0
0
mirror of https://github.com/python/cpython.git synced 2024-11-24 00:38:00 +01:00
cpython/Lib/sre_constants.py
2000-06-01 17:39:12 +00:00

173 lines
4.2 KiB
Python

#
# Secret Labs' Regular Expression Engine
# $Id$
#
# various symbols used by the regular expression engine.
# run this script to update the _sre include files!
#
# Copyright (c) 1998-2000 by Secret Labs AB. All rights reserved.
#
# This code can only be used for 1.6 alpha testing. All other use
# require explicit permission from Secret Labs AB.
#
# Portions of this engine have been developed in cooperation with
# CNRI. Hewlett-Packard provided funding for 1.6 integration and
# other compatibility work.
#
# operators
FAILURE = "failure"
SUCCESS = "success"
ANY = "any"
ASSERT = "assert"
AT = "at"
BRANCH = "branch"
CALL = "call"
CATEGORY = "category"
GROUP = "group"
GROUP_IGNORE = "group_ignore"
IN = "in"
IN_IGNORE = "in_ignore"
JUMP = "jump"
LITERAL = "literal"
LITERAL_IGNORE = "literal_ignore"
MARK = "mark"
MAX_REPEAT = "max_repeat"
MAX_REPEAT_ONE = "max_repeat_one"
MAX_UNTIL = "max_until"
MIN_REPEAT = "min_repeat"
MIN_UNTIL = "min_until"
NEGATE = "negate"
NOT_LITERAL = "not_literal"
NOT_LITERAL_IGNORE = "not_literal_ignore"
RANGE = "range"
REPEAT = "repeat"
SUBPATTERN = "subpattern"
# positions
AT_BEGINNING = "at_beginning"
AT_BEGINNING_LINE = "at_beginning_line"
AT_BOUNDARY = "at_boundary"
AT_NON_BOUNDARY = "at_non_boundary"
AT_END = "at_end"
AT_END_LINE = "at_end_line"
# categories
CATEGORY_DIGIT = "category_digit"
CATEGORY_NOT_DIGIT = "category_not_digit"
CATEGORY_SPACE = "category_space"
CATEGORY_NOT_SPACE = "category_not_space"
CATEGORY_WORD = "category_word"
CATEGORY_NOT_WORD = "category_not_word"
CATEGORY_LINEBREAK = "category_linebreak"
CATEGORY_NOT_LINEBREAK = "category_not_linebreak"
CATEGORY_LOC_DIGIT = "category_loc_digit"
CATEGORY_LOC_NOT_DIGIT = "category_loc_not_digit"
CATEGORY_LOC_SPACE = "category_loc_space"
CATEGORY_LOC_NOT_SPACE = "category_loc_not_space"
CATEGORY_LOC_WORD = "category_loc_word"
CATEGORY_LOC_NOT_WORD = "category_loc_not_word"
CATEGORY_LOC_LINEBREAK = "category_loc_linebreak"
CATEGORY_LOC_NOT_LINEBREAK = "category_loc_not_linebreak"
OPCODES = [
# failure=0 success=1 (just because it looks better that way :-)
FAILURE, SUCCESS,
ANY,
ASSERT,
AT,
BRANCH,
CALL,
CATEGORY,
GROUP, GROUP_IGNORE,
IN, IN_IGNORE,
JUMP,
LITERAL, LITERAL_IGNORE,
MARK,
MAX_REPEAT, MAX_UNTIL,
MAX_REPEAT_ONE,
MIN_REPEAT, MIN_UNTIL,
NOT_LITERAL, NOT_LITERAL_IGNORE,
NEGATE,
RANGE,
REPEAT
]
ATCODES = [
AT_BEGINNING, AT_BEGINNING_LINE, AT_BOUNDARY,
AT_NON_BOUNDARY, AT_END, AT_END_LINE
]
CHCODES = [
CATEGORY_DIGIT, CATEGORY_NOT_DIGIT, CATEGORY_SPACE,
CATEGORY_NOT_SPACE, CATEGORY_WORD, CATEGORY_NOT_WORD,
CATEGORY_LINEBREAK, CATEGORY_NOT_LINEBREAK, CATEGORY_LOC_DIGIT,
CATEGORY_LOC_NOT_DIGIT, CATEGORY_LOC_SPACE,
CATEGORY_LOC_NOT_SPACE, CATEGORY_LOC_WORD, CATEGORY_LOC_NOT_WORD,
CATEGORY_LOC_LINEBREAK, CATEGORY_LOC_NOT_LINEBREAK
]
def makedict(list):
d = {}
i = 0
for item in list:
d[item] = i
i = i + 1
return d
OPCODES = makedict(OPCODES)
ATCODES = makedict(ATCODES)
CHCODES = makedict(CHCODES)
# replacement operations for "ignore case" mode
OP_IGNORE = {
GROUP: GROUP_IGNORE,
IN: IN_IGNORE,
LITERAL: LITERAL_IGNORE,
NOT_LITERAL: NOT_LITERAL_IGNORE
}
AT_MULTILINE = {
AT_BEGINNING: AT_BEGINNING_LINE,
AT_END: AT_END_LINE
}
CH_LOCALE = {
CATEGORY_DIGIT: CATEGORY_LOC_DIGIT,
CATEGORY_NOT_DIGIT: CATEGORY_LOC_NOT_DIGIT,
CATEGORY_SPACE: CATEGORY_LOC_SPACE,
CATEGORY_NOT_SPACE: CATEGORY_LOC_NOT_SPACE,
CATEGORY_WORD: CATEGORY_LOC_WORD,
CATEGORY_NOT_WORD: CATEGORY_LOC_NOT_WORD,
CATEGORY_LINEBREAK: CATEGORY_LOC_LINEBREAK,
CATEGORY_NOT_LINEBREAK: CATEGORY_LOC_NOT_LINEBREAK
}
# flags
SRE_FLAG_TEMPLATE = 1 # NYI
SRE_FLAG_IGNORECASE = 2
SRE_FLAG_LOCALE = 4
SRE_FLAG_MULTILINE = 8
SRE_FLAG_DOTALL = 16
SRE_FLAG_VERBOSE = 32
if __name__ == "__main__":
import string
def dump(f, d, prefix):
items = d.items()
items.sort(lambda a, b: cmp(a[1], b[1]))
for k, v in items:
f.write("#define %s_%s %s\n" % (prefix, string.upper(k), v))
f = open("sre_constants.h", "w")
f.write("/* generated from sre_constants.py */\n")
dump(f, OPCODES, "SRE_OP")
dump(f, ATCODES, "SRE")
dump(f, CHCODES, "SRE")
f.close()
print "done"