| 1 | /* |
| 2 | * |
| 3 | * Copyright (c) 2004 |
| 4 | * John Maddock |
| 5 | * |
| 6 | * Use, modification and distribution are subject to the |
| 7 | * Boost Software License, Version 1.0. (See accompanying file |
| 8 | * LICENSE_1_0.txt or copy at http://www.boost.org/LICENSE_1_0.txt) |
| 9 | * |
| 10 | */ |
| 11 | |
| 12 | /* |
| 13 | * LOCATION: see http://www.boost.org for most recent version. |
| 14 | * FILE: wc_regex_traits.cpp |
| 15 | * VERSION: see <boost/version.hpp> |
| 16 | * DESCRIPTION: Implements out of line members for c_regex_traits<wchar_t> |
| 17 | */ |
| 18 | |
| 19 | |
| 20 | #define BOOST_REGEX_SOURCE |
| 21 | |
| 22 | #include <boost/detail/workaround.hpp> |
| 23 | #include <memory> |
| 24 | #include <string> |
| 25 | #include "internals.hpp" |
| 26 | |
| 27 | #if defined(_DLL_CPPLIB) && !defined(_M_CEE_PURE) && defined(_NATIVE_WCHAR_T_DEFINED) \ |
| 28 | && !(defined(__SGI_STL_PORT) || defined(_STLPORT_VERSION) || defined(__STD_RWCOMPILER_H__) || defined(_RWSTD_VER))\ |
| 29 | && BOOST_WORKAROUND(BOOST_MSVC, <1600) |
| 30 | // |
| 31 | // This is a horrible workaround, but without declaring these symbols extern we get |
| 32 | // duplicate symbol errors when linking if the application is built without |
| 33 | // /Zc:wchar_t |
| 34 | // |
| 35 | #ifdef _CRTIMP2_PURE |
| 36 | # define BOOST_REGEX_STDLIB_DECL _CRTIMP2_PURE |
| 37 | #else |
| 38 | # define BOOST_REGEX_STDLIB_DECL _CRTIMP2 |
| 39 | #endif |
| 40 | |
| 41 | namespace std{ |
| 42 | |
| 43 | #if BOOST_WORKAROUND(BOOST_MSVC, >= 1400) |
| 44 | template class BOOST_REGEX_STDLIB_DECL allocator<unsigned short>; |
| 45 | template class BOOST_REGEX_STDLIB_DECL _String_val<unsigned short, allocator<unsigned short> >; |
| 46 | template class BOOST_REGEX_STDLIB_DECL basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >; |
| 47 | #endif |
| 48 | |
| 49 | #if BOOST_WORKAROUND(BOOST_MSVC, > 1300) && BOOST_WORKAROUND(BOOST_MSVC, BOOST_TESTED_AT(1400)) |
| 50 | template<> BOOST_REGEX_STDLIB_DECL std::size_t __cdecl char_traits<unsigned short>::length(unsigned short const*); |
| 51 | #endif |
| 52 | |
| 53 | template BOOST_REGEX_STDLIB_DECL bool __cdecl operator==( |
| 54 | const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&, |
| 55 | const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&); |
| 56 | template BOOST_REGEX_STDLIB_DECL bool __cdecl operator==( |
| 57 | const unsigned short *, |
| 58 | const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&); |
| 59 | template BOOST_REGEX_STDLIB_DECL bool __cdecl operator==( |
| 60 | const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&, |
| 61 | const unsigned short *); |
| 62 | template BOOST_REGEX_STDLIB_DECL bool __cdecl operator<( |
| 63 | const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&, |
| 64 | const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&); |
| 65 | template BOOST_REGEX_STDLIB_DECL bool __cdecl operator>( |
| 66 | const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&, |
| 67 | const basic_string<unsigned short, char_traits<unsigned short>, allocator<unsigned short> >&); |
| 68 | } |
| 69 | #endif |
| 70 | |
| 71 | #include <boost/regex/config.hpp> |
| 72 | #include <boost/detail/workaround.hpp> |
| 73 | |
| 74 | #if !BOOST_WORKAROUND(__BORLANDC__, < 0x560) |
| 75 | |
| 76 | #include <boost/regex/v4/c_regex_traits.hpp> |
| 77 | #ifndef BOOST_NO_WREGEX |
| 78 | #include <boost/regex/v4/primary_transform.hpp> |
| 79 | #include <boost/regex/v4/regex_traits_defaults.hpp> |
| 80 | |
| 81 | #if defined(BOOST_NO_STDC_NAMESPACE) |
| 82 | namespace std{ |
| 83 | using ::wcstol; |
| 84 | } |
| 85 | #endif |
| 86 | |
| 87 | namespace boost{ |
| 88 | |
| 89 | c_regex_traits<wchar_t>::string_type BOOST_REGEX_CALL c_regex_traits<wchar_t>::transform(const wchar_t* p1, const wchar_t* p2) |
| 90 | { |
| 91 | std::size_t r; |
| 92 | std::size_t s = 10; |
| 93 | std::wstring src(p1, p2); |
| 94 | std::wstring result(s, L' '); |
| 95 | while(s < (r = std::wcsxfrm(&*result.begin(), src.c_str(), s))) |
| 96 | { |
| 97 | #if defined(_CPPLIB_VER) |
| 98 | // |
| 99 | // A bug in VC11 and 12 causes the program to hang if we pass a null-string |
| 100 | // to std::strxfrm, but only for certain locales :-( |
| 101 | // Probably effects Intel and Clang or any compiler using the VC std library (Dinkumware). |
| 102 | // |
| 103 | if(r == INT_MAX) |
| 104 | { |
| 105 | result.erase(); |
| 106 | result.insert(result.begin(), static_cast<wchar_t>(0)); |
| 107 | return result; |
| 108 | } |
| 109 | #endif |
| 110 | result.append(r - s + 3, L' '); |
| 111 | s = result.size(); |
| 112 | } |
| 113 | result.erase(r); |
| 114 | return result; |
| 115 | } |
| 116 | |
| 117 | c_regex_traits<wchar_t>::string_type BOOST_REGEX_CALL c_regex_traits<wchar_t>::transform_primary(const wchar_t* p1, const wchar_t* p2) |
| 118 | { |
| 119 | static wchar_t s_delim; |
| 120 | static const int s_collate_type = ::boost::BOOST_REGEX_DETAIL_NS::find_sort_syntax(static_cast<const c_regex_traits<wchar_t>*>(0), &s_delim); |
| 121 | std::wstring result; |
| 122 | // |
| 123 | // What we do here depends upon the format of the sort key returned by |
| 124 | // sort key returned by this->transform: |
| 125 | // |
| 126 | switch(s_collate_type) |
| 127 | { |
| 128 | case ::boost::BOOST_REGEX_DETAIL_NS::sort_C: |
| 129 | case ::boost::BOOST_REGEX_DETAIL_NS::sort_unknown: |
| 130 | // the best we can do is translate to lower case, then get a regular sort key: |
| 131 | { |
| 132 | result.assign(p1, p2); |
| 133 | for(std::wstring::size_type i = 0; i < result.size(); ++i) |
| 134 | result[i] = (std::towlower)(result[i]); |
| 135 | result = c_regex_traits<wchar_t>::transform(&*result.begin(), &*result.begin() + result.size()); |
| 136 | break; |
| 137 | } |
| 138 | case ::boost::BOOST_REGEX_DETAIL_NS::sort_fixed: |
| 139 | { |
| 140 | // get a regular sort key, and then truncate it: |
| 141 | result = c_regex_traits<wchar_t>::transform(&*result.begin(), &*result.begin() + result.size()); |
| 142 | result.erase(s_delim); |
| 143 | break; |
| 144 | } |
| 145 | case ::boost::BOOST_REGEX_DETAIL_NS::sort_delim: |
| 146 | // get a regular sort key, and then truncate everything after the delim: |
| 147 | result = c_regex_traits<wchar_t>::transform(&*result.begin(), &*result.begin() + result.size()); |
| 148 | if(result.size() && (result[0] == s_delim)) |
| 149 | break; |
| 150 | std::size_t i; |
| 151 | for(i = 0; i < result.size(); ++i) |
| 152 | { |
| 153 | if(result[i] == s_delim) |
| 154 | break; |
| 155 | } |
| 156 | result.erase(i); |
| 157 | break; |
| 158 | } |
| 159 | if(result.empty()) |
| 160 | result = std::wstring(1, char(0)); |
| 161 | return result; |
| 162 | } |
| 163 | |
| 164 | c_regex_traits<wchar_t>::char_class_type BOOST_REGEX_CALL c_regex_traits<wchar_t>::lookup_classname(const wchar_t* p1, const wchar_t* p2) |
| 165 | { |
| 166 | static const char_class_type masks[] = |
| 167 | { |
| 168 | 0, |
| 169 | char_class_alnum, |
| 170 | char_class_alpha, |
| 171 | char_class_blank, |
| 172 | char_class_cntrl, |
| 173 | char_class_digit, |
| 174 | char_class_digit, |
| 175 | char_class_graph, |
| 176 | char_class_horizontal, |
| 177 | char_class_lower, |
| 178 | char_class_lower, |
| 179 | char_class_print, |
| 180 | char_class_punct, |
| 181 | char_class_space, |
| 182 | char_class_space, |
| 183 | char_class_upper, |
| 184 | char_class_unicode, |
| 185 | char_class_upper, |
| 186 | char_class_vertical, |
| 187 | char_class_alnum | char_class_word, |
| 188 | char_class_alnum | char_class_word, |
| 189 | char_class_xdigit, |
| 190 | }; |
| 191 | |
| 192 | int idx = ::boost::BOOST_REGEX_DETAIL_NS::get_default_class_id(p1, p2); |
| 193 | if(idx < 0) |
| 194 | { |
| 195 | std::wstring s(p1, p2); |
| 196 | for(std::wstring::size_type i = 0; i < s.size(); ++i) |
| 197 | s[i] = (std::towlower)(s[i]); |
| 198 | idx = ::boost::BOOST_REGEX_DETAIL_NS::get_default_class_id(&*s.begin(), &*s.begin() + s.size()); |
| 199 | } |
| 200 | BOOST_ASSERT(idx+1 < static_cast<int>(sizeof(masks) / sizeof(masks[0]))); |
| 201 | return masks[idx+1]; |
| 202 | } |
| 203 | |
| 204 | bool BOOST_REGEX_CALL c_regex_traits<wchar_t>::isctype(wchar_t c, char_class_type mask) |
| 205 | { |
| 206 | return |
| 207 | ((mask & char_class_space) && (std::iswspace)(c)) |
| 208 | || ((mask & char_class_print) && (std::iswprint)(c)) |
| 209 | || ((mask & char_class_cntrl) && (std::iswcntrl)(c)) |
| 210 | || ((mask & char_class_upper) && (std::iswupper)(c)) |
| 211 | || ((mask & char_class_lower) && (std::iswlower)(c)) |
| 212 | || ((mask & char_class_alpha) && (std::iswalpha)(c)) |
| 213 | || ((mask & char_class_digit) && (std::iswdigit)(c)) |
| 214 | || ((mask & char_class_punct) && (std::iswpunct)(c)) |
| 215 | || ((mask & char_class_xdigit) && (std::iswxdigit)(c)) |
| 216 | || ((mask & char_class_blank) && (std::iswspace)(c) && !::boost::BOOST_REGEX_DETAIL_NS::is_separator(c)) |
| 217 | || ((mask & char_class_word) && (c == '_')) |
| 218 | || ((mask & char_class_unicode) && (c & ~static_cast<wchar_t>(0xff))) |
| 219 | || ((mask & char_class_vertical) && (::boost::BOOST_REGEX_DETAIL_NS::is_separator(c) || (c == L'\v'))) |
| 220 | || ((mask & char_class_horizontal) && (std::iswspace)(c) && !::boost::BOOST_REGEX_DETAIL_NS::is_separator(c) && (c != L'\v')); |
| 221 | } |
| 222 | |
| 223 | c_regex_traits<wchar_t>::string_type BOOST_REGEX_CALL c_regex_traits<wchar_t>::lookup_collatename(const wchar_t* p1, const wchar_t* p2) |
| 224 | { |
| 225 | #if !defined(BOOST_NO_TEMPLATED_ITERATOR_CONSTRUCTORS)\ |
| 226 | && !BOOST_WORKAROUND(BOOST_MSVC, < 1300)\ |
| 227 | && !BOOST_WORKAROUND(__BORLANDC__, <= 0x0551) |
| 228 | std::string name(p1, p2); |
| 229 | #else |
| 230 | std::string name; |
| 231 | const wchar_t* p0 = p1; |
| 232 | while(p0 != p2) |
| 233 | name.append(1, char(*p0++)); |
| 234 | #endif |
| 235 | name = ::boost::BOOST_REGEX_DETAIL_NS::lookup_default_collate_name(name); |
| 236 | #if !defined(BOOST_NO_TEMPLATED_ITERATOR_CONSTRUCTORS)\ |
| 237 | && !BOOST_WORKAROUND(BOOST_MSVC, < 1300)\ |
| 238 | && !BOOST_WORKAROUND(__BORLANDC__, <= 0x0551) |
| 239 | if(name.size()) |
| 240 | return string_type(name.begin(), name.end()); |
| 241 | #else |
| 242 | if(name.size()) |
| 243 | { |
| 244 | string_type result; |
| 245 | typedef std::string::const_iterator iter; |
| 246 | iter b = name.begin(); |
| 247 | iter e = name.end(); |
| 248 | while(b != e) |
| 249 | result.append(1, wchar_t(*b++)); |
| 250 | return result; |
| 251 | } |
| 252 | #endif |
| 253 | if(p2 - p1 == 1) |
| 254 | return string_type(1, *p1); |
| 255 | return string_type(); |
| 256 | } |
| 257 | |
| 258 | int BOOST_REGEX_CALL c_regex_traits<wchar_t>::value(wchar_t c, int radix) |
| 259 | { |
| 260 | #ifdef __BORLANDC__ |
| 261 | // workaround for broken wcstol: |
| 262 | if((std::iswxdigit)(c) == 0) |
| 263 | return -1; |
| 264 | #endif |
| 265 | wchar_t b[2] = { c, '\0', }; |
| 266 | wchar_t* ep; |
| 267 | int result = std::wcstol(b, &ep, radix); |
| 268 | if(ep == b) |
| 269 | return -1; |
| 270 | return result; |
| 271 | } |
| 272 | |
| 273 | #ifdef BOOST_REGEX_HAS_OTHER_WCHAR_T |
| 274 | c_regex_traits<unsigned short>::string_type BOOST_REGEX_CALL c_regex_traits<unsigned short>::transform(const unsigned short* p1, const unsigned short* p2) |
| 275 | { |
| 276 | std::wstring result = c_regex_traits<wchar_t>::transform((const wchar_t*)p1, (const wchar_t*)p2); |
| 277 | return string_type(result.begin(), result.end()); |
| 278 | } |
| 279 | |
| 280 | c_regex_traits<unsigned short>::string_type BOOST_REGEX_CALL c_regex_traits<unsigned short>::transform_primary(const unsigned short* p1, const unsigned short* p2) |
| 281 | { |
| 282 | std::wstring result = c_regex_traits<wchar_t>::transform_primary((const wchar_t*)p1, (const wchar_t*)p2); |
| 283 | return string_type(result.begin(), result.end()); |
| 284 | } |
| 285 | |
| 286 | c_regex_traits<unsigned short>::char_class_type BOOST_REGEX_CALL c_regex_traits<unsigned short>::lookup_classname(const unsigned short* p1, const unsigned short* p2) |
| 287 | { |
| 288 | return c_regex_traits<wchar_t>::lookup_classname((const wchar_t*)p1, (const wchar_t*)p2); |
| 289 | } |
| 290 | |
| 291 | c_regex_traits<unsigned short>::string_type BOOST_REGEX_CALL c_regex_traits<unsigned short>::lookup_collatename(const unsigned short* p1, const unsigned short* p2) |
| 292 | { |
| 293 | std::wstring result = c_regex_traits<wchar_t>::lookup_collatename((const wchar_t*)p1, (const wchar_t*)p2); |
| 294 | return string_type(result.begin(), result.end()); |
| 295 | } |
| 296 | |
| 297 | bool BOOST_REGEX_CALL c_regex_traits<unsigned short>::isctype(unsigned short c, char_class_type m) |
| 298 | { |
| 299 | return c_regex_traits<wchar_t>::isctype(c, m); |
| 300 | } |
| 301 | |
| 302 | int BOOST_REGEX_CALL c_regex_traits<unsigned short>::value(unsigned short c, int radix) |
| 303 | { |
| 304 | return c_regex_traits<wchar_t>::value(c, radix); |
| 305 | } |
| 306 | |
| 307 | #endif |
| 308 | |
| 309 | } |
| 310 | |
| 311 | #endif // BOOST_NO_WREGEX |
| 312 | |
| 313 | #endif // __BORLANDC__ |
| 314 | |
| 315 | |