feedc0de/boost_regex
1
0
Fork 0
forked from boostorg/regex
Code Pull Requests Activity

Make w32_regex_traits.hpp header only.

Browse Source
This commit is contained in:
jzmaddock
2020-11-28 12:49:59 +00:00
parent ddb682a208
commit 5002272ec8
7 changed files with 1039 additions and 79 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
build/Jamfile.v2
View File

@ -129,7 +129,6 @@ SOURCES =
regex_raw_buffer.cpp
regex_traits_defaults.cpp
static_mutex.cpp
w32_regex_traits.cpp
wide_posix_api.cpp
;

560
include/boost/regex/v4/w32_regex_traits.hpp
View File

@ -25,18 +25,27 @@
#include <boost/regex/pattern_except.hpp>
#endif
#ifndef BOOST_REGEX_TRAITS_DEFAULTS_HPP_INCLUDED
#include <boost/regex/v4/regex_traits_defaults.hpp>
#include <boost/regex/v5/regex_traits_defaults.hpp>
#endif
#ifdef BOOST_HAS_THREADS
#include <boost/regex/pending/static_mutex.hpp>
#endif
#ifndef BOOST_REGEX_PRIMARY_TRANSFORM
#include <boost/regex/v4/primary_transform.hpp>
#include <boost/regex/v5/primary_transform.hpp>
#endif
#ifndef BOOST_REGEX_OBJECT_CACHE_HPP
#include <boost/regex/pending/object_cache.hpp>
#endif
#define VC_EXTRALEAN
#define WIN32_LEAN_AND_MEAN
#include <windows.h>
#if defined(_MSC_VER) && !defined(_WIN32_WCE) && !defined(UNDER_CE)
#pragma comment(lib, "user32.lib")
#endif
#ifdef BOOST_MSVC
#pragma warning(push)
#pragma warning(disable: 4103)
@ -75,53 +84,35 @@ typedef ::boost::shared_ptr<void> cat_type; // placeholder for dll HANDLE.
//
// then add wrappers around the actual Win32 API's (ie implementation hiding):
//
BOOST_REGEX_DECL lcid_type BOOST_REGEX_CALL w32_get_default_locale();
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_lower(char, lcid_type);
lcid_type BOOST_REGEX_CALL w32_get_default_locale();
bool BOOST_REGEX_CALL w32_is_lower(char, lcid_type);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_lower(wchar_t, lcid_type);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_lower(unsigned short ca, lcid_type state_id);
bool BOOST_REGEX_CALL w32_is_lower(wchar_t, lcid_type);
#endif
#endif
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_upper(char, lcid_type);
bool BOOST_REGEX_CALL w32_is_upper(char, lcid_type);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_upper(wchar_t, lcid_type);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_upper(unsigned short ca, lcid_type state_id);
bool BOOST_REGEX_CALL w32_is_upper(wchar_t, lcid_type);
#endif
#endif
BOOST_REGEX_DECL cat_type BOOST_REGEX_CALL w32_cat_open(const std::string& name);
BOOST_REGEX_DECL std::string BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type state_id, int i, const std::string& def);
cat_type BOOST_REGEX_CALL w32_cat_open(const std::string& name);
std::string BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type state_id, int i, const std::string& def);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL std::wstring BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type state_id, int i, const std::wstring& def);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL std::basic_string<unsigned short> BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type, int i, const std::basic_string<unsigned short>& def);
std::wstring BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type state_id, int i, const std::wstring& def);
#endif
#endif
BOOST_REGEX_DECL std::string BOOST_REGEX_CALL w32_transform(lcid_type state_id, const char* p1, const char* p2);
std::string BOOST_REGEX_CALL w32_transform(lcid_type state_id, const char* p1, const char* p2);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL std::wstring BOOST_REGEX_CALL w32_transform(lcid_type state_id, const wchar_t* p1, const wchar_t* p2);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL std::basic_string<unsigned short> BOOST_REGEX_CALL w32_transform(lcid_type state_id, const unsigned short* p1, const unsigned short* p2);
std::wstring BOOST_REGEX_CALL w32_transform(lcid_type state_id, const wchar_t* p1, const wchar_t* p2);
#endif
#endif
BOOST_REGEX_DECL char BOOST_REGEX_CALL w32_tolower(char c, lcid_type);
char BOOST_REGEX_CALL w32_tolower(char c, lcid_type);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL wchar_t BOOST_REGEX_CALL w32_tolower(wchar_t c, lcid_type);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL unsigned short BOOST_REGEX_CALL w32_tolower(unsigned short c, lcid_type state_id);
wchar_t BOOST_REGEX_CALL w32_tolower(wchar_t c, lcid_type);
#endif
#endif
BOOST_REGEX_DECL char BOOST_REGEX_CALL w32_toupper(char c, lcid_type);
char BOOST_REGEX_CALL w32_toupper(char c, lcid_type);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL wchar_t BOOST_REGEX_CALL w32_toupper(wchar_t c, lcid_type);
wchar_t BOOST_REGEX_CALL w32_toupper(wchar_t c, lcid_type);
#endif
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is(lcid_type, boost::uint32_t mask, char c);
bool BOOST_REGEX_CALL w32_is(lcid_type, boost::uint32_t mask, char c);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is(lcid_type, boost::uint32_t mask, wchar_t c);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is(lcid_type state_id, boost::uint32_t m, unsigned short c);
#endif
bool BOOST_REGEX_CALL w32_is(lcid_type, boost::uint32_t mask, wchar_t c);
#endif
//
// class w32_regex_traits_base:
@ -253,14 +244,14 @@ typename w32_regex_traits_char_layer<charT>::string_type
// specialised version for narrow characters:
//
template <>
class BOOST_REGEX_DECL w32_regex_traits_char_layer<char> : public w32_regex_traits_base<char>
class w32_regex_traits_char_layer<char> : public w32_regex_traits_base<char>
{
typedef std::string string_type;
public:
w32_regex_traits_char_layer(::boost::BOOST_REGEX_DETAIL_NS::lcid_type l)
: w32_regex_traits_base<char>(l)
{
init();
init<char>();
}
regex_constants::syntax_type syntax_type(char c)const
@ -284,6 +275,7 @@ private:
regex_constants::syntax_type m_char_map[1u << CHAR_BIT];
char m_lower_map[1u << CHAR_BIT];
boost::uint16_t m_type_map[1u << CHAR_BIT];
template <class U>
void init();
};
@ -722,6 +714,498 @@ static_mutex& w32_regex_traits<charT>::get_mutex_inst()
}
#endif
namespace BOOST_REGEX_DETAIL_NS {
#ifdef BOOST_NO_ANSI_APIS
inline UINT get_code_page_for_locale_id(lcid_type idx)
{
WCHAR code_page_string[7];
if (::GetLocaleInfoW(idx, LOCALE_IDEFAULTANSICODEPAGE, code_page_string, 7) == 0)
return 0;
return static_cast<UINT>(_wtol(code_page_string));
}
#endif
template <class U>
inline void w32_regex_traits_char_layer<char>::init()
{
// we need to start by initialising our syntax map so we know which
// character is used for which purpose:
std::memset(m_char_map, 0, sizeof(m_char_map));
cat_type cat;
std::string cat_name(w32_regex_traits<char>::get_catalog_name());
if (cat_name.size())
{
cat = ::boost::BOOST_REGEX_DETAIL_NS::w32_cat_open(cat_name);
if (!cat)
{
std::string m("Unable to open message catalog: ");
std::runtime_error err(m + cat_name);
::boost::BOOST_REGEX_DETAIL_NS::raise_runtime_error(err);
}
}
//
// if we have a valid catalog then load our messages:
//
if (cat)
{
for (regex_constants::syntax_type i = 1; i < regex_constants::syntax_max; ++i)
{
string_type mss = ::boost::BOOST_REGEX_DETAIL_NS::w32_cat_get(cat, this->m_locale, i, get_default_syntax(i));
for (string_type::size_type j = 0; j < mss.size(); ++j)
{
m_char_map[static_cast<unsigned char>(mss[j])] = i;
}
}
}
else
{
for (regex_constants::syntax_type i = 1; i < regex_constants::syntax_max; ++i)
{
const char* ptr = get_default_syntax(i);
while (ptr && *ptr)
{
m_char_map[static_cast<unsigned char>(*ptr)] = i;
++ptr;
}
}
}
//
// finish off by calculating our escape types:
//
unsigned char i = 'A';
do
{
if (m_char_map[i] == 0)
{
if (::boost::BOOST_REGEX_DETAIL_NS::w32_is(this->m_locale, 0x0002u, (char)i))
m_char_map[i] = regex_constants::escape_type_class;
else if (::boost::BOOST_REGEX_DETAIL_NS::w32_is(this->m_locale, 0x0001u, (char)i))
m_char_map[i] = regex_constants::escape_type_not_class;
}
} while (0xFF != i++);
//
// fill in lower case map:
//
char char_map[1 << CHAR_BIT];
for (int ii = 0; ii < (1 << CHAR_BIT); ++ii)
char_map[ii] = static_cast<char>(ii);
#ifndef BOOST_NO_ANSI_APIS
int r = ::LCMapStringA(this->m_locale, LCMAP_LOWERCASE, char_map, 1 << CHAR_BIT, this->m_lower_map, 1 << CHAR_BIT);
BOOST_ASSERT(r != 0);
#else
UINT code_page = get_code_page_for_locale_id(this->m_locale);
BOOST_ASSERT(code_page != 0);
WCHAR wide_char_map[1 << CHAR_BIT];
int conv_r = ::MultiByteToWideChar(code_page, 0, char_map, 1 << CHAR_BIT, wide_char_map, 1 << CHAR_BIT);
BOOST_ASSERT(conv_r != 0);
WCHAR wide_lower_map[1 << CHAR_BIT];
int r = ::LCMapStringW(this->m_locale, LCMAP_LOWERCASE, wide_char_map, 1 << CHAR_BIT, wide_lower_map, 1 << CHAR_BIT);
BOOST_ASSERT(r != 0);
conv_r = ::WideCharToMultiByte(code_page, 0, wide_lower_map, r, this->m_lower_map, 1 << CHAR_BIT, NULL, NULL);
BOOST_ASSERT(conv_r != 0);
#endif
if (r < (1 << CHAR_BIT))
{
// if we have multibyte characters then not all may have been given
// a lower case mapping:
for (int jj = r; jj < (1 << CHAR_BIT); ++jj)
this->m_lower_map[jj] = static_cast<char>(jj);
}
#ifndef BOOST_NO_ANSI_APIS
r = ::GetStringTypeExA(this->m_locale, CT_CTYPE1, char_map, 1 << CHAR_BIT, this->m_type_map);
#else
r = ::GetStringTypeExW(this->m_locale, CT_CTYPE1, wide_char_map, 1 << CHAR_BIT, this->m_type_map);
#endif
BOOST_ASSERT(0 != r);
}
inline lcid_type BOOST_REGEX_CALL w32_get_default_locale()
{
return ::GetUserDefaultLCID();
}
inline bool BOOST_REGEX_CALL w32_is_lower(char c, lcid_type idx)
{
#ifndef BOOST_NO_ANSI_APIS
WORD mask;
if (::GetStringTypeExA(idx, CT_CTYPE1, &c, 1, &mask) && (mask & C1_LOWER))
return true;
return false;
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return false;
WCHAR wide_c;
if (::MultiByteToWideChar(code_page, 0, &c, 1, &wide_c, 1) == 0)
return false;
WORD mask;
if (::GetStringTypeExW(idx, CT_CTYPE1, &wide_c, 1, &mask) && (mask & C1_LOWER))
return true;
return false;
#endif
}
inline bool BOOST_REGEX_CALL w32_is_lower(wchar_t c, lcid_type idx)
{
WORD mask;
if (::GetStringTypeExW(idx, CT_CTYPE1, &c, 1, &mask) && (mask & C1_LOWER))
return true;
return false;
}
inline bool BOOST_REGEX_CALL w32_is_upper(char c, lcid_type idx)
{
#ifndef BOOST_NO_ANSI_APIS
WORD mask;
if (::GetStringTypeExA(idx, CT_CTYPE1, &c, 1, &mask) && (mask & C1_UPPER))
return true;
return false;
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return false;
WCHAR wide_c;
if (::MultiByteToWideChar(code_page, 0, &c, 1, &wide_c, 1) == 0)
return false;
WORD mask;
if (::GetStringTypeExW(idx, CT_CTYPE1, &wide_c, 1, &mask) && (mask & C1_UPPER))
return true;
return false;
#endif
}
inline bool BOOST_REGEX_CALL w32_is_upper(wchar_t c, lcid_type idx)
{
WORD mask;
if (::GetStringTypeExW(idx, CT_CTYPE1, &c, 1, &mask) && (mask & C1_UPPER))
return true;
return false;
}
inline void free_module(void* mod)
{
::FreeLibrary(static_cast<HMODULE>(mod));
}
inline cat_type BOOST_REGEX_CALL w32_cat_open(const std::string& name)
{
#ifndef BOOST_NO_ANSI_APIS
cat_type result(::LoadLibraryA(name.c_str()), &free_module);
return result;
#else
LPWSTR wide_name = (LPWSTR)_alloca((name.size() + 1) * sizeof(WCHAR));
if (::MultiByteToWideChar(CP_ACP, 0, name.c_str(), name.size(), wide_name, name.size() + 1) == 0)
return cat_type();
cat_type result(::LoadLibraryW(wide_name), &free_module);
return result;
#endif
}
inline std::string BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type, int i, const std::string& def)
{
#ifndef BOOST_NO_ANSI_APIS
char buf[256];
if (0 == ::LoadStringA(
static_cast<HMODULE>(cat.get()),
i,
buf,
256
))
{
return def;
}
#else
WCHAR wbuf[256];
int r = ::LoadStringW(
static_cast<HMODULE>(cat.get()),
i,
wbuf,
256
);
if (r == 0)
return def;
int buf_size = 1 + ::WideCharToMultiByte(CP_ACP, 0, wbuf, r, NULL, 0, NULL, NULL);
LPSTR buf = (LPSTR)_alloca(buf_size);
if (::WideCharToMultiByte(CP_ACP, 0, wbuf, r, buf, buf_size, NULL, NULL) == 0)
return def; // failed conversion.
#endif
return std::string(buf);
}
#ifndef BOOST_NO_WREGEX
inline std::wstring BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type, int i, const std::wstring& def)
{
wchar_t buf[256];
if (0 == ::LoadStringW(
static_cast<HMODULE>(cat.get()),
i,
buf,
256
))
{
return def;
}
return std::wstring(buf);
}
#endif
inline std::string BOOST_REGEX_CALL w32_transform(lcid_type idx, const char* p1, const char* p2)
{
#ifndef BOOST_NO_ANSI_APIS
int bytes = ::LCMapStringA(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
p1, // source string
static_cast<int>(p2 - p1), // number of characters in source string
0, // destination buffer
0 // size of destination buffer
);
if (!bytes)
return std::string(p1, p2);
std::string result(++bytes, '\0');
bytes = ::LCMapStringA(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
p1, // source string
static_cast<int>(p2 - p1), // number of characters in source string
&*result.begin(), // destination buffer
bytes // size of destination buffer
);
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return std::string(p1, p2);
int src_len = static_cast<int>(p2 - p1);
LPWSTR wide_p1 = (LPWSTR)_alloca((src_len + 1) * 2);
if (::MultiByteToWideChar(code_page, 0, p1, src_len, wide_p1, src_len + 1) == 0)
return std::string(p1, p2);
int bytes = ::LCMapStringW(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
wide_p1, // source string
src_len, // number of characters in source string
0, // destination buffer
0 // size of destination buffer
);
if (!bytes)
return std::string(p1, p2);
std::string result(++bytes, '\0');
bytes = ::LCMapStringW(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
wide_p1, // source string
src_len, // number of characters in source string
(LPWSTR) & *result.begin(), // destination buffer
bytes // size of destination buffer
);
#endif
if (bytes > static_cast<int>(result.size()))
return std::string(p1, p2);
while (result.size() && result[result.size() - 1] == '\0')
{
result.erase(result.size() - 1);
}
return result;
}
#ifndef BOOST_NO_WREGEX
inline std::wstring BOOST_REGEX_CALL w32_transform(lcid_type idx, const wchar_t* p1, const wchar_t* p2)
{
int bytes = ::LCMapStringW(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
p1, // source string
static_cast<int>(p2 - p1), // number of characters in source string
0, // destination buffer
0 // size of destination buffer
);
if (!bytes)
return std::wstring(p1, p2);
std::string result(++bytes, '\0');
bytes = ::LCMapStringW(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
p1, // source string
static_cast<int>(p2 - p1), // number of characters in source string
reinterpret_cast<wchar_t*>(&*result.begin()), // destination buffer *of bytes*
bytes // size of destination buffer
);
if (bytes > static_cast<int>(result.size()))
return std::wstring(p1, p2);
while (result.size() && result[result.size() - 1] == L'\0')
{
result.erase(result.size() - 1);
}
std::wstring r2;
for (std::string::size_type i = 0; i < result.size(); ++i)
r2.append(1, static_cast<wchar_t>(static_cast<unsigned char>(result[i])));
return r2;
}
#endif
inline char BOOST_REGEX_CALL w32_tolower(char c, lcid_type idx)
{
char result[2];
#ifndef BOOST_NO_ANSI_APIS
int b = ::LCMapStringA(
idx, // locale identifier
LCMAP_LOWERCASE, // mapping transformation type
&c, // source string
1, // number of characters in source string
result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return c;
WCHAR wide_c;
if (::MultiByteToWideChar(code_page, 0, &c, 1, &wide_c, 1) == 0)
return c;
WCHAR wide_result;
int b = ::LCMapStringW(
idx, // locale identifier
LCMAP_LOWERCASE, // mapping transformation type
&wide_c, // source string
1, // number of characters in source string
&wide_result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
if (::WideCharToMultiByte(code_page, 0, &wide_result, 1, result, 2, NULL, NULL) == 0)
return c; // No single byte lower case equivalent available
#endif
return result[0];
}
#ifndef BOOST_NO_WREGEX
inline wchar_t BOOST_REGEX_CALL w32_tolower(wchar_t c, lcid_type idx)
{
wchar_t result[2];
int b = ::LCMapStringW(
idx, // locale identifier
LCMAP_LOWERCASE, // mapping transformation type
&c, // source string
1, // number of characters in source string
result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
return result[0];
}
#endif
inline char BOOST_REGEX_CALL w32_toupper(char c, lcid_type idx)
{
char result[2];
#ifndef BOOST_NO_ANSI_APIS
int b = ::LCMapStringA(
idx, // locale identifier
LCMAP_UPPERCASE, // mapping transformation type
&c, // source string
1, // number of characters in source string
result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return c;
WCHAR wide_c;
if (::MultiByteToWideChar(code_page, 0, &c, 1, &wide_c, 1) == 0)
return c;
WCHAR wide_result;
int b = ::LCMapStringW(
idx, // locale identifier
LCMAP_UPPERCASE, // mapping transformation type
&wide_c, // source string
1, // number of characters in source string
&wide_result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
if (::WideCharToMultiByte(code_page, 0, &wide_result, 1, result, 2, NULL, NULL) == 0)
return c; // No single byte upper case equivalent available.
#endif
return result[0];
}
#ifndef BOOST_NO_WREGEX
inline wchar_t BOOST_REGEX_CALL w32_toupper(wchar_t c, lcid_type idx)
{
wchar_t result[2];
int b = ::LCMapStringW(
idx, // locale identifier
LCMAP_UPPERCASE, // mapping transformation type
&c, // source string
1, // number of characters in source string
result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
return result[0];
}
#endif
inline bool BOOST_REGEX_CALL w32_is(lcid_type idx, boost::uint32_t m, char c)
{
WORD mask;
#ifndef BOOST_NO_ANSI_APIS
if (::GetStringTypeExA(idx, CT_CTYPE1, &c, 1, &mask) && (mask & m & w32_regex_traits_implementation<char>::mask_base))
return true;
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return false;
WCHAR wide_c;
if (::MultiByteToWideChar(code_page, 0, &c, 1, &wide_c, 1) == 0)
return false;
if (::GetStringTypeExW(idx, CT_CTYPE1, &wide_c, 1, &mask) && (mask & m & w32_regex_traits_implementation<char>::mask_base))
return true;
#endif
if ((m & w32_regex_traits_implementation<char>::mask_word) && (c == '_'))
return true;
return false;
}
#ifndef BOOST_NO_WREGEX
inline bool BOOST_REGEX_CALL w32_is(lcid_type idx, boost::uint32_t m, wchar_t c)
{
WORD mask;
if (::GetStringTypeExW(idx, CT_CTYPE1, &c, 1, &mask) && (mask & m & w32_regex_traits_implementation<wchar_t>::mask_base))
return true;
if ((m & w32_regex_traits_implementation<wchar_t>::mask_word) && (c == '_'))
return true;
if ((m & w32_regex_traits_implementation<wchar_t>::mask_unicode) && (c > 0xff))
return true;
return false;
}
#endif
} // BOOST_REGEX_DETAIL_NS
} // boost

553
include/boost/regex/v5/w32_regex_traits.hpp
View File

@ -37,6 +37,12 @@
#include <boost/regex/pending/object_cache.hpp>
#endif
#include <windows.h>
#if defined(_MSC_VER) && !defined(_WIN32_WCE) && !defined(UNDER_CE)
#pragma comment(lib, "user32.lib")
#endif
#ifdef BOOST_MSVC
#pragma warning(push)
#pragma warning(disable: 4103)
@ -75,53 +81,35 @@ typedef ::boost::shared_ptr<void> cat_type; // placeholder for dll HANDLE.
//
// then add wrappers around the actual Win32 API's (ie implementation hiding):
//
BOOST_REGEX_DECL lcid_type BOOST_REGEX_CALL w32_get_default_locale();
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_lower(char, lcid_type);
lcid_type BOOST_REGEX_CALL w32_get_default_locale();
bool BOOST_REGEX_CALL w32_is_lower(char, lcid_type);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_lower(wchar_t, lcid_type);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_lower(unsigned short ca, lcid_type state_id);
bool BOOST_REGEX_CALL w32_is_lower(wchar_t, lcid_type);
#endif
#endif
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_upper(char, lcid_type);
bool BOOST_REGEX_CALL w32_is_upper(char, lcid_type);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_upper(wchar_t, lcid_type);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is_upper(unsigned short ca, lcid_type state_id);
bool BOOST_REGEX_CALL w32_is_upper(wchar_t, lcid_type);
#endif
#endif
BOOST_REGEX_DECL cat_type BOOST_REGEX_CALL w32_cat_open(const std::string& name);
BOOST_REGEX_DECL std::string BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type state_id, int i, const std::string& def);
cat_type BOOST_REGEX_CALL w32_cat_open(const std::string& name);
std::string BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type state_id, int i, const std::string& def);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL std::wstring BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type state_id, int i, const std::wstring& def);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL std::basic_string<unsigned short> BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type, int i, const std::basic_string<unsigned short>& def);
std::wstring BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type state_id, int i, const std::wstring& def);
#endif
#endif
BOOST_REGEX_DECL std::string BOOST_REGEX_CALL w32_transform(lcid_type state_id, const char* p1, const char* p2);
std::string BOOST_REGEX_CALL w32_transform(lcid_type state_id, const char* p1, const char* p2);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL std::wstring BOOST_REGEX_CALL w32_transform(lcid_type state_id, const wchar_t* p1, const wchar_t* p2);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL std::basic_string<unsigned short> BOOST_REGEX_CALL w32_transform(lcid_type state_id, const unsigned short* p1, const unsigned short* p2);
std::wstring BOOST_REGEX_CALL w32_transform(lcid_type state_id, const wchar_t* p1, const wchar_t* p2);
#endif
#endif
BOOST_REGEX_DECL char BOOST_REGEX_CALL w32_tolower(char c, lcid_type);
char BOOST_REGEX_CALL w32_tolower(char c, lcid_type);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL wchar_t BOOST_REGEX_CALL w32_tolower(wchar_t c, lcid_type);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL unsigned short BOOST_REGEX_CALL w32_tolower(unsigned short c, lcid_type state_id);
wchar_t BOOST_REGEX_CALL w32_tolower(wchar_t c, lcid_type);
#endif
#endif
BOOST_REGEX_DECL char BOOST_REGEX_CALL w32_toupper(char c, lcid_type);
char BOOST_REGEX_CALL w32_toupper(char c, lcid_type);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL wchar_t BOOST_REGEX_CALL w32_toupper(wchar_t c, lcid_type);
wchar_t BOOST_REGEX_CALL w32_toupper(wchar_t c, lcid_type);
#endif
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is(lcid_type, boost::uint32_t mask, char c);
bool BOOST_REGEX_CALL w32_is(lcid_type, boost::uint32_t mask, char c);
#ifndef BOOST_NO_WREGEX
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is(lcid_type, boost::uint32_t mask, wchar_t c);
#ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T
BOOST_REGEX_DECL bool BOOST_REGEX_CALL w32_is(lcid_type state_id, boost::uint32_t m, unsigned short c);
#endif
bool BOOST_REGEX_CALL w32_is(lcid_type, boost::uint32_t mask, wchar_t c);
#endif
//
// class w32_regex_traits_base:
@ -253,14 +241,14 @@ typename w32_regex_traits_char_layer<charT>::string_type
// specialised version for narrow characters:
//
template <>
class BOOST_REGEX_DECL w32_regex_traits_char_layer<char> : public w32_regex_traits_base<char>
class w32_regex_traits_char_layer<char> : public w32_regex_traits_base<char>
{
typedef std::string string_type;
public:
w32_regex_traits_char_layer(::boost::BOOST_REGEX_DETAIL_NS::lcid_type l)
: w32_regex_traits_base<char>(l)
{
init();
init<char>();
}
regex_constants::syntax_type syntax_type(char c)const
@ -284,6 +272,7 @@ private:
regex_constants::syntax_type m_char_map[1u << CHAR_BIT];
char m_lower_map[1u << CHAR_BIT];
boost::uint16_t m_type_map[1u << CHAR_BIT];
template <class U>
void init();
};
@ -722,6 +711,498 @@ static_mutex& w32_regex_traits<charT>::get_mutex_inst()
}
#endif
namespace BOOST_REGEX_DETAIL_NS {
#ifdef BOOST_NO_ANSI_APIS
inline UINT get_code_page_for_locale_id(lcid_type idx)
{
WCHAR code_page_string[7];
if (::GetLocaleInfoW(idx, LOCALE_IDEFAULTANSICODEPAGE, code_page_string, 7) == 0)
return 0;
return static_cast<UINT>(_wtol(code_page_string));
}
#endif
template <class U>
inline void w32_regex_traits_char_layer<char>::init()
{
// we need to start by initialising our syntax map so we know which
// character is used for which purpose:
std::memset(m_char_map, 0, sizeof(m_char_map));
cat_type cat;
std::string cat_name(w32_regex_traits<char>::get_catalog_name());
if (cat_name.size())
{
cat = ::boost::BOOST_REGEX_DETAIL_NS::w32_cat_open(cat_name);
if (!cat)
{
std::string m("Unable to open message catalog: ");
std::runtime_error err(m + cat_name);
::boost::BOOST_REGEX_DETAIL_NS::raise_runtime_error(err);
}
}
//
// if we have a valid catalog then load our messages:
//
if (cat)
{
for (regex_constants::syntax_type i = 1; i < regex_constants::syntax_max; ++i)
{
string_type mss = ::boost::BOOST_REGEX_DETAIL_NS::w32_cat_get(cat, this->m_locale, i, get_default_syntax(i));
for (string_type::size_type j = 0; j < mss.size(); ++j)
{
m_char_map[static_cast<unsigned char>(mss[j])] = i;
}
}
}
else
{
for (regex_constants::syntax_type i = 1; i < regex_constants::syntax_max; ++i)
{
const char* ptr = get_default_syntax(i);
while (ptr && *ptr)
{
m_char_map[static_cast<unsigned char>(*ptr)] = i;
++ptr;
}
}
}
//
// finish off by calculating our escape types:
//
unsigned char i = 'A';
do
{
if (m_char_map[i] == 0)
{
if (::boost::BOOST_REGEX_DETAIL_NS::w32_is(this->m_locale, 0x0002u, (char)i))
m_char_map[i] = regex_constants::escape_type_class;
else if (::boost::BOOST_REGEX_DETAIL_NS::w32_is(this->m_locale, 0x0001u, (char)i))
m_char_map[i] = regex_constants::escape_type_not_class;
}
} while (0xFF != i++);
//
// fill in lower case map:
//
char char_map[1 << CHAR_BIT];
for (int ii = 0; ii < (1 << CHAR_BIT); ++ii)
char_map[ii] = static_cast<char>(ii);
#ifndef BOOST_NO_ANSI_APIS
int r = ::LCMapStringA(this->m_locale, LCMAP_LOWERCASE, char_map, 1 << CHAR_BIT, this->m_lower_map, 1 << CHAR_BIT);
BOOST_ASSERT(r != 0);
#else
UINT code_page = get_code_page_for_locale_id(this->m_locale);
BOOST_ASSERT(code_page != 0);
WCHAR wide_char_map[1 << CHAR_BIT];
int conv_r = ::MultiByteToWideChar(code_page, 0, char_map, 1 << CHAR_BIT, wide_char_map, 1 << CHAR_BIT);
BOOST_ASSERT(conv_r != 0);
WCHAR wide_lower_map[1 << CHAR_BIT];
int r = ::LCMapStringW(this->m_locale, LCMAP_LOWERCASE, wide_char_map, 1 << CHAR_BIT, wide_lower_map, 1 << CHAR_BIT);
BOOST_ASSERT(r != 0);
conv_r = ::WideCharToMultiByte(code_page, 0, wide_lower_map, r, this->m_lower_map, 1 << CHAR_BIT, NULL, NULL);
BOOST_ASSERT(conv_r != 0);
#endif
if (r < (1 << CHAR_BIT))
{
// if we have multibyte characters then not all may have been given
// a lower case mapping:
for (int jj = r; jj < (1 << CHAR_BIT); ++jj)
this->m_lower_map[jj] = static_cast<char>(jj);
}
#ifndef BOOST_NO_ANSI_APIS
r = ::GetStringTypeExA(this->m_locale, CT_CTYPE1, char_map, 1 << CHAR_BIT, this->m_type_map);
#else
r = ::GetStringTypeExW(this->m_locale, CT_CTYPE1, wide_char_map, 1 << CHAR_BIT, this->m_type_map);
#endif
BOOST_ASSERT(0 != r);
}
inline lcid_type BOOST_REGEX_CALL w32_get_default_locale()
{
return ::GetUserDefaultLCID();
}
inline bool BOOST_REGEX_CALL w32_is_lower(char c, lcid_type idx)
{
#ifndef BOOST_NO_ANSI_APIS
WORD mask;
if (::GetStringTypeExA(idx, CT_CTYPE1, &c, 1, &mask) && (mask & C1_LOWER))
return true;
return false;
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return false;
WCHAR wide_c;
if (::MultiByteToWideChar(code_page, 0, &c, 1, &wide_c, 1) == 0)
return false;
WORD mask;
if (::GetStringTypeExW(idx, CT_CTYPE1, &wide_c, 1, &mask) && (mask & C1_LOWER))
return true;
return false;
#endif
}
inline bool BOOST_REGEX_CALL w32_is_lower(wchar_t c, lcid_type idx)
{
WORD mask;
if (::GetStringTypeExW(idx, CT_CTYPE1, &c, 1, &mask) && (mask & C1_LOWER))
return true;
return false;
}
inline bool BOOST_REGEX_CALL w32_is_upper(char c, lcid_type idx)
{
#ifndef BOOST_NO_ANSI_APIS
WORD mask;
if (::GetStringTypeExA(idx, CT_CTYPE1, &c, 1, &mask) && (mask & C1_UPPER))
return true;
return false;
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return false;
WCHAR wide_c;
if (::MultiByteToWideChar(code_page, 0, &c, 1, &wide_c, 1) == 0)
return false;
WORD mask;
if (::GetStringTypeExW(idx, CT_CTYPE1, &wide_c, 1, &mask) && (mask & C1_UPPER))
return true;
return false;
#endif
}
inline bool BOOST_REGEX_CALL w32_is_upper(wchar_t c, lcid_type idx)
{
WORD mask;
if (::GetStringTypeExW(idx, CT_CTYPE1, &c, 1, &mask) && (mask & C1_UPPER))
return true;
return false;
}
inline void free_module(void* mod)
{
::FreeLibrary(static_cast<HMODULE>(mod));
}
inline cat_type BOOST_REGEX_CALL w32_cat_open(const std::string& name)
{
#ifndef BOOST_NO_ANSI_APIS
cat_type result(::LoadLibraryA(name.c_str()), &free_module);
return result;
#else
LPWSTR wide_name = (LPWSTR)_alloca((name.size() + 1) * sizeof(WCHAR));
if (::MultiByteToWideChar(CP_ACP, 0, name.c_str(), name.size(), wide_name, name.size() + 1) == 0)
return cat_type();
cat_type result(::LoadLibraryW(wide_name), &free_module);
return result;
#endif
}
inline std::string BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type, int i, const std::string& def)
{
#ifndef BOOST_NO_ANSI_APIS
char buf[256];
if (0 == ::LoadStringA(
static_cast<HMODULE>(cat.get()),
i,
buf,
256
))
{
return def;
}
#else
WCHAR wbuf[256];
int r = ::LoadStringW(
static_cast<HMODULE>(cat.get()),
i,
wbuf,
256
);
if (r == 0)
return def;
int buf_size = 1 + ::WideCharToMultiByte(CP_ACP, 0, wbuf, r, NULL, 0, NULL, NULL);
LPSTR buf = (LPSTR)_alloca(buf_size);
if (::WideCharToMultiByte(CP_ACP, 0, wbuf, r, buf, buf_size, NULL, NULL) == 0)
return def; // failed conversion.
#endif
return std::string(buf);
}
#ifndef BOOST_NO_WREGEX
inline std::wstring BOOST_REGEX_CALL w32_cat_get(const cat_type& cat, lcid_type, int i, const std::wstring& def)
{
wchar_t buf[256];
if (0 == ::LoadStringW(
static_cast<HMODULE>(cat.get()),
i,
buf,
256
))
{
return def;
}
return std::wstring(buf);
}
#endif
inline std::string BOOST_REGEX_CALL w32_transform(lcid_type idx, const char* p1, const char* p2)
{
#ifndef BOOST_NO_ANSI_APIS
int bytes = ::LCMapStringA(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
p1, // source string
static_cast<int>(p2 - p1), // number of characters in source string
0, // destination buffer
0 // size of destination buffer
);
if (!bytes)
return std::string(p1, p2);
std::string result(++bytes, '\0');
bytes = ::LCMapStringA(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
p1, // source string
static_cast<int>(p2 - p1), // number of characters in source string
&*result.begin(), // destination buffer
bytes // size of destination buffer
);
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return std::string(p1, p2);
int src_len = static_cast<int>(p2 - p1);
LPWSTR wide_p1 = (LPWSTR)_alloca((src_len + 1) * 2);
if (::MultiByteToWideChar(code_page, 0, p1, src_len, wide_p1, src_len + 1) == 0)
return std::string(p1, p2);
int bytes = ::LCMapStringW(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
wide_p1, // source string
src_len, // number of characters in source string
0, // destination buffer
0 // size of destination buffer
);
if (!bytes)
return std::string(p1, p2);
std::string result(++bytes, '\0');
bytes = ::LCMapStringW(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
wide_p1, // source string
src_len, // number of characters in source string
(LPWSTR) & *result.begin(), // destination buffer
bytes // size of destination buffer
);
#endif
if (bytes > static_cast<int>(result.size()))
return std::string(p1, p2);
while (result.size() && result[result.size() - 1] == '\0')
{
result.erase(result.size() - 1);
}
return result;
}
#ifndef BOOST_NO_WREGEX
inline std::wstring BOOST_REGEX_CALL w32_transform(lcid_type idx, const wchar_t* p1, const wchar_t* p2)
{
int bytes = ::LCMapStringW(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
p1, // source string
static_cast<int>(p2 - p1), // number of characters in source string
0, // destination buffer
0 // size of destination buffer
);
if (!bytes)
return std::wstring(p1, p2);
std::string result(++bytes, '\0');
bytes = ::LCMapStringW(
idx, // locale identifier
LCMAP_SORTKEY, // mapping transformation type
p1, // source string
static_cast<int>(p2 - p1), // number of characters in source string
reinterpret_cast<wchar_t*>(&*result.begin()), // destination buffer *of bytes*
bytes // size of destination buffer
);
if (bytes > static_cast<int>(result.size()))
return std::wstring(p1, p2);
while (result.size() && result[result.size() - 1] == L'\0')
{
result.erase(result.size() - 1);
}
std::wstring r2;
for (std::string::size_type i = 0; i < result.size(); ++i)
r2.append(1, static_cast<wchar_t>(static_cast<unsigned char>(result[i])));
return r2;
}
#endif
inline char BOOST_REGEX_CALL w32_tolower(char c, lcid_type idx)
{
char result[2];
#ifndef BOOST_NO_ANSI_APIS
int b = ::LCMapStringA(
idx, // locale identifier
LCMAP_LOWERCASE, // mapping transformation type
&c, // source string
1, // number of characters in source string
result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return c;
WCHAR wide_c;
if (::MultiByteToWideChar(code_page, 0, &c, 1, &wide_c, 1) == 0)
return c;
WCHAR wide_result;
int b = ::LCMapStringW(
idx, // locale identifier
LCMAP_LOWERCASE, // mapping transformation type
&wide_c, // source string
1, // number of characters in source string
&wide_result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
if (::WideCharToMultiByte(code_page, 0, &wide_result, 1, result, 2, NULL, NULL) == 0)
return c; // No single byte lower case equivalent available
#endif
return result[0];
}
#ifndef BOOST_NO_WREGEX
inline wchar_t BOOST_REGEX_CALL w32_tolower(wchar_t c, lcid_type idx)
{
wchar_t result[2];
int b = ::LCMapStringW(
idx, // locale identifier
LCMAP_LOWERCASE, // mapping transformation type
&c, // source string
1, // number of characters in source string
result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
return result[0];
}
#endif
inline char BOOST_REGEX_CALL w32_toupper(char c, lcid_type idx)
{
char result[2];
#ifndef BOOST_NO_ANSI_APIS
int b = ::LCMapStringA(
idx, // locale identifier
LCMAP_UPPERCASE, // mapping transformation type
&c, // source string
1, // number of characters in source string
result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return c;
WCHAR wide_c;
if (::MultiByteToWideChar(code_page, 0, &c, 1, &wide_c, 1) == 0)
return c;
WCHAR wide_result;
int b = ::LCMapStringW(
idx, // locale identifier
LCMAP_UPPERCASE, // mapping transformation type
&wide_c, // source string
1, // number of characters in source string
&wide_result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
if (::WideCharToMultiByte(code_page, 0, &wide_result, 1, result, 2, NULL, NULL) == 0)
return c; // No single byte upper case equivalent available.
#endif
return result[0];
}
#ifndef BOOST_NO_WREGEX
inline wchar_t BOOST_REGEX_CALL w32_toupper(wchar_t c, lcid_type idx)
{
wchar_t result[2];
int b = ::LCMapStringW(
idx, // locale identifier
LCMAP_UPPERCASE, // mapping transformation type
&c, // source string
1, // number of characters in source string
result, // destination buffer
1); // size of destination buffer
if (b == 0)
return c;
return result[0];
}
#endif
inline bool BOOST_REGEX_CALL w32_is(lcid_type idx, boost::uint32_t m, char c)
{
WORD mask;
#ifndef BOOST_NO_ANSI_APIS
if (::GetStringTypeExA(idx, CT_CTYPE1, &c, 1, &mask) && (mask & m & w32_regex_traits_implementation<char>::mask_base))
return true;
#else
UINT code_page = get_code_page_for_locale_id(idx);
if (code_page == 0)
return false;
WCHAR wide_c;
if (::MultiByteToWideChar(code_page, 0, &c, 1, &wide_c, 1) == 0)
return false;
if (::GetStringTypeExW(idx, CT_CTYPE1, &wide_c, 1, &mask) && (mask & m & w32_regex_traits_implementation<char>::mask_base))
return true;
#endif
if ((m & w32_regex_traits_implementation<char>::mask_word) && (c == '_'))
return true;
return false;
}
#ifndef BOOST_NO_WREGEX
inline bool BOOST_REGEX_CALL w32_is(lcid_type idx, boost::uint32_t m, wchar_t c)
{
WORD mask;
if (::GetStringTypeExW(idx, CT_CTYPE1, &c, 1, &mask) && (mask & m & w32_regex_traits_implementation<wchar_t>::mask_base))
return true;
if ((m & w32_regex_traits_implementation<wchar_t>::mask_word) && (c == '_'))
return true;
if ((m & w32_regex_traits_implementation<wchar_t>::mask_unicode) && (c > 0xff))
return true;
return false;
}
#endif
} // BOOST_REGEX_DETAIL_NS
} // boost

1
test/Jamfile.v2
View File

@ -68,7 +68,6 @@ lib boost_regex_recursive :
../src/regex_raw_buffer.cpp
../src/regex_traits_defaults.cpp
../src/static_mutex.cpp
../src/w32_regex_traits.cpp
../src/wide_posix_api.cpp
../build//icu_options
:

1
test/captures/Jamfile.v2
View File

@ -15,7 +15,6 @@ EX_SOURCES =
regex_raw_buffer.cpp
regex_traits_defaults.cpp
static_mutex.cpp
w32_regex_traits.cpp
wide_posix_api.cpp ;
lib boost_regex_extra : $(EX_SOURCES) ../../build//icu_options

1
test/noeh_test/Jamfile.v2
View File

@ -31,7 +31,6 @@ lib boost_regex_noeh :
../../src/regex_raw_buffer.cpp
../../src/regex_traits_defaults.cpp
../../src/static_mutex.cpp
../../src/w32_regex_traits.cpp
../../src/wide_posix_api.cpp
../../build//icu_options
:

1
test/test_consolidated.cpp
View File

@ -17,5 +17,4 @@
#include <libs/regex/src/regex_raw_buffer.cpp>
#include <libs/regex/src/regex_traits_defaults.cpp>
#include <libs/regex/src/static_mutex.cpp>
#include <libs/regex/src/w32_regex_traits.cpp>
#include <libs/regex/src/wide_posix_api.cpp>