aorefkov
/
simstr
mirror of https://github.com/orefkov/simstr.git
0
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

- Updated to 1.6.6. 

- Improved UTF conversion.
- Fixed a number of bugs.
- Improved Emscripten benchmark compilation.
---
- Обновление до 1.6.6.
- Доработал UTF-конвертацию.
- Исправил ряд ошибок.
- Доработал Emscripten сборку бенчмарков.
This commit is contained in:
Aleksandr Orefkov 2026-02-10 23:35:02 +03:00
parent 2148718b32
commit 8cf026f2c6
20 changed files with 5472 additions and 5371 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

2
CMakeLists.txt
View File

@ -5,7 +5,7 @@ include(FetchContent)
project( project(
simstr simstr
VERSION 1.6.5 VERSION 1.6.6
DESCRIPTION "Yet another modern C++ string library" DESCRIPTION "Yet another modern C++ string library"
HOMEPAGE_URL "https://github.com/orefkov/simstr" HOMEPAGE_URL "https://github.com/orefkov/simstr"
LANGUAGES CXX LANGUAGES CXX

31
bench/bench_str.cpp
View File

@ -1,5 +1,5 @@
/* /*
* ver. 1.6.5 * ver. 1.6.6
* (c) Проект "SimStr", Александр Орефков orefkov@gmail.com * (c) Проект "SimStr", Александр Орефков orefkov@gmail.com
* Бенчмарки * Бенчмарки
* (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com * (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com
@ -14,15 +14,14 @@
#include <fmt/format.h> #include <fmt/format.h>
#include <fmt/compile.h> #include <fmt/compile.h>
using namespace simstr;
using namespace std::literals;
#ifdef EMSCRIPTEN #ifdef EMSCRIPTEN
#include <emscripten.h> #include <emscripten.h>
static void DoTeardown(const benchmark::State& state) { static void DoTeardown(const benchmark::State& state) {
emscripten_sleep(1); // Это нужно чтобы интерфейс браузера обновился
} // This is necessary for the browser interface to refresh.
static void DoSetup(const benchmark::State& state) {
EM_ASM({
on_begin_benchmark(UTF8ToString($0));
}, state.name().c_str());
emscripten_sleep(1); emscripten_sleep(1);
} }
#undef BENCHMARK #undef BENCHMARK
@ -31,11 +30,18 @@ static void DoSetup(const benchmark::State& state) {
(::benchmark::internal::RegisterBenchmarkInternal( \ (::benchmark::internal::RegisterBenchmarkInternal( \
::benchmark::internal::make_unique< \ ::benchmark::internal::make_unique< \
::benchmark::internal::FunctionBenchmark>(#__VA_ARGS__, \ ::benchmark::internal::FunctionBenchmark>(#__VA_ARGS__, \
__VA_ARGS__)))->Setup(DoSetup)->Teardown(DoTeardown) __VA_ARGS__)))->Teardown(DoTeardown)
#endif
using namespace simstr; #undef BENCHMARK_CAPTURE
using namespace std::literals; #define BENCHMARK_CAPTURE(func, test_case_name, ...) \
BENCHMARK_PRIVATE_DECLARE(_benchmark_) = \
(::benchmark::internal::RegisterBenchmarkInternal( \
::benchmark::internal::make_unique< \
::benchmark::internal::FunctionBenchmark>( \
#func "/" #test_case_name, \
[](::benchmark::State& st) { func(st, __VA_ARGS__); })))->Teardown(DoTeardown)
#endif
#define TEST_TEXT "Test text" #define TEST_TEXT "Test text"
#define LONG_TEXT "123456789012345678901234567890" #define LONG_TEXT "123456789012345678901234567890"
@ -769,7 +775,6 @@ void ToIntNoOverflow(benchmark::State& state, ssa t, int c) {
benchmark::DoNotOptimize(t); benchmark::DoNotOptimize(t);
} }
} }
BENCHMARK(__)->Name("----- Convert to int '1234567' ---------")->Repetitions(1); BENCHMARK(__)->Name("----- Convert to int '1234567' ---------")->Repetitions(1);
BENCHMARK_CAPTURE(ToIntStr10, , std::string{"123456789"}, 123456789) ->Name("std::string s = \"123456789\"; int res = std::strtol(s.c_str(), 0, 10);"); BENCHMARK_CAPTURE(ToIntStr10, , std::string{"123456789"}, 123456789) ->Name("std::string s = \"123456789\"; int res = std::strtol(s.c_str(), 0, 10);");
BENCHMARK_CAPTURE(ToIntFromChars10, , std::string_view{"123456789"}, 123456789) ->Name("std::string_view s = \"123456789\"; std::from_chars(s.data(), s.data() + s.size(), res, 10);"); BENCHMARK_CAPTURE(ToIntFromChars10, , std::string_view{"123456789"}, 123456789) ->Name("std::string_view s = \"123456789\"; std::from_chars(s.data(), s.data() + s.size(), res, 10);");
@ -2475,7 +2480,9 @@ int main(int argc, char** argv) {
::benchmark::RunSpecifiedBenchmarks(); ::benchmark::RunSpecifiedBenchmarks();
::benchmark::Shutdown(); ::benchmark::Shutdown();
#ifdef EMSCRIPTEN #ifdef EMSCRIPTEN
std::cout << "Done!\n";
EM_ASM({ on_done(); }); EM_ASM({ on_done(); });
#endif #endif
return 0; return 0;
} }

332
bench/emscripten/shell_minimal.html
View File

@ -1,96 +1,244 @@
<!doctype html> <!doctype html>
<html lang="en-us"> <html lang="en-us">
<head>
<meta charset="utf-8">
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<title>SimStr Bencjmarks</title>
<style>
body {
color: white;
background-color: black;
}
#output {
position: absolute;
left: 10px;
top: 80px;
right: 10px;
bottom: 10px;
margin: 0 auto;
border: 0;
display: block;
background-color: black;
color: white;
font-family: 'Cascadia Code', 'Consolas', 'Lucida Console', Monaco, monospace;
outline: none;
}
.ctr {
text-align: center;
font-size: 16pt;
}
</style>
</head>
<body>
<div class="ctr" id="bs_name"></div><div class="ctr" id="b_name"></div>
<textarea class="emscripten" id="output" rows="20"></textarea>
<script type='text/javascript'>
var outputElement = document.getElementById('output');
var bsNameEl= document.getElementById('bs_name');
var bNameEl= document.getElementById('b_name');
if (outputElement) outputElement.value = ''; // clear browser cache
var Module = {
print(...args) {
console.log(...args);
if (outputElement) {
var text = args.join(' ').replace('_mean', ' ');
var rpt = text.indexOf("/repeats");
if (rpt != -1)
text = text.substr(0, rpt);
if (text.indexOf('_median ') == -1 && text.indexOf('_stddev ') == -1 && text.indexOf('_cv ') == -1) {
outputElement.value += text + "\n";
outputElement.scrollTop = outputElement.scrollHeight; // focus on bottom
}
}
},
setStatus(text) {
if (!Module.setStatus.last) Module.setStatus.last = { time: Date.now(), text: '' };
if (text === Module.setStatus.last.text) return;
var m = text.match(/([^(]+)\((\d+(\.\d+)?)\/(\d+)\)/);
var now = Date.now();
if (m && now - Module.setStatus.last.time < 30) return; // if this is a progress update, skip it if too soon
Module.setStatus.last.time = now;
Module.setStatus.last.text = text;
if (m) {
text = m[1];
} else {
}
},
totalDependencies: 0,
monitorRunDependencies(left) {
this.totalDependencies = Math.max(this.totalDependencies, left);
Module.setStatus(left ? 'Preparing... (' + (this.totalDependencies-left) + '/' + this.totalDependencies + ')' : 'All downloads complete.');
},
};
Module.setStatus('Downloading...');
window.onerror = () => {
Module.setStatus('Exception thrown, see JavaScript console');
Module.setStatus = (text) => {
if (text) console.error('[post-exception status] ' + text);
};
};
function on_begin_benchmark(name) {
if (name.substr(0, 3) == "---") {
bsNameEl.innerText = name.match(/^[- ]+(.+?)[ -]+$/)[1];
bNameEl.innerText = "";
} else {
bNameEl.innerText = name;
}
}
function on_done() {
bsNameEl.innerText = "Done";
bNameEl.innerText = "";
}
</script> <head>
{{{ SCRIPT }}} <meta charset="utf-8">
</body> <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
<title>SimStr Benchmarks</title>
<style>
body {
margin: auto;
font-size: 12pt;
}
h2 {
text-align: center;
}
table {
border: 1px black;
text-align: right;
margin: auto;
}
table tr:nth-child(odd) {
background: #fff;
}
table tr:nth-child(even) {
background: #f2f2f2;
}
table th {
text-align: center;
}
.tooltip a:any-link {
text-decoration: none;
}
.tooltip a:link {
color: black;
}
.tooltip a:visited {
color: black;
}
.tooltip a:hover {
text-decoration: underline;
}
table tr:hover td {
background-color: lightcyan;
}
div.header {
max-width: 960px;
margin: 0 auto;
}
canvas {
max-width: 2000px;
margin: 0 auto;
}
.benchset {
max-width: 960px;
padding: 1px;
margin: 5px auto;
border-radius: 3px;
background-color: gainsboro;
}
.benchset h4 {
text-align: left;
padding-left: 20px;
}
.test_platforms {
margin: auto;
}
.test_platforms ul {
display: table;
}
.test_platforms li {
display: table-row;
}
.test_platforms li:before {
content: "\2022";
padding-right: 0.5em;
}
.test_platforms span.tooltip {
padding-left: 20px;
display: table-cell;
}
span.platform {
display: table-cell;
}
.info {
border-radius: 3px;
border: solid 1px lightgrey;
background-color: beige;
color: brown;
}
.code {
font-family: 'Cascadia Code', Consolas, 'Courier New', Courier, monospace;
}
.tooltip {
position: relative;
display: inline-block;
}
.tooltip .tooltiptext {
visibility: hidden;
background-color: beige;
color: black;
text-align: left;
border-radius: 6px;
padding: 15px;
position: absolute;
z-index: 1;
top: 35px;
right: 0;
white-space: pre;
border: 1px solid darkgray;
box-shadow: 4px 4px 8px 0px rgba(34, 60, 80, 0.2);
transform: translate(50%);
}
.tooltip:hover .tooltiptext {
visibility: visible;
}
h4 {
text-align: center;
}
#output {
position: absolute;
left: 10px;
top: 50px;
right: 10px;
bottom: 60%;
margin: 0 auto;
border: 0;
display: block;
background-color: black;
color: white;
font-family: 'Cascadia Code', 'Consolas', 'Lucida Console', Monaco, monospace;
outline: none;
}
#results {
position: absolute;
left: 10px;
top: 42%;
right: 10px;
bottom: 10px;
margin: 0 auto;
border: 0;
overflow: auto;
}
div.head {
text-align: center;
}
div.head h3 {
margin: 0;
}
div.head span {
margin: 0;
}
</style>
</head>
<body>
<div class="head">
<h3>SimStr 1.6.6 Benchmark</h3>
<span><a href="https://orefkov.github.io/simstr/results.html" target="blank">All results</a></span>
<span><a href="https://github.com/orefkov/simstr/blob/main/bench/bench_str.cpp" target="blank">Sources for
benchmarks</a></span>
<textarea id="output"></textarea>
</div>
<div id="results"></div>
<script type='text/javascript'>
var outputElement = document.getElementById('output');
var resultEl = document.getElementById('results');
if (outputElement) outputElement.value = ''; // clear browser cache
var Module = {
print(...args) {
console.log(...args);
var text = args.join(' ')
try {
if (text.match(/^--+[^-]/)) {
add_benchset(text.match(/^-+ +(.+?) +-+\/repeats/)[1]);
} else {
var im = text.indexOf("_mean ");
if (im != -1) {
add_benchmark(text.substr(0, im), text.match(/ ([^ ]+?) ns/)[1]);
}
}
} catch (e) {
console.error(e);
}
if (outputElement) {
outputElement.value += text + "\n";
outputElement.scrollTop = outputElement.scrollHeight;
}
}
};
function on_done() {
resultEl.scrollTop = 0;
}
function escapeHtml(unsafe) {
return unsafe
.replace(/&/g, "&amp;")
.replace(/</g, "&lt;")
.replace(/>/g, "&gt;")
.replace(/"/g, "&quot;")
.replace(/'/g, "&#039;");
}
function add_benchset(bs_name) {
resultEl.insertAdjacentHTML("beforeend", `<div class="benchset"><h4>${escapeHtml(bs_name)}</h4><table width="100%"><thead><tr><th>Benchmark name</th><th width="8%">Time(ns)</th></tr></thead><tbody></tbody></table></div>`);
resultEl.scrollTop = resultEl.scrollHeight;
}
function add_benchmark(name, result) {
resultEl.querySelector("div:last-child table tbody").insertAdjacentHTML("beforeend", `<tr><td>${escapeHtml(name)}</td><td>${result}</td></tr>`);
resultEl.scrollTop = resultEl.scrollHeight;
}
</script>
{{{ SCRIPT }}}
</body>
</html> </html>

1342
bench/results.html
View File

File diff suppressed because it is too large Load Diff

1758
bench/results/000-Xeon E5-2682 v4, Ubuntu 22 (WSL), Clang-21.txt
View File

File diff suppressed because it is too large Load Diff

1780
bench/results/001-Xeon E5-2682 v4, Ubuntu 22 (WSL), GCC-13.txt
View File

File diff suppressed because it is too large Load Diff

1776
bench/results/002-Xeon E5-2682 v4, Windows 10, Clang-19.txt
View File

File diff suppressed because it is too large Load Diff

1784
bench/results/003-Xeon E5-2682 v4, Windows 10, MSVC-19.txt
View File

File diff suppressed because it is too large Load Diff

1758
bench/results/004-Xeon E5-2682 v4, WASM Chrome 143, Clang-21.txt
View File

File diff suppressed because it is too large Load Diff

2
docs/Doxyfile
View File

@ -48,7 +48,7 @@ PROJECT_NAME = "simstr"
# could be handy for archiving the generated documentation or if some version # could be handy for archiving the generated documentation or if some version
# control system is used. # control system is used.
PROJECT_NUMBER = 1.6.5 PROJECT_NUMBER = 1.6.6
# Using the PROJECT_BRIEF tag one can provide an optional one line description # Using the PROJECT_BRIEF tag one can provide an optional one line description
# for a project that appears at the top of each page and should give viewers a # for a project that appears at the top of each page and should give viewers a

2
docs/Doxyfile_ru
View File

@ -48,7 +48,7 @@ PROJECT_NAME = "simstr"
# could be handy for archiving the generated documentation or if some version # could be handy for archiving the generated documentation or if some version
# control system is used. # control system is used.
PROJECT_NUMBER = 1.6.5 PROJECT_NUMBER = 1.6.6
# Using the PROJECT_BRIEF tag one can provide an optional one line description # Using the PROJECT_BRIEF tag one can provide an optional one line description
# for a project that appears at the top of each page and should give viewers a # for a project that appears at the top of each page and should give viewers a

2
include/simstr/simple_unicode.h
View File

@ -1,6 +1,6 @@
/* /*
* (c) Проект "SimStr", Александр Орефков orefkov@gmail.com * (c) Проект "SimStr", Александр Орефков orefkov@gmail.com
* ver. 1.6.5 * ver. 1.6.6
*/ */
#pragma once #pragma once

192
include/simstr/sstring.h
View File

@ -1,5 +1,5 @@
/* /*
* ver. 1.6.5 * ver. 1.6.6
* (c) Проект "SimStr", Александр Орефков orefkov@gmail.com * (c) Проект "SimStr", Александр Орефков orefkov@gmail.com
* Классы для работы со строками * Классы для работы со строками
* (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com * (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com
@ -386,7 +386,7 @@ struct simple_str : str_algs<K, simple_str<K>, simple_str<K>, false> {
* @en @brief Constructor from a string literal. * @en @brief Constructor from a string literal.
*/ */
template<typename T, size_t N = const_lit_for<K, T>::Count> template<typename T, size_t N = const_lit_for<K, T>::Count>
constexpr simple_str(T&& v) noexcept : str(v), len(N - 1) {} constexpr simple_str(T&& v) noexcept : str((const K*)v), len(N - 1) {}
/*! /*!
* @ru @brief Конструктор из указателя и длины. * @ru @brief Конструктор из указателя и длины.
* @en @brief Constructor from pointer and length. * @en @brief Constructor from pointer and length.
@ -597,6 +597,17 @@ using struu = simple_str_nt<u32s>;
template<typename Src, typename Dest> template<typename Src, typename Dest>
struct utf_convert_selector; struct utf_convert_selector;
template<typename Src>
struct utf_convert_selector<Src, Src> {
static size_t need_len(const Src* src, size_t srcLen) {
return srcLen;
}
static size_t convert(const Src* src, size_t srcLen, Src* dest) {
ch_traits<Src>::copy(dest, src, srcLen + 1);
return srcLen;
}
};
template<> template<>
struct utf_convert_selector<u8s, u16s> { struct utf_convert_selector<u8s, u16s> {
static SIMSTR_API size_t need_len(const u8s* src, size_t srcLen); static SIMSTR_API size_t need_len(const u8s* src, size_t srcLen);
@ -609,16 +620,6 @@ struct utf_convert_selector<u8s, u32s> {
static SIMSTR_API size_t convert(const u8s* src, size_t srcLen, u32s* dest); static SIMSTR_API size_t convert(const u8s* src, size_t srcLen, u32s* dest);
}; };
template<>
struct utf_convert_selector<u8s, wchar_t> {
static size_t need_len(const u8s* src, size_t srcLen) {
return utf_convert_selector<u8s, wchar_type>::need_len(src, srcLen);
}
static size_t convert(const u8s* src, size_t srcLen, wchar_t* dest) {
return utf_convert_selector<u8s, wchar_type>::convert(src, srcLen, to_w(dest));
}
};
template<> template<>
struct utf_convert_selector<u16s, u8s> { struct utf_convert_selector<u16s, u8s> {
static SIMSTR_API size_t need_len(const u16s* src, size_t srcLen); static SIMSTR_API size_t need_len(const u16s* src, size_t srcLen);
@ -631,42 +632,6 @@ struct utf_convert_selector<u16s, u32s> {
static SIMSTR_API size_t convert(const u16s* src, size_t srcLen, u32s* dest); static SIMSTR_API size_t convert(const u16s* src, size_t srcLen, u32s* dest);
}; };
template<>
struct utf_convert_selector<u16s, u16s> {
// При конвертации char16_t в wchar_t под windows будет вызываться эта реализация
// When converting char16_t to wchar_t under windows this implementation will be called
static size_t need_len(const u16s* src, size_t srcLen) {
return srcLen;
}
static size_t convert(const u16s* src, size_t srcLen, u16s* dest) {
ch_traits<u16s>::copy(dest, src, srcLen + 1);
return srcLen;
}
};
template<>
struct utf_convert_selector<u32s, u32s> {
// При конвертации char32_t в wchar_t под linux будет вызываться эта реализация
// When converting char32_t to wchar_t under Linux, this implementation will be called
static size_t need_len(const u32s* src, size_t srcLen) {
return srcLen;
}
static size_t convert(const u32s* src, size_t srcLen, u32s* dest) {
ch_traits<u32s>::copy(dest, src, srcLen + 1);
return srcLen;
}
};
template<>
struct utf_convert_selector<u16s, wchar_t> {
static size_t need_len(const u16s* src, size_t srcLen) {
return utf_convert_selector<u16s, wchar_type>::need_len(src, srcLen);
}
static size_t convert(const u16s* src, size_t srcLen, wchar_t* dest) {
return utf_convert_selector<u16s, wchar_type>::convert(src, srcLen, to_w(dest));
}
};
template<> template<>
struct utf_convert_selector<u32s, u8s> { struct utf_convert_selector<u32s, u8s> {
static SIMSTR_API size_t need_len(const u32s* src, size_t srcLen); static SIMSTR_API size_t need_len(const u32s* src, size_t srcLen);
@ -679,68 +644,6 @@ struct utf_convert_selector<u32s, u16s> {
static SIMSTR_API size_t convert(const u32s* src, size_t srcLen, u16s* dest); static SIMSTR_API size_t convert(const u32s* src, size_t srcLen, u16s* dest);
}; };
template<>
struct utf_convert_selector<u32s, wchar_t> {
static size_t need_len(const u32s* src, size_t srcLen) {
return utf_convert_selector<u32s, wchar_type>::need_len(src, srcLen);
}
static size_t convert(const u32s* src, size_t srcLen, wchar_t* dest) {
return utf_convert_selector<u32s, wchar_type>::convert(src, srcLen, to_w(dest));
}
};
template<>
struct utf_convert_selector<wchar_t, u8s> {
static size_t need_len(const wchar_t* src, size_t srcLen) {
return utf_convert_selector<wchar_type, u8s>::need_len(to_w(src), srcLen);
}
static size_t convert(const wchar_t* src, size_t srcLen, u8s* dest) {
return utf_convert_selector<wchar_type, u8s>::convert(to_w(src), srcLen, dest);
}
};
template<>
struct utf_convert_selector<wchar_t, u16s> {
static size_t need_len(const wchar_t* src, size_t srcLen) {
return utf_convert_selector<wchar_type, u16s>::need_len(to_w(src), srcLen);
}
static size_t convert(const wchar_t* src, size_t srcLen, u16s* dest) {
return utf_convert_selector<wchar_type, u16s>::convert(to_w(src), srcLen, dest);
}
};
template<>
struct utf_convert_selector<wchar_t, u32s> {
static size_t need_len(const wchar_t* src, size_t srcLen) {
return utf_convert_selector<wchar_type, u32s>::need_len(to_w(src), srcLen);
}
static size_t convert(const wchar_t* src, size_t srcLen, u32s* dest) {
return utf_convert_selector<wchar_type, u32s>::convert(to_w(src), srcLen, dest);
}
};
template<>
struct utf_convert_selector<u8s, ubs> {
static size_t need_len(const u8s* src, size_t srcLen) {
return srcLen;
}
static size_t convert(const u8s* src, size_t srcLen, ubs* dest) {
ch_traits<u8s>::copy((u8s*)dest, src, srcLen);
return srcLen;
}
};
template<>
struct utf_convert_selector<ubs, u8s> {
static size_t need_len(const ubs* src, size_t srcLen) {
return srcLen;
}
static size_t convert(const ubs* src, size_t srcLen, u8s* dest) {
ch_traits<u8s>::copy(dest, (const u8s*)src, srcLen);
return srcLen;
}
};
/*! /*!
* @ru @brief Базовый класс для строк, могущих конвертироваться из другого типа символов. * @ru @brief Базовый класс для строк, могущих конвертироваться из другого типа символов.
* @tparam K - тип символов. * @tparam K - тип символов.
@ -769,16 +672,19 @@ protected:
template<typename O> template<typename O>
requires(!std::is_same_v<O, K>) requires(!std::is_same_v<O, K>)
void init_from_utf_convertible(simple_str<O> init) { void init_from_utf_convertible(simple_str<O> init) {
using worker = utf_convert_selector<O, K>; using from_t = to_base_char_t<O>;
using to_t = to_base_char_t<K>;
using worker = utf_convert_selector<from_t, to_t>;
Impl* d = static_cast<Impl*>(this); Impl* d = static_cast<Impl*>(this);
size_t len = init.length(); size_t len = init.length();
if (!len) if (!len)
d->create_empty(); d->create_empty();
else { else {
size_t need = worker::need_len(init.symbols(), len); size_t need = worker::need_len((const from_t*)init.symbols(), len);
K* str = d->init(need); K* str = d->init(need);
str[need] = 0; str[need] = 0;
worker::convert(init.symbols(), len, str); worker::convert((const from_t*)init.symbols(), len, (to_t*)str);
} }
} }
}; };
@ -794,17 +700,19 @@ protected:
template<typename From, typename To> requires (!std::is_same_v<From, To>) template<typename From, typename To> requires (!std::is_same_v<From, To>)
struct expr_utf : expr_to_std_string<expr_utf<From, To>> { struct expr_utf : expr_to_std_string<expr_utf<From, To>> {
using symb_type = To; using symb_type = To;
using worker = utf_convert_selector<From, To>; using from_t = to_base_char_t<From>;
using to_t = to_base_char_t<To>;
using worker = utf_convert_selector<from_t, to_t>;
simple_str<From> source_; simple_str<From> source_;
constexpr expr_utf(simple_str<From> source) : source_(source){} constexpr expr_utf(simple_str<From> source) : source_(source){}
size_t length() const noexcept { size_t length() const noexcept {
return worker::need_len(source_.symbols(), source_.length()); return worker::need_len((const from_t*)source_.symbols(), source_.length());
} }
To* place(To* ptr) const noexcept { To* place(To* ptr) const noexcept {
return ptr + worker::convert(source_.symbols(), source_.length(), ptr); return ptr + worker::convert((const from_t*)source_.symbols(), source_.length(), (to_t*)ptr);
} }
}; };
@ -1376,24 +1284,6 @@ inline size_t grow2(size_t ret, size_t currentCapacity) {
return ret <= currentCapacity ? ret : ret * 2; return ret <= currentCapacity ? ret : ret * 2;
} }
template<typename K>
struct to_std_char_type : std::type_identity<K>{};
template<>
struct to_std_char_type<char8_t>{
using type = char;
};
template<>
struct to_std_char_type<char16_t>{
using type = std::conditional_t<sizeof(char16_t) == sizeof(wchar_t), wchar_t, void>;
};
template<>
struct to_std_char_type<char32_t>{
using type = std::conditional_t<sizeof(char32_t) == sizeof(wchar_t), wchar_t, void>;
};
/*! /*!
* @ru @brief Базовый класс работы с изменяемыми строками * @ru @brief Базовый класс работы с изменяемыми строками
* @tparam K - тип символов * @tparam K - тип символов
@ -2391,7 +2281,7 @@ public:
writer& operator=(writer&&) noexcept = default; writer& operator=(writer&&) noexcept = default;
using difference_type = int; using difference_type = int;
}; };
using fmt_type = typename to_std_char_type<K>::type; using fmt_type = to_std_char_t<K>;
/*! /*!
* @ru @brief Добавляет отформатированный с помощью std::format вывод, начиная с указанной позиции. * @ru @brief Добавляет отформатированный с помощью std::format вывод, начиная с указанной позиции.
* @param from - начальная позиция добавления. * @param from - начальная позиция добавления.
@ -2715,14 +2605,15 @@ protected:
} }
constexpr K* init(size_t s) { constexpr K* init(size_t s) {
size_ = s; size_t need_cap = s;
if (size_ > LocalCapacity) { if (need_cap > LocalCapacity) {
s = calc_capacity(s); need_cap = calc_capacity(s);
data_ = alloc_place(s); data_ = alloc_place(need_cap);
capacity_ = s; capacity_ = need_cap;
} else { } else {
data_ = local_; data_ = local_;
} }
size_ = s;
return str(); return str();
} }
// Методы для себя | Methods for yourself // Методы для себя | Methods for yourself
@ -2886,7 +2777,14 @@ public:
*/ */
constexpr lstring(const my_type& other) : base_storable(other.allocator()) { constexpr lstring(const my_type& other) : base_storable(other.allocator()) {
if (other.size_) { if (other.size_) {
traits::copy(init(other.size_), other.symbols(), other.size_ + 1); K* buf = init(other.size_);
const size_t short_str = 16 / sizeof(K);
if (LocalCapacity >= short_str - 1 && other.size_ < short_str) {
struct copy { char buf[16]; };
*reinterpret_cast<copy*>(buf) = *reinterpret_cast<const copy*>(other.symbols());
} else {
traits::copy(buf, other.symbols(), other.size_ + 1);
}
} }
} }
/*! /*!
@ -2917,7 +2815,7 @@ public:
constexpr lstring(T&& value, Args&&... args) : base_storable(std::forward<Args>(args)...) { constexpr lstring(T&& value, Args&&... args) : base_storable(std::forward<Args>(args)...) {
if constexpr (I > 1) { if constexpr (I > 1) {
K* ptr = init(I - 1); K* ptr = init(I - 1);
traits::copy(ptr, value, I - 1); traits::copy(ptr, (const K*)value, I - 1);
ptr[I - 1] = 0; ptr[I - 1] = 0;
} else } else
create_empty(); create_empty();
@ -3057,7 +2955,7 @@ public:
*/ */
template<typename T, size_t S = const_lit_for<K, T>::Count> template<typename T, size_t S = const_lit_for<K, T>::Count>
my_type& operator=(T&& other) { my_type& operator=(T&& other) {
return assign(other, S - 1); return assign((const K*)other, S - 1);
} }
/*! /*!
* @ru @brief Оператор присваивания строкового выражения. * @ru @brief Оператор присваивания строкового выражения.
@ -3636,7 +3534,7 @@ public:
sstring(T&& s, Args&&... args) : base_storable(std::forward<Args>(args)...) { sstring(T&& s, Args&&... args) : base_storable(std::forward<Args>(args)...) {
type_ = Constant; type_ = Constant;
localRemain_ = 0; localRemain_ = 0;
cstr_ = s; cstr_ = (const K*)s;
bigLen_ = N - 1; bigLen_ = N - 1;
} }
@ -3696,7 +3594,7 @@ public:
*/ */
template<typename T, size_t N = const_lit_for<K, T>::Count> template<typename T, size_t N = const_lit_for<K, T>::Count>
constexpr my_type& operator=(T&& other) { constexpr my_type& operator=(T&& other) {
return operator=(my_type{other, base_storable::allocator()}); return operator=(my_type{std::forward<T>(other), base_storable::allocator()});
} }
/*! /*!
* @ru @brief Оператор присвоения другой строки типа lstring. * @ru @brief Оператор присвоения другой строки типа lstring.
@ -3790,7 +3688,7 @@ public:
* @return my_type. * @return my_type.
*/ */
template<typename... T> template<typename... T>
static my_type format(const FmtString<typename to_std_char_type<K>::type, T...>& fmtString, T&&... args) { static my_type format(const FmtString<to_std_char_t<K>, T...>& fmtString, T&&... args) {
return my_type{lstring<K, 256, true, Allocator>{}.format(fmtString, std::forward<T>(args)...)}; return my_type{lstring<K, 256, true, Allocator>{}.format(fmtString, std::forward<T>(args)...)};
} }
/*! /*!
@ -3964,7 +3862,7 @@ public:
* @details In this case, we simply remember the pointer to the string and its length. * @details In this case, we simply remember the pointer to the string and its length.
*/ */
template<typename T, size_t M = const_lit_for<K, T>::Count> template<typename T, size_t M = const_lit_for<K, T>::Count>
constexpr cestring(T&& s) : base_storable(), cstr_(s), size_(M - 1), is_cstr_(true), local_{0} {} constexpr cestring(T&& s) : base_storable(), cstr_((const K*)s), size_(M - 1), is_cstr_(true), local_{0} {}
}; };
template<typename K> template<typename K>

47
include/simstr/strexpr.h
View File

@ -1,5 +1,5 @@
/* /*
* ver. 1.6.5 * ver. 1.6.6
* (c) Проект "SimStr", Александр Орефков orefkov@gmail.com * (c) Проект "SimStr", Александр Орефков orefkov@gmail.com
* База для строковых конкатенаций через выражения времени компиляции * База для строковых конкатенаций через выражения времени компиляции
* (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com * (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com
@ -116,6 +116,43 @@ auto to_one_of_std_char(const From* from) {
} }
} }
template<typename K>
struct to_std_char_type : std::type_identity<K>{};
template<>
struct to_std_char_type<char8_t>{
using type = char;
};
template<>
struct to_std_char_type<char16_t>{
using type = std::conditional_t<sizeof(char16_t) == sizeof(wchar_t), wchar_t, void>;
};
template<>
struct to_std_char_type<char32_t>{
using type = std::conditional_t<sizeof(char32_t) == sizeof(wchar_t), wchar_t, void>;
};
template<typename K>
using to_std_char_t = typename to_std_char_type<K>::type;
template<typename K>
struct to_base_char_type : std::type_identity<K>{};
template<>
struct to_base_char_type<char8_t>{
using type = char;
};
template<>
struct to_base_char_type<wchar_t>{
using type = std::conditional_t<sizeof(char16_t) == sizeof(wchar_t), char16_t, char32_t>;
};
template<typename K>
using to_base_char_t = typename to_base_char_type<K>::type;
/*! /*!
* @ingroup StrExprs * @ingroup StrExprs
* @ru @brief Проверка, являются ли два типа совместимыми строковыми типами. * @ru @brief Проверка, являются ли два типа совместимыми строковыми типами.
@ -226,7 +263,7 @@ public:
template<typename T, size_t M = const_lit_for<K, T>::Count> requires (M == N + 1) template<typename T, size_t M = const_lit_for<K, T>::Count> requires (M == N + 1)
constexpr const_lit_to_array(T&& s) constexpr const_lit_to_array(T&& s)
: symbols_(s) {} : symbols_((const K(&)[M])s) {}
constexpr bool contain(K s) const { constexpr bool contain(K s) const {
return exist<0>(s); return exist<0>(s);
@ -3569,7 +3606,7 @@ public:
*/ */
template<typename T, size_t N = const_lit_for<K, T>::Count> template<typename T, size_t N = const_lit_for<K, T>::Count>
constexpr bool operator==(T&& other) const noexcept { constexpr bool operator==(T&& other) const noexcept {
return N - 1 == _len() && traits::compare(_str(), other, N - 1) == 0; return N - 1 == _len() && traits::compare(_str(), (const K*)other, N - 1) == 0;
} }
/*! /*!
* @ru @brief Оператор сравнения строки и строкового литерала. * @ru @brief Оператор сравнения строки и строкового литерала.
@ -3580,7 +3617,7 @@ public:
template<typename T, size_t N = const_lit_for<K, T>::Count> template<typename T, size_t N = const_lit_for<K, T>::Count>
constexpr auto operator<=>(T&& other) const noexcept { constexpr auto operator<=>(T&& other) const noexcept {
size_t myLen = _len(); size_t myLen = _len();
int cmp = traits::compare(_str(), other, std::min(myLen, N - 1)); int cmp = traits::compare(_str(), (const K*)other, std::min(myLen, N - 1));
int res = cmp == 0 ? (myLen > N - 1 ? 1 : myLen == N - 1 ? 0 : -1) : cmp; int res = cmp == 0 ? (myLen > N - 1 ? 1 : myLen == N - 1 ? 0 : -1) : cmp;
return res <=> 0; return res <=> 0;
} }
@ -4799,7 +4836,7 @@ struct str_src : str_src_algs<K, str_src<K>, str_src<K>, false> {
* @en @brief Constructor from a string literal. * @en @brief Constructor from a string literal.
*/ */
template<typename T, size_t N = const_lit_for<K, T>::Count> template<typename T, size_t N = const_lit_for<K, T>::Count>
constexpr str_src(T&& v) noexcept : str(v), len(N - 1) {} constexpr str_src(T&& v) noexcept : str((const K*)v), len(N - 1) {}
/*! /*!
* @ru @brief Конструктор из указателя и длины. * @ru @brief Конструктор из указателя и длины.

9
readme.md
View File

@ -3,7 +3,7 @@
[![CMake on multiple platforms](https://github.com/orefkov/simstr/actions/workflows/cmake-multi-platform.yml/badge.svg)](https://github.com/orefkov/simstr/actions/workflows/cmake-multi-platform.yml) [![CMake on multiple platforms](https://github.com/orefkov/simstr/actions/workflows/cmake-multi-platform.yml/badge.svg)](https://github.com/orefkov/simstr/actions/workflows/cmake-multi-platform.yml)
Version 1.6.5. Version 1.6.6.
<h2>Speed up your work with strings by 2-10 times!</h2> <h2>Speed up your work with strings by 2-10 times!</h2>
@ -32,7 +32,8 @@ The library contains two parts:
```cpp ```cpp
#include "path/to file/strexpr.h" #include "path/to file/strexpr.h"
``` ```
This will allow you to use powerful and fast *"string expressions"* for concatenation and string construction for standard string types (`std::basic_string`, `std::basic_string_view`), as well as simplified versions of the `simple_str` and This will allow you to use powerful and fast *"string expressions"* for concatenation and string construction for standard string types
(`std::basic_string`, `std::basic_string_view`), as well as simplified versions of the `simple_str` and
`simple_str_nt` classes, which implement all those string algorithms of the library that do not require storing or modifying strings. `simple_str_nt` classes, which implement all those string algorithms of the library that do not require storing or modifying strings.
Since this is a header-only part, it does not include working with UTF encodings and simplified Unicode. Since this is a header-only part, it does not include working with UTF encodings and simplified Unicode.
- The full version, requiring the connection of the entire library (`"include/simstr/sstring.h"`), adds its string types with - The full version, requiring the connection of the entire library (`"include/simstr/sstring.h"`), adds its string types with
@ -322,8 +323,8 @@ function(add_simstr)
simstr simstr
GIT_REPOSITORY https://github.com/orefkov/simstr.git GIT_REPOSITORY https://github.com/orefkov/simstr.git
GIT_SHALLOW TRUE GIT_SHALLOW TRUE
GIT_TAG tags/rel1.6.5 # Specify the desired release GIT_TAG tags/rel1.6.6 # Specify the desired release
FIND_PACKAGE_ARGS NAMES simstr 1.6.5 FIND_PACKAGE_ARGS NAMES simstr 1.6.6
) )
FetchContent_MakeAvailable(simstr) FetchContent_MakeAvailable(simstr)
endfunction() endfunction()

6
readme_ru.md
View File

@ -3,7 +3,7 @@
[![CMake on multiple platforms](https://github.com/orefkov/simstr/actions/workflows/cmake-multi-platform.yml/badge.svg)](https://github.com/orefkov/simstr/actions/workflows/cmake-multi-platform.yml) [![CMake on multiple platforms](https://github.com/orefkov/simstr/actions/workflows/cmake-multi-platform.yml/badge.svg)](https://github.com/orefkov/simstr/actions/workflows/cmake-multi-platform.yml)
Версия 1.6.5. Версия 1.6.6.
<h2>Ускорь работу со строками в 2-10 раз!</h2> <h2>Ускорь работу со строками в 2-10 раз!</h2>
@ -323,8 +323,8 @@ function(add_simstr)
simstr simstr
GIT_REPOSITORY https://github.com/orefkov/simstr.git GIT_REPOSITORY https://github.com/orefkov/simstr.git
GIT_SHALLOW TRUE GIT_SHALLOW TRUE
GIT_TAG tags/rel1.6.5 # Укажите нужный релиз GIT_TAG tags/rel1.6.6 # Укажите нужный релиз
FIND_PACKAGE_ARGS NAMES simstr 1.6.5 FIND_PACKAGE_ARGS NAMES simstr 1.6.6
) )
FetchContent_MakeAvailable(simstr) FetchContent_MakeAvailable(simstr)
endfunction() endfunction()

2
src/sstring.cpp
View File

@ -1,5 +1,5 @@
/* /*
* ver. 1.6.5 * ver. 1.6.6
* (c) Проект "SimStr", Александр Орефков orefkov@gmail.com * (c) Проект "SimStr", Александр Орефков orefkov@gmail.com
* Реализация строковых функций * Реализация строковых функций
* (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com * (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com

2
tests/test_expr_only.cpp
View File

@ -1,5 +1,5 @@
/* /*
* ver. 1.6.5 * ver. 1.6.6
* (c) Проект "SimStr", Александр Орефков orefkov@gmail.com * (c) Проект "SimStr", Александр Орефков orefkov@gmail.com
* Тесты simstr * Тесты simstr
* (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com * (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com

14
tests/test_str.cpp
View File

@ -1,5 +1,5 @@
/* /*
* ver. 1.6.5 * ver. 1.6.6
* (c) Проект "SimStr", Александр Орефков orefkov@gmail.com * (c) Проект "SimStr", Александр Орефков orefkov@gmail.com
* Тесты simstr * Тесты simstr
* (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com * (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com
@ -958,13 +958,23 @@ TEST(SimStr, LStringAssign) {
TEST(SimStr, UtfConvert) { TEST(SimStr, UtfConvert) {
EXPECT_EQ(stringa{ssu{u"testпройден"}}, "testпройден"); EXPECT_EQ(stringa{ssu{u"testпройден"}}, "testпройден");
EXPECT_EQ(stringa{ssb{u8"testпройден"}}, "testпройден");
EXPECT_EQ(stringa{ssw{L"testпройден"}}, "testпройден"); EXPECT_EQ(stringa{ssw{L"testпройден"}}, "testпройден");
EXPECT_EQ(stringa{stringu{u"testпройден"}}, "testпройден"); EXPECT_EQ(stringa{stringu{u"testпройден"}}, "testпройден");
EXPECT_EQ(stringa{stringw{L"testпройден"}}, "testпройден"); EXPECT_EQ(stringa{stringw{L"testпройден"}}, "testпройден");
EXPECT_EQ(stringa{lstringu<40>{u"testпройден"}}, "testпройден"); EXPECT_EQ(stringa{lstringu<40>{u"testпройден"}}, "testпройден");
EXPECT_EQ(stringa{lstringw<40>{L"testпройден"}}, "testпройден"); EXPECT_EQ(stringa{lstringw<40>{L"testпройден"}}, "testпройден");
EXPECT_EQ(stringb{ssu{u"testпройден"}}, u8"testпройден");
EXPECT_EQ(stringb{ssa{u8"testпройден"}}, u8"testпройден");
EXPECT_EQ(stringb{ssw{L"testпройден"}}, u8"testпройден");
EXPECT_EQ(stringb{stringu{u"testпройден"}}, u8"testпройден");
EXPECT_EQ(stringb{stringw{L"testпройден"}}, u8"testпройден");
EXPECT_EQ(stringb{lstringu<40>{u"testпройден"}}, u8"testпройден");
EXPECT_EQ(stringb{lstringw<40>{L"testпройден"}}, u8"testпройден");
EXPECT_EQ(stringu{ssa{"testпройден"}}, u"testпройден"); EXPECT_EQ(stringu{ssa{"testпройден"}}, u"testпройден");
EXPECT_EQ(stringu{ssb{u8"testпройден"}}, u"testпройден");
EXPECT_EQ(stringu{ssw{L"testпройден"}}, u"testпройден"); EXPECT_EQ(stringu{ssw{L"testпройден"}}, u"testпройден");
EXPECT_EQ(stringu{stringa{"testпройден"}}, u"testпройден"); EXPECT_EQ(stringu{stringa{"testпройден"}}, u"testпройден");
EXPECT_EQ(stringu{stringw{L"testпройден"}}, u"testпройден"); EXPECT_EQ(stringu{stringw{L"testпройден"}}, u"testпройден");
@ -1384,7 +1394,7 @@ TEST(SimStr, LStrJoinAndExpressions) {
int k = 10; int k = 10;
double d = 12.1; double d = 12.1;
lstringa<10> test = ">" + e_repl(buffer.to_str(), "<>", "-") + k + ", " + d; lstringa<10> test = ">" + e_repl(buffer, "<>", "-") + k + ", " + d;
EXPECT_EQ(test, ">asd-fgh-jkl10, 12.1"); EXPECT_EQ(test, ">asd-fgh-jkl10, 12.1");
buffer.prepend(e_choice(test.length() > 2, eea + 99, eea + 12.1 + "asd")); buffer.prepend(e_choice(test.length() > 2, eea + 99, eea + 12.1 + "asd"));

2
tests/test_tostrexpr.cpp
View File

@ -1,5 +1,5 @@
/* /*
* ver. 1.6.5 * ver. 1.6.6
* (c) Проект "SimStr", Александр Орефков orefkov@gmail.com * (c) Проект "SimStr", Александр Орефков orefkov@gmail.com
* Тесты simstr * Тесты simstr
* (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com * (c) Project "SimStr", Aleksandr Orefkov orefkov@gmail.com