Mercurial Repositories > eric / file comparison
comparison: ThirdParty/Pygments/pygments/lexers/compiled.py
ThirdParty/Pygments/pygments/lexers/compiled.py
- changeset 0
- de9c2efb9d02
- child 12
- 1d8dd9706f46
equal
deleted
inserted
replaced
| -1:000000000000 | 0:de9c2efb9d02 |
|---|---|
| 1 # -*- coding: utf-8 -*- | |
| 2 """ | |
| 3 pygments.lexers.compiled | |
| 4 ~~~~~~~~~~~~~~~~~~~~~~~~ | |
| 5 | |
| 6 Lexers for compiled languages. | |
| 7 | |
| 8 :copyright: Copyright 2006-2009 by the Pygments team, see AUTHORS. | |
| 9 :license: BSD, see LICENSE for details. | |
| 10 """ | |
| 11 | |
| 12 import re | |
| 13 try: | |
| 14 set | |
| 15 except NameError: | |
| 16 from sets import Set as set | |
| 17 | |
| 18 from pygments.scanner import Scanner | |
| 19 from pygments.lexer import Lexer, RegexLexer, include, bygroups, using, \ | |
| 20 this, combined | |
| 21 from pygments.util import get_bool_opt, get_list_opt | |
| 22 from pygments.token import \ | |
| 23 Text, Comment, Operator, Keyword, Name, String, Number, Punctuation, \ | |
| 24 Error | |
| 25 | |
| 26 # backwards compatibility | |
| 27 from pygments.lexers.functional import OcamlLexer | |
| 28 | |
| 29 __all__ = ['CLexer', 'CppLexer', 'DLexer', 'DelphiLexer', 'JavaLexer', | |
| 30 'ScalaLexer', 'DylanLexer', 'OcamlLexer', 'ObjectiveCLexer', | |
| 31 'FortranLexer', 'GLShaderLexer', 'PrologLexer', 'CythonLexer', | |
| 32 'ValaLexer'] | |
| 33 | |
| 34 | |
| 35 class CLexer(RegexLexer): | |
| 36 """ | |
| 37 For C source code with preprocessor directives. | |
| 38 """ | |
| 39 name = 'C' | |
| 40 aliases = ['c'] | |
| 41 filenames = ['*.c', '*.h'] | |
| 42 mimetypes = ['text/x-chdr', 'text/x-csrc'] | |
| 43 | |
| 44 #: optional Comment or Whitespace | |
| 45 _ws = r'(?:\s|//.*?\n|/[*].*?[*]/)+' | |
| 46 | |
| 47 tokens = { | |
| 48 'whitespace': [ | |
| 49 (r'^\s*#if\s+0', Comment.Preproc, 'if0'), | |
| 50 (r'^\s*#', Comment.Preproc, 'macro'), | |
| 51 (r'\n', Text), | |
| 52 (r'\s+', Text), | |
| 53 (r'\\\n', Text), # line continuation | |
| 54 (r'//(\n|(.|\n)*?[^\\]\n)', Comment.Single), | |
| 55 (r'/(\\\n)?[*](.|\n)*?[*](\\\n)?/', Comment.Multiline), | |
| 56 ], | |
| 57 'statements': [ | |
| 58 (r'L?"', String, 'string'), | |
| 59 (r"L?'(\\.|\\[0-7]{1,3}|\\x[a-fA-F0-9]{1,2}|[^\\\'\n])'", String.Char), | |
| 60 (r'(\d+\.\d*|\.\d+|\d+)[eE][+-]?\d+[lL]?', Number.Float), | |
| 61 (r'(\d+\.\d*|\.\d+|\d+[fF])[fF]?', Number.Float), | |
| 62 (r'0x[0-9a-fA-F]+[Ll]?', Number.Hex), | |
| 63 (r'0[0-7]+[Ll]?', Number.Oct), | |
| 64 (r'\d+[Ll]?', Number.Integer), | |
| 65 (r'[~!%^&*+=|?:<>/-]', Operator), | |
| 66 (r'[()\[\],.]', Punctuation), | |
| 67 (r'\b(case)(.+?)(:)', bygroups(Keyword, using(this), Text)), | |
| 68 (r'(auto|break|case|const|continue|default|do|else|enum|extern|' | |
| 69 r'for|goto|if|register|restricted|return|sizeof|static|struct|' | |
| 70 r'switch|typedef|union|volatile|virtual|while)\b', Keyword), | |
| 71 (r'(int|long|float|short|double|char|unsigned|signed|void)\b', | |
| 72 Keyword.Type), | |
| 73 (r'(_{0,2}inline|naked|restrict|thread|typename)\b', Keyword.Reserved), | |
| 74 (r'__(asm|int8|based|except|int16|stdcall|cdecl|fastcall|int32|' | |
| 75 r'declspec|finally|int64|try|leave)\b', Keyword.Reserved), | |
| 76 (r'(true|false|NULL)\b', Name.Builtin), | |
| 77 ('[a-zA-Z_][a-zA-Z0-9_]*:(?!:)', Name.Label), | |
| 78 ('[a-zA-Z_][a-zA-Z0-9_]*', Name), | |
| 79 ], | |
| 80 'root': [ | |
| 81 include('whitespace'), | |
| 82 # functions | |
| 83 (r'((?:[a-zA-Z0-9_*\s])+?(?:\s|[*]))' # return arguments | |
| 84 r'([a-zA-Z_][a-zA-Z0-9_]*)' # method name | |
| 85 r'(\s*\([^;]*?\))' # signature | |
| 86 r'(' + _ws + r')({)', | |
| 87 bygroups(using(this), Name.Function, using(this), using(this), | |
| 88 Punctuation), | |
| 89 'function'), | |
| 90 # function declarations | |
| 91 (r'((?:[a-zA-Z0-9_*\s])+?(?:\s|[*]))' # return arguments | |
| 92 r'([a-zA-Z_][a-zA-Z0-9_]*)' # method name | |
| 93 r'(\s*\([^;]*?\))' # signature | |
| 94 r'(' + _ws + r')(;)', | |
| 95 bygroups(using(this), Name.Function, using(this), using(this), | |
| 96 Punctuation)), | |
| 97 ('', Text, 'statement'), | |
| 98 ], | |
| 99 'statement' : [ | |
| 100 include('whitespace'), | |
| 101 include('statements'), | |
| 102 ('[{}]', Punctuation), | |
| 103 (';', Punctuation, '#pop'), | |
| 104 ], | |
| 105 'function': [ | |
| 106 include('whitespace'), | |
| 107 include('statements'), | |
| 108 (';', Punctuation), | |
| 109 ('{', Punctuation, '#push'), | |
| 110 ('}', Punctuation, '#pop'), | |
| 111 ], | |
| 112 'string': [ | |
| 113 (r'"', String, '#pop'), | |
| 114 (r'\\([\\abfnrtv"\']|x[a-fA-F0-9]{2,4}|[0-7]{1,3})', String.Escape), | |
| 115 (r'[^\\"\n]+', String), # all other characters | |
| 116 (r'\\\n', String), # line continuation | |
| 117 (r'\\', String), # stray backslash | |
| 118 ], | |
| 119 'macro': [ | |
| 120 (r'[^/\n]+', Comment.Preproc), | |
| 121 (r'/[*](.|\n)*?[*]/', Comment.Multiline), | |
| 122 (r'//.*?\n', Comment.Single, '#pop'), | |
| 123 (r'/', Comment.Preproc), | |
| 124 (r'(?<=\\)\n', Comment.Preproc), | |
| 125 (r'\n', Comment.Preproc, '#pop'), | |
| 126 ], | |
| 127 'if0': [ | |
| 128 (r'^\s*#if.*?(?<!\\)\n', Comment.Preproc, '#push'), | |
| 129 (r'^\s*#el(?:se|if).*\n', Comment.Preproc, '#pop'), | |
| 130 (r'^\s*#endif.*?(?<!\\)\n', Comment.Preproc, '#pop'), | |
| 131 (r'.*?\n', Comment), | |
| 132 ] | |
| 133 } | |
| 134 | |
| 135 stdlib_types = ['size_t', 'ssize_t', 'off_t', 'wchar_t', 'ptrdiff_t', | |
| 136 'sig_atomic_t', 'fpos_t', 'clock_t', 'time_t', 'va_list', | |
| 137 'jmp_buf', 'FILE', 'DIR', 'div_t', 'ldiv_t', 'mbstate_t', | |
| 138 'wctrans_t', 'wint_t', 'wctype_t'] | |
| 139 c99_types = ['_Bool', '_Complex', 'int8_t', 'int16_t', 'int32_t', 'int64_t', | |
| 140 'uint8_t', 'uint16_t', 'uint32_t', 'uint64_t', 'int_least8_t', | |
| 141 'int_least16_t', 'int_least32_t', 'int_least64_t', | |
| 142 'uint_least8_t', 'uint_least16_t', 'uint_least32_t', | |
| 143 'uint_least64_t', 'int_fast8_t', 'int_fast16_t', 'int_fast32_t', | |
| 144 'int_fast64_t', 'uint_fast8_t', 'uint_fast16_t', 'uint_fast32_t', | |
| 145 'uint_fast64_t', 'intptr_t', 'uintptr_t', 'intmax_t', 'uintmax_t'] | |
| 146 | |
| 147 def __init__(self, **options): | |
| 148 self.stdlibhighlighting = get_bool_opt(options, | |
| 149 'stdlibhighlighting', True) | |
| 150 self.c99highlighting = get_bool_opt(options, | |
| 151 'c99highlighting', True) | |
| 152 RegexLexer.__init__(self, **options) | |
| 153 | |
| 154 def get_tokens_unprocessed(self, text): | |
| 155 for index, token, value in \ | |
| 156 RegexLexer.get_tokens_unprocessed(self, text): | |
| 157 if token is Name: | |
| 158 if self.stdlibhighlighting and value in self.stdlib_types: | |
| 159 token = Keyword.Type | |
| 160 elif self.c99highlighting and value in self.c99_types: | |
| 161 token = Keyword.Type | |
| 162 yield index, token, value | |
| 163 | |
| 164 class CppLexer(RegexLexer): | |
| 165 """ | |
| 166 For C++ source code with preprocessor directives. | |
| 167 """ | |
| 168 name = 'C++' | |
| 169 aliases = ['cpp', 'c++'] | |
| 170 filenames = ['*.cpp', '*.hpp', '*.c++', '*.h++', '*.cc', '*.hh', '*.cxx', '*.hxx'] | |
| 171 mimetypes = ['text/x-c++hdr', 'text/x-c++src'] | |
| 172 | |
| 173 tokens = { | |
| 174 'root': [ | |
| 175 (r'^\s*#if\s+0', Comment.Preproc, 'if0'), | |
| 176 (r'^\s*#', Comment.Preproc, 'macro'), | |
| 177 (r'\n', Text), | |
| 178 (r'\s+', Text), | |
| 179 (r'\\\n', Text), # line continuation | |
| 180 (r'/(\\\n)?/(\n|(.|\n)*?[^\\]\n)', Comment.Single), | |
| 181 (r'/(\\\n)?[*](.|\n)*?[*](\\\n)?/', Comment.Multiline), | |
| 182 (r'[{}]', Punctuation), | |
| 183 (r'L?"', String, 'string'), | |
| 184 (r"L?'(\\.|\\[0-7]{1,3}|\\x[a-fA-F0-9]{1,2}|[^\\\'\n])'", String.Char), | |
| 185 (r'(\d+\.\d*|\.\d+|\d+)[eE][+-]?\d+[lL]?', Number.Float), | |
| 186 (r'(\d+\.\d*|\.\d+|\d+[fF])[fF]?', Number.Float), | |
| 187 (r'0x[0-9a-fA-F]+[Ll]?', Number.Hex), | |
| 188 (r'0[0-7]+[Ll]?', Number.Oct), | |
| 189 (r'\d+[Ll]?', Number.Integer), | |
| 190 (r'[~!%^&*+=|?:<>/-]', Operator), | |
| 191 (r'[()\[\],.;]', Punctuation), | |
| 192 (r'(asm|auto|break|case|catch|const|const_cast|continue|' | |
| 193 r'default|delete|do|dynamic_cast|else|enum|explicit|export|' | |
| 194 r'extern|for|friend|goto|if|mutable|namespace|new|operator|' | |
| 195 r'private|protected|public|register|reinterpret_cast|return|' | |
| 196 r'restrict|sizeof|static|static_cast|struct|switch|template|' | |
| 197 r'this|throw|throws|try|typedef|typeid|typename|union|using|' | |
| 198 r'volatile|virtual|while)\b', Keyword), | |
| 199 (r'(class)(\s+)', bygroups(Keyword, Text), 'classname'), | |
| 200 (r'(bool|int|long|float|short|double|char|unsigned|signed|' | |
| 201 r'void|wchar_t)\b', Keyword.Type), | |
| 202 (r'(_{0,2}inline|naked|thread)\b', Keyword.Reserved), | |
| 203 (r'__(asm|int8|based|except|int16|stdcall|cdecl|fastcall|int32|' | |
| 204 r'declspec|finally|int64|try|leave|wchar_t|w64|virtual_inheritance|' | |
| 205 r'uuidof|unaligned|super|single_inheritance|raise|noop|' | |
| 206 r'multiple_inheritance|m128i|m128d|m128|m64|interface|' | |
| 207 r'identifier|forceinline|event|assume)\b', Keyword.Reserved), | |
| 208 (r'(true|false)\b', Keyword.Constant), | |
| 209 (r'NULL\b', Name.Builtin), | |
| 210 ('[a-zA-Z_][a-zA-Z0-9_]*:(?!:)', Name.Label), | |
| 211 ('[a-zA-Z_][a-zA-Z0-9_]*', Name), | |
| 212 ], | |
| 213 'classname': [ | |
| 214 (r'[a-zA-Z_][a-zA-Z0-9_]*', Name.Class, '#pop'), | |
| 215 # template specification | |
| 216 (r'\s*(?=>)', Text, '#pop'), | |
| 217 ], | |
| 218 'string': [ | |
| 219 (r'"', String, '#pop'), | |
| 220 (r'\\([\\abfnrtv"\']|x[a-fA-F0-9]{2,4}|[0-7]{1,3})', String.Escape), | |
| 221 (r'[^\\"\n]+', String), # all other characters | |
| 222 (r'\\\n', String), # line continuation | |
| 223 (r'\\', String), # stray backslash | |
| 224 ], | |
| 225 'macro': [ | |
| 226 (r'[^/\n]+', Comment.Preproc), | |
| 227 (r'/[*](.|\n)*?[*]/', Comment.Multiline), | |
| 228 (r'//.*?\n', Comment.Single, '#pop'), | |
| 229 (r'/', Comment.Preproc), | |
| 230 (r'(?<=\\)\n', Comment.Preproc), | |
| 231 (r'\n', Comment.Preproc, '#pop'), | |
| 232 ], | |
| 233 'if0': [ | |
| 234 (r'^\s*#if.*?(?<!\\)\n', Comment.Preproc, '#push'), | |
| 235 (r'^\s*#endif.*?(?<!\\)\n', Comment.Preproc, '#pop'), | |
| 236 (r'.*?\n', Comment), | |
| 237 ] | |
| 238 } | |
| 239 | |
| 240 | |
| 241 class DLexer(RegexLexer): | |
| 242 """ | |
| 243 For D source. | |
| 244 """ | |
| 245 name = 'D' | |
| 246 filenames = ['*.d', '*.di'] | |
| 247 aliases = ['d'] | |
| 248 mimetypes = ['text/x-dsrc'] | |
| 249 | |
| 250 tokens = { | |
| 251 'root': [ | |
| 252 (r'\n', Text), | |
| 253 (r'\s+', Text), | |
| 254 #(r'\\\n', Text), # line continuations | |
| 255 # Comments | |
| 256 (r'//(.*?)\n', Comment.Single), | |
| 257 (r'/(\\\n)?[*](.|\n)*?[*](\\\n)?/', Comment.Multiline), | |
| 258 (r'/\+', Comment.Multiline, 'nested_comment'), | |
| 259 # Keywords | |
| 260 (r'(abstract|alias|align|asm|assert|auto|body|break|case|cast' | |
| 261 r'|catch|class|const|continue|debug|default|delegate|delete' | |
| 262 r'|deprecated|do|else|enum|export|extern|finally|final' | |
| 263 r'|foreach_reverse|foreach|for|function|goto|if|import|inout' | |
| 264 r'|interface|invariant|in|is|lazy|mixin|module|new|nothrow|out' | |
| 265 r'|override|package|pragma|private|protected|public|pure|ref|return' | |
| 266 r'|scope|static|struct|super|switch|synchronized|template|this' | |
| 267 r'|throw|try|typedef|typeid|typeof|union|unittest|version|volatile' | |
| 268 r'|while|with|__traits)\b', Keyword | |
| 269 ), | |
| 270 (r'(bool|byte|cdouble|cent|cfloat|char|creal|dchar|double|float' | |
| 271 r'|idouble|ifloat|int|ireal|long|real|short|ubyte|ucent|uint|ulong' | |
| 272 r'|ushort|void|wchar)\b', Keyword.Type | |
| 273 ), | |
| 274 (r'(false|true|null)\b', Keyword.Constant), | |
| 275 (r'macro\b', Keyword.Reserved), | |
| 276 (r'(string|wstring|dstring)\b', Name.Builtin), | |
| 277 # FloatLiteral | |
| 278 # -- HexFloat | |
| 279 (r'0[xX]([0-9a-fA-F_]*\.[0-9a-fA-F_]+|[0-9a-fA-F_]+)' | |
| 280 r'[pP][+\-]?[0-9_]+[fFL]?[i]?', Number.Float), | |
| 281 # -- DecimalFloat | |
| 282 (r'[0-9_]+(\.[0-9_]+[eE][+\-]?[0-9_]+|' | |
| 283 r'\.[0-9_]*|[eE][+\-]?[0-9_]+)[fFL]?[i]?', Number.Float), | |
| 284 (r'\.(0|[1-9][0-9_]*)([eE][+\-]?[0-9_]+)?[fFL]?[i]?', Number.Float), | |
| 285 # IntegerLiteral | |
| 286 # -- Binary | |
| 287 (r'0[Bb][01_]+', Number), | |
| 288 # -- Octal | |
| 289 (r'0[0-7_]+', Number.Oct), | |
| 290 # -- Hexadecimal | |
| 291 (r'0[xX][0-9a-fA-F_]+', Number.Hex), | |
| 292 # -- Decimal | |
| 293 (r'(0|[1-9][0-9_]*)([LUu]|Lu|LU|uL|UL)?', Number.Integer), | |
| 294 # CharacterLiteral | |
| 295 (r"""'(\\['"?\\abfnrtv]|\\x[0-9a-fA-F]{2}|\\[0-7]{1,3}""" | |
| 296 r"""|\\u[0-9a-fA-F]{4}|\\U[0-9a-fA-F]{8}|\\&\w+;|.)'""", | |
| 297 String.Char | |
| 298 ), | |
| 299 # StringLiteral | |
| 300 # -- WysiwygString | |
| 301 (r'r"[^"]*"[cwd]?', String), | |
| 302 # -- AlternateWysiwygString | |
| 303 (r'`[^`]*`[cwd]?', String), | |
| 304 # -- DoubleQuotedString | |
| 305 (r'"(\\\\|\\"|[^"])*"[cwd]?', String), | |
| 306 # -- EscapeSequence | |
| 307 (r"""\\(['"?\\abfnrtv]|x[0-9a-fA-F]{2}|[0-7]{1,3}""" | |
| 308 r"""|u[0-9a-fA-F]{4}|U[0-9a-fA-F]{8}|&\w+;)""", | |
| 309 String | |
| 310 ), | |
| 311 # -- HexString | |
| 312 (r'x"[0-9a-fA-F_\s]*"[cwd]?', String), | |
| 313 # -- DelimitedString | |
| 314 (r'q"\[', String, 'delimited_bracket'), | |
| 315 (r'q"\(', String, 'delimited_parenthesis'), | |
| 316 (r'q"<', String, 'delimited_angle'), | |
| 317 (r'q"{', String, 'delimited_curly'), | |
| 318 (r'q"([a-zA-Z_]\w*)\n.*?\n\1"', String), | |
| 319 (r'q"(.).*?\1"', String), | |
| 320 # -- TokenString | |
| 321 (r'q{', String, 'token_string'), | |
| 322 # Tokens | |
| 323 (r'(~=|\^=|%=|\*=|==|!>=|!<=|!<>=|!<>|!<|!>|!=|>>>=|>>>|>>=|>>|>=' | |
| 324 r'|<>=|<>|<<=|<<|<=|\+\+|\+=|--|-=|\|\||\|=|&&|&=|\.\.\.|\.\.|/=)' | |
| 325 r'|[/.&|\-+<>!()\[\]{}?,;:$=*%^~]', Punctuation | |
| 326 ), | |
| 327 # Identifier | |
| 328 (r'[a-zA-Z_]\w*', Name), | |
| 329 ], | |
| 330 'nested_comment': [ | |
| 331 (r'[^+/]+', Comment.Multiline), | |
| 332 (r'/\+', Comment.Multiline, '#push'), | |
| 333 (r'\+/', Comment.Multiline, '#pop'), | |
| 334 (r'[+/]', Comment.Multiline), | |
| 335 ], | |
| 336 'token_string': [ | |
| 337 (r'{', Punctuation, 'token_string_nest'), | |
| 338 (r'}', String, '#pop'), | |
| 339 include('root'), | |
| 340 ], | |
| 341 'token_string_nest': [ | |
| 342 (r'{', Punctuation, '#push'), | |
| 343 (r'}', Punctuation, '#pop'), | |
| 344 include('root'), | |
| 345 ], | |
| 346 'delimited_bracket': [ | |
| 347 (r'[^\[\]]+', String), | |
| 348 (r'\[', String, 'delimited_inside_bracket'), | |
| 349 (r'\]"', String, '#pop'), | |
| 350 ], | |
| 351 'delimited_inside_bracket': [ | |
| 352 (r'[^\[\]]+', String), | |
| 353 (r'\[', String, '#push'), | |
| 354 (r'\]', String, '#pop'), | |
| 355 ], | |
| 356 'delimited_parenthesis': [ | |
| 357 (r'[^\(\)]+', String), | |
| 358 (r'\(', String, 'delimited_inside_parenthesis'), | |
| 359 (r'\)"', String, '#pop'), | |
| 360 ], | |
| 361 'delimited_inside_parenthesis': [ | |
| 362 (r'[^\(\)]+', String), | |
| 363 (r'\(', String, '#push'), | |
| 364 (r'\)', String, '#pop'), | |
| 365 ], | |
| 366 'delimited_angle': [ | |
| 367 (r'[^<>]+', String), | |
| 368 (r'<', String, 'delimited_inside_angle'), | |
| 369 (r'>"', String, '#pop'), | |
| 370 ], | |
| 371 'delimited_inside_angle': [ | |
| 372 (r'[^<>]+', String), | |
| 373 (r'<', String, '#push'), | |
| 374 (r'>', String, '#pop'), | |
| 375 ], | |
| 376 'delimited_curly': [ | |
| 377 (r'[^{}]+', String), | |
| 378 (r'{', String, 'delimited_inside_curly'), | |
| 379 (r'}"', String, '#pop'), | |
| 380 ], | |
| 381 'delimited_inside_curly': [ | |
| 382 (r'[^{}]+', String), | |
| 383 (r'{', String, '#push'), | |
| 384 (r'}', String, '#pop'), | |
| 385 ], | |
| 386 } | |
| 387 | |
| 388 | |
| 389 class DelphiLexer(Lexer): | |
| 390 """ | |
| 391 For `Delphi <http://www.borland.com/delphi/>`_ (Borland Object Pascal), | |
| 392 Turbo Pascal and Free Pascal source code. | |
| 393 | |
| 394 Additional options accepted: | |
| 395 | |
| 396 `turbopascal` | |
| 397 Highlight Turbo Pascal specific keywords (default: ``True``). | |
| 398 `delphi` | |
| 399 Highlight Borland Delphi specific keywords (default: ``True``). | |
| 400 `freepascal` | |
| 401 Highlight Free Pascal specific keywords (default: ``True``). | |
| 402 `units` | |
| 403 A list of units that should be considered builtin, supported are | |
| 404 ``System``, ``SysUtils``, ``Classes`` and ``Math``. | |
| 405 Default is to consider all of them builtin. | |
| 406 """ | |
| 407 name = 'Delphi' | |
| 408 aliases = ['delphi', 'pas', 'pascal', 'objectpascal'] | |
| 409 filenames = ['*.pas'] | |
| 410 mimetypes = ['text/x-pascal'] | |
| 411 | |
| 412 TURBO_PASCAL_KEYWORDS = [ | |
| 413 'absolute', 'and', 'array', 'asm', 'begin', 'break', 'case', | |
| 414 'const', 'constructor', 'continue', 'destructor', 'div', 'do', | |
| 415 'downto', 'else', 'end', 'file', 'for', 'function', 'goto', | |
| 416 'if', 'implementation', 'in', 'inherited', 'inline', 'interface', | |
| 417 'label', 'mod', 'nil', 'not', 'object', 'of', 'on', 'operator', | |
| 418 'or', 'packed', 'procedure', 'program', 'record', 'reintroduce', | |
| 419 'repeat', 'self', 'set', 'shl', 'shr', 'string', 'then', 'to', | |
| 420 'type', 'unit', 'until', 'uses', 'var', 'while', 'with', 'xor' | |
| 421 ] | |
| 422 | |
| 423 DELPHI_KEYWORDS = [ | |
| 424 'as', 'class', 'except', 'exports', 'finalization', 'finally', | |
| 425 'initialization', 'is', 'library', 'on', 'property', 'raise', | |
| 426 'threadvar', 'try' | |
| 427 ] | |
| 428 | |
| 429 FREE_PASCAL_KEYWORDS = [ | |
| 430 'dispose', 'exit', 'false', 'new', 'true' | |
| 431 ] | |
| 432 | |
| 433 BLOCK_KEYWORDS = set([ | |
| 434 'begin', 'class', 'const', 'constructor', 'destructor', 'end', | |
| 435 'finalization', 'function', 'implementation', 'initialization', | |
| 436 'label', 'library', 'operator', 'procedure', 'program', 'property', | |
| 437 'record', 'threadvar', 'type', 'unit', 'uses', 'var' | |
| 438 ]) | |
| 439 | |
| 440 FUNCTION_MODIFIERS = set([ | |
| 441 'alias', 'cdecl', 'export', 'inline', 'interrupt', 'nostackframe', | |
| 442 'pascal', 'register', 'safecall', 'softfloat', 'stdcall', | |
| 443 'varargs', 'name', 'dynamic', 'near', 'virtual', 'external', | |
| 444 'override', 'assembler' | |
| 445 ]) | |
| 446 | |
| 447 # XXX: those aren't global. but currently we know no way for defining | |
| 448 # them just for the type context. | |
| 449 DIRECTIVES = set([ | |
| 450 'absolute', 'abstract', 'assembler', 'cppdecl', 'default', 'far', | |
| 451 'far16', 'forward', 'index', 'oldfpccall', 'private', 'protected', | |
| 452 'published', 'public' | |
| 453 ]) | |
| 454 | |
| 455 BUILTIN_TYPES = set([ | |
| 456 'ansichar', 'ansistring', 'bool', 'boolean', 'byte', 'bytebool', | |
| 457 'cardinal', 'char', 'comp', 'currency', 'double', 'dword', | |
| 458 'extended', 'int64', 'integer', 'iunknown', 'longbool', 'longint', | |
| 459 'longword', 'pansichar', 'pansistring', 'pbool', 'pboolean', | |
| 460 'pbyte', 'pbytearray', 'pcardinal', 'pchar', 'pcomp', 'pcurrency', | |
| 461 'pdate', 'pdatetime', 'pdouble', 'pdword', 'pextended', 'phandle', | |
| 462 'pint64', 'pinteger', 'plongint', 'plongword', 'pointer', | |
| 463 'ppointer', 'pshortint', 'pshortstring', 'psingle', 'psmallint', | |
| 464 'pstring', 'pvariant', 'pwidechar', 'pwidestring', 'pword', | |
| 465 'pwordarray', 'pwordbool', 'real', 'real48', 'shortint', | |
| 466 'shortstring', 'single', 'smallint', 'string', 'tclass', 'tdate', | |
| 467 'tdatetime', 'textfile', 'thandle', 'tobject', 'ttime', 'variant', | |
| 468 'widechar', 'widestring', 'word', 'wordbool' | |
| 469 ]) | |
| 470 | |
| 471 BUILTIN_UNITS = { | |
| 472 'System': [ | |
| 473 'abs', 'acquireexceptionobject', 'addr', 'ansitoutf8', | |
| 474 'append', 'arctan', 'assert', 'assigned', 'assignfile', | |
| 475 'beginthread', 'blockread', 'blockwrite', 'break', 'chdir', | |
| 476 'chr', 'close', 'closefile', 'comptocurrency', 'comptodouble', | |
| 477 'concat', 'continue', 'copy', 'cos', 'dec', 'delete', | |
| 478 'dispose', 'doubletocomp', 'endthread', 'enummodules', | |
| 479 'enumresourcemodules', 'eof', 'eoln', 'erase', 'exceptaddr', | |
| 480 'exceptobject', 'exclude', 'exit', 'exp', 'filepos', 'filesize', | |
| 481 'fillchar', 'finalize', 'findclasshinstance', 'findhinstance', | |
| 482 'findresourcehinstance', 'flush', 'frac', 'freemem', | |
| 483 'get8087cw', 'getdir', 'getlasterror', 'getmem', | |
| 484 'getmemorymanager', 'getmodulefilename', 'getvariantmanager', | |
| 485 'halt', 'hi', 'high', 'inc', 'include', 'initialize', 'insert', | |
| 486 'int', 'ioresult', 'ismemorymanagerset', 'isvariantmanagerset', | |
| 487 'length', 'ln', 'lo', 'low', 'mkdir', 'move', 'new', 'odd', | |
| 488 'olestrtostring', 'olestrtostrvar', 'ord', 'paramcount', | |
| 489 'paramstr', 'pi', 'pos', 'pred', 'ptr', 'pucs4chars', 'random', | |
| 490 'randomize', 'read', 'readln', 'reallocmem', | |
| 491 'releaseexceptionobject', 'rename', 'reset', 'rewrite', 'rmdir', | |
| 492 'round', 'runerror', 'seek', 'seekeof', 'seekeoln', | |
| 493 'set8087cw', 'setlength', 'setlinebreakstyle', | |
| 494 'setmemorymanager', 'setstring', 'settextbuf', | |
| 495 'setvariantmanager', 'sin', 'sizeof', 'slice', 'sqr', 'sqrt', | |
| 496 'str', 'stringofchar', 'stringtoolestr', 'stringtowidechar', | |
| 497 'succ', 'swap', 'trunc', 'truncate', 'typeinfo', | |
| 498 'ucs4stringtowidestring', 'unicodetoutf8', 'uniquestring', | |
| 499 'upcase', 'utf8decode', 'utf8encode', 'utf8toansi', | |
| 500 'utf8tounicode', 'val', 'vararrayredim', 'varclear', | |
| 501 'widecharlentostring', 'widecharlentostrvar', | |
| 502 'widechartostring', 'widechartostrvar', | |
| 503 'widestringtoucs4string', 'write', 'writeln' | |
| 504 ], | |
| 505 'SysUtils': [ | |
| 506 'abort', 'addexitproc', 'addterminateproc', 'adjustlinebreaks', | |
| 507 'allocmem', 'ansicomparefilename', 'ansicomparestr', | |
| 508 'ansicomparetext', 'ansidequotedstr', 'ansiextractquotedstr', | |
| 509 'ansilastchar', 'ansilowercase', 'ansilowercasefilename', | |
| 510 'ansipos', 'ansiquotedstr', 'ansisamestr', 'ansisametext', | |
| 511 'ansistrcomp', 'ansistricomp', 'ansistrlastchar', 'ansistrlcomp', | |
| 512 'ansistrlicomp', 'ansistrlower', 'ansistrpos', 'ansistrrscan', | |
| 513 'ansistrscan', 'ansistrupper', 'ansiuppercase', | |
| 514 'ansiuppercasefilename', 'appendstr', 'assignstr', 'beep', | |
| 515 'booltostr', 'bytetocharindex', 'bytetocharlen', 'bytetype', | |
| 516 'callterminateprocs', 'changefileext', 'charlength', | |
| 517 'chartobyteindex', 'chartobytelen', 'comparemem', 'comparestr', | |
| 518 'comparetext', 'createdir', 'createguid', 'currentyear', | |
| 519 'currtostr', 'currtostrf', 'date', 'datetimetofiledate', | |
| 520 'datetimetostr', 'datetimetostring', 'datetimetosystemtime', | |
| 521 'datetimetotimestamp', 'datetostr', 'dayofweek', 'decodedate', | |
| 522 'decodedatefully', 'decodetime', 'deletefile', 'directoryexists', | |
| 523 'diskfree', 'disksize', 'disposestr', 'encodedate', 'encodetime', | |
| 524 'exceptionerrormessage', 'excludetrailingbackslash', | |
| 525 'excludetrailingpathdelimiter', 'expandfilename', | |
| 526 'expandfilenamecase', 'expanduncfilename', 'extractfiledir', | |
| 527 'extractfiledrive', 'extractfileext', 'extractfilename', | |
| 528 'extractfilepath', 'extractrelativepath', 'extractshortpathname', | |
| 529 'fileage', 'fileclose', 'filecreate', 'filedatetodatetime', | |
| 530 'fileexists', 'filegetattr', 'filegetdate', 'fileisreadonly', | |
| 531 'fileopen', 'fileread', 'filesearch', 'fileseek', 'filesetattr', | |
| 532 'filesetdate', 'filesetreadonly', 'filewrite', 'finalizepackage', | |
| 533 'findclose', 'findcmdlineswitch', 'findfirst', 'findnext', | |
| 534 'floattocurr', 'floattodatetime', 'floattodecimal', 'floattostr', | |
| 535 'floattostrf', 'floattotext', 'floattotextfmt', 'fmtloadstr', | |
| 536 'fmtstr', 'forcedirectories', 'format', 'formatbuf', 'formatcurr', | |
| 537 'formatdatetime', 'formatfloat', 'freeandnil', 'getcurrentdir', | |
| 538 'getenvironmentvariable', 'getfileversion', 'getformatsettings', | |
| 539 'getlocaleformatsettings', 'getmodulename', 'getpackagedescription', | |
| 540 'getpackageinfo', 'gettime', 'guidtostring', 'incamonth', | |
| 541 'includetrailingbackslash', 'includetrailingpathdelimiter', | |
| 542 'incmonth', 'initializepackage', 'interlockeddecrement', | |
| 543 'interlockedexchange', 'interlockedexchangeadd', | |
| 544 'interlockedincrement', 'inttohex', 'inttostr', 'isdelimiter', | |
| 545 'isequalguid', 'isleapyear', 'ispathdelimiter', 'isvalidident', | |
| 546 'languages', 'lastdelimiter', 'loadpackage', 'loadstr', | |
| 547 'lowercase', 'msecstotimestamp', 'newstr', 'nextcharindex', 'now', | |
| 548 'outofmemoryerror', 'quotedstr', 'raiselastoserror', | |
| 549 'raiselastwin32error', 'removedir', 'renamefile', 'replacedate', | |
| 550 'replacetime', 'safeloadlibrary', 'samefilename', 'sametext', | |
| 551 'setcurrentdir', 'showexception', 'sleep', 'stralloc', 'strbufsize', | |
| 552 'strbytetype', 'strcat', 'strcharlength', 'strcomp', 'strcopy', | |
| 553 'strdispose', 'strecopy', 'strend', 'strfmt', 'stricomp', | |
| 554 'stringreplace', 'stringtoguid', 'strlcat', 'strlcomp', 'strlcopy', | |
| 555 'strlen', 'strlfmt', 'strlicomp', 'strlower', 'strmove', 'strnew', | |
| 556 'strnextchar', 'strpas', 'strpcopy', 'strplcopy', 'strpos', | |
| 557 'strrscan', 'strscan', 'strtobool', 'strtobooldef', 'strtocurr', | |
| 558 'strtocurrdef', 'strtodate', 'strtodatedef', 'strtodatetime', | |
| 559 'strtodatetimedef', 'strtofloat', 'strtofloatdef', 'strtoint', | |
| 560 'strtoint64', 'strtoint64def', 'strtointdef', 'strtotime', | |
| 561 'strtotimedef', 'strupper', 'supports', 'syserrormessage', | |
| 562 'systemtimetodatetime', 'texttofloat', 'time', 'timestamptodatetime', | |
| 563 'timestamptomsecs', 'timetostr', 'trim', 'trimleft', 'trimright', | |
| 564 'tryencodedate', 'tryencodetime', 'tryfloattocurr', 'tryfloattodatetime', | |
| 565 'trystrtobool', 'trystrtocurr', 'trystrtodate', 'trystrtodatetime', | |
| 566 'trystrtofloat', 'trystrtoint', 'trystrtoint64', 'trystrtotime', | |
| 567 'unloadpackage', 'uppercase', 'widecomparestr', 'widecomparetext', | |
| 568 'widefmtstr', 'wideformat', 'wideformatbuf', 'widelowercase', | |
| 569 'widesamestr', 'widesametext', 'wideuppercase', 'win32check', | |
| 570 'wraptext' | |
| 571 ], | |
| 572 'Classes': [ | |
| 573 'activateclassgroup', 'allocatehwnd', 'bintohex', 'checksynchronize', | |
| 574 'collectionsequal', 'countgenerations', 'deallocatehwnd', 'equalrect', | |
| 575 'extractstrings', 'findclass', 'findglobalcomponent', 'getclass', | |
| 576 'groupdescendantswith', 'hextobin', 'identtoint', | |
| 577 'initinheritedcomponent', 'inttoident', 'invalidpoint', | |
| 578 'isuniqueglobalcomponentname', 'linestart', 'objectbinarytotext', | |
| 579 'objectresourcetotext', 'objecttexttobinary', 'objecttexttoresource', | |
| 580 'pointsequal', 'readcomponentres', 'readcomponentresex', | |
| 581 'readcomponentresfile', 'rect', 'registerclass', 'registerclassalias', | |
| 582 'registerclasses', 'registercomponents', 'registerintegerconsts', | |
| 583 'registernoicon', 'registernonactivex', 'smallpoint', 'startclassgroup', | |
| 584 'teststreamformat', 'unregisterclass', 'unregisterclasses', | |
| 585 'unregisterintegerconsts', 'unregistermoduleclasses', | |
| 586 'writecomponentresfile' | |
| 587 ], | |
| 588 'Math': [ | |
| 589 'arccos', 'arccosh', 'arccot', 'arccoth', 'arccsc', 'arccsch', 'arcsec', | |
| 590 'arcsech', 'arcsin', 'arcsinh', 'arctan2', 'arctanh', 'ceil', | |
| 591 'comparevalue', 'cosecant', 'cosh', 'cot', 'cotan', 'coth', 'csc', | |
| 592 'csch', 'cycletodeg', 'cycletograd', 'cycletorad', 'degtocycle', | |
| 593 'degtograd', 'degtorad', 'divmod', 'doubledecliningbalance', | |
| 594 'ensurerange', 'floor', 'frexp', 'futurevalue', 'getexceptionmask', | |
| 595 'getprecisionmode', 'getroundmode', 'gradtocycle', 'gradtodeg', | |
| 596 'gradtorad', 'hypot', 'inrange', 'interestpayment', 'interestrate', | |
| 597 'internalrateofreturn', 'intpower', 'isinfinite', 'isnan', 'iszero', | |
| 598 'ldexp', 'lnxp1', 'log10', 'log2', 'logn', 'max', 'maxintvalue', | |
| 599 'maxvalue', 'mean', 'meanandstddev', 'min', 'minintvalue', 'minvalue', | |
| 600 'momentskewkurtosis', 'netpresentvalue', 'norm', 'numberofperiods', | |
| 601 'payment', 'periodpayment', 'poly', 'popnstddev', 'popnvariance', | |
| 602 'power', 'presentvalue', 'radtocycle', 'radtodeg', 'radtograd', | |
| 603 'randg', 'randomrange', 'roundto', 'samevalue', 'sec', 'secant', | |
| 604 'sech', 'setexceptionmask', 'setprecisionmode', 'setroundmode', | |
| 605 'sign', 'simpleroundto', 'sincos', 'sinh', 'slndepreciation', 'stddev', | |
| 606 'sum', 'sumint', 'sumofsquares', 'sumsandsquares', 'syddepreciation', | |
| 607 'tan', 'tanh', 'totalvariance', 'variance' | |
| 608 ] | |
| 609 } | |
| 610 | |
| 611 ASM_REGISTERS = set([ | |
| 612 'ah', 'al', 'ax', 'bh', 'bl', 'bp', 'bx', 'ch', 'cl', 'cr0', | |
| 613 'cr1', 'cr2', 'cr3', 'cr4', 'cs', 'cx', 'dh', 'di', 'dl', 'dr0', | |
| 614 'dr1', 'dr2', 'dr3', 'dr4', 'dr5', 'dr6', 'dr7', 'ds', 'dx', | |
| 615 'eax', 'ebp', 'ebx', 'ecx', 'edi', 'edx', 'es', 'esi', 'esp', | |
| 616 'fs', 'gs', 'mm0', 'mm1', 'mm2', 'mm3', 'mm4', 'mm5', 'mm6', | |
| 617 'mm7', 'si', 'sp', 'ss', 'st0', 'st1', 'st2', 'st3', 'st4', 'st5', | |
| 618 'st6', 'st7', 'xmm0', 'xmm1', 'xmm2', 'xmm3', 'xmm4', 'xmm5', | |
| 619 'xmm6', 'xmm7' | |
| 620 ]) | |
| 621 | |
| 622 ASM_INSTRUCTIONS = set([ | |
| 623 'aaa', 'aad', 'aam', 'aas', 'adc', 'add', 'and', 'arpl', 'bound', | |
| 624 'bsf', 'bsr', 'bswap', 'bt', 'btc', 'btr', 'bts', 'call', 'cbw', | |
| 625 'cdq', 'clc', 'cld', 'cli', 'clts', 'cmc', 'cmova', 'cmovae', | |
| 626 'cmovb', 'cmovbe', 'cmovc', 'cmovcxz', 'cmove', 'cmovg', | |
| 627 'cmovge', 'cmovl', 'cmovle', 'cmovna', 'cmovnae', 'cmovnb', | |
| 628 'cmovnbe', 'cmovnc', 'cmovne', 'cmovng', 'cmovnge', 'cmovnl', | |
| 629 'cmovnle', 'cmovno', 'cmovnp', 'cmovns', 'cmovnz', 'cmovo', | |
| 630 'cmovp', 'cmovpe', 'cmovpo', 'cmovs', 'cmovz', 'cmp', 'cmpsb', | |
| 631 'cmpsd', 'cmpsw', 'cmpxchg', 'cmpxchg486', 'cmpxchg8b', 'cpuid', | |
| 632 'cwd', 'cwde', 'daa', 'das', 'dec', 'div', 'emms', 'enter', 'hlt', | |
| 633 'ibts', 'icebp', 'idiv', 'imul', 'in', 'inc', 'insb', 'insd', | |
| 634 'insw', 'int', 'int01', 'int03', 'int1', 'int3', 'into', 'invd', | |
| 635 'invlpg', 'iret', 'iretd', 'iretw', 'ja', 'jae', 'jb', 'jbe', | |
| 636 'jc', 'jcxz', 'jcxz', 'je', 'jecxz', 'jg', 'jge', 'jl', 'jle', | |
| 637 'jmp', 'jna', 'jnae', 'jnb', 'jnbe', 'jnc', 'jne', 'jng', 'jnge', | |
| 638 'jnl', 'jnle', 'jno', 'jnp', 'jns', 'jnz', 'jo', 'jp', 'jpe', | |
| 639 'jpo', 'js', 'jz', 'lahf', 'lar', 'lcall', 'lds', 'lea', 'leave', | |
| 640 'les', 'lfs', 'lgdt', 'lgs', 'lidt', 'ljmp', 'lldt', 'lmsw', | |
| 641 'loadall', 'loadall286', 'lock', 'lodsb', 'lodsd', 'lodsw', | |
| 642 'loop', 'loope', 'loopne', 'loopnz', 'loopz', 'lsl', 'lss', 'ltr', | |
| 643 'mov', 'movd', 'movq', 'movsb', 'movsd', 'movsw', 'movsx', | |
| 644 'movzx', 'mul', 'neg', 'nop', 'not', 'or', 'out', 'outsb', 'outsd', | |
| 645 'outsw', 'pop', 'popa', 'popad', 'popaw', 'popf', 'popfd', 'popfw', | |
| 646 'push', 'pusha', 'pushad', 'pushaw', 'pushf', 'pushfd', 'pushfw', | |
| 647 'rcl', 'rcr', 'rdmsr', 'rdpmc', 'rdshr', 'rdtsc', 'rep', 'repe', | |
| 648 'repne', 'repnz', 'repz', 'ret', 'retf', 'retn', 'rol', 'ror', | |
| 649 'rsdc', 'rsldt', 'rsm', 'sahf', 'sal', 'salc', 'sar', 'sbb', | |
| 650 'scasb', 'scasd', 'scasw', 'seta', 'setae', 'setb', 'setbe', | |
| 651 'setc', 'setcxz', 'sete', 'setg', 'setge', 'setl', 'setle', | |
| 652 'setna', 'setnae', 'setnb', 'setnbe', 'setnc', 'setne', 'setng', | |
| 653 'setnge', 'setnl', 'setnle', 'setno', 'setnp', 'setns', 'setnz', | |
| 654 'seto', 'setp', 'setpe', 'setpo', 'sets', 'setz', 'sgdt', 'shl', | |
| 655 'shld', 'shr', 'shrd', 'sidt', 'sldt', 'smi', 'smint', 'smintold', | |
| 656 'smsw', 'stc', 'std', 'sti', 'stosb', 'stosd', 'stosw', 'str', | |
| 657 'sub', 'svdc', 'svldt', 'svts', 'syscall', 'sysenter', 'sysexit', | |
| 658 'sysret', 'test', 'ud1', 'ud2', 'umov', 'verr', 'verw', 'wait', | |
| 659 'wbinvd', 'wrmsr', 'wrshr', 'xadd', 'xbts', 'xchg', 'xlat', | |
| 660 'xlatb', 'xor' | |
| 661 ]) | |
| 662 | |
| 663 def __init__(self, **options): | |
| 664 Lexer.__init__(self, **options) | |
| 665 self.keywords = set() | |
| 666 if get_bool_opt(options, 'turbopascal', True): | |
| 667 self.keywords.update(self.TURBO_PASCAL_KEYWORDS) | |
| 668 if get_bool_opt(options, 'delphi', True): | |
| 669 self.keywords.update(self.DELPHI_KEYWORDS) | |
| 670 if get_bool_opt(options, 'freepascal', True): | |
| 671 self.keywords.update(self.FREE_PASCAL_KEYWORDS) | |
| 672 self.builtins = set() | |
| 673 for unit in get_list_opt(options, 'units', self.BUILTIN_UNITS.keys()): | |
| 674 self.builtins.update(self.BUILTIN_UNITS[unit]) | |
| 675 | |
| 676 def get_tokens_unprocessed(self, text): | |
| 677 scanner = Scanner(text, re.DOTALL | re.MULTILINE | re.IGNORECASE) | |
| 678 stack = ['initial'] | |
| 679 in_function_block = False | |
| 680 in_property_block = False | |
| 681 was_dot = False | |
| 682 next_token_is_function = False | |
| 683 next_token_is_property = False | |
| 684 collect_labels = False | |
| 685 block_labels = set() | |
| 686 brace_balance = [0, 0] | |
| 687 | |
| 688 while not scanner.eos: | |
| 689 token = Error | |
| 690 | |
| 691 if stack[-1] == 'initial': | |
| 692 if scanner.scan(r'\s+'): | |
| 693 token = Text | |
| 694 elif scanner.scan(r'\{.*?\}|\(\*.*?\*\)'): | |
| 695 if scanner.match.startswith('$'): | |
| 696 token = Comment.Preproc | |
| 697 else: | |
| 698 token = Comment.Multiline | |
| 699 elif scanner.scan(r'//.*?$'): | |
| 700 token = Comment.Single | |
| 701 elif scanner.scan(r'[-+*\/=<>:;,.@\^]'): | |
| 702 token = Operator | |
| 703 # stop label highlighting on next ";" | |
| 704 if collect_labels and scanner.match == ';': | |
| 705 collect_labels = False | |
| 706 elif scanner.scan(r'[\(\)\[\]]+'): | |
| 707 token = Punctuation | |
| 708 # abort function naming ``foo = Function(...)`` | |
| 709 next_token_is_function = False | |
| 710 # if we are in a function block we count the open | |
| 711 # braces because ootherwise it's impossible to | |
| 712 # determine the end of the modifier context | |
| 713 if in_function_block or in_property_block: | |
| 714 if scanner.match == '(': | |
| 715 brace_balance[0] += 1 | |
| 716 elif scanner.match == ')': | |
| 717 brace_balance[0] -= 1 | |
| 718 elif scanner.match == '[': | |
| 719 brace_balance[1] += 1 | |
| 720 elif scanner.match == ']': | |
| 721 brace_balance[1] -= 1 | |
| 722 elif scanner.scan(r'[A-Za-z_][A-Za-z_0-9]*'): | |
| 723 lowercase_name = scanner.match.lower() | |
| 724 if lowercase_name == 'result': | |
| 725 token = Name.Builtin.Pseudo | |
| 726 elif lowercase_name in self.keywords: | |
| 727 token = Keyword | |
| 728 # if we are in a special block and a | |
| 729 # block ending keyword occours (and the parenthesis | |
| 730 # is balanced) we end the current block context | |
| 731 if (in_function_block or in_property_block) and \ | |
| 732 lowercase_name in self.BLOCK_KEYWORDS and \ | |
| 733 brace_balance[0] <= 0 and \ | |
| 734 brace_balance[1] <= 0: | |
| 735 in_function_block = False | |
| 736 in_property_block = False | |
| 737 brace_balance = [0, 0] | |
| 738 block_labels = set() | |
| 739 if lowercase_name in ('label', 'goto'): | |
| 740 collect_labels = True | |
| 741 elif lowercase_name == 'asm': | |
| 742 stack.append('asm') | |
| 743 elif lowercase_name == 'property': | |
| 744 in_property_block = True | |
| 745 next_token_is_property = True | |
| 746 elif lowercase_name in ('procedure', 'operator', | |
| 747 'function', 'constructor', | |
| 748 'destructor'): | |
| 749 in_function_block = True | |
| 750 next_token_is_function = True | |
| 751 # we are in a function block and the current name | |
| 752 # is in the set of registered modifiers. highlight | |
| 753 # it as pseudo keyword | |
| 754 elif in_function_block and \ | |
| 755 lowercase_name in self.FUNCTION_MODIFIERS: | |
| 756 token = Keyword.Pseudo | |
| 757 # if we are in a property highlight some more | |
| 758 # modifiers | |
| 759 elif in_property_block and \ | |
| 760 lowercase_name in ('read', 'write'): | |
| 761 token = Keyword.Pseudo | |
| 762 next_token_is_function = True | |
| 763 # if the last iteration set next_token_is_function | |
| 764 # to true we now want this name highlighted as | |
| 765 # function. so do that and reset the state | |
| 766 elif next_token_is_function: | |
| 767 # Look if the next token is a dot. If yes it's | |
| 768 # not a function, but a class name and the | |
| 769 # part after the dot a function name | |
| 770 if scanner.test(r'\s*\.\s*'): | |
| 771 token = Name.Class | |
| 772 # it's not a dot, our job is done | |
| 773 else: | |
| 774 token = Name.Function | |
| 775 next_token_is_function = False | |
| 776 # same for properties | |
| 777 elif next_token_is_property: | |
| 778 token = Name.Property | |
| 779 next_token_is_property = False | |
| 780 # Highlight this token as label and add it | |
| 781 # to the list of known labels | |
| 782 elif collect_labels: | |
| 783 token = Name.Label | |
| 784 block_labels.add(scanner.match.lower()) | |
| 785 # name is in list of known labels | |
| 786 elif lowercase_name in block_labels: | |
| 787 token = Name.Label | |
| 788 elif lowercase_name in self.BUILTIN_TYPES: | |
| 789 token = Keyword.Type | |
| 790 elif lowercase_name in self.DIRECTIVES: | |
| 791 token = Keyword.Pseudo | |
| 792 # builtins are just builtins if the token | |
| 793 # before isn't a dot | |
| 794 elif not was_dot and lowercase_name in self.builtins: | |
| 795 token = Name.Builtin | |
| 796 else: | |
| 797 token = Name | |
| 798 elif scanner.scan(r"'"): | |
| 799 token = String | |
| 800 stack.append('string') | |
| 801 elif scanner.scan(r'\#(\d+|\$[0-9A-Fa-f]+)'): | |
| 802 token = String.Char | |
| 803 elif scanner.scan(r'\$[0-9A-Fa-f]+'): | |
| 804 token = Number.Hex | |
| 805 elif scanner.scan(r'\d+(?![eE]|\.[^.])'): | |
| 806 token = Number.Integer | |
| 807 elif scanner.scan(r'\d+(\.\d+([eE][+-]?\d+)?|[eE][+-]?\d+)'): | |
| 808 token = Number.Float | |
| 809 else: | |
| 810 # if the stack depth is deeper than once, pop | |
| 811 if len(stack) > 1: | |
| 812 stack.pop() | |
| 813 scanner.get_char() | |
| 814 | |
| 815 elif stack[-1] == 'string': | |
| 816 if scanner.scan(r"''"): | |
| 817 token = String.Escape | |
| 818 elif scanner.scan(r"'"): | |
| 819 token = String | |
| 820 stack.pop() | |
| 821 elif scanner.scan(r"[^']*"): | |
| 822 token = String | |
| 823 else: | |
| 824 scanner.get_char() | |
| 825 stack.pop() | |
| 826 | |
| 827 elif stack[-1] == 'asm': | |
| 828 if scanner.scan(r'\s+'): | |
| 829 token = Text | |
| 830 elif scanner.scan(r'end'): | |
| 831 token = Keyword | |
| 832 stack.pop() | |
| 833 elif scanner.scan(r'\{.*?\}|\(\*.*?\*\)'): | |
| 834 if scanner.match.startswith('$'): | |
| 835 token = Comment.Preproc | |
| 836 else: | |
| 837 token = Comment.Multiline | |
| 838 elif scanner.scan(r'//.*?$'): | |
| 839 token = Comment.Single | |
| 840 elif scanner.scan(r"'"): | |
| 841 token = String | |
| 842 stack.append('string') | |
| 843 elif scanner.scan(r'@@[A-Za-z_][A-Za-z_0-9]*'): | |
| 844 token = Name.Label | |
| 845 elif scanner.scan(r'[A-Za-z_][A-Za-z_0-9]*'): | |
| 846 lowercase_name = scanner.match.lower() | |
| 847 if lowercase_name in self.ASM_INSTRUCTIONS: | |
| 848 token = Keyword | |
| 849 elif lowercase_name in self.ASM_REGISTERS: | |
| 850 token = Name.Builtin | |
| 851 else: | |
| 852 token = Name | |
| 853 elif scanner.scan(r'[-+*\/=<>:;,.@\^]+'): | |
| 854 token = Operator | |
| 855 elif scanner.scan(r'[\(\)\[\]]+'): | |
| 856 token = Punctuation | |
| 857 elif scanner.scan(r'\$[0-9A-Fa-f]+'): | |
| 858 token = Number.Hex | |
| 859 elif scanner.scan(r'\d+(?![eE]|\.[^.])'): | |
| 860 token = Number.Integer | |
| 861 elif scanner.scan(r'\d+(\.\d+([eE][+-]?\d+)?|[eE][+-]?\d+)'): | |
| 862 token = Number.Float | |
| 863 else: | |
| 864 scanner.get_char() | |
| 865 stack.pop() | |
| 866 | |
| 867 # save the dot!!!11 | |
| 868 if scanner.match.strip(): | |
| 869 was_dot = scanner.match == '.' | |
| 870 yield scanner.start_pos, token, scanner.match or '' | |
| 871 | |
| 872 | |
| 873 class JavaLexer(RegexLexer): | |
| 874 """ | |
| 875 For `Java <http://www.sun.com/java/>`_ source code. | |
| 876 """ | |
| 877 | |
| 878 name = 'Java' | |
| 879 aliases = ['java'] | |
| 880 filenames = ['*.java'] | |
| 881 mimetypes = ['text/x-java'] | |
| 882 | |
| 883 flags = re.MULTILINE | re.DOTALL | |
| 884 | |
| 885 #: optional Comment or Whitespace | |
| 886 _ws = r'(?:\s|//.*?\n|/[*].*?[*]/)+' | |
| 887 | |
| 888 tokens = { | |
| 889 'root': [ | |
| 890 # method names | |
| 891 (r'^(\s*(?:[a-zA-Z_][a-zA-Z0-9_\.\[\]]*\s+)+?)' # return arguments | |
| 892 r'([a-zA-Z_][a-zA-Z0-9_]*)' # method name | |
| 893 r'(\s*)(\()', # signature start | |
| 894 bygroups(using(this), Name.Function, Text, Operator)), | |
| 895 (r'[^\S\n]+', Text), | |
| 896 (r'//.*?\n', Comment.Single), | |
| 897 (r'/\*.*?\*/', Comment.Multiline), | |
| 898 (r'@[a-zA-Z_][a-zA-Z0-9_\.]*', Name.Decorator), | |
| 899 (r'(assert|break|case|catch|continue|default|do|else|finally|for|' | |
| 900 r'if|goto|instanceof|new|return|switch|this|throw|try|while)\b', | |
| 901 Keyword), | |
| 902 (r'(abstract|const|enum|extends|final|implements|native|private|' | |
| 903 r'protected|public|static|strictfp|super|synchronized|throws|' | |
| 904 r'transient|volatile)\b', Keyword.Declaration), | |
| 905 (r'(boolean|byte|char|double|float|int|long|short|void)\b', | |
| 906 Keyword.Type), | |
| 907 (r'(package)(\s+)', bygroups(Keyword.Namespace, Text)), | |
| 908 (r'(true|false|null)\b', Keyword.Constant), | |
| 909 (r'(class|interface)(\s+)', bygroups(Keyword.Declaration, Text), 'class'), | |
| 910 (r'(import)(\s+)', bygroups(Keyword.Namespace, Text), 'import'), | |
| 911 (r'"(\\\\|\\"|[^"])*"', String), | |
| 912 (r"'\\.'|'[^\\]'|'\\u[0-9a-f]{4}'", String.Char), | |
| 913 (r'(\.)([a-zA-Z_][a-zA-Z0-9_]*)', bygroups(Operator, Name.Attribute)), | |
| 914 (r'[a-zA-Z_][a-zA-Z0-9_]*:', Name.Label), | |
| 915 (r'[a-zA-Z_\$][a-zA-Z0-9_]*', Name), | |
| 916 (r'[~\^\*!%&\[\]\(\)\{\}<>\|+=:;,./?-]', Operator), | |
| 917 (r'[0-9][0-9]*\.[0-9]+([eE][0-9]+)?[fd]?', Number.Float), | |
| 918 (r'0x[0-9a-f]+', Number.Hex), | |
| 919 (r'[0-9]+L?', Number.Integer), | |
| 920 (r'\n', Text) | |
| 921 ], | |
| 922 'class': [ | |
| 923 (r'[a-zA-Z_][a-zA-Z0-9_]*', Name.Class, '#pop') | |
| 924 ], | |
| 925 'import': [ | |
| 926 (r'[a-zA-Z0-9_.]+\*?', Name.Namespace, '#pop') | |
| 927 ], | |
| 928 } | |
| 929 | |
| 930 class ScalaLexer(RegexLexer): | |
| 931 """ | |
| 932 For `Scala <http://www.scala-lang.org>`_ source code. | |
| 933 """ | |
| 934 | |
| 935 name = 'Scala' | |
| 936 aliases = ['scala'] | |
| 937 filenames = ['*.scala'] | |
| 938 mimetypes = ['text/x-scala'] | |
| 939 | |
| 940 flags = re.MULTILINE | re.DOTALL | |
| 941 | |
| 942 #: optional Comment or Whitespace | |
| 943 _ws = r'(?:\s|//.*?\n|/[*].*?[*]/)+' | |
| 944 | |
| 945 # don't use raw unicode strings! | |
| 946 op = u'[-~\\^\\*!%&\\\\<>\\|+=:/?@\u00a6-\u00a7\u00a9\u00ac\u00ae\u00b0-\u00b1\u00b6\u00d7\u00f7\u03f6\u0482\u0606-\u0608\u060e-\u060f\u06e9\u06fd-\u06fe\u07f6\u09fa\u0b70\u0bf3-\u0bf8\u0bfa\u0c7f\u0cf1-\u0cf2\u0d79\u0f01-\u0f03\u0f13-\u0f17\u0f1a-\u0f1f\u0f34\u0f36\u0f38\u0fbe-\u0fc5\u0fc7-\u0fcf\u109e-\u109f\u1360\u1390-\u1399\u1940\u19e0-\u19ff\u1b61-\u1b6a\u1b74-\u1b7c\u2044\u2052\u207a-\u207c\u208a-\u208c\u2100-\u2101\u2103-\u2106\u2108-\u2109\u2114\u2116-\u2118\u211e-\u2123\u2125\u2127\u2129\u212e\u213a-\u213b\u2140-\u2144\u214a-\u214d\u214f\u2190-\u2328\u232b-\u244a\u249c-\u24e9\u2500-\u2767\u2794-\u27c4\u27c7-\u27e5\u27f0-\u2982\u2999-\u29d7\u29dc-\u29fb\u29fe-\u2b54\u2ce5-\u2cea\u2e80-\u2ffb\u3004\u3012-\u3013\u3020\u3036-\u3037\u303e-\u303f\u3190-\u3191\u3196-\u319f\u31c0-\u31e3\u3200-\u321e\u322a-\u3250\u3260-\u327f\u328a-\u32b0\u32c0-\u33ff\u4dc0-\u4dff\ua490-\ua4c6\ua828-\ua82b\ufb29\ufdfd\ufe62\ufe64-\ufe66\uff0b\uff1c-\uff1e\uff5c\uff5e\uffe2\uffe4\uffe8-\uffee\ufffc-\ufffd]+' | |
| 947 | |
| 948 letter = u'[a-zA-Z\\$_\u00aa\u00b5\u00ba\u00c0-\u00d6\u00d8-\u00f6\u00f8-\u02af\u0370-\u0373\u0376-\u0377\u037b-\u037d\u0386\u0388-\u03f5\u03f7-\u0481\u048a-\u0556\u0561-\u0587\u05d0-\u05f2\u0621-\u063f\u0641-\u064a\u066e-\u066f\u0671-\u06d3\u06d5\u06ee-\u06ef\u06fa-\u06fc\u06ff\u0710\u0712-\u072f\u074d-\u07a5\u07b1\u07ca-\u07ea\u0904-\u0939\u093d\u0950\u0958-\u0961\u0972-\u097f\u0985-\u09b9\u09bd\u09ce\u09dc-\u09e1\u09f0-\u09f1\u0a05-\u0a39\u0a59-\u0a5e\u0a72-\u0a74\u0a85-\u0ab9\u0abd\u0ad0-\u0ae1\u0b05-\u0b39\u0b3d\u0b5c-\u0b61\u0b71\u0b83-\u0bb9\u0bd0\u0c05-\u0c3d\u0c58-\u0c61\u0c85-\u0cb9\u0cbd\u0cde-\u0ce1\u0d05-\u0d3d\u0d60-\u0d61\u0d7a-\u0d7f\u0d85-\u0dc6\u0e01-\u0e30\u0e32-\u0e33\u0e40-\u0e45\u0e81-\u0eb0\u0eb2-\u0eb3\u0ebd-\u0ec4\u0edc-\u0f00\u0f40-\u0f6c\u0f88-\u0f8b\u1000-\u102a\u103f\u1050-\u1055\u105a-\u105d\u1061\u1065-\u1066\u106e-\u1070\u1075-\u1081\u108e\u10a0-\u10fa\u1100-\u135a\u1380-\u138f\u13a0-\u166c\u166f-\u1676\u1681-\u169a\u16a0-\u16ea\u16ee-\u1711\u1720-\u1731\u1740-\u1751\u1760-\u1770\u1780-\u17b3\u17dc\u1820-\u1842\u1844-\u18a8\u18aa-\u191c\u1950-\u19a9\u19c1-\u19c7\u1a00-\u1a16\u1b05-\u1b33\u1b45-\u1b4b\u1b83-\u1ba0\u1bae-\u1baf\u1c00-\u1c23\u1c4d-\u1c4f\u1c5a-\u1c77\u1d00-\u1d2b\u1d62-\u1d77\u1d79-\u1d9a\u1e00-\u1fbc\u1fbe\u1fc2-\u1fcc\u1fd0-\u1fdb\u1fe0-\u1fec\u1ff2-\u1ffc\u2071\u207f\u2102\u2107\u210a-\u2113\u2115\u2119-\u211d\u2124\u2126\u2128\u212a-\u212d\u212f-\u2139\u213c-\u213f\u2145-\u2149\u214e\u2160-\u2188\u2c00-\u2c7c\u2c80-\u2ce4\u2d00-\u2d65\u2d80-\u2dde\u3006-\u3007\u3021-\u3029\u3038-\u303a\u303c\u3041-\u3096\u309f\u30a1-\u30fa\u30ff-\u318e\u31a0-\u31b7\u31f0-\u31ff\u3400-\u4db5\u4e00-\ua014\ua016-\ua48c\ua500-\ua60b\ua610-\ua61f\ua62a-\ua66e\ua680-\ua697\ua722-\ua76f\ua771-\ua787\ua78b-\ua801\ua803-\ua805\ua807-\ua80a\ua80c-\ua822\ua840-\ua873\ua882-\ua8b3\ua90a-\ua925\ua930-\ua946\uaa00-\uaa28\uaa40-\uaa42\uaa44-\uaa4b\uac00-\ud7a3\uf900-\ufb1d\ufb1f-\ufb28\ufb2a-\ufd3d\ufd50-\ufdfb\ufe70-\ufefc\uff21-\uff3a\uff41-\uff5a\uff66-\uff6f\uff71-\uff9d\uffa0-\uffdc]' | |
| 949 | |
| 950 upper = u'[A-Z\\$_\u00c0-\u00d6\u00d8-\u00de\u0100\u0102\u0104\u0106\u0108\u010a\u010c\u010e\u0110\u0112\u0114\u0116\u0118\u011a\u011c\u011e\u0120\u0122\u0124\u0126\u0128\u012a\u012c\u012e\u0130\u0132\u0134\u0136\u0139\u013b\u013d\u013f\u0141\u0143\u0145\u0147\u014a\u014c\u014e\u0150\u0152\u0154\u0156\u0158\u015a\u015c\u015e\u0160\u0162\u0164\u0166\u0168\u016a\u016c\u016e\u0170\u0172\u0174\u0176\u0178-\u0179\u017b\u017d\u0181-\u0182\u0184\u0186-\u0187\u0189-\u018b\u018e-\u0191\u0193-\u0194\u0196-\u0198\u019c-\u019d\u019f-\u01a0\u01a2\u01a4\u01a6-\u01a7\u01a9\u01ac\u01ae-\u01af\u01b1-\u01b3\u01b5\u01b7-\u01b8\u01bc\u01c4\u01c7\u01ca\u01cd\u01cf\u01d1\u01d3\u01d5\u01d7\u01d9\u01db\u01de\u01e0\u01e2\u01e4\u01e6\u01e8\u01ea\u01ec\u01ee\u01f1\u01f4\u01f6-\u01f8\u01fa\u01fc\u01fe\u0200\u0202\u0204\u0206\u0208\u020a\u020c\u020e\u0210\u0212\u0214\u0216\u0218\u021a\u021c\u021e\u0220\u0222\u0224\u0226\u0228\u022a\u022c\u022e\u0230\u0232\u023a-\u023b\u023d-\u023e\u0241\u0243-\u0246\u0248\u024a\u024c\u024e\u0370\u0372\u0376\u0386\u0388-\u038f\u0391-\u03ab\u03cf\u03d2-\u03d4\u03d8\u03da\u03dc\u03de\u03e0\u03e2\u03e4\u03e6\u03e8\u03ea\u03ec\u03ee\u03f4\u03f7\u03f9-\u03fa\u03fd-\u042f\u0460\u0462\u0464\u0466\u0468\u046a\u046c\u046e\u0470\u0472\u0474\u0476\u0478\u047a\u047c\u047e\u0480\u048a\u048c\u048e\u0490\u0492\u0494\u0496\u0498\u049a\u049c\u049e\u04a0\u04a2\u04a4\u04a6\u04a8\u04aa\u04ac\u04ae\u04b0\u04b2\u04b4\u04b6\u04b8\u04ba\u04bc\u04be\u04c0-\u04c1\u04c3\u04c5\u04c7\u04c9\u04cb\u04cd\u04d0\u04d2\u04d4\u04d6\u04d8\u04da\u04dc\u04de\u04e0\u04e2\u04e4\u04e6\u04e8\u04ea\u04ec\u04ee\u04f0\u04f2\u04f4\u04f6\u04f8\u04fa\u04fc\u04fe\u0500\u0502\u0504\u0506\u0508\u050a\u050c\u050e\u0510\u0512\u0514\u0516\u0518\u051a\u051c\u051e\u0520\u0522\u0531-\u0556\u10a0-\u10c5\u1e00\u1e02\u1e04\u1e06\u1e08\u1e0a\u1e0c\u1e0e\u1e10\u1e12\u1e14\u1e16\u1e18\u1e1a\u1e1c\u1e1e\u1e20\u1e22\u1e24\u1e26\u1e28\u1e2a\u1e2c\u1e2e\u1e30\u1e32\u1e34\u1e36\u1e38\u1e3a\u1e3c\u1e3e\u1e40\u1e42\u1e44\u1e46\u1e48\u1e4a\u1e4c\u1e4e\u1e50\u1e52\u1e54\u1e56\u1e58\u1e5a\u1e5c\u1e5e\u1e60\u1e62\u1e64\u1e66\u1e68\u1e6a\u1e6c\u1e6e\u1e70\u1e72\u1e74\u1e76\u1e78\u1e7a\u1e7c\u1e7e\u1e80\u1e82\u1e84\u1e86\u1e88\u1e8a\u1e8c\u1e8e\u1e90\u1e92\u1e94\u1e9e\u1ea0\u1ea2\u1ea4\u1ea6\u1ea8\u1eaa\u1eac\u1eae\u1eb0\u1eb2\u1eb4\u1eb6\u1eb8\u1eba\u1ebc\u1ebe\u1ec0\u1ec2\u1ec4\u1ec6\u1ec8\u1eca\u1ecc\u1ece\u1ed0\u1ed2\u1ed4\u1ed6\u1ed8\u1eda\u1edc\u1ede\u1ee0\u1ee2\u1ee4\u1ee6\u1ee8\u1eea\u1eec\u1eee\u1ef0\u1ef2\u1ef4\u1ef6\u1ef8\u1efa\u1efc\u1efe\u1f08-\u1f0f\u1f18-\u1f1d\u1f28-\u1f2f\u1f38-\u1f3f\u1f48-\u1f4d\u1f59-\u1f5f\u1f68-\u1f6f\u1fb8-\u1fbb\u1fc8-\u1fcb\u1fd8-\u1fdb\u1fe8-\u1fec\u1ff8-\u1ffb\u2102\u2107\u210b-\u210d\u2110-\u2112\u2115\u2119-\u211d\u2124\u2126\u2128\u212a-\u212d\u2130-\u2133\u213e-\u213f\u2145\u2183\u2c00-\u2c2e\u2c60\u2c62-\u2c64\u2c67\u2c69\u2c6b\u2c6d-\u2c6f\u2c72\u2c75\u2c80\u2c82\u2c84\u2c86\u2c88\u2c8a\u2c8c\u2c8e\u2c90\u2c92\u2c94\u2c96\u2c98\u2c9a\u2c9c\u2c9e\u2ca0\u2ca2\u2ca4\u2ca6\u2ca8\u2caa\u2cac\u2cae\u2cb0\u2cb2\u2cb4\u2cb6\u2cb8\u2cba\u2cbc\u2cbe\u2cc0\u2cc2\u2cc4\u2cc6\u2cc8\u2cca\u2ccc\u2cce\u2cd0\u2cd2\u2cd4\u2cd6\u2cd8\u2cda\u2cdc\u2cde\u2ce0\u2ce2\ua640\ua642\ua644\ua646\ua648\ua64a\ua64c\ua64e\ua650\ua652\ua654\ua656\ua658\ua65a\ua65c\ua65e\ua662\ua664\ua666\ua668\ua66a\ua66c\ua680\ua682\ua684\ua686\ua688\ua68a\ua68c\ua68e\ua690\ua692\ua694\ua696\ua722\ua724\ua726\ua728\ua72a\ua72c\ua72e\ua732\ua734\ua736\ua738\ua73a\ua73c\ua73e\ua740\ua742\ua744\ua746\ua748\ua74a\ua74c\ua74e\ua750\ua752\ua754\ua756\ua758\ua75a\ua75c\ua75e\ua760\ua762\ua764\ua766\ua768\ua76a\ua76c\ua76e\ua779\ua77b\ua77d-\ua77e\ua780\ua782\ua784\ua786\ua78b\uff21-\uff3a]' | |
| 951 | |
| 952 idrest = ur'%s(?:%s|[0-9])*(?:(?<=_)%s)?' % (letter, letter, op) | |
| 953 | |
| 954 tokens = { | |
| 955 'root': [ | |
| 956 # method names | |
| 957 (r'(class|trait|object)(\s+)', bygroups(Keyword, Text), 'class'), | |
| 958 (ur"'%s" % idrest, Text.Symbol), | |
| 959 (r'[^\S\n]+', Text), | |
| 960 (r'//.*?\n', Comment.Single), | |
| 961 (r'/\*', Comment.Multiline, 'comment'), | |
| 962 (ur'@%s' % idrest, Name.Decorator), | |
| 963 (ur'(abstract|ca(?:se|tch)|d(?:ef|o)|e(?:lse|xtends)|' | |
| 964 ur'f(?:inal(?:ly)?|or(?:Some)?)|i(?:f|mplicit)|' | |
| 965 ur'lazy|match|new|override|pr(?:ivate|otected)' | |
| 966 ur'|re(?:quires|turn)|s(?:ealed|uper)|' | |
| 967 ur't(?:h(?:is|row)|ry)|va[lr]|w(?:hile|ith)|yield)\b|' | |
| 968 u'(<[%:-]|=>|>:|[#=@_\u21D2\u2190])(\b|(?=\\s)|$)', Keyword), | |
| 969 (ur':(?!%s)' % op, Keyword, 'type'), | |
| 970 (ur'%s%s\b' % (upper, idrest), Name.Class), | |
| 971 (r'(true|false|null)\b', Keyword.Constant), | |
| 972 (r'(import|package)(\s+)', bygroups(Keyword, Text), 'import'), | |
| 973 (r'(type)(\s+)', bygroups(Keyword, Text), 'type'), | |
| 974 (r'"""(?:.|\n)*?"""', String), | |
| 975 (r'"(\\\\|\\"|[^"])*"', String), | |
| 976 (ur"'\\.'|'[^\\]'|'\\u[0-9a-f]{4}'", String.Char), | |
| 977 # (ur'(\.)(%s|%s|`[^`]+`)' % (idrest, op), bygroups(Operator, | |
| 978 # Name.Attribute)), | |
| 979 (idrest, Name), | |
| 980 (r'`[^`]+`', Name), | |
| 981 (r'\[', Operator, 'typeparam'), | |
| 982 (r'[\(\)\{\};,.]', Operator), | |
| 983 (op, Operator), | |
| 984 (ur'([0-9][0-9]*\.[0-9]*|\.[0-9]+)([eE][+-]?[0-9]+)?[fFdD]?', | |
| 985 Number.Float), | |
| 986 (r'0x[0-9a-f]+', Number.Hex), | |
| 987 (r'[0-9]+L?', Number.Integer), | |
| 988 (r'\n', Text) | |
| 989 ], | |
| 990 'class': [ | |
| 991 (ur'(%s|%s|`[^`]+`)(\s*)(\[)' % (idrest, op), | |
| 992 bygroups(Name.Class, Text, Operator), 'typeparam'), | |
| 993 (r'[\s\n]+', Text), | |
| 994 (r'{', Operator, '#pop'), | |
| 995 (r'\(', Operator, '#pop'), | |
| 996 (ur'%s|%s|`[^`]+`' % (idrest, op), Name.Class, '#pop'), | |
| 997 ], | |
| 998 'type': [ | |
| 999 (r'\s+', Text), | |
| 1000 (u'<[%:]|>:|[#_\u21D2]|forSome|type', Keyword), | |
| 1001 (r'([,\);}]|=>|=)([\s\n]*)', bygroups(Operator, Text), '#pop'), | |
| 1002 (r'[\(\{]', Operator, '#push'), | |
| 1003 (ur'((?:%s|%s|`[^`]+`)(?:\.(?:%s|%s|`[^`]+`))*)(\s*)(\[)' % | |
| 1004 (idrest, op, idrest, op), | |
| 1005 bygroups(Keyword.Type, Text, Operator), ('#pop', 'typeparam')), | |
| 1006 (ur'((?:%s|%s|`[^`]+`)(?:\.(?:%s|%s|`[^`]+`))*)(\s*)$' % | |
| 1007 (idrest, op, idrest, op), | |
| 1008 bygroups(Keyword.Type, Text), '#pop'), | |
| 1009 (ur'\.|%s|%s|`[^`]+`' % (idrest, op), Keyword.Type) | |
| 1010 ], | |
| 1011 'typeparam': [ | |
| 1012 (r'[\s\n,]+', Text), | |
| 1013 (u'<[%:]|=>|>:|[#_\u21D2]|forSome|type', Keyword), | |
| 1014 (r'([\]\)\}])', Operator, '#pop'), | |
| 1015 (r'[\(\[\{]', Operator, '#push'), | |
| 1016 (ur'\.|%s|%s|`[^`]+`' % (idrest, op), Keyword.Type) | |
| 1017 ], | |
| 1018 'comment': [ | |
| 1019 (r'[^/\*]+', Comment.Multiline), | |
| 1020 (r'/\*', Comment.Multiline, '#push'), | |
| 1021 (r'\*/', Comment.Multiline, '#pop'), | |
| 1022 (r'[*/]', Comment.Multiline) | |
| 1023 ], | |
| 1024 'import': [ | |
| 1025 (ur'(%s|\.)+' % idrest, Name.Namespace, '#pop') | |
| 1026 ], | |
| 1027 } | |
| 1028 | |
| 1029 | |
| 1030 class DylanLexer(RegexLexer): | |
| 1031 """ | |
| 1032 For the `Dylan <http://www.opendylan.org/>`_ language. | |
| 1033 | |
| 1034 *New in Pygments 0.7.* | |
| 1035 """ | |
| 1036 | |
| 1037 name = 'Dylan' | |
| 1038 aliases = ['dylan'] | |
| 1039 filenames = ['*.dylan'] | |
| 1040 mimetypes = ['text/x-dylan'] | |
| 1041 | |
| 1042 flags = re.DOTALL | |
| 1043 | |
| 1044 tokens = { | |
| 1045 'root': [ | |
| 1046 (r'\b(subclass|abstract|block|c(on(crete|stant)|lass)|domain' | |
| 1047 r'|ex(c(eption|lude)|port)|f(unction(|al))|generic|handler' | |
| 1048 r'|i(n(herited|line|stance|terface)|mport)|library|m(acro|ethod)' | |
| 1049 r'|open|primary|sealed|si(deways|ngleton)|slot' | |
| 1050 r'|v(ariable|irtual))\b', Name.Builtin), | |
| 1051 (r'<\w+>', Keyword.Type), | |
| 1052 (r'#?"(?:\\.|[^"])+?"', String.Double), | |
| 1053 (r'//.*?\n', Comment.Single), | |
| 1054 (r'/\*[\w\W]*?\*/', Comment.Multiline), | |
| 1055 (r'\'.*?\'', String.Single), | |
| 1056 (r'=>|\b(a(bove|fterwards)|b(e(gin|low)|y)|c(ase|leanup|reate)' | |
| 1057 r'|define|else(|if)|end|f(inally|or|rom)|i[fn]|l(et|ocal)|otherwise' | |
| 1058 r'|rename|s(elect|ignal)|t(hen|o)|u(n(less|til)|se)|wh(en|ile))\b', | |
| 1059 Keyword), | |
| 1060 (r'([ \t])([!\$%&\*\/:<=>\?~_^a-zA-Z0-9.+\-]*:)', | |
| 1061 bygroups(Text, Name.Variable)), | |
| 1062 (r'([ \t]*)(\S+[^:])([ \t]*)(\()([ \t]*)', | |
| 1063 bygroups(Text, Name.Function, Text, Punctuation, Text)), | |
| 1064 (r'-?[0-9.]+', Number), | |
| 1065 (r'[(),;]', Punctuation), | |
| 1066 (r'\$[a-zA-Z0-9-]+', Name.Constant), | |
| 1067 (r'[!$%&*/:<>=?~^.+\[\]{}-]+', Operator), | |
| 1068 (r'\s+', Text), | |
| 1069 (r'#[a-zA-Z0-9-]+', Keyword), | |
| 1070 (r'[a-zA-Z0-9-]+', Name.Variable), | |
| 1071 ], | |
| 1072 } | |
| 1073 | |
| 1074 | |
| 1075 class ObjectiveCLexer(RegexLexer): | |
| 1076 """ | |
| 1077 For Objective-C source code with preprocessor directives. | |
| 1078 """ | |
| 1079 | |
| 1080 name = 'Objective-C' | |
| 1081 aliases = ['objective-c', 'objectivec', 'obj-c', 'objc'] | |
| 1082 #XXX: objc has .h files too :-/ | |
| 1083 filenames = ['*.m'] | |
| 1084 mimetypes = ['text/x-objective-c'] | |
| 1085 | |
| 1086 #: optional Comment or Whitespace | |
| 1087 _ws = r'(?:\s|//.*?\n|/[*].*?[*]/)+' | |
| 1088 | |
| 1089 tokens = { | |
| 1090 'whitespace': [ | |
| 1091 (r'^(\s*)(#if\s+0)', bygroups(Text, Comment.Preproc), 'if0'), | |
| 1092 (r'^(\s*)(#)', bygroups(Text, Comment.Preproc), 'macro'), | |
| 1093 (r'\n', Text), | |
| 1094 (r'\s+', Text), | |
| 1095 (r'\\\n', Text), # line continuation | |
| 1096 (r'//(\n|(.|\n)*?[^\\]\n)', Comment.Single), | |
| 1097 (r'/(\\\n)?[*](.|\n)*?[*](\\\n)?/', Comment.Multiline), | |
| 1098 ], | |
| 1099 'statements': [ | |
| 1100 (r'(L|@)?"', String, 'string'), | |
| 1101 (r"(L|@)?'(\\.|\\[0-7]{1,3}|\\x[a-fA-F0-9]{1,2}|[^\\\'\n])'", | |
| 1102 String.Char), | |
| 1103 (r'(\d+\.\d*|\.\d+|\d+)[eE][+-]?\d+[lL]?', Number.Float), | |
| 1104 (r'(\d+\.\d*|\.\d+|\d+[fF])[fF]?', Number.Float), | |
| 1105 (r'0x[0-9a-fA-F]+[Ll]?', Number.Hex), | |
| 1106 (r'0[0-7]+[Ll]?', Number.Oct), | |
| 1107 (r'\d+[Ll]?', Number.Integer), | |
| 1108 (r'[~!%^&*+=|?:<>/-]', Operator), | |
| 1109 (r'[()\[\],.]', Punctuation), | |
| 1110 (r'(auto|break|case|const|continue|default|do|else|enum|extern|' | |
| 1111 r'for|goto|if|register|restricted|return|sizeof|static|struct|' | |
| 1112 r'switch|typedef|union|volatile|virtual|while|in|@selector|' | |
| 1113 r'@private|@protected|@public|@encode|' | |
| 1114 r'@synchronized|@try|@throw|@catch|@finally|@end|@property|' | |
| 1115 r'@synthesize|@dynamic)\b', Keyword), | |
| 1116 (r'(int|long|float|short|double|char|unsigned|signed|void|' | |
| 1117 r'id|BOOL|IBOutlet|IBAction|SEL)\b', Keyword.Type), | |
| 1118 (r'(_{0,2}inline|naked|restrict|thread|typename)\b', | |
| 1119 Keyword.Reserved), | |
| 1120 (r'__(asm|int8|based|except|int16|stdcall|cdecl|fastcall|int32|' | |
| 1121 r'declspec|finally|int64|try|leave)\b', Keyword.Reserved), | |
| 1122 (r'(TRUE|FALSE|nil|NULL)\b', Name.Builtin), | |
| 1123 ('[a-zA-Z_][a-zA-Z0-9_]*:(?!:)', Name.Label), | |
| 1124 ('[a-zA-Z_][a-zA-Z0-9_]*', Name), | |
| 1125 ], | |
| 1126 'root': [ | |
| 1127 include('whitespace'), | |
| 1128 # functions | |
| 1129 (r'((?:[a-zA-Z0-9_*\s])+?(?:\s|[*]))' # return arguments | |
| 1130 r'([a-zA-Z_][a-zA-Z0-9_]*)' # method name | |
| 1131 r'(\s*\([^;]*?\))' # signature | |
| 1132 r'(' + _ws + r')({)', | |
| 1133 bygroups(using(this), Name.Function, | |
| 1134 using(this), Text, Punctuation), | |
| 1135 'function'), | |
| 1136 # function declarations | |
| 1137 (r'((?:[a-zA-Z0-9_*\s])+?(?:\s|[*]))' # return arguments | |
| 1138 r'([a-zA-Z_][a-zA-Z0-9_]*)' # method name | |
| 1139 r'(\s*\([^;]*?\))' # signature | |
| 1140 r'(' + _ws + r')(;)', | |
| 1141 bygroups(using(this), Name.Function, | |
| 1142 using(this), Text, Punctuation)), | |
| 1143 (r'(@interface|@implementation)(\s+)', bygroups(Keyword, Text), | |
| 1144 'classname'), | |
| 1145 (r'(@class|@protocol)(\s+)', bygroups(Keyword, Text), | |
| 1146 'forward_classname'), | |
| 1147 (r'(\s*)(@end)(\s*)', bygroups(Text, Keyword, Text)), | |
| 1148 ('', Text, 'statement'), | |
| 1149 ], | |
| 1150 'classname' : [ | |
| 1151 # interface definition that inherits | |
| 1152 ('([a-zA-Z_][a-zA-Z0-9_]*)(\s*:\s*)([a-zA-Z_][a-zA-Z0-9_]*)?', | |
| 1153 bygroups(Name.Class, Text, Name.Class), '#pop'), | |
| 1154 # interface definition for a category | |
| 1155 ('([a-zA-Z_][a-zA-Z0-9_]*)(\s*)(\([a-zA-Z_][a-zA-Z0-9_]*\))', | |
| 1156 bygroups(Name.Class, Text, Name.Label), '#pop'), | |
| 1157 # simple interface / implementation | |
| 1158 ('([a-zA-Z_][a-zA-Z0-9_]*)', Name.Class, '#pop') | |
| 1159 ], | |
| 1160 'forward_classname' : [ | |
| 1161 ('([a-zA-Z_][a-zA-Z0-9_]*)(\s*,\s*)', | |
| 1162 bygroups(Name.Class, Text), 'forward_classname'), | |
| 1163 ('([a-zA-Z_][a-zA-Z0-9_]*)(\s*;?)', | |
| 1164 bygroups(Name.Class, Text), '#pop') | |
| 1165 ], | |
| 1166 'statement' : [ | |
| 1167 include('whitespace'), | |
| 1168 include('statements'), | |
| 1169 ('[{}]', Punctuation), | |
| 1170 (';', Punctuation, '#pop'), | |
| 1171 ], | |
| 1172 'function': [ | |
| 1173 include('whitespace'), | |
| 1174 include('statements'), | |
| 1175 (';', Punctuation), | |
| 1176 ('{', Punctuation, '#push'), | |
| 1177 ('}', Punctuation, '#pop'), | |
| 1178 ], | |
| 1179 'string': [ | |
| 1180 (r'"', String, '#pop'), | |
| 1181 (r'\\([\\abfnrtv"\']|x[a-fA-F0-9]{2,4}|[0-7]{1,3})', String.Escape), | |
| 1182 (r'[^\\"\n]+', String), # all other characters | |
| 1183 (r'\\\n', String), # line continuation | |
| 1184 (r'\\', String), # stray backslash | |
| 1185 ], | |
| 1186 'macro': [ | |
| 1187 (r'[^/\n]+', Comment.Preproc), | |
| 1188 (r'/[*](.|\n)*?[*]/', Comment.Multiline), | |
| 1189 (r'//.*?\n', Comment.Single, '#pop'), | |
| 1190 (r'/', Comment.Preproc), | |
| 1191 (r'(?<=\\)\n', Comment.Preproc), | |
| 1192 (r'\n', Comment.Preproc, '#pop'), | |
| 1193 ], | |
| 1194 'if0': [ | |
| 1195 (r'^\s*#if.*?(?<!\\)\n', Comment.Preproc, '#push'), | |
| 1196 (r'^\s*#endif.*?(?<!\\)\n', Comment.Preproc, '#pop'), | |
| 1197 (r'.*?\n', Comment), | |
| 1198 ] | |
| 1199 } | |
| 1200 | |
| 1201 def analyse_text(text): | |
| 1202 if '@"' in text: # strings | |
| 1203 return True | |
| 1204 if re.match(r'\[[a-zA-Z0-9.]:', text): # message | |
| 1205 return True | |
| 1206 return False | |
| 1207 | |
| 1208 class FortranLexer(RegexLexer): | |
| 1209 ''' | |
| 1210 Lexer for FORTRAN 90 code. | |
| 1211 | |
| 1212 *New in Pygments 0.10.* | |
| 1213 ''' | |
| 1214 name = 'Fortran' | |
| 1215 aliases = ['fortran'] | |
| 1216 filenames = ['*.f', '*.f90'] | |
| 1217 mimetypes = ['text/x-fortran'] | |
| 1218 flags = re.IGNORECASE | |
| 1219 | |
| 1220 # Data Types: INTEGER, REAL, COMPLEX, LOGICAL, CHARACTER and DOUBLE PRECISION | |
| 1221 # Operators: **, *, +, -, /, <, >, <=, >=, ==, /= | |
| 1222 # Logical (?): NOT, AND, OR, EQV, NEQV | |
| 1223 | |
| 1224 # Builtins: | |
| 1225 # http://gcc.gnu.org/onlinedocs/gcc-3.4.6/g77/Table-of-Intrinsic-Functions.html | |
| 1226 | |
| 1227 tokens = { | |
| 1228 'root': [ | |
| 1229 (r'!.*\n', Comment), | |
| 1230 include('strings'), | |
| 1231 include('core'), | |
| 1232 (r'[a-z][a-z0-9_]*', Name.Variable), | |
| 1233 include('nums'), | |
| 1234 (r'[\s]+', Text), | |
| 1235 ], | |
| 1236 'core': [ | |
| 1237 # Statements | |
| 1238 (r'\b(ACCEPT|ALLOCATABLE|ALLOCATE|ARRAY|ASSIGN|BACKSPACE|BLOCK DATA|' | |
| 1239 r'BYTE|CALL|CASE|CLOSE|COMMON|CONTAINS|CONTINUE|CYCLE|DATA|' | |
| 1240 r'DEALLOCATE|DECODE|DIMENSION|DO|ENCODE|END FILE|ENDIF|END|ENTRY|' | |
| 1241 r'EQUIVALENCE|EXIT|EXTERNAL|EXTRINSIC|FORALL|FORMAT|FUNCTION|GOTO|' | |
| 1242 r'IF|IMPLICIT|INCLUDE|INQUIRE|INTENT|INTERFACE|INTRINSIC|MODULE|' | |
| 1243 r'NAMELIST|NULLIFY|NONE|OPEN|OPTIONAL|OPTIONS|PARAMETER|PAUSE|' | |
| 1244 r'POINTER|PRINT|PRIVATE|PROGRAM|PUBLIC|PURE|READ|RECURSIVE|RETURN|' | |
| 1245 r'REWIND|SAVE|SELECT|SEQUENCE|STOP|SUBROUTINE|TARGET|TYPE|USE|' | |
| 1246 r'VOLATILE|WHERE|WRITE|WHILE|THEN|ELSE|ENDIF)\s*\b', | |
| 1247 Keyword), | |
| 1248 | |
| 1249 # Data Types | |
| 1250 (r'\b(CHARACTER|COMPLEX|DOUBLE PRECISION|DOUBLE COMPLEX|INTEGER|' | |
| 1251 r'LOGICAL|REAL)\s*\b', | |
| 1252 Keyword.Type), | |
| 1253 | |
| 1254 # Operators | |
| 1255 (r'(\*\*|\*|\+|-|\/|<|>|<=|>=|==|\/=|=)', Operator), | |
| 1256 | |
| 1257 (r'(::)', Keyword.Declaration), | |
| 1258 | |
| 1259 (r'[(),:&%;]', Punctuation), | |
| 1260 | |
| 1261 # Intrinsics | |
| 1262 (r'\b(Abort|Abs|Access|AChar|ACos|AdjustL|AdjustR|AImag|AInt|Alarm|' | |
| 1263 r'All|Allocated|ALog|AMax|AMin|AMod|And|ANInt|Any|' | |
| 1264 r'ASin|Associated|ATan|BesJ|BesJN|BesY|BesYN|' | |
| 1265 r'Bit_Size|BTest|CAbs|CCos|Ceiling|CExp|Char|ChDir|ChMod|CLog|' | |
| 1266 r'Cmplx|Complex|Conjg|Cos|CosH|Count|CPU_Time|CShift|CSin|CSqRt|' | |
| 1267 r'CTime|DAbs|DACos|DASin|DATan|Date_and_Time|DbesJ|' | |
| 1268 r'DbesJ|DbesJN|DbesY|DbesY|DbesYN|Dble|DCos|DCosH|DDiM|DErF|DErFC|' | |
| 1269 r'DExp|Digits|DiM|DInt|DLog|DLog|DMax|DMin|DMod|DNInt|Dot_Product|' | |
| 1270 r'DProd|DSign|DSinH|DSin|DSqRt|DTanH|DTan|DTime|EOShift|Epsilon|' | |
| 1271 r'ErF|ErFC|ETime|Exit|Exp|Exponent|FDate|FGet|FGetC|Float|' | |
| 1272 r'Floor|Flush|FNum|FPutC|FPut|Fraction|FSeek|FStat|FTell|' | |
| 1273 r'GError|GetArg|GetCWD|GetEnv|GetGId|GetLog|GetPId|GetUId|' | |
| 1274 r'GMTime|HostNm|Huge|IAbs|IAChar|IAnd|IArgC|IBClr|IBits|' | |
| 1275 r'IBSet|IChar|IDate|IDiM|IDInt|IDNInt|IEOr|IErrNo|IFix|Imag|' | |
| 1276 r'ImagPart|Index|Int|IOr|IRand|IsaTty|IShft|IShftC|ISign|' | |
| 1277 r'ITime|Kill|Kind|LBound|Len|Len_Trim|LGe|LGt|Link|LLe|LLt|LnBlnk|' | |
| 1278 r'Loc|Log|Log|Logical|Long|LShift|LStat|LTime|MatMul|Max|' | |
| 1279 r'MaxExponent|MaxLoc|MaxVal|MClock|Merge|Min|MinExponent|MinLoc|' | |
| 1280 r'MinVal|Mod|Modulo|MvBits|Nearest|NInt|Not|Or|Pack|PError|' | |
| 1281 r'Precision|Present|Product|Radix|Rand|Random_Number|Random_Seed|' | |
| 1282 r'Range|Real|RealPart|Rename|Repeat|Reshape|RRSpacing|RShift|Scale|' | |
| 1283 r'Scan|Second|Selected_Int_Kind|Selected_Real_Kind|Set_Exponent|' | |
| 1284 r'Shape|Short|Sign|Signal|SinH|Sin|Sleep|Sngl|Spacing|Spread|SqRt|' | |
| 1285 r'SRand|Stat|Sum|SymLnk|System|System_Clock|Tan|TanH|Time|' | |
| 1286 r'Tiny|Transfer|Transpose|Trim|TtyNam|UBound|UMask|Unlink|Unpack|' | |
| 1287 r'Verify|XOr|ZAbs|ZCos|ZExp|ZLog|ZSin|ZSqRt)\s*\b', | |
| 1288 Name.Builtin), | |
| 1289 | |
| 1290 # Booleans | |
| 1291 (r'\.(true|false)\.', Name.Builtin), | |
| 1292 # Comparing Operators | |
| 1293 (r'\.(eq|ne|lt|le|gt|ge|not|and|or|eqv|neqv)\.', Operator.Word), | |
| 1294 ], | |
| 1295 | |
| 1296 'strings': [ | |
| 1297 (r'"(\\\\|\\[0-7]+|\\.|[^"])*"', String.Double), | |
| 1298 (r"'(\\\\|\\[0-7]+|\\.|[^'])*'", String.Single), | |
| 1299 ], | |
| 1300 | |
| 1301 'nums': [ | |
| 1302 (r'\d+(?![.Ee])', Number.Integer), | |
| 1303 (r'[+-]?\d*\.\d+([eE][-+]?\d+)?', Number.Float), | |
| 1304 (r'[+-]?\d+\.\d*([eE][-+]?\d+)?', Number.Float), | |
| 1305 ], | |
| 1306 } | |
| 1307 | |
| 1308 | |
| 1309 class GLShaderLexer(RegexLexer): | |
| 1310 """ | |
| 1311 GLSL (OpenGL Shader) lexer. | |
| 1312 | |
| 1313 *New in Pygments 1.1.* | |
| 1314 """ | |
| 1315 name = 'GLSL' | |
| 1316 aliases = ['glsl'] | |
| 1317 filenames = ['*.vert', '*.frag', '*.geo'] | |
| 1318 mimetypes = ['text/x-glslsrc'] | |
| 1319 | |
| 1320 tokens = { | |
| 1321 'root': [ | |
| 1322 (r'^#.*', Comment.Preproc), | |
| 1323 (r'//.*', Comment.Single), | |
| 1324 (r'/\*[\w\W]*\*/', Comment.Multiline), | |
| 1325 (r'\+|-|~|!=?|\*|/|%|<<|>>|<=?|>=?|==?|&&?|\^|\|\|?', | |
| 1326 Operator), | |
| 1327 (r'[?:]', Operator), # quick hack for ternary | |
| 1328 (r'\bdefined\b', Operator), | |
| 1329 (r'[;{}(),\[\]]', Punctuation), | |
| 1330 #FIXME when e is present, no decimal point needed | |
| 1331 (r'[+-]?\d*\.\d+([eE][-+]?\d+)?', Number.Float), | |
| 1332 (r'[+-]?\d+\.\d*([eE][-+]?\d+)?', Number.Float), | |
| 1333 (r'0[xX][0-9a-fA-F]*', Number.Hex), | |
| 1334 (r'0[0-7]*', Number.Octal), | |
| 1335 (r'[1-9][0-9]*', Number.Integer), | |
| 1336 (r'\b(attribute|const|uniform|varying|centroid|break|continue|' | |
| 1337 r'do|for|while|if|else|in|out|inout|float|int|void|bool|true|' | |
| 1338 r'false|invariant|discard|return|mat[234]|mat[234]x[234]|' | |
| 1339 r'vec[234]|[ib]vec[234]|sampler[123]D|samplerCube|' | |
| 1340 r'sampler[12]DShadow|struct)\b', Keyword), | |
| 1341 (r'\b(asm|class|union|enum|typedef|template|this|packed|goto|' | |
| 1342 r'switch|default|inline|noinline|volatile|public|static|extern|' | |
| 1343 r'external|interface|long|short|double|half|fixed|unsigned|' | |
| 1344 r'lowp|mediump|highp|precision|input|output|hvec[234]|' | |
| 1345 r'[df]vec[234]|sampler[23]DRect|sampler2DRectShadow|sizeof|' | |
| 1346 r'cast|namespace|using)\b', Keyword), #future use | |
| 1347 (r'[a-zA-Z_][a-zA-Z_0-9]*', Name.Variable), | |
| 1348 (r'\.', Punctuation), | |
| 1349 (r'\s+', Text), | |
| 1350 ], | |
| 1351 } | |
| 1352 | |
| 1353 class PrologLexer(RegexLexer): | |
| 1354 """ | |
| 1355 Lexer for Prolog files. | |
| 1356 """ | |
| 1357 name = 'Prolog' | |
| 1358 aliases = ['prolog'] | |
| 1359 filenames = ['*.prolog', '*.pro', '*.pl'] | |
| 1360 mimetypes = ['text/x-prolog'] | |
| 1361 | |
| 1362 flags = re.UNICODE | |
| 1363 | |
| 1364 tokens = { | |
| 1365 'root': [ | |
| 1366 (r'^#.*', Comment.Single), | |
| 1367 (r'/\*', Comment.Multiline, 'nested-comment'), | |
| 1368 (r'%.*', Comment.Single), | |
| 1369 (r'[0-9]+', Number), | |
| 1370 (r'[\[\](){}|.,;!]', Punctuation), | |
| 1371 (r':-|-->', Punctuation), | |
| 1372 (r'"(?:\\x[0-9a-fA-F]+\\|\\u[0-9a-fA-F]{4}|\U[0-9a-fA-F]{8}|' | |
| 1373 r'\\[0-7]+\\|\\[\w\W]|[^"])*"', String.Double), | |
| 1374 (r"'(?:''|[^'])*'", String.Atom), # quoted atom | |
| 1375 # Needs to not be followed by an atom. | |
| 1376 #(r'=(?=\s|[a-zA-Z\[])', Operator), | |
| 1377 (r'(is|<|>|=<|>=|==|=:=|=|/|//|\*|\+|-)(?=\s|[a-zA-Z0-9\[])', | |
| 1378 Operator), | |
| 1379 (r'(mod|div|not)\b', Operator), | |
| 1380 (r'_', Keyword), # The don't-care variable | |
| 1381 (r'([a-z]+)(:)', bygroups(Name.Namespace, Punctuation)), | |
| 1382 (u'([a-z\u00c0-\u1fff\u3040-\ud7ff\ue000-\uffef]' | |
| 1383 u'[a-zA-Z0-9_$\u00c0-\u1fff\u3040-\ud7ff\ue000-\uffef]*)' | |
| 1384 u'(\\s*)(:-|-->)', | |
| 1385 bygroups(Name.Function, Text, Operator)), # function defn | |
| 1386 (u'([a-z\u00c0-\u1fff\u3040-\ud7ff\ue000-\uffef]' | |
| 1387 u'[a-zA-Z0-9_$\u00c0-\u1fff\u3040-\ud7ff\ue000-\uffef]*)' | |
| 1388 u'(\\s*)(\\()', | |
| 1389 bygroups(Name.Function, Text, Punctuation)), | |
| 1390 (u'[a-z\u00c0-\u1fff\u3040-\ud7ff\ue000-\uffef]' | |
| 1391 u'[a-zA-Z0-9_$\u00c0-\u1fff\u3040-\ud7ff\ue000-\uffef]*', | |
| 1392 String.Atom), # atom, characters | |
| 1393 # This one includes ! | |
| 1394 (u'[#&*+\\-./:<=>?@\\\\^~\u00a1-\u00bf\u2010-\u303f]+', | |
| 1395 String.Atom), # atom, graphics | |
| 1396 (r'[A-Z_][A-Za-z0-9_]*', Name.Variable), | |
| 1397 (u'\\s+|[\u2000-\u200f\ufff0-\ufffe\uffef]', Text), | |
| 1398 ], | |
| 1399 'nested-comment': [ | |
| 1400 (r'\*/', Comment.Multiline, '#pop'), | |
| 1401 (r'/\*', Comment.Multiline, '#push'), | |
| 1402 (r'[^*/]+', Comment.Multiline), | |
| 1403 (r'[*/]', Comment.Multiline), | |
| 1404 ], | |
| 1405 } | |
| 1406 | |
| 1407 def analyse_text(text): | |
| 1408 return ':-' in text | |
| 1409 | |
| 1410 | |
| 1411 class CythonLexer(RegexLexer): | |
| 1412 """ | |
| 1413 For Pyrex and `Cython <http://cython.org>`_ source code. | |
| 1414 | |
| 1415 *New in Pygments 1.1.* | |
| 1416 """ | |
| 1417 | |
| 1418 name = 'Cython' | |
| 1419 aliases = ['cython', 'pyx'] | |
| 1420 filenames = ['*.pyx', '*.pxd', '*.pxi'] | |
| 1421 mimetypes = ['text/x-cython', 'application/x-cython'] | |
| 1422 | |
| 1423 tokens = { | |
| 1424 'root': [ | |
| 1425 (r'\n', Text), | |
| 1426 (r'^(\s*)("""(?:.|\n)*?""")', bygroups(Text, String.Doc)), | |
| 1427 (r"^(\s*)('''(?:.|\n)*?''')", bygroups(Text, String.Doc)), | |
| 1428 (r'[^\S\n]+', Text), | |
| 1429 (r'#.*$', Comment), | |
| 1430 (r'[]{}:(),;[]', Punctuation), | |
| 1431 (r'\\\n', Text), | |
| 1432 (r'\\', Text), | |
| 1433 (r'(in|is|and|or|not)\b', Operator.Word), | |
| 1434 (r'(<)([a-zA-Z0-9.?]+)(>)', | |
| 1435 bygroups(Punctuation, Keyword.Type, Punctuation)), | |
| 1436 (r'!=|==|<<|>>|[-~+/*%=<>&^|.?]', Operator), | |
| 1437 (r'(from)(\d+)(<=)(\s+)(<)(\d+)(:)', | |
| 1438 bygroups(Keyword, Number.Integer, Operator, Name, Operator, | |
| 1439 Name, Punctuation)), | |
| 1440 include('keywords'), | |
| 1441 (r'(def|property)(\s+)', bygroups(Keyword, Text), 'funcname'), | |
| 1442 (r'(cp?def)(\s+)', bygroups(Keyword, Text), 'cdef'), | |
| 1443 (r'(class|struct)(\s+)', bygroups(Keyword, Text), 'classname'), | |
| 1444 (r'(from)(\s+)', bygroups(Keyword, Text), 'fromimport'), | |
| 1445 (r'(c?import)(\s+)', bygroups(Keyword, Text), 'import'), | |
| 1446 include('builtins'), | |
| 1447 include('backtick'), | |
| 1448 ('(?:[rR]|[uU][rR]|[rR][uU])"""', String, 'tdqs'), | |
| 1449 ("(?:[rR]|[uU][rR]|[rR][uU])'''", String, 'tsqs'), | |
| 1450 ('(?:[rR]|[uU][rR]|[rR][uU])"', String, 'dqs'), | |
| 1451 ("(?:[rR]|[uU][rR]|[rR][uU])'", String, 'sqs'), | |
| 1452 ('[uU]?"""', String, combined('stringescape', 'tdqs')), | |
| 1453 ("[uU]?'''", String, combined('stringescape', 'tsqs')), | |
| 1454 ('[uU]?"', String, combined('stringescape', 'dqs')), | |
| 1455 ("[uU]?'", String, combined('stringescape', 'sqs')), | |
| 1456 include('name'), | |
| 1457 include('numbers'), | |
| 1458 ], | |
| 1459 'keywords': [ | |
| 1460 (r'(assert|break|by|continue|ctypedef|del|elif|else|except\??|exec|' | |
| 1461 r'finally|for|gil|global|if|include|lambda|nogil|pass|print|raise|' | |
| 1462 r'return|try|while|yield|as|with)\b', Keyword), | |
| 1463 (r'(DEF|IF|ELIF|ELSE)\b', Comment.Preproc), | |
| 1464 ], | |
| 1465 'builtins': [ | |
| 1466 (r'(?<!\.)(__import__|abs|all|any|apply|basestring|bin|bool|buffer|' | |
| 1467 r'bytearray|bytes|callable|chr|classmethod|cmp|coerce|compile|' | |
| 1468 r'complex|delattr|dict|dir|divmod|enumerate|eval|execfile|exit|' | |
| 1469 r'file|filter|float|frozenset|getattr|globals|hasattr|hash|hex|id|' | |
| 1470 r'input|int|intern|isinstance|issubclass|iter|len|list|locals|' | |
| 1471 r'long|map|max|min|next|object|oct|open|ord|pow|property|range|' | |
| 1472 r'raw_input|reduce|reload|repr|reversed|round|set|setattr|slice|' | |
| 1473 r'sorted|staticmethod|str|sum|super|tuple|type|unichr|unicode|' | |
| 1474 r'vars|xrange|zip)\b', Name.Builtin), | |
| 1475 (r'(?<!\.)(self|None|Ellipsis|NotImplemented|False|True|NULL' | |
| 1476 r')\b', Name.Builtin.Pseudo), | |
| 1477 (r'(?<!\.)(ArithmeticError|AssertionError|AttributeError|' | |
| 1478 r'BaseException|DeprecationWarning|EOFError|EnvironmentError|' | |
| 1479 r'Exception|FloatingPointError|FutureWarning|GeneratorExit|IOError|' | |
| 1480 r'ImportError|ImportWarning|IndentationError|IndexError|KeyError|' | |
| 1481 r'KeyboardInterrupt|LookupError|MemoryError|NameError|' | |
| 1482 r'NotImplemented|NotImplementedError|OSError|OverflowError|' | |
| 1483 r'OverflowWarning|PendingDeprecationWarning|ReferenceError|' | |
| 1484 r'RuntimeError|RuntimeWarning|StandardError|StopIteration|' | |
| 1485 r'SyntaxError|SyntaxWarning|SystemError|SystemExit|TabError|' | |
| 1486 r'TypeError|UnboundLocalError|UnicodeDecodeError|' | |
| 1487 r'UnicodeEncodeError|UnicodeError|UnicodeTranslateError|' | |
| 1488 r'UnicodeWarning|UserWarning|ValueError|Warning|ZeroDivisionError' | |
| 1489 r')\b', Name.Exception), | |
| 1490 ], | |
| 1491 'numbers': [ | |
| 1492 (r'(\d+\.?\d*|\d*\.\d+)([eE][+-]?[0-9]+)?', Number.Float), | |
| 1493 (r'0\d+', Number.Oct), | |
| 1494 (r'0[xX][a-fA-F0-9]+', Number.Hex), | |
| 1495 (r'\d+L', Number.Integer.Long), | |
| 1496 (r'\d+', Number.Integer) | |
| 1497 ], | |
| 1498 'backtick': [ | |
| 1499 ('`.*?`', String.Backtick), | |
| 1500 ], | |
| 1501 'name': [ | |
| 1502 (r'@[a-zA-Z0-9_]+', Name.Decorator), | |
| 1503 ('[a-zA-Z_][a-zA-Z0-9_]*', Name), | |
| 1504 ], | |
| 1505 'funcname': [ | |
| 1506 ('[a-zA-Z_][a-zA-Z0-9_]*', Name.Function, '#pop') | |
| 1507 ], | |
| 1508 'cdef': [ | |
| 1509 (r'(public|readonly|extern|api|inline)\b', Keyword.Reserved), | |
| 1510 (r'(struct|enum|union|class)\b', Keyword), | |
| 1511 (r'([a-zA-Z_][a-zA-Z0-9_]*)(\s*)(?=[(:#=]|$)', | |
| 1512 bygroups(Name.Function, Text), '#pop'), | |
| 1513 (r'([a-zA-Z_][a-zA-Z0-9_]*)(\s*)(,)', | |
| 1514 bygroups(Name.Function, Text, Punctuation)), | |
| 1515 (r'from\b', Keyword, '#pop'), | |
| 1516 (r'as\b', Keyword), | |
| 1517 (r':', Punctuation, '#pop'), | |
| 1518 (r'(?=["\'])', Text, '#pop'), | |
| 1519 (r'[a-zA-Z_][a-zA-Z0-9_]*', Keyword.Type), | |
| 1520 (r'.', Text), | |
| 1521 ], | |
| 1522 'classname': [ | |
| 1523 ('[a-zA-Z_][a-zA-Z0-9_]*', Name.Class, '#pop') | |
| 1524 ], | |
| 1525 'import': [ | |
| 1526 (r'(\s+)(as)(\s+)', bygroups(Text, Keyword, Text)), | |
| 1527 (r'[a-zA-Z_][a-zA-Z0-9_.]*', Name.Namespace), | |
| 1528 (r'(\s*)(,)(\s*)', bygroups(Text, Operator, Text)), | |
| 1529 (r'', Text, '#pop') # all else: go back | |
| 1530 ], | |
| 1531 'fromimport': [ | |
| 1532 (r'(\s+)(c?import)\b', bygroups(Text, Keyword), '#pop'), | |
| 1533 (r'[a-zA-Z_.][a-zA-Z0-9_.]*', Name.Namespace), | |
| 1534 # ``cdef foo from "header"``, or ``for foo from 0 < i < 10`` | |
| 1535 (r'', Text, '#pop'), | |
| 1536 ], | |
| 1537 'stringescape': [ | |
| 1538 (r'\\([\\abfnrtv"\']|\n|N{.*?}|u[a-fA-F0-9]{4}|' | |
| 1539 r'U[a-fA-F0-9]{8}|x[a-fA-F0-9]{2}|[0-7]{1,3})', String.Escape) | |
| 1540 ], | |
| 1541 'strings': [ | |
| 1542 (r'%(\([a-zA-Z0-9]+\))?[-#0 +]*([0-9]+|[*])?(\.([0-9]+|[*]))?' | |
| 1543 '[hlL]?[diouxXeEfFgGcrs%]', String.Interpol), | |
| 1544 (r'[^\\\'"%\n]+', String), | |
| 1545 # quotes, percents and backslashes must be parsed one at a time | |
| 1546 (r'[\'"\\]', String), | |
| 1547 # unhandled string formatting sign | |
| 1548 (r'%', String) | |
| 1549 # newlines are an error (use "nl" state) | |
| 1550 ], | |
| 1551 'nl': [ | |
| 1552 (r'\n', String) | |
| 1553 ], | |
| 1554 'dqs': [ | |
| 1555 (r'"', String, '#pop'), | |
| 1556 (r'\\\\|\\"|\\\n', String.Escape), # included here again for raw strings | |
| 1557 include('strings') | |
| 1558 ], | |
| 1559 'sqs': [ | |
| 1560 (r"'", String, '#pop'), | |
| 1561 (r"\\\\|\\'|\\\n", String.Escape), # included here again for raw strings | |
| 1562 include('strings') | |
| 1563 ], | |
| 1564 'tdqs': [ | |
| 1565 (r'"""', String, '#pop'), | |
| 1566 include('strings'), | |
| 1567 include('nl') | |
| 1568 ], | |
| 1569 'tsqs': [ | |
| 1570 (r"'''", String, '#pop'), | |
| 1571 include('strings'), | |
| 1572 include('nl') | |
| 1573 ], | |
| 1574 } | |
| 1575 | |
| 1576 | |
| 1577 class ValaLexer(RegexLexer): | |
| 1578 """ | |
| 1579 For Vala source code with preprocessor directives. | |
| 1580 | |
| 1581 *New in Pygments 1.1.* | |
| 1582 """ | |
| 1583 name = 'Vala' | |
| 1584 aliases = ['vala', 'vapi'] | |
| 1585 filenames = ['*.vala', '*.vapi'] | |
| 1586 mimetypes = ['text/x-vala'] | |
| 1587 | |
| 1588 tokens = { | |
| 1589 'whitespace': [ | |
| 1590 (r'^\s*#if\s+0', Comment.Preproc, 'if0'), | |
| 1591 (r'\n', Text), | |
| 1592 (r'\s+', Text), | |
| 1593 (r'\\\n', Text), # line continuation | |
| 1594 (r'//(\n|(.|\n)*?[^\\]\n)', Comment.Single), | |
| 1595 (r'/(\\\n)?[*](.|\n)*?[*](\\\n)?/', Comment.Multiline), | |
| 1596 ], | |
| 1597 'statements': [ | |
| 1598 (r'L?"', String, 'string'), | |
| 1599 (r"L?'(\\.|\\[0-7]{1,3}|\\x[a-fA-F0-9]{1,2}|[^\\\'\n])'", | |
| 1600 String.Char), | |
| 1601 (r'(\d+\.\d*|\.\d+|\d+)[eE][+-]?\d+[lL]?', Number.Float), | |
| 1602 (r'(\d+\.\d*|\.\d+|\d+[fF])[fF]?', Number.Float), | |
| 1603 (r'0x[0-9a-fA-F]+[Ll]?', Number.Hex), | |
| 1604 (r'0[0-7]+[Ll]?', Number.Oct), | |
| 1605 (r'\d+[Ll]?', Number.Integer), | |
| 1606 (r'[~!%^&*+=|?:<>/-]', Operator), | |
| 1607 (r'(\[)(Compact|Immutable|(?:Boolean|Simple)Type)(\])', | |
| 1608 bygroups(Punctuation, Name.Decorator, Punctuation)), | |
| 1609 # TODO: "correctly" parse complex code attributes | |
| 1610 (r'(\[)(CCode|(?:Integer|Floating)Type)', | |
| 1611 bygroups(Punctuation, Name.Decorator)), | |
| 1612 (r'[()\[\],.]', Punctuation), | |
| 1613 (r'(as|base|break|case|catch|construct|continue|default|delete|do|' | |
| 1614 r'else|enum|finally|for|foreach|get|if|in|is|lock|new|out|params|' | |
| 1615 r'return|set|sizeof|switch|this|throw|try|typeof|while|yield)\b', | |
| 1616 Keyword), | |
| 1617 (r'(abstract|const|delegate|dynamic|ensures|extern|inline|internal|' | |
| 1618 r'override|owned|private|protected|public|ref|requires|signal|' | |
| 1619 r'static|throws|unowned|var|virtual|volatile|weak|yields)\b', | |
| 1620 Keyword.Declaration), | |
| 1621 (r'(namespace|using)(\s+)', bygroups(Keyword.Namespace, Text), | |
| 1622 'namespace'), | |
| 1623 (r'(class|errordomain|interface|struct)(\s+)', | |
| 1624 bygroups(Keyword.Declaration, Text), 'class'), | |
| 1625 (r'(\.)([a-zA-Z_][a-zA-Z0-9_]*)', | |
| 1626 bygroups(Operator, Name.Attribute)), | |
| 1627 # void is an actual keyword, others are in glib-2.0.vapi | |
| 1628 (r'(void|bool|char|double|float|int|int8|int16|int32|int64|long|' | |
| 1629 r'short|size_t|ssize_t|string|time_t|uchar|uint|uint8|uint16|' | |
| 1630 r'uint32|uint64|ulong|unichar|ushort)\b', Keyword.Type), | |
| 1631 (r'(true|false|null)\b', Name.Builtin), | |
| 1632 ('[a-zA-Z_][a-zA-Z0-9_]*', Name), | |
| 1633 ], | |
| 1634 'root': [ | |
| 1635 include('whitespace'), | |
| 1636 ('', Text, 'statement'), | |
| 1637 ], | |
| 1638 'statement' : [ | |
| 1639 include('whitespace'), | |
| 1640 include('statements'), | |
| 1641 ('[{}]', Punctuation), | |
| 1642 (';', Punctuation, '#pop'), | |
| 1643 ], | |
| 1644 'string': [ | |
| 1645 (r'"', String, '#pop'), | |
| 1646 (r'\\([\\abfnrtv"\']|x[a-fA-F0-9]{2,4}|[0-7]{1,3})', String.Escape), | |
| 1647 (r'[^\\"\n]+', String), # all other characters | |
| 1648 (r'\\\n', String), # line continuation | |
| 1649 (r'\\', String), # stray backslash | |
| 1650 ], | |
| 1651 'if0': [ | |
| 1652 (r'^\s*#if.*?(?<!\\)\n', Comment.Preproc, '#push'), | |
| 1653 (r'^\s*#el(?:se|if).*\n', Comment.Preproc, '#pop'), | |
| 1654 (r'^\s*#endif.*?(?<!\\)\n', Comment.Preproc, '#pop'), | |
| 1655 (r'.*?\n', Comment), | |
| 1656 ], | |
| 1657 'class': [ | |
| 1658 (r'[a-zA-Z_][a-zA-Z0-9_]*', Name.Class, '#pop') | |
| 1659 ], | |
| 1660 'namespace': [ | |
| 1661 (r'[a-zA-Z_][a-zA-Z0-9_.]*', Name.Namespace, '#pop') | |
| 1662 ], | |
| 1663 } |
