussions on the topic. Version changes -------------------- v0.2.0: added XXH3 specification, by Adrien Wu v0.1.1: added a note on rationale for selection of constants v0.1.0: initial release Crypt-xxHash-0.09/ext/xxHash/fuzz000755001750001750 015146654120 20474 5ustar00dchernenkodchernenko000000000000Crypt-xxHash-0.09/ext/xxHash/fuzz/fuzzer.c000444001750001750 40515146654120 22301 0ustar00dchernenkodchernenko000000000000#include #include "xxhash.h" int LLVMFuzzerTestOneInput(const uint8_t *data, size_t size) { volatile XXH64_hash_t hash64 = XXH64(data, size, 0); (void)hash64; volatile XXH32_hash_t hash32 = XXH32(data, size, 0); (void)hash32; return 0; } Crypt-xxHash-0.09/lib000755001750001750 015146654120 16201 5ustar00dchernenkodchernenko000000000000Crypt-xxHash-0.09/lib/Crypt000755001750001750 015146654120 17302 5ustar00dchernenkodchernenko000000000000Crypt-xxHash-0.09/lib/Crypt/xxHash.pm000444001750001750 1035615146654120 21265 0ustar00dchernenkodchernenko000000000000package Crypt::xxHash; use strict; use base qw(Exporter); use Config (); use XSLoader; BEGIN { our $VERSION = '0.09'; XSLoader::load __PACKAGE__, $VERSION; } our @EXPORT_OK = qw[ xxhash32 xxhash32_hex xxhash64 xxhash64_hex xxhash3_64bits xxhash3_64bits_hex xxhash3_128bits_hex xxhash3_64bits_stream xxhash3_64bits_stream_update xxhash3_64bits_stream_digest xxhash3_64bits_stream_digest_hex ]; 1; __END__ =head1 NAME Crypt::xxHash - xxHash implementation for Perl =head1 SYNOPSIS use Crypt::xxHash qw/ xxhash32 xxhash32_hex xxhash64 xxhash64_hex xxhash3_64bits xxhash3_64bits_hex xxhash3_128bits_hex /; my $hash = xxhash32( $data, $seed ); my $hex = xxhash32_hex( $data, $seed ); my $hash_64 = xxhash64( $data, $seed ); my $hex_64 = xxhash64_hex( $data, $seed ); my $hash_64 = xxhash3_64bits( $data, $seed ); my $hex_64 = xxhash3_64bits_hex( $data, $seed ); my $hex_128 = xxhash3_128bits_hex( $data, $seed ); =head1 DESCRIPTION xxHash is a super fast algorithm that claims to work at speeds close to RAM limits. This package provides 32- and 64-bit hash functions. As bonus it provides 128-bit hex method. This package was inspired by C, but it includes the fresh C code and has some optimisations. Thus all hex methods are implemented in Perl XS. Also this module doesn't use C in favor of native 64 bit digits. The used version of xxHash is v0.8.0 =head2 $h = xxhash32( $data, $seed ) Returns a 32 bit hash. =head2 $h = xxhash32_hex( $data, $seed ) Returns a 32 bit hash converted into hex string. =head2 $h = xxhash64( $data, $seed ) Returns a 64 bit hash. =head2 $h = xxhash64_hex( $data, $seed ) Returns a 64 bit hash converted into hex string =head2 $h = xxhash3_64bits( $data, $seed ) Returns a 64 bit hash which calculated by using xxHash3 algorithm. =head2 $h = xxhash3_64bits_hex( $data, $seed ) Returns a 64 bit hash which calculated by using xxHash3 algorithm. This hash is converted into hex string. =head2 $h = xxhash3_128its_hex( $data, $seed ) Returns a 128 bit hash which calculated by using xxHash3 algorithm. This hash is converted into hex string. =head2 $stream = xxhash3_64bits_stream($seed) =head2 xxhash3_64bits_stream_update($stream, $more_data) =head2 $h = xxhash3_64bits_stream_digest($stream) =head2 $h = xxhash3_64bits_stream_digest_hex($stream) Get a 64 bit hash from segmented data by calling xxhash3_64bits_stream_update multiple times. sub hash_a_file { my($fh) = @_; my $stream = xxhash3_64bits_stream(12345); my $buf; while( read $fh, $buf, 1024 ) { xxhash3_64bits_stream_update($stream, $buf); } return xxhash3_64bits_stream_digest($stream); # return xxhash3_64bits_stream_digest_hex($stream); # hex version } # the resources $stream occupied will be released here. =head1 SPEED There are some official benchmark results can be found on the project web-site L =head1 BENCHMARKS Below you can find some benchmarks in comparison with C. The "xxhash32" methods in those two packages have the same realisation. so they have the same throughput capacity. But other methods have some differences: Rate Digest::xxHash::xxhash32_hex Crypt::xxHash::xxhash32_hex Digest::xxHash::xxhash32_hex 2577320/s -- -54% Crypt::xxHash::xxhash32_hex 5543237/s 115% -- Rate Digest::xxHash::xxhash64 Crypt::xxHash::xxhash64 Digest::xxHash::xxhash64 201729/s -- -99% Crypt::xxHash::xxhash64 14893617/s 7283% -- Rate Digest::xxHash::xxhash64_hex Crypt::xxHash::xxhash64_hex Digest::xxHash::xxhash64_hex 185048/s -- -96% Crypt::xxHash::xxhash64_hex 4926108/s 2562% -- =head1 LICENSE xxHash is covered by the BSD license. =head1 AUTHOR Chernenko Dmitiry cdn@cpan.org xxHash by Yann Collet. =cut Crypt-xxHash-0.09/src000755001750001750 015146654120 16222 5ustar00dchernenkodchernenko000000000000Crypt-xxHash-0.09/src/ppport.h000444001750001750 65716515146654120 20140 0ustar00dchernenkodchernenko000000000000#if 0 <<'SKIP'; #endif /* ---------------------------------------------------------------------- ppport.h -- Perl/Pollution/Portability Version 3.52 Automatically created by Devel::PPPort running under perl 5.028000. Do NOT edit this file directly! -- Edit PPPort_pm.PL and the includes in parts/inc/ instead. Use 'perldoc ppport.h' to view the documentation below. ---------------------------------------------------------------------- SKIP =pod =head1 NAME ppport.h - Perl/Pollution/Portability version 3.52 =head1 SYNOPSIS perl ppport.h [options] [source files] Searches current directory for files if no [source files] are given --help show short help --version show version --patch=file write one patch file with changes --copy=suffix write changed copies with suffix --diff=program use diff program and options --compat-version=version provide compatibility with Perl version --cplusplus accept C++ comments --quiet don't output anything except fatal errors --nodiag don't show diagnostics --nohints don't show hints --nochanges don't suggest changes --nofilter don't filter input files --strip strip all script and doc functionality from ppport.h --list-provided list provided API --list-unsupported list unsupported API --api-info=name show Perl API portability information =head1 COMPATIBILITY This version of F is designed to support operation with Perl installations back to 5.003, and has been tested up to 5.30. =head1 OPTIONS =head2 --help Display a brief usage summary. =head2 --version Display the version of F. =head2 --patch=I If this option is given, a single patch file will be created if any changes are suggested. This requires a working diff program to be installed on your system. =head2 --copy=I If this option is given, a copy of each file will be saved with the given suffix that contains the suggested changes. This does not require any external programs. Note that this does not automagically add a dot between the original filename and the suffix. If you want the dot, you have to include it in the option argument. If neither C<--patch> or C<--copy> are given, the default is to simply print the diffs for each file. This requires either C or a C program to be installed. =head2 --diff=I Manually set the diff program and options to use. The default is to use C, when installed, and output unified context diffs. =head2 --compat-version=I Tell F to check for compatibility with the given Perl version. The default is to check for compatibility with Perl version 5.003. You can use this option to reduce the output of F if you intend to be backward compatible only down to a certain Perl version. =head2 --cplusplus Usually, F will detect C++ style comments and replace them with C style comments for portability reasons. Using this option instructs F to leave C++ comments untouched. =head2 --quiet Be quiet. Don't print anything except fatal errors. =head2 --nodiag Don't output any diagnostic messages. Only portability alerts will be printed. =head2 --nohints Don't output any hints. Hints often contain useful portability notes. Warnings will still be displayed. =head2 --nochanges Don't suggest any changes. Only give diagnostic output and hints unless these are also deactivated. =head2 --nofilter Don't filter the list of input files. By default, files not looking like source code (i.e. not *.xs, *.c, *.cc, *.cpp or *.h) are skipped. =head2 --strip Strip all script and documentation functionality from F. This reduces the size of F dramatically and may be useful if you want to include F in smaller modules without increasing their distribution size too much. The stripped F will have a C<--unstrip> option that allows you to undo the stripping, but only if an appropriate C module is installed. =head2 --list-provided Lists the API elements for which compatibility is provided by F. Also lists if it must be explicitly requested, if it has dependencies, and if there are hints or warnings for it. =head2 --list-unsupported Lists the API elements that are known not to be supported by F and below which version of Perl they probably won't be available or work. =head2 --api-info=I Show portability information for API elements matching I. If I is surrounded by slashes, it is interpreted as a regular expression. =head1 DESCRIPTION In order for a Perl extension (XS) module to be as portable as possible across differing versions of Perl itself, certain steps need to be taken. =over 4 =item * Including this header is the first major one. This alone will give you access to a large part of the Perl API that hasn't been available in earlier Perl releases. Use perl ppport.h --list-provided to see which API elements are provided by ppport.h. =item * You should avoid using deprecated parts of the API. For example, using global Perl variables without the C prefix is deprecated. Also, some API functions used to have a C prefix. Using this form is also deprecated. You can safely use the supported API, as F will provide wrappers for older Perl versions. =item * If you use one of a few functions or variables that were not present in earlier versions of Perl, and that can't be provided using a macro, you have to explicitly request support for these functions by adding one or more C<#define>s in your source code before the inclusion of F. These functions or variables will be marked C in the list shown by C<--list-provided>. Depending on whether you module has a single or multiple files that use such functions or variables, you want either C or global variants. For a C function or variable (used only in a single source file), use: #define NEED_function #define NEED_variable For a global function or variable (used in multiple source files), use: #define NEED_function_GLOBAL #define NEED_variable_GLOBAL Note that you mustn't have more than one global request for the same function or variable in your project. Function / Variable Static Request Global Request ----------------------------------------------------------------------------------------- PL_parser NEED_PL_parser NEED_PL_parser_GLOBAL PL_signals NEED_PL_signals NEED_PL_signals_GLOBAL SvRX() NEED_SvRX NEED_SvRX_GLOBAL caller_cx() NEED_caller_cx NEED_caller_cx_GLOBAL croak_xs_usage() NEED_croak_xs_usage NEED_croak_xs_usage_GLOBAL die_sv() NEED_die_sv NEED_die_sv_GLOBAL eval_pv() NEED_eval_pv NEED_eval_pv_GLOBAL grok_bin() NEED_grok_bin NEED_grok_bin_GLOBAL grok_hex() NEED_grok_hex NEED_grok_hex_GLOBAL grok_number() NEED_grok_number NEED_grok_number_GLOBAL grok_numeric_radix() NEED_grok_numeric_radix NEED_grok_numeric_radix_GLOBAL grok_oct() NEED_grok_oct NEED_grok_oct_GLOBAL gv_fetchpvn_flags() NEED_gv_fetchpvn_flags NEED_gv_fetchpvn_flags_GLOBAL load_module() NEED_load_module NEED_load_module_GLOBAL mess() NEED_mess NEED_mess_GLOBAL mess_nocontext() NEED_mess_nocontext NEED_mess_nocontext_GLOBAL mess_sv() NEED_mess_sv NEED_mess_sv_GLOBAL mg_findext() NEED_mg_findext NEED_mg_findext_GLOBAL my_snprintf() NEED_my_snprintf NEED_my_snprintf_GLOBAL my_sprintf() NEED_my_sprintf NEED_my_sprintf_GLOBAL my_strlcat() NEED_my_strlcat NEED_my_strlcat_GLOBAL my_strlcpy() NEED_my_strlcpy NEED_my_strlcpy_GLOBAL my_strnlen() NEED_my_strnlen NEED_my_strnlen_GLOBAL newCONSTSUB() NEED_newCONSTSUB NEED_newCONSTSUB_GLOBAL newRV_noinc() NEED_newRV_noinc NEED_newRV_noinc_GLOBAL newSV_type() NEED_newSV_type NEED_newSV_type_GLOBAL newSVpvn_flags() NEED_newSVpvn_flags NEED_newSVpvn_flags_GLOBAL newSVpvn_share() NEED_newSVpvn_share NEED_newSVpvn_share_GLOBAL pv_display() NEED_pv_display NEED_pv_display_GLOBAL pv_escape() NEED_pv_escape NEED_pv_escape_GLOBAL pv_pretty() NEED_pv_pretty NEED_pv_pretty_GLOBAL sv_2pv_flags() NEED_sv_2pv_flags NEED_sv_2pv_flags_GLOBAL sv_2pvbyte() NEED_sv_2pvbyte NEED_sv_2pvbyte_GLOBAL sv_catpvf_mg() NEED_sv_catpvf_mg NEED_sv_catpvf_mg_GLOBAL sv_catpvf_mg_nocontext() NEED_sv_catpvf_mg_nocontext NEED_sv_catpvf_mg_nocontext_GLOBAL sv_pvn_force_flags() NEED_sv_pvn_force_flags NEED_sv_pvn_force_flags_GLOBAL sv_setpvf_mg() NEED_sv_setpvf_mg NEED_sv_setpvf_mg_GLOBAL sv_setpvf_mg_nocontext() NEED_sv_setpvf_mg_nocontext NEED_sv_setpvf_mg_nocontext_GLOBAL sv_unmagicext() NEED_sv_unmagicext NEED_sv_unmagicext_GLOBAL utf8_to_uvchr_buf() NEED_utf8_to_uvchr_buf NEED_utf8_to_uvchr_buf_GLOBAL vload_module() NEED_vload_module NEED_vload_module_GLOBAL vmess() NEED_vmess NEED_vmess_GLOBAL vnewSVpvf() NEED_vnewSVpvf NEED_vnewSVpvf_GLOBAL warner() NEED_warner NEED_warner_GLOBAL To avoid namespace conflicts, you can change the namespace of the explicitly exported functions / variables using the C macro. Just C<#define> the macro before including C: #define DPPP_NAMESPACE MyOwnNamespace_ #include "ppport.h" The default namespace is C. =back The good thing is that most of the above can be checked by running F on your source code. See the next section for details. =head1 EXAMPLES To verify whether F is needed for your module, whether you should make any changes to your code, and whether any special defines should be used, F can be run as a Perl script to check your source code. Simply say: perl ppport.h The result will usually be a list of patches suggesting changes that should at least be acceptable, if not necessarily the most efficient solution, or a fix for all possible problems. If you know that your XS module uses features only available in newer Perl releases, if you're aware that it uses C++ comments, and if you want all suggestions as a single patch file, you could use something like this: perl ppport.h --compat-version=5.6.0 --cplusplus --patch=test.diff If you only want your code to be scanned without any suggestions for changes, use: perl ppport.h --nochanges You can specify a different C program or options, using the C<--diff> option: perl ppport.h --diff='diff -C 10' This would output context diffs with 10 lines of context. If you want to create patched copies of your files instead, use: perl ppport.h --copy=.new To display portability information for the C function, use: perl ppport.h --api-info=newSVpvn Since the argument to C<--api-info> can be a regular expression, you can use perl ppport.h --api-info=/_nomg$/ to display portability information for all C<_nomg> functions or perl ppport.h --api-info=/./ to display information for all known API elements. =head1 BUGS If this version of F is causing failure during the compilation of this module, please check if newer versions of either this module or C are available on CPAN before sending a bug report. If F was generated using the latest version of C and is causing failure of this module, please send a bug report to L. Please include the following information: =over 4 =item 1. The complete output from running "perl -V" =item 2. This file. =item 3. The name and version of the module you were trying to build. =item 4. A full log of the build that failed. =item 5. Any other information that you think could be relevant. =back For the latest version of this code, please get the C module from CPAN. =head1 COPYRIGHT Version 3.x, Copyright (c) 2004-2013, Marcus Holland-Moritz. Version 2.x, Copyright (C) 2001, Paul Marquess. Version 1.x, Copyright (C) 1999, Kenneth Albanowski. This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself. =head1 SEE ALSO See L. =cut use strict; # Disable broken TRIE-optimization BEGIN { eval '${^RE_TRIE_MAXBUF} = -1' if "$]" >= 5.009004 && "$]" <= 5.009005 } my $VERSION = 3.52; my %opt = ( quiet => 0, diag => 1, hints => 1, changes => 1, cplusplus => 0, filter => 1, strip => 0, version => 0, ); my($ppport) = $0 =~ /([\w.]+)$/; my $LF = '(?:\r\n|[\r\n])'; # line feed my $HS = "[ \t]"; # horizontal whitespace # Never use C comments in this file! my $ccs = '/'.'*'; my $cce = '*'.'/'; my $rccs = quotemeta $ccs; my $rcce = quotemeta $cce; eval { require Getopt::Long; Getopt::Long::GetOptions(\%opt, qw( help quiet diag! filter! hints! changes! cplusplus strip version patch=s copy=s diff=s compat-version=s list-provided list-unsupported api-info=s )) or usage(); }; if ($@ and grep /^-/, @ARGV) { usage() if "@ARGV" =~ /^--?h(?:elp)?$/; die "Getopt::Long not found. Please don't use any options.\n"; } if ($opt{version}) { print "This is $0 $VERSION.\n"; exit 0; } usage() if $opt{help}; strip() if $opt{strip}; if (exists $opt{'compat-version'}) { my($r,$v,$s) = eval { parse_version($opt{'compat-version'}) }; if ($@) { die "Invalid version number format: '$opt{'compat-version'}'\n"; } die "Only Perl 5 is supported\n" if $r != 5; die "Invalid version number: $opt{'compat-version'}\n" if $v >= 1000 || $s >= 1000; $opt{'compat-version'} = sprintf "%d.%03d%03d", $r, $v, $s; } else { $opt{'compat-version'} = 5; } my %API = map { /^(\w+)\|([^|]*)\|([^|]*)\|(\w*)$/ ? ( $1 => { ($2 ? ( base => $2 ) : ()), ($3 ? ( todo => $3 ) : ()), (index($4, 'v') >= 0 ? ( varargs => 1 ) : ()), (index($4, 'p') >= 0 ? ( provided => 1 ) : ()), (index($4, 'n') >= 0 ? ( nothxarg => 1 ) : ()), } ) : die "invalid spec: $_" } qw( AvFILLp|5.004050||p AvFILL||| BOM_UTF8||| BhkDISABLE||5.024000| BhkENABLE||5.024000| BhkENTRY_set||5.024000| BhkENTRY||| BhkFLAGS||| CALL_BLOCK_HOOKS||| CLASS|||n CPERLscope|5.005000||p CX_CURPAD_SAVE||| CX_CURPAD_SV||| C_ARRAY_END|5.013002||p C_ARRAY_LENGTH|5.008001||p CopFILEAV|5.006000||p CopFILEGV_set|5.006000||p CopFILEGV|5.006000||p CopFILESV|5.006000||p CopFILE_set|5.006000||p CopFILE|5.006000||p CopSTASHPV_set|5.006000||p CopSTASHPV|5.006000||p CopSTASH_eq|5.006000||p CopSTASH_set|5.006000||p CopSTASH|5.006000||p CopyD|5.009002|5.004050|p Copy||| CvPADLIST||5.008001| CvSTASH||| CvWEAKOUTSIDE||| DECLARATION_FOR_LC_NUMERIC_MANIPULATION||5.021010|n DEFSV_set|5.010001||p DEFSV|5.004050||p DO_UTF8||5.006000| END_EXTERN_C|5.005000||p ENTER||| ERRSV|5.004050||p EXTEND||| EXTERN_C|5.005000||p F0convert|||n FREETMPS||| GIMME_V||5.004000|n GIMME|||n GROK_NUMERIC_RADIX|5.007002||p G_ARRAY||| G_DISCARD||| G_EVAL||| G_METHOD|5.006001||p G_NOARGS||| G_SCALAR||| G_VOID||5.004000| GetVars||| GvAV||| GvCV||| GvHV||| GvSV||| Gv_AMupdate||5.011000| HEf_SVKEY|5.003070||p HeHASH||5.003070| HeKEY||5.003070| HeKLEN||5.003070| HePV||5.004000| HeSVKEY_force||5.003070| HeSVKEY_set||5.004000| HeSVKEY||5.003070| HeUTF8|5.010001|5.008000|p HeVAL||5.003070| HvENAMELEN||5.015004| HvENAMEUTF8||5.015004| HvENAME||5.013007| HvNAMELEN_get|5.009003||p HvNAMELEN||5.015004| HvNAMEUTF8||5.015004| HvNAME_get|5.009003||p HvNAME||| INT2PTR|5.006000||p IN_LOCALE_COMPILETIME|5.007002||p IN_LOCALE_RUNTIME|5.007002||p IN_LOCALE|5.007002||p IN_PERL_COMPILETIME|5.008001||p IS_NUMBER_GREATER_THAN_UV_MAX|5.007002||p IS_NUMBER_INFINITY|5.007002||p IS_NUMBER_IN_UV|5.007002||p IS_NUMBER_NAN|5.007003||p IS_NUMBER_NEG|5.007002||p IS_NUMBER_NOT_INT|5.007002||p IVSIZE|5.006000||p IVTYPE|5.006000||p IVdf|5.006000||p LEAVE||| LIKELY|||p LINKLIST||5.013006| LVRET||| MARK||| MULTICALL||5.024000| MUTABLE_PTR|5.010001||p MUTABLE_SV|5.010001||p MY_CXT_CLONE|5.009002||p MY_CXT_INIT|5.007003||p MY_CXT|5.007003||p MoveD|5.009002|5.004050|p Move||| NOOP|5.005000||p NUM2PTR|5.006000||p NVTYPE|5.006000||p NVef|5.006001||p NVff|5.006001||p NVgf|5.006001||p Newxc|5.009003||p Newxz|5.009003||p Newx|5.009003||p Nullav||| Nullch||| Nullcv||| Nullhv||| Nullsv||| OP_CLASS||5.013007| OP_DESC||5.007003| OP_NAME||5.007003| OP_TYPE_IS_OR_WAS||5.019010| OP_TYPE_IS||5.019007| ORIGMARK||| OpHAS_SIBLING|5.021007||p OpLASTSIB_set|5.021011||p OpMAYBESIB_set|5.021011||p OpMORESIB_set|5.021011||p OpSIBLING|5.021007||p PAD_BASE_SV||| PAD_CLONE_VARS||| PAD_COMPNAME_FLAGS||| PAD_COMPNAME_GEN_set||| PAD_COMPNAME_GEN||| PAD_COMPNAME_OURSTASH||| PAD_COMPNAME_PV||| PAD_COMPNAME_TYPE||| PAD_RESTORE_LOCAL||| PAD_SAVE_LOCAL||| PAD_SAVE_SETNULLPAD||| PAD_SETSV||| PAD_SET_CUR_NOSAVE||| PAD_SET_CUR||| PAD_SVl||| PAD_SV||| PERLIO_FUNCS_CAST|5.009003||p PERLIO_FUNCS_DECL|5.009003||p PERL_ABS|5.008001||p PERL_ARGS_ASSERT_CROAK_XS_USAGE|||p PERL_BCDVERSION|5.024000||p PERL_GCC_BRACE_GROUPS_FORBIDDEN|5.008001||p PERL_HASH|5.003070||p PERL_INT_MAX|5.003070||p PERL_INT_MIN|5.003070||p PERL_LONG_MAX|5.003070||p PERL_LONG_MIN|5.003070||p PERL_MAGIC_arylen|5.007002||p PERL_MAGIC_backref|5.007002||p PERL_MAGIC_bm|5.007002||p PERL_MAGIC_collxfrm|5.007002||p PERL_MAGIC_dbfile|5.007002||p PERL_MAGIC_dbline|5.007002||p PERL_MAGIC_defelem|5.007002||p PERL_MAGIC_envelem|5.007002||p PERL_MAGIC_env|5.007002||p PERL_MAGIC_ext|5.007002||p PERL_MAGIC_fm|5.007002||p PERL_MAGIC_glob|5.024000||p PERL_MAGIC_isaelem|5.007002||p PERL_MAGIC_isa|5.007002||p PERL_MAGIC_mutex|5.024000||p PERL_MAGIC_nkeys|5.007002||p PERL_MAGIC_overload_elem|5.024000||p PERL_MAGIC_overload_table|5.007002||p PERL_MAGIC_overload|5.024000||p PERL_MAGIC_pos|5.007002||p PERL_MAGIC_qr|5.007002||p PERL_MAGIC_regdata|5.007002||p PERL_MAGIC_regdatum|5.007002||p PERL_MAGIC_regex_global|5.007002||p PERL_MAGIC_shared_scalar|5.007003||p PERL_MAGIC_shared|5.007003||p PERL_MAGIC_sigelem|5.007002||p PERL_MAGIC_sig|5.007002||p PERL_MAGIC_substr|5.007002||p PERL_MAGIC_sv|5.007002||p PERL_MAGIC_taint|5.007002||p PERL_MAGIC_tiedelem|5.007002||p PERL_MAGIC_tiedscalar|5.007002||p PERL_MAGIC_tied|5.007002||p PERL_MAGIC_utf8|5.008001||p PERL_MAGIC_uvar_elem|5.007003||p PERL_MAGIC_uvar|5.007002||p PERL_MAGIC_vec|5.007002||p PERL_MAGIC_vstring|5.008001||p PERL_PV_ESCAPE_ALL|5.009004||p PERL_PV_ESCAPE_FIRSTCHAR|5.009004||p PERL_PV_ESCAPE_NOBACKSLASH|5.009004||p PERL_PV_ESCAPE_NOCLEAR|5.009004||p PERL_PV_ESCAPE_QUOTE|5.009004||p PERL_PV_ESCAPE_RE|5.009005||p PERL_PV_ESCAPE_UNI_DETECT|5.009004||p PERL_PV_ESCAPE_UNI|5.009004||p PERL_PV_PRETTY_DUMP|5.009004||p PERL_PV_PRETTY_ELLIPSES|5.010000||p PERL_PV_PRETTY_LTGT|5.009004||p PERL_PV_PRETTY_NOCLEAR|5.010000||p PERL_PV_PRETTY_QUOTE|5.009004||p PERL_PV_PRETTY_REGPROP|5.009004||p PERL_QUAD_MAX|5.003070||p PERL_QUAD_MIN|5.003070||p PERL_REVISION|5.006000||p PERL_SCAN_ALLOW_UNDERSCORES|5.007003||p PERL_SCAN_DISALLOW_PREFIX|5.007003||p PERL_SCAN_GREATER_THAN_UV_MAX|5.007003||p PERL_SCAN_SILENT_ILLDIGIT|5.008001||p PERL_SHORT_MAX|5.003070||p PERL_SHORT_MIN|5.003070||p PERL_SIGNALS_UNSAFE_FLAG|5.008001||p PERL_SUBVERSION|5.006000||p PERL_SYS_INIT3||5.006000| PERL_SYS_INIT||| PERL_SYS_TERM||5.024000| PERL_UCHAR_MAX|5.003070||p PERL_UCHAR_MIN|5.003070||p PERL_UINT_MAX|5.003070||p PERL_UINT_MIN|5.003070||p PERL_ULONG_MAX|5.003070||p PERL_ULONG_MIN|5.003070||p PERL_UNUSED_ARG|5.009003||p PERL_UNUSED_CONTEXT|5.009004||p PERL_UNUSED_DECL|5.007002||p PERL_UNUSED_RESULT|5.021001||p PERL_UNUSED_VAR|5.007002||p PERL_UQUAD_MAX|5.003070||p PERL_UQUAD_MIN|5.003070||p PERL_USE_GCC_BRACE_GROUPS|5.009004||p PERL_USHORT_MAX|5.003070||p PERL_USHORT_MIN|5.003070||p PERL_VERSION|5.006000||p PL_DBsignal|5.005000||p PL_DBsingle|||pn PL_DBsub|||pn PL_DBtrace|||pn PL_Sv|5.005000||p PL_bufend|5.024000||p PL_bufptr|5.024000||p PL_check||5.006000| PL_compiling|5.004050||p PL_comppad_name||5.017004| PL_comppad||5.008001| PL_copline|5.024000||p PL_curcop|5.004050||p PL_curpad||5.005000| PL_curstash|5.004050||p PL_debstash|5.004050||p PL_defgv|5.004050||p PL_diehook|5.004050||p PL_dirty|5.004050||p PL_dowarn|||pn PL_errgv|5.004050||p PL_error_count|5.024000||p PL_expect|5.024000||p PL_hexdigit|5.005000||p PL_hints|5.005000||p PL_in_my_stash|5.024000||p PL_in_my|5.024000||p PL_keyword_plugin||5.011002| PL_last_in_gv|||n PL_laststatval|5.005000||p PL_lex_state|5.024000||p PL_lex_stuff|5.024000||p PL_linestr|5.024000||p PL_modglobal||5.005000|n PL_na|5.004050||pn PL_no_modify|5.006000||p PL_ofsgv|||n PL_opfreehook||5.011000|n PL_parser|5.009005||p PL_peepp||5.007003|n PL_perl_destruct_level|5.004050||p PL_perldb|5.004050||p PL_ppaddr|5.006000||p PL_rpeepp||5.013005|n PL_rsfp_filters|5.024000||p PL_rsfp|5.024000||p PL_rs|||n PL_signals|5.008001||p PL_stack_base|5.004050||p PL_stack_sp|5.004050||p PL_statcache|5.005000||p PL_stdingv|5.004050||p PL_sv_arenaroot|5.004050||p PL_sv_no|5.004050||pn PL_sv_undef|5.004050||pn PL_sv_yes|5.004050||pn PL_sv_zero|||n PL_tainted|5.004050||p PL_tainting|5.004050||p PL_tokenbuf|5.024000||p POP_MULTICALL||5.024000| POPi|||n POPl|||n POPn|||n POPpbytex||5.007001|n POPpx||5.005030|n POPp|||n POPs|||n POPul||5.006000|n POPu||5.004000|n PTR2IV|5.006000||p PTR2NV|5.006000||p PTR2UV|5.006000||p PTR2nat|5.009003||p PTR2ul|5.007001||p PTRV|5.006000||p PUSHMARK||| PUSH_MULTICALL||5.024000| PUSHi||| PUSHmortal|5.009002||p PUSHn||| PUSHp||| PUSHs||| PUSHu|5.004000||p PUTBACK||| PadARRAY||5.024000| PadMAX||5.024000| PadlistARRAY||5.024000| PadlistMAX||5.024000| PadlistNAMESARRAY||5.024000| PadlistNAMESMAX||5.024000| PadlistNAMES||5.024000| PadlistREFCNT||5.017004| PadnameIsOUR||| PadnameIsSTATE||| PadnameLEN||5.024000| PadnameOURSTASH||| PadnameOUTER||| PadnamePV||5.024000| PadnameREFCNT_dec||5.024000| PadnameREFCNT||5.024000| PadnameSV||5.024000| PadnameTYPE||| PadnameUTF8||5.021007| PadnamelistARRAY||5.024000| PadnamelistMAX||5.024000| PadnamelistREFCNT_dec||5.024000| PadnamelistREFCNT||5.024000| PerlIO_clearerr||5.007003| PerlIO_close||5.007003| PerlIO_context_layers||5.009004| PerlIO_eof||5.007003| PerlIO_error||5.007003| PerlIO_fileno||5.007003| PerlIO_fill||5.007003| PerlIO_flush||5.007003| PerlIO_get_base||5.007003| PerlIO_get_bufsiz||5.007003| PerlIO_get_cnt||5.007003| PerlIO_get_ptr||5.007003| PerlIO_read||5.007003| PerlIO_restore_errno||| PerlIO_save_errno||| PerlIO_seek||5.007003| PerlIO_set_cnt||5.007003| PerlIO_set_ptrcnt||5.007003| PerlIO_setlinebuf||5.007003| PerlIO_stderr||5.007003| PerlIO_stdin||5.007003| PerlIO_stdout||5.007003| PerlIO_tell||5.007003| PerlIO_unread||5.007003| PerlIO_write||5.007003| PerlLIO_dup2_cloexec||| PerlLIO_dup_cloexec||| PerlLIO_open3_cloexec||| PerlLIO_open_cloexec||| PerlProc_pipe_cloexec||| PerlSock_accept_cloexec||| PerlSock_socket_cloexec||| PerlSock_socketpair_cloexec||| Perl_langinfo|||n Perl_setlocale|||n PoisonFree|5.009004||p PoisonNew|5.009004||p PoisonWith|5.009004||p Poison|5.008000||p READ_XDIGIT||5.017006| REPLACEMENT_CHARACTER_UTF8||| RESTORE_LC_NUMERIC||5.024000| RETVAL|||n Renewc||| Renew||| SAVECLEARSV||| SAVECOMPPAD||| SAVEPADSV||| SAVETMPS||| SAVE_DEFSV|5.004050||p SPAGAIN||| SP||| START_EXTERN_C|5.005000||p START_MY_CXT|5.007003||p STMT_END|||p STMT_START|||p STORE_LC_NUMERIC_FORCE_TO_UNDERLYING||5.024000| STORE_LC_NUMERIC_SET_TO_NEEDED||5.024000| STR_WITH_LEN|5.009003||p ST||| SV_CONST_RETURN|5.009003||p SV_COW_DROP_PV|5.008001||p SV_COW_SHARED_HASH_KEYS|5.009005||p SV_GMAGIC|5.007002||p SV_HAS_TRAILING_NUL|5.009004||p SV_IMMEDIATE_UNREF|5.007001||p SV_MUTABLE_RETURN|5.009003||p SV_NOSTEAL|5.009002||p SV_SMAGIC|5.009003||p SV_UTF8_NO_ENCODING|5.008001||p SVfARG|5.009005||p SVf_UTF8|5.006000||p SVf|5.006000||p SVt_INVLIST||5.019002| SVt_IV||| SVt_NULL||| SVt_NV||| SVt_PVAV||| SVt_PVCV||| SVt_PVFM||| SVt_PVGV||| SVt_PVHV||| SVt_PVIO||| SVt_PVIV||| SVt_PVLV||| SVt_PVMG||| SVt_PVNV||| SVt_PV||| SVt_REGEXP||5.011000| Safefree||| Slab_Alloc||| Slab_Free||| Slab_to_ro||| Slab_to_rw||| StructCopy||| SvCUR_set||| SvCUR||| SvEND||| SvGAMAGIC||5.006001| SvGETMAGIC|5.004050||p SvGROW||| SvIOK_UV||5.006000| SvIOK_notUV||5.006000| SvIOK_off||| SvIOK_only_UV||5.006000| SvIOK_only||| SvIOK_on||| SvIOKp||| SvIOK||| SvIVX||| SvIV_nomg|5.009001||p SvIV_set||| SvIVx||| SvIV||| SvIsCOW_shared_hash||5.008003| SvIsCOW||5.008003| SvLEN_set||| SvLEN||| SvLOCK||5.007003| SvMAGIC_set|5.009003||p SvNIOK_off||| SvNIOKp||| SvNIOK||| SvNOK_off||| SvNOK_only||| SvNOK_on||| SvNOKp||| SvNOK||| SvNVX||| SvNV_nomg||5.013002| SvNV_set||| SvNVx||| SvNV||| SvOK||| SvOOK_offset||5.011000| SvOOK||| SvPOK_off||| SvPOK_only_UTF8||5.006000| SvPOK_only||| SvPOK_on||| SvPOKp||| SvPOK||| SvPVCLEAR||| SvPVX_const|5.009003||p SvPVX_mutable|5.009003||p SvPVX||| SvPV_const|5.009003||p SvPV_flags_const_nolen|5.009003||p SvPV_flags_const|5.009003||p SvPV_flags_mutable|5.009003||p SvPV_flags|5.007002||p SvPV_force_flags_mutable|5.009003||p SvPV_force_flags_nolen|5.009003||p SvPV_force_flags|5.007002||p SvPV_force_mutable|5.009003||p SvPV_force_nolen|5.009003||p SvPV_force_nomg_nolen|5.009003||p SvPV_force_nomg|5.007002||p SvPV_force|||p SvPV_mutable|5.009003||p SvPV_nolen_const|5.009003||p SvPV_nolen|5.006000||p SvPV_nomg_const_nolen|5.009003||p SvPV_nomg_const|5.009003||p SvPV_nomg_nolen|5.013007||p SvPV_nomg|5.007002||p SvPV_renew|5.009003||p SvPV_set||| SvPVbyte_force||5.009002| SvPVbyte_nolen||5.006000| SvPVbytex_force||5.006000| SvPVbytex||5.006000| SvPVbyte|5.006000||p SvPVutf8_force||5.006000| SvPVutf8_nolen||5.006000| SvPVutf8x_force||5.006000| SvPVutf8x||5.006000| SvPVutf8||5.006000| SvPVx||| SvPV||| SvREADONLY_off||| SvREADONLY_on||| SvREADONLY||| SvREFCNT_dec_NN||5.017007| SvREFCNT_dec||| SvREFCNT_inc_NN|5.009004||p SvREFCNT_inc_simple_NN|5.009004||p SvREFCNT_inc_simple_void_NN|5.009004||p SvREFCNT_inc_simple_void|5.009004||p SvREFCNT_inc_simple|5.009004||p SvREFCNT_inc_void_NN|5.009004||p SvREFCNT_inc_void|5.009004||p SvREFCNT_inc|||p SvREFCNT||| SvROK_off||| SvROK_on||| SvROK||| SvRV_set|5.009003||p SvRV||| SvRXOK|5.009005||p SvRX|5.009005||p SvSETMAGIC||| SvSHARED_HASH|5.009003||p SvSHARE||5.007003| SvSTASH_set|5.009003||p SvSTASH||| SvSetMagicSV_nosteal||5.004000| SvSetMagicSV||5.004000| SvSetSV_nosteal||5.004000| SvSetSV||| SvTAINTED_off||5.004000| SvTAINTED_on||5.004000| SvTAINTED||5.004000| SvTAINT||| SvTHINKFIRST||| SvTRUE_nomg||5.013006| SvTRUE||| SvTYPE||| SvUNLOCK||5.007003| SvUOK|5.007001|5.006000|p SvUPGRADE||| SvUTF8_off||5.006000| SvUTF8_on||5.006000| SvUTF8||5.006000| SvUVXx|5.004000||p SvUVX|5.004000||p SvUV_nomg|5.009001||p SvUV_set|5.009003||p SvUVx|5.004000||p SvUV|5.004000||p SvVOK||5.008001| SvVSTRING_mg|5.009004||p THIS|||n UNDERBAR|5.009002||p UNICODE_REPLACEMENT|||p UNLIKELY|||p UTF8SKIP||5.006000| UTF8_IS_INVARIANT||| UTF8_IS_NONCHAR||| UTF8_IS_SUPER||| UTF8_IS_SURROGATE||| UTF8_MAXBYTES|5.009002||p UTF8_SAFE_SKIP|||p UVCHR_IS_INVARIANT||| UVCHR_SKIP||5.022000| UVSIZE|5.006000||p UVTYPE|5.006000||p UVXf|5.007001||p UVof|5.006000||p UVuf|5.006000||p UVxf|5.006000||p WARN_ALL|5.006000||p WARN_AMBIGUOUS|5.006000||p WARN_ASSERTIONS|5.024000||p WARN_BAREWORD|5.006000||p WARN_CLOSED|5.006000||p WARN_CLOSURE|5.006000||p WARN_DEBUGGING|5.006000||p WARN_DEPRECATED|5.006000||p WARN_DIGIT|5.006000||p WARN_EXEC|5.006000||p WARN_EXITING|5.006000||p WARN_GLOB|5.006000||p WARN_INPLACE|5.006000||p WARN_INTERNAL|5.006000||p WARN_IO|5.006000||p WARN_LAYER|5.008000||p WARN_MALLOC|5.006000||p WARN_MISC|5.006000||p WARN_NEWLINE|5.006000||p WARN_NUMERIC|5.006000||p WARN_ONCE|5.006000||p WARN_OVERFLOW|5.006000||p WARN_PACK|5.006000||p WARN_PARENTHESIS|5.006000||p WARN_PIPE|5.006000||p WARN_PORTABLE|5.006000||p WARN_PRECEDENCE|5.006000||p WARN_PRINTF|5.006000||p WARN_PROTOTYPE|5.006000||p WARN_QW|5.006000||p WARN_RECURSION|5.006000||p WARN_REDEFINE|5.006000||p WARN_REGEXP|5.006000||p WARN_RESERVED|5.006000||p WARN_SEMICOLON|5.006000||p WARN_SEVERE|5.006000||p WARN_SIGNAL|5.006000||p WARN_SUBSTR|5.006000||p WARN_SYNTAX|5.006000||p WARN_TAINT|5.006000||p WARN_THREADS|5.008000||p WARN_UNINITIALIZED|5.006000||p WARN_UNOPENED|5.006000||p WARN_UNPACK|5.006000||p WARN_UNTIE|5.006000||p WARN_UTF8|5.006000||p WARN_VOID|5.006000||p WIDEST_UTYPE|5.015004||p XCPT_CATCH|5.009002||p XCPT_RETHROW|5.009002||p XCPT_TRY_END|5.009002||p XCPT_TRY_START|5.009002||p XPUSHi||| XPUSHmortal|5.009002||p XPUSHn||| XPUSHp||| XPUSHs||| XPUSHu|5.004000||p XSPROTO|5.010000||p XSRETURN_EMPTY||| XSRETURN_IV||| XSRETURN_NO||| XSRETURN_NV||| XSRETURN_PV||| XSRETURN_UNDEF||| XSRETURN_UV|5.008001||p XSRETURN_YES||| XSRETURN|||p XST_mIV||| XST_mNO||| XST_mNV||| XST_mPV||| XST_mUNDEF||| XST_mUV|5.008001||p XST_mYES||| XS_APIVERSION_BOOTCHECK||5.024000| XS_EXTERNAL||5.024000| XS_INTERNAL||5.024000| XS_VERSION_BOOTCHECK||5.024000| XS_VERSION||| XSprePUSH|5.006000||p XS||| XopDISABLE||5.024000| XopENABLE||5.024000| XopENTRYCUSTOM||5.024000| XopENTRY_set||5.024000| XopENTRY||5.024000| XopFLAGS||5.013007| ZeroD|5.009002||p Zero||| __ASSERT_|||p _aMY_CXT|5.007003||p _inverse_folds||| _is_grapheme||| _is_in_locale_category||| _new_invlist_C_array||| _pMY_CXT|5.007003||p _to_fold_latin1|||n _to_upper_title_latin1||| _to_utf8_case||| _variant_byte_number|||n _warn_problematic_locale|||n aMY_CXT_|5.007003||p aMY_CXT|5.007003||p aTHXR_|5.024000||p aTHXR|5.024000||p aTHX_|5.006000||p aTHX|5.006000||p abort_execution||| add_above_Latin1_folds||| add_data|||n add_multi_match||| add_utf16_textfilter||| adjust_size_and_find_bucket|||n advance_one_LB||| advance_one_SB||| advance_one_WB||| allocmy||| amagic_call||| amagic_cmp_locale||| amagic_cmp||| amagic_deref_call||5.013007| amagic_i_ncmp||| amagic_is_enabled||| amagic_ncmp||| anonymise_cv_maybe||| any_dup||| ao||| apply_attrs_my||| apply_attrs||| apply||| argvout_final||| assert_uft8_cache_coherent||| assignment_type||| atfork_lock||5.007003|n atfork_unlock||5.007003|n av_arylen_p||5.009003| av_clear||| av_delete||5.006000| av_exists||5.006000| av_extend_guts||| av_extend||| av_fetch||| av_fill||| av_iter_p||5.011000| av_len||| av_make||| av_nonelem||| av_pop||| av_push||| av_reify||| av_shift||| av_store||| av_tindex|5.017009|5.017009|p av_top_index|5.017009|5.017009|p av_undef||| av_unshift||| ax|||n backup_one_GCB||| backup_one_LB||| backup_one_SB||| backup_one_WB||| bad_type_gv||| bad_type_pv||| bind_match||| block_end||5.004000| block_gimme||5.004000| block_start||5.004000| blockhook_register||5.013003| boolSV|5.004000||p boot_core_PerlIO||| boot_core_UNIVERSAL||| boot_core_mro||| bytes_cmp_utf8||5.013007| cBOOL|5.013000||p call_argv|5.006000||p call_atexit||5.006000| call_list||5.004000| call_method|5.006000||p call_pv|5.006000||p call_sv|5.006000||p caller_cx|5.013005|5.006000|p calloc||5.007002|n cando||| cast_i32||5.006000|n cast_iv||5.006000|n cast_ulong||5.006000|n cast_uv||5.006000|n category_name|||n change_engine_size||| check_and_deprecate||| check_type_and_open||| check_uni||| checkcomma||| ckWARN2_d||| ckWARN2||| ckWARN3_d||| ckWARN3||| ckWARN4_d||| ckWARN4||| ckWARN_d||| ckWARN|5.006000||p ck_entersub_args_core||| ck_entersub_args_list||5.013006| ck_entersub_args_proto_or_list||5.013006| ck_entersub_args_proto||5.013006| ck_warner_d||5.011001|v ck_warner||5.011001|v ckwarn_common||| ckwarn_d||5.009003| ckwarn||5.009003| clear_defarray||5.023008| clear_special_blocks||| clone_params_del|||n clone_params_new|||n closest_cop||| cntrl_to_mnemonic|||n compute_EXACTish|||n construct_ahocorasick_from_trie||| cop_free||| cop_hints_2hv||5.013007| cop_hints_fetch_pvn||5.013007| cop_hints_fetch_pvs||5.013007| cop_hints_fetch_pv||5.013007| cop_hints_fetch_sv||5.013007| cophh_2hv||5.013007| cophh_copy||5.013007| cophh_delete_pvn||5.013007| cophh_delete_pvs||5.013007| cophh_delete_pv||5.013007| cophh_delete_sv||5.013007| cophh_fetch_pvn||5.013007| cophh_fetch_pvs||5.013007| cophh_fetch_pv||5.013007| cophh_fetch_sv||5.013007| cophh_free||5.013007| cophh_new_empty||5.024000| cophh_store_pvn||5.013007| cophh_store_pvs||5.013007| cophh_store_pv||5.013007| cophh_store_sv||5.013007| core_prototype||| coresub_op||| cr_textfilter||| croak_caller|||vn croak_memory_wrap|5.019003||pn croak_no_mem|||n croak_no_modify|5.013003||pn croak_nocontext|||pvn croak_popstack|||n croak_sv|5.013001||p croak_xs_usage|5.010001||pn croak|||v csighandler||5.009003|n current_re_engine||| curse||| custom_op_desc||5.007003| custom_op_get_field||| custom_op_name||5.007003| custom_op_register||5.013007| custom_op_xop||5.013007| cv_clone_into||| cv_clone||| cv_const_sv_or_av|||n cv_const_sv||5.003070|n cv_dump||| cv_forget_slab||| cv_get_call_checker_flags||| cv_get_call_checker||5.013006| cv_name||5.021005| cv_set_call_checker_flags||5.021004| cv_set_call_checker||5.013006| cv_undef_flags||| cv_undef||| cvgv_from_hek||| cvgv_set||| cvstash_set||| cx_dump||5.005000| cx_dup||| cxinc||| dAXMARK|5.009003||p dAX|5.007002||p dITEMS|5.007002||p dMARK||| dMULTICALL||5.009003| dMY_CXT_SV|5.007003||p dMY_CXT|5.007003||p dNOOP|5.006000||p dORIGMARK||| dSP||| dTHR|5.004050||p dTHXR|5.024000||p dTHXa|5.006000||p dTHXoa|5.006000||p dTHX|5.006000||p dUNDERBAR|5.009002||p dVAR|5.009003||p dXCPT|5.009002||p dXSARGS||| dXSI32||| dXSTARG|5.006000||p deb_curcv||| deb_nocontext|||vn deb_stack_all||| deb_stack_n||| debop||5.005000| debprofdump||5.005000| debprof||| debstackptrs||5.007003| debstack||5.007003| debug_start_match||| deb||5.007003|v defelem_target||| del_sv||| delimcpy_no_escape|||n delimcpy||5.004000|n despatch_signals||5.007001| destroy_matcher||| die_nocontext|||vn die_sv|5.013001||p die_unwind||| die|||v dirp_dup||| div128||| djSP||| do_aexec5||| do_aexec||| do_aspawn||| do_binmode||5.004050| do_chomp||| do_close||| do_delete_local||| do_dump_pad||| do_eof||| do_exec3||| do_exec||| do_gv_dump||5.006000| do_gvgv_dump||5.006000| do_hv_dump||5.006000| do_ipcctl||| do_ipcget||| do_join||| do_magic_dump||5.006000| do_msgrcv||| do_msgsnd||| do_ncmp||| do_oddball||| do_op_dump||5.006000| do_open9||5.006000| do_openn||5.007001| do_open||5.003070| do_pmop_dump||5.006000| do_print||| do_readline||| do_seek||| do_semop||| do_shmio||| do_smartmatch||| do_spawn_nowait||| do_spawn||| do_sprintf||| do_sv_dump||5.006000| do_sysseek||| do_tell||| do_trans_complex_utf8||| do_trans_complex||| do_trans_count_utf8||| do_trans_count||| do_trans_simple_utf8||| do_trans_simple||| do_trans||| do_vecget||| do_vecset||| do_vop||| docatch||| does_utf8_overflow|||n doeval_compile||| dofile||| dofindlabel||| doform||| doing_taint||5.008001|n dooneliner||| doopen_pm||| doparseform||| dopoptoeval||| dopoptogivenfor||| dopoptolabel||| dopoptoloop||| dopoptosub_at||| dopoptowhen||| doref||5.009003| dounwind||| dowantarray||| drand48_init_r|||n drand48_r|||n dtrace_probe_call||| dtrace_probe_load||| dtrace_probe_op||| dtrace_probe_phase||| dump_all_perl||| dump_all||5.006000| dump_c_backtrace||| dump_eval||5.006000| dump_exec_pos||| dump_form||5.006000| dump_indent||5.006000|v dump_mstats||| dump_packsubs_perl||| dump_packsubs||5.006000| dump_regex_sets_structures||| dump_sub_perl||| dump_sub||5.006000| dump_sv_child||| dump_trie_interim_list||| dump_trie_interim_table||| dump_trie||| dump_vindent||5.006000| dumpuntil||| dup_attrlist||| dup_warnings||| edit_distance|||n emulate_setlocale|||n eval_pv|5.006000||p eval_sv|5.006000||p exec_failed||| expect_number||| fbm_compile||5.005000| fbm_instr||5.005000| feature_is_enabled||| filter_add||| filter_del||| filter_gets||| filter_read||| finalize_optree||| finalize_op||| find_and_forget_pmops||| find_array_subscript||| find_beginning||| find_byclass||| find_default_stash||| find_hash_subscript||| find_in_my_stash||| find_lexical_cv||| find_next_masked|||n find_runcv_where||| find_runcv||5.008001| find_rundefsv||5.013002| find_script||| find_span_end_mask|||n find_span_end|||n first_symbol|||n fixup_errno_string||| foldEQ_latin1_s2_folded|||n foldEQ_latin1||5.013008|n foldEQ_locale||5.013002|n foldEQ_utf8||5.013002| foldEQ||5.013002|n fold_constants||| forbid_setid||| force_ident_maybe_lex||| force_ident||| force_list||| force_next||| force_strict_version||| force_version||| force_word||| forget_pmop||| form_nocontext|||vn form||5.004000|v fp_dup||| fprintf_nocontext|||vn free_c_backtrace||| free_global_struct||| free_tied_hv_pool||| free_tmps||| gen_constant_list||| get_ANYOFM_contents||| get_ANYOF_cp_list_for_ssc||| get_and_check_backslash_N_name_wrapper||| get_and_check_backslash_N_name||| get_aux_mg||| get_av|5.006000||p get_c_backtrace_dump||| get_c_backtrace||| get_context||5.006000|n get_cvn_flags||| get_cvs|5.011000||p get_cv|5.006000||p get_db_sub||| get_debug_opts||| get_hash_seed||| get_hv|5.006000||p get_mstats||| get_no_modify||| get_num||| get_op_descs||5.005000| get_op_names||5.005000| get_opargs||| get_ppaddr||5.006000| get_sv|5.006000||p get_vtbl||5.005030| getcwd_sv||5.007002| getenv_len||| glob_2number||| glob_assign_glob||| gp_dup||| gp_free||| gp_ref||| grok_atoUV|||n grok_bin|5.007003||p grok_bslash_N||| grok_hex|5.007003||p grok_infnan||5.021004| grok_number_flags||5.021002| grok_number|5.007002||p grok_numeric_radix|5.007002||p grok_oct|5.007003||p group_end||| gv_AVadd||| gv_HVadd||| gv_IOadd||| gv_SVadd||| gv_add_by_type||5.011000| gv_autoload4||5.004000| gv_autoload_pvn||5.015004| gv_autoload_pv||5.015004| gv_autoload_sv||5.015004| gv_check||| gv_const_sv||5.009003| gv_dump||5.006000| gv_efullname3||5.003070| gv_efullname4||5.006001| gv_efullname||| gv_fetchfile_flags||5.009005| gv_fetchfile||| gv_fetchmeth_autoload||5.007003| gv_fetchmeth_internal||| gv_fetchmeth_pv_autoload||5.015004| gv_fetchmeth_pvn_autoload||5.015004| gv_fetchmeth_pvn||5.015004| gv_fetchmeth_pv||5.015004| gv_fetchmeth_sv_autoload||5.015004| gv_fetchmeth_sv||5.015004| gv_fetchmethod_autoload||5.004000| gv_fetchmethod||| gv_fetchmeth||| gv_fetchpvn_flags|5.009002||p gv_fetchpvs|5.009004||p gv_fetchpv||| gv_fetchsv||| gv_fullname3||5.003070| gv_fullname4||5.006001| gv_fullname||| gv_handler||5.007001| gv_init_pvn||| gv_init_pv||5.015004| gv_init_svtype||| gv_init_sv||5.015004| gv_init||| gv_is_in_main||| gv_magicalize_isa||| gv_magicalize||| gv_name_set||5.009004| gv_override||| gv_setref||| gv_stashpvn_internal||| gv_stashpvn|5.003070||p gv_stashpvs|5.009003||p gv_stashpv||| gv_stashsvpvn_cached||| gv_stashsv||| handle_named_backref||| handle_possible_posix||| handle_regex_sets||| handle_user_defined_property||| he_dup||| hek_dup||| hfree_next_entry||| hsplit||| hv_assert||| hv_auxinit_internal|||n hv_auxinit||| hv_clear_placeholders||5.009001| hv_clear||| hv_common_key_len||5.010000| hv_common||5.010000| hv_copy_hints_hv||5.009004| hv_delayfree_ent||5.004000| hv_delete_ent||5.003070| hv_delete||| hv_eiter_p||5.009003| hv_eiter_set||5.009003| hv_ename_add||| hv_ename_delete||| hv_exists_ent||5.003070| hv_exists||| hv_fetch_ent||5.003070| hv_fetchs|5.009003||p hv_fetch||| hv_fill||5.013002| hv_free_ent_ret||| hv_free_entries||| hv_free_ent||5.004000| hv_iterinit||| hv_iterkeysv||5.003070| hv_iterkey||| hv_iternextsv||| hv_iternext||| hv_iterval||| hv_ksplit||5.003070| hv_magic_check|||n hv_magic||| hv_name_set||5.009003| hv_notallowed||| hv_placeholders_get||5.009003| hv_placeholders_p||| hv_placeholders_set||5.009003| hv_pushkv||| hv_rand_set||5.018000| hv_riter_p||5.009003| hv_riter_set||5.009003| hv_scalar||5.009001| hv_store_ent||5.003070| hv_stores|5.009004||p hv_store||| hv_undef_flags||| hv_undef||| ibcmp_locale||5.004000| ibcmp_utf8||5.007003| ibcmp||| incline||| incpush_if_exists||| incpush_use_sep||| incpush||| ingroup||| init_argv_symbols||| init_constants||| init_dbargs||| init_debugger||| init_global_struct||| init_ids||| init_interp||| init_main_stash||| init_named_cv||| init_perllib||| init_postdump_symbols||| init_predump_symbols||| init_stacks||5.005000| init_tm||5.007002| init_uniprops||| inplace_aassign||| instr|||n intro_my||5.004000| intuit_method||| intuit_more||| invert||| invoke_exception_hook||| io_close||| isALNUMC_A|||p isALNUMC|5.006000||p isALNUM_A|||p isALNUM|||p isALPHANUMERIC_A|||p isALPHANUMERIC|5.017008|5.017008|p isALPHA_A|||p isALPHA|||p isASCII_A|||p isASCII|5.006000||p isBLANK_A|||p isBLANK|5.006001||p isC9_STRICT_UTF8_CHAR|||n isCNTRL_A|||p isCNTRL|5.006000||p isDIGIT_A|||p isDIGIT|||p isFF_OVERLONG|||n isFOO_utf8_lc||| isGCB||| isGRAPH_A|||p isGRAPH|5.006000||p isIDCONT_A|||p isIDCONT|5.017008|5.017008|p isIDFIRST_A|||p isIDFIRST|||p isLB||| isLOWER_A|||p isLOWER|||p isOCTAL_A|||p isOCTAL|5.013005|5.013005|p isPRINT_A|||p isPRINT|5.004000||p isPSXSPC_A|||p isPSXSPC|5.006001||p isPUNCT_A|||p isPUNCT|5.006000||p isSB||| isSCRIPT_RUN||| isSPACE_A|||p isSPACE|||p isSTRICT_UTF8_CHAR|||n isUPPER_A|||p isUPPER|||p isUTF8_CHAR_flags||| isUTF8_CHAR||5.021001|n isWB||| isWORDCHAR_A|||p isWORDCHAR|5.013006|5.013006|p isXDIGIT_A|||p isXDIGIT|5.006000||p is_an_int||| is_ascii_string||5.011000|n is_c9strict_utf8_string_loclen|||n is_c9strict_utf8_string_loc|||n is_c9strict_utf8_string|||n is_handle_constructor|||n is_invariant_string||5.021007|n is_lvalue_sub||5.007001| is_safe_syscall||5.019004| is_ssc_worth_it|||n is_strict_utf8_string_loclen|||n is_strict_utf8_string_loc|||n is_strict_utf8_string|||n is_utf8_char_buf||5.015008|n is_utf8_common_with_len||| is_utf8_common||| is_utf8_cp_above_31_bits|||n is_utf8_fixed_width_buf_flags|||n is_utf8_fixed_width_buf_loc_flags|||n is_utf8_fixed_width_buf_loclen_flags|||n is_utf8_invariant_string_loc|||n is_utf8_invariant_string|||n is_utf8_non_invariant_string|||n is_utf8_overlong_given_start_byte_ok|||n is_utf8_string_flags|||n is_utf8_string_loc_flags|||n is_utf8_string_loclen_flags|||n is_utf8_string_loclen||5.009003|n is_utf8_string_loc||5.008001|n is_utf8_string||5.006001|n is_utf8_valid_partial_char_flags|||n is_utf8_valid_partial_char|||n isa_lookup||| isinfnansv||| isinfnan||5.021004|n items|||n ix|||n jmaybe||| join_exact||| keyword_plugin_standard||| keyword||| leave_scope||| lex_stuff_pvs||5.013005| listkids||| list||| load_module_nocontext|||vn load_module|5.006000||pv localize||| looks_like_bool||| looks_like_number||| lop||| mPUSHi|5.009002||p mPUSHn|5.009002||p mPUSHp|5.009002||p mPUSHs|5.010001||p mPUSHu|5.009002||p mXPUSHi|5.009002||p mXPUSHn|5.009002||p mXPUSHp|5.009002||p mXPUSHs|5.010001||p mXPUSHu|5.009002||p magic_clear_all_env||| magic_cleararylen_p||| magic_clearenv||| magic_clearhints||| magic_clearhint||| magic_clearisa||| magic_clearpack||| magic_clearsig||| magic_copycallchecker||| magic_dump||5.006000| magic_existspack||| magic_freearylen_p||| magic_freeovrld||| magic_getarylen||| magic_getdebugvar||| magic_getdefelem||| magic_getnkeys||| magic_getpack||| magic_getpos||| magic_getsig||| magic_getsubstr||| magic_gettaint||| magic_getuvar||| magic_getvec||| magic_get||| magic_killbackrefs||| magic_methcall1||| magic_methcall|||v magic_methpack||| magic_nextpack||| magic_regdata_cnt||| magic_regdatum_get||| magic_regdatum_set||| magic_scalarpack||| magic_set_all_env||| magic_setarylen||| magic_setcollxfrm||| magic_setdbline||| magic_setdebugvar||| magic_setdefelem||| magic_setenv||| magic_sethint||| magic_setisa||| magic_setlvref||| magic_setmglob||| magic_setnkeys||| magic_setnonelem||| magic_setpack||| magic_setpos||| magic_setregexp||| magic_setsig||| magic_setsubstr||| magic_settaint||| magic_setutf8||| magic_setuvar||| magic_setvec||| magic_set||| magic_sizepack||| magic_wipepack||| make_matcher||| make_trie||| malloc_good_size|||n malloced_size|||n malloc||5.007002|n markstack_grow||5.021001| matcher_matches_sv||| maybe_multimagic_gv||| mayberelocate||| measure_struct||| memEQs|5.009005||p memEQ|5.004000||p memNEs|5.009005||p memNE|5.004000||p mem_collxfrm||| mem_log_alloc|||n mem_log_common|||n mem_log_free|||n mem_log_realloc|||n mess_alloc||| mess_nocontext|||pvn mess_sv|5.013001||p mess|5.006000||pv mfree||5.007002|n mg_clear||| mg_copy||| mg_dup||| mg_find_mglob||| mg_findext|5.013008||pn mg_find|||n mg_free_type||5.013006| mg_freeext||| mg_free||| mg_get||| mg_localize||| mg_magical|||n mg_set||| mg_size||5.005000| mini_mktime||5.007002|n minus_v||| missingterm||| mode_from_discipline||| modkids||| more_bodies||| more_sv||| moreswitches||| move_proto_attr||| mro_clean_isarev||| mro_gather_and_rename||| mro_get_from_name||5.010001| mro_get_linear_isa_dfs||| mro_get_linear_isa||5.009005| mro_get_private_data||5.010001| mro_isa_changed_in||| mro_meta_dup||| mro_meta_init||| mro_method_changed_in||5.009005| mro_package_moved||| mro_register||5.010001| mro_set_mro||5.010001| mro_set_private_data||5.010001| mul128||| multiconcat_stringify||| multideref_stringify||| my_atof2||5.007002| my_atof3||| my_atof||5.006000| my_attrs||| my_bytes_to_utf8|||n my_chsize||| my_clearenv||| my_cxt_index||| my_cxt_init||| my_dirfd||5.009005|n my_exit_jump||| my_exit||| my_failure_exit||5.004000| my_fflush_all||5.006000| my_fork||5.007003|n my_kid||| my_lstat_flags||| my_lstat||5.024000| my_memrchr|||n my_mkostemp|||n my_mkstemp_cloexec|||n my_mkstemp|||n my_nl_langinfo|||n my_pclose||5.003070| my_popen_list||5.007001| my_popen||5.003070| my_setenv||| my_snprintf|5.009004||pvn my_socketpair||5.007003|n my_sprintf|5.009003||pvn my_stat_flags||| my_stat||5.024000| my_strerror||| my_strftime||5.007002| my_strlcat|5.009004||pn my_strlcpy|5.009004||pn my_strnlen|||pn my_strtod|||n my_unexec||| my_vsnprintf||5.009004|n need_utf8|||n newANONATTRSUB||5.006000| newANONHASH||| newANONLIST||| newANONSUB||| newASSIGNOP||| newATTRSUB_x||| newATTRSUB||5.006000| newAVREF||| newAV||| newBINOP||| newCONDOP||| newCONSTSUB_flags||5.015006| newCONSTSUB|5.004050||p newCVREF||| newDEFSVOP||5.021006| newFORM||| newFOROP||5.013007| newGIVENOP||5.009003| newGIVWHENOP||| newGVOP||| newGVREF||| newGVgen_flags||5.015004| newGVgen||| newHVREF||| newHVhv||5.005000| newHV||| newIO||| newLISTOP||| newLOGOP||| newLOOPEX||| newLOOPOP||| newMETHOP_internal||| newMETHOP_named||5.021005| newMETHOP||5.021005| newMYSUB||5.017004| newNULLLIST||| newOP||| newPADOP||| newPMOP||| newPROG||| newPVOP||| newRANGE||| newRV_inc|5.004000||p newRV_noinc|5.004000||p newRV||| newSLICEOP||| newSTATEOP||| newSTUB||| newSUB||| newSVOP||| newSVREF||| newSV_type|5.009005||p newSVavdefelem||| newSVhek||5.009003| newSViv||| newSVnv||| newSVpadname||5.017004| newSVpv_share||5.013006| newSVpvf_nocontext|||vn newSVpvf||5.004000|v newSVpvn_flags|5.010001||p newSVpvn_share|5.007001||p newSVpvn_utf8|5.010001||p newSVpvn|5.004050||p newSVpvs_flags|5.010001||p newSVpvs_share|5.009003||p newSVpvs|5.009003||p newSVpv||| newSVrv||| newSVsv_flags||| newSVsv_nomg||| newSVsv||| newSVuv|5.006000||p newSV||| newUNOP_AUX||5.021007| newUNOP||| newWHENOP||5.009003| newWHILEOP||5.013007| newXS_deffile||| newXS_len_flags||| newXSproto||5.006000| newXS||5.006000| new_collate||| new_constant||| new_ctype||| new_he||| new_logop||| new_msg_hv||| new_numeric||| new_regcurly|||n new_stackinfo||5.005000| new_version||5.009000| next_symbol||| nextargv||| nextchar||| ninstr|||n no_bareword_allowed||| no_fh_allowed||| no_op||| noperl_die|||vn not_a_number||| not_incrementable||| nothreadhook||5.008000| notify_parser_that_changed_to_utf8||| nuke_stacks||| num_overflow|||n oopsAV||| oopsHV||| op_append_elem||5.013006| op_append_list||5.013006| op_class||| op_clear||| op_contextualize||5.013006| op_convert_list||5.021006| op_dump||5.006000| op_free||| op_integerize||| op_linklist||5.013006| op_lvalue_flags||| op_null||5.007002| op_parent|||n op_prepend_elem||5.013006| op_refcnt_lock||5.009002| op_refcnt_unlock||5.009002| op_relocate_sv||| op_sibling_splice||5.021002|n op_std_init||| open_script||| openn_cleanup||| openn_setup||| opmethod_stash||| opslab_force_free||| opslab_free_nopad||| opslab_free||| optimize_optree||| optimize_op||| output_posix_warnings||| pMY_CXT_|5.007003||p pMY_CXT|5.007003||p pTHX_|5.006000||p pTHX|5.006000||p packWARN|5.007003||p pack_cat||5.007003| pack_rec||| package_version||| package||| packlist||5.008001| pad_add_anon||5.008001| pad_add_name_pvn||5.015001| pad_add_name_pvs||5.015001| pad_add_name_pv||5.015001| pad_add_name_sv||5.015001| pad_add_weakref||| pad_alloc_name||| pad_block_start||| pad_check_dup||| pad_compname_type||5.009003| pad_findlex||| pad_findmy_pvn||5.015001| pad_findmy_pvs||5.015001| pad_findmy_pv||5.015001| pad_findmy_sv||5.015001| pad_fixup_inner_anons||| pad_free||| pad_leavemy||| pad_new||5.008001| pad_push||| pad_reset||| pad_setsv||| pad_sv||| pad_swipe||| padlist_dup||| padlist_store||| padname_dup||| padname_free||| padnamelist_dup||| padnamelist_free||| parse_body||| parse_gv_stash_name||| parse_ident||| parse_lparen_question_flags||| parse_unicode_opts||| parse_uniprop_string||| parser_dup||| parser_free_nexttoke_ops||| parser_free||| path_is_searchable|||n peep||| pending_ident||| perl_alloc_using|||n perl_alloc|||n perl_clone_using|||n perl_clone|||n perl_construct|||n perl_destruct||5.007003|n perl_free|||n perl_parse||5.006000|n perl_run|||n pidgone||| pm_description||| pmop_dump||5.006000| pmruntime||| pmtrans||| pop_scope||| populate_ANYOF_from_invlist||| populate_isa|||v pregcomp||5.009005| pregexec||| pregfree2||5.011000| pregfree||| prescan_version||5.011004| print_bytes_for_locale||| print_collxfrm_input_and_return||| printbuf||| printf_nocontext|||vn process_special_blocks||| ptr_hash|||n ptr_table_fetch||5.009005| ptr_table_find|||n ptr_table_free||5.009005| ptr_table_new||5.009005| ptr_table_split||5.009005| ptr_table_store||5.009005| push_scope||| put_charclass_bitmap_innards_common||| put_charclass_bitmap_innards_invlist||| put_charclass_bitmap_innards||| put_code_point||| put_range||| pv_display|5.006000||p pv_escape|5.009004||p pv_pretty|5.009004||p pv_uni_display||5.007003| qerror||| quadmath_format_needed|||n quadmath_format_single|||n re_compile||5.009005| re_croak2||| re_dup_guts||| re_exec_indentf|||v re_indentf|||v re_intuit_start||5.019001| re_intuit_string||5.006000| re_op_compile||| re_printf|||v realloc||5.007002|n reentrant_free||5.024000| reentrant_init||5.024000| reentrant_retry||5.024000|vn reentrant_size||5.024000| ref_array_or_hash||| refcounted_he_chain_2hv||| refcounted_he_fetch_pvn||| refcounted_he_fetch_pvs||| refcounted_he_fetch_pv||| refcounted_he_fetch_sv||| refcounted_he_free||| refcounted_he_inc||| refcounted_he_new_pvn||| refcounted_he_new_pvs||| refcounted_he_new_pv||| refcounted_he_new_sv||| refcounted_he_value||| refkids||| refto||| ref||5.024000| reg2Lanode||| reg_check_named_buff_matched|||n reg_named_buff_all||5.009005| reg_named_buff_exists||5.009005| reg_named_buff_fetch||5.009005| reg_named_buff_firstkey||5.009005| reg_named_buff_iter||| reg_named_buff_nextkey||5.009005| reg_named_buff_scalar||5.009005| reg_named_buff||| reg_node||| reg_numbered_buff_fetch||| reg_numbered_buff_length||| reg_numbered_buff_store||| reg_qr_package||| reg_scan_name||| reg_skipcomment|||n reg_temp_copy||| reganode||| regatom||| regbranch||| regclass||| regcp_restore||| regcppop||| regcppush||| regcurly|||n regdump_extflags||| regdump_intflags||| regdump||5.005000| regdupe_internal||| regex_set_precedence|||n regexec_flags||5.005000| regfree_internal||5.009005| reghop3|||n reghop4|||n reghopmaybe3|||n reginclass||| reginitcolors||5.006000| reginsert||| regmatch||| regnext||5.005000| regnode_guts||| regpiece||| regprop||| regrepeat||| regtail_study||| regtail||| regtry||| reg||| repeatcpy|||n report_evil_fh||| report_redefined_cv||| report_uninit||| report_wrongway_fh||| require_pv||5.006000| require_tie_mod||| restore_magic||| restore_switched_locale||| rninstr|||n rpeep||| rsignal_restore||| rsignal_save||| rsignal_state||5.004000| rsignal||5.004000| run_body||| run_user_filter||| runops_debug||5.005000| runops_standard||5.005000| rv2cv_op_cv||5.013006| rvpv_dup||| rxres_free||| rxres_restore||| rxres_save||| safesyscalloc||5.006000|n safesysfree||5.006000|n safesysmalloc||5.006000|n safesysrealloc||5.006000|n same_dirent||| save_I16||5.004000| save_I32||| save_I8||5.006000| save_adelete||5.011000| save_aelem_flags||5.011000| save_aelem||5.004050| save_alloc||5.006000| save_aptr||| save_ary||| save_bool||5.008001| save_clearsv||| save_delete||| save_destructor_x||5.006000| save_destructor||5.006000| save_freeop||| save_freepv||| save_freesv||| save_generic_pvref||5.006001| save_generic_svref||5.005030| save_gp||5.004000| save_hash||| save_hdelete||5.011000| save_hek_flags|||n save_helem_flags||5.011000| save_helem||5.004050| save_hints||5.010001| save_hptr||| save_int||| save_item||| save_iv||5.005000| save_lines||| save_list||| save_long||| save_magic_flags||| save_mortalizesv||5.007001| save_nogv||| save_op||5.005000| save_padsv_and_mortalize||5.010001| save_pptr||| save_pushi32ptr||5.010001| save_pushptri32ptr||| save_pushptrptr||5.010001| save_pushptr||5.010001| save_re_context||5.006000| save_scalar_at||| save_scalar||| save_set_svflags||5.009000| save_shared_pvref||5.007003| save_sptr||| save_strlen||| save_svref||| save_to_buffer|||n save_vptr||5.006000| savepvn||| savepvs||5.009003| savepv||| savesharedpvn||5.009005| savesharedpvs||5.013006| savesharedpv||5.007003| savesharedsvpv||5.013006| savestack_grow_cnt||5.008001| savestack_grow||| savesvpv||5.009002| sawparens||| scalar_mod_type|||n scalarboolean||| scalarkids||| scalarseq||| scalarvoid||| scalar||| scan_bin||5.006000| scan_commit||| scan_const||| scan_formline||| scan_heredoc||| scan_hex||| scan_ident||| scan_inputsymbol||| scan_num||5.007001| scan_oct||| scan_pat||| scan_subst||| scan_trans||| scan_version||5.009001| scan_vstring||5.009005| search_const||| seed||5.008001| sequence_num||| set_ANYOF_arg||| set_caret_X||| set_context||5.006000|n set_numeric_radix||5.006000| set_numeric_standard||5.006000| set_numeric_underlying||| set_padlist|||n set_regex_pv||| setdefout||| setfd_cloexec_for_nonsysfd||| setfd_cloexec_or_inhexec_by_sysfdness||| setfd_cloexec|||n setfd_inhexec_for_sysfd||| setfd_inhexec|||n setlocale_debug_string|||n share_hek_flags||| share_hek||5.004000| should_warn_nl|||n si_dup||| sighandler|||n simplify_sort||| skip_to_be_ignored_text||| softref2xv||| sortcv_stacked||| sortcv_xsub||| sortcv||| sortsv_flags||5.009003| sortsv||5.007003| space_join_names_mortal||| ss_dup||| ssc_add_range||| ssc_and||| ssc_anything||| ssc_clear_locale|||n ssc_cp_and||| ssc_finalize||| ssc_init||| ssc_intersection||| ssc_is_anything|||n ssc_is_cp_posixl_init|||n ssc_or||| ssc_union||| stack_grow||| start_subparse||5.004000| stdize_locale||| strEQ||| strGE||| strGT||| strLE||| strLT||| strNE||| str_to_version||5.006000| strip_return||| strnEQ||| strnNE||| study_chunk||| sub_crush_depth||| sublex_done||| sublex_push||| sublex_start||| sv_2bool_flags||5.013006| sv_2bool||| sv_2cv||| sv_2io||| sv_2iuv_common||| sv_2iuv_non_preserve||| sv_2iv_flags||5.009001| sv_2iv||| sv_2mortal||| sv_2nv_flags||5.013001| sv_2pv_flags|5.007002||p sv_2pv_nolen|5.006000||p sv_2pvbyte_nolen|5.006000||p sv_2pvbyte|5.006000||p sv_2pvutf8_nolen||5.006000| sv_2pvutf8||5.006000| sv_2pv||| sv_2uv_flags||5.009001| sv_2uv|5.004000||p sv_add_arena||| sv_add_backref||| sv_backoff|||n sv_bless||| sv_buf_to_ro||| sv_buf_to_rw||| sv_cat_decode||5.008001| sv_catpv_flags||5.013006| sv_catpv_mg|5.004050||p sv_catpv_nomg||5.013006| sv_catpvf_mg_nocontext|||pvn sv_catpvf_mg|5.006000|5.004000|pv sv_catpvf_nocontext|||vn sv_catpvf||5.004000|v sv_catpvn_flags||5.007002| sv_catpvn_mg|5.004050||p sv_catpvn_nomg|5.007002||p sv_catpvn||| sv_catpvs_flags||5.013006| sv_catpvs_mg||5.013006| sv_catpvs_nomg||5.013006| sv_catpvs|5.009003||p sv_catpv||| sv_catsv_flags||5.007002| sv_catsv_mg|5.004050||p sv_catsv_nomg|5.007002||p sv_catsv||| sv_chop||| sv_clean_all||| sv_clean_objs||| sv_clear||| sv_cmp_flags||5.013006| sv_cmp_locale_flags||5.013006| sv_cmp_locale||5.004000| sv_cmp||| sv_collxfrm_flags||5.013006| sv_collxfrm||| sv_copypv_flags||5.017002| sv_copypv_nomg||5.017002| sv_copypv||| sv_dec_nomg||5.013002| sv_dec||| sv_del_backref||| sv_derived_from_pvn||5.015004| sv_derived_from_pv||5.015004| sv_derived_from_sv||5.015004| sv_derived_from||5.004000| sv_destroyable||5.010000| sv_display||| sv_does_pvn||5.015004| sv_does_pv||5.015004| sv_does_sv||5.015004| sv_does||5.009004| sv_dump||| sv_dup_common||| sv_dup_inc_multiple||| sv_dup_inc||| sv_dup||| sv_eq_flags||5.013006| sv_eq||| sv_exp_grow||| sv_force_normal_flags||5.007001| sv_force_normal||5.006000| sv_free_arenas||| sv_free||| sv_gets||5.003070| sv_grow||| sv_i_ncmp||| sv_inc_nomg||5.013002| sv_inc||| sv_insert_flags||5.010001| sv_insert||| sv_isa||| sv_isobject||| sv_iv||5.005000| sv_len_utf8_nomg||| sv_len_utf8||5.006000| sv_len||| sv_magic_portable|5.024000|5.004000|p sv_magicext_mglob||| sv_magicext||5.007003| sv_magic||| sv_mortalcopy_flags||| sv_mortalcopy||| sv_ncmp||| sv_newmortal||| sv_newref||| sv_nolocking||5.007003| sv_nosharing||5.007003| sv_nounlocking||| sv_nv||5.005000| sv_only_taint_gmagic|||n sv_or_pv_pos_u2b||| sv_peek||5.005000| sv_pos_b2u_flags||5.019003| sv_pos_b2u_midway||| sv_pos_b2u||5.006000| sv_pos_u2b_cached||| sv_pos_u2b_flags||5.011005| sv_pos_u2b_forwards|||n sv_pos_u2b_midway|||n sv_pos_u2b||5.006000| sv_pvbyten_force||5.006000| sv_pvbyten||5.006000| sv_pvbyte||5.006000| sv_pvn_force_flags|5.007002||p sv_pvn_force||| sv_pvn_nomg|5.007003|5.005000|p sv_pvn||5.005000| sv_pvutf8n_force||5.006000| sv_pvutf8n||5.006000| sv_pvutf8||5.006000| sv_pv||5.006000| sv_recode_to_utf8||5.007003| sv_reftype||| sv_ref||5.015004| sv_replace||| sv_report_used||| sv_resetpvn||| sv_reset||| sv_rvunweaken||| sv_rvweaken||5.006000| sv_set_undef||| sv_sethek||| sv_setiv_mg|5.004050||p sv_setiv||| sv_setnv_mg|5.006000||p sv_setnv||| sv_setpv_bufsize||| sv_setpv_mg|5.004050||p sv_setpvf_mg_nocontext|||pvn sv_setpvf_mg|5.006000|5.004000|pv sv_setpvf_nocontext|||vn sv_setpvf||5.004000|v sv_setpviv_mg||5.008001| sv_setpviv||5.008001| sv_setpvn_mg|5.004050||p sv_setpvn||| sv_setpvs_mg||5.013006| sv_setpvs|5.009004||p sv_setpv||| sv_setref_iv||| sv_setref_nv||| sv_setref_pvn||| sv_setref_pvs||5.024000| sv_setref_pv||| sv_setref_uv||5.007001| sv_setsv_flags||5.007002| sv_setsv_mg|5.004050||p sv_setsv_nomg|5.007002||p sv_setsv||| sv_setuv_mg|5.004050||p sv_setuv|5.004000||p sv_string_from_errnum||| sv_tainted||5.004000| sv_taint||5.004000| sv_true||5.005000| sv_unglob||| sv_uni_display||5.007003| sv_unmagicext|5.013008||p sv_unmagic||| sv_unref_flags||5.007001| sv_unref||| sv_untaint||5.004000| sv_upgrade||| sv_usepvn_flags||5.009004| sv_usepvn_mg|5.004050||p sv_usepvn||| sv_utf8_decode||| sv_utf8_downgrade||| sv_utf8_encode||5.006000| sv_utf8_upgrade_flags_grow||5.011000| sv_utf8_upgrade_flags||5.007002| sv_utf8_upgrade_nomg||5.007002| sv_utf8_upgrade||5.007001| sv_uv|5.005000||p sv_vcatpvf_mg|5.006000|5.004000|p sv_vcatpvfn_flags||5.017002| sv_vcatpvfn||5.004000| sv_vcatpvf|5.006000|5.004000|p sv_vsetpvf_mg|5.006000|5.004000|p sv_vsetpvfn||5.004000| sv_vsetpvf|5.006000|5.004000|p svtype||| swallow_bom||| swatch_get||| switch_category_locale_to_template||| switch_to_global_locale|||n sync_locale||5.021004|n sys_init3||5.010000|n sys_init||5.010000|n sys_intern_clear||| sys_intern_dup||| sys_intern_init||| sys_term||5.010000|n taint_env||| taint_proper||| tied_method|||v tmps_grow_p||| toFOLD_utf8_safe||| toFOLD_utf8||5.019001| toFOLD_uvchr||5.023009| toFOLD||5.019001| toLOWER_L1||5.019001| toLOWER_LC||5.004000| toLOWER_utf8_safe||| toLOWER_utf8||5.015007| toLOWER_uvchr||5.023009| toLOWER||| toTITLE_utf8_safe||| toTITLE_utf8||5.015007| toTITLE_uvchr||5.023009| toTITLE||5.019001| toUPPER_utf8_safe||| toUPPER_utf8||5.015007| toUPPER_uvchr||5.023009| toUPPER||| to_byte_substr||| to_lower_latin1|||n to_utf8_substr||| tokenize_use||| tokeq||| tokereport||| too_few_arguments_pv||| too_many_arguments_pv||| translate_substr_offsets|||n traverse_op_tree||| try_amagic_bin||| try_amagic_un||| turkic_fc||| turkic_lc||| turkic_uc||| uiv_2buf|||n unlnk||| unpack_rec||| unpack_str||5.007003| unpackstring||5.008001| unreferenced_to_tmp_stack||| unshare_hek_or_pvn||| unshare_hek||| unsharepvn||5.003070| unwind_handler_stack||| update_debugger_info||| upg_version||5.009005| usage||| utf16_textfilter||| utf16_to_utf8_reversed||5.006001| utf16_to_utf8||5.006001| utf8_distance||5.006000| utf8_hop_back|||n utf8_hop_forward|||n utf8_hop_safe|||n utf8_hop||5.006000|n utf8_length||5.007001| utf8_mg_len_cache_update||| utf8_mg_pos_cache_update||| utf8_to_uvchr_buf|5.015009|5.015009|p utf8_to_uvchr|||p utf8n_to_uvchr_error|||n utf8n_to_uvchr||5.007001|n utf8n_to_uvuni||5.007001| utilize||| uvchr_to_utf8_flags||5.007003| uvchr_to_utf8||5.007001| uvoffuni_to_utf8_flags||5.019004| uvuni_to_utf8_flags||5.007003| uvuni_to_utf8||5.007001| valid_utf8_to_uvchr|||n validate_suid||| variant_under_utf8_count|||n varname||| vcmp||5.009000| vcroak||5.006000| vdeb||5.007003| vform||5.006000| visit||| vivify_defelem||| vivify_ref||| vload_module|5.006000||p vmess|5.006000|5.006000|p vnewSVpvf|5.006000|5.004000|p vnormal||5.009002| vnumify||5.009000| vstringify||5.009000| vverify||5.009003| vwarner||5.006000| vwarn||5.006000| wait4pid||| warn_nocontext|||pvn warn_on_first_deprecated_use||| warn_sv|5.013001||p warner_nocontext|||vn warner|5.006000|5.004000|pv warn|||v was_lvalue_sub||| watch||| whichsig_pvn||5.015004| whichsig_pv||5.015004| whichsig_sv||5.015004| whichsig||| win32_croak_not_implemented|||n win32_setlocale||| with_queued_errors||| wrap_op_checker||5.015008| write_to_stderr||| xs_boot_epilog||| xs_handshake|||vn xs_version_bootcheck||| yyerror_pvn||| yyerror_pv||| yyerror||| yylex||| yyparse||| yyquit||| yyunlex||| yywarn||| ); if (exists $opt{'list-unsupported'}) { my $f; for $f (sort { lc $a cmp lc $b } keys %API) { next unless $API{$f}{todo}; print "$f ", '.'x(40-length($f)), " ", format_version($API{$f}{todo}), "\n"; } exit 0; } # Scan for possible replacement candidates my(%replace, %need, %hints, %warnings, %depends); my $replace = 0; my($hint, $define, $function); sub find_api { my $code = shift; $code =~ s{ / (?: \*[^*]*\*+(?:[^$ccs][^*]*\*+)* / | /[^\r\n]*) | "[^"\\]*(?:\\.[^"\\]*)*" | '[^'\\]*(?:\\.[^'\\]*)*' }{}egsx; grep { exists $API{$_} } $code =~ /(\w+)/mg; } while () { if ($hint) { my $h = $hint->[0] eq 'Hint' ? \%hints : \%warnings; if (m{^\s*\*\s(.*?)\s*$}) { for (@{$hint->[1]}) { $h->{$_} ||= ''; # suppress warning with older perls $h->{$_} .= "$1\n"; } } else { undef $hint } } $hint = [$1, [split /,?\s+/, $2]] if m{^\s*$rccs\s+(Hint|Warning):\s+(\w+(?:,?\s+\w+)*)\s*$}; if ($define) { if ($define->[1] =~ /\\$/) { $define->[1] .= $_; } else { if (exists $API{$define->[0]} && $define->[1] !~ /^DPPP_\(/) { my @n = find_api($define->[1]); push @{$depends{$define->[0]}}, @n if @n } undef $define; } } $define = [$1, $2] if m{^\s*#\s*define\s+(\w+)(?:\([^)]*\))?\s+(.*)}; if ($function) { if (/^}/) { if (exists $API{$function->[0]}) { my @n = find_api($function->[1]); push @{$depends{$function->[0]}}, @n if @n } undef $function; } else { $function->[1] .= $_; } } $function = [$1, ''] if m{^DPPP_\(my_(\w+)\)}; $replace = $1 if m{^\s*$rccs\s+Replace:\s+(\d+)\s+$rcce\s*$}; $replace{$2} = $1 if $replace and m{^\s*#\s*define\s+(\w+)(?:\([^)]*\))?\s+(\w+)}; $replace{$2} = $1 if m{^\s*#\s*define\s+(\w+)(?:\([^)]*\))?\s+(\w+).*$rccs\s+Replace\s+$rcce}; $replace{$1} = $2 if m{^\s*$rccs\s+Replace (\w+) with (\w+)\s+$rcce\s*$}; if (m{^\s*$rccs\s+(\w+(\s*,\s*\w+)*)\s+depends\s+on\s+(\w+(\s*,\s*\w+)*)\s+$rcce\s*$}) { my @deps = map { s/\s+//g; $_ } split /,/, $3; my $d; for $d (map { s/\s+//g; $_ } split /,/, $1) { push @{$depends{$d}}, @deps; } } $need{$1} = 1 if m{^#if\s+defined\(NEED_(\w+)(?:_GLOBAL)?\)}; } for (values %depends) { my %s; $_ = [sort grep !$s{$_}++, @$_]; } if (exists $opt{'api-info'}) { my $f; my $count = 0; my $match = $opt{'api-info'} =~ m!^/(.*)/$! ? $1 : "^\Q$opt{'api-info'}\E\$"; for $f (sort { lc $a cmp lc $b } keys %API) { next unless $f =~ /$match/; print "\n=== $f ===\n\n"; my $info = 0; if ($API{$f}{base} || $API{$f}{todo}) { my $base = format_version($API{$f}{base} || $API{$f}{todo}); print "Supported at least starting from perl-$base.\n"; $info++; } if ($API{$f}{provided}) { my $todo = $API{$f}{todo} ? format_version($API{$f}{todo}) : "5.003"; print "Support by $ppport provided back to perl-$todo.\n"; print "Support needs to be explicitly requested by NEED_$f.\n" if exists $need{$f}; print "Depends on: ", join(', ', @{$depends{$f}}), ".\n" if exists $depends{$f}; print "\n$hints{$f}" if exists $hints{$f}; print "\nWARNING:\n$warnings{$f}" if exists $warnings{$f}; $info++; } print "No portability information available.\n" unless $info; $count++; } $count or print "Found no API matching '$opt{'api-info'}'."; print "\n"; exit 0; } if (exists $opt{'list-provided'}) { my $f; for $f (sort { lc $a cmp lc $b } keys %API) { next unless $API{$f}{provided}; my @flags; push @flags, 'explicit' if exists $need{$f}; push @flags, 'depend' if exists $depends{$f}; push @flags, 'hint' if exists $hints{$f}; push @flags, 'warning' if exists $warnings{$f}; my $flags = @flags ? ' ['.join(', ', @flags).']' : ''; print "$f$flags\n"; } exit 0; } my @files; my @srcext = qw( .xs .c .h .cc .cpp -c.inc -xs.inc ); my $srcext = join '|', map { quotemeta $_ } @srcext; if (@ARGV) { my %seen; for (@ARGV) { if (-e) { if (-f) { push @files, $_ unless $seen{$_}++; } else { warn "'$_' is not a file.\n" } } else { my @new = grep { -f } glob $_ or warn "'$_' does not exist.\n"; push @files, grep { !$seen{$_}++ } @new; } } } else { eval { require File::Find; File::Find::find(sub { $File::Find::name =~ /($srcext)$/i and push @files, $File::Find::name; }, '.'); }; if ($@) { @files = map { glob "*$_" } @srcext; } } if (!@ARGV || $opt{filter}) { my(@in, @out); my %xsc = map { /(.*)\.xs$/ ? ("$1.c" => 1, "$1.cc" => 1) : () } @files; for (@files) { my $out = exists $xsc{$_} || /\b\Q$ppport\E$/i || !/($srcext)$/i; push @{ $out ? \@out : \@in }, $_; } if (@ARGV && @out) { warning("Skipping the following files (use --nofilter to avoid this):\n| ", join "\n| ", @out); } @files = @in; } die "No input files given!\n" unless @files; my(%files, %global, %revreplace); %revreplace = reverse %replace; my $filename; my $patch_opened = 0; for $filename (@files) { unless (open IN, "<$filename") { warn "Unable to read from $filename: $!\n"; next; } info("Scanning $filename ..."); my $c = do { local $/; }; close IN; my %file = (orig => $c, changes => 0); # Temporarily remove C/XS comments and strings from the code my @ccom; $c =~ s{ ( ^$HS*\#$HS*include\b[^\r\n]+\b(?:\Q$ppport\E|XSUB\.h)\b[^\r\n]* | ^$HS*\#$HS*(?:define|elif|if(?:def)?)\b[^\r\n]* ) | ( ^$HS*\#[^\r\n]* | "[^"\\]*(?:\\.[^"\\]*)*" | '[^'\\]*(?:\\.[^'\\]*)*' | / (?: \*[^*]*\*+(?:[^$ccs][^*]*\*+)* / | /[^\r\n]* ) ) }{ defined $2 and push @ccom, $2; defined $1 ? $1 : "$ccs$#ccom$cce" }mgsex; $file{ccom} = \@ccom; $file{code} = $c; $file{has_inc_ppport} = $c =~ /^$HS*#$HS*include[^\r\n]+\b\Q$ppport\E\b/m; my $func; for $func (keys %API) { my $match = $func; $match .= "|$revreplace{$func}" if exists $revreplace{$func}; if ($c =~ /\b(?:Perl_)?($match)\b/) { $file{uses_replace}{$1}++ if exists $revreplace{$func} && $1 eq $revreplace{$func}; $file{uses_Perl}{$func}++ if $c =~ /\bPerl_$func\b/; if (exists $API{$func}{provided}) { $file{uses_provided}{$func}++; if (!exists $API{$func}{base} || $API{$func}{base} > $opt{'compat-version'}) { $file{uses}{$func}++; my @deps = rec_depend($func); if (@deps) { $file{uses_deps}{$func} = \@deps; for (@deps) { $file{uses}{$_} = 0 unless exists $file{uses}{$_}; } } for ($func, @deps) { $file{needs}{$_} = 'static' if exists $need{$_}; } } } if (exists $API{$func}{todo} && $API{$func}{todo} > $opt{'compat-version'}) { if ($c =~ /\b$func\b/) { $file{uses_todo}{$func}++; } } } } while ($c =~ /^$HS*#$HS*define$HS+(NEED_(\w+?)(_GLOBAL)?)\b/mg) { if (exists $need{$2}) { $file{defined $3 ? 'needed_global' : 'needed_static'}{$2}++; } else { warning("Possibly wrong #define $1 in $filename") } } for (qw(uses needs uses_todo needed_global needed_static)) { for $func (keys %{$file{$_}}) { push @{$global{$_}{$func}}, $filename; } } $files{$filename} = \%file; } # Globally resolve NEED_'s my $need; for $need (keys %{$global{needs}}) { if (@{$global{needs}{$need}} > 1) { my @targets = @{$global{needs}{$need}}; my @t = grep $files{$_}{needed_global}{$need}, @targets; @targets = @t if @t; @t = grep /\.xs$/i, @targets; @targets = @t if @t; my $target = shift @targets; $files{$target}{needs}{$need} = 'global'; for (@{$global{needs}{$need}}) { $files{$_}{needs}{$need} = 'extern' if $_ ne $target; } } } for $filename (@files) { exists $files{$filename} or next; info("=== Analyzing $filename ==="); my %file = %{$files{$filename}}; my $func; my $c = $file{code}; my $warnings = 0; for $func (sort keys %{$file{uses_Perl}}) { if ($API{$func}{varargs}) { unless ($API{$func}{nothxarg}) { my $changes = ($c =~ s{\b(Perl_$func\s*\(\s*)(?!aTHX_?)(\)|[^\s)]*\))} { $1 . ($2 eq ')' ? 'aTHX' : 'aTHX_ ') . $2 }ge); if ($changes) { warning("Doesn't pass interpreter argument aTHX to Perl_$func"); $file{changes} += $changes; } } } else { warning("Uses Perl_$func instead of $func"); $file{changes} += ($c =~ s{\bPerl_$func(\s*)\((\s*aTHX_?)?\s*} {$func$1(}g); } } for $func (sort keys %{$file{uses_replace}}) { warning("Uses $func instead of $replace{$func}"); $file{changes} += ($c =~ s/\b$func\b/$replace{$func}/g); } for $func (sort keys %{$file{uses_provided}}) { if ($file{uses}{$func}) { if (exists $file{uses_deps}{$func}) { diag("Uses $func, which depends on ", join(', ', @{$file{uses_deps}{$func}})); } else { diag("Uses $func"); } } $warnings += hint($func); } unless ($opt{quiet}) { for $func (sort keys %{$file{uses_todo}}) { print "*** WARNING: Uses $func, which may not be portable below perl ", format_version($API{$func}{todo}), ", even with '$ppport'\n"; $warnings++; } } for $func (sort keys %{$file{needed_static}}) { my $message = ''; if (not exists $file{uses}{$func}) { $message = "No need to define NEED_$func if $func is never used"; } elsif (exists $file{needs}{$func} && $file{needs}{$func} ne 'static') { $message = "No need to define NEED_$func when already needed globally"; } if ($message) { diag($message); $file{changes} += ($c =~ s/^$HS*#$HS*define$HS+NEED_$func\b.*$LF//mg); } } for $func (sort keys %{$file{needed_global}}) { my $message = ''; if (not exists $global{uses}{$func}) { $message = "No need to define NEED_${func}_GLOBAL if $func is never used"; } elsif (exists $file{needs}{$func}) { if ($file{needs}{$func} eq 'extern') { $message = "No need to define NEED_${func}_GLOBAL when already needed globally"; } elsif ($file{needs}{$func} eq 'static') { $message = "No need to define NEED_${func}_GLOBAL when only used in this file"; } } if ($message) { diag($message); $file{changes} += ($c =~ s/^$HS*#$HS*define$HS+NEED_${func}_GLOBAL\b.*$LF//mg); } } $file{needs_inc_ppport} = keys %{$file{uses}}; if ($file{needs_inc_ppport}) { my $pp = ''; for $func (sort keys %{$file{needs}}) { my $type = $file{needs}{$func}; next if $type eq 'extern'; my $suffix = $type eq 'global' ? '_GLOBAL' : ''; unless (exists $file{"needed_$type"}{$func}) { if ($type eq 'global') { diag("Files [@{$global{needs}{$func}}] need $func, adding global request"); } else { diag("File needs $func, adding static request"); } $pp .= "#define NEED_$func$suffix\n"; } } if ($pp && ($c =~ s/^(?=$HS*#$HS*define$HS+NEED_\w+)/$pp/m)) { $pp = ''; $file{changes}++; } unless ($file{has_inc_ppport}) { diag("Needs to include '$ppport'"); $pp .= qq(#include "$ppport"\n) } if ($pp) { $file{changes} += ($c =~ s/^($HS*#$HS*define$HS+NEED_\w+.*?)^/$1$pp/ms) || ($c =~ s/^(?=$HS*#$HS*include.*\Q$ppport\E)/$pp/m) || ($c =~ s/^($HS*#$HS*include.*XSUB.*\s*?)^/$1$pp/m) || ($c =~ s/^/$pp/); } } else { if ($file{has_inc_ppport}) { diag("No need to include '$ppport'"); $file{changes} += ($c =~ s/^$HS*?#$HS*include.*\Q$ppport\E.*?$LF//m); } } # put back in our C comments my $ix; my $cppc = 0; my @ccom = @{$file{ccom}}; for $ix (0 .. $#ccom) { if (!$opt{cplusplus} && $ccom[$ix] =~ s!^//!!) { $cppc++; $file{changes} += $c =~ s/$rccs$ix$rcce/$ccs$ccom[$ix] $cce/; } else { $c =~ s/$rccs$ix$rcce/$ccom[$ix]/; } } if ($cppc) { my $s = $cppc != 1 ? 's' : ''; warning("Uses $cppc C++ style comment$s, which is not portable"); } my $s = $warnings != 1 ? 's' : ''; my $warn = $warnings ? " ($warnings warning$s)" : ''; info("Analysis completed$warn"); if ($file{changes}) { if (exists $opt{copy}) { my $newfile = "$filename$opt{copy}"; if (-e $newfile) { error("'$newfile' already exists, refusing to write copy of '$filename'"); } else { local *F; if (open F, ">$newfile") { info("Writing copy of '$filename' with changes to '$newfile'"); print F $c; close F; } else { error("Cannot open '$newfile' for writing: $!"); } } } elsif (exists $opt{patch} || $opt{changes}) { if (exists $opt{patch}) { unless ($patch_opened) { if (open PATCH, ">$opt{patch}") { $patch_opened = 1; } else { error("Cannot open '$opt{patch}' for writing: $!"); delete $opt{patch}; $opt{changes} = 1; goto fallback; } } mydiff(\*PATCH, $filename, $c); } else { fallback: info("Suggested changes:"); mydiff(\*STDOUT, $filename, $c); } } else { my $s = $file{changes} == 1 ? '' : 's'; info("$file{changes} potentially required change$s detected"); } } else { info("Looks good"); } } close PATCH if $patch_opened; exit 0; sub try_use { eval "use @_;"; return $@ eq '' } sub mydiff { local *F = shift; my($file, $str) = @_; my $diff; if (exists $opt{diff}) { $diff = run_diff($opt{diff}, $file, $str); } if (!defined $diff and try_use('Text::Diff')) { $diff = Text::Diff::diff($file, \$str, { STYLE => 'Unified' }); $diff = <
$tmp") { print F $str; close F; if (open F, "$prog $file $tmp |") { while () { s/\Q$tmp\E/$file.patched/; $diff .= $_; } close F; unlink $tmp; return $diff; } unlink $tmp; } else { error("Cannot open '$tmp' for writing: $!"); } return undef; } sub rec_depend { my($func, $seen) = @_; return () unless exists $depends{$func}; $seen = {%{$seen||{}}}; return () if $seen->{$func}++; my %s; grep !$s{$_}++, map { ($_, rec_depend($_, $seen)) } @{$depends{$func}}; } sub parse_version { my $ver = shift; if ($ver =~ /^(\d+)\.(\d+)\.(\d+)$/) { return ($1, $2, $3); } elsif ($ver !~ /^\d+\.[\d_]+$/) { die "cannot parse version '$ver'\n"; } $ver =~ s/_//g; $ver =~ s/$/000000/; my($r,$v,$s) = $ver =~ /(\d+)\.(\d{3})(\d{3})/; $v = int $v; $s = int $s; if ($r < 5 || ($r == 5 && $v < 6)) { if ($s % 10) { die "cannot parse version '$ver'\n"; } } return ($r, $v, $s); } sub format_version { my $ver = shift; $ver =~ s/$/000000/; my($r,$v,$s) = $ver =~ /(\d+)\.(\d{3})(\d{3})/; $v = int $v; $s = int $s; if ($r < 5 || ($r == 5 && $v < 6)) { if ($s % 10) { die "invalid version '$ver'\n"; } $s /= 10; $ver = sprintf "%d.%03d", $r, $v; $s > 0 and $ver .= sprintf "_%02d", $s; return $ver; } return sprintf "%d.%d.%d", $r, $v, $s; } sub info { $opt{quiet} and return; print @_, "\n"; } sub diag { $opt{quiet} and return; $opt{diag} and print @_, "\n"; } sub warning { $opt{quiet} and return; print "*** ", @_, "\n"; } sub error { print "*** ERROR: ", @_, "\n"; } my %given_hints; my %given_warnings; sub hint { $opt{quiet} and return; my $func = shift; my $rv = 0; if (exists $warnings{$func} && !$given_warnings{$func}++) { my $warn = $warnings{$func}; $warn =~ s!^!*** !mg; print "*** WARNING: $func\n", $warn; $rv++; } if ($opt{hints} && exists $hints{$func} && !$given_hints{$func}++) { my $hint = $hints{$func}; $hint =~ s/^/ /mg; print " --- hint for $func ---\n", $hint; } $rv; } sub usage { my($usage) = do { local(@ARGV,$/)=($0); <> } =~ /^=head\d$HS+SYNOPSIS\s*^(.*?)\s*^=/ms; my %M = ( 'I' => '*' ); $usage =~ s/^\s*perl\s+\S+/$^X $0/; $usage =~ s/([A-Z])<([^>]+)>/$M{$1}$2$M{$1}/g; print < }; my($copy) = $self =~ /^=head\d\s+COPYRIGHT\s*^(.*?)^=\w+/ms; $copy =~ s/^(?=\S+)/ /gms; $self =~ s/^$HS+Do NOT edit.*?(?=^-)/$copy/ms; $self =~ s/^SKIP.*(?=^__DATA__)/SKIP if (\@ARGV && \$ARGV[0] eq '--unstrip') { eval { require Devel::PPPort }; \$@ and die "Cannot require Devel::PPPort, please install.\\n"; if (eval \$Devel::PPPort::VERSION < $VERSION) { die "$0 was originally generated with Devel::PPPort $VERSION.\\n" . "Your Devel::PPPort is only version \$Devel::PPPort::VERSION.\\n" . "Please install a newer version, or --unstrip will not work.\\n"; } Devel::PPPort::WriteFile(\$0); exit 0; } print <$0" or die "cannot strip $0: $!\n"; print OUT "$pl$c\n"; exit 0; } __DATA__ */ #ifndef _P_P_PORTABILITY_H_ #define _P_P_PORTABILITY_H_ #ifndef DPPP_NAMESPACE # define DPPP_NAMESPACE DPPP_ #endif #define DPPP_CAT2(x,y) CAT2(x,y) #define DPPP_(name) DPPP_CAT2(DPPP_NAMESPACE, name) #ifndef PERL_REVISION # if !defined(__PATCHLEVEL_H_INCLUDED__) && !(defined(PATCHLEVEL) && defined(SUBVERSION)) # define PERL_PATCHLEVEL_H_IMPLICIT # include # endif # if !(defined(PERL_VERSION) || (defined(SUBVERSION) && defined(PATCHLEVEL))) # include # endif # ifndef PERL_REVISION # define PERL_REVISION (5) /* Replace: 1 */ # define PERL_VERSION PATCHLEVEL # define PERL_SUBVERSION SUBVERSION /* Replace PERL_PATCHLEVEL with PERL_VERSION */ /* Replace: 0 */ # endif #endif #define D_PPP_DEC2BCD(dec) ((((dec)/100)<<8)|((((dec)%100)/10)<<4)|((dec)%10)) #define PERL_BCDVERSION ((D_PPP_DEC2BCD(PERL_REVISION)<<24)|(D_PPP_DEC2BCD(PERL_VERSION)<<12)|D_PPP_DEC2BCD(PERL_SUBVERSION)) /* It is very unlikely that anyone will try to use this with Perl 6 (or greater), but who knows. */ #if PERL_REVISION != 5 # error ppport.h only works with Perl version 5 #endif /* PERL_REVISION != 5 */ #ifndef dTHR # define dTHR dNOOP #endif #ifndef dTHX # define dTHX dNOOP #endif #ifndef dTHXa # define dTHXa(x) dNOOP #endif #ifndef pTHX # define pTHX void #endif #ifndef pTHX_ # define pTHX_ #endif #ifndef aTHX # define aTHX #endif #ifndef aTHX_ # define aTHX_ #endif #if (PERL_BCDVERSION < 0x5006000) # ifdef USE_THREADS # define aTHXR thr # define aTHXR_ thr, # else # define aTHXR # define aTHXR_ # endif # define dTHXR dTHR #else # define aTHXR aTHX # define aTHXR_ aTHX_ # define dTHXR dTHX #endif #ifndef dTHXoa # define dTHXoa(x) dTHXa(x) #endif #ifdef I_LIMITS # include #endif #ifndef PERL_UCHAR_MIN # define PERL_UCHAR_MIN ((unsigned char)0) #endif #ifndef PERL_UCHAR_MAX # ifdef UCHAR_MAX # define PERL_UCHAR_MAX ((unsigned char)UCHAR_MAX) # else # ifdef MAXUCHAR # define PERL_UCHAR_MAX ((unsigned char)MAXUCHAR) # else # define PERL_UCHAR_MAX ((unsigned char)~(unsigned)0) # endif # endif #endif #ifndef PERL_USHORT_MIN # define PERL_USHORT_MIN ((unsigned short)0) #endif #ifndef PERL_USHORT_MAX # ifdef USHORT_MAX # define PERL_USHORT_MAX ((unsigned short)USHORT_MAX) # else # ifdef MAXUSHORT # define PERL_USHORT_MAX ((unsigned short)MAXUSHORT) # else # ifdef USHRT_MAX # define PERL_USHORT_MAX ((unsigned short)USHRT_MAX) # else # define PERL_USHORT_MAX ((unsigned short)~(unsigned)0) # endif # endif # endif #endif #ifndef PERL_SHORT_MAX # ifdef SHORT_MAX # define PERL_SHORT_MAX ((short)SHORT_MAX) # else # ifdef MAXSHORT /* Often used in */ # define PERL_SHORT_MAX ((short)MAXSHORT) # else # ifdef SHRT_MAX # define PERL_SHORT_MAX ((short)SHRT_MAX) # else # define PERL_SHORT_MAX ((short) (PERL_USHORT_MAX >> 1)) # endif # endif # endif #endif #ifndef PERL_SHORT_MIN # ifdef SHORT_MIN # define PERL_SHORT_MIN ((short)SHORT_MIN) # else # ifdef MINSHORT # define PERL_SHORT_MIN ((short)MINSHORT) # else # ifdef SHRT_MIN # define PERL_SHORT_MIN ((short)SHRT_MIN) # else # define PERL_SHORT_MIN (-PERL_SHORT_MAX - ((3 & -1) == 3)) # endif # endif # endif #endif #ifndef PERL_UINT_MAX # ifdef UINT_MAX # define PERL_UINT_MAX ((unsigned int)UINT_MAX) # else # ifdef MAXUINT # define PERL_UINT_MAX ((unsigned int)MAXUINT) # else # define PERL_UINT_MAX (~(unsigned int)0) # endif # endif #endif #ifndef PERL_UINT_MIN # define PERL_UINT_MIN ((unsigned int)0) #endif #ifndef PERL_INT_MAX # ifdef INT_MAX # define PERL_INT_MAX ((int)INT_MAX) # else # ifdef MAXINT /* Often used in */ # define PERL_INT_MAX ((int)MAXINT) # else # define PERL_INT_MAX ((int)(PERL_UINT_MAX >> 1)) # endif # endif #endif #ifndef PERL_INT_MIN # ifdef INT_MIN # define PERL_INT_MIN ((int)INT_MIN) # else # ifdef MININT # define PERL_INT_MIN ((int)MININT) # else # define PERL_INT_MIN (-PERL_INT_MAX - ((3 & -1) == 3)) # endif # endif #endif #ifndef PERL_ULONG_MAX # ifdef ULONG_MAX # define PERL_ULONG_MAX ((unsigned long)ULONG_MAX) # else # ifdef MAXULONG # define PERL_ULONG_MAX ((unsigned long)MAXULONG) # else # define PERL_ULONG_MAX (~(unsigned long)0) # endif # endif #endif #ifndef PERL_ULONG_MIN # define PERL_ULONG_MIN ((unsigned long)0L) #endif #ifndef PERL_LONG_MAX # ifdef LONG_MAX # define PERL_LONG_MAX ((long)LONG_MAX) # else # ifdef MAXLONG # define PERL_LONG_MAX ((long)MAXLONG) # else # define PERL_LONG_MAX ((long) (PERL_ULONG_MAX >> 1)) # endif # endif #endif #ifndef PERL_LONG_MIN # ifdef LONG_MIN # define PERL_LONG_MIN ((long)LONG_MIN) # else # ifdef MINLONG # define PERL_LONG_MIN ((long)MINLONG) # else # define PERL_LONG_MIN (-PERL_LONG_MAX - ((3 & -1) == 3)) # endif # endif #endif #if defined(HAS_QUAD) && (defined(convex) || defined(uts)) # ifndef PERL_UQUAD_MAX # ifdef ULONGLONG_MAX # define PERL_UQUAD_MAX ((unsigned long long)ULONGLONG_MAX) # else # ifdef MAXULONGLONG # define PERL_UQUAD_MAX ((unsigned long long)MAXULONGLONG) # else # define PERL_UQUAD_MAX (~(unsigned long long)0) # endif # endif # endif # ifndef PERL_UQUAD_MIN # define PERL_UQUAD_MIN ((unsigned long long)0L) # endif # ifndef PERL_QUAD_MAX # ifdef LONGLONG_MAX # define PERL_QUAD_MAX ((long long)LONGLONG_MAX) # else # ifdef MAXLONGLONG # define PERL_QUAD_MAX ((long long)MAXLONGLONG) # else # define PERL_QUAD_MAX ((long long) (PERL_UQUAD_MAX >> 1)) # endif # endif # endif # ifndef PERL_QUAD_MIN # ifdef LONGLONG_MIN # define PERL_QUAD_MIN ((long long)LONGLONG_MIN) # else # ifdef MINLONGLONG # define PERL_QUAD_MIN ((long long)MINLONGLONG) # else # define PERL_QUAD_MIN (-PERL_QUAD_MAX - ((3 & -1) == 3)) # endif # endif # endif #endif /* This is based on code from 5.003 perl.h */ #ifdef HAS_QUAD # ifdef cray #ifndef IVTYPE # define IVTYPE int #endif #ifndef IV_MIN # define IV_MIN PERL_INT_MIN #endif #ifndef IV_MAX # define IV_MAX PERL_INT_MAX #endif #ifndef UV_MIN # define UV_MIN PERL_UINT_MIN #endif #ifndef UV_MAX # define UV_MAX PERL_UINT_MAX #endif # ifdef INTSIZE #ifndef IVSIZE # define IVSIZE INTSIZE #endif # endif # else # if defined(convex) || defined(uts) #ifndef IVTYPE # define IVTYPE long long #endif #ifndef IV_MIN # define IV_MIN PERL_QUAD_MIN #endif #ifndef IV_MAX # define IV_MAX PERL_QUAD_MAX #endif #ifndef UV_MIN # define UV_MIN PERL_UQUAD_MIN #endif #ifndef UV_MAX # define UV_MAX PERL_UQUAD_MAX #endif # ifdef LONGLONGSIZE #ifndef IVSIZE # define IVSIZE LONGLONGSIZE #endif # endif # else #ifndef IVTYPE # define IVTYPE long #endif #ifndef IV_MIN # define IV_MIN PERL_LONG_MIN #endif #ifndef IV_MAX # define IV_MAX PERL_LONG_MAX #endif #ifndef UV_MIN # define UV_MIN PERL_ULONG_MIN #endif #ifndef UV_MAX # define UV_MAX PERL_ULONG_MAX #endif # ifdef LONGSIZE #ifndef IVSIZE # define IVSIZE LONGSIZE #endif # endif # endif # endif #ifndef IVSIZE # define IVSIZE 8 #endif #ifndef LONGSIZE # define LONGSIZE 8 #endif #ifndef PERL_QUAD_MIN # define PERL_QUAD_MIN IV_MIN #endif #ifndef PERL_QUAD_MAX # define PERL_QUAD_MAX IV_MAX #endif #ifndef PERL_UQUAD_MIN # define PERL_UQUAD_MIN UV_MIN #endif #ifndef PERL_UQUAD_MAX # define PERL_UQUAD_MAX UV_MAX #endif #else #ifndef IVTYPE # define IVTYPE long #endif #ifndef LONGSIZE # define LONGSIZE 4 #endif #ifndef IV_MIN # define IV_MIN PERL_LONG_MIN #endif #ifndef IV_MAX # define IV_MAX PERL_LONG_MAX #endif #ifndef UV_MIN # define UV_MIN PERL_ULONG_MIN #endif #ifndef UV_MAX # define UV_MAX PERL_ULONG_MAX #endif #endif #ifndef IVSIZE # ifdef LONGSIZE # define IVSIZE LONGSIZE # else # define IVSIZE 4 /* A bold guess, but the best we can make. */ # endif #endif #ifndef UVTYPE # define UVTYPE unsigned IVTYPE #endif #ifndef UVSIZE # define UVSIZE IVSIZE #endif #ifndef cBOOL # define cBOOL(cbool) ((cbool) ? (bool)1 : (bool)0) #endif #ifndef OpHAS_SIBLING # define OpHAS_SIBLING(o) (cBOOL((o)->op_sibling)) #endif #ifndef OpSIBLING # define OpSIBLING(o) (0 + (o)->op_sibling) #endif #ifndef OpMORESIB_set # define OpMORESIB_set(o, sib) ((o)->op_sibling = (sib)) #endif #ifndef OpLASTSIB_set # define OpLASTSIB_set(o, parent) ((o)->op_sibling = NULL) #endif #ifndef OpMAYBESIB_set # define OpMAYBESIB_set(o, sib, parent) ((o)->op_sibling = (sib)) #endif #ifndef HEf_SVKEY # define HEf_SVKEY -2 #endif #if defined(DEBUGGING) && !defined(__COVERITY__) #ifndef __ASSERT_ # define __ASSERT_(statement) assert(statement), #endif #else #ifndef __ASSERT_ # define __ASSERT_(statement) #endif #endif #ifndef SvRX #if defined(NEED_SvRX) static void * DPPP_(my_SvRX)(pTHX_ SV *rv); static #else extern void * DPPP_(my_SvRX)(pTHX_ SV *rv); #endif #if defined(NEED_SvRX) || defined(NEED_SvRX_GLOBAL) #ifdef SvRX # undef SvRX #endif #define SvRX(a) DPPP_(my_SvRX)(aTHX_ a) void * DPPP_(my_SvRX)(pTHX_ SV *rv) { if (SvROK(rv)) { SV *sv = SvRV(rv); if (SvMAGICAL(sv)) { MAGIC *mg = mg_find(sv, PERL_MAGIC_qr); if (mg && mg->mg_obj) { return mg->mg_obj; } } } return 0; } #endif #endif #ifndef SvRXOK # define SvRXOK(sv) (!!SvRX(sv)) #endif #ifndef PERL_UNUSED_DECL # ifdef HASATTRIBUTE # if (defined(__GNUC__) && defined(__cplusplus)) || defined(__INTEL_COMPILER) # define PERL_UNUSED_DECL # else # define PERL_UNUSED_DECL __attribute__((unused)) # endif # else # define PERL_UNUSED_DECL # endif #endif #ifndef PERL_UNUSED_ARG # if defined(lint) && defined(S_SPLINT_S) /* www.splint.org */ # include # define PERL_UNUSED_ARG(x) NOTE(ARGUNUSED(x)) # else # define PERL_UNUSED_ARG(x) ((void)x) # endif #endif #ifndef PERL_UNUSED_VAR # define PERL_UNUSED_VAR(x) ((void)x) #endif #ifndef PERL_UNUSED_CONTEXT # ifdef USE_ITHREADS # define PERL_UNUSED_CONTEXT PERL_UNUSED_ARG(my_perl) # else # define PERL_UNUSED_CONTEXT # endif #endif #ifndef PERL_UNUSED_RESULT # if defined(__GNUC__) && defined(HASATTRIBUTE_WARN_UNUSED_RESULT) # define PERL_UNUSED_RESULT(v) STMT_START { __typeof__(v) z = (v); (void)sizeof(z); } STMT_END # else # define PERL_UNUSED_RESULT(v) ((void)(v)) # endif #endif #ifndef NOOP # define NOOP /*EMPTY*/(void)0 #endif #ifndef dNOOP # define dNOOP extern int /*@unused@*/ Perl___notused PERL_UNUSED_DECL #endif #ifndef NVTYPE # if defined(USE_LONG_DOUBLE) && defined(HAS_LONG_DOUBLE) # define NVTYPE long double # else # define NVTYPE double # endif typedef NVTYPE NV; #endif #ifndef INT2PTR # if (IVSIZE == PTRSIZE) && (UVSIZE == PTRSIZE) # define PTRV UV # define INT2PTR(any,d) (any)(d) # else # if PTRSIZE == LONGSIZE # define PTRV unsigned long # else # define PTRV unsigned # endif # define INT2PTR(any,d) (any)(PTRV)(d) # endif #endif #ifndef PTR2ul # if PTRSIZE == LONGSIZE # define PTR2ul(p) (unsigned long)(p) # else # define PTR2ul(p) INT2PTR(unsigned long,p) # endif #endif #ifndef PTR2nat # define PTR2nat(p) (PTRV)(p) #endif #ifndef NUM2PTR # define NUM2PTR(any,d) (any)PTR2nat(d) #endif #ifndef PTR2IV # define PTR2IV(p) INT2PTR(IV,p) #endif #ifndef PTR2UV # define PTR2UV(p) INT2PTR(UV,p) #endif #ifndef PTR2NV # define PTR2NV(p) NUM2PTR(NV,p) #endif #undef START_EXTERN_C #undef END_EXTERN_C #undef EXTERN_C #ifdef __cplusplus # define START_EXTERN_C extern "C" { # define END_EXTERN_C } # define EXTERN_C extern "C" #else # define START_EXTERN_C # define END_EXTERN_C # define EXTERN_C extern #endif #if defined(PERL_GCC_PEDANTIC) # ifndef PERL_GCC_BRACE_GROUPS_FORBIDDEN # define PERL_GCC_BRACE_GROUPS_FORBIDDEN # endif #endif #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN) && !defined(__cplusplus) # ifndef PERL_USE_GCC_BRACE_GROUPS # define PERL_USE_GCC_BRACE_GROUPS # endif #endif #undef STMT_START #undef STMT_END #ifdef PERL_USE_GCC_BRACE_GROUPS # define STMT_START (void)( /* gcc supports ``({ STATEMENTS; })'' */ # define STMT_END ) #else # if defined(VOIDFLAGS) && (VOIDFLAGS) && (defined(sun) || defined(__sun__)) && !defined(__GNUC__) # define STMT_START if (1) # define STMT_END else (void)0 # else # define STMT_START do # define STMT_END while (0) # endif #endif #ifndef boolSV # define boolSV(b) ((b) ? &PL_sv_yes : &PL_sv_no) #endif /* DEFSV appears first in 5.004_56 */ #ifndef DEFSV # define DEFSV GvSV(PL_defgv) #endif #ifndef SAVE_DEFSV # define SAVE_DEFSV SAVESPTR(GvSV(PL_defgv)) #endif #ifndef DEFSV_set # define DEFSV_set(sv) (DEFSV = (sv)) #endif /* Older perls (<=5.003) lack AvFILLp */ #ifndef AvFILLp # define AvFILLp AvFILL #endif #ifndef av_tindex # define av_tindex AvFILL #endif #ifndef av_top_index # define av_top_index AvFILL #endif #ifndef ERRSV # define ERRSV get_sv("@",FALSE) #endif /* Hint: gv_stashpvn * This function's backport doesn't support the length parameter, but * rather ignores it. Portability can only be ensured if the length * parameter is used for speed reasons, but the length can always be * correctly computed from the string argument. */ #ifndef gv_stashpvn # define gv_stashpvn(str,len,create) gv_stashpv(str,create) #endif /* Replace: 1 */ #ifndef get_cv # define get_cv perl_get_cv #endif #ifndef get_sv # define get_sv perl_get_sv #endif #ifndef get_av # define get_av perl_get_av #endif #ifndef get_hv # define get_hv perl_get_hv #endif /* Replace: 0 */ #ifndef dUNDERBAR # define dUNDERBAR dNOOP #endif #ifndef UNDERBAR # define UNDERBAR DEFSV #endif #ifndef dAX # define dAX I32 ax = MARK - PL_stack_base + 1 #endif #ifndef dITEMS # define dITEMS I32 items = SP - MARK #endif #ifndef dXSTARG # define dXSTARG SV * targ = sv_newmortal() #endif #ifndef dAXMARK # define dAXMARK I32 ax = POPMARK; \ register SV ** const mark = PL_stack_base + ax++ #endif #ifndef XSprePUSH # define XSprePUSH (sp = PL_stack_base + ax - 1) #endif #if (PERL_BCDVERSION < 0x5005000) # undef XSRETURN # define XSRETURN(off) \ STMT_START { \ PL_stack_sp = PL_stack_base + ax + ((off) - 1); \ return; \ } STMT_END #endif #ifndef XSPROTO # define XSPROTO(name) void name(pTHX_ CV* cv) #endif #ifndef SVfARG # define SVfARG(p) ((void*)(p)) #endif #ifndef PERL_ABS # define PERL_ABS(x) ((x) < 0 ? -(x) : (x)) #endif #ifndef dVAR # define dVAR dNOOP #endif #ifndef SVf # define SVf "_" #endif #ifndef UTF8_MAXBYTES # define UTF8_MAXBYTES UTF8_MAXLEN #endif #ifndef CPERLscope # define CPERLscope(x) x #endif #ifndef PERL_HASH # define PERL_HASH(hash,str,len) \ STMT_START { \ const char *s_PeRlHaSh = str; \ I32 i_PeRlHaSh = len; \ U32 hash_PeRlHaSh = 0; \ while (i_PeRlHaSh--) \ hash_PeRlHaSh = hash_PeRlHaSh * 33 + *s_PeRlHaSh++; \ (hash) = hash_PeRlHaSh; \ } STMT_END #endif #ifndef PERLIO_FUNCS_DECL # ifdef PERLIO_FUNCS_CONST # define PERLIO_FUNCS_DECL(funcs) const PerlIO_funcs funcs # define PERLIO_FUNCS_CAST(funcs) (PerlIO_funcs*)(funcs) # else # define PERLIO_FUNCS_DECL(funcs) PerlIO_funcs funcs # define PERLIO_FUNCS_CAST(funcs) (funcs) # endif #endif /* provide these typedefs for older perls */ #if (PERL_BCDVERSION < 0x5009003) # ifdef ARGSproto typedef OP* (CPERLscope(*Perl_ppaddr_t))(ARGSproto); # else typedef OP* (CPERLscope(*Perl_ppaddr_t))(pTHX); # endif typedef OP* (CPERLscope(*Perl_check_t)) (pTHX_ OP*); #endif #ifndef WIDEST_UTYPE # ifdef QUADKIND # ifdef U64TYPE # define WIDEST_UTYPE U64TYPE # else # define WIDEST_UTYPE Quad_t # endif # else # define WIDEST_UTYPE U32 # endif #endif #ifdef EBCDIC /* This is the first version where these macros are fully correct. Relying on * the C library functions, as earlier releases did, causes problems with * locales */ # if (PERL_BCDVERSION < 0x5022000) # undef isALNUM # undef isALNUM_A # undef isALNUMC # undef isALNUMC_A # undef isALPHA # undef isALPHA_A # undef isALPHANUMERIC # undef isALPHANUMERIC_A # undef isASCII # undef isASCII_A # undef isBLANK # undef isBLANK_A # undef isCNTRL # undef isCNTRL_A # undef isDIGIT # undef isDIGIT_A # undef isGRAPH # undef isGRAPH_A # undef isIDCONT # undef isIDCONT_A # undef isIDFIRST # undef isIDFIRST_A # undef isLOWER # undef isLOWER_A # undef isOCTAL # undef isOCTAL_A # undef isPRINT # undef isPRINT_A # undef isPSXSPC # undef isPSXSPC_A # undef isPUNCT # undef isPUNCT_A # undef isSPACE # undef isSPACE_A # undef isUPPER # undef isUPPER_A # undef isWORDCHAR # undef isWORDCHAR_A # undef isXDIGIT # undef isXDIGIT_A # endif #ifndef isASCII # define isASCII(c) (isCNTRL(c) || isPRINT(c)) #endif /* The below is accurate for all EBCDIC code pages supported by * all the versions of Perl overridden by this */ #ifndef isCNTRL # define isCNTRL(c) ( (c) == '\0' || (c) == '\a' || (c) == '\b' \ || (c) == '\f' || (c) == '\n' || (c) == '\r' \ || (c) == '\t' || (c) == '\v' \ || ((c) <= 3 && (c) >= 1) /* SOH, STX, ETX */ \ || (c) == 7 /* U+7F DEL */ \ || ((c) <= 0x13 && (c) >= 0x0E) /* SO, SI */ \ /* DLE, DC[1-3] */ \ || (c) == 0x18 /* U+18 CAN */ \ || (c) == 0x19 /* U+19 EOM */ \ || ((c) <= 0x1F && (c) >= 0x1C) /* [FGRU]S */ \ || (c) == 0x26 /* U+17 ETB */ \ || (c) == 0x27 /* U+1B ESC */ \ || (c) == 0x2D /* U+05 ENQ */ \ || (c) == 0x2E /* U+06 ACK */ \ || (c) == 0x32 /* U+16 SYN */ \ || (c) == 0x37 /* U+04 EOT */ \ || (c) == 0x3C /* U+14 DC4 */ \ || (c) == 0x3D /* U+15 NAK */ \ || (c) == 0x3F /* U+1A SUB */ \ ) #endif /* The ordering of the tests in this and isUPPER are to exclude most characters * early */ #ifndef isLOWER # define isLOWER(c) ( (c) >= 'a' && (c) <= 'z' \ && ( (c) <= 'i' \ || ((c) >= 'j' && (c) <= 'r') \ || (c) >= 's')) #endif #ifndef isUPPER # define isUPPER(c) ( (c) >= 'A' && (c) <= 'Z' \ && ( (c) <= 'I' \ || ((c) >= 'J' && (c) <= 'R') \ || (c) >= 'S')) #endif #else /* Above is EBCDIC; below is ASCII */ # if (PERL_BCDVERSION < 0x5004000) /* The implementation of these in older perl versions can give wrong results if * the C program locale is set to other than the C locale */ # undef isALNUM # undef isALNUM_A # undef isALPHA # undef isALPHA_A # undef isDIGIT # undef isDIGIT_A # undef isIDFIRST # undef isIDFIRST_A # undef isLOWER # undef isLOWER_A # undef isUPPER # undef isUPPER_A # endif # if (PERL_BCDVERSION < 0x5008000) /* Hint: isCNTRL * Earlier perls omitted DEL */ # undef isCNTRL # endif # if (PERL_BCDVERSION < 0x5010000) /* Hint: isPRINT * The implementation in older perl versions includes all of the * isSPACE() characters, which is wrong. The version provided by * Devel::PPPort always overrides a present buggy version. */ # undef isPRINT # undef isPRINT_A # endif # if (PERL_BCDVERSION < 0x5014000) /* Hint: isASCII * The implementation in older perl versions always returned true if the * parameter was a signed char */ # undef isASCII # undef isASCII_A # endif # if (PERL_BCDVERSION < 0x5020000) /* Hint: isSPACE * The implementation in older perl versions didn't include \v */ # undef isSPACE # undef isSPACE_A # endif #ifndef isASCII # define isASCII(c) ((WIDEST_UTYPE) (c) <= 127) #endif #ifndef isCNTRL # define isCNTRL(c) ((WIDEST_UTYPE) (c) < ' ' || (c) == 127) #endif #ifndef isLOWER # define isLOWER(c) ((c) >= 'a' && (c) <= 'z') #endif #ifndef isUPPER # define isUPPER(c) ((c) <= 'Z' && (c) >= 'A') #endif #endif /* Below are definitions common to EBCDIC and ASCII */ #ifndef isALNUM # define isALNUM(c) isWORDCHAR(c) #endif #ifndef isALNUMC # define isALNUMC(c) isALPHANUMERIC(c) #endif #ifndef isALPHA # define isALPHA(c) (isUPPER(c) || isLOWER(c)) #endif #ifndef isALPHANUMERIC # define isALPHANUMERIC(c) (isALPHA(c) || isDIGIT(c)) #endif #ifndef isBLANK # define isBLANK(c) ((c) == ' ' || (c) == '\t') #endif #ifndef isDIGIT # define isDIGIT(c) ((c) <= '9' && (c) >= '0') #endif #ifndef isGRAPH # define isGRAPH(c) (isWORDCHAR(c) || isPUNCT(c)) #endif #ifndef isIDCONT # define isIDCONT(c) isWORDCHAR(c) #endif #ifndef isIDFIRST # define isIDFIRST(c) (isALPHA(c) || (c) == '_') #endif #ifndef isOCTAL # define isOCTAL(c) (((WIDEST_UTYPE)((c)) & ~7) == '0') #endif #ifndef isPRINT # define isPRINT(c) (isGRAPH(c) || (c) == ' ') #endif #ifndef isPSXSPC # define isPSXSPC(c) isSPACE(c) #endif #ifndef isPUNCT # define isPUNCT(c) ( (c) == '-' || (c) == '!' || (c) == '"' \ || (c) == '#' || (c) == '$' || (c) == '%' \ || (c) == '&' || (c) == '\'' || (c) == '(' \ || (c) == ')' || (c) == '*' || (c) == '+' \ || (c) == ',' || (c) == '.' || (c) == '/' \ || (c) == ':' || (c) == ';' || (c) == '<' \ || (c) == '=' || (c) == '>' || (c) == '?' \ || (c) == '@' || (c) == '[' || (c) == '\\' \ || (c) == ']' || (c) == '^' || (c) == '_' \ || (c) == '`' || (c) == '{' || (c) == '|' \ || (c) == '}' || (c) == '~') #endif #ifndef isSPACE # define isSPACE(c) ( isBLANK(c) || (c) == '\n' || (c) == '\r' \ || (c) == '\v' || (c) == '\f') #endif #ifndef isWORDCHAR # define isWORDCHAR(c) (isALPHANUMERIC(c) || (c) == '_') #endif #ifndef isXDIGIT # define isXDIGIT(c) ( isDIGIT(c) \ || ((c) >= 'a' && (c) <= 'f') \ || ((c) >= 'A' && (c) <= 'F')) #endif #ifndef isALNUM_A # define isALNUM_A isALNUM #endif #ifndef isALNUMC_A # define isALNUMC_A isALNUMC #endif #ifndef isALPHA_A # define isALPHA_A isALPHA #endif #ifndef isALPHANUMERIC_A # define isALPHANUMERIC_A isALPHANUMERIC #endif #ifndef isASCII_A # define isASCII_A isASCII #endif #ifndef isBLANK_A # define isBLANK_A isBLANK #endif #ifndef isCNTRL_A # define isCNTRL_A isCNTRL #endif #ifndef isDIGIT_A # define isDIGIT_A isDIGIT #endif #ifndef isGRAPH_A # define isGRAPH_A isGRAPH #endif #ifndef isIDCONT_A # define isIDCONT_A isIDCONT #endif #ifndef isIDFIRST_A # define isIDFIRST_A isIDFIRST #endif #ifndef isLOWER_A # define isLOWER_A isLOWER #endif #ifndef isOCTAL_A # define isOCTAL_A isOCTAL #endif #ifndef isPRINT_A # define isPRINT_A isPRINT #endif #ifndef isPSXSPC_A # define isPSXSPC_A isPSXSPC #endif #ifndef isPUNCT_A # define isPUNCT_A isPUNCT #endif #ifndef isSPACE_A # define isSPACE_A isSPACE #endif #ifndef isUPPER_A # define isUPPER_A isUPPER #endif #ifndef isWORDCHAR_A # define isWORDCHAR_A isWORDCHAR #endif #ifndef isXDIGIT_A # define isXDIGIT_A isXDIGIT #endif /* Until we figure out how to support this in older perls... */ #if (PERL_BCDVERSION >= 0x5008000) #ifndef HeUTF8 # define HeUTF8(he) ((HeKLEN(he) == HEf_SVKEY) ? \ SvUTF8(HeKEY_sv(he)) : \ (U32)HeKUTF8(he)) #endif #endif #ifndef C_ARRAY_LENGTH # define C_ARRAY_LENGTH(a) (sizeof(a)/sizeof((a)[0])) #endif #ifndef C_ARRAY_END # define C_ARRAY_END(a) ((a) + C_ARRAY_LENGTH(a)) #endif #ifndef LIKELY # define LIKELY(x) (x) #endif #ifndef UNLIKELY # define UNLIKELY(x) (x) #endif #ifndef UNICODE_REPLACEMENT # define UNICODE_REPLACEMENT 0xFFFD #endif #ifndef MUTABLE_PTR #if defined(__GNUC__) && !defined(PERL_GCC_BRACE_GROUPS_FORBIDDEN) # define MUTABLE_PTR(p) ({ void *_p = (p); _p; }) #else # define MUTABLE_PTR(p) ((void *) (p)) #endif #endif #ifndef MUTABLE_SV # define MUTABLE_SV(p) ((SV *)MUTABLE_PTR(p)) #endif #ifndef WARN_ALL # define WARN_ALL 0 #endif #ifndef WARN_CLOSURE # define WARN_CLOSURE 1 #endif #ifndef WARN_DEPRECATED # define WARN_DEPRECATED 2 #endif #ifndef WARN_EXITING # define WARN_EXITING 3 #endif #ifndef WARN_GLOB # define WARN_GLOB 4 #endif #ifndef WARN_IO # define WARN_IO 5 #endif #ifndef WARN_CLOSED # define WARN_CLOSED 6 #endif #ifndef WARN_EXEC # define WARN_EXEC 7 #endif #ifndef WARN_LAYER # define WARN_LAYER 8 #endif #ifndef WARN_NEWLINE # define WARN_NEWLINE 9 #endif #ifndef WARN_PIPE # define WARN_PIPE 10 #endif #ifndef WARN_UNOPENED # define WARN_UNOPENED 11 #endif #ifndef WARN_MISC # define WARN_MISC 12 #endif #ifndef WARN_NUMERIC # define WARN_NUMERIC 13 #endif #ifndef WARN_ONCE # define WARN_ONCE 14 #endif #ifndef WARN_OVERFLOW # define WARN_OVERFLOW 15 #endif #ifndef WARN_PACK # define WARN_PACK 16 #endif #ifndef WARN_PORTABLE # define WARN_PORTABLE 17 #endif #ifndef WARN_RECURSION # define WARN_RECURSION 18 #endif #ifndef WARN_REDEFINE # define WARN_REDEFINE 19 #endif #ifndef WARN_REGEXP # define WARN_REGEXP 20 #endif #ifndef WARN_SEVERE # define WARN_SEVERE 21 #endif #ifndef WARN_DEBUGGING # define WARN_DEBUGGING 22 #endif #ifndef WARN_INPLACE # define WARN_INPLACE 23 #endif #ifndef WARN_INTERNAL # define WARN_INTERNAL 24 #endif #ifndef WARN_MALLOC # define WARN_MALLOC 25 #endif #ifndef WARN_SIGNAL # define WARN_SIGNAL 26 #endif #ifndef WARN_SUBSTR # define WARN_SUBSTR 27 #endif #ifndef WARN_SYNTAX # define WARN_SYNTAX 28 #endif #ifndef WARN_AMBIGUOUS # define WARN_AMBIGUOUS 29 #endif #ifndef WARN_BAREWORD # define WARN_BAREWORD 30 #endif #ifndef WARN_DIGIT # define WARN_DIGIT 31 #endif #ifndef WARN_PARENTHESIS # define WARN_PARENTHESIS 32 #endif #ifndef WARN_PRECEDENCE # define WARN_PRECEDENCE 33 #endif #ifndef WARN_PRINTF # define WARN_PRINTF 34 #endif #ifndef WARN_PROTOTYPE # define WARN_PROTOTYPE 35 #endif #ifndef WARN_QW # define WARN_QW 36 #endif #ifndef WARN_RESERVED # define WARN_RESERVED 37 #endif #ifndef WARN_SEMICOLON # define WARN_SEMICOLON 38 #endif #ifndef WARN_TAINT # define WARN_TAINT 39 #endif #ifndef WARN_THREADS # define WARN_THREADS 40 #endif #ifndef WARN_UNINITIALIZED # define WARN_UNINITIALIZED 41 #endif #ifndef WARN_UNPACK # define WARN_UNPACK 42 #endif #ifndef WARN_UNTIE # define WARN_UNTIE 43 #endif #ifndef WARN_UTF8 # define WARN_UTF8 44 #endif #ifndef WARN_VOID # define WARN_VOID 45 #endif #ifndef WARN_ASSERTIONS # define WARN_ASSERTIONS 46 #endif #ifndef packWARN # define packWARN(a) (a) #endif #ifndef ckWARN # ifdef G_WARN_ON # define ckWARN(a) (PL_dowarn & G_WARN_ON) # else # define ckWARN(a) PL_dowarn # endif #endif #if (PERL_BCDVERSION >= 0x5004000) && !defined(warner) #if defined(NEED_warner) static void DPPP_(my_warner)(U32 err, const char *pat, ...); static #else extern void DPPP_(my_warner)(U32 err, const char *pat, ...); #endif #if defined(NEED_warner) || defined(NEED_warner_GLOBAL) #define Perl_warner DPPP_(my_warner) void DPPP_(my_warner)(U32 err, const char *pat, ...) { SV *sv; va_list args; PERL_UNUSED_ARG(err); va_start(args, pat); sv = vnewSVpvf(pat, &args); va_end(args); sv_2mortal(sv); warn("%s", SvPV_nolen(sv)); } #define warner Perl_warner #define Perl_warner_nocontext Perl_warner #endif #endif #define _ppport_MIN(a,b) (((a) <= (b)) ? (a) : (b)) #ifndef sv_setuv # define sv_setuv(sv, uv) \ STMT_START { \ UV TeMpUv = uv; \ if (TeMpUv <= IV_MAX) \ sv_setiv(sv, TeMpUv); \ else \ sv_setnv(sv, (double)TeMpUv); \ } STMT_END #endif #ifndef newSVuv # define newSVuv(uv) ((uv) <= IV_MAX ? newSViv((IV)uv) : newSVnv((NV)uv)) #endif #ifndef sv_2uv # define sv_2uv(sv) ((PL_Sv = (sv)), (UV) (SvNOK(PL_Sv) ? SvNV(PL_Sv) : sv_2nv(PL_Sv))) #endif #ifndef SvUVX # define SvUVX(sv) ((UV)SvIVX(sv)) #endif #ifndef SvUVXx # define SvUVXx(sv) SvUVX(sv) #endif #ifndef SvUV # define SvUV(sv) (SvIOK(sv) ? SvUVX(sv) : sv_2uv(sv)) #endif #ifndef SvUVx # define SvUVx(sv) ((PL_Sv = (sv)), SvUV(PL_Sv)) #endif /* Hint: sv_uv * Always use the SvUVx() macro instead of sv_uv(). */ #ifndef sv_uv # define sv_uv(sv) SvUVx(sv) #endif #if !defined(SvUOK) && defined(SvIOK_UV) # define SvUOK(sv) SvIOK_UV(sv) #endif #ifndef XST_mUV # define XST_mUV(i,v) (ST(i) = sv_2mortal(newSVuv(v)) ) #endif #ifndef XSRETURN_UV # define XSRETURN_UV(v) STMT_START { XST_mUV(0,v); XSRETURN(1); } STMT_END #endif #ifndef PUSHu # define PUSHu(u) STMT_START { sv_setuv(TARG, (UV)(u)); PUSHTARG; } STMT_END #endif #ifndef XPUSHu # define XPUSHu(u) STMT_START { sv_setuv(TARG, (UV)(u)); XPUSHTARG; } STMT_END #endif #if defined UTF8SKIP /* Don't use official version because it uses MIN, which may not be available */ #undef UTF8_SAFE_SKIP #ifndef UTF8_SAFE_SKIP # define UTF8_SAFE_SKIP(s, e) ( \ ((((e) - (s)) <= 0) \ ? 0 \ : _ppport_MIN(((e) - (s)), UTF8SKIP(s)))) #endif #endif #if !defined(my_strnlen) #if defined(NEED_my_strnlen) static STRLEN DPPP_(my_my_strnlen)(const char *str, Size_t maxlen); static #else extern STRLEN DPPP_(my_my_strnlen)(const char *str, Size_t maxlen); #endif #if defined(NEED_my_strnlen) || defined(NEED_my_strnlen_GLOBAL) #define my_strnlen DPPP_(my_my_strnlen) #define Perl_my_strnlen DPPP_(my_my_strnlen) STRLEN DPPP_(my_my_strnlen)(const char *str, Size_t maxlen) { const char *p = str; while(maxlen-- && *p) p++; return p - str; } #endif #endif #if (PERL_BCDVERSION < 0x5031002) /* Versions prior to this accepted things that are now considered * malformations, and didn't return -1 on error with warnings enabled * */ # undef utf8_to_uvchr_buf #endif /* This implementation brings modern, generally more restricted standards to * utf8_to_uvchr_buf. Some of these are security related, and clearly must * be done. But its arguable that the others need not, and hence should not. * The reason they're here is that a module that intends to play with the * latest perls shoud be able to work the same in all releases. An example is * that perl no longer accepts any UV for a code point, but limits them to * IV_MAX or below. This is for future internal use of the larger code points. * If it turns out that some of these changes are breaking code that isn't * intended to work with modern perls, the tighter restrictions could be * relaxed. khw thinks this is unlikely, but has been wrong in the past. */ #ifndef utf8_to_uvchr_buf /* Choose which underlying implementation to use. At least one must be * present or the perl is too early to handle this function */ # if defined(utf8n_to_uvchr) || defined(utf8_to_uv) # if defined(utf8n_to_uvchr) /* This is the preferred implementation */ # define _ppport_utf8_to_uvchr_buf_callee utf8n_to_uvchr # else # define _ppport_utf8_to_uvchr_buf_callee utf8_to_uv # endif # endif #ifdef _ppport_utf8_to_uvchr_buf_callee # if defined(NEED_utf8_to_uvchr_buf) static UV DPPP_(my_utf8_to_uvchr_buf)(pTHX_ const U8 * s, const U8 * send, STRLEN * retlen); static #else extern UV DPPP_(my_utf8_to_uvchr_buf)(pTHX_ const U8 * s, const U8 * send, STRLEN * retlen); #endif #if defined(NEED_utf8_to_uvchr_buf) || defined(NEED_utf8_to_uvchr_buf_GLOBAL) #ifdef utf8_to_uvchr_buf # undef utf8_to_uvchr_buf #endif #define utf8_to_uvchr_buf(a,b,c) DPPP_(my_utf8_to_uvchr_buf)(aTHX_ a,b,c) #define Perl_utf8_to_uvchr_buf DPPP_(my_utf8_to_uvchr_buf) UV DPPP_(my_utf8_to_uvchr_buf)(pTHX_ const U8 *s, const U8 *send, STRLEN *retlen) { UV ret; STRLEN curlen; bool overflows = 0; const U8 *cur_s = s; const bool do_warnings = ckWARN_d(WARN_UTF8); if (send > s) { curlen = send - s; } else { assert(0); /* Modern perls die under this circumstance */ curlen = 0; if (! do_warnings) { /* Handle empty here if no warnings needed */ if (retlen) *retlen = 0; return UNICODE_REPLACEMENT; } } /* The modern version allows anything that evaluates to a legal UV, but not * overlongs nor an empty input */ ret = _ppport_utf8_to_uvchr_buf_callee( s, curlen, retlen, (UTF8_ALLOW_ANYUV & ~(UTF8_ALLOW_LONG|UTF8_ALLOW_EMPTY))); /* But actually, modern versions restrict the UV to being no more than what * an IV can hold */ if (ret > PERL_INT_MAX) { overflows = 1; } # if (PERL_BCDVERSION < 0x5026000) # ifndef EBCDIC /* There are bugs in versions earlier than this on non-EBCDIC platforms * in which it did not detect all instances of overflow, which could be * a security hole. Also, earlier versions did not allow the overflow * malformation under any circumstances, and modern ones do. So we * need to check here. */ else if (curlen > 0 && *s >= 0xFE) { /* If the main routine detected overflow, great; it returned 0. But if the * input's first byte indicates it could overflow, we need to verify. * First, on a 32-bit machine the first byte being at least \xFE * automatically is overflow */ if (sizeof(ret) < 8) { overflows = 1; } else { const U8 highest[] = /* 2*63-1 */ "\xFF\x80\x87\xBF\xBF\xBF\xBF\xBF\xBF\xBF\xBF\xBF\xBF"; const U8 *cur_h = highest; for (cur_s = s; cur_s < send; cur_s++, cur_h++) { if (UNLIKELY(*cur_s == *cur_h)) { continue; } /* If this byte is larger than the corresponding highest UTF-8 * byte, the sequence overflows; otherwise the byte is less than * (as we handled the equality case above), and so the sequence * doesn't overflow */ overflows = *cur_s > *cur_h; break; } /* Here, either we set the bool and broke out of the loop, or got * to the end and all bytes are the same which indicates it doesn't * overflow. */ } } # endif # endif /* < 5.26 */ if (UNLIKELY(overflows)) { if (! do_warnings) { if (retlen) { *retlen = _ppport_MIN(*retlen, UTF8SKIP(s)); *retlen = _ppport_MIN(*retlen, curlen); } return UNICODE_REPLACEMENT; } else { /* On versions that correctly detect overflow, but forbid it * always, 0 will be returned, but also a warning will have been * raised. Don't repeat it */ if (ret != 0) { /* We use the error message in use from 5.8-5.14 */ Perl_warner(aTHX_ packWARN(WARN_UTF8), "Malformed UTF-8 character (overflow at 0x%" UVxf ", byte 0x%02x, after start byte 0x%02x)", ret, *cur_s, *s); } if (retlen) { *retlen = (STRLEN) -1; } return 0; } } /* If failed and warnings are off, to emulate the behavior of the real * utf8_to_uvchr(), try again, allowing anything. (Note a return of 0 is * ok if the input was '\0') */ if (UNLIKELY(ret == 0 && (curlen == 0 || *s != '\0'))) { /* If curlen is 0, we already handled the case where warnings are * disabled, so this 'if' will be true, and we won't look at the * contents of 's' */ if (do_warnings) { *retlen = (STRLEN) -1; } else { ret = _ppport_utf8_to_uvchr_buf_callee( s, curlen, retlen, UTF8_ALLOW_ANY); /* Override with the REPLACEMENT character, as that is what the * modern version of this function returns */ ret = UNICODE_REPLACEMENT; # if (PERL_BCDVERSION < 0x5016000) /* Versions earlier than this don't necessarily return the proper * length. It should not extend past the end of string, nor past * what the first byte indicates the length is, nor past the * continuation characters */ if (retlen && *retlen >= 0) { *retlen = _ppport_MIN(*retlen, curlen); *retlen = _ppport_MIN(*retlen, UTF8SKIP(s)); unsigned int i = 1; do { if (s[i] < 0x80 || s[i] > 0xBF) { *retlen = i; break; } } while (++i < *retlen); } # endif } } return ret; } # endif #endif #endif #if defined(UTF8SKIP) && defined(utf8_to_uvchr_buf) #undef utf8_to_uvchr /* Always redefine this unsafe function so that it refuses to read past a NUL, making it much less likely to read off the end of the buffer. A NUL indicates the start of the next character anyway. If the input isn't NUL-terminated, the function remains unsafe, as it always has been. */ #ifndef utf8_to_uvchr # define utf8_to_uvchr(s, lp) \ ((*(s) == '\0') \ ? utf8_to_uvchr_buf(s,((s)+1), lp) /* Handle single NUL specially */ \ : utf8_to_uvchr_buf(s, (s) + my_strnlen((char *) (s), UTF8SKIP(s)), (lp))) #endif #endif #ifdef HAS_MEMCMP #ifndef memNE # define memNE(s1,s2,l) (memcmp(s1,s2,l)) #endif #ifndef memEQ # define memEQ(s1,s2,l) (!memcmp(s1,s2,l)) #endif #else #ifndef memNE # define memNE(s1,s2,l) (bcmp(s1,s2,l)) #endif #ifndef memEQ # define memEQ(s1,s2,l) (!bcmp(s1,s2,l)) #endif #endif #ifndef memEQs # define memEQs(s1, l, s2) \ (sizeof(s2)-1 == l && memEQ(s1, (s2 ""), (sizeof(s2)-1))) #endif #ifndef memNEs # define memNEs(s1, l, s2) !memEQs(s1, l, s2) #endif #ifndef MoveD # define MoveD(s,d,n,t) memmove((char*)(d),(char*)(s), (n) * sizeof(t)) #endif #ifndef CopyD # define CopyD(s,d,n,t) memcpy((char*)(d),(char*)(s), (n) * sizeof(t)) #endif #ifdef HAS_MEMSET #ifndef ZeroD # define ZeroD(d,n,t) memzero((char*)(d), (n) * sizeof(t)) #endif #else #ifndef ZeroD # define ZeroD(d,n,t) ((void)memzero((char*)(d), (n) * sizeof(t)), d) #endif #endif #ifndef PoisonWith # define PoisonWith(d,n,t,b) (void)memset((char*)(d), (U8)(b), (n) * sizeof(t)) #endif #ifndef PoisonNew # define PoisonNew(d,n,t) PoisonWith(d,n,t,0xAB) #endif #ifndef PoisonFree # define PoisonFree(d,n,t) PoisonWith(d,n,t,0xEF) #endif #ifndef Poison # define Poison(d,n,t) PoisonFree(d,n,t) #endif #ifndef Newx # define Newx(v,n,t) New(0,v,n,t) #endif #ifndef Newxc # define Newxc(v,n,t,c) Newc(0,v,n,t,c) #endif #ifndef Newxz # define Newxz(v,n,t) Newz(0,v,n,t) #endif #ifndef PERL_MAGIC_sv # define PERL_MAGIC_sv '\0' #endif #ifndef PERL_MAGIC_overload # define PERL_MAGIC_overload 'A' #endif #ifndef PERL_MAGIC_overload_elem # define PERL_MAGIC_overload_elem 'a' #endif #ifndef PERL_MAGIC_overload_table # define PERL_MAGIC_overload_table 'c' #endif #ifndef PERL_MAGIC_bm # define PERL_MAGIC_bm 'B' #endif #ifndef PERL_MAGIC_regdata # define PERL_MAGIC_regdata 'D' #endif #ifndef PERL_MAGIC_regdatum # define PERL_MAGIC_regdatum 'd' #endif #ifndef PERL_MAGIC_env # define PERL_MAGIC_env 'E' #endif #ifndef PERL_MAGIC_envelem # define PERL_MAGIC_envelem 'e' #endif #ifndef PERL_MAGIC_fm # define PERL_MAGIC_fm 'f' #endif #ifndef PERL_MAGIC_regex_global # define PERL_MAGIC_regex_global 'g' #endif #ifndef PERL_MAGIC_isa # define PERL_MAGIC_isa 'I' #endif #ifndef PERL_MAGIC_isaelem # define PERL_MAGIC_isaelem 'i' #endif #ifndef PERL_MAGIC_nkeys # define PERL_MAGIC_nkeys 'k' #endif #ifndef PERL_MAGIC_dbfile # define PERL_MAGIC_dbfile 'L' #endif #ifndef PERL_MAGIC_dbline # define PERL_MAGIC_dbline 'l' #endif #ifndef PERL_MAGIC_mutex # define PERL_MAGIC_mutex 'm' #endif #ifndef PERL_MAGIC_shared # define PERL_MAGIC_shared 'N' #endif #ifndef PERL_MAGIC_shared_scalar # define PERL_MAGIC_shared_scalar 'n' #endif #ifndef PERL_MAGIC_collxfrm # define PERL_MAGIC_collxfrm 'o' #endif #ifndef PERL_MAGIC_tied # define PERL_MAGIC_tied 'P' #endif #ifndef PERL_MAGIC_tiedelem # define PERL_MAGIC_tiedelem 'p' #endif #ifndef PERL_MAGIC_tiedscalar # define PERL_MAGIC_tiedscalar 'q' #endif #ifndef PERL_MAGIC_qr # define PERL_MAGIC_qr 'r' #endif #ifndef PERL_MAGIC_sig # define PERL_MAGIC_sig 'S' #endif #ifndef PERL_MAGIC_sigelem # define PERL_MAGIC_sigelem 's' #endif #ifndef PERL_MAGIC_taint # define PERL_MAGIC_taint 't' #endif #ifndef PERL_MAGIC_uvar # define PERL_MAGIC_uvar 'U' #endif #ifndef PERL_MAGIC_uvar_elem # define PERL_MAGIC_uvar_elem 'u' #endif #ifndef PERL_MAGIC_vstring # define PERL_MAGIC_vstring 'V' #endif #ifndef PERL_MAGIC_vec # define PERL_MAGIC_vec 'v' #endif #ifndef PERL_MAGIC_utf8 # define PERL_MAGIC_utf8 'w' #endif #ifndef PERL_MAGIC_substr # define PERL_MAGIC_substr 'x' #endif #ifndef PERL_MAGIC_defelem # define PERL_MAGIC_defelem 'y' #endif #ifndef PERL_MAGIC_glob # define PERL_MAGIC_glob '*' #endif #ifndef PERL_MAGIC_arylen # define PERL_MAGIC_arylen '#' #endif #ifndef PERL_MAGIC_pos # define PERL_MAGIC_pos '.' #endif #ifndef PERL_MAGIC_backref # define PERL_MAGIC_backref '<' #endif #ifndef PERL_MAGIC_ext # define PERL_MAGIC_ext '~' #endif #ifdef NEED_mess_sv #define NEED_mess #endif #ifdef NEED_mess #define NEED_mess_nocontext #define NEED_vmess #endif #ifndef croak_sv #if (PERL_BCDVERSION >= 0x5007003) || ( (PERL_BCDVERSION >= 0x5006001) && (PERL_BCDVERSION < 0x5007000) ) # if ( (PERL_BCDVERSION >= 0x5008000) && (PERL_BCDVERSION < 0x5008009) ) || ( (PERL_BCDVERSION >= 0x5009000) && (PERL_BCDVERSION < 0x5010001) ) # define D_PPP_FIX_UTF8_ERRSV(errsv, sv) \ STMT_START { \ if (sv != errsv) \ SvFLAGS(errsv) = (SvFLAGS(errsv) & ~SVf_UTF8) | \ (SvFLAGS(sv) & SVf_UTF8); \ } STMT_END # else # define D_PPP_FIX_UTF8_ERRSV(errsv, sv) STMT_START {} STMT_END # endif # define croak_sv(sv) \ STMT_START { \ if (SvROK(sv)) { \ sv_setsv(ERRSV, sv); \ croak(NULL); \ } else { \ D_PPP_FIX_UTF8_ERRSV(ERRSV, sv); \ croak("%" SVf, SVfARG(sv)); \ } \ } STMT_END #elif (PERL_BCDVERSION >= 0x5004000) # define croak_sv(sv) croak("%" SVf, SVfARG(sv)) #else # define croak_sv(sv) croak("%s", SvPV_nolen(sv)) #endif #endif #ifndef die_sv #if defined(NEED_die_sv) static OP * DPPP_(my_die_sv)(pTHX_ SV *sv); static #else extern OP * DPPP_(my_die_sv)(pTHX_ SV *sv); #endif #if defined(NEED_die_sv) || defined(NEED_die_sv_GLOBAL) #ifdef die_sv # undef die_sv #endif #define die_sv(a) DPPP_(my_die_sv)(aTHX_ a) #define Perl_die_sv DPPP_(my_die_sv) OP * DPPP_(my_die_sv)(pTHX_ SV *sv) { croak_sv(sv); return (OP *)NULL; } #endif #endif #ifndef warn_sv #if (PERL_BCDVERSION >= 0x5004000) # define warn_sv(sv) warn("%" SVf, SVfARG(sv)) #else # define warn_sv(sv) warn("%s", SvPV_nolen(sv)) #endif #endif #ifndef vmess #if defined(NEED_vmess) static SV * DPPP_(my_vmess)(pTHX_ const char * pat, va_list * args); static #else extern SV * DPPP_(my_vmess)(pTHX_ const char * pat, va_list * args); #endif #if defined(NEED_vmess) || defined(NEED_vmess_GLOBAL) #ifdef vmess # undef vmess #endif #define vmess(a,b) DPPP_(my_vmess)(aTHX_ a,b) #define Perl_vmess DPPP_(my_vmess) SV* DPPP_(my_vmess)(pTHX_ const char* pat, va_list* args) { mess(pat, args); return PL_mess_sv; } #endif #endif #if (PERL_BCDVERSION < 0x5006000) #undef mess #endif #if !defined(mess_nocontext) && !defined(Perl_mess_nocontext) #if defined(NEED_mess_nocontext) static SV * DPPP_(my_mess_nocontext)(const char * pat, ...); static #else extern SV * DPPP_(my_mess_nocontext)(const char * pat, ...); #endif #if defined(NEED_mess_nocontext) || defined(NEED_mess_nocontext_GLOBAL) #define mess_nocontext DPPP_(my_mess_nocontext) #define Perl_mess_nocontext DPPP_(my_mess_nocontext) SV* DPPP_(my_mess_nocontext)(const char* pat, ...) { dTHX; SV *sv; va_list args; va_start(args, pat); sv = vmess(pat, &args); va_end(args); return sv; } #endif #endif #ifndef mess #if defined(NEED_mess) static SV * DPPP_(my_mess)(pTHX_ const char * pat, ...); static #else extern SV * DPPP_(my_mess)(pTHX_ const char * pat, ...); #endif #if defined(NEED_mess) || defined(NEED_mess_GLOBAL) #define Perl_mess DPPP_(my_mess) SV* DPPP_(my_mess)(pTHX_ const char* pat, ...) { SV *sv; va_list args; va_start(args, pat); sv = vmess(pat, &args); va_end(args); return sv; } #ifdef mess_nocontext #define mess mess_nocontext #else #define mess Perl_mess_nocontext #endif #endif #endif #ifndef mess_sv #if defined(NEED_mess_sv) static SV * DPPP_(my_mess_sv)(pTHX_ SV * basemsg, bool consume); static #else extern SV * DPPP_(my_mess_sv)(pTHX_ SV * basemsg, bool consume); #endif #if defined(NEED_mess_sv) || defined(NEED_mess_sv_GLOBAL) #ifdef mess_sv # undef mess_sv #endif #define mess_sv(a,b) DPPP_(my_mess_sv)(aTHX_ a,b) #define Perl_mess_sv DPPP_(my_mess_sv) SV * DPPP_(my_mess_sv)(pTHX_ SV *basemsg, bool consume) { SV *tmp; SV *ret; if (SvPOK(basemsg) && SvCUR(basemsg) && *(SvEND(basemsg)-1) == '\n') { if (consume) return basemsg; ret = mess(""); SvSetSV_nosteal(ret, basemsg); return ret; } if (consume) { sv_catsv(basemsg, mess("")); return basemsg; } ret = mess(""); tmp = newSVsv(ret); SvSetSV_nosteal(ret, basemsg); sv_catsv(ret, tmp); sv_dec(tmp); return ret; } #endif #endif #ifndef warn_nocontext #define warn_nocontext warn #endif #ifndef croak_nocontext #define croak_nocontext croak #endif #ifndef croak_no_modify #define croak_no_modify() croak_nocontext("%s", PL_no_modify) #define Perl_croak_no_modify() croak_no_modify() #endif #ifndef croak_memory_wrap #if (PERL_BCDVERSION >= 0x5009002) || ( (PERL_BCDVERSION >= 0x5008006) && (PERL_BCDVERSION < 0x5009000) ) # define croak_memory_wrap() croak_nocontext("%s", PL_memory_wrap) #else # define croak_memory_wrap() croak_nocontext("panic: memory wrap") #endif #endif #ifndef croak_xs_usage #if defined(NEED_croak_xs_usage) static void DPPP_(my_croak_xs_usage)(const CV * const cv, const char * const params); static #else extern void DPPP_(my_croak_xs_usage)(const CV * const cv, const char * const params); #endif #if defined(NEED_croak_xs_usage) || defined(NEED_croak_xs_usage_GLOBAL) #define croak_xs_usage DPPP_(my_croak_xs_usage) #define Perl_croak_xs_usage DPPP_(my_croak_xs_usage) #ifndef PERL_ARGS_ASSERT_CROAK_XS_USAGE #define PERL_ARGS_ASSERT_CROAK_XS_USAGE assert(cv); assert(params) #endif void DPPP_(my_croak_xs_usage)(const CV *const cv, const char *const params) { dTHX; const GV *const gv = CvGV(cv); PERL_ARGS_ASSERT_CROAK_XS_USAGE; if (gv) { const char *const gvname = GvNAME(gv); const HV *const stash = GvSTASH(gv); const char *const hvname = stash ? HvNAME(stash) : NULL; if (hvname) croak("Usage: %s::%s(%s)", hvname, gvname, params); else croak("Usage: %s(%s)", gvname, params); } else { /* Pants. I don't think that it should be possible to get here. */ croak("Usage: CODE(0x%" UVxf ")(%s)", PTR2UV(cv), params); } } #endif #endif #ifndef PERL_SIGNALS_UNSAFE_FLAG #define PERL_SIGNALS_UNSAFE_FLAG 0x0001 #if (PERL_BCDVERSION < 0x5008000) # define D_PPP_PERL_SIGNALS_INIT PERL_SIGNALS_UNSAFE_FLAG #else # define D_PPP_PERL_SIGNALS_INIT 0 #endif #if defined(NEED_PL_signals) static U32 DPPP_(my_PL_signals) = D_PPP_PERL_SIGNALS_INIT; #elif defined(NEED_PL_signals_GLOBAL) U32 DPPP_(my_PL_signals) = D_PPP_PERL_SIGNALS_INIT; #else extern U32 DPPP_(my_PL_signals); #endif #define PL_signals DPPP_(my_PL_signals) #endif /* Hint: PL_ppaddr * Calling an op via PL_ppaddr requires passing a context argument * for threaded builds. Since the context argument is different for * 5.005 perls, you can use aTHXR (supplied by ppport.h), which will * automatically be defined as the correct argument. */ #if (PERL_BCDVERSION <= 0x5005005) /* Replace: 1 */ # define PL_ppaddr ppaddr # define PL_no_modify no_modify /* Replace: 0 */ #endif #if (PERL_BCDVERSION <= 0x5004005) /* Replace: 1 */ # define PL_DBsignal DBsignal # define PL_DBsingle DBsingle # define PL_DBsub DBsub # define PL_DBtrace DBtrace # define PL_Sv Sv # define PL_bufend bufend # define PL_bufptr bufptr # define PL_compiling compiling # define PL_copline copline # define PL_curcop curcop # define PL_curstash curstash # define PL_debstash debstash # define PL_defgv defgv # define PL_diehook diehook # define PL_dirty dirty # define PL_dowarn dowarn # define PL_errgv errgv # define PL_error_count error_count # define PL_expect expect # define PL_hexdigit hexdigit # define PL_hints hints # define PL_in_my in_my # define PL_laststatval laststatval # define PL_lex_state lex_state # define PL_lex_stuff lex_stuff # define PL_linestr linestr # define PL_na na # define PL_perl_destruct_level perl_destruct_level # define PL_perldb perldb # define PL_rsfp_filters rsfp_filters # define PL_rsfp rsfp # define PL_stack_base stack_base # define PL_stack_sp stack_sp # define PL_statcache statcache # define PL_stdingv stdingv # define PL_sv_arenaroot sv_arenaroot # define PL_sv_no sv_no # define PL_sv_undef sv_undef # define PL_sv_yes sv_yes # define PL_tainted tainted # define PL_tainting tainting # define PL_tokenbuf tokenbuf /* Replace: 0 */ #endif /* Warning: PL_parser * For perl versions earlier than 5.9.5, this is an always * non-NULL dummy. Also, it cannot be dereferenced. Don't * use it if you can avoid is and unless you absolutely know * what you're doing. * If you always check that PL_parser is non-NULL, you can * define DPPP_PL_parser_NO_DUMMY to avoid the creation of * a dummy parser structure. */ #if (PERL_BCDVERSION >= 0x5009005) # ifdef DPPP_PL_parser_NO_DUMMY # define D_PPP_my_PL_parser_var(var) ((PL_parser ? PL_parser : \ (croak("panic: PL_parser == NULL in %s:%d", \ __FILE__, __LINE__), (yy_parser *) NULL))->var) # else # ifdef DPPP_PL_parser_NO_DUMMY_WARNING # define D_PPP_parser_dummy_warning(var) # else # define D_PPP_parser_dummy_warning(var) \ warn("warning: dummy PL_" #var " used in %s:%d", __FILE__, __LINE__), # endif # define D_PPP_my_PL_parser_var(var) ((PL_parser ? PL_parser : \ (D_PPP_parser_dummy_warning(var) &DPPP_(dummy_PL_parser)))->var) #if defined(NEED_PL_parser) static yy_parser DPPP_(dummy_PL_parser); #elif defined(NEED_PL_parser_GLOBAL) yy_parser DPPP_(dummy_PL_parser); #else extern yy_parser DPPP_(dummy_PL_parser); #endif # endif /* PL_expect, PL_copline, PL_rsfp, PL_rsfp_filters, PL_linestr, PL_bufptr, PL_bufend, PL_lex_state, PL_lex_stuff, PL_tokenbuf depends on PL_parser */ /* Warning: PL_expect, PL_copline, PL_rsfp, PL_rsfp_filters, PL_linestr, PL_bufptr, PL_bufend, PL_lex_state, PL_lex_stuff, PL_tokenbuf * Do not use this variable unless you know exactly what you're * doing. It is internal to the perl parser and may change or even * be removed in the future. As of perl 5.9.5, you have to check * for (PL_parser != NULL) for this variable to have any effect. * An always non-NULL PL_parser dummy is provided for earlier * perl versions. * If PL_parser is NULL when you try to access this variable, a * dummy is being accessed instead and a warning is issued unless * you define DPPP_PL_parser_NO_DUMMY_WARNING. * If DPPP_PL_parser_NO_DUMMY is defined, the code trying to access * this variable will croak with a panic message. */ # define PL_expect D_PPP_my_PL_parser_var(expect) # define PL_copline D_PPP_my_PL_parser_var(copline) # define PL_rsfp D_PPP_my_PL_parser_var(rsfp) # define PL_rsfp_filters D_PPP_my_PL_parser_var(rsfp_filters) # define PL_linestr D_PPP_my_PL_parser_var(linestr) # define PL_bufptr D_PPP_my_PL_parser_var(bufptr) # define PL_bufend D_PPP_my_PL_parser_var(bufend) # define PL_lex_state D_PPP_my_PL_parser_var(lex_state) # define PL_lex_stuff D_PPP_my_PL_parser_var(lex_stuff) # define PL_tokenbuf D_PPP_my_PL_parser_var(tokenbuf) # define PL_in_my D_PPP_my_PL_parser_var(in_my) # define PL_in_my_stash D_PPP_my_PL_parser_var(in_my_stash) # define PL_error_count D_PPP_my_PL_parser_var(error_count) #else /* ensure that PL_parser != NULL and cannot be dereferenced */ # define PL_parser ((void *) 1) #endif #ifndef mPUSHs # define mPUSHs(s) PUSHs(sv_2mortal(s)) #endif #ifndef PUSHmortal # define PUSHmortal PUSHs(sv_newmortal()) #endif #ifndef mPUSHp # define mPUSHp(p,l) sv_setpvn(PUSHmortal, (p), (l)) #endif #ifndef mPUSHn # define mPUSHn(n) sv_setnv(PUSHmortal, (NV)(n)) #endif #ifndef mPUSHi # define mPUSHi(i) sv_setiv(PUSHmortal, (IV)(i)) #endif #ifndef mPUSHu # define mPUSHu(u) sv_setuv(PUSHmortal, (UV)(u)) #endif #ifndef mXPUSHs # define mXPUSHs(s) XPUSHs(sv_2mortal(s)) #endif #ifndef XPUSHmortal # define XPUSHmortal XPUSHs(sv_newmortal()) #endif #ifndef mXPUSHp # define mXPUSHp(p,l) STMT_START { EXTEND(sp,1); sv_setpvn(PUSHmortal, (p), (l)); } STMT_END #endif #ifndef mXPUSHn # define mXPUSHn(n) STMT_START { EXTEND(sp,1); sv_setnv(PUSHmortal, (NV)(n)); } STMT_END #endif #ifndef mXPUSHi # define mXPUSHi(i) STMT_START { EXTEND(sp,1); sv_setiv(PUSHmortal, (IV)(i)); } STMT_END #endif #ifndef mXPUSHu # define mXPUSHu(u) STMT_START { EXTEND(sp,1); sv_setuv(PUSHmortal, (UV)(u)); } STMT_END #endif /* Replace: 1 */ #ifndef call_sv # define call_sv perl_call_sv #endif #ifndef call_pv # define call_pv perl_call_pv #endif #ifndef call_argv # define call_argv perl_call_argv #endif #ifndef call_method # define call_method perl_call_method #endif #ifndef eval_sv # define eval_sv perl_eval_sv #endif /* Replace: 0 */ #ifndef PERL_LOADMOD_DENY # define PERL_LOADMOD_DENY 0x1 #endif #ifndef PERL_LOADMOD_NOIMPORT # define PERL_LOADMOD_NOIMPORT 0x2 #endif #ifndef PERL_LOADMOD_IMPORT_OPS # define PERL_LOADMOD_IMPORT_OPS 0x4 #endif #ifndef G_METHOD # define G_METHOD 64 # ifdef call_sv # undef call_sv # endif # if (PERL_BCDVERSION < 0x5006000) # define call_sv(sv, flags) ((flags) & G_METHOD ? perl_call_method((char *) SvPV_nolen_const(sv), \ (flags) & ~G_METHOD) : perl_call_sv(sv, flags)) # else # define call_sv(sv, flags) ((flags) & G_METHOD ? Perl_call_method(aTHX_ (char *) SvPV_nolen_const(sv), \ (flags) & ~G_METHOD) : Perl_call_sv(aTHX_ sv, flags)) # endif #endif /* Replace perl_eval_pv with eval_pv */ #ifndef eval_pv #if defined(NEED_eval_pv) static SV* DPPP_(my_eval_pv)(char *p, I32 croak_on_error); static #else extern SV* DPPP_(my_eval_pv)(char *p, I32 croak_on_error); #endif #if defined(NEED_eval_pv) || defined(NEED_eval_pv_GLOBAL) #ifdef eval_pv # undef eval_pv #endif #define eval_pv(a,b) DPPP_(my_eval_pv)(aTHX_ a,b) #define Perl_eval_pv DPPP_(my_eval_pv) SV* DPPP_(my_eval_pv)(char *p, I32 croak_on_error) { dSP; SV* sv = newSVpv(p, 0); PUSHMARK(sp); eval_sv(sv, G_SCALAR); SvREFCNT_dec(sv); SPAGAIN; sv = POPs; PUTBACK; if (croak_on_error && SvTRUEx(ERRSV)) croak_sv(ERRSV); return sv; } #endif #endif #ifndef vload_module #if defined(NEED_vload_module) static void DPPP_(my_vload_module)(U32 flags, SV *name, SV *ver, va_list *args); static #else extern void DPPP_(my_vload_module)(U32 flags, SV *name, SV *ver, va_list *args); #endif #if defined(NEED_vload_module) || defined(NEED_vload_module_GLOBAL) #ifdef vload_module # undef vload_module #endif #define vload_module(a,b,c,d) DPPP_(my_vload_module)(aTHX_ a,b,c,d) #define Perl_vload_module DPPP_(my_vload_module) void DPPP_(my_vload_module)(U32 flags, SV *name, SV *ver, va_list *args) { dTHR; dVAR; OP *veop, *imop; OP * const modname = newSVOP(OP_CONST, 0, name); /* 5.005 has a somewhat hacky force_normal that doesn't croak on SvREADONLY() if PL_compling is true. Current perls take care in ck_require() to correctly turn off SvREADONLY before calling force_normal_flags(). This seems a better fix than fudging PL_compling */ SvREADONLY_off(((SVOP*)modname)->op_sv); modname->op_private |= OPpCONST_BARE; if (ver) { veop = newSVOP(OP_CONST, 0, ver); } else veop = NULL; if (flags & PERL_LOADMOD_NOIMPORT) { imop = sawparens(newNULLLIST()); } else if (flags & PERL_LOADMOD_IMPORT_OPS) { imop = va_arg(*args, OP*); } else { SV *sv; imop = NULL; sv = va_arg(*args, SV*); while (sv) { imop = append_elem(OP_LIST, imop, newSVOP(OP_CONST, 0, sv)); sv = va_arg(*args, SV*); } } { const line_t ocopline = PL_copline; COP * const ocurcop = PL_curcop; const int oexpect = PL_expect; #if (PERL_BCDVERSION >= 0x5004000) utilize(!(flags & PERL_LOADMOD_DENY), start_subparse(FALSE, 0), veop, modname, imop); #elif (PERL_BCDVERSION > 0x5003000) utilize(!(flags & PERL_LOADMOD_DENY), start_subparse(), veop, modname, imop); #else utilize(!(flags & PERL_LOADMOD_DENY), start_subparse(), modname, imop); #endif PL_expect = oexpect; PL_copline = ocopline; PL_curcop = ocurcop; } } #endif #endif #ifndef load_module #if defined(NEED_load_module) static void DPPP_(my_load_module)(U32 flags, SV *name, SV *ver, ...); static #else extern void DPPP_(my_load_module)(U32 flags, SV *name, SV *ver, ...); #endif #if defined(NEED_load_module) || defined(NEED_load_module_GLOBAL) #ifdef load_module # undef load_module #endif #define load_module DPPP_(my_load_module) #define Perl_load_module DPPP_(my_load_module) void DPPP_(my_load_module)(U32 flags, SV *name, SV *ver, ...) { va_list args; va_start(args, ver); vload_module(flags, name, ver, &args); va_end(args); } #endif #endif #ifndef newRV_inc # define newRV_inc(sv) newRV(sv) /* Replace */ #endif #ifndef newRV_noinc #if defined(NEED_newRV_noinc) static SV * DPPP_(my_newRV_noinc)(SV *sv); static #else extern SV * DPPP_(my_newRV_noinc)(SV *sv); #endif #if defined(NEED_newRV_noinc) || defined(NEED_newRV_noinc_GLOBAL) #ifdef newRV_noinc # undef newRV_noinc #endif #define newRV_noinc(a) DPPP_(my_newRV_noinc)(aTHX_ a) #define Perl_newRV_noinc DPPP_(my_newRV_noinc) SV * DPPP_(my_newRV_noinc)(SV *sv) { SV *rv = (SV *)newRV(sv); SvREFCNT_dec(sv); return rv; } #endif #endif /* Hint: newCONSTSUB * Returns a CV* as of perl-5.7.1. This return value is not supported * by Devel::PPPort. */ /* newCONSTSUB from IO.xs is in the core starting with 5.004_63 */ #if (PERL_BCDVERSION < 0x5004063) && (PERL_BCDVERSION != 0x5004005) #if defined(NEED_newCONSTSUB) static void DPPP_(my_newCONSTSUB)(HV *stash, const char *name, SV *sv); static #else extern void DPPP_(my_newCONSTSUB)(HV *stash, const char *name, SV *sv); #endif #if defined(NEED_newCONSTSUB) || defined(NEED_newCONSTSUB_GLOBAL) #ifdef newCONSTSUB # undef newCONSTSUB #endif #define newCONSTSUB(a,b,c) DPPP_(my_newCONSTSUB)(aTHX_ a,b,c) #define Perl_newCONSTSUB DPPP_(my_newCONSTSUB) /* This is just a trick to avoid a dependency of newCONSTSUB on PL_parser */ /* (There's no PL_parser in perl < 5.005, so this is completely safe) */ #define D_PPP_PL_copline PL_copline void DPPP_(my_newCONSTSUB)(HV *stash, const char *name, SV *sv) { U32 oldhints = PL_hints; HV *old_cop_stash = PL_curcop->cop_stash; HV *old_curstash = PL_curstash; line_t oldline = PL_curcop->cop_line; PL_curcop->cop_line = D_PPP_PL_copline; PL_hints &= ~HINT_BLOCK_SCOPE; if (stash) PL_curstash = PL_curcop->cop_stash = stash; newSUB( #if (PERL_BCDVERSION < 0x5003022) start_subparse(), #elif (PERL_BCDVERSION == 0x5003022) start_subparse(0), #else /* 5.003_23 onwards */ start_subparse(FALSE, 0), #endif newSVOP(OP_CONST, 0, newSVpv((char *) name, 0)), newSVOP(OP_CONST, 0, &PL_sv_no), /* SvPV(&PL_sv_no) == "" -- GMB */ newSTATEOP(0, Nullch, newSVOP(OP_CONST, 0, sv)) ); PL_hints = oldhints; PL_curcop->cop_stash = old_cop_stash; PL_curstash = old_curstash; PL_curcop->cop_line = oldline; } #endif #endif /* * Boilerplate macros for initializing and accessing interpreter-local * data from C. All statics in extensions should be reworked to use * this, if you want to make the extension thread-safe. See ext/re/re.xs * for an example of the use of these macros. * * Code that uses these macros is responsible for the following: * 1. #define MY_CXT_KEY to a unique string, e.g. "DynaLoader_guts" * 2. Declare a typedef named my_cxt_t that is a structure that contains * all the data that needs to be interpreter-local. * 3. Use the START_MY_CXT macro after the declaration of my_cxt_t. * 4. Use the MY_CXT_INIT macro such that it is called exactly once * (typically put in the BOOT: section). * 5. Use the members of the my_cxt_t structure everywhere as * MY_CXT.member. * 6. Use the dMY_CXT macro (a declaration) in all the functions that * access MY_CXT. */ #if defined(MULTIPLICITY) || defined(PERL_OBJECT) || \ defined(PERL_CAPI) || defined(PERL_IMPLICIT_CONTEXT) #ifndef START_MY_CXT /* This must appear in all extensions that define a my_cxt_t structure, * right after the definition (i.e. at file scope). The non-threads * case below uses it to declare the data as static. */ #define START_MY_CXT #if (PERL_BCDVERSION < 0x5004068) /* Fetches the SV that keeps the per-interpreter data. */ #define dMY_CXT_SV \ SV *my_cxt_sv = get_sv(MY_CXT_KEY, FALSE) #else /* >= perl5.004_68 */ #define dMY_CXT_SV \ SV *my_cxt_sv = *hv_fetch(PL_modglobal, MY_CXT_KEY, \ sizeof(MY_CXT_KEY)-1, TRUE) #endif /* < perl5.004_68 */ /* This declaration should be used within all functions that use the * interpreter-local data. */ #define dMY_CXT \ dMY_CXT_SV; \ my_cxt_t *my_cxtp = INT2PTR(my_cxt_t*,SvUV(my_cxt_sv)) /* Creates and zeroes the per-interpreter data. * (We allocate my_cxtp in a Perl SV so that it will be released when * the interpreter goes away.) */ #define MY_CXT_INIT \ dMY_CXT_SV; \ /* newSV() allocates one more than needed */ \ my_cxt_t *my_cxtp = (my_cxt_t*)SvPVX(newSV(sizeof(my_cxt_t)-1));\ Zero(my_cxtp, 1, my_cxt_t); \ sv_setuv(my_cxt_sv, PTR2UV(my_cxtp)) /* This macro must be used to access members of the my_cxt_t structure. * e.g. MYCXT.some_data */ #define MY_CXT (*my_cxtp) /* Judicious use of these macros can reduce the number of times dMY_CXT * is used. Use is similar to pTHX, aTHX etc. */ #define pMY_CXT my_cxt_t *my_cxtp #define pMY_CXT_ pMY_CXT, #define _pMY_CXT ,pMY_CXT #define aMY_CXT my_cxtp #define aMY_CXT_ aMY_CXT, #define _aMY_CXT ,aMY_CXT #endif /* START_MY_CXT */ #ifndef MY_CXT_CLONE /* Clones the per-interpreter data. */ #define MY_CXT_CLONE \ dMY_CXT_SV; \ my_cxt_t *my_cxtp = (my_cxt_t*)SvPVX(newSV(sizeof(my_cxt_t)-1));\ Copy(INT2PTR(my_cxt_t*, SvUV(my_cxt_sv)), my_cxtp, 1, my_cxt_t);\ sv_setuv(my_cxt_sv, PTR2UV(my_cxtp)) #endif #else /* single interpreter */ #ifndef START_MY_CXT #define START_MY_CXT static my_cxt_t my_cxt; #define dMY_CXT_SV dNOOP #define dMY_CXT dNOOP #define MY_CXT_INIT NOOP #define MY_CXT my_cxt #define pMY_CXT void #define pMY_CXT_ #define _pMY_CXT #define aMY_CXT #define aMY_CXT_ #define _aMY_CXT #endif /* START_MY_CXT */ #ifndef MY_CXT_CLONE #define MY_CXT_CLONE NOOP #endif #endif #ifndef IVdf # if IVSIZE == LONGSIZE # define IVdf "ld" # define UVuf "lu" # define UVof "lo" # define UVxf "lx" # define UVXf "lX" # elif IVSIZE == INTSIZE # define IVdf "d" # define UVuf "u" # define UVof "o" # define UVxf "x" # define UVXf "X" # else # error "cannot define IV/UV formats" # endif #endif #ifndef NVef # if defined(USE_LONG_DOUBLE) && defined(HAS_LONG_DOUBLE) && \ defined(PERL_PRIfldbl) && (PERL_BCDVERSION != 0x5006000) /* Not very likely, but let's try anyway. */ # define NVef PERL_PRIeldbl # define NVff PERL_PRIfldbl # define NVgf PERL_PRIgldbl # else # define NVef "e" # define NVff "f" # define NVgf "g" # endif #endif #ifndef SvREFCNT_inc # ifdef PERL_USE_GCC_BRACE_GROUPS # define SvREFCNT_inc(sv) \ ({ \ SV * const _sv = (SV*)(sv); \ if (_sv) \ (SvREFCNT(_sv))++; \ _sv; \ }) # else # define SvREFCNT_inc(sv) \ ((PL_Sv=(SV*)(sv)) ? (++(SvREFCNT(PL_Sv)),PL_Sv) : NULL) # endif #endif #ifndef SvREFCNT_inc_simple # ifdef PERL_USE_GCC_BRACE_GROUPS # define SvREFCNT_inc_simple(sv) \ ({ \ if (sv) \ (SvREFCNT(sv))++; \ (SV *)(sv); \ }) # else # define SvREFCNT_inc_simple(sv) \ ((sv) ? (SvREFCNT(sv)++,(SV*)(sv)) : NULL) # endif #endif #ifndef SvREFCNT_inc_NN # ifdef PERL_USE_GCC_BRACE_GROUPS # define SvREFCNT_inc_NN(sv) \ ({ \ SV * const _sv = (SV*)(sv); \ SvREFCNT(_sv)++; \ _sv; \ }) # else # define SvREFCNT_inc_NN(sv) \ (PL_Sv=(SV*)(sv),++(SvREFCNT(PL_Sv)),PL_Sv) # endif #endif #ifndef SvREFCNT_inc_void # ifdef PERL_USE_GCC_BRACE_GROUPS # define SvREFCNT_inc_void(sv) \ ({ \ SV * const _sv = (SV*)(sv); \ if (_sv) \ (void)(SvREFCNT(_sv)++); \ }) # else # define SvREFCNT_inc_void(sv) \ (void)((PL_Sv=(SV*)(sv)) ? ++(SvREFCNT(PL_Sv)) : 0) # endif #endif #ifndef SvREFCNT_inc_simple_void # define SvREFCNT_inc_simple_void(sv) STMT_START { if (sv) SvREFCNT(sv)++; } STMT_END #endif #ifndef SvREFCNT_inc_simple_NN # define SvREFCNT_inc_simple_NN(sv) (++SvREFCNT(sv), (SV*)(sv)) #endif #ifndef SvREFCNT_inc_void_NN # define SvREFCNT_inc_void_NN(sv) (void)(++SvREFCNT((SV*)(sv))) #endif #ifndef SvREFCNT_inc_simple_void_NN # define SvREFCNT_inc_simple_void_NN(sv) (void)(++SvREFCNT((SV*)(sv))) #endif #ifndef newSV_type #if defined(NEED_newSV_type) static SV* DPPP_(my_newSV_type)(pTHX_ svtype const t); static #else extern SV* DPPP_(my_newSV_type)(pTHX_ svtype const t); #endif #if defined(NEED_newSV_type) || defined(NEED_newSV_type_GLOBAL) #ifdef newSV_type # undef newSV_type #endif #define newSV_type(a) DPPP_(my_newSV_type)(aTHX_ a) #define Perl_newSV_type DPPP_(my_newSV_type) SV* DPPP_(my_newSV_type)(pTHX_ svtype const t) { SV* const sv = newSV(0); sv_upgrade(sv, t); return sv; } #endif #endif #if (PERL_BCDVERSION < 0x5006000) # define D_PPP_CONSTPV_ARG(x) ((char *) (x)) #else # define D_PPP_CONSTPV_ARG(x) (x) #endif #ifndef newSVpvn # define newSVpvn(data,len) ((data) \ ? ((len) ? newSVpv((data), (len)) : newSVpv("", 0)) \ : newSV(0)) #endif #ifndef newSVpvn_utf8 # define newSVpvn_utf8(s, len, u) newSVpvn_flags((s), (len), (u) ? SVf_UTF8 : 0) #endif #ifndef SVf_UTF8 # define SVf_UTF8 0 #endif #ifndef newSVpvn_flags #if defined(NEED_newSVpvn_flags) static SV * DPPP_(my_newSVpvn_flags)(pTHX_ const char *s, STRLEN len, U32 flags); static #else extern SV * DPPP_(my_newSVpvn_flags)(pTHX_ const char *s, STRLEN len, U32 flags); #endif #if defined(NEED_newSVpvn_flags) || defined(NEED_newSVpvn_flags_GLOBAL) #ifdef newSVpvn_flags # undef newSVpvn_flags #endif #define newSVpvn_flags(a,b,c) DPPP_(my_newSVpvn_flags)(aTHX_ a,b,c) #define Perl_newSVpvn_flags DPPP_(my_newSVpvn_flags) SV * DPPP_(my_newSVpvn_flags)(pTHX_ const char *s, STRLEN len, U32 flags) { SV *sv = newSVpvn(D_PPP_CONSTPV_ARG(s), len); SvFLAGS(sv) |= (flags & SVf_UTF8); return (flags & SVs_TEMP) ? sv_2mortal(sv) : sv; } #endif #endif /* Backwards compatibility stuff... :-( */ #if !defined(NEED_sv_2pv_flags) && defined(NEED_sv_2pv_nolen) # define NEED_sv_2pv_flags #endif #if !defined(NEED_sv_2pv_flags_GLOBAL) && defined(NEED_sv_2pv_nolen_GLOBAL) # define NEED_sv_2pv_flags_GLOBAL #endif /* Hint: sv_2pv_nolen * Use the SvPV_nolen() or SvPV_nolen_const() macros instead of sv_2pv_nolen(). */ #ifndef sv_2pv_nolen # define sv_2pv_nolen(sv) SvPV_nolen(sv) #endif #ifdef SvPVbyte /* Hint: SvPVbyte * Does not work in perl-5.6.1, ppport.h implements a version * borrowed from perl-5.7.3. */ #if (PERL_BCDVERSION < 0x5007000) #if defined(NEED_sv_2pvbyte) static char * DPPP_(my_sv_2pvbyte)(pTHX_ SV *sv, STRLEN *lp); static #else extern char * DPPP_(my_sv_2pvbyte)(pTHX_ SV *sv, STRLEN *lp); #endif #if defined(NEED_sv_2pvbyte) || defined(NEED_sv_2pvbyte_GLOBAL) #ifdef sv_2pvbyte # undef sv_2pvbyte #endif #define sv_2pvbyte(a,b) DPPP_(my_sv_2pvbyte)(aTHX_ a,b) #define Perl_sv_2pvbyte DPPP_(my_sv_2pvbyte) char * DPPP_(my_sv_2pvbyte)(pTHX_ SV *sv, STRLEN *lp) { sv_utf8_downgrade(sv,0); return SvPV(sv,*lp); } #endif /* Hint: sv_2pvbyte * Use the SvPVbyte() macro instead of sv_2pvbyte(). */ #undef SvPVbyte #define SvPVbyte(sv, lp) \ ((SvFLAGS(sv) & (SVf_POK|SVf_UTF8)) == (SVf_POK) \ ? ((lp = SvCUR(sv)), SvPVX(sv)) : sv_2pvbyte(sv, &lp)) #endif #else # define SvPVbyte SvPV # define sv_2pvbyte sv_2pv #endif #ifndef sv_2pvbyte_nolen # define sv_2pvbyte_nolen(sv) sv_2pv_nolen(sv) #endif /* Hint: sv_pvn * Always use the SvPV() macro instead of sv_pvn(). */ /* Hint: sv_pvn_force * Always use the SvPV_force() macro instead of sv_pvn_force(). */ /* If these are undefined, they're not handled by the core anyway */ #ifndef SV_IMMEDIATE_UNREF # define SV_IMMEDIATE_UNREF 0 #endif #ifndef SV_GMAGIC # define SV_GMAGIC 0 #endif #ifndef SV_COW_DROP_PV # define SV_COW_DROP_PV 0 #endif #ifndef SV_UTF8_NO_ENCODING # define SV_UTF8_NO_ENCODING 0 #endif #ifndef SV_NOSTEAL # define SV_NOSTEAL 0 #endif #ifndef SV_CONST_RETURN # define SV_CONST_RETURN 0 #endif #ifndef SV_MUTABLE_RETURN # define SV_MUTABLE_RETURN 0 #endif #ifndef SV_SMAGIC # define SV_SMAGIC 0 #endif #ifndef SV_HAS_TRAILING_NUL # define SV_HAS_TRAILING_NUL 0 #endif #ifndef SV_COW_SHARED_HASH_KEYS # define SV_COW_SHARED_HASH_KEYS 0 #endif #if (PERL_BCDVERSION < 0x5007002) #if defined(NEED_sv_2pv_flags) static char * DPPP_(my_sv_2pv_flags)(pTHX_ SV *sv, STRLEN *lp, I32 flags); static #else extern char * DPPP_(my_sv_2pv_flags)(pTHX_ SV *sv, STRLEN *lp, I32 flags); #endif #if defined(NEED_sv_2pv_flags) || defined(NEED_sv_2pv_flags_GLOBAL) #ifdef sv_2pv_flags # undef sv_2pv_flags #endif #define sv_2pv_flags(a,b,c) DPPP_(my_sv_2pv_flags)(aTHX_ a,b,c) #define Perl_sv_2pv_flags DPPP_(my_sv_2pv_flags) char * DPPP_(my_sv_2pv_flags)(pTHX_ SV *sv, STRLEN *lp, I32 flags) { STRLEN n_a = (STRLEN) flags; return sv_2pv(sv, lp ? lp : &n_a); } #endif #if defined(NEED_sv_pvn_force_flags) static char * DPPP_(my_sv_pvn_force_flags)(pTHX_ SV *sv, STRLEN *lp, I32 flags); static #else extern char * DPPP_(my_sv_pvn_force_flags)(pTHX_ SV *sv, STRLEN *lp, I32 flags); #endif #if defined(NEED_sv_pvn_force_flags) || defined(NEED_sv_pvn_force_flags_GLOBAL) #ifdef sv_pvn_force_flags # undef sv_pvn_force_flags #endif #define sv_pvn_force_flags(a,b,c) DPPP_(my_sv_pvn_force_flags)(aTHX_ a,b,c) #define Perl_sv_pvn_force_flags DPPP_(my_sv_pvn_force_flags) char * DPPP_(my_sv_pvn_force_flags)(pTHX_ SV *sv, STRLEN *lp, I32 flags) { STRLEN n_a = (STRLEN) flags; return sv_pvn_force(sv, lp ? lp : &n_a); } #endif #endif #if (PERL_BCDVERSION < 0x5008008) || ( (PERL_BCDVERSION >= 0x5009000) && (PERL_BCDVERSION < 0x5009003) ) # define D_PPP_SVPV_NOLEN_LP_ARG &PL_na #else # define D_PPP_SVPV_NOLEN_LP_ARG 0 #endif #ifndef SvPV_const # define SvPV_const(sv, lp) SvPV_flags_const(sv, lp, SV_GMAGIC) #endif #ifndef SvPV_mutable # define SvPV_mutable(sv, lp) SvPV_flags_mutable(sv, lp, SV_GMAGIC) #endif #ifndef SvPV_flags # define SvPV_flags(sv, lp, flags) \ ((SvFLAGS(sv) & (SVf_POK)) == SVf_POK \ ? ((lp = SvCUR(sv)), SvPVX(sv)) : sv_2pv_flags(sv, &lp, flags)) #endif #ifndef SvPV_flags_const # define SvPV_flags_const(sv, lp, flags) \ ((SvFLAGS(sv) & (SVf_POK)) == SVf_POK \ ? ((lp = SvCUR(sv)), SvPVX_const(sv)) : \ (const char*) sv_2pv_flags(sv, &lp, flags|SV_CONST_RETURN)) #endif #ifndef SvPV_flags_const_nolen # define SvPV_flags_const_nolen(sv, flags) \ ((SvFLAGS(sv) & (SVf_POK)) == SVf_POK \ ? SvPVX_const(sv) : \ (const char*) sv_2pv_flags(sv, D_PPP_SVPV_NOLEN_LP_ARG, flags|SV_CONST_RETURN)) #endif #ifndef SvPV_flags_mutable # define SvPV_flags_mutable(sv, lp, flags) \ ((SvFLAGS(sv) & (SVf_POK)) == SVf_POK \ ? ((lp = SvCUR(sv)), SvPVX_mutable(sv)) : \ sv_2pv_flags(sv, &lp, flags|SV_MUTABLE_RETURN)) #endif #ifndef SvPV_force # define SvPV_force(sv, lp) SvPV_force_flags(sv, lp, SV_GMAGIC) #endif #ifndef SvPV_force_nolen # define SvPV_force_nolen(sv) SvPV_force_flags_nolen(sv, SV_GMAGIC) #endif #ifndef SvPV_force_mutable # define SvPV_force_mutable(sv, lp) SvPV_force_flags_mutable(sv, lp, SV_GMAGIC) #endif #ifndef SvPV_force_nomg # define SvPV_force_nomg(sv, lp) SvPV_force_flags(sv, lp, 0) #endif #ifndef SvPV_force_nomg_nolen # define SvPV_force_nomg_nolen(sv) SvPV_force_flags_nolen(sv, 0) #endif #ifndef SvPV_force_flags # define SvPV_force_flags(sv, lp, flags) \ ((SvFLAGS(sv) & (SVf_POK|SVf_THINKFIRST)) == SVf_POK \ ? ((lp = SvCUR(sv)), SvPVX(sv)) : sv_pvn_force_flags(sv, &lp, flags)) #endif #ifndef SvPV_force_flags_nolen # define SvPV_force_flags_nolen(sv, flags) \ ((SvFLAGS(sv) & (SVf_POK|SVf_THINKFIRST)) == SVf_POK \ ? SvPVX(sv) : sv_pvn_force_flags(sv, D_PPP_SVPV_NOLEN_LP_ARG, flags)) #endif #ifndef SvPV_force_flags_mutable # define SvPV_force_flags_mutable(sv, lp, flags) \ ((SvFLAGS(sv) & (SVf_POK|SVf_THINKFIRST)) == SVf_POK \ ? ((lp = SvCUR(sv)), SvPVX_mutable(sv)) \ : sv_pvn_force_flags(sv, &lp, flags|SV_MUTABLE_RETURN)) #endif #ifndef SvPV_nolen # define SvPV_nolen(sv) \ ((SvFLAGS(sv) & (SVf_POK)) == SVf_POK \ ? SvPVX(sv) : sv_2pv_flags(sv, D_PPP_SVPV_NOLEN_LP_ARG, SV_GMAGIC)) #endif #ifndef SvPV_nolen_const # define SvPV_nolen_const(sv) \ ((SvFLAGS(sv) & (SVf_POK)) == SVf_POK \ ? SvPVX_const(sv) : sv_2pv_flags(sv, D_PPP_SVPV_NOLEN_LP_ARG, SV_GMAGIC|SV_CONST_RETURN)) #endif #ifndef SvPV_nomg # define SvPV_nomg(sv, lp) SvPV_flags(sv, lp, 0) #endif #ifndef SvPV_nomg_const # define SvPV_nomg_const(sv, lp) SvPV_flags_const(sv, lp, 0) #endif #ifndef SvPV_nomg_const_nolen # define SvPV_nomg_const_nolen(sv) SvPV_flags_const_nolen(sv, 0) #endif #ifndef SvPV_nomg_nolen # define SvPV_nomg_nolen(sv) ((SvFLAGS(sv) & (SVf_POK)) == SVf_POK \ ? SvPVX(sv) : sv_2pv_flags(sv, D_PPP_SVPV_NOLEN_LP_ARG, 0)) #endif #ifndef SvPV_renew # define SvPV_renew(sv,n) STMT_START { SvLEN_set(sv, n); \ SvPV_set((sv), (char *) saferealloc( \ (Malloc_t)SvPVX(sv), (MEM_SIZE)((n)))); \ } STMT_END #endif #ifndef SvMAGIC_set # define SvMAGIC_set(sv, val) \ STMT_START { assert(SvTYPE(sv) >= SVt_PVMG); \ (((XPVMG*) SvANY(sv))->xmg_magic = (val)); } STMT_END #endif #if (PERL_BCDVERSION < 0x5009003) #ifndef SvPVX_const # define SvPVX_const(sv) ((const char*) (0 + SvPVX(sv))) #endif #ifndef SvPVX_mutable # define SvPVX_mutable(sv) (0 + SvPVX(sv)) #endif #ifndef SvRV_set # define SvRV_set(sv, val) \ STMT_START { assert(SvTYPE(sv) >= SVt_RV); \ (((XRV*) SvANY(sv))->xrv_rv = (val)); } STMT_END #endif #else #ifndef SvPVX_const # define SvPVX_const(sv) ((const char*)((sv)->sv_u.svu_pv)) #endif #ifndef SvPVX_mutable # define SvPVX_mutable(sv) ((sv)->sv_u.svu_pv) #endif #ifndef SvRV_set # define SvRV_set(sv, val) \ STMT_START { assert(SvTYPE(sv) >= SVt_RV); \ ((sv)->sv_u.svu_rv = (val)); } STMT_END #endif #endif #ifndef SvSTASH_set # define SvSTASH_set(sv, val) \ STMT_START { assert(SvTYPE(sv) >= SVt_PVMG); \ (((XPVMG*) SvANY(sv))->xmg_stash = (val)); } STMT_END #endif #if (PERL_BCDVERSION < 0x5004000) #ifndef SvUV_set # define SvUV_set(sv, val) \ STMT_START { assert(SvTYPE(sv) == SVt_IV || SvTYPE(sv) >= SVt_PVIV); \ (((XPVIV*) SvANY(sv))->xiv_iv = (IV) (val)); } STMT_END #endif #else #ifndef SvUV_set # define SvUV_set(sv, val) \ STMT_START { assert(SvTYPE(sv) == SVt_IV || SvTYPE(sv) >= SVt_PVIV); \ (((XPVUV*) SvANY(sv))->xuv_uv = (val)); } STMT_END #endif #endif #if (PERL_BCDVERSION >= 0x5004000) && !defined(vnewSVpvf) #if defined(NEED_vnewSVpvf) static SV * DPPP_(my_vnewSVpvf)(pTHX_ const char *pat, va_list *args); static #else extern SV * DPPP_(my_vnewSVpvf)(pTHX_ const char *pat, va_list *args); #endif #if defined(NEED_vnewSVpvf) || defined(NEED_vnewSVpvf_GLOBAL) #ifdef vnewSVpvf # undef vnewSVpvf #endif #define vnewSVpvf(a,b) DPPP_(my_vnewSVpvf)(aTHX_ a,b) #define Perl_vnewSVpvf DPPP_(my_vnewSVpvf) SV * DPPP_(my_vnewSVpvf)(pTHX_ const char *pat, va_list *args) { register SV *sv = newSV(0); sv_vsetpvfn(sv, pat, strlen(pat), args, Null(SV**), 0, Null(bool*)); return sv; } #endif #endif #if (PERL_BCDVERSION >= 0x5004000) && !defined(sv_vcatpvf) # define sv_vcatpvf(sv, pat, args) sv_vcatpvfn(sv, pat, strlen(pat), args, Null(SV**), 0, Null(bool*)) #endif #if (PERL_BCDVERSION >= 0x5004000) && !defined(sv_vsetpvf) # define sv_vsetpvf(sv, pat, args) sv_vsetpvfn(sv, pat, strlen(pat), args, Null(SV**), 0, Null(bool*)) #endif #if (PERL_BCDVERSION >= 0x5004000) && !defined(sv_catpvf_mg) #if defined(NEED_sv_catpvf_mg) static void DPPP_(my_sv_catpvf_mg)(pTHX_ SV *sv, const char *pat, ...); static #else extern void DPPP_(my_sv_catpvf_mg)(pTHX_ SV *sv, const char *pat, ...); #endif #if defined(NEED_sv_catpvf_mg) || defined(NEED_sv_catpvf_mg_GLOBAL) #define Perl_sv_catpvf_mg DPPP_(my_sv_catpvf_mg) void DPPP_(my_sv_catpvf_mg)(pTHX_ SV *sv, const char *pat, ...) { va_list args; va_start(args, pat); sv_vcatpvfn(sv, pat, strlen(pat), &args, Null(SV**), 0, Null(bool*)); SvSETMAGIC(sv); va_end(args); } #endif #endif #ifdef PERL_IMPLICIT_CONTEXT #if (PERL_BCDVERSION >= 0x5004000) && !defined(sv_catpvf_mg_nocontext) #if defined(NEED_sv_catpvf_mg_nocontext) static void DPPP_(my_sv_catpvf_mg_nocontext)(SV *sv, const char *pat, ...); static #else extern void DPPP_(my_sv_catpvf_mg_nocontext)(SV *sv, const char *pat, ...); #endif #if defined(NEED_sv_catpvf_mg_nocontext) || defined(NEED_sv_catpvf_mg_nocontext_GLOBAL) #define sv_catpvf_mg_nocontext DPPP_(my_sv_catpvf_mg_nocontext) #define Perl_sv_catpvf_mg_nocontext DPPP_(my_sv_catpvf_mg_nocontext) void DPPP_(my_sv_catpvf_mg_nocontext)(SV *sv, const char *pat, ...) { dTHX; va_list args; va_start(args, pat); sv_vcatpvfn(sv, pat, strlen(pat), &args, Null(SV**), 0, Null(bool*)); SvSETMAGIC(sv); va_end(args); } #endif #endif #endif /* sv_catpvf_mg depends on sv_catpvf_mg_nocontext */ #ifndef sv_catpvf_mg # ifdef PERL_IMPLICIT_CONTEXT # define sv_catpvf_mg Perl_sv_catpvf_mg_nocontext # else # define sv_catpvf_mg Perl_sv_catpvf_mg # endif #endif #if (PERL_BCDVERSION >= 0x5004000) && !defined(sv_vcatpvf_mg) # define sv_vcatpvf_mg(sv, pat, args) \ STMT_START { \ sv_vcatpvfn(sv, pat, strlen(pat), args, Null(SV**), 0, Null(bool*)); \ SvSETMAGIC(sv); \ } STMT_END #endif #if (PERL_BCDVERSION >= 0x5004000) && !defined(sv_setpvf_mg) #if defined(NEED_sv_setpvf_mg) static void DPPP_(my_sv_setpvf_mg)(pTHX_ SV *sv, const char *pat, ...); static #else extern void DPPP_(my_sv_setpvf_mg)(pTHX_ SV *sv, const char *pat, ...); #endif #if defined(NEED_sv_setpvf_mg) || defined(NEED_sv_setpvf_mg_GLOBAL) #define Perl_sv_setpvf_mg DPPP_(my_sv_setpvf_mg) void DPPP_(my_sv_setpvf_mg)(pTHX_ SV *sv, const char *pat, ...) { va_list args; va_start(args, pat); sv_vsetpvfn(sv, pat, strlen(pat), &args, Null(SV**), 0, Null(bool*)); SvSETMAGIC(sv); va_end(args); } #endif #endif #ifdef PERL_IMPLICIT_CONTEXT #if (PERL_BCDVERSION >= 0x5004000) && !defined(sv_setpvf_mg_nocontext) #if defined(NEED_sv_setpvf_mg_nocontext) static void DPPP_(my_sv_setpvf_mg_nocontext)(SV *sv, const char *pat, ...); static #else extern void DPPP_(my_sv_setpvf_mg_nocontext)(SV *sv, const char *pat, ...); #endif #if defined(NEED_sv_setpvf_mg_nocontext) || defined(NEED_sv_setpvf_mg_nocontext_GLOBAL) #define sv_setpvf_mg_nocontext DPPP_(my_sv_setpvf_mg_nocontext) #define Perl_sv_setpvf_mg_nocontext DPPP_(my_sv_setpvf_mg_nocontext) void DPPP_(my_sv_setpvf_mg_nocontext)(SV *sv, const char *pat, ...) { dTHX; va_list args; va_start(args, pat); sv_vsetpvfn(sv, pat, strlen(pat), &args, Null(SV**), 0, Null(bool*)); SvSETMAGIC(sv); va_end(args); } #endif #endif #endif /* sv_setpvf_mg depends on sv_setpvf_mg_nocontext */ #ifndef sv_setpvf_mg # ifdef PERL_IMPLICIT_CONTEXT # define sv_setpvf_mg Perl_sv_setpvf_mg_nocontext # else # define sv_setpvf_mg Perl_sv_setpvf_mg # endif #endif #if (PERL_BCDVERSION >= 0x5004000) && !defined(sv_vsetpvf_mg) # define sv_vsetpvf_mg(sv, pat, args) \ STMT_START { \ sv_vsetpvfn(sv, pat, strlen(pat), args, Null(SV**), 0, Null(bool*)); \ SvSETMAGIC(sv); \ } STMT_END #endif /* Hint: newSVpvn_share * The SVs created by this function only mimic the behaviour of * shared PVs without really being shared. Only use if you know * what you're doing. */ #ifndef newSVpvn_share #if defined(NEED_newSVpvn_share) static SV * DPPP_(my_newSVpvn_share)(pTHX_ const char *src, I32 len, U32 hash); static #else extern SV * DPPP_(my_newSVpvn_share)(pTHX_ const char *src, I32 len, U32 hash); #endif #if defined(NEED_newSVpvn_share) || defined(NEED_newSVpvn_share_GLOBAL) #ifdef newSVpvn_share # undef newSVpvn_share #endif #define newSVpvn_share(a,b,c) DPPP_(my_newSVpvn_share)(aTHX_ a,b,c) #define Perl_newSVpvn_share DPPP_(my_newSVpvn_share) SV * DPPP_(my_newSVpvn_share)(pTHX_ const char *src, I32 len, U32 hash) { SV *sv; if (len < 0) len = -len; if (!hash) PERL_HASH(hash, (char*) src, len); sv = newSVpvn((char *) src, len); sv_upgrade(sv, SVt_PVIV); SvIVX(sv) = hash; SvREADONLY_on(sv); SvPOK_on(sv); return sv; } #endif #endif #ifndef SvSHARED_HASH # define SvSHARED_HASH(sv) (0 + SvUVX(sv)) #endif #ifndef HvNAME_get # define HvNAME_get(hv) HvNAME(hv) #endif #ifndef HvNAMELEN_get # define HvNAMELEN_get(hv) (HvNAME_get(hv) ? (I32)strlen(HvNAME_get(hv)) : 0) #endif #ifndef gv_fetchpvn_flags #if defined(NEED_gv_fetchpvn_flags) static GV* DPPP_(my_gv_fetchpvn_flags)(pTHX_ const char* name, STRLEN len, int flags, int types); static #else extern GV* DPPP_(my_gv_fetchpvn_flags)(pTHX_ const char* name, STRLEN len, int flags, int types); #endif #if defined(NEED_gv_fetchpvn_flags) || defined(NEED_gv_fetchpvn_flags_GLOBAL) #ifdef gv_fetchpvn_flags # undef gv_fetchpvn_flags #endif #define gv_fetchpvn_flags(a,b,c,d) DPPP_(my_gv_fetchpvn_flags)(aTHX_ a,b,c,d) #define Perl_gv_fetchpvn_flags DPPP_(my_gv_fetchpvn_flags) GV* DPPP_(my_gv_fetchpvn_flags)(pTHX_ const char* name, STRLEN len, int flags, int types) { char *namepv = savepvn(name, len); GV* stash = gv_fetchpv(namepv, TRUE, SVt_PVHV); Safefree(namepv); return stash; } #endif #endif #ifndef GvSVn # define GvSVn(gv) GvSV(gv) #endif #ifndef isGV_with_GP # define isGV_with_GP(gv) isGV(gv) #endif #ifndef gv_fetchsv # define gv_fetchsv(name, flags, svt) gv_fetchpv(SvPV_nolen_const(name), flags, svt) #endif #ifndef get_cvn_flags # define get_cvn_flags(name, namelen, flags) get_cv(name, flags) #endif #ifndef gv_init_pvn # define gv_init_pvn(gv, stash, ptr, len, flags) gv_init(gv, stash, ptr, len, flags & GV_ADDMULTI ? TRUE : FALSE) #endif /* concatenating with "" ensures that only literal strings are accepted as argument * note that STR_WITH_LEN() can't be used as argument to macros or functions that * under some configurations might be macros */ #ifndef STR_WITH_LEN # define STR_WITH_LEN(s) (s ""), (sizeof(s)-1) #endif #ifndef newSVpvs # define newSVpvs(str) newSVpvn(str "", sizeof(str) - 1) #endif #ifndef newSVpvs_flags # define newSVpvs_flags(str, flags) newSVpvn_flags(str "", sizeof(str) - 1, flags) #endif #ifndef newSVpvs_share # define newSVpvs_share(str) newSVpvn_share(str "", sizeof(str) - 1, 0) #endif #ifndef sv_catpvs # define sv_catpvs(sv, str) sv_catpvn(sv, str "", sizeof(str) - 1) #endif #ifndef sv_setpvs # define sv_setpvs(sv, str) sv_setpvn(sv, str "", sizeof(str) - 1) #endif #ifndef hv_fetchs # define hv_fetchs(hv, key, lval) hv_fetch(hv, key "", sizeof(key) - 1, lval) #endif #ifndef hv_stores # define hv_stores(hv, key, val) hv_store(hv, key "", sizeof(key) - 1, val, 0) #endif #ifndef gv_fetchpvs # define gv_fetchpvs(name, flags, svt) gv_fetchpvn_flags(name "", sizeof(name) - 1, flags, svt) #endif #ifndef gv_stashpvs # define gv_stashpvs(name, flags) gv_stashpvn(name "", sizeof(name) - 1, flags) #endif #ifndef get_cvs # define get_cvs(name, flags) get_cvn_flags(name "", sizeof(name)-1, flags) #endif #ifndef SvGETMAGIC # define SvGETMAGIC(x) STMT_START { if (SvGMAGICAL(x)) mg_get(x); } STMT_END #endif /* That's the best we can do... */ #ifndef sv_catpvn_nomg # define sv_catpvn_nomg sv_catpvn #endif #ifndef sv_catsv_nomg # define sv_catsv_nomg sv_catsv #endif #ifndef sv_setsv_nomg # define sv_setsv_nomg sv_setsv #endif #ifndef sv_pvn_nomg # define sv_pvn_nomg sv_pvn #endif #ifndef SvIV_nomg # define SvIV_nomg SvIV #endif #ifndef SvUV_nomg # define SvUV_nomg SvUV #endif #ifndef sv_catpv_mg # define sv_catpv_mg(sv, ptr) \ STMT_START { \ SV *TeMpSv = sv; \ sv_catpv(TeMpSv,ptr); \ SvSETMAGIC(TeMpSv); \ } STMT_END #endif #ifndef sv_catpvn_mg # define sv_catpvn_mg(sv, ptr, len) \ STMT_START { \ SV *TeMpSv = sv; \ sv_catpvn(TeMpSv,ptr,len); \ SvSETMAGIC(TeMpSv); \ } STMT_END #endif #ifndef sv_catsv_mg # define sv_catsv_mg(dsv, ssv) \ STMT_START { \ SV *TeMpSv = dsv; \ sv_catsv(TeMpSv,ssv); \ SvSETMAGIC(TeMpSv); \ } STMT_END #endif #ifndef sv_setiv_mg # define sv_setiv_mg(sv, i) \ STMT_START { \ SV *TeMpSv = sv; \ sv_setiv(TeMpSv,i); \ SvSETMAGIC(TeMpSv); \ } STMT_END #endif #ifndef sv_setnv_mg # define sv_setnv_mg(sv, num) \ STMT_START { \ SV *TeMpSv = sv; \ sv_setnv(TeMpSv,num); \ SvSETMAGIC(TeMpSv); \ } STMT_END #endif #ifndef sv_setpv_mg # define sv_setpv_mg(sv, ptr) \ STMT_START { \ SV *TeMpSv = sv; \ sv_setpv(TeMpSv,ptr); \ SvSETMAGIC(TeMpSv); \ } STMT_END #endif #ifndef sv_setpvn_mg # define sv_setpvn_mg(sv, ptr, len) \ STMT_START { \ SV *TeMpSv = sv; \ sv_setpvn(TeMpSv,ptr,len); \ SvSETMAGIC(TeMpSv); \ } STMT_END #endif #ifndef sv_setsv_mg # define sv_setsv_mg(dsv, ssv) \ STMT_START { \ SV *TeMpSv = dsv; \ sv_setsv(TeMpSv,ssv); \ SvSETMAGIC(TeMpSv); \ } STMT_END #endif #ifndef sv_setuv_mg # define sv_setuv_mg(sv, i) \ STMT_START { \ SV *TeMpSv = sv; \ sv_setuv(TeMpSv,i); \ SvSETMAGIC(TeMpSv); \ } STMT_END #endif #ifndef sv_usepvn_mg # define sv_usepvn_mg(sv, ptr, len) \ STMT_START { \ SV *TeMpSv = sv; \ sv_usepvn(TeMpSv,ptr,len); \ SvSETMAGIC(TeMpSv); \ } STMT_END #endif #ifndef SvVSTRING_mg # define SvVSTRING_mg(sv) (SvMAGICAL(sv) ? mg_find(sv, PERL_MAGIC_vstring) : NULL) #endif /* Hint: sv_magic_portable * This is a compatibility function that is only available with * Devel::PPPort. It is NOT in the perl core. * Its purpose is to mimic the 5.8.0 behaviour of sv_magic() when * it is being passed a name pointer with namlen == 0. In that * case, perl 5.8.0 and later store the pointer, not a copy of it. * The compatibility can be provided back to perl 5.004. With * earlier versions, the code will not compile. */ #if (PERL_BCDVERSION < 0x5004000) /* code that uses sv_magic_portable will not compile */ #elif (PERL_BCDVERSION < 0x5008000) # define sv_magic_portable(sv, obj, how, name, namlen) \ STMT_START { \ SV *SvMp_sv = (sv); \ char *SvMp_name = (char *) (name); \ I32 SvMp_namlen = (namlen); \ if (SvMp_name && SvMp_namlen == 0) \ { \ MAGIC *mg; \ sv_magic(SvMp_sv, obj, how, 0, 0); \ mg = SvMAGIC(SvMp_sv); \ mg->mg_len = -42; /* XXX: this is the tricky part */ \ mg->mg_ptr = SvMp_name; \ } \ else \ { \ sv_magic(SvMp_sv, obj, how, SvMp_name, SvMp_namlen); \ } \ } STMT_END #else # define sv_magic_portable(a, b, c, d, e) sv_magic(a, b, c, d, e) #endif #if !defined(mg_findext) #if defined(NEED_mg_findext) static MAGIC * DPPP_(my_mg_findext)(SV * sv, int type, const MGVTBL *vtbl); static #else extern MAGIC * DPPP_(my_mg_findext)(SV * sv, int type, const MGVTBL *vtbl); #endif #if defined(NEED_mg_findext) || defined(NEED_mg_findext_GLOBAL) #define mg_findext DPPP_(my_mg_findext) #define Perl_mg_findext DPPP_(my_mg_findext) MAGIC * DPPP_(my_mg_findext)(SV * sv, int type, const MGVTBL *vtbl) { if (sv) { MAGIC *mg; #ifdef AvPAD_NAMELIST assert(!(SvTYPE(sv) == SVt_PVAV && AvPAD_NAMELIST(sv))); #endif for (mg = SvMAGIC (sv); mg; mg = mg->mg_moremagic) { if (mg->mg_type == type && mg->mg_virtual == vtbl) return mg; } } return NULL; } #endif #endif #if !defined(sv_unmagicext) #if defined(NEED_sv_unmagicext) static int DPPP_(my_sv_unmagicext)(pTHX_ SV * const sv, const int type, MGVTBL * vtbl); static #else extern int DPPP_(my_sv_unmagicext)(pTHX_ SV * const sv, const int type, MGVTBL * vtbl); #endif #if defined(NEED_sv_unmagicext) || defined(NEED_sv_unmagicext_GLOBAL) #ifdef sv_unmagicext # undef sv_unmagicext #endif #define sv_unmagicext(a,b,c) DPPP_(my_sv_unmagicext)(aTHX_ a,b,c) #define Perl_sv_unmagicext DPPP_(my_sv_unmagicext) int DPPP_(my_sv_unmagicext)(pTHX_ SV *const sv, const int type, MGVTBL *vtbl) { MAGIC* mg; MAGIC** mgp; if (SvTYPE(sv) < SVt_PVMG || !SvMAGIC(sv)) return 0; mgp = &(SvMAGIC(sv)); for (mg = *mgp; mg; mg = *mgp) { const MGVTBL* const virt = mg->mg_virtual; if (mg->mg_type == type && virt == vtbl) { *mgp = mg->mg_moremagic; if (virt && virt->svt_free) virt->svt_free(aTHX_ sv, mg); if (mg->mg_ptr && mg->mg_type != PERL_MAGIC_regex_global) { if (mg->mg_len > 0) Safefree(mg->mg_ptr); else if (mg->mg_len == HEf_SVKEY) /* Questionable on older perls... */ SvREFCNT_dec(MUTABLE_SV(mg->mg_ptr)); else if (mg->mg_type == PERL_MAGIC_utf8) Safefree(mg->mg_ptr); } if (mg->mg_flags & MGf_REFCOUNTED) SvREFCNT_dec(mg->mg_obj); Safefree(mg); } else mgp = &mg->mg_moremagic; } if (SvMAGIC(sv)) { if (SvMAGICAL(sv)) /* if we're under save_magic, wait for restore_magic; */ mg_magical(sv); /* else fix the flags now */ } else { SvMAGICAL_off(sv); SvFLAGS(sv) |= (SvFLAGS(sv) & (SVp_IOK|SVp_NOK|SVp_POK)) >> PRIVSHIFT; } return 0; } #endif #endif #ifdef USE_ITHREADS #ifndef CopFILE # define CopFILE(c) ((c)->cop_file) #endif #ifndef CopFILEGV # define CopFILEGV(c) (CopFILE(c) ? gv_fetchfile(CopFILE(c)) : Nullgv) #endif #ifndef CopFILE_set # define CopFILE_set(c,pv) ((c)->cop_file = savepv(pv)) #endif #ifndef CopFILESV # define CopFILESV(c) (CopFILE(c) ? GvSV(gv_fetchfile(CopFILE(c))) : Nullsv) #endif #ifndef CopFILEAV # define CopFILEAV(c) (CopFILE(c) ? GvAV(gv_fetchfile(CopFILE(c))) : Nullav) #endif #ifndef CopSTASHPV # define CopSTASHPV(c) ((c)->cop_stashpv) #endif #ifndef CopSTASHPV_set # define CopSTASHPV_set(c,pv) ((c)->cop_stashpv = ((pv) ? savepv(pv) : Nullch)) #endif #ifndef CopSTASH # define CopSTASH(c) (CopSTASHPV(c) ? gv_stashpv(CopSTASHPV(c),GV_ADD) : Nullhv) #endif #ifndef CopSTASH_set # define CopSTASH_set(c,hv) CopSTASHPV_set(c, (hv) ? HvNAME(hv) : Nullch) #endif #ifndef CopSTASH_eq # define CopSTASH_eq(c,hv) ((hv) && (CopSTASHPV(c) == HvNAME(hv) \ || (CopSTASHPV(c) && HvNAME(hv) \ && strEQ(CopSTASHPV(c), HvNAME(hv))))) #endif #else #ifndef CopFILEGV # define CopFILEGV(c) ((c)->cop_filegv) #endif #ifndef CopFILEGV_set # define CopFILEGV_set(c,gv) ((c)->cop_filegv = (GV*)SvREFCNT_inc(gv)) #endif #ifndef CopFILE_set # define CopFILE_set(c,pv) CopFILEGV_set((c), gv_fetchfile(pv)) #endif #ifndef CopFILESV # define CopFILESV(c) (CopFILEGV(c) ? GvSV(CopFILEGV(c)) : Nullsv) #endif #ifndef CopFILEAV # define CopFILEAV(c) (CopFILEGV(c) ? GvAV(CopFILEGV(c)) : Nullav) #endif #ifndef CopFILE # define CopFILE(c) (CopFILESV(c) ? SvPVX(CopFILESV(c)) : Nullch) #endif #ifndef CopSTASH # define CopSTASH(c) ((c)->cop_stash) #endif #ifndef CopSTASH_set # define CopSTASH_set(c,hv) ((c)->cop_stash = (hv)) #endif #ifndef CopSTASHPV # define CopSTASHPV(c) (CopSTASH(c) ? HvNAME(CopSTASH(c)) : Nullch) #endif #ifndef CopSTASHPV_set # define CopSTASHPV_set(c,pv) CopSTASH_set((c), gv_stashpv(pv,GV_ADD)) #endif #ifndef CopSTASH_eq # define CopSTASH_eq(c,hv) (CopSTASH(c) == (hv)) #endif #endif /* USE_ITHREADS */ #if (PERL_BCDVERSION >= 0x5006000) #ifndef caller_cx # if defined(NEED_caller_cx) || defined(NEED_caller_cx_GLOBAL) static I32 DPPP_dopoptosub_at(const PERL_CONTEXT *cxstk, I32 startingblock) { I32 i; for (i = startingblock; i >= 0; i--) { register const PERL_CONTEXT * const cx = &cxstk[i]; switch (CxTYPE(cx)) { default: continue; case CXt_EVAL: case CXt_SUB: case CXt_FORMAT: return i; } } return i; } # endif # if defined(NEED_caller_cx) static const PERL_CONTEXT * DPPP_(my_caller_cx)(pTHX_ I32 count, const PERL_CONTEXT **dbcxp); static #else extern const PERL_CONTEXT * DPPP_(my_caller_cx)(pTHX_ I32 count, const PERL_CONTEXT **dbcxp); #endif #if defined(NEED_caller_cx) || defined(NEED_caller_cx_GLOBAL) #ifdef caller_cx # undef caller_cx #endif #define caller_cx(a,b) DPPP_(my_caller_cx)(aTHX_ a,b) #define Perl_caller_cx DPPP_(my_caller_cx) const PERL_CONTEXT * DPPP_(my_caller_cx)(pTHX_ I32 count, const PERL_CONTEXT **dbcxp) { register I32 cxix = DPPP_dopoptosub_at(cxstack, cxstack_ix); register const PERL_CONTEXT *cx; register const PERL_CONTEXT *ccstack = cxstack; const PERL_SI *top_si = PL_curstackinfo; for (;;) { /* we may be in a higher stacklevel, so dig down deeper */ while (cxix < 0 && top_si->si_type != PERLSI_MAIN) { top_si = top_si->si_prev; ccstack = top_si->si_cxstack; cxix = DPPP_dopoptosub_at(ccstack, top_si->si_cxix); } if (cxix < 0) return NULL; /* caller() should not report the automatic calls to &DB::sub */ if (PL_DBsub && GvCV(PL_DBsub) && cxix >= 0 && ccstack[cxix].blk_sub.cv == GvCV(PL_DBsub)) count++; if (!count--) break; cxix = DPPP_dopoptosub_at(ccstack, cxix - 1); } cx = &ccstack[cxix]; if (dbcxp) *dbcxp = cx; if (CxTYPE(cx) == CXt_SUB || CxTYPE(cx) == CXt_FORMAT) { const I32 dbcxix = DPPP_dopoptosub_at(ccstack, cxix - 1); /* We expect that ccstack[dbcxix] is CXt_SUB, anyway, the field below is defined for any cx. */ /* caller() should not report the automatic calls to &DB::sub */ if (PL_DBsub && GvCV(PL_DBsub) && dbcxix >= 0 && ccstack[dbcxix].blk_sub.cv == GvCV(PL_DBsub)) cx = &ccstack[dbcxix]; } return cx; } # endif #endif /* caller_cx */ #endif /* 5.6.0 */ #ifndef IN_PERL_COMPILETIME # define IN_PERL_COMPILETIME (PL_curcop == &PL_compiling) #endif #ifndef IN_LOCALE_RUNTIME # define IN_LOCALE_RUNTIME (PL_curcop->op_private & HINT_LOCALE) #endif #ifndef IN_LOCALE_COMPILETIME # define IN_LOCALE_COMPILETIME (PL_hints & HINT_LOCALE) #endif #ifndef IN_LOCALE # define IN_LOCALE (IN_PERL_COMPILETIME ? IN_LOCALE_COMPILETIME : IN_LOCALE_RUNTIME) #endif #ifndef IS_NUMBER_IN_UV # define IS_NUMBER_IN_UV 0x01 #endif #ifndef IS_NUMBER_GREATER_THAN_UV_MAX # define IS_NUMBER_GREATER_THAN_UV_MAX 0x02 #endif #ifndef IS_NUMBER_NOT_INT # define IS_NUMBER_NOT_INT 0x04 #endif #ifndef IS_NUMBER_NEG # define IS_NUMBER_NEG 0x08 #endif #ifndef IS_NUMBER_INFINITY # define IS_NUMBER_INFINITY 0x10 #endif #ifndef IS_NUMBER_NAN # define IS_NUMBER_NAN 0x20 #endif #ifndef GROK_NUMERIC_RADIX # define GROK_NUMERIC_RADIX(sp, send) grok_numeric_radix(sp, send) #endif #ifndef PERL_SCAN_GREATER_THAN_UV_MAX # define PERL_SCAN_GREATER_THAN_UV_MAX 0x02 #endif #ifndef PERL_SCAN_SILENT_ILLDIGIT # define PERL_SCAN_SILENT_ILLDIGIT 0x04 #endif #ifndef PERL_SCAN_ALLOW_UNDERSCORES # define PERL_SCAN_ALLOW_UNDERSCORES 0x01 #endif #ifndef PERL_SCAN_DISALLOW_PREFIX # define PERL_SCAN_DISALLOW_PREFIX 0x02 #endif #ifndef grok_numeric_radix #if defined(NEED_grok_numeric_radix) static bool DPPP_(my_grok_numeric_radix)(pTHX_ const char ** sp, const char * send); static #else extern bool DPPP_(my_grok_numeric_radix)(pTHX_ const char ** sp, const char * send); #endif #if defined(NEED_grok_numeric_radix) || defined(NEED_grok_numeric_radix_GLOBAL) #ifdef grok_numeric_radix # undef grok_numeric_radix #endif #define grok_numeric_radix(a,b) DPPP_(my_grok_numeric_radix)(aTHX_ a,b) #define Perl_grok_numeric_radix DPPP_(my_grok_numeric_radix) bool DPPP_(my_grok_numeric_radix)(pTHX_ const char **sp, const char *send) { #ifdef USE_LOCALE_NUMERIC #ifdef PL_numeric_radix_sv if (PL_numeric_radix_sv && IN_LOCALE) { STRLEN len; char* radix = SvPV(PL_numeric_radix_sv, len); if (*sp + len <= send && memEQ(*sp, radix, len)) { *sp += len; return TRUE; } } #else /* older perls don't have PL_numeric_radix_sv so the radix * must manually be requested from locale.h */ #include dTHR; /* needed for older threaded perls */ struct lconv *lc = localeconv(); char *radix = lc->decimal_point; if (radix && IN_LOCALE) { STRLEN len = strlen(radix); if (*sp + len <= send && memEQ(*sp, radix, len)) { *sp += len; return TRUE; } } #endif #endif /* USE_LOCALE_NUMERIC */ /* always try "." if numeric radix didn't match because * we may have data from different locales mixed */ if (*sp < send && **sp == '.') { ++*sp; return TRUE; } return FALSE; } #endif #endif #ifndef grok_number #if defined(NEED_grok_number) static int DPPP_(my_grok_number)(pTHX_ const char * pv, STRLEN len, UV * valuep); static #else extern int DPPP_(my_grok_number)(pTHX_ const char * pv, STRLEN len, UV * valuep); #endif #if defined(NEED_grok_number) || defined(NEED_grok_number_GLOBAL) #ifdef grok_number # undef grok_number #endif #define grok_number(a,b,c) DPPP_(my_grok_number)(aTHX_ a,b,c) #define Perl_grok_number DPPP_(my_grok_number) int DPPP_(my_grok_number)(pTHX_ const char *pv, STRLEN len, UV *valuep) { const char *s = pv; const char *send = pv + len; const UV max_div_10 = UV_MAX / 10; const char max_mod_10 = UV_MAX % 10; int numtype = 0; int sawinf = 0; int sawnan = 0; while (s < send && isSPACE(*s)) s++; if (s == send) { return 0; } else if (*s == '-') { s++; numtype = IS_NUMBER_NEG; } else if (*s == '+') s++; if (s == send) return 0; /* next must be digit or the radix separator or beginning of infinity */ if (isDIGIT(*s)) { /* UVs are at least 32 bits, so the first 9 decimal digits cannot overflow. */ UV value = *s - '0'; /* This construction seems to be more optimiser friendly. (without it gcc does the isDIGIT test and the *s - '0' separately) With it gcc on arm is managing 6 instructions (6 cycles) per digit. In theory the optimiser could deduce how far to unroll the loop before checking for overflow. */ if (++s < send) { int digit = *s - '0'; if (digit >= 0 && digit <= 9) { value = value * 10 + digit; if (++s < send) { digit = *s - '0'; if (digit >= 0 && digit <= 9) { value = value * 10 + digit; if (++s < send) { digit = *s - '0'; if (digit >= 0 && digit <= 9) { value = value * 10 + digit; if (++s < send) { digit = *s - '0'; if (digit >= 0 && digit <= 9) { value = value * 10 + digit; if (++s < send) { digit = *s - '0'; if (digit >= 0 && digit <= 9) { value = value * 10 + digit; if (++s < send) { digit = *s - '0'; if (digit >= 0 && digit <= 9) { value = value * 10 + digit; if (++s < send) { digit = *s - '0'; if (digit >= 0 && digit <= 9) { value = value * 10 + digit; if (++s < send) { digit = *s - '0'; if (digit >= 0 && digit <= 9) { value = value * 10 + digit; if (++s < send) { /* Now got 9 digits, so need to check each time for overflow. */ digit = *s - '0'; while (digit >= 0 && digit <= 9 && (value < max_div_10 || (value == max_div_10 && digit <= max_mod_10))) { value = value * 10 + digit; if (++s < send) digit = *s - '0'; else break; } if (digit >= 0 && digit <= 9 && (s < send)) { /* value overflowed. skip the remaining digits, don't worry about setting *valuep. */ do { s++; } while (s < send && isDIGIT(*s)); numtype |= IS_NUMBER_GREATER_THAN_UV_MAX; goto skip_value; } } } } } } } } } } } } } } } } } } numtype |= IS_NUMBER_IN_UV; if (valuep) *valuep = value; skip_value: if (GROK_NUMERIC_RADIX(&s, send)) { numtype |= IS_NUMBER_NOT_INT; while (s < send && isDIGIT(*s)) /* optional digits after the radix */ s++; } } else if (GROK_NUMERIC_RADIX(&s, send)) { numtype |= IS_NUMBER_NOT_INT | IS_NUMBER_IN_UV; /* valuep assigned below */ /* no digits before the radix means we need digits after it */ if (s < send && isDIGIT(*s)) { do { s++; } while (s < send && isDIGIT(*s)); if (valuep) { /* integer approximation is valid - it's 0. */ *valuep = 0; } } else return 0; } else if (*s == 'I' || *s == 'i') { s++; if (s == send || (*s != 'N' && *s != 'n')) return 0; s++; if (s == send || (*s != 'F' && *s != 'f')) return 0; s++; if (s < send && (*s == 'I' || *s == 'i')) { s++; if (s == send || (*s != 'N' && *s != 'n')) return 0; s++; if (s == send || (*s != 'I' && *s != 'i')) return 0; s++; if (s == send || (*s != 'T' && *s != 't')) return 0; s++; if (s == send || (*s != 'Y' && *s != 'y')) return 0; s++; } sawinf = 1; } else if (*s == 'N' || *s == 'n') { /* XXX TODO: There are signaling NaNs and quiet NaNs. */ s++; if (s == send || (*s != 'A' && *s != 'a')) return 0; s++; if (s == send || (*s != 'N' && *s != 'n')) return 0; s++; sawnan = 1; } else return 0; if (sawinf) { numtype &= IS_NUMBER_NEG; /* Keep track of sign */ numtype |= IS_NUMBER_INFINITY | IS_NUMBER_NOT_INT; } else if (sawnan) { numtype &= IS_NUMBER_NEG; /* Keep track of sign */ numtype |= IS_NUMBER_NAN | IS_NUMBER_NOT_INT; } else if (s < send) { /* we can have an optional exponent part */ if (*s == 'e' || *s == 'E') { /* The only flag we keep is sign. Blow away any "it's UV" */ numtype &= IS_NUMBER_NEG; numtype |= IS_NUMBER_NOT_INT; s++; if (s < send && (*s == '-' || *s == '+')) s++; if (s < send && isDIGIT(*s)) { do { s++; } while (s < send && isDIGIT(*s)); } else return 0; } } while (s < send && isSPACE(*s)) s++; if (s >= send) return numtype; if (len == 10 && memEQ(pv, "0 but true", 10)) { if (valuep) *valuep = 0; return IS_NUMBER_IN_UV; } return 0; } #endif #endif /* * The grok_* routines have been modified to use warn() instead of * Perl_warner(). Also, 'hexdigit' was the former name of PL_hexdigit, * which is why the stack variable has been renamed to 'xdigit'. */ #ifndef grok_bin #if defined(NEED_grok_bin) static UV DPPP_(my_grok_bin)(pTHX_ const char * start, STRLEN * len_p, I32 * flags, NV * result); static #else extern UV DPPP_(my_grok_bin)(pTHX_ const char * start, STRLEN * len_p, I32 * flags, NV * result); #endif #if defined(NEED_grok_bin) || defined(NEED_grok_bin_GLOBAL) #ifdef grok_bin # undef grok_bin #endif #define grok_bin(a,b,c,d) DPPP_(my_grok_bin)(aTHX_ a,b,c,d) #define Perl_grok_bin DPPP_(my_grok_bin) UV DPPP_(my_grok_bin)(pTHX_ const char *start, STRLEN *len_p, I32 *flags, NV *result) { const char *s = start; STRLEN len = *len_p; UV value = 0; NV value_nv = 0; const UV max_div_2 = UV_MAX / 2; bool allow_underscores = *flags & PERL_SCAN_ALLOW_UNDERSCORES; bool overflowed = FALSE; if (!(*flags & PERL_SCAN_DISALLOW_PREFIX)) { /* strip off leading b or 0b. for compatibility silently suffer "b" and "0b" as valid binary numbers. */ if (len >= 1) { if (s[0] == 'b') { s++; len--; } else if (len >= 2 && s[0] == '0' && s[1] == 'b') { s+=2; len-=2; } } } for (; len-- && *s; s++) { char bit = *s; if (bit == '0' || bit == '1') { /* Write it in this wonky order with a goto to attempt to get the compiler to make the common case integer-only loop pretty tight. With gcc seems to be much straighter code than old scan_bin. */ redo: if (!overflowed) { if (value <= max_div_2) { value = (value << 1) | (bit - '0'); continue; } /* Bah. We're just overflowed. */ warn("Integer overflow in binary number"); overflowed = TRUE; value_nv = (NV) value; } value_nv *= 2.0; /* If an NV has not enough bits in its mantissa to * represent a UV this summing of small low-order numbers * is a waste of time (because the NV cannot preserve * the low-order bits anyway): we could just remember when * did we overflow and in the end just multiply value_nv by the * right amount. */ value_nv += (NV)(bit - '0'); continue; } if (bit == '_' && len && allow_underscores && (bit = s[1]) && (bit == '0' || bit == '1')) { --len; ++s; goto redo; } if (!(*flags & PERL_SCAN_SILENT_ILLDIGIT)) warn("Illegal binary digit '%c' ignored", *s); break; } if ( ( overflowed && value_nv > 4294967295.0) #if UVSIZE > 4 || (!overflowed && value > 0xffffffff ) #endif ) { warn("Binary number > 0b11111111111111111111111111111111 non-portable"); } *len_p = s - start; if (!overflowed) { *flags = 0; return value; } *flags = PERL_SCAN_GREATER_THAN_UV_MAX; if (result) *result = value_nv; return UV_MAX; } #endif #endif #ifndef grok_hex #if defined(NEED_grok_hex) static UV DPPP_(my_grok_hex)(pTHX_ const char * start, STRLEN * len_p, I32 * flags, NV * result); static #else extern UV DPPP_(my_grok_hex)(pTHX_ const char * start, STRLEN * len_p, I32 * flags, NV * result); #endif #if defined(NEED_grok_hex) || defined(NEED_grok_hex_GLOBAL) #ifdef grok_hex # undef grok_hex #endif #define grok_hex(a,b,c,d) DPPP_(my_grok_hex)(aTHX_ a,b,c,d) #define Perl_grok_hex DPPP_(my_grok_hex) UV DPPP_(my_grok_hex)(pTHX_ const char *start, STRLEN *len_p, I32 *flags, NV *result) { const char *s = start; STRLEN len = *len_p; UV value = 0; NV value_nv = 0; const UV max_div_16 = UV_MAX / 16; bool allow_underscores = *flags & PERL_SCAN_ALLOW_UNDERSCORES; bool overflowed = FALSE; const char *xdigit; if (!(*flags & PERL_SCAN_DISALLOW_PREFIX)) { /* strip off leading x or 0x. for compatibility silently suffer "x" and "0x" as valid hex numbers. */ if (len >= 1) { if (s[0] == 'x') { s++; len--; } else if (len >= 2 && s[0] == '0' && s[1] == 'x') { s+=2; len-=2; } } } for (; len-- && *s; s++) { xdigit = strchr((char *) PL_hexdigit, *s); if (xdigit) { /* Write it in this wonky order with a goto to attempt to get the compiler to make the common case integer-only loop pretty tight. With gcc seems to be much straighter code than old scan_hex. */ redo: if (!overflowed) { if (value <= max_div_16) { value = (value << 4) | ((xdigit - PL_hexdigit) & 15); continue; } warn("Integer overflow in hexadecimal number"); overflowed = TRUE; value_nv = (NV) value; } value_nv *= 16.0; /* If an NV has not enough bits in its mantissa to * represent a UV this summing of small low-order numbers * is a waste of time (because the NV cannot preserve * the low-order bits anyway): we could just remember when * did we overflow and in the end just multiply value_nv by the * right amount of 16-tuples. */ value_nv += (NV)((xdigit - PL_hexdigit) & 15); continue; } if (*s == '_' && len && allow_underscores && s[1] && (xdigit = strchr((char *) PL_hexdigit, s[1]))) { --len; ++s; goto redo; } if (!(*flags & PERL_SCAN_SILENT_ILLDIGIT)) warn("Illegal hexadecimal digit '%c' ignored", *s); break; } if ( ( overflowed && value_nv > 4294967295.0) #if UVSIZE > 4 || (!overflowed && value > 0xffffffff ) #endif ) { warn("Hexadecimal number > 0xffffffff non-portable"); } *len_p = s - start; if (!overflowed) { *flags = 0; return value; } *flags = PERL_SCAN_GREATER_THAN_UV_MAX; if (result) *result = value_nv; return UV_MAX; } #endif #endif #ifndef grok_oct #if defined(NEED_grok_oct) static UV DPPP_(my_grok_oct)(pTHX_ const char * start, STRLEN * len_p, I32 * flags, NV * result); static #else extern UV DPPP_(my_grok_oct)(pTHX_ const char * start, STRLEN * len_p, I32 * flags, NV * result); #endif #if defined(NEED_grok_oct) || defined(NEED_grok_oct_GLOBAL) #ifdef grok_oct # undef grok_oct #endif #define grok_oct(a,b,c,d) DPPP_(my_grok_oct)(aTHX_ a,b,c,d) #define Perl_grok_oct DPPP_(my_grok_oct) UV DPPP_(my_grok_oct)(pTHX_ const char *start, STRLEN *len_p, I32 *flags, NV *result) { const char *s = start; STRLEN len = *len_p; UV value = 0; NV value_nv = 0; const UV max_div_8 = UV_MAX / 8; bool allow_underscores = *flags & PERL_SCAN_ALLOW_UNDERSCORES; bool overflowed = FALSE; for (; len-- && *s; s++) { /* gcc 2.95 optimiser not smart enough to figure that this subtraction out front allows slicker code. */ int digit = *s - '0'; if (digit >= 0 && digit <= 7) { /* Write it in this wonky order with a goto to attempt to get the compiler to make the common case integer-only loop pretty tight. */ redo: if (!overflowed) { if (value <= max_div_8) { value = (value << 3) | digit; continue; } /* Bah. We're just overflowed. */ warn("Integer overflow in octal number"); overflowed = TRUE; value_nv = (NV) value; } value_nv *= 8.0; /* If an NV has not enough bits in its mantissa to * represent a UV this summing of small low-order numbers * is a waste of time (because the NV cannot preserve * the low-order bits anyway): we could just remember when * did we overflow and in the end just multiply value_nv by the * right amount of 8-tuples. */ value_nv += (NV)digit; continue; } if (digit == ('_' - '0') && len && allow_underscores && (digit = s[1] - '0') && (digit >= 0 && digit <= 7)) { --len; ++s; goto redo; } /* Allow \octal to work the DWIM way (that is, stop scanning * as soon as non-octal characters are seen, complain only iff * someone seems to want to use the digits eight and nine). */ if (digit == 8 || digit == 9) { if (!(*flags & PERL_SCAN_SILENT_ILLDIGIT)) warn("Illegal octal digit '%c' ignored", *s); } break; } if ( ( overflowed && value_nv > 4294967295.0) #if UVSIZE > 4 || (!overflowed && value > 0xffffffff ) #endif ) { warn("Octal number > 037777777777 non-portable"); } *len_p = s - start; if (!overflowed) { *flags = 0; return value; } *flags = PERL_SCAN_GREATER_THAN_UV_MAX; if (result) *result = value_nv; return UV_MAX; } #endif #endif #if !defined(my_snprintf) #if defined(NEED_my_snprintf) static int DPPP_(my_my_snprintf)(char * buffer, const Size_t len, const char * format, ...); static #else extern int DPPP_(my_my_snprintf)(char * buffer, const Size_t len, const char * format, ...); #endif #if defined(NEED_my_snprintf) || defined(NEED_my_snprintf_GLOBAL) #define my_snprintf DPPP_(my_my_snprintf) #define Perl_my_snprintf DPPP_(my_my_snprintf) int DPPP_(my_my_snprintf)(char *buffer, const Size_t len, const char *format, ...) { dTHX; int retval; va_list ap; va_start(ap, format); #ifdef HAS_VSNPRINTF retval = vsnprintf(buffer, len, format, ap); #else retval = vsprintf(buffer, format, ap); #endif va_end(ap); if (retval < 0 || (len > 0 && (Size_t)retval >= len)) Perl_croak(aTHX_ "panic: my_snprintf buffer overflow"); return retval; } #endif #endif #if !defined(my_sprintf) #if defined(NEED_my_sprintf) static int DPPP_(my_my_sprintf)(char * buffer, const char * pat, ...); static #else extern int DPPP_(my_my_sprintf)(char * buffer, const char * pat, ...); #endif #if defined(NEED_my_sprintf) || defined(NEED_my_sprintf_GLOBAL) #define my_sprintf DPPP_(my_my_sprintf) #define Perl_my_sprintf DPPP_(my_my_sprintf) int DPPP_(my_my_sprintf)(char *buffer, const char* pat, ...) { va_list args; va_start(args, pat); vsprintf(buffer, pat, args); va_end(args); return strlen(buffer); } #endif #endif #ifdef NO_XSLOCKS # ifdef dJMPENV # define dXCPT dJMPENV; int rEtV = 0 # define XCPT_TRY_START JMPENV_PUSH(rEtV); if (rEtV == 0) # define XCPT_TRY_END JMPENV_POP; # define XCPT_CATCH if (rEtV != 0) # define XCPT_RETHROW JMPENV_JUMP(rEtV) # else # define dXCPT Sigjmp_buf oldTOP; int rEtV = 0 # define XCPT_TRY_START Copy(top_env, oldTOP, 1, Sigjmp_buf); rEtV = Sigsetjmp(top_env, 1); if (rEtV == 0) # define XCPT_TRY_END Copy(oldTOP, top_env, 1, Sigjmp_buf); # define XCPT_CATCH if (rEtV != 0) # define XCPT_RETHROW Siglongjmp(top_env, rEtV) # endif #endif #if !defined(my_strlcat) #if defined(NEED_my_strlcat) static Size_t DPPP_(my_my_strlcat)(char * dst, const char * src, Size_t size); static #else extern Size_t DPPP_(my_my_strlcat)(char * dst, const char * src, Size_t size); #endif #if defined(NEED_my_strlcat) || defined(NEED_my_strlcat_GLOBAL) #define my_strlcat DPPP_(my_my_strlcat) #define Perl_my_strlcat DPPP_(my_my_strlcat) Size_t DPPP_(my_my_strlcat)(char *dst, const char *src, Size_t size) { Size_t used, length, copy; used = strlen(dst); length = strlen(src); if (size > 0 && used < size - 1) { copy = (length >= size - used) ? size - used - 1 : length; memcpy(dst + used, src, copy); dst[used + copy] = '\0'; } return used + length; } #endif #endif #if !defined(my_strlcpy) #if defined(NEED_my_strlcpy) static Size_t DPPP_(my_my_strlcpy)(char * dst, const char * src, Size_t size); static #else extern Size_t DPPP_(my_my_strlcpy)(char * dst, const char * src, Size_t size); #endif #if defined(NEED_my_strlcpy) || defined(NEED_my_strlcpy_GLOBAL) #define my_strlcpy DPPP_(my_my_strlcpy) #define Perl_my_strlcpy DPPP_(my_my_strlcpy) Size_t DPPP_(my_my_strlcpy)(char *dst, const char *src, Size_t size) { Size_t length, copy; length = strlen(src); if (size > 0) { copy = (length >= size) ? size - 1 : length; memcpy(dst, src, copy); dst[copy] = '\0'; } return length; } #endif #endif #ifndef PERL_PV_ESCAPE_QUOTE # define PERL_PV_ESCAPE_QUOTE 0x0001 #endif #ifndef PERL_PV_PRETTY_QUOTE # define PERL_PV_PRETTY_QUOTE PERL_PV_ESCAPE_QUOTE #endif #ifndef PERL_PV_PRETTY_ELLIPSES # define PERL_PV_PRETTY_ELLIPSES 0x0002 #endif #ifndef PERL_PV_PRETTY_LTGT # define PERL_PV_PRETTY_LTGT 0x0004 #endif #ifndef PERL_PV_ESCAPE_FIRSTCHAR # define PERL_PV_ESCAPE_FIRSTCHAR 0x0008 #endif #ifndef PERL_PV_ESCAPE_UNI # define PERL_PV_ESCAPE_UNI 0x0100 #endif #ifndef PERL_PV_ESCAPE_UNI_DETECT # define PERL_PV_ESCAPE_UNI_DETECT 0x0200 #endif #ifndef PERL_PV_ESCAPE_ALL # define PERL_PV_ESCAPE_ALL 0x1000 #endif #ifndef PERL_PV_ESCAPE_NOBACKSLASH # define PERL_PV_ESCAPE_NOBACKSLASH 0x2000 #endif #ifndef PERL_PV_ESCAPE_NOCLEAR # define PERL_PV_ESCAPE_NOCLEAR 0x4000 #endif #ifndef PERL_PV_ESCAPE_RE # define PERL_PV_ESCAPE_RE 0x8000 #endif #ifndef PERL_PV_PRETTY_NOCLEAR # define PERL_PV_PRETTY_NOCLEAR PERL_PV_ESCAPE_NOCLEAR #endif #ifndef PERL_PV_PRETTY_DUMP # define PERL_PV_PRETTY_DUMP PERL_PV_PRETTY_ELLIPSES|PERL_PV_PRETTY_QUOTE #endif #ifndef PERL_PV_PRETTY_REGPROP # define PERL_PV_PRETTY_REGPROP PERL_PV_PRETTY_ELLIPSES|PERL_PV_PRETTY_LTGT|PERL_PV_ESCAPE_RE #endif /* Hint: pv_escape * Note that unicode functionality is only backported to * those perl versions that support it. For older perl * versions, the implementation will fall back to bytes. */ #ifndef pv_escape #if defined(NEED_pv_escape) static char * DPPP_(my_pv_escape)(pTHX_ SV * dsv, char const * const str, const STRLEN count, const STRLEN max, STRLEN * const escaped, const U32 flags); static #else extern char * DPPP_(my_pv_escape)(pTHX_ SV * dsv, char const * const str, const STRLEN count, const STRLEN max, STRLEN * const escaped, const U32 flags); #endif #if defined(NEED_pv_escape) || defined(NEED_pv_escape_GLOBAL) #ifdef pv_escape # undef pv_escape #endif #define pv_escape(a,b,c,d,e,f) DPPP_(my_pv_escape)(aTHX_ a,b,c,d,e,f) #define Perl_pv_escape DPPP_(my_pv_escape) char * DPPP_(my_pv_escape)(pTHX_ SV *dsv, char const * const str, const STRLEN count, const STRLEN max, STRLEN * const escaped, const U32 flags) { const char esc = flags & PERL_PV_ESCAPE_RE ? '%' : '\\'; const char dq = flags & PERL_PV_ESCAPE_QUOTE ? '"' : esc; char octbuf[32] = "%123456789ABCDF"; STRLEN wrote = 0; STRLEN chsize = 0; STRLEN readsize = 1; #if defined(is_utf8_string) && defined(utf8_to_uvchr_buf) bool isuni = flags & PERL_PV_ESCAPE_UNI ? 1 : 0; #endif const char *pv = str; const char * const end = pv + count; octbuf[0] = esc; if (!(flags & PERL_PV_ESCAPE_NOCLEAR)) sv_setpvs(dsv, ""); #if defined(is_utf8_string) && defined(utf8_to_uvchr_buf) if ((flags & PERL_PV_ESCAPE_UNI_DETECT) && is_utf8_string((U8*)pv, count)) isuni = 1; #endif for (; pv < end && (!max || wrote < max) ; pv += readsize) { const UV u = #if defined(is_utf8_string) && defined(utf8_to_uvchr_buf) isuni ? utf8_to_uvchr_buf((U8*)pv, end, &readsize) : #endif (U8)*pv; const U8 c = (U8)u & 0xFF; if (u > 255 || (flags & PERL_PV_ESCAPE_ALL)) { if (flags & PERL_PV_ESCAPE_FIRSTCHAR) chsize = my_snprintf(octbuf, sizeof octbuf, "%" UVxf, u); else chsize = my_snprintf(octbuf, sizeof octbuf, "%cx{%" UVxf "}", esc, u); } else if (flags & PERL_PV_ESCAPE_NOBACKSLASH) { chsize = 1; } else { if (c == dq || c == esc || !isPRINT(c)) { chsize = 2; switch (c) { case '\\' : /* fallthrough */ case '%' : if (c == esc) octbuf[1] = esc; else chsize = 1; break; case '\v' : octbuf[1] = 'v'; break; case '\t' : octbuf[1] = 't'; break; case '\r' : octbuf[1] = 'r'; break; case '\n' : octbuf[1] = 'n'; break; case '\f' : octbuf[1] = 'f'; break; case '"' : if (dq == '"') octbuf[1] = '"'; else chsize = 1; break; default: chsize = my_snprintf(octbuf, sizeof octbuf, pv < end && isDIGIT((U8)*(pv+readsize)) ? "%c%03o" : "%c%o", esc, c); } } else { chsize = 1; } } if (max && wrote + chsize > max) { break; } else if (chsize > 1) { sv_catpvn(dsv, octbuf, chsize); wrote += chsize; } else { char tmp[2]; my_snprintf(tmp, sizeof tmp, "%c", c); sv_catpvn(dsv, tmp, 1); wrote++; } if (flags & PERL_PV_ESCAPE_FIRSTCHAR) break; } if (escaped != NULL) *escaped= pv - str; return SvPVX(dsv); } #endif #endif #ifndef pv_pretty #if defined(NEED_pv_pretty) static char * DPPP_(my_pv_pretty)(pTHX_ SV * dsv, char const * const str, const STRLEN count, const STRLEN max, char const * const start_color, char const * const end_color, const U32 flags); static #else extern char * DPPP_(my_pv_pretty)(pTHX_ SV * dsv, char const * const str, const STRLEN count, const STRLEN max, char const * const start_color, char const * const end_color, const U32 flags); #endif #if defined(NEED_pv_pretty) || defined(NEED_pv_pretty_GLOBAL) #ifdef pv_pretty # undef pv_pretty #endif #define pv_pretty(a,b,c,d,e,f,g) DPPP_(my_pv_pretty)(aTHX_ a,b,c,d,e,f,g) #define Perl_pv_pretty DPPP_(my_pv_pretty) char * DPPP_(my_pv_pretty)(pTHX_ SV *dsv, char const * const str, const STRLEN count, const STRLEN max, char const * const start_color, char const * const end_color, const U32 flags) { const U8 dq = (flags & PERL_PV_PRETTY_QUOTE) ? '"' : '%'; STRLEN escaped; if (!(flags & PERL_PV_PRETTY_NOCLEAR)) sv_setpvs(dsv, ""); if (dq == '"') sv_catpvs(dsv, "\""); else if (flags & PERL_PV_PRETTY_LTGT) sv_catpvs(dsv, "<"); if (start_color != NULL) sv_catpv(dsv, D_PPP_CONSTPV_ARG(start_color)); pv_escape(dsv, str, count, max, &escaped, flags | PERL_PV_ESCAPE_NOCLEAR); if (end_color != NULL) sv_catpv(dsv, D_PPP_CONSTPV_ARG(end_color)); if (dq == '"') sv_catpvs(dsv, "\""); else if (flags & PERL_PV_PRETTY_LTGT) sv_catpvs(dsv, ">"); if ((flags & PERL_PV_PRETTY_ELLIPSES) && escaped < count) sv_catpvs(dsv, "..."); return SvPVX(dsv); } #endif #endif #ifndef pv_display #if defined(NEED_pv_display) static char * DPPP_(my_pv_display)(pTHX_ SV * dsv, const char * pv, STRLEN cur, STRLEN len, STRLEN pvlim); static #else extern char * DPPP_(my_pv_display)(pTHX_ SV * dsv, const char * pv, STRLEN cur, STRLEN len, STRLEN pvlim); #endif #if defined(NEED_pv_display) || defined(NEED_pv_display_GLOBAL) #ifdef pv_display # undef pv_display #endif #define pv_display(a,b,c,d,e) DPPP_(my_pv_display)(aTHX_ a,b,c,d,e) #define Perl_pv_display DPPP_(my_pv_display) char * DPPP_(my_pv_display)(pTHX_ SV *dsv, const char *pv, STRLEN cur, STRLEN len, STRLEN pvlim) { pv_pretty(dsv, pv, cur, pvlim, NULL, NULL, PERL_PV_PRETTY_DUMP); if (len > cur && pv[cur] == '\0') sv_catpvs(dsv, "\\0"); return SvPVX(dsv); } #endif #endif #endif /* _P_P_PORTABILITY_H_ */ /* End of File ppport.h */ Crypt-xxHash-0.09/t000755001750001750 015146654120 15676 5ustar00dchernenkodchernenko000000000000Crypt-xxHash-0.09/t/Crypt-xxHash.t000444001750001750 1132215146654120 20601 0ustar00dchernenkodchernenko000000000000package main; use utf8; use strict; use warnings; use lib '../lib', 'blib', 'lib'; use Test::More ('import' => [qw/ is done_testing /]);; use Math::Int64 qw/ uint64 hex_to_uint64 /; no warnings 'portable'; # Support for 64-bit ints required use Crypt::xxHash qw/ xxhash32 xxhash32_hex xxhash64 xxhash64_hex xxhash3_64bits xxhash3_64bits_hex xxhash3_128bits_hex xxhash3_64bits_stream xxhash3_64bits_stream_update xxhash3_64bits_stream_digest xxhash3_64bits_stream_digest_hex /; sub testSequence; sub testSequence64; my $b1 = join '', map {chr} 0xB8, 0x1E, 0x85, 0xEB, 0x51, 0xB8, 0x9E, 0x3F, 0xB8, 0x1E, 0x85, 0xEB, 0x51, 0xB8, 0x9E, 0x3F, 0xB8, 0x1E, 0x85, 0xEB, 0x51, 0xB8, 0x9E, 0x3F, 0xB8, 0x1E, 0x85, 0xEB, 0x51, 0xB8, 0x9E, 0x3F, 0xB8, 0x1E, 0x85, 0xEB, 0x51, 0xB8, 0x9E, 0x3F, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x50, 0xC3, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x20, 0x13, 0x08, 0x12, 0x65, 0xE3, 0x6A, 0xC0; my $b2 = join '', map {chr} 0xD7, 0xA3, 0x70, 0x3D, 0x0A, 0x57, 0x21, 0x40, 0x9A, 0x99, 0x99, 0x99, 0x99, 0x99, 0x21, 0x40, 0xA4, 0x70, 0x3D, 0x0A, 0xD7, 0x23, 0x21, 0x40, 0x14, 0xAE, 0x47, 0xE1, 0x7A, 0x94, 0x21, 0x40, 0x14, 0xAE, 0x47, 0xE1, 0x7A, 0x94, 0x21, 0x40, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0xD8, 0x3C, 0x09, 0x00, 0x00, 0x00, 0x00, 0x00, 0x20, 0x13, 0x08, 0x13, 0xA9, 0xF1, 0xE2, 0x2A; # 32bit is( xxhash32($b1, 0x5262), xxhash32($b2, 0x5262), 'Known 32 bit collision' ); is( xxhash32('test', 123), 2758658570, 'basic' ); is( xxhash32('test', 12345), 3834992036, 'basic w/ different seed' ); is( xxhash32_hex('test', 12345), 'e49555a4', 'xxhash32_hex' ); # 64bit is( uint64(xxhash64('test64', 1123)), uint64('18300740539230391133') ); is( uint64(xxhash64('test64', 5813)), uint64('470519085964830776') ); is( uint64(xxhash3_64bits('test64', 1123)), uint64('855843287810803924') ); is( uint64(xxhash3_64bits('test64', 5813)), uint64('9266226278152659498') ); is( xxhash64_hex('test64', 5813), '06879e9da2f9a838', 'xxhash64_hex' ); is( xxhash3_64bits_hex('test64', 5813), '80983fb495ade22a', 'xxhash3_64bits_hex' ); my $SANITY_BUFFER_SIZE = 101; my $prime = 2654435761; my $sanityBuffer = pack('H*', '9eff1f4b5e532fddb5544d2a952b57ae5dba74e9d3a64c983060c080'); testSequence('', 0, 0, 0x02CC5D05); testSequence('', 0, $prime, 0x36B78AE7); testSequence($sanityBuffer, 1, 0, 0xB85CBEE5); testSequence($sanityBuffer, 1, $prime, 0xD5845D64); testSequence($sanityBuffer, 14, 0, 0xE5AA0AB4); testSequence($sanityBuffer, 14, $prime, 0x4481951D); testSequence($sanityBuffer, $SANITY_BUFFER_SIZE, 0, 0x1F1AA412); testSequence($sanityBuffer, $SANITY_BUFFER_SIZE, $prime, 0x498EC8E2); testSequence64('', 0, 0, 'EF46DB3751D8E999'); testSequence64('', 0, $prime, 'AC75FDA2929B17EF'); testSequence64($sanityBuffer, 1, 0, '4FCE394CC88952D8'); testSequence64($sanityBuffer, 1, $prime, '739840CB819FA723'); testSequence64($sanityBuffer, 14, 0, 'CFFA8DB881BC3A3D'); testSequence64($sanityBuffer, 14, $prime, '5B9611585EFCC9CB'); testSequence64($sanityBuffer, $SANITY_BUFFER_SIZE, 0, '0EAB543384F878AD'); testSequence64($sanityBuffer, $SANITY_BUFFER_SIZE, $prime, 'CAA65939306F1E21'); is( xxhash64_hex("b" x 100000, 890272), 'df8fee94dbf20a9d', 'uint64' ); is( xxhash64_hex("b" x 100000, 89), '01aae2582443bbf0', 'expect leading zeros' ); { my $stream = xxhash3_64bits_stream(1123); xxhash3_64bits_stream_update($stream, 'te'); xxhash3_64bits_stream_update($stream, 'st'); xxhash3_64bits_stream_update($stream, '64'); is( xxhash3_64bits_stream_digest($stream), uint64('18300740539230391133'), 'uint64 streaming' ); } { my $stream = xxhash3_64bits_stream(890272); xxhash3_64bits_stream_update($stream, "b" x 1000) for(1..100); is( xxhash3_64bits_stream_digest_hex($stream), 'df8fee94dbf20a9d', 'uint64 streaming hex' ); } { my $stream = xxhash3_64bits_stream(89); xxhash3_64bits_stream_update($stream, "b" x 1000) for(1..100); is( xxhash3_64bits_stream_digest_hex($stream), '01aae2582443bbf0', 'uint64 streaming hex leading zeros' ); } # 128 bits is( xxhash3_128bits_hex('test64', 5813), '2675a6f40fe52f19601fc692c37d6e4d', 'xxhash3_128bits_hex' ); # TODO: 128, 256 done_testing; #### sub testSequence { my ($sentence, $len, $seed, $result) = @_; is xxhash32(pack('a' . $len, $sentence), $seed), $result, 'line ' . (caller())[2]; } sub testSequence64 { my ($sentence, $len, $seed, $result) = @_; is uint64(xxhash64(pack('a' . $len, $sentence), $seed)), hex_to_uint64($result), 'line ' . (caller())[2]; } 1; __END__ Crypt-xxHash-0.09/xs000755001750001750 015146654120 16065 5ustar00dchernenkodchernenko000000000000Crypt-xxHash-0.09/xs/xxHash.xs000444001750001750 1317115146654120 20064 0ustar00dchernenkodchernenko000000000000// PERL_NO_GET_CONTEXT is not used here, so it's OK to define it after inculding these files #include "EXTERN.h" #include "perl.h" // There are a lot of macro about threads: USE_ITHREADS, USE_5005THREADS, I_PTHREAD, I_MACH_CTHREADS, OLD_PTHREADS_API // This symbol, if defined, indicates that Perl should be built to use the interpreter-based threading implementation. #ifndef USE_ITHREADS # define PERL_NO_GET_CONTEXT #endif #include "XSUB.h" #include "ppport.h" #ifdef I_PTHREAD # include "pthread.h" #endif #ifdef I_MACH_CTHREADS # include "mach/cthreads.h" #endif /* define int64_t and uint64_t when using MinGW compiler */ #ifdef __MINGW32__ #include #endif /* define int64_t and uint64_t when using MS compiler */ #ifdef _MSC_VER #include typedef __int64 int64_t; typedef unsigned __int64 uint64_t; #endif #include "xxhash.h" #include #include #ifdef HAS_INT64_T #endif #ifdef CPP using namespace std; #endif static int canceller_free (pTHX_ SV *cv, MAGIC *mg) { XXH64_freeState((XXH64_state_t*)mg->mg_ptr); return 0; } static MGVTBL canceller_vtbl = { 0, 0, 0, 0, canceller_free }; MODULE = Crypt::xxHash PACKAGE = Crypt::xxHash PROTOTYPES: DISABLE U32 xxhash32( const char *input, int length(input), UV seed ) CODE: #ifdef CPP RETVAL = XXH32(input, STRLEN_length_of_input, move(seed)); #else RETVAL = XXH32(input, STRLEN_length_of_input, seed); #endif OUTPUT: RETVAL UV xxhash64( const char *input, int length(input), UV seed ) CODE: #ifdef CPP RETVAL = (UV) XXH64(input, STRLEN_length_of_input, move(seed)); #else RETVAL = (UV) XXH64(input, STRLEN_length_of_input, seed); #endif OUTPUT: RETVAL UV xxhash3_64bits( const char *input, int length(input), UV seed ) CODE: #ifdef CPP RETVAL = (UV) XXH3_64bits_withSeed(input, STRLEN_length_of_input, move(seed)); #else RETVAL = (UV) XXH3_64bits_withSeed(input, STRLEN_length_of_input, seed); #endif OUTPUT: RETVAL char* xxhash32_hex ( const char *input, int length(input), UV seed ) CODE: static char value32[9]; static const char *format = "%08x"; #ifdef CPP sprintf(value32, format, (uint32_t) XXH32(input, STRLEN_length_of_input, move(seed)) ); #else sprintf(value32, format, (uint32_t) XXH32(input, STRLEN_length_of_input, seed) ); #endif RETVAL = value32; OUTPUT: RETVAL char* xxhash64_hex( const char *input, int length(input), UV seed ) CODE: static char value64[17]; static const char *format = "%016" PRIx64; #ifdef CPP sprintf(value64, format, (uint64_t) XXH64(input, STRLEN_length_of_input, move(seed)) ); #else sprintf(value64, format, (uint64_t) XXH64(input, STRLEN_length_of_input, seed) ); #endif RETVAL = value64; OUTPUT: RETVAL char* xxhash3_64bits_hex( const char *input, int length(input), UV seed ) CODE: static char value64[17]; static const char *format = "%016" PRIx64; #ifdef CPP sprintf(value64, format, (uint64_t) XXH3_64bits_withSeed(input, STRLEN_length_of_input, move(seed)) ); #else sprintf(value64, format, (uint64_t) XXH3_64bits_withSeed(input, STRLEN_length_of_input, seed) ); #endif RETVAL = value64; OUTPUT: RETVAL char* xxhash3_128bits_hex( const char *input, int length(input), UV seed ) CODE: static char value64[33]; static const char *format = "%016" PRIx64 "%016" PRIx64; #ifdef CPP XXH128_hash_t hash = XXH3_128bits_withSeed(input, STRLEN_length_of_input, move(seed)); #else XXH128_hash_t hash = XXH3_128bits_withSeed(input, STRLEN_length_of_input, seed); #endif sprintf(value64, format, (uint64_t) hash.high64, (uint64_t) hash.low64 ); RETVAL = value64; OUTPUT: RETVAL void xxhash3_64bits_stream( UV seed ) PPCODE: XXH64_state_t * state = XXH64_createState(); if( !state ) croak("Allocate 64bits xxHash state failed."); if( XXH64_reset(state, seed) == XXH_ERROR ) croak("Initialize 64bits xxHash state failed."); SV *canceller = NEWSV(0, 0); SvUPGRADE(canceller, SVt_PVMG); sv_magicext(canceller, NULL, PERL_MAGIC_ext, &canceller_vtbl, (const char*)state, 0); mPUSHs(newRV_noinc(canceller)); void xxhash3_64bits_stream_update(SV * state_RV, SV * data_SV) PPCODE: if( !SvROK(state_RV) ) croak("This is not a 64bits xxHash state variable."); MAGIC * mg = mg_findext(SvRV(state_RV), PERL_MAGIC_ext, &canceller_vtbl); if( !mg ) croak("This is not a 64bits xxHash state variable."); STRLEN len; char * buf = SvPV(data_SV, len); if( XXH64_update((XXH64_state_t*)mg->mg_ptr, buf, len) == XXH_ERROR ) croak("Update 64bits xxHash state failed."); void xxhash3_64bits_stream_digest(SV * state_RV) PPCODE: dXSTARG; if( !SvROK(state_RV) ) croak("This is not a 64bits xxHash state variable."); MAGIC * mg = mg_findext(SvRV(state_RV), PERL_MAGIC_ext, &canceller_vtbl); if( !mg ) croak("This is not a 64bits xxHash state variable."); PUSHu((UV) XXH64_digest((XXH64_state_t*)mg->mg_ptr)); void xxhash3_64bits_stream_digest_hex(SV * state_RV) PPCODE: if( !SvROK(state_RV) ) croak("This is not a 64bits xxHash state variable."); MAGIC * mg = mg_findext(SvRV(state_RV), PERL_MAGIC_ext, &canceller_vtbl); if( !mg ) croak("This is not a 64bits xxHash state variable."); mPUSHs(newSVpvf("%016" PRIx64, (uint64_t)XXH64_digest((XXH64_state_t*)mg->mg_ptr)));