Home | History | Annotate | Download | only in src
      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 #pragma clang diagnostic push
     73 #pragma clang diagnostic ignored "-Wpadded"
     74 
     75 struct collationnames
     76 {
     77     const char* elem_;
     78     char char_;
     79 };
     80 
     81 #pragma clang diagnostic pop
     82 
     83 const collationnames collatenames[] =
     84 {
     85     {"A", 0x41},
     86     {"B", 0x42},
     87     {"C", 0x43},
     88     {"D", 0x44},
     89     {"E", 0x45},
     90     {"F", 0x46},
     91     {"G", 0x47},
     92     {"H", 0x48},
     93     {"I", 0x49},
     94     {"J", 0x4a},
     95     {"K", 0x4b},
     96     {"L", 0x4c},
     97     {"M", 0x4d},
     98     {"N", 0x4e},
     99     {"NUL", 0x00},
    100     {"O", 0x4f},
    101     {"P", 0x50},
    102     {"Q", 0x51},
    103     {"R", 0x52},
    104     {"S", 0x53},
    105     {"T", 0x54},
    106     {"U", 0x55},
    107     {"V", 0x56},
    108     {"W", 0x57},
    109     {"X", 0x58},
    110     {"Y", 0x59},
    111     {"Z", 0x5a},
    112     {"a", 0x61},
    113     {"alert", 0x07},
    114     {"ampersand", 0x26},
    115     {"apostrophe", 0x27},
    116     {"asterisk", 0x2a},
    117     {"b", 0x62},
    118     {"backslash", 0x5c},
    119     {"backspace", 0x08},
    120     {"c", 0x63},
    121     {"carriage-return", 0x0d},
    122     {"circumflex", 0x5e},
    123     {"circumflex-accent", 0x5e},
    124     {"colon", 0x3a},
    125     {"comma", 0x2c},
    126     {"commercial-at", 0x40},
    127     {"d", 0x64},
    128     {"dollar-sign", 0x24},
    129     {"e", 0x65},
    130     {"eight", 0x38},
    131     {"equals-sign", 0x3d},
    132     {"exclamation-mark", 0x21},
    133     {"f", 0x66},
    134     {"five", 0x35},
    135     {"form-feed", 0x0c},
    136     {"four", 0x34},
    137     {"full-stop", 0x2e},
    138     {"g", 0x67},
    139     {"grave-accent", 0x60},
    140     {"greater-than-sign", 0x3e},
    141     {"h", 0x68},
    142     {"hyphen", 0x2d},
    143     {"hyphen-minus", 0x2d},
    144     {"i", 0x69},
    145     {"j", 0x6a},
    146     {"k", 0x6b},
    147     {"l", 0x6c},
    148     {"left-brace", 0x7b},
    149     {"left-curly-bracket", 0x7b},
    150     {"left-parenthesis", 0x28},
    151     {"left-square-bracket", 0x5b},
    152     {"less-than-sign", 0x3c},
    153     {"low-line", 0x5f},
    154     {"m", 0x6d},
    155     {"n", 0x6e},
    156     {"newline", 0x0a},
    157     {"nine", 0x39},
    158     {"number-sign", 0x23},
    159     {"o", 0x6f},
    160     {"one", 0x31},
    161     {"p", 0x70},
    162     {"percent-sign", 0x25},
    163     {"period", 0x2e},
    164     {"plus-sign", 0x2b},
    165     {"q", 0x71},
    166     {"question-mark", 0x3f},
    167     {"quotation-mark", 0x22},
    168     {"r", 0x72},
    169     {"reverse-solidus", 0x5c},
    170     {"right-brace", 0x7d},
    171     {"right-curly-bracket", 0x7d},
    172     {"right-parenthesis", 0x29},
    173     {"right-square-bracket", 0x5d},
    174     {"s", 0x73},
    175     {"semicolon", 0x3b},
    176     {"seven", 0x37},
    177     {"six", 0x36},
    178     {"slash", 0x2f},
    179     {"solidus", 0x2f},
    180     {"space", 0x20},
    181     {"t", 0x74},
    182     {"tab", 0x09},
    183     {"three", 0x33},
    184     {"tilde", 0x7e},
    185     {"two", 0x32},
    186     {"u", 0x75},
    187     {"underscore", 0x5f},
    188     {"v", 0x76},
    189     {"vertical-line", 0x7c},
    190     {"vertical-tab", 0x0b},
    191     {"w", 0x77},
    192     {"x", 0x78},
    193     {"y", 0x79},
    194     {"z", 0x7a},
    195     {"zero", 0x30}
    196 };
    197 
    198 #pragma clang diagnostic push
    199 #pragma clang diagnostic ignored "-Wpadded"
    200 
    201 struct classnames
    202 {
    203     const char* elem_;
    204     ctype_base::mask mask_;
    205 };
    206 
    207 #pragma clang diagnostic pop
    208 
    209 const classnames ClassNames[] =
    210 {
    211     {"alnum",  ctype_base::alnum},
    212     {"alpha",  ctype_base::alpha},
    213     {"blank",  ctype_base::blank},
    214     {"cntrl",  ctype_base::cntrl},
    215     {"d",      ctype_base::digit},
    216     {"digit",  ctype_base::digit},
    217     {"graph",  ctype_base::graph},
    218     {"lower",  ctype_base::lower},
    219     {"print",  ctype_base::print},
    220     {"punct",  ctype_base::punct},
    221     {"s",      ctype_base::space},
    222     {"space",  ctype_base::space},
    223     {"upper",  ctype_base::upper},
    224     {"w",      regex_traits<char>::__regex_word},
    225     {"xdigit", ctype_base::xdigit}
    226 };
    227 
    228 struct use_strcmp
    229 {
    230     bool operator()(const collationnames& x, const char* y)
    231         {return strcmp(x.elem_, y) < 0;}
    232     bool operator()(const classnames& x, const char* y)
    233         {return strcmp(x.elem_, y) < 0;}
    234 };
    235 
    236 }
    237 
    238 string
    239 __get_collation_name(const char* s)
    240 {
    241     const collationnames* i =
    242             _VSTD::lower_bound(begin(collatenames), end(collatenames), s, use_strcmp());
    243     string r;
    244     if (i != end(collatenames) && strcmp(s, i->elem_) == 0)
    245         r = char(i->char_);
    246     return r;
    247 }
    248 
    249 ctype_base::mask
    250 __get_classname(const char* s, bool __icase)
    251 {
    252     const classnames* i =
    253             _VSTD::lower_bound(begin(ClassNames), end(ClassNames), s, use_strcmp());
    254     ctype_base::mask r = 0;
    255     if (i != end(ClassNames) && strcmp(s, i->elem_) == 0)
    256     {
    257         r = i->mask_;
    258         if (r == regex_traits<char>::__regex_word)
    259             r |= ctype_base::alnum | ctype_base::upper | ctype_base::lower;
    260         else if (__icase)
    261         {
    262             if (r & (ctype_base::lower | ctype_base::upper))
    263                 r |= ctype_base::alpha;
    264         }
    265     }
    266     return r;
    267 }
    268 
    269 template <>
    270 void
    271 __match_any_but_newline<char>::__exec(__state& __s) const
    272 {
    273     if (__s.__current_ != __s.__last_)
    274     {
    275         switch (*__s.__current_)
    276         {
    277         case '\r':
    278         case '\n':
    279             __s.__do_ = __state::__reject;
    280             __s.__node_ = nullptr;
    281             break;
    282         default:
    283             __s.__do_ = __state::__accept_and_consume;
    284             ++__s.__current_;
    285             __s.__node_ = this->first();
    286             break;
    287         }
    288     }
    289     else
    290     {
    291         __s.__do_ = __state::__reject;
    292         __s.__node_ = nullptr;
    293     }
    294 }
    295 
    296 template <>
    297 void
    298 __match_any_but_newline<wchar_t>::__exec(__state& __s) const
    299 {
    300     if (__s.__current_ != __s.__last_)
    301     {
    302         switch (*__s.__current_)
    303         {
    304         case '\r':
    305         case '\n':
    306         case 0x2028:
    307         case 0x2029:
    308             __s.__do_ = __state::__reject;
    309             __s.__node_ = nullptr;
    310             break;
    311         default:
    312             __s.__do_ = __state::__accept_and_consume;
    313             ++__s.__current_;
    314             __s.__node_ = this->first();
    315             break;
    316         }
    317     }
    318     else
    319     {
    320         __s.__do_ = __state::__reject;
    321         __s.__node_ = nullptr;
    322     }
    323 }
    324 
    325 _LIBCPP_END_NAMESPACE_STD
    326