diff options
| author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-27 18:24:20 +0000 |
|---|---|---|
| committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-27 18:24:20 +0000 |
| commit | 483eb2f56657e8e7f419ab1a4fab8dce9ade8609 (patch) | |
| tree | e5d88d25d870d5dedacb6bbdbe2a966086a0a5cf /src/boost/libs/beast/test/bench | |
| parent | Initial commit. (diff) | |
| download | ceph-483eb2f56657e8e7f419ab1a4fab8dce9ade8609.tar.xz ceph-483eb2f56657e8e7f419ab1a4fab8dce9ade8609.zip | |
Adding upstream version 14.2.21.upstream/14.2.21upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'src/boost/libs/beast/test/bench')
25 files changed, 5620 insertions, 0 deletions
diff --git a/src/boost/libs/beast/test/bench/CMakeLists.txt b/src/boost/libs/beast/test/bench/CMakeLists.txt new file mode 100644 index 00000000..849f032a --- /dev/null +++ b/src/boost/libs/beast/test/bench/CMakeLists.txt @@ -0,0 +1,14 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +add_subdirectory (buffers) +add_subdirectory (parser) +add_subdirectory (utf8_checker) +add_subdirectory (wsload) +add_subdirectory (zlib) diff --git a/src/boost/libs/beast/test/bench/Jamfile b/src/boost/libs/beast/test/bench/Jamfile new file mode 100644 index 00000000..44bae113 --- /dev/null +++ b/src/boost/libs/beast/test/bench/Jamfile @@ -0,0 +1,16 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +alias run-tests : + buffers//run-tests + parser//run-tests + wsload//run-tests + utf8_checker//run-tests + #zlib//run-tests # Not built, too slow + ; diff --git a/src/boost/libs/beast/test/bench/buffers/CMakeLists.txt b/src/boost/libs/beast/test/bench/buffers/CMakeLists.txt new file mode 100644 index 00000000..d05db11f --- /dev/null +++ b/src/boost/libs/beast/test/bench/buffers/CMakeLists.txt @@ -0,0 +1,25 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +GroupSources (include/boost/beast beast) +GroupSources (test/bench/buffers "/") + +add_executable (bench-buffers + ${BOOST_BEAST_FILES} + Jamfile + bench_buffers.cpp +) + +target_link_libraries(bench-buffers + lib-asio + lib-beast + lib-test + ) + +set_property(TARGET bench-buffers PROPERTY FOLDER "tests-bench") diff --git a/src/boost/libs/beast/test/bench/buffers/Jamfile b/src/boost/libs/beast/test/bench/buffers/Jamfile new file mode 100644 index 00000000..c73d8a11 --- /dev/null +++ b/src/boost/libs/beast/test/bench/buffers/Jamfile @@ -0,0 +1,19 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +exe bench-buffers : + bench_buffers.cpp + /boost/beast/test//lib-test + ; + +explicit bench-buffers ; + +alias run-tests : + [ compile bench_buffers.cpp ] + ; diff --git a/src/boost/libs/beast/test/bench/buffers/bench_buffers.cpp b/src/boost/libs/beast/test/bench/buffers/bench_buffers.cpp new file mode 100644 index 00000000..fdd00a47 --- /dev/null +++ b/src/boost/libs/beast/test/bench/buffers/bench_buffers.cpp @@ -0,0 +1,241 @@ +// +// Copyright (c) 2016-2019 Vinnie Falco (vinnie dot falco at gmail dot com) +// +// Distributed under the Boost Software License, Version 1.0. (See accompanying +// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +// +// Official repository: https://github.com/boostorg/beast +// + +#include <boost/beast/core/buffers_range.hpp> +#include <boost/beast/core/flat_buffer.hpp> +#include <boost/beast/core/multi_buffer.hpp> +#include <boost/beast/core/read_size.hpp> +#include <boost/beast/core/string.hpp> +#include <boost/beast/_experimental/unit_test/suite.hpp> +#include <boost/asio/streambuf.hpp> +#include <algorithm> +#include <chrono> +#include <cstdlib> +#include <iomanip> +#include <utility> +#include <vector> + +namespace boost { +namespace beast { + +class buffers_test : public beast::unit_test::suite +{ +public: + using size_type = std::uint64_t; + + class timer + { + using clock_type = + std::chrono::system_clock; + + clock_type::time_point when_; + + public: + using duration = + clock_type::duration; + + timer() + : when_(clock_type::now()) + { + } + + duration + elapsed() const + { + return clock_type::now() - when_; + } + }; + + inline + size_type + throughput(std::chrono::duration< + double> const& elapsed, size_type items) + { + using namespace std::chrono; + return static_cast<size_type>( + 1 / (elapsed/items).count()); + } + + template<class MutableBufferSequence> + static + std::size_t + fill(MutableBufferSequence const& buffers) + { + std::size_t n = 0; + for(auto b : beast::buffers_range_ref(buffers)) + { + std::fill( + static_cast<char*>(b.data()), + static_cast<char*>(b.data()) + + b.size(), '\0'); + n += b.size(); + } + return n; + } + + template<class DynamicBuffer> + size_type + do_prepares(std::size_t repeat, + std::size_t count, std::size_t size) + { + timer t; + size_type total = 0; + for(auto i = repeat; i--;) + { + DynamicBuffer b; + for(auto j = count; j--;) + { + auto const n = fill(b.prepare(size)); + b.commit(n); + total += n; + } + } + return throughput(t.elapsed(), total); + } + + template<class DynamicBuffer> + size_type + do_hints(std::size_t repeat, + std::size_t count, std::size_t size) + { + timer t; + size_type total = 0; + for(auto i = repeat; i--;) + { + DynamicBuffer b; + for(auto j = count; j--;) + { + for(auto remain = size; remain;) + { + auto const n = fill(b.prepare( + read_size(b, remain))); + b.commit(n); + remain -= n; + total += n; + } + } + } + return throughput(t.elapsed(), total); + } + + template<class DynamicBuffer> + size_type + do_random(std::size_t repeat, + std::size_t count, std::size_t size) + { + timer t; + size_type total = 0; + for(auto i = repeat; i--;) + { + DynamicBuffer b; + for(auto j = count; j--;) + { + auto const n = fill(b.prepare( + 1 + (rand()%(2*size)))); + b.commit(n); + total += n; + } + } + return throughput(t.elapsed(), total); + } + + static + inline + void + do_trials_1(bool) + { + } + + template<class F0, class... FN> + void + do_trials_1(bool print, F0&& f, FN... fn) + { + timer t; + using namespace std::chrono; + static size_type constexpr den = 1024 * 1024; + if(print) + { + log << std::right << std::setw(10) << + ((f() + (den / 2)) / den) << " MB/s"; + log.flush(); + } + else + { + f(); + } + do_trials_1(print, fn...); + } + + template<class F0, class... FN> + void + do_trials(string_view name, + std::size_t trials, F0&& f0, FN... fn) + { + using namespace std::chrono; + // warm-up + do_trials_1(false, f0, fn...); + do_trials_1(false, f0, fn...); + while(trials--) + { + timer t; + log << std::left << std::setw(24) << name << ":"; + log.flush(); + do_trials_1(true, f0, fn...); + log << " " << + duration_cast<milliseconds>(t.elapsed()).count() << "ms"; + log << std::endl; + } + } + + void + run() override + { + static std::size_t constexpr trials = 1; + static std::size_t constexpr repeat = 250; + std::vector<std::pair<std::size_t, std::size_t>> params; + params.emplace_back(1024, 1024); + params.emplace_back(512, 4096); + params.emplace_back(256, 32768); + log << std::endl; + for(auto const& param : params) + { + auto const count = param.first; + auto const size = param.second; + auto const s = std::string("count=") + std::to_string(count) + + ", size=" + std::to_string(size); + log << std::left << std::setw(24) << s << " " << + std::right << std::setw(15) << "prepare" << + std::right << std::setw(15) << "with hint" << + std::right << std::setw(15) << "random" << + std::endl; + do_trials("multi_buffer", trials, + [&](){ return do_prepares<multi_buffer>(repeat, count, size); } + ,[&](){ return do_hints <multi_buffer>(repeat, count, size); } + ,[&](){ return do_random <multi_buffer>(repeat, count, size); } + ); + do_trials("flat_buffer", trials, + [&](){ return do_prepares<flat_buffer>(repeat, count, size); } + ,[&](){ return do_hints <flat_buffer>(repeat, count, size); } + ,[&](){ return do_random <flat_buffer>(repeat, count, size); } + ); + do_trials("net::streambuf", trials, + [&](){ return do_prepares<net::streambuf>(repeat, count, size); } + ,[&](){ return do_hints <net::streambuf>(repeat, count, size); } + ,[&](){ return do_random <net::streambuf>(repeat, count, size); } + ); + log << std::endl; + } + pass(); + } +}; + +BEAST_DEFINE_TESTSUITE(beast,benchmarks,buffers); + +} // beast +} // boost diff --git a/src/boost/libs/beast/test/bench/parser/CMakeLists.txt b/src/boost/libs/beast/test/bench/parser/CMakeLists.txt new file mode 100644 index 00000000..4a4eb2ce --- /dev/null +++ b/src/boost/libs/beast/test/bench/parser/CMakeLists.txt @@ -0,0 +1,29 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +GroupSources (include/boost/beast beast) +GroupSources (test/beast/http "/") +GroupSources (test/bench/parser "/") + +add_executable (bench-parser + ${BOOST_BEAST_FILES} + Jamfile + ${PROJECT_SOURCE_DIR}/test/beast/http/message_fuzz.hpp + nodejs_parser.hpp + nodejs_parser.cpp + bench_parser.cpp +) + +target_link_libraries(bench-parser + lib-asio + lib-beast + lib-test + ) + +set_property(TARGET bench-parser PROPERTY FOLDER "tests-bench") diff --git a/src/boost/libs/beast/test/bench/parser/Jamfile b/src/boost/libs/beast/test/bench/parser/Jamfile new file mode 100644 index 00000000..43629fc3 --- /dev/null +++ b/src/boost/libs/beast/test/bench/parser/Jamfile @@ -0,0 +1,22 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +run + nodejs_parser.cpp + bench_parser.cpp + /boost/beast/test//lib-test + : : : : + bench-parser ; + +explicit bench-parser ; + +alias run-tests : + [ compile nodejs_parser.cpp ] + [ compile bench_parser.cpp ] + ; diff --git a/src/boost/libs/beast/test/bench/parser/bench_parser.cpp b/src/boost/libs/beast/test/bench/parser/bench_parser.cpp new file mode 100644 index 00000000..67cc99a0 --- /dev/null +++ b/src/boost/libs/beast/test/bench/parser/bench_parser.cpp @@ -0,0 +1,346 @@ +// +// Copyright (c) 2016-2019 Vinnie Falco (vinnie dot falco at gmail dot com) +// +// Distributed under the Boost Software License, Version 1.0. (See accompanying +// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +// +// Official repository: https://github.com/boostorg/beast +// + +#include "nodejs_parser.hpp" + +#include "test/beast/http/message_fuzz.hpp" + +#include <boost/beast/http.hpp> +#include <boost/beast/core/buffer_traits.hpp> +#include <boost/beast/core/buffers_suffix.hpp> +#include <boost/beast/core/buffers_to_string.hpp> +#include <boost/beast/core/ostream.hpp> +#include <boost/beast/core/flat_buffer.hpp> +#include <boost/beast/core/multi_buffer.hpp> +#include <boost/beast/_experimental/unit_test/suite.hpp> +#include <chrono> +#include <iostream> +#include <vector> + +namespace boost { +namespace beast { +namespace http { + +class parser_test : public beast::unit_test::suite +{ +public: + static std::size_t constexpr N = 2000; + + //using corpus = std::vector<multi_buffer>; + using corpus = std::vector<flat_buffer>; + + corpus creq_; + corpus cres_; + std::size_t size_ = 0; + + corpus + build_corpus(std::size_t n, std::true_type) + { + corpus v; + v.resize(n); + message_fuzz mg; + for(std::size_t i = 0; i < n; ++i) + { + mg.request(v[i]); + size_ += v[i].size(); + BEAST_EXPECT(v[i].size() > 0); + } + return v; + } + + corpus + build_corpus(std::size_t n, std::false_type) + { + corpus v; + v.resize(n); + message_fuzz mg; + for(std::size_t i = 0; i < n; ++i) + { + mg.response(v[i]); + size_ += v[i].size(); + BEAST_EXPECT(v[i].size() > 0); + } + return v; + } + + template<class ConstBufferSequence, + bool isRequest> + static + std::size_t + feed(ConstBufferSequence const& buffers, + basic_parser<isRequest>& parser, + error_code& ec) + { + beast::buffers_suffix< + ConstBufferSequence> cb{buffers}; + std::size_t used = 0; + for(;;) + { + auto const n = + parser.put(cb, ec); + if(ec) + return 0; + if(n == 0) + break; + cb.consume(n); + used += n; + if(parser.is_done()) + break; + if(buffer_bytes(cb) == 0) + break; + } + return used; + } + + template<class Parser> + void + testParser1(std::size_t repeat, corpus const& v) + { + while(repeat--) + for(auto const& b : v) + { + Parser p; + error_code ec; + p.write(b.data(), ec); + if(! BEAST_EXPECTS(! ec, ec.message())) + log << buffers_to_string(b.data()) << std::endl; + } + } + + template<class Parser> + void + testParser2(std::size_t repeat, corpus const& v) + { + while(repeat--) + for(auto const& b : v) + { + Parser p; + p.header_limit((std::numeric_limits<std::uint32_t>::max)()); + error_code ec; + feed(b.data(), p, ec); + if(! BEAST_EXPECTS(! ec, ec.message())) + log << buffers_to_string(b.data()) << std::endl; + } + } + + template<class Function> + void + timedTest(std::size_t repeat, std::string const& name, Function&& f) + { + using namespace std::chrono; + using clock_type = std::chrono::high_resolution_clock; + log << name << std::endl; + for(std::size_t trial = 1; trial <= repeat; ++trial) + { + auto const t0 = clock_type::now(); + f(); + auto const elapsed = clock_type::now() - t0; + log << + "Trial " << trial << ": " << + duration_cast<milliseconds>(elapsed).count() << " ms" << std::endl; + } + } + + template<bool isRequest> + struct null_parser : + basic_parser<isRequest> + { + void + on_request_impl( + verb, string_view, string_view, + int, error_code&) override + { + } + + void + on_response_impl( + int, string_view, int, + error_code&) override + { + } + + void + on_field_impl( + field, string_view, string_view, + error_code&) override + { + } + + void + on_header_impl(error_code&) override + { + } + + void + on_body_init_impl( + boost::optional<std::uint64_t> const&, + error_code&) override + { + } + + std::size_t + on_body_impl( + string_view, + error_code&) override + { + return 0; + } + + void + on_chunk_header_impl( + std::uint64_t, + string_view, + error_code&) override + { + } + + std::size_t + on_chunk_body_impl( + std::uint64_t, + string_view, + error_code&) override + { + return 0; + } + + void + on_finish_impl(error_code&) override + { + } + }; + + template<bool isRequest, class Body, class Fields> + struct bench_parser : basic_parser<isRequest> + { + using mutable_buffers_type = + net::mutable_buffer; + + void + on_request_impl(verb, string_view, + string_view, int, error_code&) override + { + } + + void + on_response_impl(int, + string_view, int, error_code&) override + { + } + + void + on_field_impl(field, + string_view, string_view, error_code&) override + { + } + + void + on_header_impl(error_code&) override + { + } + + void + on_body_init_impl( + boost::optional<std::uint64_t> const&, + error_code&) override + { + } + + std::size_t + on_body_impl( + string_view s, error_code&) override + { + return s.size(); + } + + void + on_chunk_header_impl(std::uint64_t, + string_view, error_code&) override + { + } + + std::size_t + on_chunk_body_impl(std::uint64_t, + string_view s, error_code&) override + { + return s.size(); + } + + void + on_finish_impl(error_code&) override + { + } + }; + + void + testSpeed() + { + static std::size_t constexpr Trials = 5; + static std::size_t constexpr Repeat = 500; + + creq_ = build_corpus(N/2, std::true_type{}); + cres_ = build_corpus(N/2, std::false_type{}); + + log << "sizeof(request parser) == " << + sizeof(null_parser<true>) << '\n'; + + log << "sizeof(response parser) == " << + sizeof(null_parser<false>)<< '\n'; + + testcase << "Parser speed test, " << + ((Repeat * size_ + 512) / 1024) << "KB in " << + (Repeat * (creq_.size() + cres_.size())) << " messages"; + +#if 0 + timedTest(Trials, "http::parser", + [&] + { + testParser2<request_parser<dynamic_body>>(Repeat, creq_); + testParser2<response_parser<dynamic_body>>(Repeat, cres_); + }); +#endif +#if 1 + timedTest(Trials, "http::basic_parser", + [&] + { + testParser2<bench_parser< + true, dynamic_body, fields> >( + Repeat, creq_); + testParser2<bench_parser< + false, dynamic_body, fields>>( + Repeat, cres_); + }); +#if 1 + timedTest(Trials, "nodejs_parser", + [&] + { + testParser1<nodejs_parser< + true, dynamic_body, fields>>( + Repeat, creq_); + testParser1<nodejs_parser< + false, dynamic_body, fields>>( + Repeat, cres_); + }); +#endif +#endif + pass(); + } + + void run() override + { + pass(); + testSpeed(); + } +}; + +BEAST_DEFINE_TESTSUITE(beast,benchmarks,parser); + +} // http +} // beast +} // boost + diff --git a/src/boost/libs/beast/test/bench/parser/nodejs-parser/AUTHORS b/src/boost/libs/beast/test/bench/parser/nodejs-parser/AUTHORS new file mode 100644 index 00000000..5323b685 --- /dev/null +++ b/src/boost/libs/beast/test/bench/parser/nodejs-parser/AUTHORS @@ -0,0 +1,68 @@ +# Authors ordered by first contribution. +Ryan Dahl <ry@tinyclouds.org> +Jeremy Hinegardner <jeremy@hinegardner.org> +Sergey Shepelev <temotor@gmail.com> +Joe Damato <ice799@gmail.com> +tomika <tomika_nospam@freemail.hu> +Phoenix Sol <phoenix@burninglabs.com> +Cliff Frey <cliff@meraki.com> +Ewen Cheslack-Postava <ewencp@cs.stanford.edu> +Santiago Gala <sgala@apache.org> +Tim Becker <tim.becker@syngenio.de> +Jeff Terrace <jterrace@gmail.com> +Ben Noordhuis <info@bnoordhuis.nl> +Nathan Rajlich <nathan@tootallnate.net> +Mark Nottingham <mnot@mnot.net> +Aman Gupta <aman@tmm1.net> +Tim Becker <tim.becker@kuriositaet.de> +Sean Cunningham <sean.cunningham@mandiant.com> +Peter Griess <pg@std.in> +Salman Haq <salman.haq@asti-usa.com> +Cliff Frey <clifffrey@gmail.com> +Jon Kolb <jon@b0g.us> +Fouad Mardini <f.mardini@gmail.com> +Paul Querna <pquerna@apache.org> +Felix Geisendörfer <felix@debuggable.com> +koichik <koichik@improvement.jp> +Andre Caron <andre.l.caron@gmail.com> +Ivo Raisr <ivosh@ivosh.net> +James McLaughlin <jamie@lacewing-project.org> +David Gwynne <loki@animata.net> +Thomas LE ROUX <thomas@november-eleven.fr> +Randy Rizun <rrizun@ortivawireless.com> +Andre Louis Caron <andre.louis.caron@usherbrooke.ca> +Simon Zimmermann <simonz05@gmail.com> +Erik Dubbelboer <erik@dubbelboer.com> +Martell Malone <martellmalone@gmail.com> +Bertrand Paquet <bpaquet@octo.com> +BogDan Vatra <bogdan@kde.org> +Peter Faiman <peter@thepicard.org> +Corey Richardson <corey@octayn.net> +Tóth Tamás <tomika_nospam@freemail.hu> +Cam Swords <cam.swords@gmail.com> +Chris Dickinson <christopher.s.dickinson@gmail.com> +Uli Köhler <ukoehler@btronik.de> +Charlie Somerville <charlie@charliesomerville.com> +Patrik Stutz <patrik.stutz@gmail.com> +Fedor Indutny <fedor.indutny@gmail.com> +runner <runner.mei@gmail.com> +Alexis Campailla <alexis@janeasystems.com> +David Wragg <david@wragg.org> +Vinnie Falco <vinnie.falco@gmail.com> +Alex Butum <alexbutum@linux.com> +Rex Feng <rexfeng@gmail.com> +Alex Kocharin <alex@kocharin.ru> +Mark Koopman <markmontymark@yahoo.com> +Helge Heß <me@helgehess.eu> +Alexis La Goutte <alexis.lagoutte@gmail.com> +George Miroshnykov <george.miroshnykov@gmail.com> +Maciej Małecki <me@mmalecki.com> +Marc O'Morain <github.com@marcomorain.com> +Jeff Pinner <jpinner@twitter.com> +Timothy J Fontaine <tjfontaine@gmail.com> +Akagi201 <akagi201@gmail.com> +Romain Giraud <giraud.romain@gmail.com> +Jay Satiro <raysatiro@yahoo.com> +Arne Steen <Arne.Steen@gmx.de> +Kjell Schubert <kjell.schubert@gmail.com> +Olivier Mengué <dolmen@cpan.org> diff --git a/src/boost/libs/beast/test/bench/parser/nodejs-parser/LICENSE-MIT b/src/boost/libs/beast/test/bench/parser/nodejs-parser/LICENSE-MIT new file mode 100644 index 00000000..58010b38 --- /dev/null +++ b/src/boost/libs/beast/test/bench/parser/nodejs-parser/LICENSE-MIT @@ -0,0 +1,23 @@ +http_parser.c is based on src/http/ngx_http_parse.c from NGINX copyright +Igor Sysoev. + +Additional changes are licensed under the same terms as NGINX and +copyright Joyent, Inc. and other Node contributors. All rights reserved. + +Permission is hereby granted, free of charge, to any person obtaining a copy +of this software and associated documentation files (the "Software"), to +deal in the Software without restriction, including without limitation the +rights to use, copy, modify, merge, publish, distribute, sublicense, and/or +sell copies of the Software, and to permit persons to whom the Software is +furnished to do so, subject to the following conditions: + +The above copyright notice and this permission notice shall be included in +all copies or substantial portions of the Software. + +THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR +IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, +FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE +AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER +LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING +FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS +IN THE SOFTWARE. diff --git a/src/boost/libs/beast/test/bench/parser/nodejs-parser/README.md b/src/boost/libs/beast/test/bench/parser/nodejs-parser/README.md new file mode 100644 index 00000000..eedd7f8c --- /dev/null +++ b/src/boost/libs/beast/test/bench/parser/nodejs-parser/README.md @@ -0,0 +1,246 @@ +HTTP Parser +=========== + +[](https://travis-ci.org/nodejs/http-parser) + +This is a parser for HTTP messages written in C. It parses both requests and +responses. The parser is designed to be used in performance HTTP +applications. It does not make any syscalls nor allocations, it does not +buffer data, it can be interrupted at anytime. Depending on your +architecture, it only requires about 40 bytes of data per message +stream (in a web server that is per connection). + +Features: + + * No dependencies + * Handles persistent streams (keep-alive). + * Decodes chunked encoding. + * Upgrade support + * Defends against buffer overflow attacks. + +The parser extracts the following information from HTTP messages: + + * Header fields and values + * Content-Length + * Request method + * Response status code + * Transfer-Encoding + * HTTP version + * Request URL + * Message body + + +Usage +----- + +One `http_parser` object is used per TCP connection. Initialize the struct +using `http_parser_init()` and set the callbacks. That might look something +like this for a request parser: +```c +http_parser_settings settings; +settings.on_url = my_url_callback; +settings.on_header_field = my_header_field_callback; +/* ... */ + +http_parser *parser = malloc(sizeof(http_parser)); +http_parser_init(parser, HTTP_REQUEST); +parser->data = my_socket; +``` + +When data is received on the socket execute the parser and check for errors. + +```c +size_t len = 80*1024, nparsed; +char buf[len]; +ssize_t recved; + +recved = recv(fd, buf, len, 0); + +if (recved < 0) { + /* Handle error. */ +} + +/* Start up / continue the parser. + * Note we pass recved==0 to signal that EOF has been received. + */ +nparsed = http_parser_execute(parser, &settings, buf, recved); + +if (parser->upgrade) { + /* handle new protocol */ +} else if (nparsed != recved) { + /* Handle error. Usually just close the connection. */ +} +``` + +HTTP needs to know where the end of the stream is. For example, sometimes +servers send responses without Content-Length and expect the client to +consume input (for the body) until EOF. To tell http_parser about EOF, give +`0` as the fourth parameter to `http_parser_execute()`. Callbacks and errors +can still be encountered during an EOF, so one must still be prepared +to receive them. + +Scalar valued message information such as `status_code`, `method`, and the +HTTP version are stored in the parser structure. This data is only +temporally stored in `http_parser` and gets reset on each new message. If +this information is needed later, copy it out of the structure during the +`headers_complete` callback. + +The parser decodes the transfer-encoding for both requests and responses +transparently. That is, a chunked encoding is decoded before being sent to +the on_body callback. + + +The Special Problem of Upgrade +------------------------------ + +HTTP supports upgrading the connection to a different protocol. An +increasingly common example of this is the WebSocket protocol which sends +a request like + + GET /demo HTTP/1.1 + Upgrade: WebSocket + Connection: Upgrade + Host: example.com + Origin: http://example.com + WebSocket-Protocol: sample + +followed by non-HTTP data. + +(See [RFC6455](https://tools.ietf.org/html/rfc6455) for more information the +WebSocket protocol.) + +To support this, the parser will treat this as a normal HTTP message without a +body, issuing both on_headers_complete and on_message_complete callbacks. However +http_parser_execute() will stop parsing at the end of the headers and return. + +The user is expected to check if `parser->upgrade` has been set to 1 after +`http_parser_execute()` returns. Non-HTTP data begins at the buffer supplied +offset by the return value of `http_parser_execute()`. + + +Callbacks +--------- + +During the `http_parser_execute()` call, the callbacks set in +`http_parser_settings` will be executed. The parser maintains state and +never looks behind, so buffering the data is not necessary. If you need to +save certain data for later usage, you can do that from the callbacks. + +There are two types of callbacks: + +* notification `typedef int (*http_cb) (http_parser*);` + Callbacks: on_message_begin, on_headers_complete, on_message_complete. +* data `typedef int (*http_data_cb) (http_parser*, const char *at, size_t length);` + Callbacks: (requests only) on_url, + (common) on_header_field, on_header_value, on_body; + +Callbacks must return 0 on success. Returning a non-zero value indicates +error to the parser, making it exit immediately. + +For cases where it is necessary to pass local information to/from a callback, +the `http_parser` object's `data` field can be used. +An example of such a case is when using threads to handle a socket connection, +parse a request, and then give a response over that socket. By instantiation +of a thread-local struct containing relevant data (e.g. accepted socket, +allocated memory for callbacks to write into, etc), a parser's callbacks are +able to communicate data between the scope of the thread and the scope of the +callback in a threadsafe manner. This allows http-parser to be used in +multi-threaded contexts. + +Example: +``` + typedef struct { + socket_t sock; + void* buffer; + int buf_len; + } custom_data_t; + + +int my_url_callback(http_parser* parser, const char *at, size_t length) { + /* access to thread local custom_data_t struct. + Use this access save parsed data for later use into thread local + buffer, or communicate over socket + */ + parser->data; + ... + return 0; +} + +... + +void http_parser_thread(socket_t sock) { + int nparsed = 0; + /* allocate memory for user data */ + custom_data_t *my_data = malloc(sizeof(custom_data_t)); + + /* some information for use by callbacks. + * achieves thread -> callback information flow */ + my_data->sock = sock; + + /* instantiate a thread-local parser */ + http_parser *parser = malloc(sizeof(http_parser)); + http_parser_init(parser, HTTP_REQUEST); /* initialise parser */ + /* this custom data reference is accessible through the reference to the + parser supplied to callback functions */ + parser->data = my_data; + + http_parser_settings settings; / * set up callbacks */ + settings.on_url = my_url_callback; + + /* execute parser */ + nparsed = http_parser_execute(parser, &settings, buf, recved); + + ... + /* parsed information copied from callback. + can now perform action on data copied into thread-local memory from callbacks. + achieves callback -> thread information flow */ + my_data->buffer; + ... +} + +``` + +In case you parse HTTP message in chunks (i.e. `read()` request line +from socket, parse, read half headers, parse, etc) your data callbacks +may be called more than once. Http-parser guarantees that data pointer is only +valid for the lifetime of callback. You can also `read()` into a heap allocated +buffer to avoid copying memory around if this fits your application. + +Reading headers may be a tricky task if you read/parse headers partially. +Basically, you need to remember whether last header callback was field or value +and apply the following logic: + + (on_header_field and on_header_value shortened to on_h_*) + ------------------------ ------------ -------------------------------------------- + | State (prev. callback) | Callback | Description/action | + ------------------------ ------------ -------------------------------------------- + | nothing (first call) | on_h_field | Allocate new buffer and copy callback data | + | | | into it | + ------------------------ ------------ -------------------------------------------- + | value | on_h_field | New header started. | + | | | Copy current name,value buffers to headers | + | | | list and allocate new buffer for new name | + ------------------------ ------------ -------------------------------------------- + | field | on_h_field | Previous name continues. Reallocate name | + | | | buffer and append callback data to it | + ------------------------ ------------ -------------------------------------------- + | field | on_h_value | Value for current header started. Allocate | + | | | new buffer and copy callback data to it | + ------------------------ ------------ -------------------------------------------- + | value | on_h_value | Value continues. Reallocate value buffer | + | | | and append callback data to it | + ------------------------ ------------ -------------------------------------------- + + +Parsing URLs +------------ + +A simplistic zero-copy URL parser is provided as `http_parser_parse_url()`. +Users of this library may wish to use it to parse URLs constructed from +consecutive `on_url` callbacks. + +See examples of reading in headers: + +* [partial example](http://gist.github.com/155877) in C +* [from http-parser tests](http://github.com/joyent/http-parser/blob/37a0ff8/test.c#L403) in C +* [from Node library](http://github.com/joyent/node/blob/842eaf4/src/http.js#L284) in Javascript diff --git a/src/boost/libs/beast/test/bench/parser/nodejs-parser/http_parser.c b/src/boost/libs/beast/test/bench/parser/nodejs-parser/http_parser.c new file mode 100644 index 00000000..71961754 --- /dev/null +++ b/src/boost/libs/beast/test/bench/parser/nodejs-parser/http_parser.c @@ -0,0 +1,2468 @@ +/* Based on src/http/ngx_http_parse.c from NGINX copyright Igor Sysoev + * + * Additional changes are licensed under the same terms as NGINX and + * copyright Joyent, Inc. and other Node contributors. All rights reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + */ +#include "http_parser.h" +#include <assert.h> +#include <stddef.h> +#include <ctype.h> +#include <stdlib.h> +#include <string.h> +#include <limits.h> + +#ifndef ULLONG_MAX +# define ULLONG_MAX ((uint64_t) -1) /* 2^64-1 */ +#endif + +#ifndef MIN +# define MIN(a,b) ((a) < (b) ? (a) : (b)) +#endif + +#ifndef ARRAY_SIZE +# define ARRAY_SIZE(a) (sizeof(a) / sizeof((a)[0])) +#endif + +#ifndef BIT_AT +# define BIT_AT(a, i) \ + (!!((unsigned int) (a)[(unsigned int) (i) >> 3] & \ + (1 << ((unsigned int) (i) & 7)))) +#endif + +#ifndef ELEM_AT +# define ELEM_AT(a, i, v) ((unsigned int) (i) < ARRAY_SIZE(a) ? (a)[(i)] : (v)) +#endif + +#define SET_ERRNO(e) \ +do { \ + parser->http_errno = (e); \ +} while(0) + +#define CURRENT_STATE() p_state +#define UPDATE_STATE(V) p_state = (enum state) (V); +#define RETURN(V) \ +do { \ + parser->state = CURRENT_STATE(); \ + return (V); \ +} while (0); +#define REEXECUTE() \ + goto reexecute; \ + + +#ifdef __GNUC__ +# define LIKELY(X) __builtin_expect(!!(X), 1) +# define UNLIKELY(X) __builtin_expect(!!(X), 0) +#else +# define LIKELY(X) (X) +# define UNLIKELY(X) (X) +#endif + + +/* Run the notify callback FOR, returning ER if it fails */ +#define CALLBACK_NOTIFY_(FOR, ER) \ +do { \ + assert(HTTP_PARSER_ERRNO(parser) == HPE_OK); \ + \ + if (LIKELY(settings->on_##FOR)) { \ + parser->state = CURRENT_STATE(); \ + if (UNLIKELY(0 != settings->on_##FOR(parser))) { \ + SET_ERRNO(HPE_CB_##FOR); \ + } \ + UPDATE_STATE(parser->state); \ + \ + /* We either errored above or got paused; get out */ \ + if (UNLIKELY(HTTP_PARSER_ERRNO(parser) != HPE_OK)) { \ + return (ER); \ + } \ + } \ +} while (0) + +/* Run the notify callback FOR and consume the current byte */ +#define CALLBACK_NOTIFY(FOR) CALLBACK_NOTIFY_(FOR, p - data + 1) + +/* Run the notify callback FOR and don't consume the current byte */ +#define CALLBACK_NOTIFY_NOADVANCE(FOR) CALLBACK_NOTIFY_(FOR, p - data) + +/* Run data callback FOR with LEN bytes, returning ER if it fails */ +#define CALLBACK_DATA_(FOR, LEN, ER) \ +do { \ + assert(HTTP_PARSER_ERRNO(parser) == HPE_OK); \ + \ + if (FOR##_mark) { \ + if (LIKELY(settings->on_##FOR)) { \ + parser->state = CURRENT_STATE(); \ + if (UNLIKELY(0 != \ + settings->on_##FOR(parser, FOR##_mark, (LEN)))) { \ + SET_ERRNO(HPE_CB_##FOR); \ + } \ + UPDATE_STATE(parser->state); \ + \ + /* We either errored above or got paused; get out */ \ + if (UNLIKELY(HTTP_PARSER_ERRNO(parser) != HPE_OK)) { \ + return (ER); \ + } \ + } \ + FOR##_mark = NULL; \ + } \ +} while (0) + +/* Run the data callback FOR and consume the current byte */ +#define CALLBACK_DATA(FOR) \ + CALLBACK_DATA_(FOR, p - FOR##_mark, p - data + 1) + +/* Run the data callback FOR and don't consume the current byte */ +#define CALLBACK_DATA_NOADVANCE(FOR) \ + CALLBACK_DATA_(FOR, p - FOR##_mark, p - data) + +/* Set the mark FOR; non-destructive if mark is already set */ +#define MARK(FOR) \ +do { \ + if (!FOR##_mark) { \ + FOR##_mark = p; \ + } \ +} while (0) + +/* Don't allow the total size of the HTTP headers (including the status + * line) to exceed HTTP_MAX_HEADER_SIZE. This check is here to protect + * embedders against denial-of-service attacks where the attacker feeds + * us a never-ending header that the embedder keeps buffering. + * + * This check is arguably the responsibility of embedders but we're doing + * it on the embedder's behalf because most won't bother and this way we + * make the web a little safer. HTTP_MAX_HEADER_SIZE is still far bigger + * than any reasonable request or response so this should never affect + * day-to-day operation. + */ +#define COUNT_HEADER_SIZE(V) \ +do { \ + parser->nread += (V); \ + if (UNLIKELY(parser->nread > (HTTP_MAX_HEADER_SIZE))) { \ + SET_ERRNO(HPE_HEADER_OVERFLOW); \ + goto error; \ + } \ +} while (0) + + +#define PROXY_CONNECTION "proxy-connection" +#define CONNECTION "connection" +#define CONTENT_LENGTH "content-length" +#define TRANSFER_ENCODING "transfer-encoding" +#define UPGRADE "upgrade" +#define CHUNKED "chunked" +#define KEEP_ALIVE "keep-alive" +#define CLOSE "close" + + +static const char *method_strings[] = + { +#define XX(num, name, string) #string, + HTTP_METHOD_MAP(XX) +#undef XX + }; + + +/* Tokens as defined by rfc 2616. Also lowercases them. + * token = 1*<any CHAR except CTLs or separators> + * separators = "(" | ")" | "<" | ">" | "@" + * | "," | ";" | ":" | "\" | <"> + * | "/" | "[" | "]" | "?" | "=" + * | "{" | "}" | SP | HT + */ +static const char tokens[256] = { +/* 0 nul 1 soh 2 stx 3 etx 4 eot 5 enq 6 ack 7 bel */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 8 bs 9 ht 10 nl 11 vt 12 np 13 cr 14 so 15 si */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 16 dle 17 dc1 18 dc2 19 dc3 20 dc4 21 nak 22 syn 23 etb */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 24 can 25 em 26 sub 27 esc 28 fs 29 gs 30 rs 31 us */ + 0, 0, 0, 0, 0, 0, 0, 0, +/* 32 sp 33 ! 34 " 35 # 36 $ 37 % 38 & 39 ' */ + 0, '!', 0, '#', '$', '%', '&', '\'', +/* 40 ( 41 ) 42 * 43 + 44 , 45 - 46 . 47 / */ + 0, 0, '*', '+', 0, '-', '.', 0, +/* 48 0 49 1 50 2 51 3 52 4 53 5 54 6 55 7 */ + '0', '1', '2', '3', '4', '5', '6', '7', +/* 56 8 57 9 58 : 59 ; 60 < 61 = 62 > 63 ? */ + '8', '9', 0, 0, 0, 0, 0, 0, +/* 64 @ 65 A 66 B 67 C 68 D 69 E 70 F 71 G */ + 0, 'a', 'b', 'c', 'd', 'e', 'f', 'g', +/* 72 H 73 I 74 J 75 K 76 L 77 M 78 N 79 O */ + 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', +/* 80 P 81 Q 82 R 83 S 84 T 85 U 86 V 87 W */ + 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', +/* 88 X 89 Y 90 Z 91 [ 92 \ 93 ] 94 ^ 95 _ */ + 'x', 'y', 'z', 0, 0, 0, '^', '_', +/* 96 ` 97 a 98 b 99 c 100 d 101 e 102 f 103 g */ + '`', 'a', 'b', 'c', 'd', 'e', 'f', 'g', +/* 104 h 105 i 106 j 107 k 108 l 109 m 110 n 111 o */ + 'h', 'i', 'j', 'k', 'l', 'm', 'n', 'o', +/* 112 p 113 q 114 r 115 s 116 t 117 u 118 v 119 w */ + 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', +/* 120 x 121 y 122 z 123 { 124 | 125 } 126 ~ 127 del */ + 'x', 'y', 'z', 0, '|', 0, '~', 0 }; + + +static const int8_t unhex[256] = + {-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 + ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 + ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 + , 0, 1, 2, 3, 4, 5, 6, 7, 8, 9,-1,-1,-1,-1,-1,-1 + ,-1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1 + ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 + ,-1,10,11,12,13,14,15,-1,-1,-1,-1,-1,-1,-1,-1,-1 + ,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1,-1 + }; + + +#if HTTP_PARSER_STRICT +# define T(v) 0 +#else +# define T(v) v +#endif + + +static const uint8_t normal_url_char[32] = { +/* 0 nul 1 soh 2 stx 3 etx 4 eot 5 enq 6 ack 7 bel */ + 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0, +/* 8 bs 9 ht 10 nl 11 vt 12 np 13 cr 14 so 15 si */ + 0 | T(2) | 0 | 0 | T(16) | 0 | 0 | 0, +/* 16 dle 17 dc1 18 dc2 19 dc3 20 dc4 21 nak 22 syn 23 etb */ + 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0, +/* 24 can 25 em 26 sub 27 esc 28 fs 29 gs 30 rs 31 us */ + 0 | 0 | 0 | 0 | 0 | 0 | 0 | 0, +/* 32 sp 33 ! 34 " 35 # 36 $ 37 % 38 & 39 ' */ + 0 | 2 | 4 | 0 | 16 | 32 | 64 | 128, +/* 40 ( 41 ) 42 * 43 + 44 , 45 - 46 . 47 / */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128, +/* 48 0 49 1 50 2 51 3 52 4 53 5 54 6 55 7 */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128, +/* 56 8 57 9 58 : 59 ; 60 < 61 = 62 > 63 ? */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 0, +/* 64 @ 65 A 66 B 67 C 68 D 69 E 70 F 71 G */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128, +/* 72 H 73 I 74 J 75 K 76 L 77 M 78 N 79 O */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128, +/* 80 P 81 Q 82 R 83 S 84 T 85 U 86 V 87 W */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128, +/* 88 X 89 Y 90 Z 91 [ 92 \ 93 ] 94 ^ 95 _ */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128, +/* 96 ` 97 a 98 b 99 c 100 d 101 e 102 f 103 g */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128, +/* 104 h 105 i 106 j 107 k 108 l 109 m 110 n 111 o */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128, +/* 112 p 113 q 114 r 115 s 116 t 117 u 118 v 119 w */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 128, +/* 120 x 121 y 122 z 123 { 124 | 125 } 126 ~ 127 del */ + 1 | 2 | 4 | 8 | 16 | 32 | 64 | 0, }; + +#undef T + +enum state + { s_dead = 1 /* important that this is > 0 */ + + , s_start_req_or_res + , s_res_or_resp_H + , s_start_res + , s_res_H + , s_res_HT + , s_res_HTT + , s_res_HTTP + , s_res_first_http_major + , s_res_http_major + , s_res_first_http_minor + , s_res_http_minor + , s_res_first_status_code + , s_res_status_code + , s_res_status_start + , s_res_status + , s_res_line_almost_done + + , s_start_req + + , s_req_method + , s_req_spaces_before_url + , s_req_schema + , s_req_schema_slash + , s_req_schema_slash_slash + , s_req_server_start + , s_req_server + , s_req_server_with_at + , s_req_path + , s_req_query_string_start + , s_req_query_string + , s_req_fragment_start + , s_req_fragment + , s_req_http_start + , s_req_http_H + , s_req_http_HT + , s_req_http_HTT + , s_req_http_HTTP + , s_req_first_http_major + , s_req_http_major + , s_req_first_http_minor + , s_req_http_minor + , s_req_line_almost_done + + , s_header_field_start + , s_header_field + , s_header_value_discard_ws + , s_header_value_discard_ws_almost_done + , s_header_value_discard_lws + , s_header_value_start + , s_header_value + , s_header_value_lws + + , s_header_almost_done + + , s_chunk_size_start + , s_chunk_size + , s_chunk_parameters + , s_chunk_size_almost_done + + , s_headers_almost_done + , s_headers_done + + /* Important: 's_headers_done' must be the last 'header' state. All + * states beyond this must be 'body' states. It is used for overflow + * checking. See the PARSING_HEADER() macro. + */ + + , s_chunk_data + , s_chunk_data_almost_done + , s_chunk_data_done + + , s_body_identity + , s_body_identity_eof + + , s_message_done + }; + + +#define PARSING_HEADER(state) (state <= s_headers_done) + + +enum header_states + { h_general = 0 + , h_C + , h_CO + , h_CON + + , h_matching_connection + , h_matching_proxy_connection + , h_matching_content_length + , h_matching_transfer_encoding + , h_matching_upgrade + + , h_connection + , h_content_length + , h_transfer_encoding + , h_upgrade + + , h_matching_transfer_encoding_chunked + , h_matching_connection_token_start + , h_matching_connection_keep_alive + , h_matching_connection_close + , h_matching_connection_upgrade + , h_matching_connection_token + + , h_transfer_encoding_chunked + , h_connection_keep_alive + , h_connection_close + , h_connection_upgrade + }; + +enum http_host_state + { + s_http_host_dead = 1 + , s_http_userinfo_start + , s_http_userinfo + , s_http_host_start + , s_http_host_v6_start + , s_http_host + , s_http_host_v6 + , s_http_host_v6_end + , s_http_host_v6_zone_start + , s_http_host_v6_zone + , s_http_host_port_start + , s_http_host_port +}; + +/* Macros for character classes; depends on strict-mode */ +#define CR '\r' +#define LF '\n' +#define LOWER(c) (unsigned char)(c | 0x20) +#define IS_ALPHA(c) (LOWER(c) >= 'a' && LOWER(c) <= 'z') +#define IS_NUM(c) ((c) >= '0' && (c) <= '9') +#define IS_ALPHANUM(c) (IS_ALPHA(c) || IS_NUM(c)) +#define IS_HEX(c) (IS_NUM(c) || (LOWER(c) >= 'a' && LOWER(c) <= 'f')) +#define IS_MARK(c) ((c) == '-' || (c) == '_' || (c) == '.' || \ + (c) == '!' || (c) == '~' || (c) == '*' || (c) == '\'' || (c) == '(' || \ + (c) == ')') +#define IS_USERINFO_CHAR(c) (IS_ALPHANUM(c) || IS_MARK(c) || (c) == '%' || \ + (c) == ';' || (c) == ':' || (c) == '&' || (c) == '=' || (c) == '+' || \ + (c) == '$' || (c) == ',') + +#define STRICT_TOKEN(c) (tokens[(unsigned char)c]) + +#if HTTP_PARSER_STRICT +#define TOKEN(c) (tokens[(unsigned char)c]) +#define IS_URL_CHAR(c) (BIT_AT(normal_url_char, (unsigned char)c)) +#define IS_HOST_CHAR(c) (IS_ALPHANUM(c) || (c) == '.' || (c) == '-') +#else +#define TOKEN(c) ((c == ' ') ? ' ' : tokens[(unsigned char)c]) +#define IS_URL_CHAR(c) \ + (BIT_AT(normal_url_char, (unsigned char)c) || ((c) & 0x80)) +#define IS_HOST_CHAR(c) \ + (IS_ALPHANUM(c) || (c) == '.' || (c) == '-' || (c) == '_') +#endif + +/** + * Verify that a char is a valid visible (printable) US-ASCII + * character or %x80-FF + **/ +#define IS_HEADER_CHAR(ch) \ + (ch == CR || ch == LF || ch == 9 || ((unsigned char)ch > 31 && ch != 127)) + +#define start_state (parser->type == HTTP_REQUEST ? s_start_req : s_start_res) + + +#if HTTP_PARSER_STRICT +# define STRICT_CHECK(cond) \ +do { \ + if (cond) { \ + SET_ERRNO(HPE_STRICT); \ + goto error; \ + } \ +} while (0) +# define NEW_MESSAGE() (http_should_keep_alive(parser) ? start_state : s_dead) +#else +# define STRICT_CHECK(cond) +# define NEW_MESSAGE() start_state +#endif + + +/* Map errno values to strings for human-readable output */ +#define HTTP_STRERROR_GEN(n, s) { "HPE_" #n, s }, +static struct { + const char *name; + const char *description; +} http_strerror_tab[] = { + HTTP_ERRNO_MAP(HTTP_STRERROR_GEN) +}; +#undef HTTP_STRERROR_GEN + +int http_message_needs_eof(const http_parser *parser); + +/* Our URL parser. + * + * This is designed to be shared by http_parser_execute() for URL validation, + * hence it has a state transition + byte-for-byte interface. In addition, it + * is meant to be embedded in http_parser_parse_url(), which does the dirty + * work of turning state transitions URL components for its API. + * + * This function should only be invoked with non-space characters. It is + * assumed that the caller cares about (and can detect) the transition between + * URL and non-URL states by looking for these. + */ +static enum state +parse_url_char(enum state s, const char ch) +{ + if (ch == ' ' || ch == '\r' || ch == '\n') { + return s_dead; + } + +#if HTTP_PARSER_STRICT + if (ch == '\t' || ch == '\f') { + return s_dead; + } +#endif + + switch (s) { + case s_req_spaces_before_url: + /* Proxied requests are followed by scheme of an absolute URI (alpha). + * All methods except CONNECT are followed by '/' or '*'. + */ + + if (ch == '/' || ch == '*') { + return s_req_path; + } + + if (IS_ALPHA(ch)) { + return s_req_schema; + } + + break; + + case s_req_schema: + if (IS_ALPHA(ch)) { + return s; + } + + if (ch == ':') { + return s_req_schema_slash; + } + + break; + + case s_req_schema_slash: + if (ch == '/') { + return s_req_schema_slash_slash; + } + + break; + + case s_req_schema_slash_slash: + if (ch == '/') { + return s_req_server_start; + } + + break; + + case s_req_server_with_at: + if (ch == '@') { + return s_dead; + } + + /* FALLTHROUGH */ + case s_req_server_start: + case s_req_server: + if (ch == '/') { + return s_req_path; + } + + if (ch == '?') { + return s_req_query_string_start; + } + + if (ch == '@') { + return s_req_server_with_at; + } + + if (IS_USERINFO_CHAR(ch) || ch == '[' || ch == ']') { + return s_req_server; + } + + break; + + case s_req_path: + if (IS_URL_CHAR(ch)) { + return s; + } + + switch (ch) { + case '?': + return s_req_query_string_start; + + case '#': + return s_req_fragment_start; + } + + break; + + case s_req_query_string_start: + case s_req_query_string: + if (IS_URL_CHAR(ch)) { + return s_req_query_string; + } + + switch (ch) { + case '?': + /* allow extra '?' in query string */ + return s_req_query_string; + + case '#': + return s_req_fragment_start; + } + + break; + + case s_req_fragment_start: + if (IS_URL_CHAR(ch)) { + return s_req_fragment; + } + + switch (ch) { + case '?': + return s_req_fragment; + + case '#': + return s; + } + + break; + + case s_req_fragment: + if (IS_URL_CHAR(ch)) { + return s; + } + + switch (ch) { + case '?': + case '#': + return s; + } + + break; + + default: + break; + } + + /* We should never fall out of the switch above unless there's an error */ + return s_dead; +} + +size_t http_parser_execute (http_parser *parser, + const http_parser_settings *settings, + const char *data, + size_t len) +{ + char c, ch; + int8_t unhex_val; + const char *p = data; + const char *header_field_mark = 0; + const char *header_value_mark = 0; + const char *url_mark = 0; + const char *body_mark = 0; + const char *status_mark = 0; + enum state p_state = (enum state) parser->state; + const unsigned int lenient = parser->lenient_http_headers; + + /* We're in an error state. Don't bother doing anything. */ + if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { + return 0; + } + + if (len == 0) { + switch (CURRENT_STATE()) { + case s_body_identity_eof: + /* Use of CALLBACK_NOTIFY() here would erroneously return 1 byte read if + * we got paused. + */ + CALLBACK_NOTIFY_NOADVANCE(message_complete); + return 0; + + case s_dead: + case s_start_req_or_res: + case s_start_res: + case s_start_req: + return 0; + + default: + SET_ERRNO(HPE_INVALID_EOF_STATE); + return 1; + } + } + + + if (CURRENT_STATE() == s_header_field) + header_field_mark = data; + if (CURRENT_STATE() == s_header_value) + header_value_mark = data; + switch (CURRENT_STATE()) { + case s_req_path: + case s_req_schema: + case s_req_schema_slash: + case s_req_schema_slash_slash: + case s_req_server_start: + case s_req_server: + case s_req_server_with_at: + case s_req_query_string_start: + case s_req_query_string: + case s_req_fragment_start: + case s_req_fragment: + url_mark = data; + break; + case s_res_status: + status_mark = data; + break; + default: + break; + } + + for (p=data; p != data + len; p++) { + ch = *p; + + if (PARSING_HEADER(CURRENT_STATE())) + COUNT_HEADER_SIZE(1); + +reexecute: + switch (CURRENT_STATE()) { + + case s_dead: + /* this state is used after a 'Connection: close' message + * the parser will error out if it reads another message + */ + if (LIKELY(ch == CR || ch == LF)) + break; + + SET_ERRNO(HPE_CLOSED_CONNECTION); + goto error; + + case s_start_req_or_res: + { + if (ch == CR || ch == LF) + break; + parser->flags = 0; + parser->content_length = ULLONG_MAX; + + if (ch == 'H') { + UPDATE_STATE(s_res_or_resp_H); + + CALLBACK_NOTIFY(message_begin); + } else { + parser->type = HTTP_REQUEST; + UPDATE_STATE(s_start_req); + REEXECUTE(); + } + + break; + } + + case s_res_or_resp_H: + if (ch == 'T') { + parser->type = HTTP_RESPONSE; + UPDATE_STATE(s_res_HT); + } else { + if (UNLIKELY(ch != 'E')) { + SET_ERRNO(HPE_INVALID_CONSTANT); + goto error; + } + + parser->type = HTTP_REQUEST; + parser->method = HTTP_HEAD; + parser->index = 2; + UPDATE_STATE(s_req_method); + } + break; + + case s_start_res: + { + parser->flags = 0; + parser->content_length = ULLONG_MAX; + + switch (ch) { + case 'H': + UPDATE_STATE(s_res_H); + break; + + case CR: + case LF: + break; + + default: + SET_ERRNO(HPE_INVALID_CONSTANT); + goto error; + } + + CALLBACK_NOTIFY(message_begin); + break; + } + + case s_res_H: + STRICT_CHECK(ch != 'T'); + UPDATE_STATE(s_res_HT); + break; + + case s_res_HT: + STRICT_CHECK(ch != 'T'); + UPDATE_STATE(s_res_HTT); + break; + + case s_res_HTT: + STRICT_CHECK(ch != 'P'); + UPDATE_STATE(s_res_HTTP); + break; + + case s_res_HTTP: + STRICT_CHECK(ch != '/'); + UPDATE_STATE(s_res_first_http_major); + break; + + case s_res_first_http_major: + if (UNLIKELY(ch < '0' || ch > '9')) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_major = ch - '0'; + UPDATE_STATE(s_res_http_major); + break; + + /* major HTTP version or dot */ + case s_res_http_major: + { + if (ch == '.') { + UPDATE_STATE(s_res_first_http_minor); + break; + } + + if (!IS_NUM(ch)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_major *= 10; + parser->http_major += ch - '0'; + + if (UNLIKELY(parser->http_major > 999)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + break; + } + + /* first digit of minor HTTP version */ + case s_res_first_http_minor: + if (UNLIKELY(!IS_NUM(ch))) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_minor = ch - '0'; + UPDATE_STATE(s_res_http_minor); + break; + + /* minor HTTP version or end of request line */ + case s_res_http_minor: + { + if (ch == ' ') { + UPDATE_STATE(s_res_first_status_code); + break; + } + + if (UNLIKELY(!IS_NUM(ch))) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_minor *= 10; + parser->http_minor += ch - '0'; + + if (UNLIKELY(parser->http_minor > 999)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + break; + } + + case s_res_first_status_code: + { + if (!IS_NUM(ch)) { + if (ch == ' ') { + break; + } + + SET_ERRNO(HPE_INVALID_STATUS); + goto error; + } + parser->status_code = ch - '0'; + UPDATE_STATE(s_res_status_code); + break; + } + + case s_res_status_code: + { + if (!IS_NUM(ch)) { + switch (ch) { + case ' ': + UPDATE_STATE(s_res_status_start); + break; + case CR: + UPDATE_STATE(s_res_line_almost_done); + break; + case LF: + UPDATE_STATE(s_header_field_start); + break; + default: + SET_ERRNO(HPE_INVALID_STATUS); + goto error; + } + break; + } + + parser->status_code *= 10; + parser->status_code += ch - '0'; + + if (UNLIKELY(parser->status_code > 999)) { + SET_ERRNO(HPE_INVALID_STATUS); + goto error; + } + + break; + } + + case s_res_status_start: + { + if (ch == CR) { + UPDATE_STATE(s_res_line_almost_done); + break; + } + + if (ch == LF) { + UPDATE_STATE(s_header_field_start); + break; + } + + MARK(status); + UPDATE_STATE(s_res_status); + parser->index = 0; + break; + } + + case s_res_status: + if (ch == CR) { + UPDATE_STATE(s_res_line_almost_done); + CALLBACK_DATA(status); + break; + } + + if (ch == LF) { + UPDATE_STATE(s_header_field_start); + CALLBACK_DATA(status); + break; + } + + break; + + case s_res_line_almost_done: + STRICT_CHECK(ch != LF); + UPDATE_STATE(s_header_field_start); + break; + + case s_start_req: + { + if (ch == CR || ch == LF) + break; + parser->flags = 0; + parser->content_length = ULLONG_MAX; + + if (UNLIKELY(!IS_ALPHA(ch))) { + SET_ERRNO(HPE_INVALID_METHOD); + goto error; + } + + parser->method = (enum http_method) 0; + parser->index = 1; + switch (ch) { + case 'A': parser->method = HTTP_ACL; break; + case 'B': parser->method = HTTP_BIND; break; + case 'C': parser->method = HTTP_CONNECT; /* or COPY, CHECKOUT */ break; + case 'D': parser->method = HTTP_DELETE; break; + case 'G': parser->method = HTTP_GET; break; + case 'H': parser->method = HTTP_HEAD; break; + case 'L': parser->method = HTTP_LOCK; /* or LINK */ break; + case 'M': parser->method = HTTP_MKCOL; /* or MOVE, MKACTIVITY, MERGE, M-SEARCH, MKCALENDAR */ break; + case 'N': parser->method = HTTP_NOTIFY; break; + case 'O': parser->method = HTTP_OPTIONS; break; + case 'P': parser->method = HTTP_POST; + /* or PROPFIND|PROPPATCH|PUT|PATCH|PURGE */ + break; + case 'R': parser->method = HTTP_REPORT; /* or REBIND */ break; + case 'S': parser->method = HTTP_SUBSCRIBE; /* or SEARCH */ break; + case 'T': parser->method = HTTP_TRACE; break; + case 'U': parser->method = HTTP_UNLOCK; /* or UNSUBSCRIBE, UNBIND, UNLINK */ break; + default: + SET_ERRNO(HPE_INVALID_METHOD); + goto error; + } + UPDATE_STATE(s_req_method); + + CALLBACK_NOTIFY(message_begin); + + break; + } + + case s_req_method: + { + const char *matcher; + if (UNLIKELY(ch == '\0')) { + SET_ERRNO(HPE_INVALID_METHOD); + goto error; + } + + matcher = method_strings[parser->method]; + if (ch == ' ' && matcher[parser->index] == '\0') { + UPDATE_STATE(s_req_spaces_before_url); + } else if (ch == matcher[parser->index]) { + ; /* nada */ + } else if (IS_ALPHA(ch)) { + + switch (parser->method << 16 | parser->index << 8 | ch) { +#define XX(meth, pos, ch, new_meth) \ + case (HTTP_##meth << 16 | pos << 8 | ch): \ + parser->method = HTTP_##new_meth; break; + + XX(POST, 1, 'U', PUT) + XX(POST, 1, 'A', PATCH) + XX(CONNECT, 1, 'H', CHECKOUT) + XX(CONNECT, 2, 'P', COPY) + XX(MKCOL, 1, 'O', MOVE) + XX(MKCOL, 1, 'E', MERGE) + XX(MKCOL, 2, 'A', MKACTIVITY) + XX(MKCOL, 3, 'A', MKCALENDAR) + XX(SUBSCRIBE, 1, 'E', SEARCH) + XX(REPORT, 2, 'B', REBIND) + XX(POST, 1, 'R', PROPFIND) + XX(PROPFIND, 4, 'P', PROPPATCH) + XX(PUT, 2, 'R', PURGE) + XX(LOCK, 1, 'I', LINK) + XX(UNLOCK, 2, 'S', UNSUBSCRIBE) + XX(UNLOCK, 2, 'B', UNBIND) + XX(UNLOCK, 3, 'I', UNLINK) +#undef XX + + default: + SET_ERRNO(HPE_INVALID_METHOD); + goto error; + } + } else if (ch == '-' && + parser->index == 1 && + parser->method == HTTP_MKCOL) { + parser->method = HTTP_MSEARCH; + } else { + SET_ERRNO(HPE_INVALID_METHOD); + goto error; + } + + ++parser->index; + break; + } + + case s_req_spaces_before_url: + { + if (ch == ' ') break; + + MARK(url); + if (parser->method == HTTP_CONNECT) { + UPDATE_STATE(s_req_server_start); + } + + UPDATE_STATE(parse_url_char(CURRENT_STATE(), ch)); + if (UNLIKELY(CURRENT_STATE() == s_dead)) { + SET_ERRNO(HPE_INVALID_URL); + goto error; + } + + break; + } + + case s_req_schema: + case s_req_schema_slash: + case s_req_schema_slash_slash: + case s_req_server_start: + { + switch (ch) { + /* No whitespace allowed here */ + case ' ': + case CR: + case LF: + SET_ERRNO(HPE_INVALID_URL); + goto error; + default: + UPDATE_STATE(parse_url_char(CURRENT_STATE(), ch)); + if (UNLIKELY(CURRENT_STATE() == s_dead)) { + SET_ERRNO(HPE_INVALID_URL); + goto error; + } + } + + break; + } + + case s_req_server: + case s_req_server_with_at: + case s_req_path: + case s_req_query_string_start: + case s_req_query_string: + case s_req_fragment_start: + case s_req_fragment: + { + switch (ch) { + case ' ': + UPDATE_STATE(s_req_http_start); + CALLBACK_DATA(url); + break; + case CR: + case LF: + parser->http_major = 0; + parser->http_minor = 9; + UPDATE_STATE((ch == CR) ? + s_req_line_almost_done : + s_header_field_start); + CALLBACK_DATA(url); + break; + default: + UPDATE_STATE(parse_url_char(CURRENT_STATE(), ch)); + if (UNLIKELY(CURRENT_STATE() == s_dead)) { + SET_ERRNO(HPE_INVALID_URL); + goto error; + } + } + break; + } + + case s_req_http_start: + switch (ch) { + case 'H': + UPDATE_STATE(s_req_http_H); + break; + case ' ': + break; + default: + SET_ERRNO(HPE_INVALID_CONSTANT); + goto error; + } + break; + + case s_req_http_H: + STRICT_CHECK(ch != 'T'); + UPDATE_STATE(s_req_http_HT); + break; + + case s_req_http_HT: + STRICT_CHECK(ch != 'T'); + UPDATE_STATE(s_req_http_HTT); + break; + + case s_req_http_HTT: + STRICT_CHECK(ch != 'P'); + UPDATE_STATE(s_req_http_HTTP); + break; + + case s_req_http_HTTP: + STRICT_CHECK(ch != '/'); + UPDATE_STATE(s_req_first_http_major); + break; + + /* first digit of major HTTP version */ + case s_req_first_http_major: + if (UNLIKELY(ch < '1' || ch > '9')) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_major = ch - '0'; + UPDATE_STATE(s_req_http_major); + break; + + /* major HTTP version or dot */ + case s_req_http_major: + { + if (ch == '.') { + UPDATE_STATE(s_req_first_http_minor); + break; + } + + if (UNLIKELY(!IS_NUM(ch))) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_major *= 10; + parser->http_major += ch - '0'; + + if (UNLIKELY(parser->http_major > 999)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + break; + } + + /* first digit of minor HTTP version */ + case s_req_first_http_minor: + if (UNLIKELY(!IS_NUM(ch))) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_minor = ch - '0'; + UPDATE_STATE(s_req_http_minor); + break; + + /* minor HTTP version or end of request line */ + case s_req_http_minor: + { + if (ch == CR) { + UPDATE_STATE(s_req_line_almost_done); + break; + } + + if (ch == LF) { + UPDATE_STATE(s_header_field_start); + break; + } + + /* XXX allow spaces after digit? */ + + if (UNLIKELY(!IS_NUM(ch))) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + parser->http_minor *= 10; + parser->http_minor += ch - '0'; + + if (UNLIKELY(parser->http_minor > 999)) { + SET_ERRNO(HPE_INVALID_VERSION); + goto error; + } + + break; + } + + /* end of request line */ + case s_req_line_almost_done: + { + if (UNLIKELY(ch != LF)) { + SET_ERRNO(HPE_LF_EXPECTED); + goto error; + } + + UPDATE_STATE(s_header_field_start); + break; + } + + case s_header_field_start: + { + if (ch == CR) { + UPDATE_STATE(s_headers_almost_done); + break; + } + + if (ch == LF) { + /* they might be just sending \n instead of \r\n so this would be + * the second \n to denote the end of headers*/ + UPDATE_STATE(s_headers_almost_done); + REEXECUTE(); + } + + c = TOKEN(ch); + + if (UNLIKELY(!c)) { + SET_ERRNO(HPE_INVALID_HEADER_TOKEN); + goto error; + } + + MARK(header_field); + + parser->index = 0; + UPDATE_STATE(s_header_field); + + switch (c) { + case 'c': + parser->header_state = h_C; + break; + + case 'p': + parser->header_state = h_matching_proxy_connection; + break; + + case 't': + parser->header_state = h_matching_transfer_encoding; + break; + + case 'u': + parser->header_state = h_matching_upgrade; + break; + + default: + parser->header_state = h_general; + break; + } + break; + } + + case s_header_field: + { + const char* start = p; + for (; p != data + len; p++) { + ch = *p; + c = TOKEN(ch); + + if (!c) + break; + + switch (parser->header_state) { + case h_general: + break; + + case h_C: + parser->index++; + parser->header_state = (c == 'o' ? h_CO : h_general); + break; + + case h_CO: + parser->index++; + parser->header_state = (c == 'n' ? h_CON : h_general); + break; + + case h_CON: + parser->index++; + switch (c) { + case 'n': + parser->header_state = h_matching_connection; + break; + case 't': + parser->header_state = h_matching_content_length; + break; + default: + parser->header_state = h_general; + break; + } + break; + + /* connection */ + + case h_matching_connection: + parser->index++; + if (parser->index > sizeof(CONNECTION)-1 + || c != CONNECTION[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(CONNECTION)-2) { + parser->header_state = h_connection; + } + break; + + /* proxy-connection */ + + case h_matching_proxy_connection: + parser->index++; + if (parser->index > sizeof(PROXY_CONNECTION)-1 + || c != PROXY_CONNECTION[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(PROXY_CONNECTION)-2) { + parser->header_state = h_connection; + } + break; + + /* content-length */ + + case h_matching_content_length: + parser->index++; + if (parser->index > sizeof(CONTENT_LENGTH)-1 + || c != CONTENT_LENGTH[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(CONTENT_LENGTH)-2) { + if (parser->flags & F_CONTENTLENGTH) { + SET_ERRNO(HPE_UNEXPECTED_CONTENT_LENGTH); + goto error; + } + parser->header_state = h_content_length; + parser->flags |= F_CONTENTLENGTH; + } + break; + + /* transfer-encoding */ + + case h_matching_transfer_encoding: + parser->index++; + if (parser->index > sizeof(TRANSFER_ENCODING)-1 + || c != TRANSFER_ENCODING[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(TRANSFER_ENCODING)-2) { + parser->header_state = h_transfer_encoding; + } + break; + + /* upgrade */ + + case h_matching_upgrade: + parser->index++; + if (parser->index > sizeof(UPGRADE)-1 + || c != UPGRADE[parser->index]) { + parser->header_state = h_general; + } else if (parser->index == sizeof(UPGRADE)-2) { + parser->header_state = h_upgrade; + } + break; + + case h_connection: + case h_content_length: + case h_transfer_encoding: + case h_upgrade: + if (ch != ' ') parser->header_state = h_general; + break; + + default: + assert(0 && "Unknown header_state"); + break; + } + } + + COUNT_HEADER_SIZE(p - start); + + if (p == data + len) { + --p; + break; + } + + if (ch == ':') { + UPDATE_STATE(s_header_value_discard_ws); + CALLBACK_DATA(header_field); + break; + } + + SET_ERRNO(HPE_INVALID_HEADER_TOKEN); + goto error; + } + + case s_header_value_discard_ws: + if (ch == ' ' || ch == '\t') break; + + if (ch == CR) { + UPDATE_STATE(s_header_value_discard_ws_almost_done); + break; + } + + if (ch == LF) { + UPDATE_STATE(s_header_value_discard_lws); + break; + } + + /* FALLTHROUGH */ + + case s_header_value_start: + { + MARK(header_value); + + UPDATE_STATE(s_header_value); + parser->index = 0; + + c = LOWER(ch); + + switch (parser->header_state) { + case h_upgrade: + parser->flags |= F_UPGRADE; + parser->header_state = h_general; + break; + + case h_transfer_encoding: + /* looking for 'Transfer-Encoding: chunked' */ + if ('c' == c) { + parser->header_state = h_matching_transfer_encoding_chunked; + } else { + parser->header_state = h_general; + } + break; + + case h_content_length: + if (UNLIKELY(!IS_NUM(ch))) { + SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); + goto error; + } + + parser->content_length = ch - '0'; + break; + + case h_connection: + /* looking for 'Connection: keep-alive' */ + if (c == 'k') { + parser->header_state = h_matching_connection_keep_alive; + /* looking for 'Connection: close' */ + } else if (c == 'c') { + parser->header_state = h_matching_connection_close; + } else if (c == 'u') { + parser->header_state = h_matching_connection_upgrade; + } else { + parser->header_state = h_matching_connection_token; + } + break; + + /* Multi-value `Connection` header */ + case h_matching_connection_token_start: + break; + + default: + parser->header_state = h_general; + break; + } + break; + } + + case s_header_value: + { + const char* start = p; + enum header_states h_state = (enum header_states) parser->header_state; + for (; p != data + len; p++) { + ch = *p; + if (ch == CR) { + UPDATE_STATE(s_header_almost_done); + parser->header_state = h_state; + CALLBACK_DATA(header_value); + break; + } + + if (ch == LF) { + UPDATE_STATE(s_header_almost_done); + COUNT_HEADER_SIZE(p - start); + parser->header_state = h_state; + CALLBACK_DATA_NOADVANCE(header_value); + REEXECUTE(); + } + + if (!lenient && !IS_HEADER_CHAR(ch)) { + SET_ERRNO(HPE_INVALID_HEADER_TOKEN); + goto error; + } + + c = LOWER(ch); + + switch (h_state) { + case h_general: + { + const char* p_cr; + const char* p_lf; + size_t limit = data + len - p; + + limit = MIN(limit, HTTP_MAX_HEADER_SIZE); + + p_cr = (const char*) memchr(p, CR, limit); + p_lf = (const char*) memchr(p, LF, limit); + if (p_cr != NULL) { + if (p_lf != NULL && p_cr >= p_lf) + p = p_lf; + else + p = p_cr; + } else if (UNLIKELY(p_lf != NULL)) { + p = p_lf; + } else { + p = data + len; + } + --p; + + break; + } + + case h_connection: + case h_transfer_encoding: + assert(0 && "Shouldn't get here."); + break; + + case h_content_length: + { + uint64_t t; + + if (ch == ' ') break; + + if (UNLIKELY(!IS_NUM(ch))) { + SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); + parser->header_state = h_state; + goto error; + } + + t = parser->content_length; + t *= 10; + t += ch - '0'; + + /* Overflow? Test against a conservative limit for simplicity. */ + if (UNLIKELY((ULLONG_MAX - 10) / 10 < parser->content_length)) { + SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); + parser->header_state = h_state; + goto error; + } + + parser->content_length = t; + break; + } + + /* Transfer-Encoding: chunked */ + case h_matching_transfer_encoding_chunked: + parser->index++; + if (parser->index > sizeof(CHUNKED)-1 + || c != CHUNKED[parser->index]) { + h_state = h_general; + } else if (parser->index == sizeof(CHUNKED)-2) { + h_state = h_transfer_encoding_chunked; + } + break; + + case h_matching_connection_token_start: + /* looking for 'Connection: keep-alive' */ + if (c == 'k') { + h_state = h_matching_connection_keep_alive; + /* looking for 'Connection: close' */ + } else if (c == 'c') { + h_state = h_matching_connection_close; + } else if (c == 'u') { + h_state = h_matching_connection_upgrade; + } else if (STRICT_TOKEN(c)) { + h_state = h_matching_connection_token; + } else if (c == ' ' || c == '\t') { + /* Skip lws */ + } else { + h_state = h_general; + } + break; + + /* looking for 'Connection: keep-alive' */ + case h_matching_connection_keep_alive: + parser->index++; + if (parser->index > sizeof(KEEP_ALIVE)-1 + || c != KEEP_ALIVE[parser->index]) { + h_state = h_matching_connection_token; + } else if (parser->index == sizeof(KEEP_ALIVE)-2) { + h_state = h_connection_keep_alive; + } + break; + + /* looking for 'Connection: close' */ + case h_matching_connection_close: + parser->index++; + if (parser->index > sizeof(CLOSE)-1 || c != CLOSE[parser->index]) { + h_state = h_matching_connection_token; + } else if (parser->index == sizeof(CLOSE)-2) { + h_state = h_connection_close; + } + break; + + /* looking for 'Connection: upgrade' */ + case h_matching_connection_upgrade: + parser->index++; + if (parser->index > sizeof(UPGRADE) - 1 || + c != UPGRADE[parser->index]) { + h_state = h_matching_connection_token; + } else if (parser->index == sizeof(UPGRADE)-2) { + h_state = h_connection_upgrade; + } + break; + + case h_matching_connection_token: + if (ch == ',') { + h_state = h_matching_connection_token_start; + parser->index = 0; + } + break; + + case h_transfer_encoding_chunked: + if (ch != ' ') h_state = h_general; + break; + + case h_connection_keep_alive: + case h_connection_close: + case h_connection_upgrade: + if (ch == ',') { + if (h_state == h_connection_keep_alive) { + parser->flags |= F_CONNECTION_KEEP_ALIVE; + } else if (h_state == h_connection_close) { + parser->flags |= F_CONNECTION_CLOSE; + } else if (h_state == h_connection_upgrade) { + parser->flags |= F_CONNECTION_UPGRADE; + } + h_state = h_matching_connection_token_start; + parser->index = 0; + } else if (ch != ' ') { + h_state = h_matching_connection_token; + } + break; + + default: + UPDATE_STATE(s_header_value); + h_state = h_general; + break; + } + } + parser->header_state = h_state; + + COUNT_HEADER_SIZE(p - start); + + if (p == data + len) + --p; + break; + } + + case s_header_almost_done: + { + if (UNLIKELY(ch != LF)) { + SET_ERRNO(HPE_LF_EXPECTED); + goto error; + } + + UPDATE_STATE(s_header_value_lws); + break; + } + + case s_header_value_lws: + { + if (ch == ' ' || ch == '\t') { + UPDATE_STATE(s_header_value_start); + REEXECUTE(); + } + + /* finished the header */ + switch (parser->header_state) { + case h_connection_keep_alive: + parser->flags |= F_CONNECTION_KEEP_ALIVE; + break; + case h_connection_close: + parser->flags |= F_CONNECTION_CLOSE; + break; + case h_transfer_encoding_chunked: + parser->flags |= F_CHUNKED; + break; + case h_connection_upgrade: + parser->flags |= F_CONNECTION_UPGRADE; + break; + default: + break; + } + + UPDATE_STATE(s_header_field_start); + REEXECUTE(); + } + + case s_header_value_discard_ws_almost_done: + { + STRICT_CHECK(ch != LF); + UPDATE_STATE(s_header_value_discard_lws); + break; + } + + case s_header_value_discard_lws: + { + if (ch == ' ' || ch == '\t') { + UPDATE_STATE(s_header_value_discard_ws); + break; + } else { + switch (parser->header_state) { + case h_connection_keep_alive: + parser->flags |= F_CONNECTION_KEEP_ALIVE; + break; + case h_connection_close: + parser->flags |= F_CONNECTION_CLOSE; + break; + case h_connection_upgrade: + parser->flags |= F_CONNECTION_UPGRADE; + break; + case h_transfer_encoding_chunked: + parser->flags |= F_CHUNKED; + break; + default: + break; + } + + /* header value was empty */ + MARK(header_value); + UPDATE_STATE(s_header_field_start); + CALLBACK_DATA_NOADVANCE(header_value); + REEXECUTE(); + } + } + + case s_headers_almost_done: + { + STRICT_CHECK(ch != LF); + + if (parser->flags & F_TRAILING) { + /* End of a chunked request */ + UPDATE_STATE(s_message_done); + CALLBACK_NOTIFY_NOADVANCE(chunk_complete); + REEXECUTE(); + } + + /* Cannot use chunked encoding and a content-length header together + per the HTTP specification. */ + if ((parser->flags & F_CHUNKED) && + (parser->flags & F_CONTENTLENGTH)) { + SET_ERRNO(HPE_UNEXPECTED_CONTENT_LENGTH); + goto error; + } + + UPDATE_STATE(s_headers_done); + + /* Set this here so that on_headers_complete() callbacks can see it */ + parser->upgrade = + ((parser->flags & (F_UPGRADE | F_CONNECTION_UPGRADE)) == + (F_UPGRADE | F_CONNECTION_UPGRADE) || + parser->method == HTTP_CONNECT); + + /* Here we call the headers_complete callback. This is somewhat + * different than other callbacks because if the user returns 1, we + * will interpret that as saying that this message has no body. This + * is needed for the annoying case of recieving a response to a HEAD + * request. + * + * We'd like to use CALLBACK_NOTIFY_NOADVANCE() here but we cannot, so + * we have to simulate it by handling a change in errno below. + */ + if (settings->on_headers_complete) { + switch (settings->on_headers_complete(parser)) { + case 0: + break; + + case 2: + parser->upgrade = 1; + + case 1: + parser->flags |= F_SKIPBODY; + break; + + default: + SET_ERRNO(HPE_CB_headers_complete); + RETURN(p - data); /* Error */ + } + } + + if (HTTP_PARSER_ERRNO(parser) != HPE_OK) { + RETURN(p - data); + } + + REEXECUTE(); + } + + case s_headers_done: + { + int hasBody; + STRICT_CHECK(ch != LF); + + parser->nread = 0; + + hasBody = parser->flags & F_CHUNKED || + (parser->content_length > 0 && parser->content_length != ULLONG_MAX); + if (parser->upgrade && (parser->method == HTTP_CONNECT || + (parser->flags & F_SKIPBODY) || !hasBody)) { + /* Exit, the rest of the message is in a different protocol. */ + UPDATE_STATE(NEW_MESSAGE()); + CALLBACK_NOTIFY(message_complete); + RETURN((p - data) + 1); + } + + if (parser->flags & F_SKIPBODY) { + UPDATE_STATE(NEW_MESSAGE()); + CALLBACK_NOTIFY(message_complete); + } else if (parser->flags & F_CHUNKED) { + /* chunked encoding - ignore Content-Length header */ + UPDATE_STATE(s_chunk_size_start); + } else { + if (parser->content_length == 0) { + /* Content-Length header given but zero: Content-Length: 0\r\n */ + UPDATE_STATE(NEW_MESSAGE()); + CALLBACK_NOTIFY(message_complete); + } else if (parser->content_length != ULLONG_MAX) { + /* Content-Length header given and non-zero */ + UPDATE_STATE(s_body_identity); + } else { + if (!http_message_needs_eof(parser)) { + /* Assume content-length 0 - read the next */ + UPDATE_STATE(NEW_MESSAGE()); + CALLBACK_NOTIFY(message_complete); + } else { + /* Read body until EOF */ + UPDATE_STATE(s_body_identity_eof); + } + } + } + + break; + } + + case s_body_identity: + { + uint64_t to_read = MIN(parser->content_length, + (uint64_t) ((data + len) - p)); + + assert(parser->content_length != 0 + && parser->content_length != ULLONG_MAX); + + /* The difference between advancing content_length and p is because + * the latter will automaticaly advance on the next loop iteration. + * Further, if content_length ends up at 0, we want to see the last + * byte again for our message complete callback. + */ + MARK(body); + parser->content_length -= to_read; + p += to_read - 1; + + if (parser->content_length == 0) { + UPDATE_STATE(s_message_done); + + /* Mimic CALLBACK_DATA_NOADVANCE() but with one extra byte. + * + * The alternative to doing this is to wait for the next byte to + * trigger the data callback, just as in every other case. The + * problem with this is that this makes it difficult for the test + * harness to distinguish between complete-on-EOF and + * complete-on-length. It's not clear that this distinction is + * important for applications, but let's keep it for now. + */ + CALLBACK_DATA_(body, p - body_mark + 1, p - data); + REEXECUTE(); + } + + break; + } + + /* read until EOF */ + case s_body_identity_eof: + MARK(body); + p = data + len - 1; + + break; + + case s_message_done: + UPDATE_STATE(NEW_MESSAGE()); + CALLBACK_NOTIFY(message_complete); + if (parser->upgrade) { + /* Exit, the rest of the message is in a different protocol. */ + RETURN((p - data) + 1); + } + break; + + case s_chunk_size_start: + { + assert(parser->nread == 1); + assert(parser->flags & F_CHUNKED); + + unhex_val = unhex[(unsigned char)ch]; + if (UNLIKELY(unhex_val == -1)) { + SET_ERRNO(HPE_INVALID_CHUNK_SIZE); + goto error; + } + + parser->content_length = unhex_val; + UPDATE_STATE(s_chunk_size); + break; + } + + case s_chunk_size: + { + uint64_t t; + + assert(parser->flags & F_CHUNKED); + + if (ch == CR) { + UPDATE_STATE(s_chunk_size_almost_done); + break; + } + + unhex_val = unhex[(unsigned char)ch]; + + if (unhex_val == -1) { + if (ch == ';' || ch == ' ') { + UPDATE_STATE(s_chunk_parameters); + break; + } + + SET_ERRNO(HPE_INVALID_CHUNK_SIZE); + goto error; + } + + t = parser->content_length; + t *= 16; + t += unhex_val; + + /* Overflow? Test against a conservative limit for simplicity. */ + if (UNLIKELY((ULLONG_MAX - 16) / 16 < parser->content_length)) { + SET_ERRNO(HPE_INVALID_CONTENT_LENGTH); + goto error; + } + + parser->content_length = t; + break; + } + + case s_chunk_parameters: + { + assert(parser->flags & F_CHUNKED); + /* just ignore this shit. TODO check for overflow */ + if (ch == CR) { + UPDATE_STATE(s_chunk_size_almost_done); + break; + } + break; + } + + case s_chunk_size_almost_done: + { + assert(parser->flags & F_CHUNKED); + STRICT_CHECK(ch != LF); + + parser->nread = 0; + + if (parser->content_length == 0) { + parser->flags |= F_TRAILING; + UPDATE_STATE(s_header_field_start); + } else { + UPDATE_STATE(s_chunk_data); + } + CALLBACK_NOTIFY(chunk_header); + break; + } + + case s_chunk_data: + { + uint64_t to_read = MIN(parser->content_length, + (uint64_t) ((data + len) - p)); + + assert(parser->flags & F_CHUNKED); + assert(parser->content_length != 0 + && parser->content_length != ULLONG_MAX); + + /* See the explanation in s_body_identity for why the content + * length and data pointers are managed this way. + */ + MARK(body); + parser->content_length -= to_read; + p += to_read - 1; + + if (parser->content_length == 0) { + UPDATE_STATE(s_chunk_data_almost_done); + } + + break; + } + + case s_chunk_data_almost_done: + assert(parser->flags & F_CHUNKED); + assert(parser->content_length == 0); + STRICT_CHECK(ch != CR); + UPDATE_STATE(s_chunk_data_done); + CALLBACK_DATA(body); + break; + + case s_chunk_data_done: + assert(parser->flags & F_CHUNKED); + STRICT_CHECK(ch != LF); + parser->nread = 0; + UPDATE_STATE(s_chunk_size_start); + CALLBACK_NOTIFY(chunk_complete); + break; + + default: + assert(0 && "unhandled state"); + SET_ERRNO(HPE_INVALID_INTERNAL_STATE); + goto error; + } + } + + /* Run callbacks for any marks that we have leftover after we ran our of + * bytes. There should be at most one of these set, so it's OK to invoke + * them in series (unset marks will not result in callbacks). + * + * We use the NOADVANCE() variety of callbacks here because 'p' has already + * overflowed 'data' and this allows us to correct for the off-by-one that + * we'd otherwise have (since CALLBACK_DATA() is meant to be run with a 'p' + * value that's in-bounds). + */ + + assert(((header_field_mark ? 1 : 0) + + (header_value_mark ? 1 : 0) + + (url_mark ? 1 : 0) + + (body_mark ? 1 : 0) + + (status_mark ? 1 : 0)) <= 1); + + CALLBACK_DATA_NOADVANCE(header_field); + CALLBACK_DATA_NOADVANCE(header_value); + CALLBACK_DATA_NOADVANCE(url); + CALLBACK_DATA_NOADVANCE(body); + CALLBACK_DATA_NOADVANCE(status); + + RETURN(len); + +error: + if (HTTP_PARSER_ERRNO(parser) == HPE_OK) { + SET_ERRNO(HPE_UNKNOWN); + } + + RETURN(p - data); +} + + +/* Does the parser need to see an EOF to find the end of the message? */ +int +http_message_needs_eof (const http_parser *parser) +{ + if (parser->type == HTTP_REQUEST) { + return 0; + } + + /* See RFC 2616 section 4.4 */ + if (parser->status_code / 100 == 1 || /* 1xx e.g. Continue */ + parser->status_code == 204 || /* No Content */ + parser->status_code == 304 || /* Not Modified */ + parser->flags & F_SKIPBODY) { /* response to a HEAD request */ + return 0; + } + + if ((parser->flags & F_CHUNKED) || parser->content_length != ULLONG_MAX) { + return 0; + } + + return 1; +} + + +int +http_should_keep_alive (const http_parser *parser) +{ + if (parser->http_major > 0 && parser->http_minor > 0) { + /* HTTP/1.1 */ + if (parser->flags & F_CONNECTION_CLOSE) { + return 0; + } + } else { + /* HTTP/1.0 or earlier */ + if (!(parser->flags & F_CONNECTION_KEEP_ALIVE)) { + return 0; + } + } + + return !http_message_needs_eof(parser); +} + + +const char * +http_method_str (enum http_method m) +{ + return ELEM_AT(method_strings, m, "<unknown>"); +} + + +void +http_parser_init (http_parser *parser, enum http_parser_type t) +{ + void *data = parser->data; /* preserve application data */ + memset(parser, 0, sizeof(*parser)); + parser->data = data; + parser->type = t; + parser->state = (t == HTTP_REQUEST ? s_start_req : (t == HTTP_RESPONSE ? s_start_res : s_start_req_or_res)); + parser->http_errno = HPE_OK; +} + +void +http_parser_settings_init(http_parser_settings *settings) +{ + memset(settings, 0, sizeof(*settings)); +} + +const char * +http_errno_name(enum http_errno err) { + assert(((size_t) err) < ARRAY_SIZE(http_strerror_tab)); + return http_strerror_tab[err].name; +} + +const char * +http_errno_description(enum http_errno err) { + assert(((size_t) err) < ARRAY_SIZE(http_strerror_tab)); + return http_strerror_tab[err].description; +} + +static enum http_host_state +http_parse_host_char(enum http_host_state s, const char ch) { + switch(s) { + case s_http_userinfo: + case s_http_userinfo_start: + if (ch == '@') { + return s_http_host_start; + } + + if (IS_USERINFO_CHAR(ch)) { + return s_http_userinfo; + } + break; + + case s_http_host_start: + if (ch == '[') { + return s_http_host_v6_start; + } + + if (IS_HOST_CHAR(ch)) { + return s_http_host; + } + + break; + + case s_http_host: + if (IS_HOST_CHAR(ch)) { + return s_http_host; + } + + /* FALLTHROUGH */ + case s_http_host_v6_end: + if (ch == ':') { + return s_http_host_port_start; + } + + break; + + case s_http_host_v6: + if (ch == ']') { + return s_http_host_v6_end; + } + + /* FALLTHROUGH */ + case s_http_host_v6_start: + if (IS_HEX(ch) || ch == ':' || ch == '.') { + return s_http_host_v6; + } + + if (s == s_http_host_v6 && ch == '%') { + return s_http_host_v6_zone_start; + } + break; + + case s_http_host_v6_zone: + if (ch == ']') { + return s_http_host_v6_end; + } + + /* FALLTHROUGH */ + case s_http_host_v6_zone_start: + /* RFC 6874 Zone ID consists of 1*( unreserved / pct-encoded) */ + if (IS_ALPHANUM(ch) || ch == '%' || ch == '.' || ch == '-' || ch == '_' || + ch == '~') { + return s_http_host_v6_zone; + } + break; + + case s_http_host_port: + case s_http_host_port_start: + if (IS_NUM(ch)) { + return s_http_host_port; + } + + break; + + default: + break; + } + return s_http_host_dead; +} + +static int +http_parse_host(const char * buf, struct http_parser_url *u, int found_at) { + assert(u->field_set & (1 << UF_HOST)); + enum http_host_state s; + + const char *p; + size_t buflen = u->field_data[UF_HOST].off + u->field_data[UF_HOST].len; + + u->field_data[UF_HOST].len = 0; + + s = found_at ? s_http_userinfo_start : s_http_host_start; + + for (p = buf + u->field_data[UF_HOST].off; p < buf + buflen; p++) { + enum http_host_state new_s = http_parse_host_char(s, *p); + + if (new_s == s_http_host_dead) { + return 1; + } + + switch(new_s) { + case s_http_host: + if (s != s_http_host) { + u->field_data[UF_HOST].off = p - buf; + } + u->field_data[UF_HOST].len++; + break; + + case s_http_host_v6: + if (s != s_http_host_v6) { + u->field_data[UF_HOST].off = p - buf; + } + u->field_data[UF_HOST].len++; + break; + + case s_http_host_v6_zone_start: + case s_http_host_v6_zone: + u->field_data[UF_HOST].len++; + break; + + case s_http_host_port: + if (s != s_http_host_port) { + u->field_data[UF_PORT].off = p - buf; + u->field_data[UF_PORT].len = 0; + u->field_set |= (1 << UF_PORT); + } + u->field_data[UF_PORT].len++; + break; + + case s_http_userinfo: + if (s != s_http_userinfo) { + u->field_data[UF_USERINFO].off = p - buf ; + u->field_data[UF_USERINFO].len = 0; + u->field_set |= (1 << UF_USERINFO); + } + u->field_data[UF_USERINFO].len++; + break; + + default: + break; + } + s = new_s; + } + + /* Make sure we don't end somewhere unexpected */ + switch (s) { + case s_http_host_start: + case s_http_host_v6_start: + case s_http_host_v6: + case s_http_host_v6_zone_start: + case s_http_host_v6_zone: + case s_http_host_port_start: + case s_http_userinfo: + case s_http_userinfo_start: + return 1; + default: + break; + } + + return 0; +} + +void +http_parser_url_init(struct http_parser_url *u) { + memset(u, 0, sizeof(*u)); +} + +int +http_parser_parse_url(const char *buf, size_t buflen, int is_connect, + struct http_parser_url *u) +{ + enum state s; + const char *p; + enum http_parser_url_fields uf, old_uf; + int found_at = 0; + + u->port = u->field_set = 0; + s = is_connect ? s_req_server_start : s_req_spaces_before_url; + old_uf = UF_MAX; + + for (p = buf; p < buf + buflen; p++) { + s = parse_url_char(s, *p); + + /* Figure out the next field that we're operating on */ + switch (s) { + case s_dead: + return 1; + + /* Skip delimeters */ + case s_req_schema_slash: + case s_req_schema_slash_slash: + case s_req_server_start: + case s_req_query_string_start: + case s_req_fragment_start: + continue; + + case s_req_schema: + uf = UF_SCHEMA; + break; + + case s_req_server_with_at: + found_at = 1; + + /* FALLTROUGH */ + case s_req_server: + uf = UF_HOST; + break; + + case s_req_path: + uf = UF_PATH; + break; + + case s_req_query_string: + uf = UF_QUERY; + break; + + case s_req_fragment: + uf = UF_FRAGMENT; + break; + + default: + assert(!"Unexpected state"); + return 1; + } + + /* Nothing's changed; soldier on */ + if (uf == old_uf) { + u->field_data[uf].len++; + continue; + } + + u->field_data[uf].off = p - buf; + u->field_data[uf].len = 1; + + u->field_set |= (1 << uf); + old_uf = uf; + } + + /* host must be present if there is a schema */ + /* parsing http:///toto will fail */ + if ((u->field_set & (1 << UF_SCHEMA)) && + (u->field_set & (1 << UF_HOST)) == 0) { + return 1; + } + + if (u->field_set & (1 << UF_HOST)) { + if (http_parse_host(buf, u, found_at) != 0) { + return 1; + } + } + + /* CONNECT requests can only contain "hostname:port" */ + if (is_connect && u->field_set != ((1 << UF_HOST)|(1 << UF_PORT))) { + return 1; + } + + if (u->field_set & (1 << UF_PORT)) { + /* Don't bother with endp; we've already validated the string */ + unsigned long v = strtoul(buf + u->field_data[UF_PORT].off, NULL, 10); + + /* Ports have a max value of 2^16 */ + if (v > 0xffff) { + return 1; + } + + u->port = (uint16_t) v; + } + + return 0; +} + +void +http_parser_pause(http_parser *parser, int paused) { + /* Users should only be pausing/unpausing a parser that is not in an error + * state. In non-debug builds, there's not much that we can do about this + * other than ignore it. + */ + if (HTTP_PARSER_ERRNO(parser) == HPE_OK || + HTTP_PARSER_ERRNO(parser) == HPE_PAUSED) { + SET_ERRNO((paused) ? HPE_PAUSED : HPE_OK); + } else { + assert(0 && "Attempting to pause parser in error state"); + } +} + +int +http_body_is_final(const struct http_parser *parser) { + return parser->state == s_message_done; +} + +unsigned long +http_parser_version(void) { + return HTTP_PARSER_VERSION_MAJOR * 0x10000 | + HTTP_PARSER_VERSION_MINOR * 0x00100 | + HTTP_PARSER_VERSION_PATCH * 0x00001; +} diff --git a/src/boost/libs/beast/test/bench/parser/nodejs-parser/http_parser.h b/src/boost/libs/beast/test/bench/parser/nodejs-parser/http_parser.h new file mode 100644 index 00000000..105ae510 --- /dev/null +++ b/src/boost/libs/beast/test/bench/parser/nodejs-parser/http_parser.h @@ -0,0 +1,362 @@ +/* Copyright Joyent, Inc. and other Node contributors. All rights reserved. + * + * Permission is hereby granted, free of charge, to any person obtaining a copy + * of this software and associated documentation files (the "Software"), to + * deal in the Software without restriction, including without limitation the + * rights to use, copy, modify, merge, publish, distribute, sublicense, and/or + * sell copies of the Software, and to permit persons to whom the Software is + * furnished to do so, subject to the following conditions: + * + * The above copyright notice and this permission notice shall be included in + * all copies or substantial portions of the Software. + * + * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR + * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, + * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE + * AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER + * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING + * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS + * IN THE SOFTWARE. + */ +#ifndef http_parser_h +#define http_parser_h +#ifdef __cplusplus +extern "C" { +#endif + +/* Also update SONAME in the Makefile whenever you change these. */ +#define HTTP_PARSER_VERSION_MAJOR 2 +#define HTTP_PARSER_VERSION_MINOR 7 +#define HTTP_PARSER_VERSION_PATCH 0 + +#include <sys/types.h> +#if defined(_WIN32) && !defined(__MINGW32__) && \ + (!defined(_MSC_VER) || _MSC_VER<1600) && !defined(__WINE__) +#include <BaseTsd.h> +#include <stddef.h> +typedef __int8 int8_t; +typedef unsigned __int8 uint8_t; +typedef __int16 int16_t; +typedef unsigned __int16 uint16_t; +typedef __int32 int32_t; +typedef unsigned __int32 uint32_t; +typedef __int64 int64_t; +typedef unsigned __int64 uint64_t; +#else +#include <stdint.h> +#endif + +/* Compile with -DHTTP_PARSER_STRICT=0 to make less checks, but run + * faster + */ +#ifndef HTTP_PARSER_STRICT +# define HTTP_PARSER_STRICT 1 +#endif + +/* Maximium header size allowed. If the macro is not defined + * before including this header then the default is used. To + * change the maximum header size, define the macro in the build + * environment (e.g. -DHTTP_MAX_HEADER_SIZE=<value>). To remove + * the effective limit on the size of the header, define the macro + * to a very large number (e.g. -DHTTP_MAX_HEADER_SIZE=0x7fffffff) + */ +#ifndef HTTP_MAX_HEADER_SIZE +# define HTTP_MAX_HEADER_SIZE (80*1024) +#endif + +typedef struct http_parser http_parser; +typedef struct http_parser_settings http_parser_settings; + + +/* Callbacks should return non-zero to indicate an error. The parser will + * then halt execution. + * + * The one exception is on_headers_complete. In a HTTP_RESPONSE parser + * returning '1' from on_headers_complete will tell the parser that it + * should not expect a body. This is used when receiving a response to a + * HEAD request which may contain 'Content-Length' or 'Transfer-Encoding: + * chunked' headers that indicate the presence of a body. + * + * Returning `2` from on_headers_complete will tell parser that it should not + * expect neither a body nor any futher responses on this connection. This is + * useful for handling responses to a CONNECT request which may not contain + * `Upgrade` or `Connection: upgrade` headers. + * + * http_data_cb does not return data chunks. It will be called arbitrarily + * many times for each string. E.G. you might get 10 callbacks for "on_url" + * each providing just a few characters more data. + */ +typedef int (*http_data_cb) (http_parser*, const char *at, size_t length); +typedef int (*http_cb) (http_parser*); + + +/* Request Methods */ +#define HTTP_METHOD_MAP(XX) \ + XX(0, DELETE, DELETE) \ + XX(1, GET, GET) \ + XX(2, HEAD, HEAD) \ + XX(3, POST, POST) \ + XX(4, PUT, PUT) \ + /* pathological */ \ + XX(5, CONNECT, CONNECT) \ + XX(6, OPTIONS, OPTIONS) \ + XX(7, TRACE, TRACE) \ + /* WebDAV */ \ + XX(8, COPY, COPY) \ + XX(9, LOCK, LOCK) \ + XX(10, MKCOL, MKCOL) \ + XX(11, MOVE, MOVE) \ + XX(12, PROPFIND, PROPFIND) \ + XX(13, PROPPATCH, PROPPATCH) \ + XX(14, SEARCH, SEARCH) \ + XX(15, UNLOCK, UNLOCK) \ + XX(16, BIND, BIND) \ + XX(17, REBIND, REBIND) \ + XX(18, UNBIND, UNBIND) \ + XX(19, ACL, ACL) \ + /* subversion */ \ + XX(20, REPORT, REPORT) \ + XX(21, MKACTIVITY, MKACTIVITY) \ + XX(22, CHECKOUT, CHECKOUT) \ + XX(23, MERGE, MERGE) \ + /* upnp */ \ + XX(24, MSEARCH, M-SEARCH) \ + XX(25, NOTIFY, NOTIFY) \ + XX(26, SUBSCRIBE, SUBSCRIBE) \ + XX(27, UNSUBSCRIBE, UNSUBSCRIBE) \ + /* RFC-5789 */ \ + XX(28, PATCH, PATCH) \ + XX(29, PURGE, PURGE) \ + /* CalDAV */ \ + XX(30, MKCALENDAR, MKCALENDAR) \ + /* RFC-2068, section 19.6.1.2 */ \ + XX(31, LINK, LINK) \ + XX(32, UNLINK, UNLINK) \ + +enum http_method + { +#define XX(num, name, string) HTTP_##name = num, + HTTP_METHOD_MAP(XX) +#undef XX + }; + + +enum http_parser_type { HTTP_REQUEST, HTTP_RESPONSE, HTTP_BOTH }; + + +/* Flag values for http_parser.flags field */ +enum flags + { F_CHUNKED = 1 << 0 + , F_CONNECTION_KEEP_ALIVE = 1 << 1 + , F_CONNECTION_CLOSE = 1 << 2 + , F_CONNECTION_UPGRADE = 1 << 3 + , F_TRAILING = 1 << 4 + , F_UPGRADE = 1 << 5 + , F_SKIPBODY = 1 << 6 + , F_CONTENTLENGTH = 1 << 7 + }; + + +/* Map for errno-related constants + * + * The provided argument should be a macro that takes 2 arguments. + */ +#define HTTP_ERRNO_MAP(XX) \ + /* No error */ \ + XX(OK, "success") \ + \ + /* Callback-related errors */ \ + XX(CB_message_begin, "the on_message_begin callback failed") \ + XX(CB_url, "the on_url callback failed") \ + XX(CB_header_field, "the on_header_field callback failed") \ + XX(CB_header_value, "the on_header_value callback failed") \ + XX(CB_headers_complete, "the on_headers_complete callback failed") \ + XX(CB_body, "the on_body callback failed") \ + XX(CB_message_complete, "the on_message_complete callback failed") \ + XX(CB_status, "the on_status callback failed") \ + XX(CB_chunk_header, "the on_chunk_header callback failed") \ + XX(CB_chunk_complete, "the on_chunk_complete callback failed") \ + \ + /* Parsing-related errors */ \ + XX(INVALID_EOF_STATE, "stream ended at an unexpected time") \ + XX(HEADER_OVERFLOW, \ + "too many header bytes seen; overflow detected") \ + XX(CLOSED_CONNECTION, \ + "data received after completed connection: close message") \ + XX(INVALID_VERSION, "invalid HTTP version") \ + XX(INVALID_STATUS, "invalid HTTP status code") \ + XX(INVALID_METHOD, "invalid HTTP method") \ + XX(INVALID_URL, "invalid URL") \ + XX(INVALID_HOST, "invalid host") \ + XX(INVALID_PORT, "invalid port") \ + XX(INVALID_PATH, "invalid path") \ + XX(INVALID_QUERY_STRING, "invalid query string") \ + XX(INVALID_FRAGMENT, "invalid fragment") \ + XX(LF_EXPECTED, "LF character expected") \ + XX(INVALID_HEADER_TOKEN, "invalid character in header") \ + XX(INVALID_CONTENT_LENGTH, \ + "invalid character in content-length header") \ + XX(UNEXPECTED_CONTENT_LENGTH, \ + "unexpected content-length header") \ + XX(INVALID_CHUNK_SIZE, \ + "invalid character in chunk size header") \ + XX(INVALID_CONSTANT, "invalid constant string") \ + XX(INVALID_INTERNAL_STATE, "encountered unexpected internal state")\ + XX(STRICT, "strict mode assertion failed") \ + XX(PAUSED, "parser is paused") \ + XX(UNKNOWN, "an unknown error occurred") + + +/* Define HPE_* values for each errno value above */ +#define HTTP_ERRNO_GEN(n, s) HPE_##n, +enum http_errno { + HTTP_ERRNO_MAP(HTTP_ERRNO_GEN) +}; +#undef HTTP_ERRNO_GEN + + +/* Get an http_errno value from an http_parser */ +#define HTTP_PARSER_ERRNO(p) ((enum http_errno) (p)->http_errno) + + +struct http_parser { + /** PRIVATE **/ + unsigned int type : 2; /* enum http_parser_type */ + unsigned int flags : 8; /* F_* values from 'flags' enum; semi-public */ + unsigned int state : 7; /* enum state from http_parser.c */ + unsigned int header_state : 7; /* enum header_state from http_parser.c */ + unsigned int index : 7; /* index into current matcher */ + unsigned int lenient_http_headers : 1; + + uint32_t nread; /* # bytes read in various scenarios */ + uint64_t content_length; /* # bytes in body (0 if no Content-Length header) */ + + /** READ-ONLY **/ + unsigned short http_major; + unsigned short http_minor; + unsigned int status_code : 16; /* responses only */ + unsigned int method : 8; /* requests only */ + unsigned int http_errno : 7; + + /* 1 = Upgrade header was present and the parser has exited because of that. + * 0 = No upgrade header present. + * Should be checked when http_parser_execute() returns in addition to + * error checking. + */ + unsigned int upgrade : 1; + + /** PUBLIC **/ + void *data; /* A pointer to get hook to the "connection" or "socket" object */ +}; + + +struct http_parser_settings { + http_cb on_message_begin; + http_data_cb on_url; + http_data_cb on_status; + http_data_cb on_header_field; + http_data_cb on_header_value; + http_cb on_headers_complete; + http_data_cb on_body; + http_cb on_message_complete; + /* When on_chunk_header is called, the current chunk length is stored + * in parser->content_length. + */ + http_cb on_chunk_header; + http_cb on_chunk_complete; +}; + + +enum http_parser_url_fields + { UF_SCHEMA = 0 + , UF_HOST = 1 + , UF_PORT = 2 + , UF_PATH = 3 + , UF_QUERY = 4 + , UF_FRAGMENT = 5 + , UF_USERINFO = 6 + , UF_MAX = 7 + }; + + +/* Result structure for http_parser_parse_url(). + * + * Callers should index into field_data[] with UF_* values iff field_set + * has the relevant (1 << UF_*) bit set. As a courtesy to clients (and + * because we probably have padding left over), we convert any port to + * a uint16_t. + */ +struct http_parser_url { + uint16_t field_set; /* Bitmask of (1 << UF_*) values */ + uint16_t port; /* Converted UF_PORT string */ + + struct { + uint16_t off; /* Offset into buffer in which field starts */ + uint16_t len; /* Length of run in buffer */ + } field_data[UF_MAX]; +}; + + +/* Returns the library version. Bits 16-23 contain the major version number, + * bits 8-15 the minor version number and bits 0-7 the patch level. + * Usage example: + * + * unsigned long version = http_parser_version(); + * unsigned major = (version >> 16) & 255; + * unsigned minor = (version >> 8) & 255; + * unsigned patch = version & 255; + * printf("http_parser v%u.%u.%u\n", major, minor, patch); + */ +unsigned long http_parser_version(void); + +void http_parser_init(http_parser *parser, enum http_parser_type type); + + +/* Initialize http_parser_settings members to 0 + */ +void http_parser_settings_init(http_parser_settings *settings); + + +/* Executes the parser. Returns number of parsed bytes. Sets + * `parser->http_errno` on error. */ +size_t http_parser_execute(http_parser *parser, + const http_parser_settings *settings, + const char *data, + size_t len); + + +/* If http_should_keep_alive() in the on_headers_complete or + * on_message_complete callback returns 0, then this should be + * the last message on the connection. + * If you are the server, respond with the "Connection: close" header. + * If you are the client, close the connection. + */ +int http_should_keep_alive(const http_parser *parser); + +/* Returns a string version of the HTTP method. */ +const char *http_method_str(enum http_method m); + +/* Return a string name of the given error */ +const char *http_errno_name(enum http_errno err); + +/* Return a string description of the given error */ +const char *http_errno_description(enum http_errno err); + +/* Initialize all http_parser_url members to 0 */ +void http_parser_url_init(struct http_parser_url *u); + +/* Parse a URL; return nonzero on failure */ +int http_parser_parse_url(const char *buf, size_t buflen, + int is_connect, + struct http_parser_url *u); + +/* Pause or un-pause the parser; a nonzero value pauses */ +void http_parser_pause(http_parser *parser, int paused); + +/* Checks if this is the final chunk of the body. */ +int http_body_is_final(const http_parser *parser); + +#ifdef __cplusplus +} +#endif +#endif diff --git a/src/boost/libs/beast/test/bench/parser/nodejs_parser.cpp b/src/boost/libs/beast/test/bench/parser/nodejs_parser.cpp new file mode 100644 index 00000000..cc14b981 --- /dev/null +++ b/src/boost/libs/beast/test/bench/parser/nodejs_parser.cpp @@ -0,0 +1,31 @@ +// +// Copyright (c) 2016-2019 Vinnie Falco (vinnie dot falco at gmail dot com) +// +// Distributed under the Boost Software License, Version 1.0. (See accompanying +// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +// +// Official repository: https://github.com/boostorg/beast +// + +#if defined(__GNUC__) && (__GNUC__ >= 7) +#pragma GCC diagnostic push +#pragma GCC diagnostic ignored "-Wimplicit-fallthrough" +#endif + +#include <boost/config.hpp> + +#ifdef BOOST_MSVC +# pragma warning (push) +# pragma warning (disable: 4127) // conditional expression is constant +# pragma warning (disable: 4244) // integer conversion, possible loss of data +#endif + +#include "nodejs-parser/http_parser.c" + +#ifdef BOOST_MSVC +# pragma warning (pop) +#endif + +#if defined(__GNUC__) && (__GNUC__ >= 7) +#pragma GCC diagnostic pop +#endif diff --git a/src/boost/libs/beast/test/bench/parser/nodejs_parser.hpp b/src/boost/libs/beast/test/bench/parser/nodejs_parser.hpp new file mode 100644 index 00000000..68afd6f5 --- /dev/null +++ b/src/boost/libs/beast/test/bench/parser/nodejs_parser.hpp @@ -0,0 +1,640 @@ +// +// Copyright (c) 2016-2019 Vinnie Falco (vinnie dot falco at gmail dot com) +// +// Distributed under the Boost Software License, Version 1.0. (See accompanying +// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +// +// Official repository: https://github.com/boostorg/beast +// + +#ifndef BOOST_BEAST_HTTP_NODEJS_PARSER_HPP +#define BOOST_BEAST_HTTP_NODEJS_PARSER_HPP + +#include "nodejs-parser/http_parser.h" + +#include <boost/beast/http/message.hpp> +#include <boost/beast/http/rfc7230.hpp> +#include <boost/beast/core/buffers_range.hpp> +#include <boost/beast/core/error.hpp> +#include <boost/asio/buffer.hpp> +#include <boost/system/error_code.hpp> +#include <boost/throw_exception.hpp> +#include <cstdint> +#include <string> +#include <type_traits> +#include <utility> + +namespace boost { +namespace beast { +namespace http { + +namespace detail { + +class nodejs_message_category final + : public boost::system::error_category +{ +public: + const char* + name() const noexcept override + { + return "nodejs-http-error"; + } + + std::string + message(int ev) const override + { + return http_errno_description( + static_cast<http_errno>(ev)); + } + + boost::system::error_condition + default_error_condition(int ev) const noexcept override + { + return boost::system::error_condition{ev, *this}; + } + + bool + equivalent(int ev, + boost::system::error_condition const& condition + ) const noexcept override + { + return condition.value() == ev && + &condition.category() == this; + } + + bool + equivalent(boost::system::error_code const& error, + int ev) const noexcept override + { + return error.value() == ev && + &error.category() == this; + } +}; + +template<class = void> +boost::system::error_code +make_nodejs_error(int http_errno) +{ + static nodejs_message_category const mc{}; + return boost::system::error_code{http_errno, mc}; +} + +inline +char const* +method_to_string(unsigned method) +{ + using namespace beast; + switch(static_cast<http_method>(method)) + { + case HTTP_DELETE: return "DELETE"; + case HTTP_GET: return "GET"; + case HTTP_HEAD: return "HEAD"; + case HTTP_POST: return "POST"; + case HTTP_PUT: return "PUT"; + + // pathological + case HTTP_CONNECT: return "CONNECT"; + case HTTP_OPTIONS: return "OPTIONS"; + case HTTP_TRACE: return "TRACE"; + + // webdav + case HTTP_COPY: return "COPY"; + case HTTP_LOCK: return "LOCK"; + case HTTP_MKCOL: return "MKCOL"; + case HTTP_MOVE: return "MOVE"; + case HTTP_PROPFIND: return "PROPFIND"; + case HTTP_PROPPATCH: return "PROPPATCH"; + case HTTP_SEARCH: return "SEARCH"; + case HTTP_UNLOCK: return "UNLOCK"; + case HTTP_BIND: return "BIND"; + case HTTP_REBIND: return "REBIND"; + case HTTP_UNBIND: return "UNBIND"; + case HTTP_ACL: return "ACL"; + + // subversion + case HTTP_REPORT: return "REPORT"; + case HTTP_MKACTIVITY: return "MKACTIVITY"; + case HTTP_CHECKOUT: return "CHECKOUT"; + case HTTP_MERGE: return "MERGE"; + + // upnp + case HTTP_MSEARCH: return "MSEARCH"; + case HTTP_NOTIFY: return "NOTIFY"; + case HTTP_SUBSCRIBE: return "SUBSCRIBE"; + case HTTP_UNSUBSCRIBE: return "UNSUBSCRIBE"; + + // RFC-5789 + case HTTP_PATCH: return "PATCH"; + case HTTP_PURGE: return "PURGE"; + + // CalDav + case HTTP_MKCALENDAR: return "MKCALENDAR"; + + // RFC-2068, section 19.6.1.2 + case HTTP_LINK: return "LINK"; + case HTTP_UNLINK: return "UNLINK"; + }; + + return "<unknown>"; +} + +} // detail + +template<class Derived> +class nodejs_basic_parser +{ + http_parser state_; + boost::system::error_code* ec_; + bool complete_ = false; + std::string url_; + std::string status_; + std::string field_; + std::string value_; + +public: + using error_code = boost::system::error_code; + + nodejs_basic_parser(nodejs_basic_parser&& other); + + nodejs_basic_parser& + operator=(nodejs_basic_parser&& other); + + nodejs_basic_parser(nodejs_basic_parser const& other); + + nodejs_basic_parser& operator=(nodejs_basic_parser const& other); + + explicit + nodejs_basic_parser(bool request) noexcept; + + bool + complete() const noexcept + { + return complete_; + } + + std::size_t + write(void const* data, std::size_t size) + { + error_code ec; + auto const used = write(data, size, ec); + if(ec) + BOOST_THROW_EXCEPTION(system_error{ec}); + return used; + } + + std::size_t + write(void const* data, std::size_t size, + error_code& ec); + + template<class ConstBufferSequence> + std::size_t + write(ConstBufferSequence const& buffers) + { + error_code ec; + auto const used = write(buffers, ec); + if(ec) + BOOST_THROW_EXCEPTION(system_error{ec}); + return used; + } + + template<class ConstBufferSequence> + std::size_t + write(ConstBufferSequence const& buffers, + error_code& ec); + + void + write_eof() + { + error_code ec; + write_eof(ec); + if(ec) + BOOST_THROW_EXCEPTION(system_error{ec}); + } + + void + write_eof(error_code& ec); + + void + check_header(); + +private: + Derived& + impl() + { + return *static_cast<Derived*>(this); + } + + static int cb_message_start(http_parser*); + static int cb_url(http_parser*, char const*, std::size_t); + static int cb_status(http_parser*, char const*, std::size_t); + static int cb_header_field(http_parser*, char const*, std::size_t); + static int cb_header_value(http_parser*, char const*, std::size_t); + static int cb_headers_complete(http_parser*); + static int cb_body(http_parser*, char const*, std::size_t); + static int cb_message_complete(http_parser*); + static int cb_chunk_header(http_parser*); + static int cb_chunk_complete(http_parser*); + + struct hooks_t : http_parser_settings + { + hooks_t() + { + http_parser_settings_init(this); + on_message_begin = &nodejs_basic_parser::cb_message_start; + on_url = &nodejs_basic_parser::cb_url; + on_status = &nodejs_basic_parser::cb_status; + on_header_field = &nodejs_basic_parser::cb_header_field; + on_header_value = &nodejs_basic_parser::cb_header_value; + on_headers_complete = &nodejs_basic_parser::cb_headers_complete; + on_body = &nodejs_basic_parser::cb_body; + on_message_complete = &nodejs_basic_parser::cb_message_complete; + on_chunk_header = &nodejs_basic_parser::cb_chunk_header; + on_chunk_complete = &nodejs_basic_parser::cb_chunk_complete; + } + }; + + static + http_parser_settings const* + hooks(); +}; + +template<class Derived> +template<class ConstBufferSequence> +std::size_t +nodejs_basic_parser<Derived>::write( + ConstBufferSequence const& buffers, error_code& ec) +{ + static_assert(net::is_const_buffer_sequence< + ConstBufferSequence>::value, + "ConstBufferSequence type requirements not met"); + std::size_t bytes_used = 0; + for(auto buffer : beast::buffers_range_ref(buffers)) + { + auto const n = write( + static_cast<void const*>(buffer.data()), + buffer.size(), ec); + if(ec) + return 0; + bytes_used += n; + if(complete()) + break; + } + return bytes_used; +} + +template<class Derived> +http_parser_settings const* +nodejs_basic_parser<Derived>::hooks() +{ + static hooks_t const h; + return &h; +} + +template<class Derived> +nodejs_basic_parser<Derived>:: +nodejs_basic_parser(nodejs_basic_parser&& other) +{ + state_ = other.state_; + state_.data = this; + complete_ = other.complete_; + url_ = std::move(other.url_); + status_ = std::move(other.status_); + field_ = std::move(other.field_); + value_ = std::move(other.value_); +} + +template<class Derived> +auto +nodejs_basic_parser<Derived>:: +operator=(nodejs_basic_parser&& other) -> + nodejs_basic_parser& +{ + state_ = other.state_; + state_.data = this; + complete_ = other.complete_; + url_ = std::move(other.url_); + status_ = std::move(other.status_); + field_ = std::move(other.field_); + value_ = std::move(other.value_); + return *this; +} + +template<class Derived> +nodejs_basic_parser<Derived>:: +nodejs_basic_parser(nodejs_basic_parser const& other) +{ + state_ = other.state_; + state_.data = this; + complete_ = other.complete_; + url_ = other.url_; + status_ = other.status_; + field_ = other.field_; + value_ = other.value_; +} + +template<class Derived> +auto +nodejs_basic_parser<Derived>:: +operator=(nodejs_basic_parser const& other) -> + nodejs_basic_parser& +{ + state_ = other.state_; + state_.data = this; + complete_ = other.complete_; + url_ = other.url_; + status_ = other.status_; + field_ = other.field_; + value_ = other.value_; + return *this; +} + +template<class Derived> +nodejs_basic_parser<Derived>:: +nodejs_basic_parser(bool request) noexcept +{ + state_.data = this; + http_parser_init(&state_, request + ? http_parser_type::HTTP_REQUEST + : http_parser_type::HTTP_RESPONSE); +} + +template<class Derived> +std::size_t +nodejs_basic_parser<Derived>:: +write(void const* data, + std::size_t size, error_code& ec) +{ + ec_ = &ec; + auto const n = http_parser_execute( + &state_, hooks(), + static_cast<const char*>(data), size); + if(! ec) + ec = detail::make_nodejs_error( + static_cast<int>(state_.http_errno)); + if(ec) + return 0; + return n; +} + +template<class Derived> +void +nodejs_basic_parser<Derived>:: +write_eof(error_code& ec) +{ + ec_ = &ec; + http_parser_execute(&state_, hooks(), nullptr, 0); + if(! ec) + ec = detail::make_nodejs_error( + static_cast<int>(state_.http_errno)); +} + +template<class Derived> +void +nodejs_basic_parser<Derived>:: +check_header() +{ + if(! value_.empty()) + { + impl().on_field(field_, value_); + field_.clear(); + value_.clear(); + } +} + +template<class Derived> +int +nodejs_basic_parser<Derived>:: +cb_message_start(http_parser* p) +{ + auto& t = *reinterpret_cast<nodejs_basic_parser*>(p->data); + t.complete_ = false; + t.url_.clear(); + t.status_.clear(); + t.field_.clear(); + t.value_.clear(); + t.impl().on_start(); + return 0; +} + +template<class Derived> +int +nodejs_basic_parser<Derived>:: +cb_url(http_parser* p, + char const* in, std::size_t bytes) +{ + auto& t = *reinterpret_cast<nodejs_basic_parser*>(p->data); + t.url_.append(in, bytes); + return 0; +} + +template<class Derived> +int +nodejs_basic_parser<Derived>:: +cb_status(http_parser* p, + char const* in, std::size_t bytes) +{ + auto& t = *reinterpret_cast<nodejs_basic_parser*>(p->data); + t.status_.append(in, bytes); + return 0; +} + +template<class Derived> +int +nodejs_basic_parser<Derived>:: +cb_header_field(http_parser* p, + char const* in, std::size_t bytes) +{ + auto& t = *reinterpret_cast<nodejs_basic_parser*>(p->data); + t.check_header(); + t.field_.append(in, bytes); + return 0; +} + +template<class Derived> +int +nodejs_basic_parser<Derived>:: +cb_header_value(http_parser* p, + char const* in, std::size_t bytes) +{ + auto& t = *reinterpret_cast<nodejs_basic_parser*>(p->data); + t.value_.append(in, bytes); + return 0; +} + +template<class Derived> +int +nodejs_basic_parser<Derived>:: +cb_headers_complete(http_parser* p) +{ + auto& t = *reinterpret_cast<nodejs_basic_parser*>(p->data); + t.check_header(); + t.impl().on_headers_complete(*t.ec_); + if(*t.ec_) + return 1; + bool const keep_alive = + http_should_keep_alive(p) != 0; + if(p->type == http_parser_type::HTTP_REQUEST) + { + t.impl().on_request(p->method, t.url_, + p->http_major, p->http_minor, keep_alive, + p->upgrade); + return 0; + } + return t.impl().on_response(p->status_code, t.status_, + p->http_major, p->http_minor, keep_alive, + p->upgrade) ? 0 : 1; +} + +template<class Derived> +int +nodejs_basic_parser<Derived>:: +cb_body(http_parser* p, + char const* in, std::size_t bytes) +{ + auto& t = *reinterpret_cast<nodejs_basic_parser*>(p->data); + t.impl().on_body(in, bytes, *t.ec_); + return *t.ec_ ? 1 : 0; +} + +template<class Derived> +int +nodejs_basic_parser<Derived>:: +cb_message_complete(http_parser* p) +{ + auto& t = *reinterpret_cast<nodejs_basic_parser*>(p->data); + t.complete_ = true; + t.impl().on_complete(); + return 0; +} + +template<class Derived> +int +nodejs_basic_parser<Derived>:: +cb_chunk_header(http_parser*) +{ + return 0; +} + +template<class Derived> +int +nodejs_basic_parser<Derived>:: +cb_chunk_complete(http_parser*) +{ + return 0; +} + +//------------------------------------------------------------------------------ + +/** A HTTP parser. + + The parser may only be used once. +*/ +template<bool isRequest, class Body, class Fields> +class nodejs_parser + : public nodejs_basic_parser<nodejs_parser<isRequest, Body, Fields>> +{ + bool started_ = false; + +public: + nodejs_parser(nodejs_parser&&) = default; + + nodejs_parser() + : http::nodejs_basic_parser<nodejs_parser>(isRequest) + { + } + + /// Returns `true` if at least one byte has been processed + bool + started() + { + return started_; + } + +private: + friend class http::nodejs_basic_parser<nodejs_parser>; + + void + on_start() + { + started_ = true; + } + + void + on_field(std::string const&, std::string const&) + { + } + + void + on_headers_complete(error_code& ec) + { + // vFALCO TODO Decode the Content-Length and + // Transfer-Encoding, see if we can reserve the buffer. + // + // r_.reserve(content_length) + ec = {}; + } + + bool + on_request(unsigned, std::string const&, + int, int, bool, bool, std::true_type) + { + return true; + } + + bool + on_request(unsigned, std::string const&, + int, int, bool, bool, std::false_type) + { + return true; + } + + bool + on_request(unsigned method, std::string const& url, + int major, int minor, bool keep_alive, bool upgrade) + { + return on_request(method, url, + major, minor, keep_alive, upgrade, + std::integral_constant< + bool, isRequest>{}); + } + + bool + on_response(int, std::string const&, + int, int, bool, bool, std::true_type) + { + return true; + } + + bool + on_response(int, std::string const&, int, int, bool, bool, + std::false_type) + { + return true; + } + + bool + on_response(int status, std::string const& reason, + int major, int minor, bool keep_alive, bool upgrade) + { + return on_response( + status, reason, major, minor, keep_alive, upgrade, + std::integral_constant<bool, ! isRequest>{}); + } + + void + on_body(void const*, std::size_t, error_code& ec) + { + ec = {}; + } + + void + on_complete() + { + } +}; + +} // http +} // beast +} // boost + +#endif diff --git a/src/boost/libs/beast/test/bench/utf8_checker/CMakeLists.txt b/src/boost/libs/beast/test/bench/utf8_checker/CMakeLists.txt new file mode 100644 index 00000000..6164127a --- /dev/null +++ b/src/boost/libs/beast/test/bench/utf8_checker/CMakeLists.txt @@ -0,0 +1,25 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +GroupSources (include/boost/beast beast) +GroupSources (test/bench/utf8_checker "/") + +add_executable (bench-utf8-checker + ${BOOST_BEAST_FILES} + Jamfile + bench_utf8_checker.cpp +) + +target_link_libraries(bench-utf8-checker + lib-asio + lib-beast + lib-test + ) + +set_property(TARGET bench-utf8-checker PROPERTY FOLDER "tests-bench") diff --git a/src/boost/libs/beast/test/bench/utf8_checker/Jamfile b/src/boost/libs/beast/test/bench/utf8_checker/Jamfile new file mode 100644 index 00000000..1d352f38 --- /dev/null +++ b/src/boost/libs/beast/test/bench/utf8_checker/Jamfile @@ -0,0 +1,19 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +exe bench-utf8-checker : bench_utf8_checker.cpp + : requirements + <library>/boost/beast/test//lib-test + ; + +explicit bench-utf8-checker ; + +alias run-tests : + [ compile bench_utf8_checker.cpp ] + ; diff --git a/src/boost/libs/beast/test/bench/utf8_checker/bench_utf8_checker.cpp b/src/boost/libs/beast/test/bench/utf8_checker/bench_utf8_checker.cpp new file mode 100644 index 00000000..e7f81004 --- /dev/null +++ b/src/boost/libs/beast/test/bench/utf8_checker/bench_utf8_checker.cpp @@ -0,0 +1,157 @@ +// +// Copyright (c) 2016-2019 Vinnie Falco (vinnie dot falco at gmail dot com) +// +// Distributed under the Boost Software License, Version 1.0. (See accompanying +// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +// +// Official repository: https://github.com/boostorg/beast +// + +#include <boost/beast/websocket/detail/utf8_checker.hpp> +#include <boost/beast/_experimental/unit_test/suite.hpp> +#include <chrono> +#include <random> + +#ifndef BEAST_USE_BOOST_LOCALE_BENCHMARK +#define BEAST_USE_BOOST_LOCALE_BENCHMARK 0 +#endif + +#if BEAST_USE_BOOST_LOCALE_BENCHMARK +#include <boost/locale.hpp> +#endif + +namespace boost { +namespace beast { + +class utf8_checker_test : public beast::unit_test::suite +{ + std::mt19937 rng_; + +public: + using size_type = std::uint64_t; + + class timer + { + public: + using clock_type = + std::chrono::system_clock; + + private: + clock_type::time_point when_; + + public: + using duration = + clock_type::duration; + + timer() + : when_(clock_type::now()) + { + } + + duration + elapsed() const + { + return clock_type::now() - when_; + } + }; + + template<class UInt = std::size_t> + UInt + rand(std::size_t n) + { + return static_cast<UInt>( + std::uniform_int_distribution< + std::size_t>{0, n-1}(rng_)); + } + + static + inline + size_type + throughput(std::chrono::duration< + double> const& elapsed, size_type items) + { + using namespace std::chrono; + return static_cast<size_type>( + 1 / (elapsed/items).count()); + } + + std::string + corpus(std::size_t n) + { + std::string s; + s.reserve(n); + while(n--) + s.push_back(static_cast<char>( + ' ' + rand(95))); + return s; + } + + void + checkBeast(std::string const& s) + { + beast::websocket::detail::check_utf8( + s.data(), s.size()); + } + +#if BEAST_USE_BOOST_LOCALE_BENCHMARK + void + checkLocale(std::string const& s) + { + using namespace boost::locale; + auto p = s.begin(); + auto const e = s.end(); + while(p != e) + { + auto cp = utf::utf_traits<char>::decode(p, e); + if(cp == utf::illegal) + break; + } + } +#endif + + template<class F> + typename timer::clock_type::duration + test(F const& f) + { + timer t; + f(); + return t.elapsed(); + } + + void + run() override + { + auto const s = corpus(32 * 1024 * 1024); + for(int i = 0; i < 5; ++ i) + { + auto const elapsed = test([&]{ + checkBeast(s); + checkBeast(s); + checkBeast(s); + checkBeast(s); + checkBeast(s); + }); + log << "beast: " << throughput(elapsed, s.size()) << " char/s" << std::endl; + } + #if BEAST_USE_BOOST_LOCALE_BENCHMARK + for(int i = 0; i < 5; ++ i) + { + auto const elapsed = test([&]{ + checkLocale(s); + checkLocale(s); + checkLocale(s); + checkLocale(s); + checkLocale(s); + }); + log << "locale: " << throughput(elapsed, s.size()) << " char/s" << std::endl; + } + #endif + pass(); + } +}; + +BEAST_DEFINE_TESTSUITE(beast,benchmarks,utf8_checker); + +} // beast +} // boost + diff --git a/src/boost/libs/beast/test/bench/wsload/CMakeLists.txt b/src/boost/libs/beast/test/bench/wsload/CMakeLists.txt new file mode 100644 index 00000000..7a6c5876 --- /dev/null +++ b/src/boost/libs/beast/test/bench/wsload/CMakeLists.txt @@ -0,0 +1,24 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +GroupSources (include/boost/beast beast) +GroupSources (test/bench/wsload "/") + +add_executable (bench-wsload + ${BOOST_BEAST_FILES} + Jamfile + wsload.cpp + ) + +target_link_libraries(bench-wsload + lib-asio + lib-beast + ) + +set_property(TARGET bench-wsload PROPERTY FOLDER "tests-bench") diff --git a/src/boost/libs/beast/test/bench/wsload/Jamfile b/src/boost/libs/beast/test/bench/wsload/Jamfile new file mode 100644 index 00000000..82706e15 --- /dev/null +++ b/src/boost/libs/beast/test/bench/wsload/Jamfile @@ -0,0 +1,18 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +exe wsload : + wsload.cpp + ; + +explicit wsload ; + +alias run-tests : + [ compile wsload.cpp : : wsload-compile ] + ; diff --git a/src/boost/libs/beast/test/bench/wsload/wsload.cpp b/src/boost/libs/beast/test/bench/wsload/wsload.cpp new file mode 100644 index 00000000..f697d9b8 --- /dev/null +++ b/src/boost/libs/beast/test/bench/wsload/wsload.cpp @@ -0,0 +1,338 @@ +// +// Copyright (c) 2016-2019 Vinnie Falco (vinnie dot falco at gmail dot com) +// +// Distributed under the Boost Software License, Version 1.0. (See accompanying +// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +// +// Official repository: https://github.com/boostorg/beast +// + +//------------------------------------------------------------------------------ +// +// wsload +// +// Measure the performance of a WebSocket server +// +//------------------------------------------------------------------------------ + +#include <boost/beast/core.hpp> +#include <boost/beast/websocket.hpp> +#include <boost/beast/_experimental/unit_test/dstream.hpp> +#include <boost/asio.hpp> +#include <atomic> +#include <chrono> +#include <cstdlib> +#include <functional> +#include <iostream> +#include <memory> +#include <mutex> +#include <random> +#include <thread> +#include <vector> + +namespace beast = boost::beast; // from <boost/beast.hpp> +namespace http = beast::http; // from <boost/beast/http.hpp> +namespace websocket = beast::websocket; // from <boost/beast/websocket.hpp> +namespace net = boost::asio; // from <boost/asio.hpp> +using tcp = boost::asio::ip::tcp; // from <boost/asio/ip/tcp.hpp> + +class test_buffer +{ + char data_[4096]; + net::const_buffer b_; + +public: + using const_iterator = + net::const_buffer const*; + + using value_type = net::const_buffer; + + test_buffer() + : b_(data_, sizeof(data_)) + { + std::mt19937_64 rng; + std::uniform_int_distribution<unsigned short> dist; + for(auto& c : data_) + c = static_cast<unsigned char>(dist(rng)); + } + + const_iterator + begin() const + { + return &b_; + } + + const_iterator + end() const + { + return begin() + 1; + } +}; + +class report +{ + std::mutex m_; + std::size_t bytes_ = 0; + std::size_t messages_ = 0; + +public: + void + insert(std::size_t messages, std::size_t bytes) + { + std::lock_guard<std::mutex> lock(m_); + bytes_ += bytes; + messages_ += messages; + } + + std::size_t + bytes() const + { + return bytes_; + } + + std::size_t + messages() const + { + return messages_; + } +}; + +void +fail(beast::error_code ec, char const* what) +{ + std::cerr << what << ": " << ec.message() << "\n"; +} + +class connection + : public std::enable_shared_from_this<connection> +{ + websocket::stream<tcp::socket> ws_; + tcp::endpoint ep_; + std::size_t messages_; + report& rep_; + test_buffer const& tb_; + net::strand< + net::io_context::executor_type> strand_; + beast::flat_buffer buffer_; + std::mt19937_64 rng_; + std::size_t count_ = 0; + std::size_t bytes_ = 0; + +public: + connection( + net::io_context& ioc, + tcp::endpoint const& ep, + std::size_t messages, + bool deflate, + report& rep, + test_buffer const& tb) + : ws_(ioc) + , ep_(ep) + , messages_(messages) + , rep_(rep) + , tb_(tb) + , strand_(ioc.get_executor()) + { + websocket::permessage_deflate pmd; + pmd.client_enable = deflate; + ws_.set_option(pmd); + ws_.binary(true); + ws_.auto_fragment(false); + ws_.write_buffer_bytes(64 * 1024); + } + + ~connection() + { + rep_.insert(count_, bytes_); + } + + void + run() + { + ws_.next_layer().async_connect(ep_, + beast::bind_front_handler( + &connection::on_connect, + this->shared_from_this())); + } + +private: + void + on_connect(beast::error_code ec) + { + if(ec) + return fail(ec, "on_connect"); + + ws_.async_handshake( + ep_.address().to_string() + ":" + std::to_string(ep_.port()), + "/", + beast::bind_front_handler( + &connection::on_handshake, + this->shared_from_this())); + } + + void + on_handshake(beast::error_code ec) + { + if(ec) + return fail(ec, "handshake"); + + do_write(); + } + + void + do_write() + { + std::geometric_distribution<std::size_t> dist{ + double(4) / beast::buffer_bytes(tb_)}; + ws_.async_write_some(true, + beast::buffers_prefix(dist(rng_), tb_), + beast::bind_front_handler( + &connection::on_write, + this->shared_from_this())); + } + + void + on_write(beast::error_code ec, std::size_t) + { + if(ec) + return fail(ec, "write"); + + if(messages_--) + return do_read(); + + ws_.async_close({}, + beast::bind_front_handler( + &connection::on_close, + this->shared_from_this())); + } + + void + do_read() + { + ws_.async_read(buffer_, + beast::bind_front_handler( + &connection::on_read, + this->shared_from_this())); + } + + void + on_read(beast::error_code ec, std::size_t) + { + if(ec) + return fail(ec, "read"); + + ++count_; + bytes_ += buffer_.size(); + buffer_.consume(buffer_.size()); + do_write(); + } + + void + on_close(beast::error_code ec) + { + if(ec) + return fail(ec, "close"); + } +}; + +class timer +{ + using clock_type = + std::chrono::system_clock; + + clock_type::time_point when_; + +public: + using duration = + clock_type::duration; + + timer() + : when_(clock_type::now()) + { + } + + duration + elapsed() const + { + return clock_type::now() - when_; + } +}; + +inline +std::uint64_t +throughput( + std::chrono::duration<double> const& elapsed, + std::uint64_t items) +{ + using namespace std::chrono; + return static_cast<std::uint64_t>( + 1 / (elapsed/items).count()); +} + +int +main(int argc, char** argv) +{ + beast::unit_test::dstream dout(std::cerr); + + try + { + // Check command line arguments. + if(argc != 8) + { + std::cerr << + "Usage: bench-wsload <address> <port> <trials> <messages> <workers> <threads> <compression:0|1>"; + return EXIT_FAILURE; + } + + auto const address = net::ip::make_address(argv[1]); + auto const port = static_cast<unsigned short>(std::atoi(argv[2])); + auto const trials = static_cast<std::size_t>(std::atoi(argv[3])); + auto const messages= static_cast<std::size_t>(std::atoi(argv[4])); + auto const workers = static_cast<std::size_t>(std::atoi(argv[5])); + auto const threads = static_cast<std::size_t>(std::atoi(argv[6])); + auto const deflate = std::atoi(argv[7]) != 0; + auto const work = (messages + workers - 1) / workers; + test_buffer tb; + for(auto i = trials; i != 0; --i) + { + report rep; + net::io_context ioc{1}; + for(auto j = workers; j; --j) + { + auto sp = + std::make_shared<connection>( + ioc, + tcp::endpoint{address, port}, + work, + deflate, + rep, + tb); + sp->run(); + } + timer clock; + std::vector<std::thread> tv; + if(threads > 1) + { + tv.reserve(threads); + tv.emplace_back([&ioc]{ ioc.run(); }); + } + ioc.run(); + for(auto& t : tv) + t.join(); + auto const elapsed = clock.elapsed(); + dout << + throughput(elapsed, rep.bytes()) << " bytes/s in " << + (std::chrono::duration_cast< + std::chrono::milliseconds>( + elapsed).count() / 1000.) << "ms and " << + rep.bytes() << " bytes" << std::endl; + } + } + catch(std::exception const& e) + { + std::cerr << "Error: " << e.what() << std::endl; + return EXIT_FAILURE; + } + + return EXIT_SUCCESS; +} diff --git a/src/boost/libs/beast/test/bench/zlib/CMakeLists.txt b/src/boost/libs/beast/test/bench/zlib/CMakeLists.txt new file mode 100644 index 00000000..a030ed93 --- /dev/null +++ b/src/boost/libs/beast/test/bench/zlib/CMakeLists.txt @@ -0,0 +1,31 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +GroupSources (include/boost/beast beast) +GroupSources (test/extern/zlib-1.2.11 zlib) +GroupSources (test/bench/zlib "/") + +if (MSVC) + set_source_files_properties (${ZLIB_SOURCES} PROPERTIES COMPILE_FLAGS "/wd4127 /wd4131 /wd4244") +endif() + +add_executable (bench-zlib + ${BOOST_BEAST_FILES} + ${ZLIB_SOURCES} + Jamfile + deflate_stream.cpp + inflate_stream.cpp +) +target_link_libraries(bench-zlib + lib-asio + lib-beast + lib-test + ) + +set_property(TARGET bench-zlib PROPERTY FOLDER "tests-bench") diff --git a/src/boost/libs/beast/test/bench/zlib/Jamfile b/src/boost/libs/beast/test/bench/zlib/Jamfile new file mode 100644 index 00000000..6e210c29 --- /dev/null +++ b/src/boost/libs/beast/test/bench/zlib/Jamfile @@ -0,0 +1,15 @@ +# +# Copyright (c) 2016-2017 Vinnie Falco (vinnie dot falco at gmail dot com) +# +# Distributed under the Boost Software License, Version 1.0. (See accompanying +# file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +# +# Official repository: https://github.com/boostorg/beast +# + +exe bench-zlib : + $(ZLIB_SOURCES) + deflate_stream.cpp + inflate_stream.cpp + /boost/beast/test//lib-test + ; diff --git a/src/boost/libs/beast/test/bench/zlib/deflate_stream.cpp b/src/boost/libs/beast/test/bench/zlib/deflate_stream.cpp new file mode 100644 index 00000000..321fa98c --- /dev/null +++ b/src/boost/libs/beast/test/bench/zlib/deflate_stream.cpp @@ -0,0 +1,199 @@ +// +// Copyright (c) 2016-2019 Vinnie Falco (vinnie dot falco at gmail dot com) +// +// Distributed under the Boost Software License, Version 1.0. (See accompanying +// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +// +// Official repository: https://github.com/boostorg/beast +// + +#include <boost/beast/core/string.hpp> +#include <boost/beast/zlib/deflate_stream.hpp> +#include <boost/beast/test/throughput.hpp> +#include <boost/beast/_experimental/unit_test/dstream.hpp> +#include <boost/beast/_experimental/unit_test/suite.hpp> +#include <iomanip> +#include <random> +#include <string> + +#include "zlib-1.2.11/zlib.h" + +namespace boost { +namespace beast { +namespace zlib { + +class deflate_stream_test : public beast::unit_test::suite +{ +public: + // Lots of repeats, limited char range + static + std::string + corpus1(std::size_t n) + { + static std::string const alphabet{ + "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz" + }; + std::string s; + s.reserve(n + 5); + std::mt19937 g; + std::uniform_int_distribution<std::size_t> d0{ + 0, alphabet.size() - 1}; + std::uniform_int_distribution<std::size_t> d1{ + 1, 5}; + while(s.size() < n) + { + auto const rep = d1(g); + auto const ch = alphabet[d0(g)]; + s.insert(s.end(), rep, ch); + } + s.resize(n); + return s; + } + + // Random data + static + std::string + corpus2(std::size_t n) + { + std::string s; + s.reserve(n); + std::mt19937 g; + std::uniform_int_distribution<std::uint32_t> d0{0, 255}; + while(n--) + s.push_back(static_cast<char>(d0(g))); + return s; + } + + std::string + doDeflateBeast(string_view const& in) + { + z_params zs; + deflate_stream ds; + ds.reset( + Z_DEFAULT_COMPRESSION, + 15, + 4, + Strategy::normal); + std::string out; + out.resize(deflate_upper_bound(in.size())); + zs.next_in = in.data(); + zs.avail_in = in.size(); + zs.next_out = &out[0]; + zs.avail_out = out.size(); + error_code ec; + ds.write(zs, Flush::full, ec); + BEAST_EXPECTS(! ec, ec.message()); + out.resize(zs.total_out); + return out; + } + + std::string + doDeflateZLib(string_view const& in) + { + int result; + z_stream zs; + memset(&zs, 0, sizeof(zs)); + result = deflateInit2( + &zs, + Z_DEFAULT_COMPRESSION, + Z_DEFLATED, + -15, + 4, + Z_DEFAULT_STRATEGY + ); + if(result != Z_OK) + throw std::logic_error("deflateInit2 failed"); + std::string out; + out.resize(deflateBound(&zs, + static_cast<uLong>(in.size()))); + zs.next_in = (Bytef*)in.data(); + zs.avail_in = static_cast<uInt>(in.size()); + zs.next_out = (Bytef*)&out[0]; + zs.avail_out = static_cast<uInt>(out.size()); + result = deflate(&zs, Z_FULL_FLUSH); + if(result != Z_OK) + throw std::logic_error("deflate failed"); + out.resize(zs.total_out); + deflateEnd(&zs); + return out; + } + + void + doCorpus( + std::size_t size, + std::size_t repeat) + { + std::size_t constexpr trials = 3; + auto const c1 = corpus1(size); + auto const c2 = corpus2(size); + log << + std::left << std::setw(10) << (std::to_string(size) + "B") << + std::right << std::setw(12) << "Beast" << " " << + std::right << std::setw(12) << "ZLib" << + std::endl; + for(std::size_t i = 0; i < trials; ++i) + { + test::timer t; + log << std::left << std::setw(10) << "corpus1"; + std::string out1; + for(std::size_t j = 0; j < repeat; ++j) + out1 = doDeflateBeast(c1); + auto const t1 = + test::throughput(t.elapsed(), size * repeat); + log << std::right << std::setw(12) << t1 << " B/s "; + std::string out2; + for(std::size_t j = 0; j < repeat; ++j) + out2 = doDeflateZLib(c1); + BEAST_EXPECT(out1 == out2); + auto const t2 = + test::throughput(t.elapsed(), size * repeat); + log << std::right << std::setw(12) << t2 << " B/s"; + log << std::right << std::setw(12) << + unsigned(double(t1)*100/t2-100) << "%"; + log << std::endl; + } + for(std::size_t i = 0; i < trials; ++i) + { + test::timer t; + log << std::left << std::setw(10) << "corpus2"; + std::string out1; + for(std::size_t j = 0; j < repeat; ++j) + out1 = doDeflateBeast(c2); + auto const t1 = + test::throughput(t.elapsed(), size * repeat); + log << std::right << std::setw(12) << t1 << " B/s "; + std::string out2; + for(std::size_t j = 0; j < repeat; ++j) + out2 = doDeflateZLib(c2); + BEAST_EXPECT(out1 == out2); + auto const t2 = + test::throughput(t.elapsed(), size * repeat); + log << std::right << std::setw(12) << t2 << " B/s"; + log << std::right << std::setw(12) << + unsigned(double(t1)*100/t2-100) << "%"; + log << std::endl; + } + log << std::endl; + } + + void + doBench() + { + doCorpus( 16 * 1024, 512); + doCorpus( 1024 * 1024, 8); + doCorpus(8 * 1024 * 1024, 1); + } + + void + run() override + { + doBench(); + pass(); + } +}; + +BEAST_DEFINE_TESTSUITE(beast,zlib,deflate_stream); + +} // zlib +} // beast +} // boost diff --git a/src/boost/libs/beast/test/bench/zlib/inflate_stream.cpp b/src/boost/libs/beast/test/bench/zlib/inflate_stream.cpp new file mode 100644 index 00000000..ea8d6d98 --- /dev/null +++ b/src/boost/libs/beast/test/bench/zlib/inflate_stream.cpp @@ -0,0 +1,244 @@ +// +// Copyright (c) 2016-2019 Vinnie Falco (vinnie dot falco at gmail dot com) +// +// Distributed under the Boost Software License, Version 1.0. (See accompanying +// file LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) +// +// Official repository: https://github.com/boostorg/beast +// + +#include <boost/beast/core/string.hpp> +#include <boost/beast/zlib/inflate_stream.hpp> +#include <boost/beast/test/throughput.hpp> +#include <boost/beast/_experimental/unit_test/dstream.hpp> +#include <boost/beast/_experimental/unit_test/suite.hpp> +#include <iomanip> +#include <random> +#include <string> + +#include "zlib-1.2.11/zlib.h" + +namespace boost { +namespace beast { +namespace zlib { + +class inflate_stream_test : public beast::unit_test::suite +{ +public: + // Lots of repeats, limited char range + static + std::string + corpus1(std::size_t n) + { + static std::string const alphabet{ + "0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz" + }; + std::string s; + s.reserve(n + 5); + std::mt19937 g; + std::uniform_int_distribution<std::size_t> d0{ + 0, alphabet.size() - 1}; + std::uniform_int_distribution<std::size_t> d1{ + 1, 5}; + while(s.size() < n) + { + auto const rep = d1(g); + auto const ch = alphabet[d0(g)]; + s.insert(s.end(), rep, ch); + } + s.resize(n); + return s; + } + + // Random data + static + std::string + corpus2(std::size_t n) + { + std::string s; + s.reserve(n); + std::mt19937 g; + std::uniform_int_distribution<std::uint32_t> d0{0, 255}; + while(n--) + s.push_back(static_cast<char>(d0(g))); + return s; + } + + static + std::string + compress(string_view const& in) + { + int result; + z_stream zs; + memset(&zs, 0, sizeof(zs)); + result = deflateInit2( + &zs, + Z_DEFAULT_COMPRESSION, + Z_DEFLATED, + -15, + 4, + Z_DEFAULT_STRATEGY); + if(result != Z_OK) + throw std::logic_error("deflateInit2 failed"); + zs.next_in = (Bytef*)in.data(); + zs.avail_in = static_cast<uInt>(in.size()); + std::string out; + out.resize(deflateBound(&zs, + static_cast<uLong>(in.size()))); + zs.next_in = (Bytef*)in.data(); + zs.avail_in = static_cast<uInt>(in.size()); + zs.next_out = (Bytef*)&out[0]; + zs.avail_out = static_cast<uInt>(out.size()); + result = deflate(&zs, Z_FULL_FLUSH); + if(result != Z_OK) + throw std::logic_error("deflate failed"); + out.resize(zs.total_out); + deflateEnd(&zs); + return out; + } + + std::string + doInflateBeast(string_view const& in) + { + z_params zs; + std::string out; + inflate_stream is; + zs.next_in = &in[0]; + zs.avail_in = in.size(); + out.resize(in.size()); + zs.next_out = &out[0]; + zs.avail_out = out.size(); + for(;;) + { + error_code ec; + is.write(zs, Flush::sync, ec); + if(ec) + throw std::logic_error("inflate_stream failed"); + if(zs.avail_out > 0) + break; + out.resize(2 * zs.total_out); + zs.next_out = &out[zs.total_out]; + zs.avail_out = out.size() - zs.total_out; + } + out.resize(zs.total_out); + return out; + } + + std::string + doInflateZLib(string_view const& in) + { + int result; + z_stream zs; + std::string out; + memset(&zs, 0, sizeof(zs)); + result = inflateInit2(&zs, -15); + zs.next_in = (Bytef*)in.data(); + zs.avail_in = static_cast<uInt>(in.size()); + out.resize(in.size()); + zs.next_out = (Bytef*)&out[0]; + zs.avail_out = static_cast<uInt>(out.size()); + for(;;) + { + result = inflate(&zs, Z_SYNC_FLUSH); + if( result == Z_NEED_DICT || + result == Z_DATA_ERROR || + result == Z_MEM_ERROR) + { + throw std::logic_error("inflate failed"); + } + if(zs.avail_out > 0) + break; + if(result == Z_STREAM_END) + break; + out.resize(2 * zs.total_out); + zs.next_out = (Bytef*)&out[zs.total_out]; + zs.avail_out = static_cast<uInt>( + out.size() - zs.total_out); + } + out.resize(zs.total_out); + inflateEnd(&zs); + return out; + } + + void + doCorpus( + std::size_t size, + std::size_t repeat) + { + std::size_t constexpr trials = 3; + std::uint64_t constexpr scale = 16; + auto const c1 = corpus1(size); + auto const c2 = corpus2(size * scale); + auto const in1 = compress(c1); + auto const in2 = compress(c2); + log << + std::left << std::setw(10) << (std::to_string(size) + "B") << + std::right << std::setw(12) << "Beast" << " " << + std::right << std::setw(12) << "ZLib" << + std::endl; + for(std::size_t i = 0; i < trials; ++i) + { + test::timer t; + log << std::left << std::setw(10) << "corpus1"; + std::string out; + for(std::size_t j = 0; j < repeat; ++j) + out = doInflateBeast(in1); + BEAST_EXPECT(out == c1); + auto const t1 = + test::throughput(t.elapsed(), size * repeat); + log << std::right << std::setw(12) << t1 << " B/s "; + for(std::size_t j = 0; j < repeat; ++j) + out = doInflateZLib(in1); + BEAST_EXPECT(out == c1); + auto const t2 = + test::throughput(t.elapsed(), size * repeat); + log << std::right << std::setw(12) << t2 << " B/s"; + log << std::right << std::setw(12) << + unsigned(double(t1)*100/t2-100) << "%"; + log << std::endl; + } + for(std::size_t i = 0; i < trials; ++i) + { + test::timer t; + log << std::left << std::setw(10) << "corpus2"; + std::string out; + for(std::size_t j = 0; j < repeat; ++j) + out = doInflateBeast(in2); + BEAST_EXPECT(out == c2); + auto const t1 = + test::throughput(t.elapsed(), size * scale * repeat); + log << std::right << std::setw(12) << t1 << " B/s "; + for(std::size_t j = 0; j < repeat; ++j) + out = doInflateZLib(in2); + BEAST_EXPECT(out == c2); + auto const t2 = + test::throughput(t.elapsed(), size * scale * repeat); + log << std::right << std::setw(12) << t2 << " B/s"; + log << std::right << std::setw(12) << + unsigned(double(t1)*100/t2-100) << "%"; + log << std::endl; + } + log << std::endl; + } + + void + doBench() + { + doCorpus( 1 * 1024 * 1024, 64); + doCorpus( 4 * 1024 * 1024, 16); + doCorpus( 16 * 1024 * 1024, 8); + } + + void + run() override + { + doBench(); + pass(); + } +}; + +BEAST_DEFINE_TESTSUITE(beast,zlib,inflate_stream); + +} // zlib +} // beast +} // boost |