diff options
author | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-27 16:51:28 +0000 |
---|---|---|
committer | Daniel Baumann <daniel.baumann@progress-linux.org> | 2024-04-27 16:51:28 +0000 |
commit | 940b4d1848e8c70ab7642901a68594e8016caffc (patch) | |
tree | eb72f344ee6c3d9b80a7ecc079ea79e9fba8676d /include/rtl/stringutils.hxx | |
parent | Initial commit. (diff) | |
download | libreoffice-940b4d1848e8c70ab7642901a68594e8016caffc.tar.xz libreoffice-940b4d1848e8c70ab7642901a68594e8016caffc.zip |
Adding upstream version 1:7.0.4.upstream/1%7.0.4upstream
Signed-off-by: Daniel Baumann <daniel.baumann@progress-linux.org>
Diffstat (limited to 'include/rtl/stringutils.hxx')
-rw-r--r-- | include/rtl/stringutils.hxx | 354 |
1 files changed, 354 insertions, 0 deletions
diff --git a/include/rtl/stringutils.hxx b/include/rtl/stringutils.hxx new file mode 100644 index 000000000..b96884546 --- /dev/null +++ b/include/rtl/stringutils.hxx @@ -0,0 +1,354 @@ +/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*- */ +/* + * This file is part of the LibreOffice project. + * + * This Source Code Form is subject to the terms of the Mozilla Public + * License, v. 2.0. If a copy of the MPL was not distributed with this + * file, You can obtain one at http://mozilla.org/MPL/2.0/. + */ + +#ifndef INCLUDED_RTL_STRINGUTILS_HXX +#define INCLUDED_RTL_STRINGUTILS_HXX + +#include "sal/config.h" + +#include <cassert> +#include <cstddef> + +#include "sal/types.h" + +// The unittest uses slightly different code to help check that the proper +// calls are made. The class is put into a different namespace to make +// sure the compiler generates a different (if generating also non-inline) +// copy of the function and does not merge them together. The class +// is "brought" into the proper rtl namespace by a typedef below. +#ifdef RTL_STRING_UNITTEST +#define rtl rtlunittest +#endif + +namespace rtl +{ + +#ifdef RTL_STRING_UNITTEST +#undef rtl +#endif + +#if defined LIBO_INTERNAL_ONLY +/// @cond INTERNAL + +// A simple wrapper around a single char. Can be useful in string concatenation contexts, like in +// +// OString s = ...; +// char c = ...; +// s += OStringChar(c); +// +struct SAL_WARN_UNUSED OStringChar { + constexpr OStringChar(char theC): c(theC) {} + template<typename T> OStringChar(T &&) = delete; + char const c; +}; + +/** A simple wrapper around a single sal_Unicode character. + + Can be useful to pass a sal_Unicode constant into an OUString-related + function that is optimized for UTF-16 string literal arguments. That is, + instead of + + sal_Unicode const WILDCARD = '%'; + ... + if (s[i] == WILDCARD) ... + ... + if (s.endsWith(OUString(WILDCARD))) ... + + use + + sal_Unicode const WILDCARD = '%'; + ... + if (s[i] == WILDCARD) ... + ... + if (s.endsWith(OUStringChar(WILDCARD))) ... + + to avoid creating a temporary OUString instance, and instead pick the + endsWith overload actually designed to take an argument of type + sal_Unicode const[N]. + + (Because of the above use case, + instances of OUStringChar need to be const, as those literal-optimized + functions take the literal argument by non-const lvalue reference, for + technical reasons. + + For actual arrays, it is important to distinguish string literals from other char or sal_Unicode + arrays, which may contain junk after the first NUL character or may be non-ASCII in the case of + char arrays. This is not so much a concern for single char and sal_Unicode values, where NUL is + assumed to always be meant as an actual character.) + + Can also be useful in string concatenation contexts, like in + + sal_Unicode const * s = ...; + sal_Unicode c = ...; + OUString t = s + OUStringChar(c); + + @since LibreOffice 5.0 +*/ +struct SAL_WARN_UNUSED OUStringChar_ { + constexpr OUStringChar_(sal_Unicode theC): c(theC) {} + constexpr OUStringChar_(char theC): c(theC) { assert(c <= 0x7F); } + template<typename T> OUStringChar_(T &&) = delete; + sal_Unicode const c; +}; +using OUStringChar = OUStringChar_ const; + +/// @endcond +#endif + +namespace libreoffice_internal +{ +/* +These templates use SFINAE (Substitution failure is not an error) to help distinguish the various +plain C string types: char*, const char*, char[N], const char[N], char[] and const char[]. +There are 2 cases: +1) Only string literal (i.e. const char[N]) is wanted, not any of the others. + In this case it is necessary to distinguish between const char[N] and char[N], as the latter + would be automatically converted to the const variant, which is not wanted (not a string literal + with known size of the content). In this case ConstCharArrayDetector is used to ensure the function + is called only with const char[N] arguments. There's no other plain C string type overload. +2) All plain C string types are wanted, and const char[N] needs to be handled differently. + In this case const char[N] would match const char* argument type (not exactly sure why, but it's + consistent in all of gcc, clang and msvc). Using a template with a reference to const of the type + avoids this problem, and CharPtrDetector ensures that the function is called only with char pointer + arguments. The const in the argument is necessary to handle the case when something is explicitly + cast to const char*. Additionally (non-const) char[N] needs to be handled, but with the reference + being const, it would also match const char[N], so another overload with a reference to non-const + and NonConstCharArrayDetector are used to ensure the function is called only with (non-const) char[N]. +Additionally, char[] and const char[] (i.e. size unknown) are rather tricky. Their usage with 'T&' would +mean it would be 'char(&)[]', which seems to be invalid. But gcc and clang somehow manage when it is +a template. while msvc complains about no conversion from char[] to char[1]. And the reference cannot +be avoided, because 'const char[]' as argument type would match also 'const char[N]' +So char[] and const char[] should always be used with their contents specified (which automatically +turns them into char[N] or const char[N]), or char* and const char* should be used. +*/ +struct Dummy {}; +template< typename T1, typename T2 = void > +struct CharPtrDetector +{ + static const bool ok = false; +}; +template< typename T > +struct CharPtrDetector< const char*, T > +{ + typedef T Type; + static const bool ok = true; +}; +template< typename T > +struct CharPtrDetector< char*, T > +{ + typedef T Type; + static const bool ok = true; +}; +#if defined LIBO_INTERNAL_ONLY +template<typename T> struct CharPtrDetector<sal_Unicode *, T> { using TypeUtf16 = T; }; +template<typename T> struct CharPtrDetector<sal_Unicode const *, T> { using TypeUtf16 = T; }; +template<typename T> struct CharPtrDetector<sal_Unicode[], T> { using TypeUtf16 = T; }; +template<typename T> struct CharPtrDetector<sal_Unicode const[], T> { using TypeUtf16 = T; }; +#endif + +template< typename T1, typename T2 > +struct NonConstCharArrayDetector +{ +}; +template< typename T, int N > +struct NonConstCharArrayDetector< char[ N ], T > +{ + typedef T Type; +}; +#ifdef RTL_STRING_UNITTEST +// never use, until all compilers handle this +template< typename T > +struct NonConstCharArrayDetector< char[], T > +{ + typedef T Type; +}; +template< typename T > +struct NonConstCharArrayDetector< const char[], T > +{ + typedef T Type; +}; +#endif +#if defined LIBO_INTERNAL_ONLY +template<typename T, std::size_t N> struct NonConstCharArrayDetector<sal_Unicode[N], T> { + using TypeUtf16 = T; +}; +#endif + +template< typename T1, typename T2 = void > +struct ConstCharArrayDetector +{ + static const bool ok = false; +}; +template< std::size_t N, typename T > +struct ConstCharArrayDetector< const char[ N ], T > +{ + typedef T Type; + static const std::size_t length = N - 1; + static const bool ok = true; +#if defined LIBO_INTERNAL_ONLY + constexpr +#endif + static bool isValid(char const (& literal)[N]) { + for (std::size_t i = 0; i != N - 1; ++i) { + if (literal[i] == '\0') { + return false; + } + } + return literal[N - 1] == '\0'; + } +#if defined LIBO_INTERNAL_ONLY + constexpr +#endif + static char const * toPointer(char const (& literal)[N]) { return literal; } +}; + +#if defined(__COVERITY__) +//to silence over zealous warnings that the loop is logically dead +//for the single char case +template< typename T > +struct ConstCharArrayDetector< const char[ 1 ], T > +{ + typedef T Type; + static const std::size_t length = 0; + static const bool ok = true; +#if defined LIBO_INTERNAL_ONLY + constexpr +#endif + static bool isValid(char const (& literal)[1]) { + return literal[0] == '\0'; + } +#if defined LIBO_INTERNAL_ONLY + constexpr +#endif + static char const * toPointer(char const (& literal)[1]) { return literal; } +}; +#endif + +#if defined LIBO_INTERNAL_ONLY && defined __cpp_char8_t +template<std::size_t N, typename T> +struct ConstCharArrayDetector<char8_t const [N], T> { + using Type = T; + static constexpr bool const ok = true; + static constexpr std::size_t const length = N - 1; + static constexpr bool isValid(char8_t const (& literal)[N]) { + for (std::size_t i = 0; i != N - 1; ++i) { + if (literal[i] == u8'\0') { + return false; + } + } + return literal[N - 1] == u8'\0'; + } + static constexpr char const * toPointer(char8_t const (& literal)[N]) + { return reinterpret_cast<char const *>(literal); } +}; +#endif + +#if defined LIBO_INTERNAL_ONLY +template<std::size_t N, typename T> +struct ConstCharArrayDetector<sal_Unicode const [N], T> { + using TypeUtf16 = T; + static constexpr bool const ok = true; + static constexpr std::size_t const length = N - 1; + static constexpr sal_Unicode const * toPointer( + sal_Unicode const (& literal)[N]) + { return literal; } +}; +template<typename T> struct ConstCharArrayDetector< + OUStringChar, + T> +{ + using TypeUtf16 = T; + static constexpr bool const ok = true; + static constexpr std::size_t const length = 1; + static constexpr sal_Unicode const * toPointer( + OUStringChar_ const & literal) + { return &literal.c; } +}; +#endif + +// this one is used to rule out only const char[N] +template< typename T > +struct ExceptConstCharArrayDetector +{ + typedef Dummy Type; +}; +template< int N > +struct ExceptConstCharArrayDetector< const char[ N ] > +{ +}; +#if defined LIBO_INTERNAL_ONLY +template<std::size_t N> +struct ExceptConstCharArrayDetector<sal_Unicode const[N]> {}; +template<> struct ExceptConstCharArrayDetector< + OUStringChar + > +{}; +#endif + +// this one is used to rule out only const char[N] +// (const will be brought in by 'const T&' in the function call) +// msvc needs const char[N] here (not sure whether gcc or msvc +// are right, it doesn't matter). +template< typename T > +struct ExceptCharArrayDetector +{ + typedef Dummy Type; +}; +template< int N > +struct ExceptCharArrayDetector< char[ N ] > +{ +}; +template< int N > +struct ExceptCharArrayDetector< const char[ N ] > +{ +}; +#if defined LIBO_INTERNAL_ONLY +template<std::size_t N> struct ExceptCharArrayDetector<sal_Unicode[N]> {}; +template<std::size_t N> struct ExceptCharArrayDetector<sal_Unicode const[N]> {}; +template<> struct ExceptCharArrayDetector<OUStringChar_> {}; +#endif + +template< typename T1, typename T2 = void > +struct SalUnicodePtrDetector +{ + static const bool ok = false; +}; +template< typename T > +struct SalUnicodePtrDetector< const sal_Unicode*, T > +{ + typedef T Type; + static const bool ok = true; +}; +template< typename T > +struct SalUnicodePtrDetector< sal_Unicode*, T > +{ + typedef T Type; + static const bool ok = true; +}; + +// SFINAE helper class +template< typename T, bool > +struct Enable + { + }; + +template< typename T > +struct Enable< T, true > + { + typedef T Type; + }; + + +} /* Namespace */ + +} /* Namespace */ + +#endif // INCLUDED_RTL_STRINGUTILS_HXX + +/* vim:set shiftwidth=4 softtabstop=4 expandtab: */ |