cpython/Lib/re/_constants.py

220 lines
5.7 KiB
Python

#
# Secret Labs' Regular Expression Engine
#
# various symbols used by the regular expression engine.
# run this script to update the _sre include files!
#
# Copyright (c) 1998-2001 by Secret Labs AB. All rights reserved.
#
# See the __init__.py file for information on usage and redistribution.
#
"""Internal support module for sre"""
# update when constants are added or removed
MAGIC = 20230612
from _sre import MAXREPEAT, MAXGROUPS
# SRE standard exception (access as sre.error)
# should this really be here?
class error(Exception):
"""Exception raised for invalid regular expressions.
Attributes:
msg: The unformatted error message
pattern: The regular expression pattern
pos: The index in the pattern where compilation failed (may be None)
lineno: The line corresponding to pos (may be None)
colno: The column corresponding to pos (may be None)
"""
__module__ = 're'
def __init__(self, msg, pattern=None, pos=None):
self.msg = msg
self.pattern = pattern
self.pos = pos
if pattern is not None and pos is not None:
msg = '%s at position %d' % (msg, pos)
if isinstance(pattern, str):
newline = '\n'
else:
newline = b'\n'
self.lineno = pattern.count(newline, 0, pos) + 1
self.colno = pos - pattern.rfind(newline, 0, pos)
if newline in pattern:
msg = '%s (line %d, column %d)' % (msg, self.lineno, self.colno)
else:
self.lineno = self.colno = None
super().__init__(msg)
class _NamedIntConstant(int):
def __new__(cls, value, name):
self = super(_NamedIntConstant, cls).__new__(cls, value)
self.name = name
return self
def __repr__(self):
return self.name
__reduce__ = None
MAXREPEAT = _NamedIntConstant(MAXREPEAT, 'MAXREPEAT')
def _makecodes(*names):
items = [_NamedIntConstant(i, name) for i, name in enumerate(names)]
globals().update({item.name: item for item in items})
return items
# operators
OPCODES = _makecodes(
# failure=0 success=1 (just because it looks better that way :-)
'FAILURE', 'SUCCESS',
'ANY', 'ANY_ALL',
'ASSERT', 'ASSERT_NOT',
'AT',
'BRANCH',
'CATEGORY',
'CHARSET', 'BIGCHARSET',
'GROUPREF', 'GROUPREF_EXISTS',
'IN',
'INFO',
'JUMP',
'LITERAL',
'MARK',
'MAX_UNTIL',
'MIN_UNTIL',
'NOT_LITERAL',
'NEGATE',
'RANGE',
'REPEAT',
'REPEAT_ONE',
'SUBPATTERN',
'MIN_REPEAT_ONE',
'ATOMIC_GROUP',
'POSSESSIVE_REPEAT',
'POSSESSIVE_REPEAT_ONE',
'GROUPREF_IGNORE',
'IN_IGNORE',
'LITERAL_IGNORE',
'NOT_LITERAL_IGNORE',
'GROUPREF_LOC_IGNORE',
'IN_LOC_IGNORE',
'LITERAL_LOC_IGNORE',
'NOT_LITERAL_LOC_IGNORE',
'GROUPREF_UNI_IGNORE',
'IN_UNI_IGNORE',
'LITERAL_UNI_IGNORE',
'NOT_LITERAL_UNI_IGNORE',
'RANGE_UNI_IGNORE',
# The following opcodes are only occurred in the parser output,
# but not in the compiled code.
'MIN_REPEAT', 'MAX_REPEAT',
)
del OPCODES[-2:] # remove MIN_REPEAT and MAX_REPEAT
# positions
ATCODES = _makecodes(
'AT_BEGINNING', 'AT_BEGINNING_LINE', 'AT_BEGINNING_STRING',
'AT_BOUNDARY', 'AT_NON_BOUNDARY',
'AT_END', 'AT_END_LINE', 'AT_END_STRING',
'AT_LOC_BOUNDARY', 'AT_LOC_NON_BOUNDARY',
'AT_UNI_BOUNDARY', 'AT_UNI_NON_BOUNDARY',
)
# categories
CHCODES = _makecodes(
'CATEGORY_DIGIT', 'CATEGORY_NOT_DIGIT',
'CATEGORY_SPACE', 'CATEGORY_NOT_SPACE',
'CATEGORY_WORD', 'CATEGORY_NOT_WORD',
'CATEGORY_LINEBREAK', 'CATEGORY_NOT_LINEBREAK',
'CATEGORY_LOC_WORD', 'CATEGORY_LOC_NOT_WORD',
'CATEGORY_UNI_DIGIT', 'CATEGORY_UNI_NOT_DIGIT',
'CATEGORY_UNI_SPACE', 'CATEGORY_UNI_NOT_SPACE',
'CATEGORY_UNI_WORD', 'CATEGORY_UNI_NOT_WORD',
'CATEGORY_UNI_LINEBREAK', 'CATEGORY_UNI_NOT_LINEBREAK',
)
# replacement operations for "ignore case" mode
OP_IGNORE = {
LITERAL: LITERAL_IGNORE,
NOT_LITERAL: NOT_LITERAL_IGNORE,
}
OP_LOCALE_IGNORE = {
LITERAL: LITERAL_LOC_IGNORE,
NOT_LITERAL: NOT_LITERAL_LOC_IGNORE,
}
OP_UNICODE_IGNORE = {
LITERAL: LITERAL_UNI_IGNORE,
NOT_LITERAL: NOT_LITERAL_UNI_IGNORE,
}
AT_MULTILINE = {
AT_BEGINNING: AT_BEGINNING_LINE,
AT_END: AT_END_LINE
}
AT_LOCALE = {
AT_BOUNDARY: AT_LOC_BOUNDARY,
AT_NON_BOUNDARY: AT_LOC_NON_BOUNDARY
}
AT_UNICODE = {
AT_BOUNDARY: AT_UNI_BOUNDARY,
AT_NON_BOUNDARY: AT_UNI_NON_BOUNDARY
}
CH_LOCALE = {
CATEGORY_DIGIT: CATEGORY_DIGIT,
CATEGORY_NOT_DIGIT: CATEGORY_NOT_DIGIT,
CATEGORY_SPACE: CATEGORY_SPACE,
CATEGORY_NOT_SPACE: CATEGORY_NOT_SPACE,
CATEGORY_WORD: CATEGORY_LOC_WORD,
CATEGORY_NOT_WORD: CATEGORY_LOC_NOT_WORD,
CATEGORY_LINEBREAK: CATEGORY_LINEBREAK,
CATEGORY_NOT_LINEBREAK: CATEGORY_NOT_LINEBREAK
}
CH_UNICODE = {
CATEGORY_DIGIT: CATEGORY_UNI_DIGIT,
CATEGORY_NOT_DIGIT: CATEGORY_UNI_NOT_DIGIT,
CATEGORY_SPACE: CATEGORY_UNI_SPACE,
CATEGORY_NOT_SPACE: CATEGORY_UNI_NOT_SPACE,
CATEGORY_WORD: CATEGORY_UNI_WORD,
CATEGORY_NOT_WORD: CATEGORY_UNI_NOT_WORD,
CATEGORY_LINEBREAK: CATEGORY_UNI_LINEBREAK,
CATEGORY_NOT_LINEBREAK: CATEGORY_UNI_NOT_LINEBREAK
}
# flags
SRE_FLAG_IGNORECASE = 2 # case insensitive
SRE_FLAG_LOCALE = 4 # honour system locale
SRE_FLAG_MULTILINE = 8 # treat target as multiline string
SRE_FLAG_DOTALL = 16 # treat target as a single string
SRE_FLAG_UNICODE = 32 # use unicode "locale"
SRE_FLAG_VERBOSE = 64 # ignore whitespace and comments
SRE_FLAG_DEBUG = 128 # debugging
SRE_FLAG_ASCII = 256 # use ascii "locale"
# flags for INFO primitive
SRE_INFO_PREFIX = 1 # has prefix
SRE_INFO_LITERAL = 2 # entire pattern is literal (given by prefix)
SRE_INFO_CHARSET = 4 # pattern starts with character from given set