blob: c6398bfb83a5763df1ce9b383274c229e4faa567 [file] [log] [blame]
Olivier Deprezf4ef2d02021-04-20 13:36:24 +02001#
2# Secret Labs' Regular Expression Engine
3#
4# convert template to internal format
5#
6# Copyright (c) 1997-2001 by Secret Labs AB. All rights reserved.
7#
8# See the sre.py file for information on usage and redistribution.
9#
10
11"""Internal support module for sre"""
12
13import _sre
14import sre_parse
15from sre_constants import *
16
17assert _sre.MAGIC == MAGIC, "SRE module mismatch"
18
19_LITERAL_CODES = {LITERAL, NOT_LITERAL}
20_REPEATING_CODES = {REPEAT, MIN_REPEAT, MAX_REPEAT}
21_SUCCESS_CODES = {SUCCESS, FAILURE}
22_ASSERT_CODES = {ASSERT, ASSERT_NOT}
23_UNIT_CODES = _LITERAL_CODES | {ANY, IN}
24
25# Sets of lowercase characters which have the same uppercase.
26_equivalences = (
27 # LATIN SMALL LETTER I, LATIN SMALL LETTER DOTLESS I
28 (0x69, 0x131), # iı
29 # LATIN SMALL LETTER S, LATIN SMALL LETTER LONG S
30 (0x73, 0x17f), # sſ
31 # MICRO SIGN, GREEK SMALL LETTER MU
32 (0xb5, 0x3bc), # µμ
33 # COMBINING GREEK YPOGEGRAMMENI, GREEK SMALL LETTER IOTA, GREEK PROSGEGRAMMENI
34 (0x345, 0x3b9, 0x1fbe), # \u0345ιι
35 # GREEK SMALL LETTER IOTA WITH DIALYTIKA AND TONOS, GREEK SMALL LETTER IOTA WITH DIALYTIKA AND OXIA
36 (0x390, 0x1fd3), # ΐΐ
37 # GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND TONOS, GREEK SMALL LETTER UPSILON WITH DIALYTIKA AND OXIA
38 (0x3b0, 0x1fe3), # ΰΰ
39 # GREEK SMALL LETTER BETA, GREEK BETA SYMBOL
40 (0x3b2, 0x3d0), # βϐ
41 # GREEK SMALL LETTER EPSILON, GREEK LUNATE EPSILON SYMBOL
42 (0x3b5, 0x3f5), # εϵ
43 # GREEK SMALL LETTER THETA, GREEK THETA SYMBOL
44 (0x3b8, 0x3d1), # θϑ
45 # GREEK SMALL LETTER KAPPA, GREEK KAPPA SYMBOL
46 (0x3ba, 0x3f0), # κϰ
47 # GREEK SMALL LETTER PI, GREEK PI SYMBOL
48 (0x3c0, 0x3d6), # πϖ
49 # GREEK SMALL LETTER RHO, GREEK RHO SYMBOL
50 (0x3c1, 0x3f1), # ρϱ
51 # GREEK SMALL LETTER FINAL SIGMA, GREEK SMALL LETTER SIGMA
52 (0x3c2, 0x3c3), # ςσ
53 # GREEK SMALL LETTER PHI, GREEK PHI SYMBOL
54 (0x3c6, 0x3d5), # φϕ
55 # LATIN SMALL LETTER S WITH DOT ABOVE, LATIN SMALL LETTER LONG S WITH DOT ABOVE
56 (0x1e61, 0x1e9b), # ṡẛ
57 # LATIN SMALL LIGATURE LONG S T, LATIN SMALL LIGATURE ST
58 (0xfb05, 0xfb06), # ſtst
59)
60
61# Maps the lowercase code to lowercase codes which have the same uppercase.
62_ignorecase_fixes = {i: tuple(j for j in t if i != j)
63 for t in _equivalences for i in t}
64
65def _combine_flags(flags, add_flags, del_flags,
66 TYPE_FLAGS=sre_parse.TYPE_FLAGS):
67 if add_flags & TYPE_FLAGS:
68 flags &= ~TYPE_FLAGS
69 return (flags | add_flags) & ~del_flags
70
71def _compile(code, pattern, flags):
72 # internal: compile a (sub)pattern
73 emit = code.append
74 _len = len
75 LITERAL_CODES = _LITERAL_CODES
76 REPEATING_CODES = _REPEATING_CODES
77 SUCCESS_CODES = _SUCCESS_CODES
78 ASSERT_CODES = _ASSERT_CODES
79 iscased = None
80 tolower = None
81 fixes = None
82 if flags & SRE_FLAG_IGNORECASE and not flags & SRE_FLAG_LOCALE:
83 if flags & SRE_FLAG_UNICODE:
84 iscased = _sre.unicode_iscased
85 tolower = _sre.unicode_tolower
86 fixes = _ignorecase_fixes
87 else:
88 iscased = _sre.ascii_iscased
89 tolower = _sre.ascii_tolower
90 for op, av in pattern:
91 if op in LITERAL_CODES:
92 if not flags & SRE_FLAG_IGNORECASE:
93 emit(op)
94 emit(av)
95 elif flags & SRE_FLAG_LOCALE:
96 emit(OP_LOCALE_IGNORE[op])
97 emit(av)
98 elif not iscased(av):
99 emit(op)
100 emit(av)
101 else:
102 lo = tolower(av)
103 if not fixes: # ascii
104 emit(OP_IGNORE[op])
105 emit(lo)
106 elif lo not in fixes:
107 emit(OP_UNICODE_IGNORE[op])
108 emit(lo)
109 else:
110 emit(IN_UNI_IGNORE)
111 skip = _len(code); emit(0)
112 if op is NOT_LITERAL:
113 emit(NEGATE)
114 for k in (lo,) + fixes[lo]:
115 emit(LITERAL)
116 emit(k)
117 emit(FAILURE)
118 code[skip] = _len(code) - skip
119 elif op is IN:
120 charset, hascased = _optimize_charset(av, iscased, tolower, fixes)
121 if flags & SRE_FLAG_IGNORECASE and flags & SRE_FLAG_LOCALE:
122 emit(IN_LOC_IGNORE)
123 elif not hascased:
124 emit(IN)
125 elif not fixes: # ascii
126 emit(IN_IGNORE)
127 else:
128 emit(IN_UNI_IGNORE)
129 skip = _len(code); emit(0)
130 _compile_charset(charset, flags, code)
131 code[skip] = _len(code) - skip
132 elif op is ANY:
133 if flags & SRE_FLAG_DOTALL:
134 emit(ANY_ALL)
135 else:
136 emit(ANY)
137 elif op in REPEATING_CODES:
138 if flags & SRE_FLAG_TEMPLATE:
139 raise error("internal: unsupported template operator %r" % (op,))
140 if _simple(av[2]):
141 if op is MAX_REPEAT:
142 emit(REPEAT_ONE)
143 else:
144 emit(MIN_REPEAT_ONE)
145 skip = _len(code); emit(0)
146 emit(av[0])
147 emit(av[1])
148 _compile(code, av[2], flags)
149 emit(SUCCESS)
150 code[skip] = _len(code) - skip
151 else:
152 emit(REPEAT)
153 skip = _len(code); emit(0)
154 emit(av[0])
155 emit(av[1])
156 _compile(code, av[2], flags)
157 code[skip] = _len(code) - skip
158 if op is MAX_REPEAT:
159 emit(MAX_UNTIL)
160 else:
161 emit(MIN_UNTIL)
162 elif op is SUBPATTERN:
163 group, add_flags, del_flags, p = av
164 if group:
165 emit(MARK)
166 emit((group-1)*2)
167 # _compile_info(code, p, _combine_flags(flags, add_flags, del_flags))
168 _compile(code, p, _combine_flags(flags, add_flags, del_flags))
169 if group:
170 emit(MARK)
171 emit((group-1)*2+1)
172 elif op in SUCCESS_CODES:
173 emit(op)
174 elif op in ASSERT_CODES:
175 emit(op)
176 skip = _len(code); emit(0)
177 if av[0] >= 0:
178 emit(0) # look ahead
179 else:
180 lo, hi = av[1].getwidth()
181 if lo != hi:
182 raise error("look-behind requires fixed-width pattern")
183 emit(lo) # look behind
184 _compile(code, av[1], flags)
185 emit(SUCCESS)
186 code[skip] = _len(code) - skip
187 elif op is CALL:
188 emit(op)
189 skip = _len(code); emit(0)
190 _compile(code, av, flags)
191 emit(SUCCESS)
192 code[skip] = _len(code) - skip
193 elif op is AT:
194 emit(op)
195 if flags & SRE_FLAG_MULTILINE:
196 av = AT_MULTILINE.get(av, av)
197 if flags & SRE_FLAG_LOCALE:
198 av = AT_LOCALE.get(av, av)
199 elif flags & SRE_FLAG_UNICODE:
200 av = AT_UNICODE.get(av, av)
201 emit(av)
202 elif op is BRANCH:
203 emit(op)
204 tail = []
205 tailappend = tail.append
206 for av in av[1]:
207 skip = _len(code); emit(0)
208 # _compile_info(code, av, flags)
209 _compile(code, av, flags)
210 emit(JUMP)
211 tailappend(_len(code)); emit(0)
212 code[skip] = _len(code) - skip
213 emit(FAILURE) # end of branch
214 for tail in tail:
215 code[tail] = _len(code) - tail
216 elif op is CATEGORY:
217 emit(op)
218 if flags & SRE_FLAG_LOCALE:
219 av = CH_LOCALE[av]
220 elif flags & SRE_FLAG_UNICODE:
221 av = CH_UNICODE[av]
222 emit(av)
223 elif op is GROUPREF:
224 if not flags & SRE_FLAG_IGNORECASE:
225 emit(op)
226 elif flags & SRE_FLAG_LOCALE:
227 emit(GROUPREF_LOC_IGNORE)
228 elif not fixes: # ascii
229 emit(GROUPREF_IGNORE)
230 else:
231 emit(GROUPREF_UNI_IGNORE)
232 emit(av-1)
233 elif op is GROUPREF_EXISTS:
234 emit(op)
235 emit(av[0]-1)
236 skipyes = _len(code); emit(0)
237 _compile(code, av[1], flags)
238 if av[2]:
239 emit(JUMP)
240 skipno = _len(code); emit(0)
241 code[skipyes] = _len(code) - skipyes + 1
242 _compile(code, av[2], flags)
243 code[skipno] = _len(code) - skipno
244 else:
245 code[skipyes] = _len(code) - skipyes + 1
246 else:
247 raise error("internal: unsupported operand type %r" % (op,))
248
249def _compile_charset(charset, flags, code):
250 # compile charset subprogram
251 emit = code.append
252 for op, av in charset:
253 emit(op)
254 if op is NEGATE:
255 pass
256 elif op is LITERAL:
257 emit(av)
258 elif op is RANGE or op is RANGE_UNI_IGNORE:
259 emit(av[0])
260 emit(av[1])
261 elif op is CHARSET:
262 code.extend(av)
263 elif op is BIGCHARSET:
264 code.extend(av)
265 elif op is CATEGORY:
266 if flags & SRE_FLAG_LOCALE:
267 emit(CH_LOCALE[av])
268 elif flags & SRE_FLAG_UNICODE:
269 emit(CH_UNICODE[av])
270 else:
271 emit(av)
272 else:
273 raise error("internal: unsupported set operator %r" % (op,))
274 emit(FAILURE)
275
276def _optimize_charset(charset, iscased=None, fixup=None, fixes=None):
277 # internal: optimize character set
278 out = []
279 tail = []
280 charmap = bytearray(256)
281 hascased = False
282 for op, av in charset:
283 while True:
284 try:
285 if op is LITERAL:
286 if fixup:
287 lo = fixup(av)
288 charmap[lo] = 1
289 if fixes and lo in fixes:
290 for k in fixes[lo]:
291 charmap[k] = 1
292 if not hascased and iscased(av):
293 hascased = True
294 else:
295 charmap[av] = 1
296 elif op is RANGE:
297 r = range(av[0], av[1]+1)
298 if fixup:
299 if fixes:
300 for i in map(fixup, r):
301 charmap[i] = 1
302 if i in fixes:
303 for k in fixes[i]:
304 charmap[k] = 1
305 else:
306 for i in map(fixup, r):
307 charmap[i] = 1
308 if not hascased:
309 hascased = any(map(iscased, r))
310 else:
311 for i in r:
312 charmap[i] = 1
313 elif op is NEGATE:
314 out.append((op, av))
315 else:
316 tail.append((op, av))
317 except IndexError:
318 if len(charmap) == 256:
319 # character set contains non-UCS1 character codes
320 charmap += b'\0' * 0xff00
321 continue
322 # Character set contains non-BMP character codes.
323 if fixup:
324 hascased = True
325 # There are only two ranges of cased non-BMP characters:
326 # 10400-1044F (Deseret) and 118A0-118DF (Warang Citi),
327 # and for both ranges RANGE_UNI_IGNORE works.
328 if op is RANGE:
329 op = RANGE_UNI_IGNORE
330 tail.append((op, av))
331 break
332
333 # compress character map
334 runs = []
335 q = 0
336 while True:
337 p = charmap.find(1, q)
338 if p < 0:
339 break
340 if len(runs) >= 2:
341 runs = None
342 break
343 q = charmap.find(0, p)
344 if q < 0:
345 runs.append((p, len(charmap)))
346 break
347 runs.append((p, q))
348 if runs is not None:
349 # use literal/range
350 for p, q in runs:
351 if q - p == 1:
352 out.append((LITERAL, p))
353 else:
354 out.append((RANGE, (p, q - 1)))
355 out += tail
356 # if the case was changed or new representation is more compact
357 if hascased or len(out) < len(charset):
358 return out, hascased
359 # else original character set is good enough
360 return charset, hascased
361
362 # use bitmap
363 if len(charmap) == 256:
364 data = _mk_bitmap(charmap)
365 out.append((CHARSET, data))
366 out += tail
367 return out, hascased
368
369 # To represent a big charset, first a bitmap of all characters in the
370 # set is constructed. Then, this bitmap is sliced into chunks of 256
371 # characters, duplicate chunks are eliminated, and each chunk is
372 # given a number. In the compiled expression, the charset is
373 # represented by a 32-bit word sequence, consisting of one word for
374 # the number of different chunks, a sequence of 256 bytes (64 words)
375 # of chunk numbers indexed by their original chunk position, and a
376 # sequence of 256-bit chunks (8 words each).
377
378 # Compression is normally good: in a typical charset, large ranges of
379 # Unicode will be either completely excluded (e.g. if only cyrillic
380 # letters are to be matched), or completely included (e.g. if large
381 # subranges of Kanji match). These ranges will be represented by
382 # chunks of all one-bits or all zero-bits.
383
384 # Matching can be also done efficiently: the more significant byte of
385 # the Unicode character is an index into the chunk number, and the
386 # less significant byte is a bit index in the chunk (just like the
387 # CHARSET matching).
388
389 charmap = bytes(charmap) # should be hashable
390 comps = {}
391 mapping = bytearray(256)
392 block = 0
393 data = bytearray()
394 for i in range(0, 65536, 256):
395 chunk = charmap[i: i + 256]
396 if chunk in comps:
397 mapping[i // 256] = comps[chunk]
398 else:
399 mapping[i // 256] = comps[chunk] = block
400 block += 1
401 data += chunk
402 data = _mk_bitmap(data)
403 data[0:0] = [block] + _bytes_to_codes(mapping)
404 out.append((BIGCHARSET, data))
405 out += tail
406 return out, hascased
407
408_CODEBITS = _sre.CODESIZE * 8
409MAXCODE = (1 << _CODEBITS) - 1
410_BITS_TRANS = b'0' + b'1' * 255
411def _mk_bitmap(bits, _CODEBITS=_CODEBITS, _int=int):
412 s = bits.translate(_BITS_TRANS)[::-1]
413 return [_int(s[i - _CODEBITS: i], 2)
414 for i in range(len(s), 0, -_CODEBITS)]
415
416def _bytes_to_codes(b):
417 # Convert block indices to word array
418 a = memoryview(b).cast('I')
419 assert a.itemsize == _sre.CODESIZE
420 assert len(a) * a.itemsize == len(b)
421 return a.tolist()
422
423def _simple(p):
424 # check if this subpattern is a "simple" operator
425 if len(p) != 1:
426 return False
427 op, av = p[0]
428 if op is SUBPATTERN:
429 return av[0] is None and _simple(av[-1])
430 return op in _UNIT_CODES
431
432def _generate_overlap_table(prefix):
433 """
434 Generate an overlap table for the following prefix.
435 An overlap table is a table of the same size as the prefix which
436 informs about the potential self-overlap for each index in the prefix:
437 - if overlap[i] == 0, prefix[i:] can't overlap prefix[0:...]
438 - if overlap[i] == k with 0 < k <= i, prefix[i-k+1:i+1] overlaps with
439 prefix[0:k]
440 """
441 table = [0] * len(prefix)
442 for i in range(1, len(prefix)):
443 idx = table[i - 1]
444 while prefix[i] != prefix[idx]:
445 if idx == 0:
446 table[i] = 0
447 break
448 idx = table[idx - 1]
449 else:
450 table[i] = idx + 1
451 return table
452
453def _get_iscased(flags):
454 if not flags & SRE_FLAG_IGNORECASE:
455 return None
456 elif flags & SRE_FLAG_UNICODE:
457 return _sre.unicode_iscased
458 else:
459 return _sre.ascii_iscased
460
461def _get_literal_prefix(pattern, flags):
462 # look for literal prefix
463 prefix = []
464 prefixappend = prefix.append
465 prefix_skip = None
466 iscased = _get_iscased(flags)
467 for op, av in pattern.data:
468 if op is LITERAL:
469 if iscased and iscased(av):
470 break
471 prefixappend(av)
472 elif op is SUBPATTERN:
473 group, add_flags, del_flags, p = av
474 flags1 = _combine_flags(flags, add_flags, del_flags)
475 if flags1 & SRE_FLAG_IGNORECASE and flags1 & SRE_FLAG_LOCALE:
476 break
477 prefix1, prefix_skip1, got_all = _get_literal_prefix(p, flags1)
478 if prefix_skip is None:
479 if group is not None:
480 prefix_skip = len(prefix)
481 elif prefix_skip1 is not None:
482 prefix_skip = len(prefix) + prefix_skip1
483 prefix.extend(prefix1)
484 if not got_all:
485 break
486 else:
487 break
488 else:
489 return prefix, prefix_skip, True
490 return prefix, prefix_skip, False
491
492def _get_charset_prefix(pattern, flags):
493 while True:
494 if not pattern.data:
495 return None
496 op, av = pattern.data[0]
497 if op is not SUBPATTERN:
498 break
499 group, add_flags, del_flags, pattern = av
500 flags = _combine_flags(flags, add_flags, del_flags)
501 if flags & SRE_FLAG_IGNORECASE and flags & SRE_FLAG_LOCALE:
502 return None
503
504 iscased = _get_iscased(flags)
505 if op is LITERAL:
506 if iscased and iscased(av):
507 return None
508 return [(op, av)]
509 elif op is BRANCH:
510 charset = []
511 charsetappend = charset.append
512 for p in av[1]:
513 if not p:
514 return None
515 op, av = p[0]
516 if op is LITERAL and not (iscased and iscased(av)):
517 charsetappend((op, av))
518 else:
519 return None
520 return charset
521 elif op is IN:
522 charset = av
523 if iscased:
524 for op, av in charset:
525 if op is LITERAL:
526 if iscased(av):
527 return None
528 elif op is RANGE:
529 if av[1] > 0xffff:
530 return None
531 if any(map(iscased, range(av[0], av[1]+1))):
532 return None
533 return charset
534 return None
535
536def _compile_info(code, pattern, flags):
537 # internal: compile an info block. in the current version,
538 # this contains min/max pattern width, and an optional literal
539 # prefix or a character map
540 lo, hi = pattern.getwidth()
541 if hi > MAXCODE:
542 hi = MAXCODE
543 if lo == 0:
544 code.extend([INFO, 4, 0, lo, hi])
545 return
546 # look for a literal prefix
547 prefix = []
548 prefix_skip = 0
549 charset = [] # not used
550 if not (flags & SRE_FLAG_IGNORECASE and flags & SRE_FLAG_LOCALE):
551 # look for literal prefix
552 prefix, prefix_skip, got_all = _get_literal_prefix(pattern, flags)
553 # if no prefix, look for charset prefix
554 if not prefix:
555 charset = _get_charset_prefix(pattern, flags)
556## if prefix:
557## print("*** PREFIX", prefix, prefix_skip)
558## if charset:
559## print("*** CHARSET", charset)
560 # add an info block
561 emit = code.append
562 emit(INFO)
563 skip = len(code); emit(0)
564 # literal flag
565 mask = 0
566 if prefix:
567 mask = SRE_INFO_PREFIX
568 if prefix_skip is None and got_all:
569 mask = mask | SRE_INFO_LITERAL
570 elif charset:
571 mask = mask | SRE_INFO_CHARSET
572 emit(mask)
573 # pattern length
574 if lo < MAXCODE:
575 emit(lo)
576 else:
577 emit(MAXCODE)
578 prefix = prefix[:MAXCODE]
579 emit(min(hi, MAXCODE))
580 # add literal prefix
581 if prefix:
582 emit(len(prefix)) # length
583 if prefix_skip is None:
584 prefix_skip = len(prefix)
585 emit(prefix_skip) # skip
586 code.extend(prefix)
587 # generate overlap table
588 code.extend(_generate_overlap_table(prefix))
589 elif charset:
590 charset, hascased = _optimize_charset(charset)
591 assert not hascased
592 _compile_charset(charset, flags, code)
593 code[skip] = len(code) - skip
594
595def isstring(obj):
596 return isinstance(obj, (str, bytes))
597
598def _code(p, flags):
599
600 flags = p.state.flags | flags
601 code = []
602
603 # compile info block
604 _compile_info(code, p, flags)
605
606 # compile the pattern
607 _compile(code, p.data, flags)
608
609 code.append(SUCCESS)
610
611 return code
612
613def _hex_code(code):
614 return '[%s]' % ', '.join('%#0*x' % (_sre.CODESIZE*2+2, x) for x in code)
615
616def dis(code):
617 import sys
618
619 labels = set()
620 level = 0
621 offset_width = len(str(len(code) - 1))
622
623 def dis_(start, end):
624 def print_(*args, to=None):
625 if to is not None:
626 labels.add(to)
627 args += ('(to %d)' % (to,),)
628 print('%*d%s ' % (offset_width, start, ':' if start in labels else '.'),
629 end=' '*(level-1))
630 print(*args)
631
632 def print_2(*args):
633 print(end=' '*(offset_width + 2*level))
634 print(*args)
635
636 nonlocal level
637 level += 1
638 i = start
639 while i < end:
640 start = i
641 op = code[i]
642 i += 1
643 op = OPCODES[op]
644 if op in (SUCCESS, FAILURE, ANY, ANY_ALL,
645 MAX_UNTIL, MIN_UNTIL, NEGATE):
646 print_(op)
647 elif op in (LITERAL, NOT_LITERAL,
648 LITERAL_IGNORE, NOT_LITERAL_IGNORE,
649 LITERAL_UNI_IGNORE, NOT_LITERAL_UNI_IGNORE,
650 LITERAL_LOC_IGNORE, NOT_LITERAL_LOC_IGNORE):
651 arg = code[i]
652 i += 1
653 print_(op, '%#02x (%r)' % (arg, chr(arg)))
654 elif op is AT:
655 arg = code[i]
656 i += 1
657 arg = str(ATCODES[arg])
658 assert arg[:3] == 'AT_'
659 print_(op, arg[3:])
660 elif op is CATEGORY:
661 arg = code[i]
662 i += 1
663 arg = str(CHCODES[arg])
664 assert arg[:9] == 'CATEGORY_'
665 print_(op, arg[9:])
666 elif op in (IN, IN_IGNORE, IN_UNI_IGNORE, IN_LOC_IGNORE):
667 skip = code[i]
668 print_(op, skip, to=i+skip)
669 dis_(i+1, i+skip)
670 i += skip
671 elif op in (RANGE, RANGE_UNI_IGNORE):
672 lo, hi = code[i: i+2]
673 i += 2
674 print_(op, '%#02x %#02x (%r-%r)' % (lo, hi, chr(lo), chr(hi)))
675 elif op is CHARSET:
676 print_(op, _hex_code(code[i: i + 256//_CODEBITS]))
677 i += 256//_CODEBITS
678 elif op is BIGCHARSET:
679 arg = code[i]
680 i += 1
681 mapping = list(b''.join(x.to_bytes(_sre.CODESIZE, sys.byteorder)
682 for x in code[i: i + 256//_sre.CODESIZE]))
683 print_(op, arg, mapping)
684 i += 256//_sre.CODESIZE
685 level += 1
686 for j in range(arg):
687 print_2(_hex_code(code[i: i + 256//_CODEBITS]))
688 i += 256//_CODEBITS
689 level -= 1
690 elif op in (MARK, GROUPREF, GROUPREF_IGNORE, GROUPREF_UNI_IGNORE,
691 GROUPREF_LOC_IGNORE):
692 arg = code[i]
693 i += 1
694 print_(op, arg)
695 elif op is JUMP:
696 skip = code[i]
697 print_(op, skip, to=i+skip)
698 i += 1
699 elif op is BRANCH:
700 skip = code[i]
701 print_(op, skip, to=i+skip)
702 while skip:
703 dis_(i+1, i+skip)
704 i += skip
705 start = i
706 skip = code[i]
707 if skip:
708 print_('branch', skip, to=i+skip)
709 else:
710 print_(FAILURE)
711 i += 1
712 elif op in (REPEAT, REPEAT_ONE, MIN_REPEAT_ONE):
713 skip, min, max = code[i: i+3]
714 if max == MAXREPEAT:
715 max = 'MAXREPEAT'
716 print_(op, skip, min, max, to=i+skip)
717 dis_(i+3, i+skip)
718 i += skip
719 elif op is GROUPREF_EXISTS:
720 arg, skip = code[i: i+2]
721 print_(op, arg, skip, to=i+skip)
722 i += 2
723 elif op in (ASSERT, ASSERT_NOT):
724 skip, arg = code[i: i+2]
725 print_(op, skip, arg, to=i+skip)
726 dis_(i+2, i+skip)
727 i += skip
728 elif op is INFO:
729 skip, flags, min, max = code[i: i+4]
730 if max == MAXREPEAT:
731 max = 'MAXREPEAT'
732 print_(op, skip, bin(flags), min, max, to=i+skip)
733 start = i+4
734 if flags & SRE_INFO_PREFIX:
735 prefix_len, prefix_skip = code[i+4: i+6]
736 print_2(' prefix_skip', prefix_skip)
737 start = i + 6
738 prefix = code[start: start+prefix_len]
739 print_2(' prefix',
740 '[%s]' % ', '.join('%#02x' % x for x in prefix),
741 '(%r)' % ''.join(map(chr, prefix)))
742 start += prefix_len
743 print_2(' overlap', code[start: start+prefix_len])
744 start += prefix_len
745 if flags & SRE_INFO_CHARSET:
746 level += 1
747 print_2('in')
748 dis_(start, i+skip)
749 level -= 1
750 i += skip
751 else:
752 raise ValueError(op)
753
754 level -= 1
755
756 dis_(0, len(code))
757
758
759def compile(p, flags=0):
760 # internal: convert pattern list to internal format
761
762 if isstring(p):
763 pattern = p
764 p = sre_parse.parse(p, flags)
765 else:
766 pattern = None
767
768 code = _code(p, flags)
769
770 if flags & SRE_FLAG_DEBUG:
771 print()
772 dis(code)
773
774 # map in either direction
775 groupindex = p.state.groupdict
776 indexgroup = [None] * p.state.groups
777 for k, i in groupindex.items():
778 indexgroup[i] = k
779
780 return _sre.compile(
781 pattern, flags | p.state.flags, code,
782 p.state.groups-1,
783 groupindex, tuple(indexgroup)
784 )