logo

oasis-root

Compiled tree of Oasis Linux based on own branch at <https://hacktivis.me/git/oasis/> git clone https://anongit.hacktivis.me/git/oasis-root.git

sre_constants.py (7154B)


  1. #
  2. # Secret Labs' Regular Expression Engine
  3. #
  4. # various symbols used by the regular expression engine.
  5. # run this script to update the _sre include files!
  6. #
  7. # Copyright (c) 1998-2001 by Secret Labs AB. All rights reserved.
  8. #
  9. # See the sre.py file for information on usage and redistribution.
  10. #
  11. """Internal support module for sre"""
  12. # update when constants are added or removed
  13. MAGIC = 20171005
  14. from _sre import MAXREPEAT, MAXGROUPS
  15. # SRE standard exception (access as sre.error)
  16. # should this really be here?
  17. class error(Exception):
  18. """Exception raised for invalid regular expressions.
  19. Attributes:
  20. msg: The unformatted error message
  21. pattern: The regular expression pattern
  22. pos: The index in the pattern where compilation failed (may be None)
  23. lineno: The line corresponding to pos (may be None)
  24. colno: The column corresponding to pos (may be None)
  25. """
  26. __module__ = 're'
  27. def __init__(self, msg, pattern=None, pos=None):
  28. self.msg = msg
  29. self.pattern = pattern
  30. self.pos = pos
  31. if pattern is not None and pos is not None:
  32. msg = '%s at position %d' % (msg, pos)
  33. if isinstance(pattern, str):
  34. newline = '\n'
  35. else:
  36. newline = b'\n'
  37. self.lineno = pattern.count(newline, 0, pos) + 1
  38. self.colno = pos - pattern.rfind(newline, 0, pos)
  39. if newline in pattern:
  40. msg = '%s (line %d, column %d)' % (msg, self.lineno, self.colno)
  41. else:
  42. self.lineno = self.colno = None
  43. super().__init__(msg)
  44. class _NamedIntConstant(int):
  45. def __new__(cls, value, name):
  46. self = super(_NamedIntConstant, cls).__new__(cls, value)
  47. self.name = name
  48. return self
  49. def __repr__(self):
  50. return self.name
  51. MAXREPEAT = _NamedIntConstant(MAXREPEAT, 'MAXREPEAT')
  52. def _makecodes(names):
  53. names = names.strip().split()
  54. items = [_NamedIntConstant(i, name) for i, name in enumerate(names)]
  55. globals().update({item.name: item for item in items})
  56. return items
  57. # operators
  58. # failure=0 success=1 (just because it looks better that way :-)
  59. OPCODES = _makecodes("""
  60. FAILURE SUCCESS
  61. ANY ANY_ALL
  62. ASSERT ASSERT_NOT
  63. AT
  64. BRANCH
  65. CALL
  66. CATEGORY
  67. CHARSET BIGCHARSET
  68. GROUPREF GROUPREF_EXISTS
  69. IN
  70. INFO
  71. JUMP
  72. LITERAL
  73. MARK
  74. MAX_UNTIL
  75. MIN_UNTIL
  76. NOT_LITERAL
  77. NEGATE
  78. RANGE
  79. REPEAT
  80. REPEAT_ONE
  81. SUBPATTERN
  82. MIN_REPEAT_ONE
  83. GROUPREF_IGNORE
  84. IN_IGNORE
  85. LITERAL_IGNORE
  86. NOT_LITERAL_IGNORE
  87. GROUPREF_LOC_IGNORE
  88. IN_LOC_IGNORE
  89. LITERAL_LOC_IGNORE
  90. NOT_LITERAL_LOC_IGNORE
  91. GROUPREF_UNI_IGNORE
  92. IN_UNI_IGNORE
  93. LITERAL_UNI_IGNORE
  94. NOT_LITERAL_UNI_IGNORE
  95. RANGE_UNI_IGNORE
  96. MIN_REPEAT MAX_REPEAT
  97. """)
  98. del OPCODES[-2:] # remove MIN_REPEAT and MAX_REPEAT
  99. # positions
  100. ATCODES = _makecodes("""
  101. AT_BEGINNING AT_BEGINNING_LINE AT_BEGINNING_STRING
  102. AT_BOUNDARY AT_NON_BOUNDARY
  103. AT_END AT_END_LINE AT_END_STRING
  104. AT_LOC_BOUNDARY AT_LOC_NON_BOUNDARY
  105. AT_UNI_BOUNDARY AT_UNI_NON_BOUNDARY
  106. """)
  107. # categories
  108. CHCODES = _makecodes("""
  109. CATEGORY_DIGIT CATEGORY_NOT_DIGIT
  110. CATEGORY_SPACE CATEGORY_NOT_SPACE
  111. CATEGORY_WORD CATEGORY_NOT_WORD
  112. CATEGORY_LINEBREAK CATEGORY_NOT_LINEBREAK
  113. CATEGORY_LOC_WORD CATEGORY_LOC_NOT_WORD
  114. CATEGORY_UNI_DIGIT CATEGORY_UNI_NOT_DIGIT
  115. CATEGORY_UNI_SPACE CATEGORY_UNI_NOT_SPACE
  116. CATEGORY_UNI_WORD CATEGORY_UNI_NOT_WORD
  117. CATEGORY_UNI_LINEBREAK CATEGORY_UNI_NOT_LINEBREAK
  118. """)
  119. # replacement operations for "ignore case" mode
  120. OP_IGNORE = {
  121. LITERAL: LITERAL_IGNORE,
  122. NOT_LITERAL: NOT_LITERAL_IGNORE,
  123. }
  124. OP_LOCALE_IGNORE = {
  125. LITERAL: LITERAL_LOC_IGNORE,
  126. NOT_LITERAL: NOT_LITERAL_LOC_IGNORE,
  127. }
  128. OP_UNICODE_IGNORE = {
  129. LITERAL: LITERAL_UNI_IGNORE,
  130. NOT_LITERAL: NOT_LITERAL_UNI_IGNORE,
  131. }
  132. AT_MULTILINE = {
  133. AT_BEGINNING: AT_BEGINNING_LINE,
  134. AT_END: AT_END_LINE
  135. }
  136. AT_LOCALE = {
  137. AT_BOUNDARY: AT_LOC_BOUNDARY,
  138. AT_NON_BOUNDARY: AT_LOC_NON_BOUNDARY
  139. }
  140. AT_UNICODE = {
  141. AT_BOUNDARY: AT_UNI_BOUNDARY,
  142. AT_NON_BOUNDARY: AT_UNI_NON_BOUNDARY
  143. }
  144. CH_LOCALE = {
  145. CATEGORY_DIGIT: CATEGORY_DIGIT,
  146. CATEGORY_NOT_DIGIT: CATEGORY_NOT_DIGIT,
  147. CATEGORY_SPACE: CATEGORY_SPACE,
  148. CATEGORY_NOT_SPACE: CATEGORY_NOT_SPACE,
  149. CATEGORY_WORD: CATEGORY_LOC_WORD,
  150. CATEGORY_NOT_WORD: CATEGORY_LOC_NOT_WORD,
  151. CATEGORY_LINEBREAK: CATEGORY_LINEBREAK,
  152. CATEGORY_NOT_LINEBREAK: CATEGORY_NOT_LINEBREAK
  153. }
  154. CH_UNICODE = {
  155. CATEGORY_DIGIT: CATEGORY_UNI_DIGIT,
  156. CATEGORY_NOT_DIGIT: CATEGORY_UNI_NOT_DIGIT,
  157. CATEGORY_SPACE: CATEGORY_UNI_SPACE,
  158. CATEGORY_NOT_SPACE: CATEGORY_UNI_NOT_SPACE,
  159. CATEGORY_WORD: CATEGORY_UNI_WORD,
  160. CATEGORY_NOT_WORD: CATEGORY_UNI_NOT_WORD,
  161. CATEGORY_LINEBREAK: CATEGORY_UNI_LINEBREAK,
  162. CATEGORY_NOT_LINEBREAK: CATEGORY_UNI_NOT_LINEBREAK
  163. }
  164. # flags
  165. SRE_FLAG_TEMPLATE = 1 # template mode (disable backtracking)
  166. SRE_FLAG_IGNORECASE = 2 # case insensitive
  167. SRE_FLAG_LOCALE = 4 # honour system locale
  168. SRE_FLAG_MULTILINE = 8 # treat target as multiline string
  169. SRE_FLAG_DOTALL = 16 # treat target as a single string
  170. SRE_FLAG_UNICODE = 32 # use unicode "locale"
  171. SRE_FLAG_VERBOSE = 64 # ignore whitespace and comments
  172. SRE_FLAG_DEBUG = 128 # debugging
  173. SRE_FLAG_ASCII = 256 # use ascii "locale"
  174. # flags for INFO primitive
  175. SRE_INFO_PREFIX = 1 # has prefix
  176. SRE_INFO_LITERAL = 2 # entire pattern is literal (given by prefix)
  177. SRE_INFO_CHARSET = 4 # pattern starts with character from given set
  178. if __name__ == "__main__":
  179. def dump(f, d, prefix):
  180. items = sorted(d)
  181. for item in items:
  182. f.write("#define %s_%s %d\n" % (prefix, item, item))
  183. with open("sre_constants.h", "w") as f:
  184. f.write("""\
  185. /*
  186. * Secret Labs' Regular Expression Engine
  187. *
  188. * regular expression matching engine
  189. *
  190. * NOTE: This file is generated by sre_constants.py. If you need
  191. * to change anything in here, edit sre_constants.py and run it.
  192. *
  193. * Copyright (c) 1997-2001 by Secret Labs AB. All rights reserved.
  194. *
  195. * See the _sre.c file for information on usage and redistribution.
  196. */
  197. """)
  198. f.write("#define SRE_MAGIC %d\n" % MAGIC)
  199. dump(f, OPCODES, "SRE_OP")
  200. dump(f, ATCODES, "SRE")
  201. dump(f, CHCODES, "SRE")
  202. f.write("#define SRE_FLAG_TEMPLATE %d\n" % SRE_FLAG_TEMPLATE)
  203. f.write("#define SRE_FLAG_IGNORECASE %d\n" % SRE_FLAG_IGNORECASE)
  204. f.write("#define SRE_FLAG_LOCALE %d\n" % SRE_FLAG_LOCALE)
  205. f.write("#define SRE_FLAG_MULTILINE %d\n" % SRE_FLAG_MULTILINE)
  206. f.write("#define SRE_FLAG_DOTALL %d\n" % SRE_FLAG_DOTALL)
  207. f.write("#define SRE_FLAG_UNICODE %d\n" % SRE_FLAG_UNICODE)
  208. f.write("#define SRE_FLAG_VERBOSE %d\n" % SRE_FLAG_VERBOSE)
  209. f.write("#define SRE_FLAG_DEBUG %d\n" % SRE_FLAG_DEBUG)
  210. f.write("#define SRE_FLAG_ASCII %d\n" % SRE_FLAG_ASCII)
  211. f.write("#define SRE_INFO_PREFIX %d\n" % SRE_INFO_PREFIX)
  212. f.write("#define SRE_INFO_LITERAL %d\n" % SRE_INFO_LITERAL)
  213. f.write("#define SRE_INFO_CHARSET %d\n" % SRE_INFO_CHARSET)
  214. print("done")