1 /* -*- buffer-read-only: t -*-
2 !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
3 This file is built by regen/regcomp.pl from regcomp.sym.
4 Any changes made here will be lost!
7 /* Regops and State definitions */
9 #define REGNODE_MAX 109
10 #define REGMATCH_STATE_MAX 149
12 #define END 0 /* 0000 End of program. */
13 #define SUCCEED 1 /* 0x01 Return from a subroutine, basically. */
14 #define SBOL 2 /* 0x02 Match "" at beginning of line: /^/, /\A/ */
15 #define BOL 2 /* 0x02 type alias */
16 #define MBOL 3 /* 0x03 Same, assuming multiline: /^/m */
17 #define SEOL 4 /* 0x04 Match "" at end of line: /$/ */
18 #define EOL 4 /* 0x04 type alias */
19 #define MEOL 5 /* 0x05 Same, assuming multiline: /$/m */
20 #define EOS 6 /* 0x06 Match "" at end of string: /\z/ */
21 #define GPOS 7 /* 0x07 Matches where last m//g left off. */
22 #define BOUND 8 /* 0x08 Like BOUNDA for non-utf8, otherwise like BOUNDU */
23 #define BOUNDL 9 /* 0x09 Like BOUND/BOUNDU, but \w and \W are defined by current locale */
24 #define BOUNDU 10 /* 0x0a Match "" at any boundary of a given type using /u rules. */
25 #define BOUNDA 11 /* 0x0b Match "" at any boundary between \w\W or \W\w, where \w is [_a-zA-Z0-9] */
26 #define NBOUND 12 /* 0x0c Like NBOUNDA for non-utf8, otherwise like BOUNDU */
27 #define NBOUNDL 13 /* 0x0d Like NBOUND/NBOUNDU, but \w and \W are defined by current locale */
28 #define NBOUNDU 14 /* 0x0e Match "" at any non-boundary of a given type using using /u rules. */
29 #define NBOUNDA 15 /* 0x0f Match "" betweeen any \w\w or \W\W, where \w is [_a-zA-Z0-9] */
30 #define REG_ANY 16 /* 0x10 Match any one character (except newline). */
31 #define SANY 17 /* 0x11 Match any one character. */
32 #define ANYOF 18 /* 0x12 Match character in (or not in) this class, single char match only */
33 #define ANYOFD 19 /* 0x13 Like ANYOF, but /d is in effect */
34 #define ANYOFL 20 /* 0x14 Like ANYOF, but /l is in effect */
35 #define ANYOFPOSIXL 21 /* 0x15 Like ANYOFL, but matches [[:posix:]] classes */
36 #define ANYOFH 22 /* 0x16 Like ANYOF, but only has "High" matches, none in the bitmap; the flags field contains the lowest matchable UTF-8 start byte */
37 #define ANYOFHb 23 /* 0x17 Like ANYOFH, but all matches share the same UTF-8 start byte, given in the flags field */
38 #define ANYOFHr 24 /* 0x18 Like ANYOFH, but the flags field contains packed bounds for all matchable UTF-8 start bytes. */
39 #define ANYOFHs 25 /* 0x19 Like ANYOFHb, but has a string field that gives the leading matchable UTF-8 bytes; flags field is len */
40 #define ANYOFR 26 /* 0x1a Matches any character in the range given by its packed args: upper 12 bits is the max delta from the base lower 20; the flags field contains the lowest matchable UTF-8 start byte */
41 #define ANYOFRb 27 /* 0x1b Like ANYOFR, but all matches share the same UTF-8 start byte, given in the flags field */
42 #define ANYOFM 28 /* 0x1c Like ANYOF, but matches an invariant byte as determined by the mask and arg */
43 #define NANYOFM 29 /* 0x1d complement of ANYOFM */
44 #define POSIXD 30 /* 0x1e Some [[:class:]] under /d; the FLAGS field gives which one */
45 #define POSIXL 31 /* 0x1f Some [[:class:]] under /l; the FLAGS field gives which one */
46 #define POSIXU 32 /* 0x20 Some [[:class:]] under /u; the FLAGS field gives which one */
47 #define POSIXA 33 /* 0x21 Some [[:class:]] under /a; the FLAGS field gives which one */
48 #define NPOSIXD 34 /* 0x22 complement of POSIXD, [[:^class:]] */
49 #define NPOSIXL 35 /* 0x23 complement of POSIXL, [[:^class:]] */
50 #define NPOSIXU 36 /* 0x24 complement of POSIXU, [[:^class:]] */
51 #define NPOSIXA 37 /* 0x25 complement of POSIXA, [[:^class:]] */
52 #define CLUMP 38 /* 0x26 Match any extended grapheme cluster sequence */
53 #define BRANCH 39 /* 0x27 Match this alternative, or the next... */
54 #define EXACT 40 /* 0x28 Match this string (flags field is the length). */
55 #define LEXACT 41 /* 0x29 Match this long string (preceded by length; flags unused). */
56 #define EXACTL 42 /* 0x2a Like EXACT, but /l is in effect (used so locale-related warnings can be checked for) */
57 #define EXACTF 43 /* 0x2b Like EXACT, but match using /id rules; (string not UTF-8, ASCII folded; non-ASCII not) */
58 #define EXACTFL 44 /* 0x2c Like EXACT, but match using /il rules; (string not likely to be folded) */
59 #define EXACTFU 45 /* 0x2d Like EXACT, but match using /iu rules; (string folded) */
60 #define EXACTFAA 46 /* 0x2e Like EXACT, but match using /iaa rules; (string folded except in non-UTF8 patterns: MICRO, SHARP S; folded length <= unfolded) */
61 #define EXACTFUP 47 /* 0x2f Like EXACT, but match using /iu rules; (string not UTF-8, folded except MICRO, SHARP S: hence Problematic) */
62 #define EXACTFLU8 48 /* 0x30 Like EXACTFU, but use /il, UTF-8, (string is folded, and everything in it is above 255 */
63 #define EXACTFAA_NO_TRIE 49 /* 0x31 Like EXACT, but match using /iaa rules (string not UTF-8, not guaranteed to be folded, not currently trie-able) */
64 #define EXACT_REQ8 50 /* 0x32 Like EXACT, but only UTF-8 encoded targets can match */
65 #define LEXACT_REQ8 51 /* 0x33 Like LEXACT, but only UTF-8 encoded targets can match */
66 #define EXACTFU_REQ8 52 /* 0x34 Like EXACTFU, but only UTF-8 encoded targets can match */
67 #define EXACTFU_S_EDGE 53 /* 0x35 /di rules, but nothing in it precludes /ui, except begins and/or ends with [Ss]; (string not UTF-8; compile-time only) */
68 #define NOTHING 54 /* 0x36 Match empty string. */
69 #define TAIL 55 /* 0x37 Match empty string. Can jump here from outside. */
70 #define STAR 56 /* 0x38 Match this (simple) thing 0 or more times. */
71 #define PLUS 57 /* 0x39 Match this (simple) thing 1 or more times. */
72 #define CURLY 58 /* 0x3a Match this simple thing {n,m} times. */
73 #define CURLYN 59 /* 0x3b Capture next-after-this simple thing */
74 #define CURLYM 60 /* 0x3c Capture this medium-complex thing {n,m} times. */
75 #define CURLYX 61 /* 0x3d Match this complex thing {n,m} times. */
76 #define WHILEM 62 /* 0x3e Do curly processing and see if rest matches. */
77 #define OPEN 63 /* 0x3f Mark this point in input as start of #n. */
78 #define CLOSE 64 /* 0x40 Close corresponding OPEN of #n. */
79 #define SROPEN 65 /* 0x41 Same as OPEN, but for script run */
80 #define SRCLOSE 66 /* 0x42 Close preceding SROPEN */
81 #define REF 67 /* 0x43 Match some already matched string */
82 #define REFF 68 /* 0x44 Match already matched string, using /di rules. */
83 #define REFFL 69 /* 0x45 Match already matched string, using /li rules. */
84 #define REFFU 70 /* 0x46 Match already matched string, usng /ui. */
85 #define REFFA 71 /* 0x47 Match already matched string, using /aai rules. */
86 #define REFN 72 /* 0x48 Match some already matched string */
87 #define REFFN 73 /* 0x49 Match already matched string, using /di rules. */
88 #define REFFLN 74 /* 0x4a Match already matched string, using /li rules. */
89 #define REFFUN 75 /* 0x4b Match already matched string, using /ui rules. */
90 #define REFFAN 76 /* 0x4c Match already matched string, using /aai rules. */
91 #define LONGJMP 77 /* 0x4d Jump far away. */
92 #define BRANCHJ 78 /* 0x4e BRANCH with long offset. */
93 #define IFMATCH 79 /* 0x4f Succeeds if the following matches; non-zero flags "f", next_off "o" means lookbehind assertion starting "f..(f-o)" characters before current */
94 #define UNLESSM 80 /* 0x50 Fails if the following matches; non-zero flags "f", next_off "o" means lookbehind assertion starting "f..(f-o)" characters before current */
95 #define SUSPEND 81 /* 0x51 "Independent" sub-RE. */
96 #define IFTHEN 82 /* 0x52 Switch, should be preceded by switcher. */
97 #define GROUPP 83 /* 0x53 Whether the group matched. */
98 #define EVAL 84 /* 0x54 Execute some Perl code. */
99 #define MINMOD 85 /* 0x55 Next operator is not greedy. */
100 #define LOGICAL 86 /* 0x56 Next opcode should set the flag only. */
101 #define RENUM 87 /* 0x57 Group with independently numbered parens. */
102 #define TRIE 88 /* 0x58 Match many EXACT(F[ALU]?)? at once. flags==type */
103 #define TRIEC 89 /* 0x59 Same as TRIE, but with embedded charclass data */
104 #define AHOCORASICK 90 /* 0x5a Aho Corasick stclass. flags==type */
105 #define AHOCORASICKC 91 /* 0x5b Same as AHOCORASICK, but with embedded charclass data */
106 #define GOSUB 92 /* 0x5c recurse to paren arg1 at (signed) ofs arg2 */
107 #define GROUPPN 93 /* 0x5d Whether the group matched. */
108 #define INSUBP 94 /* 0x5e Whether we are in a specific recurse. */
109 #define DEFINEP 95 /* 0x5f Never execute directly. */
110 #define ENDLIKE 96 /* 0x60 Used only for the type field of verbs */
111 #define OPFAIL 97 /* 0x61 Same as (?!), but with verb arg */
112 #define ACCEPT 98 /* 0x62 Accepts the current matched string, with verbar */
113 #define VERB 99 /* 0x63 Used only for the type field of verbs */
114 #define PRUNE 100 /* 0x64 Pattern fails at this startpoint if no-backtracking through this */
115 #define MARKPOINT 101 /* 0x65 Push the current location for rollback by cut. */
116 #define SKIP 102 /* 0x66 On failure skip forward (to the mark) before retrying */
117 #define COMMIT 103 /* 0x67 Pattern fails outright if backtracking through this */
118 #define CUTGROUP 104 /* 0x68 On failure go to the next alternation in the group */
119 #define KEEPS 105 /* 0x69 $& begins here. */
120 #define LNBREAK 106 /* 0x6a generic newline pattern */
121 #define OPTIMIZED 107 /* 0x6b Placeholder for dump. */
122 #define PSEUDO 108 /* 0x6c Pseudo opcode for internal use. */
123 #define REGEX_SET 109 /* 0x6d Regex set, temporary node used in pre-optimization compilation */
124 /* ------------ States ------------- */
125 #define TRIE_next (REGNODE_MAX + 1) /* state for TRIE */
126 #define TRIE_next_fail (REGNODE_MAX + 2) /* state for TRIE */
127 #define EVAL_B (REGNODE_MAX + 3) /* state for EVAL */
128 #define EVAL_B_fail (REGNODE_MAX + 4) /* state for EVAL */
129 #define EVAL_postponed_AB (REGNODE_MAX + 5) /* state for EVAL */
130 #define EVAL_postponed_AB_fail (REGNODE_MAX + 6) /* state for EVAL */
131 #define CURLYX_end (REGNODE_MAX + 7) /* state for CURLYX */
132 #define CURLYX_end_fail (REGNODE_MAX + 8) /* state for CURLYX */
133 #define WHILEM_A_pre (REGNODE_MAX + 9) /* state for WHILEM */
134 #define WHILEM_A_pre_fail (REGNODE_MAX + 10) /* state for WHILEM */
135 #define WHILEM_A_min (REGNODE_MAX + 11) /* state for WHILEM */
136 #define WHILEM_A_min_fail (REGNODE_MAX + 12) /* state for WHILEM */
137 #define WHILEM_A_max (REGNODE_MAX + 13) /* state for WHILEM */
138 #define WHILEM_A_max_fail (REGNODE_MAX + 14) /* state for WHILEM */
139 #define WHILEM_B_min (REGNODE_MAX + 15) /* state for WHILEM */
140 #define WHILEM_B_min_fail (REGNODE_MAX + 16) /* state for WHILEM */
141 #define WHILEM_B_max (REGNODE_MAX + 17) /* state for WHILEM */
142 #define WHILEM_B_max_fail (REGNODE_MAX + 18) /* state for WHILEM */
143 #define BRANCH_next (REGNODE_MAX + 19) /* state for BRANCH */
144 #define BRANCH_next_fail (REGNODE_MAX + 20) /* state for BRANCH */
145 #define CURLYM_A (REGNODE_MAX + 21) /* state for CURLYM */
146 #define CURLYM_A_fail (REGNODE_MAX + 22) /* state for CURLYM */
147 #define CURLYM_B (REGNODE_MAX + 23) /* state for CURLYM */
148 #define CURLYM_B_fail (REGNODE_MAX + 24) /* state for CURLYM */
149 #define IFMATCH_A (REGNODE_MAX + 25) /* state for IFMATCH */
150 #define IFMATCH_A_fail (REGNODE_MAX + 26) /* state for IFMATCH */
151 #define CURLY_B_min (REGNODE_MAX + 27) /* state for CURLY */
152 #define CURLY_B_min_fail (REGNODE_MAX + 28) /* state for CURLY */
153 #define CURLY_B_max (REGNODE_MAX + 29) /* state for CURLY */
154 #define CURLY_B_max_fail (REGNODE_MAX + 30) /* state for CURLY */
155 #define COMMIT_next (REGNODE_MAX + 31) /* state for COMMIT */
156 #define COMMIT_next_fail (REGNODE_MAX + 32) /* state for COMMIT */
157 #define MARKPOINT_next (REGNODE_MAX + 33) /* state for MARKPOINT */
158 #define MARKPOINT_next_fail (REGNODE_MAX + 34) /* state for MARKPOINT */
159 #define SKIP_next (REGNODE_MAX + 35) /* state for SKIP */
160 #define SKIP_next_fail (REGNODE_MAX + 36) /* state for SKIP */
161 #define CUTGROUP_next (REGNODE_MAX + 37) /* state for CUTGROUP */
162 #define CUTGROUP_next_fail (REGNODE_MAX + 38) /* state for CUTGROUP */
163 #define KEEPS_next (REGNODE_MAX + 39) /* state for KEEPS */
164 #define KEEPS_next_fail (REGNODE_MAX + 40) /* state for KEEPS */
166 /* PL_regkind[] What type of regop or state is this. */
169 EXTCONST U8 PL_regkind[];
171 EXTCONST U8 PL_regkind[] = {
185 NBOUND, /* NBOUNDL */
186 NBOUND, /* NBOUNDU */
187 NBOUND, /* NBOUNDA */
188 REG_ANY, /* REG_ANY */
193 ANYOF, /* ANYOFPOSIXL */
199 ANYOFR, /* ANYOFRb */
201 ANYOFM, /* NANYOFM */
206 NPOSIXD, /* NPOSIXD */
207 NPOSIXD, /* NPOSIXL */
208 NPOSIXD, /* NPOSIXU */
209 NPOSIXD, /* NPOSIXA */
218 EXACT, /* EXACTFAA */
219 EXACT, /* EXACTFUP */
220 EXACT, /* EXACTFLU8 */
221 EXACT, /* EXACTFAA_NO_TRIE */
222 EXACT, /* EXACT_REQ8 */
223 EXACT, /* LEXACT_REQ8 */
224 EXACT, /* EXACTFU_REQ8 */
225 EXACT, /* EXACTFU_S_EDGE */
226 NOTHING, /* NOTHING */
238 SRCLOSE, /* SRCLOSE */
249 LONGJMP, /* LONGJMP */
250 BRANCHJ, /* BRANCHJ */
251 BRANCHJ, /* IFMATCH */
252 BRANCHJ, /* UNLESSM */
253 BRANCHJ, /* SUSPEND */
254 BRANCHJ, /* IFTHEN */
258 LOGICAL, /* LOGICAL */
262 TRIE, /* AHOCORASICK */
263 TRIE, /* AHOCORASICKC */
265 GROUPPN, /* GROUPPN */
267 DEFINEP, /* DEFINEP */
268 ENDLIKE, /* ENDLIKE */
269 ENDLIKE, /* OPFAIL */
270 ENDLIKE, /* ACCEPT */
273 VERB, /* MARKPOINT */
278 LNBREAK, /* LNBREAK */
279 NOTHING, /* OPTIMIZED */
281 REGEX_SET, /* REGEX_SET */
282 /* ------------ States ------------- */
283 TRIE, /* TRIE_next */
284 TRIE, /* TRIE_next_fail */
286 EVAL, /* EVAL_B_fail */
287 EVAL, /* EVAL_postponed_AB */
288 EVAL, /* EVAL_postponed_AB_fail */
289 CURLYX, /* CURLYX_end */
290 CURLYX, /* CURLYX_end_fail */
291 WHILEM, /* WHILEM_A_pre */
292 WHILEM, /* WHILEM_A_pre_fail */
293 WHILEM, /* WHILEM_A_min */
294 WHILEM, /* WHILEM_A_min_fail */
295 WHILEM, /* WHILEM_A_max */
296 WHILEM, /* WHILEM_A_max_fail */
297 WHILEM, /* WHILEM_B_min */
298 WHILEM, /* WHILEM_B_min_fail */
299 WHILEM, /* WHILEM_B_max */
300 WHILEM, /* WHILEM_B_max_fail */
301 BRANCH, /* BRANCH_next */
302 BRANCH, /* BRANCH_next_fail */
303 CURLYM, /* CURLYM_A */
304 CURLYM, /* CURLYM_A_fail */
305 CURLYM, /* CURLYM_B */
306 CURLYM, /* CURLYM_B_fail */
307 IFMATCH, /* IFMATCH_A */
308 IFMATCH, /* IFMATCH_A_fail */
309 CURLY, /* CURLY_B_min */
310 CURLY, /* CURLY_B_min_fail */
311 CURLY, /* CURLY_B_max */
312 CURLY, /* CURLY_B_max_fail */
313 COMMIT, /* COMMIT_next */
314 COMMIT, /* COMMIT_next_fail */
315 MARKPOINT, /* MARKPOINT_next */
316 MARKPOINT, /* MARKPOINT_next_fail */
317 SKIP, /* SKIP_next */
318 SKIP, /* SKIP_next_fail */
319 CUTGROUP, /* CUTGROUP_next */
320 CUTGROUP, /* CUTGROUP_next_fail */
321 KEEPS, /* KEEPS_next */
322 KEEPS, /* KEEPS_next_fail */
328 /* regarglen[] - How large is the argument part of the node (in regnodes) */
330 static const U8 regarglen[] = {
349 EXTRA_SIZE(struct regnode_charclass), /* ANYOF */
350 EXTRA_SIZE(struct regnode_charclass), /* ANYOFD */
351 EXTRA_SIZE(struct regnode_charclass), /* ANYOFL */
352 EXTRA_SIZE(struct regnode_charclass_posixl), /* ANYOFPOSIXL */
353 EXTRA_SIZE(struct regnode_1), /* ANYOFH */
354 EXTRA_SIZE(struct regnode_1), /* ANYOFHb */
355 EXTRA_SIZE(struct regnode_1), /* ANYOFHr */
356 EXTRA_SIZE(struct regnode_1), /* ANYOFHs */
357 EXTRA_SIZE(struct regnode_1), /* ANYOFR */
358 EXTRA_SIZE(struct regnode_1), /* ANYOFRb */
359 EXTRA_SIZE(struct regnode_1), /* ANYOFM */
360 EXTRA_SIZE(struct regnode_1), /* NANYOFM */
372 EXTRA_SIZE(struct regnode_1), /* LEXACT */
380 0, /* EXACTFAA_NO_TRIE */
382 EXTRA_SIZE(struct regnode_1), /* LEXACT_REQ8 */
383 0, /* EXACTFU_REQ8 */
384 0, /* EXACTFU_S_EDGE */
389 EXTRA_SIZE(struct regnode_2), /* CURLY */
390 EXTRA_SIZE(struct regnode_2), /* CURLYN */
391 EXTRA_SIZE(struct regnode_2), /* CURLYM */
392 EXTRA_SIZE(struct regnode_2), /* CURLYX */
394 EXTRA_SIZE(struct regnode_1), /* OPEN */
395 EXTRA_SIZE(struct regnode_1), /* CLOSE */
398 EXTRA_SIZE(struct regnode_1), /* REF */
399 EXTRA_SIZE(struct regnode_1), /* REFF */
400 EXTRA_SIZE(struct regnode_1), /* REFFL */
401 EXTRA_SIZE(struct regnode_1), /* REFFU */
402 EXTRA_SIZE(struct regnode_1), /* REFFA */
403 EXTRA_SIZE(struct regnode_1), /* REFN */
404 EXTRA_SIZE(struct regnode_1), /* REFFN */
405 EXTRA_SIZE(struct regnode_1), /* REFFLN */
406 EXTRA_SIZE(struct regnode_1), /* REFFUN */
407 EXTRA_SIZE(struct regnode_1), /* REFFAN */
408 EXTRA_SIZE(struct regnode_1), /* LONGJMP */
409 EXTRA_SIZE(struct regnode_1), /* BRANCHJ */
410 EXTRA_SIZE(struct regnode_1), /* IFMATCH */
411 EXTRA_SIZE(struct regnode_1), /* UNLESSM */
412 EXTRA_SIZE(struct regnode_1), /* SUSPEND */
413 EXTRA_SIZE(struct regnode_1), /* IFTHEN */
414 EXTRA_SIZE(struct regnode_1), /* GROUPP */
415 EXTRA_SIZE(struct regnode_2L), /* EVAL */
418 EXTRA_SIZE(struct regnode_1), /* RENUM */
419 EXTRA_SIZE(struct regnode_1), /* TRIE */
420 EXTRA_SIZE(struct regnode_charclass), /* TRIEC */
421 EXTRA_SIZE(struct regnode_1), /* AHOCORASICK */
422 EXTRA_SIZE(struct regnode_charclass), /* AHOCORASICKC */
423 EXTRA_SIZE(struct regnode_2L), /* GOSUB */
424 EXTRA_SIZE(struct regnode_1), /* GROUPPN */
425 EXTRA_SIZE(struct regnode_1), /* INSUBP */
426 EXTRA_SIZE(struct regnode_1), /* DEFINEP */
428 EXTRA_SIZE(struct regnode_1), /* OPFAIL */
429 EXTRA_SIZE(struct regnode_2L), /* ACCEPT */
430 EXTRA_SIZE(struct regnode_1), /* VERB */
431 EXTRA_SIZE(struct regnode_1), /* PRUNE */
432 EXTRA_SIZE(struct regnode_1), /* MARKPOINT */
433 EXTRA_SIZE(struct regnode_1), /* SKIP */
434 EXTRA_SIZE(struct regnode_1), /* COMMIT */
435 EXTRA_SIZE(struct regnode_1), /* CUTGROUP */
440 EXTRA_SIZE(struct regnode_p), /* REGEX_SET */
443 /* reg_off_by_arg[] - Which argument holds the offset to the next node */
445 static const char reg_off_by_arg[] = {
495 0, /* EXACTFAA_NO_TRIE */
498 0, /* EXACTFU_REQ8 */
499 0, /* EXACTFU_S_EDGE */
537 0, /* AHOCORASICKC */
558 #endif /* REG_COMP_C */
561 /* reg_name[] - Opcode/state names in string form, for debugging */
564 EXTCONST char * PL_reg_name[];
566 EXTCONST char * const PL_reg_name[] = {
568 "SUCCEED", /* 0x01 */
580 "NBOUNDL", /* 0x0d */
581 "NBOUNDU", /* 0x0e */
582 "NBOUNDA", /* 0x0f */
583 "REG_ANY", /* 0x10 */
588 "ANYOFPOSIXL", /* 0x15 */
590 "ANYOFHb", /* 0x17 */
591 "ANYOFHr", /* 0x18 */
592 "ANYOFHs", /* 0x19 */
594 "ANYOFRb", /* 0x1b */
596 "NANYOFM", /* 0x1d */
601 "NPOSIXD", /* 0x22 */
602 "NPOSIXL", /* 0x23 */
603 "NPOSIXU", /* 0x24 */
604 "NPOSIXA", /* 0x25 */
611 "EXACTFL", /* 0x2c */
612 "EXACTFU", /* 0x2d */
613 "EXACTFAA", /* 0x2e */
614 "EXACTFUP", /* 0x2f */
615 "EXACTFLU8", /* 0x30 */
616 "EXACTFAA_NO_TRIE", /* 0x31 */
617 "EXACT_REQ8", /* 0x32 */
618 "LEXACT_REQ8", /* 0x33 */
619 "EXACTFU_REQ8", /* 0x34 */
620 "EXACTFU_S_EDGE", /* 0x35 */
621 "NOTHING", /* 0x36 */
633 "SRCLOSE", /* 0x42 */
644 "LONGJMP", /* 0x4d */
645 "BRANCHJ", /* 0x4e */
646 "IFMATCH", /* 0x4f */
647 "UNLESSM", /* 0x50 */
648 "SUSPEND", /* 0x51 */
653 "LOGICAL", /* 0x56 */
657 "AHOCORASICK", /* 0x5a */
658 "AHOCORASICKC", /* 0x5b */
660 "GROUPPN", /* 0x5d */
662 "DEFINEP", /* 0x5f */
663 "ENDLIKE", /* 0x60 */
668 "MARKPOINT", /* 0x65 */
671 "CUTGROUP", /* 0x68 */
673 "LNBREAK", /* 0x6a */
674 "OPTIMIZED", /* 0x6b */
676 "REGEX_SET", /* 0x6d */
677 /* ------------ States ------------- */
678 "TRIE_next", /* REGNODE_MAX +0x01 */
679 "TRIE_next_fail", /* REGNODE_MAX +0x02 */
680 "EVAL_B", /* REGNODE_MAX +0x03 */
681 "EVAL_B_fail", /* REGNODE_MAX +0x04 */
682 "EVAL_postponed_AB", /* REGNODE_MAX +0x05 */
683 "EVAL_postponed_AB_fail", /* REGNODE_MAX +0x06 */
684 "CURLYX_end", /* REGNODE_MAX +0x07 */
685 "CURLYX_end_fail", /* REGNODE_MAX +0x08 */
686 "WHILEM_A_pre", /* REGNODE_MAX +0x09 */
687 "WHILEM_A_pre_fail", /* REGNODE_MAX +0x0a */
688 "WHILEM_A_min", /* REGNODE_MAX +0x0b */
689 "WHILEM_A_min_fail", /* REGNODE_MAX +0x0c */
690 "WHILEM_A_max", /* REGNODE_MAX +0x0d */
691 "WHILEM_A_max_fail", /* REGNODE_MAX +0x0e */
692 "WHILEM_B_min", /* REGNODE_MAX +0x0f */
693 "WHILEM_B_min_fail", /* REGNODE_MAX +0x10 */
694 "WHILEM_B_max", /* REGNODE_MAX +0x11 */
695 "WHILEM_B_max_fail", /* REGNODE_MAX +0x12 */
696 "BRANCH_next", /* REGNODE_MAX +0x13 */
697 "BRANCH_next_fail", /* REGNODE_MAX +0x14 */
698 "CURLYM_A", /* REGNODE_MAX +0x15 */
699 "CURLYM_A_fail", /* REGNODE_MAX +0x16 */
700 "CURLYM_B", /* REGNODE_MAX +0x17 */
701 "CURLYM_B_fail", /* REGNODE_MAX +0x18 */
702 "IFMATCH_A", /* REGNODE_MAX +0x19 */
703 "IFMATCH_A_fail", /* REGNODE_MAX +0x1a */
704 "CURLY_B_min", /* REGNODE_MAX +0x1b */
705 "CURLY_B_min_fail", /* REGNODE_MAX +0x1c */
706 "CURLY_B_max", /* REGNODE_MAX +0x1d */
707 "CURLY_B_max_fail", /* REGNODE_MAX +0x1e */
708 "COMMIT_next", /* REGNODE_MAX +0x1f */
709 "COMMIT_next_fail", /* REGNODE_MAX +0x20 */
710 "MARKPOINT_next", /* REGNODE_MAX +0x21 */
711 "MARKPOINT_next_fail", /* REGNODE_MAX +0x22 */
712 "SKIP_next", /* REGNODE_MAX +0x23 */
713 "SKIP_next_fail", /* REGNODE_MAX +0x24 */
714 "CUTGROUP_next", /* REGNODE_MAX +0x25 */
715 "CUTGROUP_next_fail", /* REGNODE_MAX +0x26 */
716 "KEEPS_next", /* REGNODE_MAX +0x27 */
717 "KEEPS_next_fail", /* REGNODE_MAX +0x28 */
721 /* PL_reg_extflags_name[] - Opcode/state names in string form, for debugging */
724 EXTCONST char * PL_reg_extflags_name[];
726 EXTCONST char * const PL_reg_extflags_name[] = {
727 /* Bits in extflags defined: 11111111111111110000111111111111 */
728 "MULTILINE", /* 0x00000001 */
729 "SINGLELINE", /* 0x00000002 */
730 "FOLD", /* 0x00000004 */
731 "EXTENDED", /* 0x00000008 */
732 "EXTENDED_MORE", /* 0x00000010 */
733 "NOCAPTURE", /* 0x00000020 */
734 "KEEPCOPY", /* 0x00000040 */
735 "CHARSET0", /* 0x00000080 : "CHARSET" - 0x00000380 */
736 "CHARSET1", /* 0x00000100 : "CHARSET" - 0x00000380 */
737 "CHARSET2", /* 0x00000200 : "CHARSET" - 0x00000380 */
738 "STRICT", /* 0x00000400 */
739 "SPLIT", /* 0x00000800 */
740 "UNUSED_BIT_12", /* 0x00001000 */
741 "UNUSED_BIT_13", /* 0x00002000 */
742 "UNUSED_BIT_14", /* 0x00004000 */
743 "UNUSED_BIT_15", /* 0x00008000 */
744 "NO_INPLACE_SUBST", /* 0x00010000 */
745 "EVAL_SEEN", /* 0x00020000 */
746 "UNBOUNDED_QUANTIFIER_SEEN",/* 0x00040000 */
747 "CHECK_ALL", /* 0x00080000 */
748 "MATCH_UTF8", /* 0x00100000 */
749 "USE_INTUIT_NOML", /* 0x00200000 */
750 "USE_INTUIT_ML", /* 0x00400000 */
751 "INTUIT_TAIL", /* 0x00800000 */
752 "IS_ANCHORED", /* 0x01000000 */
753 "COPY_DONE", /* 0x02000000 */
754 "TAINTED_SEEN", /* 0x04000000 */
755 "TAINTED", /* 0x08000000 */
756 "START_ONLY", /* 0x10000000 */
757 "SKIPWHITE", /* 0x20000000 */
758 "WHITE", /* 0x40000000 */
759 "NULL", /* 0x80000000 */
764 # define REG_EXTFLAGS_NAME_SIZE 32
767 /* PL_reg_intflags_name[] - Opcode/state names in string form, for debugging */
770 EXTCONST char * PL_reg_intflags_name[];
772 EXTCONST char * const PL_reg_intflags_name[] = {
773 "SKIP", /* 0x00000001 - PREGf_SKIP */
774 "IMPLICIT", /* 0x00000002 - PREGf_IMPLICIT - Converted .* to ^.* */
775 "NAUGHTY", /* 0x00000004 - PREGf_NAUGHTY - how exponential is this pattern? */
776 "VERBARG_SEEN", /* 0x00000008 - PREGf_VERBARG_SEEN */
777 "CUTGROUP_SEEN", /* 0x00000010 - PREGf_CUTGROUP_SEEN */
778 "USE_RE_EVAL", /* 0x00000020 - PREGf_USE_RE_EVAL - compiled with "use re 'eval'" */
779 "NOSCAN", /* 0x00000040 - PREGf_NOSCAN */
780 "GPOS_SEEN", /* 0x00000100 - PREGf_GPOS_SEEN */
781 "GPOS_FLOAT", /* 0x00000200 - PREGf_GPOS_FLOAT */
782 "ANCH_MBOL", /* 0x00000400 - PREGf_ANCH_MBOL */
783 "ANCH_SBOL", /* 0x00000800 - PREGf_ANCH_SBOL */
784 "ANCH_GPOS", /* 0x00001000 - PREGf_ANCH_GPOS */
785 "RECURSE_SEEN", /* 0x00002000 - PREGf_RECURSE_SEEN */
790 # define REG_INTFLAGS_NAME_SIZE 13
793 /* The following have no fixed length. U8 so we can do strchr() on it. */
794 #define REGNODE_VARIES(node) (PL_varies_bitmask[(node) >> 3] & (1 << ((node) & 7)))
797 EXTCONST U8 PL_varies[] __attribute__deprecated__;
799 EXTCONST U8 PL_varies[] __attribute__deprecated__ = {
800 CLUMP, BRANCH, STAR, PLUS, CURLY, CURLYN, CURLYM, CURLYX, WHILEM, REF,
801 REFF, REFFL, REFFU, REFFA, REFN, REFFN, REFFLN, REFFUN, REFFAN,
802 BRANCHJ, SUSPEND, IFTHEN,
808 EXTCONST U8 PL_varies_bitmask[];
810 EXTCONST U8 PL_varies_bitmask[] = {
811 0x00, 0x00, 0x00, 0x00, 0xC0, 0x00, 0x00, 0x7F, 0xF8, 0x5F, 0x06, 0x00, 0x00, 0x00
815 /* The following always have a length of 1. U8 we can do strchr() on it. */
816 /* (Note that length 1 means "one character" under UTF8, not "one octet".) */
817 #define REGNODE_SIMPLE(node) (PL_simple_bitmask[(node) >> 3] & (1 << ((node) & 7)))
820 EXTCONST U8 PL_simple[] __attribute__deprecated__;
822 EXTCONST U8 PL_simple[] __attribute__deprecated__ = {
823 REG_ANY, SANY, ANYOF, ANYOFD, ANYOFL, ANYOFPOSIXL, ANYOFH, ANYOFHb,
824 ANYOFHr, ANYOFHs, ANYOFR, ANYOFRb, ANYOFM, NANYOFM, POSIXD, POSIXL,
825 POSIXU, POSIXA, NPOSIXD, NPOSIXL, NPOSIXU, NPOSIXA, REGEX_SET,
831 EXTCONST U8 PL_simple_bitmask[];
833 EXTCONST U8 PL_simple_bitmask[] = {
834 0x00, 0x00, 0xFF, 0xFF, 0x3F, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x20