The usual :)
diff --git a/Lib/dos-8x3/sre_comp.py b/Lib/dos-8x3/sre_comp.py
index 8738061..c042375 100644
--- a/Lib/dos-8x3/sre_comp.py
+++ b/Lib/dos-8x3/sre_comp.py
@@ -1,24 +1,16 @@
#
# Secret Labs' Regular Expression Engine
-# $Id$
#
# convert template to internal format
#
# Copyright (c) 1997-2000 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.
#
-# FIXME: <fl> formalize (objectify?) and document the compiler code
-# format, so that other frontends can use this compiler
-
-import array, string, sys
-
+import array
import _sre
from sre_constants import *
@@ -30,158 +22,153 @@
else:
raise RuntimeError, "cannot find a useable array type"
-# FIXME: <fl> should move some optimizations from the parser to here!
-
-class Code:
- def __init__(self):
- self.data = []
- def __len__(self):
- return len(self.data)
- def __getitem__(self, index):
- return self.data[index]
- def __setitem__(self, index, code):
- self.data[index] = code
- def append(self, code):
- self.data.append(code)
- def todata(self):
- # print self.data
- return array.array(WORDSIZE, self.data).tostring()
-
-def _lower(literal):
- # return _sre._lower(literal) # FIXME
- return string.lower(literal)
-
def _compile(code, pattern, flags):
- append = code.append
+ emit = code.append
for op, av in pattern:
if op is ANY:
- if "s" in flags:
- append(CODES[op]) # any character at all!
+ if flags & SRE_FLAG_DOTALL:
+ emit(OPCODES[op])
else:
- append(CODES[NOT_LITERAL])
- append(10)
+ emit(OPCODES[CATEGORY])
+ emit(CHCODES[CATEGORY_NOT_LINEBREAK])
elif op in (SUCCESS, FAILURE):
- append(CODES[op])
+ emit(OPCODES[op])
elif op is AT:
- append(CODES[op])
- append(POSITIONS[av])
+ emit(OPCODES[op])
+ if flags & SRE_FLAG_MULTILINE:
+ emit(ATCODES[AT_MULTILINE[av]])
+ else:
+ emit(ATCODES[av])
elif op is BRANCH:
- append(CODES[op])
+ emit(OPCODES[op])
tail = []
for av in av[1]:
- skip = len(code); append(0)
+ skip = len(code); emit(0)
_compile(code, av, flags)
- append(CODES[JUMP])
- tail.append(len(code)); append(0)
+ emit(OPCODES[JUMP])
+ tail.append(len(code)); emit(0)
code[skip] = len(code) - skip
- append(0) # end of branch
- for tail in tail:
+ emit(0) # end of branch
+ for tail in tail:
code[tail] = len(code) - tail
elif op is CALL:
- append(CODES[op])
- skip = len(code); append(0)
+ emit(OPCODES[op])
+ skip = len(code); emit(0)
_compile(code, av, flags)
- append(CODES[SUCCESS])
+ emit(OPCODES[SUCCESS])
code[skip] = len(code) - skip
- elif op is CATEGORY: # not used by current parser
- append(CODES[op])
- append(CATEGORIES[av])
+ elif op is CATEGORY:
+ emit(OPCODES[op])
+ if flags & SRE_FLAG_LOCALE:
+ emit(CH_LOCALE[CHCODES[av]])
+ elif flags & SRE_FLAG_UNICODE:
+ emit(CH_UNICODE[CHCODES[av]])
+ else:
+ emit(CHCODES[av])
elif op is GROUP:
- if "i" in flags:
- append(CODES[MAP_IGNORE[op]])
+ if flags & SRE_FLAG_IGNORECASE:
+ emit(OPCODES[OP_IGNORE[op]])
else:
- append(CODES[op])
- append(av)
+ emit(OPCODES[op])
+ emit(av-1)
elif op is IN:
- if "i" in flags:
- append(CODES[MAP_IGNORE[op]])
- def fixup(literal):
- return ord(_lower(literal))
+ if flags & SRE_FLAG_IGNORECASE:
+ emit(OPCODES[OP_IGNORE[op]])
+ def fixup(literal, flags=flags):
+ return _sre.getlower(ord(literal), flags)
else:
- append(CODES[op])
+ emit(OPCODES[op])
fixup = ord
- skip = len(code); append(0)
+ skip = len(code); emit(0)
for op, av in av:
- append(CODES[op])
+ emit(OPCODES[op])
if op is NEGATE:
pass
elif op is LITERAL:
- append(fixup(av))
+ emit(fixup(av))
elif op is RANGE:
- append(fixup(av[0]))
- append(fixup(av[1]))
+ emit(fixup(av[0]))
+ emit(fixup(av[1]))
elif op is CATEGORY:
- append(CATEGORIES[av])
+ if flags & SRE_FLAG_LOCALE:
+ emit(CH_LOCALE[CHCODES[av]])
+ elif flags & SRE_FLAG_UNICODE:
+ emit(CH_UNICODE[CHCODES[av]])
+ else:
+ emit(CHCODES[av])
else:
- raise ValueError, "unsupported set operator"
- append(CODES[FAILURE])
+ raise error, "internal: unsupported set operator"
+ emit(OPCODES[FAILURE])
code[skip] = len(code) - skip
elif op in (LITERAL, NOT_LITERAL):
- if "i" in flags:
- append(CODES[MAP_IGNORE[op]])
- append(ord(_lower(av)))
+ if flags & SRE_FLAG_IGNORECASE:
+ emit(OPCODES[OP_IGNORE[op]])
else:
- append(CODES[op])
- append(ord(av))
+ emit(OPCODES[op])
+ emit(ord(av))
elif op is MARK:
- append(CODES[op])
- append(av)
+ emit(OPCODES[op])
+ emit(av)
elif op in (REPEAT, MIN_REPEAT, MAX_REPEAT):
- lo, hi = av[2].getwidth()
- if lo == 0:
- raise SyntaxError, "cannot repeat zero-width items"
- if lo == hi == 1 and op is MAX_REPEAT:
- append(CODES[MAX_REPEAT_ONE])
- skip = len(code); append(0)
- append(av[0])
- append(av[1])
+ if flags & SRE_FLAG_TEMPLATE:
+ emit(OPCODES[REPEAT])
+ skip = len(code); emit(0)
+ emit(av[0])
+ emit(av[1])
_compile(code, av[2], flags)
- append(CODES[SUCCESS])
+ emit(OPCODES[SUCCESS])
code[skip] = len(code) - skip
else:
- append(CODES[op])
- skip = len(code); append(0)
- append(av[0])
- append(av[1])
- _compile(code, av[2], flags)
- if op is MIN_REPEAT:
- append(CODES[MIN_UNTIL])
+ lo, hi = av[2].getwidth()
+ if lo == 0:
+ raise error, "nothing to repeat"
+ if 0 and lo == hi == 1 and op is MAX_REPEAT:
+ # FIXME: <fl> need a better way to figure out when
+ # it's safe to use this one (in the parser, probably)
+ emit(OPCODES[MAX_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
else:
- # FIXME: MAX_REPEAT PROBABLY DOESN'T WORK (?)
- append(CODES[MAX_UNTIL])
- code[skip] = len(code) - skip
+ emit(OPCODES[op])
+ skip = len(code); emit(0)
+ emit(av[0])
+ emit(av[1])
+ _compile(code, av[2], flags)
+ emit(OPCODES[SUCCESS])
+ code[skip] = len(code) - skip
elif op is SUBPATTERN:
-## group = av[0]
-## if group:
-## append(CODES[MARK])
-## append((group-1)*2)
+ group = av[0]
+ if group:
+ emit(OPCODES[MARK])
+ emit((group-1)*2)
_compile(code, av[1], flags)
-## if group:
-## append(CODES[MARK])
-## append((group-1)*2+1)
+ if group:
+ emit(OPCODES[MARK])
+ emit((group-1)*2+1)
else:
raise ValueError, ("unsupported operand type", op)
-def compile(p, flags=()):
- # convert pattern list to internal format
+def compile(p, flags=0):
+ # internal: convert pattern list to internal format
if type(p) in (type(""), type(u"")):
import sre_parse
pattern = p
p = sre_parse.parse(p)
else:
pattern = None
- # print p.getwidth()
- # print p
- code = Code()
- _compile(code, p.data, p.pattern.flags)
- code.append(CODES[SUCCESS])
- # print list(code.data)
- data = code.todata()
- if 0: # debugging
- print
- print "-" * 68
- import sre_disasm
- sre_disasm.disasm(data)
- print "-" * 68
- # print len(data), p.pattern.groups, len(p.pattern.groupdict)
- return _sre.compile(pattern, data, p.pattern.groups-1, p.pattern.groupdict)
+ flags = p.pattern.flags | flags
+ code = []
+ _compile(code, p.data, flags)
+ code.append(OPCODES[SUCCESS])
+ # FIXME: <fl> get rid of this limitation
+ assert p.pattern.groups <= 100,\
+ "sorry, but this version only supports 100 named groups"
+ return _sre.compile(
+ pattern, flags,
+ array.array(WORDSIZE, code).tostring(),
+ p.pattern.groups-1, p.pattern.groupdict
+ )