contrib/libc++/src/regex.cpp

   1 //===-------------------------- regex.cpp ---------------------------------===//
   2 //
   3 //                     The LLVM Compiler Infrastructure
   4 //
   5 // This file is dual licensed under the MIT and the University of Illinois Open
   6 // Source Licenses. See LICENSE.TXT for details.
   7 //
   8 //===----------------------------------------------------------------------===//
   9
  10 #include "regex"
  11 #include "algorithm"
  12 #include "iterator"
  13
  14 _LIBCPP_BEGIN_NAMESPACE_STD
  15
  16 static
  17 const char*
  18 make_error_type_string(regex_constants::error_type ecode)
  19 {
  20     switch (ecode)
  21     {
  22     case regex_constants::error_collate:
  23         return "The expression contained an invalid collating element name.";
  24     case regex_constants::error_ctype:
  25         return "The expression contained an invalid character class name.";
  26     case regex_constants::error_escape:
  27         return "The expression contained an invalid escaped character, or a "
  28                "trailing escape.";
  29     case regex_constants::error_backref:
  30         return "The expression contained an invalid back reference.";
  31     case regex_constants::error_brack:
  32         return "The expression contained mismatched [ and ].";
  33     case regex_constants::error_paren:
  34         return "The expression contained mismatched ( and ).";
  35     case regex_constants::error_brace:
  36         return "The expression contained mismatched { and }.";
  37     case regex_constants::error_badbrace:
  38         return "The expression contained an invalid range in a {} expression.";
  39     case regex_constants::error_range:
  40         return "The expression contained an invalid character range, "
  41                "such as [b-a] in most encodings.";
  42     case regex_constants::error_space:
  43         return "There was insufficient memory to convert the expression into "
  44                "a finite state machine.";
  45     case regex_constants::error_badrepeat:
  46         return "One of *?+{ was not preceded by a valid regular expression.";
  47     case regex_constants::error_complexity:
  48         return "The complexity of an attempted match against a regular "
  49                "expression exceeded a pre-set level.";
  50     case regex_constants::error_stack:
  51         return "There was insufficient memory to determine whether the regular "
  52                "expression could match the specified character sequence.";
  53     case regex_constants::__re_err_grammar:
  54         return "An invalid regex grammar has been requested.";
  55     case regex_constants::__re_err_empty:
  56         return "An empty regex is not allowed in the POSIX grammar.";
  57     default:
  58         break;
  59     }
  60     return "Unknown error type";
  61 }
  62
  63 regex_error::regex_error(regex_constants::error_type ecode)
  64     : runtime_error(make_error_type_string(ecode)),
  65       __code_(ecode)
  66 {}
  67
  68 regex_error::~regex_error() throw() {}
  69
  70 namespace {
  71
  72 struct collationnames
  73 {
  74     const char* elem_;
  75     char char_;
  76 };
  77
  78 const collationnames collatenames[] =
  79 {
  80     {"A", 0x41},
  81     {"B", 0x42},
  82     {"C", 0x43},
  83     {"D", 0x44},
  84     {"E", 0x45},
  85     {"F", 0x46},
  86     {"G", 0x47},
  87     {"H", 0x48},
  88     {"I", 0x49},
  89     {"J", 0x4a},
  90     {"K", 0x4b},
  91     {"L", 0x4c},
  92     {"M", 0x4d},
  93     {"N", 0x4e},
  94     {"NUL", 0x00},
  95     {"O", 0x4f},
  96     {"P", 0x50},
  97     {"Q", 0x51},
  98     {"R", 0x52},
  99     {"S", 0x53},
 100     {"T", 0x54},
 101     {"U", 0x55},
 102     {"V", 0x56},
 103     {"W", 0x57},
 104     {"X", 0x58},
 105     {"Y", 0x59},
 106     {"Z", 0x5a},
 107     {"a", 0x61},
 108     {"alert", 0x07},
 109     {"ampersand", 0x26},
 110     {"apostrophe", 0x27},
 111     {"asterisk", 0x2a},
 112     {"b", 0x62},
 113     {"backslash", 0x5c},
 114     {"backspace", 0x08},
 115     {"c", 0x63},
 116     {"carriage-return", 0x0d},
 117     {"circumflex", 0x5e},
 118     {"circumflex-accent", 0x5e},
 119     {"colon", 0x3a},
 120     {"comma", 0x2c},
 121     {"commercial-at", 0x40},
 122     {"d", 0x64},
 123     {"dollar-sign", 0x24},
 124     {"e", 0x65},
 125     {"eight", 0x38},
 126     {"equals-sign", 0x3d},
 127     {"exclamation-mark", 0x21},
 128     {"f", 0x66},
 129     {"five", 0x35},
 130     {"form-feed", 0x0c},
 131     {"four", 0x34},
 132     {"full-stop", 0x2e},
 133     {"g", 0x67},
 134     {"grave-accent", 0x60},
 135     {"greater-than-sign", 0x3e},
 136     {"h", 0x68},
 137     {"hyphen", 0x2d},
 138     {"hyphen-minus", 0x2d},
 139     {"i", 0x69},
 140     {"j", 0x6a},
 141     {"k", 0x6b},
 142     {"l", 0x6c},
 143     {"left-brace", 0x7b},
 144     {"left-curly-bracket", 0x7b},
 145     {"left-parenthesis", 0x28},
 146     {"left-square-bracket", 0x5b},
 147     {"less-than-sign", 0x3c},
 148     {"low-line", 0x5f},
 149     {"m", 0x6d},
 150     {"n", 0x6e},
 151     {"newline", 0x0a},
 152     {"nine", 0x39},
 153     {"number-sign", 0x23},
 154     {"o", 0x6f},
 155     {"one", 0x31},
 156     {"p", 0x70},
 157     {"percent-sign", 0x25},
 158     {"period", 0x2e},
 159     {"plus-sign", 0x2b},
 160     {"q", 0x71},
 161     {"question-mark", 0x3f},
 162     {"quotation-mark", 0x22},
 163     {"r", 0x72},
 164     {"reverse-solidus", 0x5c},
 165     {"right-brace", 0x7d},
 166     {"right-curly-bracket", 0x7d},
 167     {"right-parenthesis", 0x29},
 168     {"right-square-bracket", 0x5d},
 169     {"s", 0x73},
 170     {"semicolon", 0x3b},
 171     {"seven", 0x37},
 172     {"six", 0x36},
 173     {"slash", 0x2f},
 174     {"solidus", 0x2f},
 175     {"space", 0x20},
 176     {"t", 0x74},
 177     {"tab", 0x09},
 178     {"three", 0x33},
 179     {"tilde", 0x7e},
 180     {"two", 0x32},
 181     {"u", 0x75},
 182     {"underscore", 0x5f},
 183     {"v", 0x76},
 184     {"vertical-line", 0x7c},
 185     {"vertical-tab", 0x0b},
 186     {"w", 0x77},
 187     {"x", 0x78},
 188     {"y", 0x79},
 189     {"z", 0x7a},
 190     {"zero", 0x30}
 191 };
 192
 193 struct classnames
 194 {
 195     const char* elem_;
 196     regex_traits<char>::char_class_type mask_;
 197 };
 198
 199 const classnames ClassNames[] =
 200 {
 201     {"alnum",  ctype_base::alnum},
 202     {"alpha",  ctype_base::alpha},
 203     {"blank",  ctype_base::blank},
 204     {"cntrl",  ctype_base::cntrl},
 205     {"d",      ctype_base::digit},
 206     {"digit",  ctype_base::digit},
 207     {"graph",  ctype_base::graph},
 208     {"lower",  ctype_base::lower},
 209     {"print",  ctype_base::print},
 210     {"punct",  ctype_base::punct},
 211     {"s",      ctype_base::space},
 212     {"space",  ctype_base::space},
 213     {"upper",  ctype_base::upper},
 214     {"w",      regex_traits<char>::__regex_word},
 215     {"xdigit", ctype_base::xdigit}
 216 };
 217
 218 struct use_strcmp
 219 {
 220     bool operator()(const collationnames& x, const char* y)
 221         {return strcmp(x.elem_, y) < 0;}
 222     bool operator()(const classnames& x, const char* y)
 223         {return strcmp(x.elem_, y) < 0;}
 224 };
 225
 226 }
 227
 228 string
 229 __get_collation_name(const char* s)
 230 {
 231     const collationnames* i =
 232             _VSTD::lower_bound(begin(collatenames), end(collatenames), s, use_strcmp());
 233     string r;
 234     if (i != end(collatenames) && strcmp(s, i->elem_) == 0)
 235         r = char(i->char_);
 236     return r;
 237 }
 238
 239 regex_traits<char>::char_class_type
 240 __get_classname(const char* s, bool __icase)
 241 {
 242     const classnames* i =
 243             _VSTD::lower_bound(begin(ClassNames), end(ClassNames), s, use_strcmp());
 244     regex_traits<char>::char_class_type r = 0;
 245     if (i != end(ClassNames) && strcmp(s, i->elem_) == 0)
 246     {
 247         r = i->mask_;
 248         if (r == regex_traits<char>::__regex_word)
 249             r |= ctype_base::alnum | ctype_base::upper | ctype_base::lower;
 250         else if (__icase)
 251         {
 252             if (r & (ctype_base::lower | ctype_base::upper))
 253                 r |= ctype_base::alpha;
 254         }
 255     }
 256     return r;
 257 }
 258
 259 template <>
 260 void
 261 __match_any_but_newline<char>::__exec(__state& __s) const
 262 {
 263     if (__s.__current_ != __s.__last_)
 264     {
 265         switch (*__s.__current_)
 266         {
 267         case '\r':
 268         case '\n':
 269             __s.__do_ = __state::__reject;
 270             __s.__node_ = nullptr;
 271             break;
 272         default:
 273             __s.__do_ = __state::__accept_and_consume;
 274             ++__s.__current_;
 275             __s.__node_ = this->first();
 276             break;
 277         }
 278     }
 279     else
 280     {
 281         __s.__do_ = __state::__reject;
 282         __s.__node_ = nullptr;
 283     }
 284 }
 285
 286 template <>
 287 void
 288 __match_any_but_newline<wchar_t>::__exec(__state& __s) const
 289 {
 290     if (__s.__current_ != __s.__last_)
 291     {
 292         switch (*__s.__current_)
 293         {
 294         case '\r':
 295         case '\n':
 296         case 0x2028:
 297         case 0x2029:
 298             __s.__do_ = __state::__reject;
 299             __s.__node_ = nullptr;
 300             break;
 301         default:
 302             __s.__do_ = __state::__accept_and_consume;
 303             ++__s.__current_;
 304             __s.__node_ = this->first();
 305             break;
 306         }
 307     }
 308     else
 309     {
 310         __s.__do_ = __state::__reject;
 311         __s.__node_ = nullptr;
 312     }
 313 }
 314
 315 _LIBCPP_END_NAMESPACE_STD