cpython/Lib/dos-8x3/sre_comp.py

382 lines
11 KiB
Python
Raw Normal View History

2000-05-08 14:31:04 -03:00
#
# Secret Labs' Regular Expression Engine
#
# convert template to internal format
#
# Copyright (c) 1997-2000 by Secret Labs AB. All rights reserved.
#
2000-09-01 16:25:51 -03:00
# See the sre.py file for information on usage and redistribution.
2000-05-08 14:31:04 -03:00
#
import _sre
from sre_constants import *
2000-09-01 16:25:51 -03:00
MAXCODE = 65535
2000-05-08 14:31:04 -03:00
def _compile(code, pattern, flags):
2000-06-30 13:13:37 -03:00
# internal: compile a (sub)pattern
2000-06-29 16:35:29 -03:00
emit = code.append
2000-05-08 14:31:04 -03:00
for op, av in pattern:
2000-06-30 13:13:37 -03:00
if op in (LITERAL, NOT_LITERAL):
if flags & SRE_FLAG_IGNORECASE:
emit(OPCODES[OP_IGNORE[op]])
else:
emit(OPCODES[op])
emit(av)
elif op is IN:
if flags & SRE_FLAG_IGNORECASE:
emit(OPCODES[OP_IGNORE[op]])
def fixup(literal, flags=flags):
return _sre.getlower(literal, flags)
else:
emit(OPCODES[op])
2000-07-01 01:23:47 -03:00
fixup = lambda x: x
2000-06-30 13:13:37 -03:00
skip = len(code); emit(0)
2000-09-01 16:25:51 -03:00
_compile_charset(av, flags, code, fixup)
2000-06-30 13:13:37 -03:00
code[skip] = len(code) - skip
elif op is ANY:
if flags & SRE_FLAG_DOTALL:
2000-09-01 16:25:51 -03:00
emit(OPCODES[ANY_ALL])
2000-06-30 13:13:37 -03:00
else:
2000-09-01 16:25:51 -03:00
emit(OPCODES[ANY])
2000-06-30 13:13:37 -03:00
elif op in (REPEAT, MIN_REPEAT, MAX_REPEAT):
if flags & SRE_FLAG_TEMPLATE:
2000-09-01 16:25:51 -03:00
raise error, "internal: unsupported template operator"
2000-06-30 13:13:37 -03:00
emit(OPCODES[REPEAT])
skip = len(code); emit(0)
emit(av[0])
emit(av[1])
_compile(code, av[2], flags)
emit(OPCODES[SUCCESS])
code[skip] = len(code) - skip
2000-09-01 16:25:51 -03:00
elif _simple(av) and op == MAX_REPEAT:
emit(OPCODES[REPEAT_ONE])
skip = len(code); emit(0)
emit(av[0])
emit(av[1])
_compile(code, av[2], flags)
emit(OPCODES[SUCCESS])
code[skip] = len(code) - skip
2000-06-30 13:13:37 -03:00
else:
2000-09-01 16:25:51 -03:00
emit(OPCODES[REPEAT])
skip = len(code); emit(0)
emit(av[0])
emit(av[1])
_compile(code, av[2], flags)
code[skip] = len(code) - skip
if op == MAX_REPEAT:
emit(OPCODES[MAX_UNTIL])
2000-06-30 13:13:37 -03:00
else:
2000-09-01 16:25:51 -03:00
emit(OPCODES[MIN_UNTIL])
2000-06-30 13:13:37 -03:00
elif op is SUBPATTERN:
2000-09-01 16:25:51 -03:00
if av[0]:
2000-06-30 13:13:37 -03:00
emit(OPCODES[MARK])
2000-09-01 16:25:51 -03:00
emit((av[0]-1)*2)
# _compile_info(code, av[1], flags)
2000-06-30 13:13:37 -03:00
_compile(code, av[1], flags)
2000-09-01 16:25:51 -03:00
if av[0]:
2000-06-30 13:13:37 -03:00
emit(OPCODES[MARK])
2000-09-01 16:25:51 -03:00
emit((av[0]-1)*2+1)
2000-06-30 13:13:37 -03:00
elif op in (SUCCESS, FAILURE):
emit(OPCODES[op])
2000-09-01 16:25:51 -03:00
elif op in (ASSERT, ASSERT_NOT):
emit(OPCODES[op])
skip = len(code); emit(0)
if av[0] >= 0:
emit(0) # look ahead
else:
lo, hi = av[1].getwidth()
if lo != hi:
raise error, "look-behind requires fixed-width pattern"
emit(lo) # look behind
_compile(code, av[1], flags)
emit(OPCODES[SUCCESS])
code[skip] = len(code) - skip
elif op is CALL:
2000-06-30 13:13:37 -03:00
emit(OPCODES[op])
skip = len(code); emit(0)
_compile(code, av, flags)
emit(OPCODES[SUCCESS])
code[skip] = len(code) - skip
elif op is AT:
emit(OPCODES[op])
if flags & SRE_FLAG_MULTILINE:
2000-07-01 01:23:47 -03:00
emit(ATCODES[AT_MULTILINE.get(av, av)])
2000-06-30 13:13:37 -03:00
else:
emit(ATCODES[av])
elif op is BRANCH:
emit(OPCODES[op])
tail = []
for av in av[1]:
skip = len(code); emit(0)
2000-09-01 16:25:51 -03:00
# _compile_info(code, av, flags)
2000-06-30 13:13:37 -03:00
_compile(code, av, flags)
emit(OPCODES[JUMP])
tail.append(len(code)); emit(0)
code[skip] = len(code) - skip
emit(0) # end of branch
for tail in tail:
code[tail] = len(code) - tail
elif op is CATEGORY:
emit(OPCODES[op])
if flags & SRE_FLAG_LOCALE:
emit(CHCODES[CH_LOCALE[av]])
elif flags & SRE_FLAG_UNICODE:
emit(CHCODES[CH_UNICODE[av]])
else:
emit(CHCODES[av])
2000-09-01 16:25:51 -03:00
elif op is GROUPREF:
2000-06-30 13:13:37 -03:00
if flags & SRE_FLAG_IGNORECASE:
emit(OPCODES[OP_IGNORE[op]])
else:
emit(OPCODES[op])
emit(av-1)
else:
raise ValueError, ("unsupported operand type", op)
2000-09-01 16:25:51 -03:00
def _compile_charset(charset, flags, code, fixup=None):
# compile charset subprogram
emit = code.append
if not fixup:
fixup = lambda x: x
for op, av in _optimize_charset(charset, fixup):
emit(OPCODES[op])
if op is NEGATE:
pass
elif op is LITERAL:
emit(fixup(av))
elif op is RANGE:
emit(fixup(av[0]))
emit(fixup(av[1]))
elif op is CHARSET:
code.extend(av)
elif op is CATEGORY:
if flags & SRE_FLAG_LOCALE:
emit(CHCODES[CH_LOCALE[av]])
elif flags & SRE_FLAG_UNICODE:
emit(CHCODES[CH_UNICODE[av]])
else:
emit(CHCODES[av])
else:
raise error, "internal: unsupported set operator"
emit(OPCODES[FAILURE])
def _optimize_charset(charset, fixup):
# internal: optimize character set
out = []
charmap = [0]*256
try:
for op, av in charset:
if op is NEGATE:
out.append((op, av))
elif op is LITERAL:
charmap[fixup(av)] = 1
elif op is RANGE:
for i in range(fixup(av[0]), fixup(av[1])+1):
charmap[i] = 1
elif op is CATEGORY:
# FIXME: could append to charmap tail
return charset # cannot compress
except IndexError:
# character set contains unicode characters
return charset
# compress character map
i = p = n = 0
runs = []
for c in charmap:
if c:
if n == 0:
p = i
n = n + 1
elif n:
runs.append((p, n))
n = 0
i = i + 1
if n:
runs.append((p, n))
if len(runs) <= 2:
# use literal/range
for p, n in runs:
if n == 1:
out.append((LITERAL, p))
else:
out.append((RANGE, (p, p+n-1)))
if len(out) < len(charset):
return out
else:
# use bitmap
data = []
m = 1; v = 0
for c in charmap:
if c:
v = v + m
m = m << 1
if m > MAXCODE:
data.append(v)
m = 1; v = 0
out.append((CHARSET, data))
return out
return charset
def _simple(av):
# check if av is a "simple" operator
lo, hi = av[2].getwidth()
if lo == 0:
raise error, "nothing to repeat"
return lo == hi == 1 and av[2][0][0] != SUBPATTERN
2000-06-30 13:13:37 -03:00
def _compile_info(code, pattern, flags):
# internal: compile an info block. in the current version,
2000-09-01 16:25:51 -03:00
# this contains min/max pattern width, and an optional literal
# prefix or a character map
2000-06-30 13:13:37 -03:00
lo, hi = pattern.getwidth()
if lo == 0:
return # not worth it
# look for a literal prefix
prefix = []
2000-09-01 16:25:51 -03:00
prefix_skip = 0
charset = [] # not used
2000-06-30 13:13:37 -03:00
if not (flags & SRE_FLAG_IGNORECASE):
2000-09-01 16:25:51 -03:00
# look for literal prefix
2000-06-30 13:13:37 -03:00
for op, av in pattern.data:
if op is LITERAL:
2000-09-01 16:25:51 -03:00
if len(prefix) == prefix_skip:
prefix_skip = prefix_skip + 1
2000-06-30 13:13:37 -03:00
prefix.append(av)
2000-09-01 16:25:51 -03:00
elif op is SUBPATTERN and len(av[1]) == 1:
op, av = av[1][0]
if op is LITERAL:
prefix.append(av)
else:
break
2000-06-30 13:13:37 -03:00
else:
break
2000-09-01 16:25:51 -03:00
# if no prefix, look for charset prefix
if not prefix and pattern.data:
op, av = pattern.data[0]
if op is SUBPATTERN and av[1]:
op, av = av[1][0]
if op is LITERAL:
charset.append((op, av))
elif op is BRANCH:
c = []
for p in av[1]:
if not p:
break
op, av = p[0]
if op is LITERAL:
c.append((op, av))
else:
break
else:
charset = c
elif op is BRANCH:
c = []
for p in av[1]:
if not p:
break
op, av = p[0]
if op is LITERAL:
c.append((op, av))
else:
break
else:
charset = c
elif op is IN:
charset = av
## if prefix:
## print "*** PREFIX", prefix, prefix_skip
## if charset:
## print "*** CHARSET", charset
2000-06-30 13:13:37 -03:00
# add an info block
emit = code.append
emit(OPCODES[INFO])
skip = len(code); emit(0)
# literal flag
mask = 0
2000-09-01 16:25:51 -03:00
if prefix:
mask = SRE_INFO_PREFIX
if len(prefix) == prefix_skip == len(pattern.data):
mask = mask + SRE_INFO_LITERAL
elif charset:
mask = mask + SRE_INFO_CHARSET
2000-06-30 13:13:37 -03:00
emit(mask)
# pattern length
2000-09-01 16:25:51 -03:00
if lo < MAXCODE:
emit(lo)
else:
emit(MAXCODE)
prefix = prefix[:MAXCODE]
if hi < MAXCODE:
2000-06-30 13:13:37 -03:00
emit(hi)
else:
emit(0)
# add literal prefix
if prefix:
2000-09-01 16:25:51 -03:00
emit(len(prefix)) # length
emit(prefix_skip) # skip
2000-06-30 13:13:37 -03:00
code.extend(prefix)
# generate overlap table
table = [-1] + ([0]*len(prefix))
for i in range(len(prefix)):
table[i+1] = table[i]+1
while table[i+1] > 0 and prefix[i] != prefix[table[i+1]-1]:
table[i+1] = table[table[i+1]-1]+1
code.extend(table[1:]) # don't store first entry
2000-09-01 16:25:51 -03:00
elif charset:
_compile_charset(charset, 0, code)
2000-06-30 13:13:37 -03:00
code[skip] = len(code) - skip
2000-05-08 14:31:04 -03:00
2000-09-01 16:25:51 -03:00
STRING_TYPES = [type("")]
2000-06-30 13:13:37 -03:00
2000-09-01 16:25:51 -03:00
try:
STRING_TYPES.append(type(unicode("")))
except NameError:
pass
def _code(p, flags):
2000-06-30 13:13:37 -03:00
2000-06-29 16:35:29 -03:00
flags = p.pattern.flags | flags
code = []
2000-06-30 13:13:37 -03:00
# compile info block
_compile_info(code, p, flags)
# compile the pattern
2000-06-29 16:35:29 -03:00
_compile(code, p.data, flags)
2000-06-30 13:13:37 -03:00
2000-06-29 16:35:29 -03:00
code.append(OPCODES[SUCCESS])
2000-06-30 13:13:37 -03:00
2000-09-01 16:25:51 -03:00
return code
def compile(p, flags=0):
# internal: convert pattern list to internal format
if type(p) in STRING_TYPES:
import sre_parse
pattern = p
p = sre_parse.parse(p, flags)
else:
pattern = None
code = _code(p, flags)
# print code
2000-06-30 13:13:37 -03:00
# FIXME: <fl> get rid of this limitation!
2000-06-29 16:35:29 -03:00
assert p.pattern.groups <= 100,\
2000-06-30 13:13:37 -03:00
"sorry, but this version only supports 100 named groups"
2000-09-01 16:25:51 -03:00
# map in either direction
groupindex = p.pattern.groupdict
indexgroup = [None] * p.pattern.groups
for k, i in groupindex.items():
indexgroup[i] = k
2000-06-29 16:35:29 -03:00
return _sre.compile(
2000-09-01 16:25:51 -03:00
pattern, flags, code,
p.pattern.groups-1,
groupindex, indexgroup
2000-06-30 13:13:37 -03:00
)