This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
regcomp.c: White-space only
[perl5.git] / regnodes.h
CommitLineData
37442d52
RGS
1/* -*- buffer-read-only: t -*-
2 !!!!!!! DO NOT EDIT THIS FILE !!!!!!!
eb2624c9 3 This file is built by regen/regcomp.pl from regcomp.sym.
d09b2d29 4 Any changes made here will be lost!
78102347 5 */
d09b2d29 6
6bda09f9
YO
7/* Regops and State definitions */
8
39d24220
KW
9#define REGNODE_MAX 96
10#define REGMATCH_STATE_MAX 138
03363afd 11
f9f4320a
YO
12#define END 0 /* 0000 End of program. */
13#define SUCCEED 1 /* 0x01 Return from a subroutine, basically. */
d3d47aac
YO
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. */
c440a570
KW
22#define BOUND 8 /* 0x08 Like BOUNDA for non-utf8, otherwise match "" between any Unicode \w\W or \W\w */
23#define BOUNDL 9 /* 0x09 Like BOUND/BOUNDU, but \w and \W are defined by current locale */
64935bc6 24#define BOUNDU 10 /* 0x0a Match "" at any boundary of a given type using Unicode rules */
c440a570
KW
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 match "" between any Unicode \w\w or \W\W */
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 Unicode rules */
29#define NBOUNDA 15 /* 0x0f Match "" betweeen any \w\w or \W\W, where \w is [_a-zA-Z0-9] */
d3d47aac
YO
30#define REG_ANY 16 /* 0x10 Match any one character (except newline). */
31#define SANY 17 /* 0x11 Match any one character. */
33c28ab2 32#define ANYOF 18 /* 0x12 Match character in (or not in) this class, single char match only */
ac44c12e
KW
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 POSIXD 21 /* 0x15 Some [[:class:]] under /d; the FLAGS field gives which one */
36#define POSIXL 22 /* 0x16 Some [[:class:]] under /l; the FLAGS field gives which one */
37#define POSIXU 23 /* 0x17 Some [[:class:]] under /u; the FLAGS field gives which one */
38#define POSIXA 24 /* 0x18 Some [[:class:]] under /a; the FLAGS field gives which one */
39#define NPOSIXD 25 /* 0x19 complement of POSIXD, [[:^class:]] */
40#define NPOSIXL 26 /* 0x1a complement of POSIXL, [[:^class:]] */
41#define NPOSIXU 27 /* 0x1b complement of POSIXU, [[:^class:]] */
42#define NPOSIXA 28 /* 0x1c complement of POSIXA, [[:^class:]] */
39d24220
KW
43#define ASCII 29 /* 0x1d [[:ascii:]] */
44#define NASCII 30 /* 0x1e [[:^ascii:]] */
45#define CLUMP 31 /* 0x1f Match any extended grapheme cluster sequence */
46#define BRANCH 32 /* 0x20 Match this alternative, or the next... */
47#define EXACT 33 /* 0x21 Match this string (preceded by length). */
48#define EXACTL 34 /* 0x22 Like EXACT, but /l is in effect (used so locale-related warnings can be checked for). */
49#define EXACTF 35 /* 0x23 Match this non-UTF-8 string (not guaranteed to be folded) using /id rules (w/len). */
50#define EXACTFL 36 /* 0x24 Match this string (not guaranteed to be folded) using /il rules (w/len). */
51#define EXACTFU 37 /* 0x25 Match this string (folded iff in UTF-8, length in folding doesn't change if not in UTF-8) using /iu rules (w/len). */
52#define EXACTFA 38 /* 0x26 Match this string (not guaranteed to be folded) using /iaa rules (w/len). */
53#define EXACTFU_SS 39 /* 0x27 Match this string (folded iff in UTF-8, length in folding may change even if not in UTF-8) using /iu rules (w/len). */
54#define EXACTFLU8 40 /* 0x28 Rare cirucmstances: like EXACTFU, but is under /l, UTF-8, folded, and everything in it is above 255. */
55#define EXACTFA_NO_TRIE 41 /* 0x29 Match this string (which is not trie-able; not guaranteed to be folded) using /iaa rules (w/len). */
56#define NOTHING 42 /* 0x2a Match empty string. */
57#define TAIL 43 /* 0x2b Match empty string. Can jump here from outside. */
58#define STAR 44 /* 0x2c Match this (simple) thing 0 or more times. */
59#define PLUS 45 /* 0x2d Match this (simple) thing 1 or more times. */
60#define CURLY 46 /* 0x2e Match this simple thing {n,m} times. */
61#define CURLYN 47 /* 0x2f Capture next-after-this simple thing */
62#define CURLYM 48 /* 0x30 Capture this medium-complex thing {n,m} times. */
63#define CURLYX 49 /* 0x31 Match this complex thing {n,m} times. */
64#define WHILEM 50 /* 0x32 Do curly processing and see if rest matches. */
65#define OPEN 51 /* 0x33 Mark this point in input as start of #n. */
66#define CLOSE 52 /* 0x34 Close corresponding OPEN of #n. */
67#define SROPEN 53 /* 0x35 Same as OPEN, but for script run */
68#define SRCLOSE 54 /* 0x36 Close preceding SROPEN */
69#define REF 55 /* 0x37 Match some already matched string */
70#define REFF 56 /* 0x38 Match already matched string, folded using native charset rules for non-utf8 */
71#define REFFL 57 /* 0x39 Match already matched string, folded in loc. */
72#define REFFU 58 /* 0x3a Match already matched string, folded using unicode rules for non-utf8 */
73#define REFFA 59 /* 0x3b Match already matched string, folded using unicode rules for non-utf8, no mixing ASCII, non-ASCII */
74#define NREF 60 /* 0x3c Match some already matched string */
75#define NREFF 61 /* 0x3d Match already matched string, folded using native charset rules for non-utf8 */
76#define NREFFL 62 /* 0x3e Match already matched string, folded in loc. */
77#define NREFFU 63 /* 0x3f Match already matched string, folded using unicode rules for non-utf8 */
78#define NREFFA 64 /* 0x40 Match already matched string, folded using unicode rules for non-utf8, no mixing ASCII, non-ASCII */
79#define LONGJMP 65 /* 0x41 Jump far away. */
80#define BRANCHJ 66 /* 0x42 BRANCH with long offset. */
81#define IFMATCH 67 /* 0x43 Succeeds if the following matches. */
82#define UNLESSM 68 /* 0x44 Fails if the following matches. */
83#define SUSPEND 69 /* 0x45 "Independent" sub-RE. */
84#define IFTHEN 70 /* 0x46 Switch, should be preceded by switcher. */
85#define GROUPP 71 /* 0x47 Whether the group matched. */
86#define EVAL 72 /* 0x48 Execute some Perl code. */
87#define MINMOD 73 /* 0x49 Next operator is not greedy. */
88#define LOGICAL 74 /* 0x4a Next opcode should set the flag only. */
89#define RENUM 75 /* 0x4b Group with independently numbered parens. */
90#define TRIE 76 /* 0x4c Match many EXACT(F[ALU]?)? at once. flags==type */
91#define TRIEC 77 /* 0x4d Same as TRIE, but with embedded charclass data */
92#define AHOCORASICK 78 /* 0x4e Aho Corasick stclass. flags==type */
93#define AHOCORASICKC 79 /* 0x4f Same as AHOCORASICK, but with embedded charclass data */
94#define GOSUB 80 /* 0x50 recurse to paren arg1 at (signed) ofs arg2 */
95#define NGROUPP 81 /* 0x51 Whether the group matched. */
96#define INSUBP 82 /* 0x52 Whether we are in a specific recurse. */
97#define DEFINEP 83 /* 0x53 Never execute directly. */
98#define ENDLIKE 84 /* 0x54 Used only for the type field of verbs */
99#define OPFAIL 85 /* 0x55 Same as (?!), but with verb arg */
100#define ACCEPT 86 /* 0x56 Accepts the current matched string, with verbar */
101#define VERB 87 /* 0x57 Used only for the type field of verbs */
102#define PRUNE 88 /* 0x58 Pattern fails at this startpoint if no-backtracking through this */
103#define MARKPOINT 89 /* 0x59 Push the current location for rollback by cut. */
104#define SKIP 90 /* 0x5a On failure skip forward (to the mark) before retrying */
105#define COMMIT 91 /* 0x5b Pattern fails outright if backtracking through this */
106#define CUTGROUP 92 /* 0x5c On failure go to the next alternation in the group */
107#define KEEPS 93 /* 0x5d $& begins here. */
108#define LNBREAK 94 /* 0x5e generic newline pattern */
109#define OPTIMIZED 95 /* 0x5f Placeholder for dump. */
110#define PSEUDO 96 /* 0x60 Pseudo opcode for internal use. */
03363afd 111 /* ------------ States ------------- */
24b23f37
YO
112#define TRIE_next (REGNODE_MAX + 1) /* state for TRIE */
113#define TRIE_next_fail (REGNODE_MAX + 2) /* state for TRIE */
4ee16520
DM
114#define EVAL_B (REGNODE_MAX + 3) /* state for EVAL */
115#define EVAL_B_fail (REGNODE_MAX + 4) /* state for EVAL */
116#define EVAL_postponed_AB (REGNODE_MAX + 5) /* state for EVAL */
117#define EVAL_postponed_AB_fail (REGNODE_MAX + 6) /* state for EVAL */
118#define CURLYX_end (REGNODE_MAX + 7) /* state for CURLYX */
119#define CURLYX_end_fail (REGNODE_MAX + 8) /* state for CURLYX */
120#define WHILEM_A_pre (REGNODE_MAX + 9) /* state for WHILEM */
121#define WHILEM_A_pre_fail (REGNODE_MAX + 10) /* state for WHILEM */
122#define WHILEM_A_min (REGNODE_MAX + 11) /* state for WHILEM */
123#define WHILEM_A_min_fail (REGNODE_MAX + 12) /* state for WHILEM */
124#define WHILEM_A_max (REGNODE_MAX + 13) /* state for WHILEM */
125#define WHILEM_A_max_fail (REGNODE_MAX + 14) /* state for WHILEM */
126#define WHILEM_B_min (REGNODE_MAX + 15) /* state for WHILEM */
127#define WHILEM_B_min_fail (REGNODE_MAX + 16) /* state for WHILEM */
128#define WHILEM_B_max (REGNODE_MAX + 17) /* state for WHILEM */
129#define WHILEM_B_max_fail (REGNODE_MAX + 18) /* state for WHILEM */
130#define BRANCH_next (REGNODE_MAX + 19) /* state for BRANCH */
131#define BRANCH_next_fail (REGNODE_MAX + 20) /* state for BRANCH */
132#define CURLYM_A (REGNODE_MAX + 21) /* state for CURLYM */
133#define CURLYM_A_fail (REGNODE_MAX + 22) /* state for CURLYM */
134#define CURLYM_B (REGNODE_MAX + 23) /* state for CURLYM */
135#define CURLYM_B_fail (REGNODE_MAX + 24) /* state for CURLYM */
136#define IFMATCH_A (REGNODE_MAX + 25) /* state for IFMATCH */
137#define IFMATCH_A_fail (REGNODE_MAX + 26) /* state for IFMATCH */
138#define CURLY_B_min_known (REGNODE_MAX + 27) /* state for CURLY */
139#define CURLY_B_min_known_fail (REGNODE_MAX + 28) /* state for CURLY */
140#define CURLY_B_min (REGNODE_MAX + 29) /* state for CURLY */
141#define CURLY_B_min_fail (REGNODE_MAX + 30) /* state for CURLY */
142#define CURLY_B_max (REGNODE_MAX + 31) /* state for CURLY */
143#define CURLY_B_max_fail (REGNODE_MAX + 32) /* state for CURLY */
144#define COMMIT_next (REGNODE_MAX + 33) /* state for COMMIT */
145#define COMMIT_next_fail (REGNODE_MAX + 34) /* state for COMMIT */
146#define MARKPOINT_next (REGNODE_MAX + 35) /* state for MARKPOINT */
147#define MARKPOINT_next_fail (REGNODE_MAX + 36) /* state for MARKPOINT */
148#define SKIP_next (REGNODE_MAX + 37) /* state for SKIP */
149#define SKIP_next_fail (REGNODE_MAX + 38) /* state for SKIP */
150#define CUTGROUP_next (REGNODE_MAX + 39) /* state for CUTGROUP */
151#define CUTGROUP_next_fail (REGNODE_MAX + 40) /* state for CUTGROUP */
152#define KEEPS_next (REGNODE_MAX + 41) /* state for KEEPS */
153#define KEEPS_next_fail (REGNODE_MAX + 42) /* state for KEEPS */
03363afd 154
6bda09f9 155/* PL_regkind[] What type of regop or state is this. */
d09b2d29
IZ
156
157#ifndef DOINIT
22c35a8c 158EXTCONST U8 PL_regkind[];
d09b2d29 159#else
22c35a8c 160EXTCONST U8 PL_regkind[] = {
e2e6a0f1
YO
161 END, /* END */
162 END, /* SUCCEED */
e2e6a0f1 163 BOL, /* SBOL */
d3d47aac 164 BOL, /* MBOL */
e2e6a0f1 165 EOL, /* SEOL */
d3d47aac
YO
166 EOL, /* MEOL */
167 EOL, /* EOS */
168 GPOS, /* GPOS */
e2e6a0f1
YO
169 BOUND, /* BOUND */
170 BOUND, /* BOUNDL */
1e355c70 171 BOUND, /* BOUNDU */
0c6e81eb 172 BOUND, /* BOUNDA */
e2e6a0f1
YO
173 NBOUND, /* NBOUND */
174 NBOUND, /* NBOUNDL */
1e355c70 175 NBOUND, /* NBOUNDU */
0c6e81eb 176 NBOUND, /* NBOUNDA */
e2e6a0f1
YO
177 REG_ANY, /* REG_ANY */
178 REG_ANY, /* SANY */
e2e6a0f1 179 ANYOF, /* ANYOF */
ac44c12e 180 ANYOF, /* ANYOFD */
a4525e78 181 ANYOF, /* ANYOFL */
3615ea58
KW
182 POSIXD, /* POSIXD */
183 POSIXD, /* POSIXL */
184 POSIXD, /* POSIXU */
185 POSIXD, /* POSIXA */
9e84774b
KW
186 NPOSIXD, /* NPOSIXD */
187 NPOSIXD, /* NPOSIXL */
188 NPOSIXD, /* NPOSIXU */
189 NPOSIXD, /* NPOSIXA */
39d24220
KW
190 ASCII, /* ASCII */
191 ASCII, /* NASCII */
e2e6a0f1
YO
192 CLUMP, /* CLUMP */
193 BRANCH, /* BRANCH */
e2e6a0f1 194 EXACT, /* EXACT */
a4525e78 195 EXACT, /* EXACTL */
e2e6a0f1
YO
196 EXACT, /* EXACTF */
197 EXACT, /* EXACTFL */
01f98ec2 198 EXACT, /* EXACTFU */
8c1182fd 199 EXACT, /* EXACTFA */
3c760661 200 EXACT, /* EXACTFU_SS */
a4525e78 201 EXACT, /* EXACTFLU8 */
098b07d5 202 EXACT, /* EXACTFA_NO_TRIE */
e2e6a0f1
YO
203 NOTHING, /* NOTHING */
204 NOTHING, /* TAIL */
205 STAR, /* STAR */
206 PLUS, /* PLUS */
207 CURLY, /* CURLY */
208 CURLY, /* CURLYN */
209 CURLY, /* CURLYM */
210 CURLY, /* CURLYX */
211 WHILEM, /* WHILEM */
212 OPEN, /* OPEN */
213 CLOSE, /* CLOSE */
07093db4
KW
214 SROPEN, /* SROPEN */
215 SRCLOSE, /* SRCLOSE */
e2e6a0f1
YO
216 REF, /* REF */
217 REF, /* REFF */
218 REF, /* REFFL */
01f98ec2 219 REF, /* REFFU */
781aab5c 220 REF, /* REFFA */
01f98ec2
KW
221 REF, /* NREF */
222 REF, /* NREFF */
223 REF, /* NREFFL */
224 REF, /* NREFFU */
781aab5c 225 REF, /* NREFFA */
d3d47aac
YO
226 LONGJMP, /* LONGJMP */
227 BRANCHJ, /* BRANCHJ */
e2e6a0f1
YO
228 BRANCHJ, /* IFMATCH */
229 BRANCHJ, /* UNLESSM */
230 BRANCHJ, /* SUSPEND */
231 BRANCHJ, /* IFTHEN */
232 GROUPP, /* GROUPP */
e2e6a0f1
YO
233 EVAL, /* EVAL */
234 MINMOD, /* MINMOD */
235 LOGICAL, /* LOGICAL */
236 BRANCHJ, /* RENUM */
237 TRIE, /* TRIE */
238 TRIE, /* TRIEC */
239 TRIE, /* AHOCORASICK */
240 TRIE, /* AHOCORASICKC */
241 GOSUB, /* GOSUB */
e2e6a0f1
YO
242 NGROUPP, /* NGROUPP */
243 INSUBP, /* INSUBP */
244 DEFINEP, /* DEFINEP */
245 ENDLIKE, /* ENDLIKE */
246 ENDLIKE, /* OPFAIL */
247 ENDLIKE, /* ACCEPT */
248 VERB, /* VERB */
5d458dd8 249 VERB, /* PRUNE */
e2e6a0f1 250 VERB, /* MARKPOINT */
5d458dd8 251 VERB, /* SKIP */
e2e6a0f1 252 VERB, /* COMMIT */
5d458dd8 253 VERB, /* CUTGROUP */
ee9b8eae 254 KEEPS, /* KEEPS */
e1d1eefb 255 LNBREAK, /* LNBREAK */
e2e6a0f1
YO
256 NOTHING, /* OPTIMIZED */
257 PSEUDO, /* PSEUDO */
03363afd 258 /* ------------ States ------------- */
e2e6a0f1
YO
259 TRIE, /* TRIE_next */
260 TRIE, /* TRIE_next_fail */
4ee16520
DM
261 EVAL, /* EVAL_B */
262 EVAL, /* EVAL_B_fail */
263 EVAL, /* EVAL_postponed_AB */
264 EVAL, /* EVAL_postponed_AB_fail */
e2e6a0f1
YO
265 CURLYX, /* CURLYX_end */
266 CURLYX, /* CURLYX_end_fail */
267 WHILEM, /* WHILEM_A_pre */
268 WHILEM, /* WHILEM_A_pre_fail */
269 WHILEM, /* WHILEM_A_min */
270 WHILEM, /* WHILEM_A_min_fail */
271 WHILEM, /* WHILEM_A_max */
272 WHILEM, /* WHILEM_A_max_fail */
273 WHILEM, /* WHILEM_B_min */
274 WHILEM, /* WHILEM_B_min_fail */
275 WHILEM, /* WHILEM_B_max */
276 WHILEM, /* WHILEM_B_max_fail */
277 BRANCH, /* BRANCH_next */
278 BRANCH, /* BRANCH_next_fail */
279 CURLYM, /* CURLYM_A */
280 CURLYM, /* CURLYM_A_fail */
281 CURLYM, /* CURLYM_B */
282 CURLYM, /* CURLYM_B_fail */
283 IFMATCH, /* IFMATCH_A */
284 IFMATCH, /* IFMATCH_A_fail */
285 CURLY, /* CURLY_B_min_known */
286 CURLY, /* CURLY_B_min_known_fail */
287 CURLY, /* CURLY_B_min */
288 CURLY, /* CURLY_B_min_fail */
289 CURLY, /* CURLY_B_max */
290 CURLY, /* CURLY_B_max_fail */
291 COMMIT, /* COMMIT_next */
292 COMMIT, /* COMMIT_next_fail */
293 MARKPOINT, /* MARKPOINT_next */
294 MARKPOINT, /* MARKPOINT_next_fail */
5d458dd8
YO
295 SKIP, /* SKIP_next */
296 SKIP, /* SKIP_next_fail */
297 CUTGROUP, /* CUTGROUP_next */
298 CUTGROUP, /* CUTGROUP_next_fail */
ee9b8eae
YO
299 KEEPS, /* KEEPS_next */
300 KEEPS, /* KEEPS_next_fail */
d09b2d29
IZ
301};
302#endif
303
f83e001e
YO
304#ifdef REG_COMP_C
305
6bda09f9 306/* regarglen[] - How large is the argument part of the node (in regnodes) */
d09b2d29 307
29de9391 308static const U8 regarglen[] = {
03363afd
YO
309 0, /* END */
310 0, /* SUCCEED */
03363afd 311 0, /* SBOL */
d3d47aac 312 0, /* MBOL */
03363afd 313 0, /* SEOL */
d3d47aac
YO
314 0, /* MEOL */
315 0, /* EOS */
316 0, /* GPOS */
03363afd
YO
317 0, /* BOUND */
318 0, /* BOUNDL */
1e355c70 319 0, /* BOUNDU */
0c6e81eb 320 0, /* BOUNDA */
03363afd
YO
321 0, /* NBOUND */
322 0, /* NBOUNDL */
1e355c70 323 0, /* NBOUNDU */
0c6e81eb 324 0, /* NBOUNDA */
03363afd
YO
325 0, /* REG_ANY */
326 0, /* SANY */
975a06f7 327 EXTRA_SIZE(struct regnode_1), /* ANYOF */
ac44c12e 328 EXTRA_SIZE(struct regnode_1), /* ANYOFD */
a4525e78 329 EXTRA_SIZE(struct regnode_1), /* ANYOFL */
3615ea58
KW
330 0, /* POSIXD */
331 0, /* POSIXL */
332 0, /* POSIXU */
333 0, /* POSIXA */
334 0, /* NPOSIXD */
335 0, /* NPOSIXL */
336 0, /* NPOSIXU */
337 0, /* NPOSIXA */
39d24220
KW
338 0, /* ASCII */
339 0, /* NASCII */
03363afd
YO
340 0, /* CLUMP */
341 0, /* BRANCH */
03363afd 342 0, /* EXACT */
a4525e78 343 0, /* EXACTL */
03363afd
YO
344 0, /* EXACTF */
345 0, /* EXACTFL */
01f98ec2 346 0, /* EXACTFU */
8c1182fd 347 0, /* EXACTFA */
3c760661 348 0, /* EXACTFU_SS */
a4525e78 349 0, /* EXACTFLU8 */
098b07d5 350 0, /* EXACTFA_NO_TRIE */
03363afd
YO
351 0, /* NOTHING */
352 0, /* TAIL */
353 0, /* STAR */
354 0, /* PLUS */
355 EXTRA_SIZE(struct regnode_2), /* CURLY */
356 EXTRA_SIZE(struct regnode_2), /* CURLYN */
357 EXTRA_SIZE(struct regnode_2), /* CURLYM */
358 EXTRA_SIZE(struct regnode_2), /* CURLYX */
359 0, /* WHILEM */
360 EXTRA_SIZE(struct regnode_1), /* OPEN */
361 EXTRA_SIZE(struct regnode_1), /* CLOSE */
07093db4
KW
362 0, /* SROPEN */
363 0, /* SRCLOSE */
03363afd
YO
364 EXTRA_SIZE(struct regnode_1), /* REF */
365 EXTRA_SIZE(struct regnode_1), /* REFF */
366 EXTRA_SIZE(struct regnode_1), /* REFFL */
01f98ec2 367 EXTRA_SIZE(struct regnode_1), /* REFFU */
781aab5c 368 EXTRA_SIZE(struct regnode_1), /* REFFA */
01f98ec2
KW
369 EXTRA_SIZE(struct regnode_1), /* NREF */
370 EXTRA_SIZE(struct regnode_1), /* NREFF */
371 EXTRA_SIZE(struct regnode_1), /* NREFFL */
372 EXTRA_SIZE(struct regnode_1), /* NREFFU */
781aab5c 373 EXTRA_SIZE(struct regnode_1), /* NREFFA */
d3d47aac
YO
374 EXTRA_SIZE(struct regnode_1), /* LONGJMP */
375 EXTRA_SIZE(struct regnode_1), /* BRANCHJ */
03363afd
YO
376 EXTRA_SIZE(struct regnode_1), /* IFMATCH */
377 EXTRA_SIZE(struct regnode_1), /* UNLESSM */
378 EXTRA_SIZE(struct regnode_1), /* SUSPEND */
379 EXTRA_SIZE(struct regnode_1), /* IFTHEN */
380 EXTRA_SIZE(struct regnode_1), /* GROUPP */
13f27704 381 EXTRA_SIZE(struct regnode_2L), /* EVAL */
03363afd
YO
382 0, /* MINMOD */
383 0, /* LOGICAL */
384 EXTRA_SIZE(struct regnode_1), /* RENUM */
385 EXTRA_SIZE(struct regnode_1), /* TRIE */
386 EXTRA_SIZE(struct regnode_charclass), /* TRIEC */
387 EXTRA_SIZE(struct regnode_1), /* AHOCORASICK */
388 EXTRA_SIZE(struct regnode_charclass), /* AHOCORASICKC */
1a147d38 389 EXTRA_SIZE(struct regnode_2L), /* GOSUB */
0a4db386 390 EXTRA_SIZE(struct regnode_1), /* NGROUPP */
1a147d38 391 EXTRA_SIZE(struct regnode_1), /* INSUBP */
0a4db386 392 EXTRA_SIZE(struct regnode_1), /* DEFINEP */
e2e6a0f1 393 0, /* ENDLIKE */
fee50582
YO
394 EXTRA_SIZE(struct regnode_1), /* OPFAIL */
395 EXTRA_SIZE(struct regnode_2L), /* ACCEPT */
20832bc5 396 EXTRA_SIZE(struct regnode_1), /* VERB */
5d458dd8 397 EXTRA_SIZE(struct regnode_1), /* PRUNE */
e2e6a0f1 398 EXTRA_SIZE(struct regnode_1), /* MARKPOINT */
5d458dd8 399 EXTRA_SIZE(struct regnode_1), /* SKIP */
e2e6a0f1 400 EXTRA_SIZE(struct regnode_1), /* COMMIT */
5d458dd8 401 EXTRA_SIZE(struct regnode_1), /* CUTGROUP */
ee9b8eae 402 0, /* KEEPS */
e1d1eefb 403 0, /* LNBREAK */
03363afd
YO
404 0, /* OPTIMIZED */
405 0, /* PSEUDO */
d09b2d29
IZ
406};
407
6bda09f9
YO
408/* reg_off_by_arg[] - Which argument holds the offset to the next node */
409
29de9391 410static const char reg_off_by_arg[] = {
03363afd
YO
411 0, /* END */
412 0, /* SUCCEED */
03363afd 413 0, /* SBOL */
d3d47aac 414 0, /* MBOL */
03363afd 415 0, /* SEOL */
d3d47aac
YO
416 0, /* MEOL */
417 0, /* EOS */
418 0, /* GPOS */
03363afd
YO
419 0, /* BOUND */
420 0, /* BOUNDL */
1e355c70 421 0, /* BOUNDU */
0c6e81eb 422 0, /* BOUNDA */
03363afd
YO
423 0, /* NBOUND */
424 0, /* NBOUNDL */
1e355c70 425 0, /* NBOUNDU */
0c6e81eb 426 0, /* NBOUNDA */
03363afd
YO
427 0, /* REG_ANY */
428 0, /* SANY */
03363afd 429 0, /* ANYOF */
ac44c12e 430 0, /* ANYOFD */
a4525e78 431 0, /* ANYOFL */
3615ea58
KW
432 0, /* POSIXD */
433 0, /* POSIXL */
434 0, /* POSIXU */
435 0, /* POSIXA */
436 0, /* NPOSIXD */
437 0, /* NPOSIXL */
438 0, /* NPOSIXU */
439 0, /* NPOSIXA */
39d24220
KW
440 0, /* ASCII */
441 0, /* NASCII */
03363afd
YO
442 0, /* CLUMP */
443 0, /* BRANCH */
03363afd 444 0, /* EXACT */
a4525e78 445 0, /* EXACTL */
03363afd
YO
446 0, /* EXACTF */
447 0, /* EXACTFL */
01f98ec2 448 0, /* EXACTFU */
8c1182fd 449 0, /* EXACTFA */
3c760661 450 0, /* EXACTFU_SS */
a4525e78 451 0, /* EXACTFLU8 */
098b07d5 452 0, /* EXACTFA_NO_TRIE */
03363afd
YO
453 0, /* NOTHING */
454 0, /* TAIL */
455 0, /* STAR */
456 0, /* PLUS */
457 0, /* CURLY */
458 0, /* CURLYN */
459 0, /* CURLYM */
460 0, /* CURLYX */
461 0, /* WHILEM */
462 0, /* OPEN */
463 0, /* CLOSE */
07093db4
KW
464 0, /* SROPEN */
465 0, /* SRCLOSE */
03363afd
YO
466 0, /* REF */
467 0, /* REFF */
468 0, /* REFFL */
01f98ec2 469 0, /* REFFU */
781aab5c 470 0, /* REFFA */
01f98ec2
KW
471 0, /* NREF */
472 0, /* NREFF */
473 0, /* NREFFL */
474 0, /* NREFFU */
781aab5c 475 0, /* NREFFA */
d3d47aac
YO
476 1, /* LONGJMP */
477 1, /* BRANCHJ */
03363afd
YO
478 2, /* IFMATCH */
479 2, /* UNLESSM */
480 1, /* SUSPEND */
481 1, /* IFTHEN */
482 0, /* GROUPP */
03363afd
YO
483 0, /* EVAL */
484 0, /* MINMOD */
485 0, /* LOGICAL */
486 1, /* RENUM */
487 0, /* TRIE */
488 0, /* TRIEC */
489 0, /* AHOCORASICK */
490 0, /* AHOCORASICKC */
1a147d38 491 0, /* GOSUB */
0a4db386 492 0, /* NGROUPP */
1a147d38 493 0, /* INSUBP */
0a4db386 494 0, /* DEFINEP */
e2e6a0f1 495 0, /* ENDLIKE */
7f69552c 496 0, /* OPFAIL */
e2e6a0f1
YO
497 0, /* ACCEPT */
498 0, /* VERB */
5d458dd8 499 0, /* PRUNE */
e2e6a0f1 500 0, /* MARKPOINT */
5d458dd8 501 0, /* SKIP */
e2e6a0f1 502 0, /* COMMIT */
5d458dd8 503 0, /* CUTGROUP */
ee9b8eae 504 0, /* KEEPS */
e1d1eefb 505 0, /* LNBREAK */
03363afd
YO
506 0, /* OPTIMIZED */
507 0, /* PSEUDO */
d09b2d29 508};
885f9e59 509
13d6edb4
NC
510#endif /* REG_COMP_C */
511
f83e001e 512
6bda09f9
YO
513/* reg_name[] - Opcode/state names in string form, for debugging */
514
22429478 515#ifndef DOINIT
13d6edb4 516EXTCONST char * PL_reg_name[];
22429478 517#else
4764e399 518EXTCONST char * const PL_reg_name[] = {
03363afd
YO
519 "END", /* 0000 */
520 "SUCCEED", /* 0x01 */
d3d47aac 521 "SBOL", /* 0x02 */
03363afd 522 "MBOL", /* 0x03 */
d3d47aac
YO
523 "SEOL", /* 0x04 */
524 "MEOL", /* 0x05 */
525 "EOS", /* 0x06 */
526 "GPOS", /* 0x07 */
527 "BOUND", /* 0x08 */
528 "BOUNDL", /* 0x09 */
529 "BOUNDU", /* 0x0a */
530 "BOUNDA", /* 0x0b */
531 "NBOUND", /* 0x0c */
532 "NBOUNDL", /* 0x0d */
533 "NBOUNDU", /* 0x0e */
534 "NBOUNDA", /* 0x0f */
535 "REG_ANY", /* 0x10 */
536 "SANY", /* 0x11 */
33c28ab2 537 "ANYOF", /* 0x12 */
ac44c12e
KW
538 "ANYOFD", /* 0x13 */
539 "ANYOFL", /* 0x14 */
540 "POSIXD", /* 0x15 */
541 "POSIXL", /* 0x16 */
542 "POSIXU", /* 0x17 */
543 "POSIXA", /* 0x18 */
544 "NPOSIXD", /* 0x19 */
545 "NPOSIXL", /* 0x1a */
546 "NPOSIXU", /* 0x1b */
547 "NPOSIXA", /* 0x1c */
39d24220
KW
548 "ASCII", /* 0x1d */
549 "NASCII", /* 0x1e */
550 "CLUMP", /* 0x1f */
551 "BRANCH", /* 0x20 */
552 "EXACT", /* 0x21 */
553 "EXACTL", /* 0x22 */
554 "EXACTF", /* 0x23 */
555 "EXACTFL", /* 0x24 */
556 "EXACTFU", /* 0x25 */
557 "EXACTFA", /* 0x26 */
558 "EXACTFU_SS", /* 0x27 */
559 "EXACTFLU8", /* 0x28 */
560 "EXACTFA_NO_TRIE", /* 0x29 */
561 "NOTHING", /* 0x2a */
562 "TAIL", /* 0x2b */
563 "STAR", /* 0x2c */
564 "PLUS", /* 0x2d */
565 "CURLY", /* 0x2e */
566 "CURLYN", /* 0x2f */
567 "CURLYM", /* 0x30 */
568 "CURLYX", /* 0x31 */
569 "WHILEM", /* 0x32 */
570 "OPEN", /* 0x33 */
571 "CLOSE", /* 0x34 */
572 "SROPEN", /* 0x35 */
573 "SRCLOSE", /* 0x36 */
574 "REF", /* 0x37 */
575 "REFF", /* 0x38 */
576 "REFFL", /* 0x39 */
577 "REFFU", /* 0x3a */
578 "REFFA", /* 0x3b */
579 "NREF", /* 0x3c */
580 "NREFF", /* 0x3d */
581 "NREFFL", /* 0x3e */
582 "NREFFU", /* 0x3f */
583 "NREFFA", /* 0x40 */
584 "LONGJMP", /* 0x41 */
585 "BRANCHJ", /* 0x42 */
586 "IFMATCH", /* 0x43 */
587 "UNLESSM", /* 0x44 */
588 "SUSPEND", /* 0x45 */
589 "IFTHEN", /* 0x46 */
590 "GROUPP", /* 0x47 */
591 "EVAL", /* 0x48 */
592 "MINMOD", /* 0x49 */
593 "LOGICAL", /* 0x4a */
594 "RENUM", /* 0x4b */
595 "TRIE", /* 0x4c */
596 "TRIEC", /* 0x4d */
597 "AHOCORASICK", /* 0x4e */
598 "AHOCORASICKC", /* 0x4f */
599 "GOSUB", /* 0x50 */
600 "NGROUPP", /* 0x51 */
601 "INSUBP", /* 0x52 */
602 "DEFINEP", /* 0x53 */
603 "ENDLIKE", /* 0x54 */
604 "OPFAIL", /* 0x55 */
605 "ACCEPT", /* 0x56 */
606 "VERB", /* 0x57 */
607 "PRUNE", /* 0x58 */
608 "MARKPOINT", /* 0x59 */
609 "SKIP", /* 0x5a */
610 "COMMIT", /* 0x5b */
611 "CUTGROUP", /* 0x5c */
612 "KEEPS", /* 0x5d */
613 "LNBREAK", /* 0x5e */
614 "OPTIMIZED", /* 0x5f */
615 "PSEUDO", /* 0x60 */
03363afd 616 /* ------------ States ------------- */
24b23f37
YO
617 "TRIE_next", /* REGNODE_MAX +0x01 */
618 "TRIE_next_fail", /* REGNODE_MAX +0x02 */
4ee16520
DM
619 "EVAL_B", /* REGNODE_MAX +0x03 */
620 "EVAL_B_fail", /* REGNODE_MAX +0x04 */
621 "EVAL_postponed_AB", /* REGNODE_MAX +0x05 */
622 "EVAL_postponed_AB_fail", /* REGNODE_MAX +0x06 */
623 "CURLYX_end", /* REGNODE_MAX +0x07 */
624 "CURLYX_end_fail", /* REGNODE_MAX +0x08 */
625 "WHILEM_A_pre", /* REGNODE_MAX +0x09 */
626 "WHILEM_A_pre_fail", /* REGNODE_MAX +0x0a */
627 "WHILEM_A_min", /* REGNODE_MAX +0x0b */
628 "WHILEM_A_min_fail", /* REGNODE_MAX +0x0c */
629 "WHILEM_A_max", /* REGNODE_MAX +0x0d */
630 "WHILEM_A_max_fail", /* REGNODE_MAX +0x0e */
631 "WHILEM_B_min", /* REGNODE_MAX +0x0f */
632 "WHILEM_B_min_fail", /* REGNODE_MAX +0x10 */
633 "WHILEM_B_max", /* REGNODE_MAX +0x11 */
634 "WHILEM_B_max_fail", /* REGNODE_MAX +0x12 */
635 "BRANCH_next", /* REGNODE_MAX +0x13 */
636 "BRANCH_next_fail", /* REGNODE_MAX +0x14 */
637 "CURLYM_A", /* REGNODE_MAX +0x15 */
638 "CURLYM_A_fail", /* REGNODE_MAX +0x16 */
639 "CURLYM_B", /* REGNODE_MAX +0x17 */
640 "CURLYM_B_fail", /* REGNODE_MAX +0x18 */
641 "IFMATCH_A", /* REGNODE_MAX +0x19 */
642 "IFMATCH_A_fail", /* REGNODE_MAX +0x1a */
643 "CURLY_B_min_known", /* REGNODE_MAX +0x1b */
644 "CURLY_B_min_known_fail", /* REGNODE_MAX +0x1c */
645 "CURLY_B_min", /* REGNODE_MAX +0x1d */
646 "CURLY_B_min_fail", /* REGNODE_MAX +0x1e */
647 "CURLY_B_max", /* REGNODE_MAX +0x1f */
648 "CURLY_B_max_fail", /* REGNODE_MAX +0x20 */
649 "COMMIT_next", /* REGNODE_MAX +0x21 */
650 "COMMIT_next_fail", /* REGNODE_MAX +0x22 */
651 "MARKPOINT_next", /* REGNODE_MAX +0x23 */
652 "MARKPOINT_next_fail", /* REGNODE_MAX +0x24 */
653 "SKIP_next", /* REGNODE_MAX +0x25 */
654 "SKIP_next_fail", /* REGNODE_MAX +0x26 */
655 "CUTGROUP_next", /* REGNODE_MAX +0x27 */
656 "CUTGROUP_next_fail", /* REGNODE_MAX +0x28 */
657 "KEEPS_next", /* REGNODE_MAX +0x29 */
658 "KEEPS_next_fail", /* REGNODE_MAX +0x2a */
885f9e59 659};
22429478 660#endif /* DOINIT */
d09b2d29 661
f7819f85
A
662/* PL_reg_extflags_name[] - Opcode/state names in string form, for debugging */
663
664#ifndef DOINIT
665EXTCONST char * PL_reg_extflags_name[];
666#else
667EXTCONST char * const PL_reg_extflags_name[] = {
d262c0c7 668 /* Bits in extflags defined: 11111111111111110000111111111111 */
52d81aa8
NC
669 "MULTILINE", /* 0x00000001 */
670 "SINGLELINE", /* 0x00000002 */
671 "FOLD", /* 0x00000004 */
672 "EXTENDED", /* 0x00000008 */
334afb3e 673 "EXTENDED_MORE", /* 0x00000010 */
e3b64d84
KW
674 "NOCAPTURE", /* 0x00000020 */
675 "KEEPCOPY", /* 0x00000040 */
676 "CHARSET0", /* 0x00000080 : "CHARSET" - 0x00000380 */
677 "CHARSET1", /* 0x00000100 : "CHARSET" - 0x00000380 */
678 "CHARSET2", /* 0x00000200 : "CHARSET" - 0x00000380 */
d262c0c7
KW
679 "STRICT", /* 0x00000400 */
680 "SPLIT", /* 0x00000800 */
1d32d911
KW
681 "UNUSED_BIT_12", /* 0x00001000 */
682 "UNUSED_BIT_13", /* 0x00002000 */
683 "UNUSED_BIT_14", /* 0x00004000 */
a3b51d37
KW
684 "UNUSED_BIT_15", /* 0x00008000 */
685 "NO_INPLACE_SUBST", /* 0x00010000 */
686 "EVAL_SEEN", /* 0x00020000 */
ee273784 687 "UNBOUNDED_QUANTIFIER_SEEN",/* 0x00040000 */
e795e964 688 "CHECK_ALL", /* 0x00080000 */
52d81aa8
NC
689 "MATCH_UTF8", /* 0x00100000 */
690 "USE_INTUIT_NOML", /* 0x00200000 */
691 "USE_INTUIT_ML", /* 0x00400000 */
692 "INTUIT_TAIL", /* 0x00800000 */
a3b51d37 693 "IS_ANCHORED", /* 0x01000000 */
52d81aa8
NC
694 "COPY_DONE", /* 0x02000000 */
695 "TAINTED_SEEN", /* 0x04000000 */
696 "TAINTED", /* 0x08000000 */
697 "START_ONLY", /* 0x10000000 */
dbc200c5 698 "SKIPWHITE", /* 0x20000000 */
52d81aa8
NC
699 "WHITE", /* 0x40000000 */
700 "NULL", /* 0x80000000 */
f7819f85
A
701};
702#endif /* DOINIT */
703
adc2d0c9
JH
704#ifdef DEBUGGING
705# define REG_EXTFLAGS_NAME_SIZE 32
706#endif
707
337ff307
YO
708/* PL_reg_intflags_name[] - Opcode/state names in string form, for debugging */
709
710#ifndef DOINIT
711EXTCONST char * PL_reg_intflags_name[];
712#else
713EXTCONST char * const PL_reg_intflags_name[] = {
b8f6efdd
YO
714 "SKIP", /* 0x00000001 - PREGf_SKIP */
715 "IMPLICIT", /* 0x00000002 - PREGf_IMPLICIT - Converted .* to ^.* */
716 "NAUGHTY", /* 0x00000004 - PREGf_NAUGHTY - how exponential is this pattern? */
717 "VERBARG_SEEN", /* 0x00000008 - PREGf_VERBARG_SEEN */
718 "CUTGROUP_SEEN", /* 0x00000010 - PREGf_CUTGROUP_SEEN */
719 "USE_RE_EVAL", /* 0x00000020 - PREGf_USE_RE_EVAL - compiled with "use re 'eval'" */
58430ea8 720 "NOSCAN", /* 0x00000040 - PREGf_NOSCAN */
58430ea8
YO
721 "GPOS_SEEN", /* 0x00000100 - PREGf_GPOS_SEEN */
722 "GPOS_FLOAT", /* 0x00000200 - PREGf_GPOS_FLOAT */
d3d47aac
YO
723 "ANCH_MBOL", /* 0x00000400 - PREGf_ANCH_MBOL */
724 "ANCH_SBOL", /* 0x00000800 - PREGf_ANCH_SBOL */
725 "ANCH_GPOS", /* 0x00001000 - PREGf_ANCH_GPOS */
d5a00e4a 726 "RECURSE_SEEN", /* 0x00002000 - PREGf_RECURSE_SEEN */
337ff307
YO
727};
728#endif /* DOINIT */
729
adc2d0c9 730#ifdef DEBUGGING
d5a00e4a 731# define REG_INTFLAGS_NAME_SIZE 13
adc2d0c9
JH
732#endif
733
f9ef50a7 734/* The following have no fixed length. U8 so we can do strchr() on it. */
ded4dd2a 735#define REGNODE_VARIES(node) (PL_varies_bitmask[(node) >> 3] & (1 << ((node) & 7)))
e52fc539 736
f9ef50a7 737#ifndef DOINIT
ded4dd2a 738EXTCONST U8 PL_varies[] __attribute__deprecated__;
f9ef50a7 739#else
ded4dd2a 740EXTCONST U8 PL_varies[] __attribute__deprecated__ = {
62e6ef33
AC
741 CLUMP, BRANCH, STAR, PLUS, CURLY, CURLYN, CURLYM, CURLYX, WHILEM, REF,
742 REFF, REFFL, REFFU, REFFA, NREF, NREFF, NREFFL, NREFFU, NREFFA,
d3d47aac 743 BRANCHJ, SUSPEND, IFTHEN,
f9ef50a7
NC
744 0
745};
746#endif /* DOINIT */
747
ded4dd2a
NC
748#ifndef DOINIT
749EXTCONST U8 PL_varies_bitmask[];
750#else
751EXTCONST U8 PL_varies_bitmask[] = {
39d24220 752 0x00, 0x00, 0x00, 0x80, 0x01, 0xF0, 0x87, 0xFF, 0x65, 0x00, 0x00, 0x00, 0x00
ded4dd2a
NC
753};
754#endif /* DOINIT */
755
f9ef50a7
NC
756/* The following always have a length of 1. U8 we can do strchr() on it. */
757/* (Note that length 1 means "one character" under UTF8, not "one octet".) */
ded4dd2a 758#define REGNODE_SIMPLE(node) (PL_simple_bitmask[(node) >> 3] & (1 << ((node) & 7)))
e52fc539 759
f9ef50a7 760#ifndef DOINIT
ded4dd2a 761EXTCONST U8 PL_simple[] __attribute__deprecated__;
f9ef50a7 762#else
ded4dd2a 763EXTCONST U8 PL_simple[] __attribute__deprecated__ = {
ac44c12e 764 REG_ANY, SANY, ANYOF, ANYOFD, ANYOFL, POSIXD, POSIXL, POSIXU, POSIXA,
39d24220 765 NPOSIXD, NPOSIXL, NPOSIXU, NPOSIXA, ASCII, NASCII,
f9ef50a7
NC
766 0
767};
768#endif /* DOINIT */
769
ded4dd2a
NC
770#ifndef DOINIT
771EXTCONST U8 PL_simple_bitmask[];
772#else
773EXTCONST U8 PL_simple_bitmask[] = {
39d24220 774 0x00, 0x00, 0xFF, 0x7F, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
ded4dd2a
NC
775};
776#endif /* DOINIT */
777
37442d52 778/* ex: set ro: */