This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
make gimme consistently U8
[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
ac44c12e
KW
9#define REGNODE_MAX 93
10#define REGMATCH_STATE_MAX 133
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:]] */
43#define CLUMP 29 /* 0x1d Match any extended grapheme cluster sequence */
44#define BRANCH 30 /* 0x1e Match this alternative, or the next... */
45#define EXACT 31 /* 0x1f Match this string (preceded by length). */
46#define EXACTL 32 /* 0x20 Like EXACT, but /l is in effect (used so locale-related warnings can be checked for). */
47#define EXACTF 33 /* 0x21 Match this non-UTF-8 string (not guaranteed to be folded) using /id rules (w/len). */
48#define EXACTFL 34 /* 0x22 Match this string (not guaranteed to be folded) using /il rules (w/len). */
49#define EXACTFU 35 /* 0x23 Match this string (folded iff in UTF-8, length in folding doesn't change if not in UTF-8) using /iu rules (w/len). */
50#define EXACTFA 36 /* 0x24 Match this string (not guaranteed to be folded) using /iaa rules (w/len). */
51#define EXACTFU_SS 37 /* 0x25 Match this string (folded iff in UTF-8, length in folding may change even if not in UTF-8) using /iu rules (w/len). */
52#define EXACTFLU8 38 /* 0x26 Rare cirucmstances: like EXACTFU, but is under /l, UTF-8, folded, and everything in it is above 255. */
53#define EXACTFA_NO_TRIE 39 /* 0x27 Match this string (which is not trie-able; not guaranteed to be folded) using /iaa rules (w/len). */
54#define NOTHING 40 /* 0x28 Match empty string. */
55#define TAIL 41 /* 0x29 Match empty string. Can jump here from outside. */
56#define STAR 42 /* 0x2a Match this (simple) thing 0 or more times. */
57#define PLUS 43 /* 0x2b Match this (simple) thing 1 or more times. */
58#define CURLY 44 /* 0x2c Match this simple thing {n,m} times. */
59#define CURLYN 45 /* 0x2d Capture next-after-this simple thing */
60#define CURLYM 46 /* 0x2e Capture this medium-complex thing {n,m} times. */
61#define CURLYX 47 /* 0x2f Match this complex thing {n,m} times. */
62#define WHILEM 48 /* 0x30 Do curly processing and see if rest matches. */
63#define OPEN 49 /* 0x31 Mark this point in input as start of #n. */
64#define CLOSE 50 /* 0x32 Analogous to OPEN. */
65#define REF 51 /* 0x33 Match some already matched string */
66#define REFF 52 /* 0x34 Match already matched string, folded using native charset rules for non-utf8 */
67#define REFFL 53 /* 0x35 Match already matched string, folded in loc. */
68#define REFFU 54 /* 0x36 Match already matched string, folded using unicode rules for non-utf8 */
69#define REFFA 55 /* 0x37 Match already matched string, folded using unicode rules for non-utf8, no mixing ASCII, non-ASCII */
70#define NREF 56 /* 0x38 Match some already matched string */
71#define NREFF 57 /* 0x39 Match already matched string, folded using native charset rules for non-utf8 */
72#define NREFFL 58 /* 0x3a Match already matched string, folded in loc. */
73#define NREFFU 59 /* 0x3b Match already matched string, folded using unicode rules for non-utf8 */
74#define NREFFA 60 /* 0x3c Match already matched string, folded using unicode rules for non-utf8, no mixing ASCII, non-ASCII */
75#define LONGJMP 61 /* 0x3d Jump far away. */
76#define BRANCHJ 62 /* 0x3e BRANCH with long offset. */
77#define IFMATCH 63 /* 0x3f Succeeds if the following matches. */
78#define UNLESSM 64 /* 0x40 Fails if the following matches. */
79#define SUSPEND 65 /* 0x41 "Independent" sub-RE. */
80#define IFTHEN 66 /* 0x42 Switch, should be preceded by switcher. */
81#define GROUPP 67 /* 0x43 Whether the group matched. */
82#define EVAL 68 /* 0x44 Execute some Perl code. */
83#define MINMOD 69 /* 0x45 Next operator is not greedy. */
84#define LOGICAL 70 /* 0x46 Next opcode should set the flag only. */
85#define RENUM 71 /* 0x47 Group with independently numbered parens. */
86#define TRIE 72 /* 0x48 Match many EXACT(F[ALU]?)? at once. flags==type */
87#define TRIEC 73 /* 0x49 Same as TRIE, but with embedded charclass data */
88#define AHOCORASICK 74 /* 0x4a Aho Corasick stclass. flags==type */
89#define AHOCORASICKC 75 /* 0x4b Same as AHOCORASICK, but with embedded charclass data */
90#define GOSUB 76 /* 0x4c recurse to paren arg1 at (signed) ofs arg2 */
91#define GOSTART 77 /* 0x4d recurse to start of pattern */
92#define NGROUPP 78 /* 0x4e Whether the group matched. */
93#define INSUBP 79 /* 0x4f Whether we are in a specific recurse. */
94#define DEFINEP 80 /* 0x50 Never execute directly. */
95#define ENDLIKE 81 /* 0x51 Used only for the type field of verbs */
fee50582
YO
96#define OPFAIL 82 /* 0x52 Same as (?!), but with verb arg */
97#define ACCEPT 83 /* 0x53 Accepts the current matched string, with verbar */
ac44c12e
KW
98#define VERB 84 /* 0x54 Used only for the type field of verbs */
99#define PRUNE 85 /* 0x55 Pattern fails at this startpoint if no-backtracking through this */
100#define MARKPOINT 86 /* 0x56 Push the current location for rollback by cut. */
101#define SKIP 87 /* 0x57 On failure skip forward (to the mark) before retrying */
102#define COMMIT 88 /* 0x58 Pattern fails outright if backtracking through this */
103#define CUTGROUP 89 /* 0x59 On failure go to the next alternation in the group */
104#define KEEPS 90 /* 0x5a $& begins here. */
105#define LNBREAK 91 /* 0x5b generic newline pattern */
106#define OPTIMIZED 92 /* 0x5c Placeholder for dump. */
107#define PSEUDO 93 /* 0x5d Pseudo opcode for internal use. */
03363afd 108 /* ------------ States ------------- */
24b23f37
YO
109#define TRIE_next (REGNODE_MAX + 1) /* state for TRIE */
110#define TRIE_next_fail (REGNODE_MAX + 2) /* state for TRIE */
111#define EVAL_AB (REGNODE_MAX + 3) /* state for EVAL */
112#define EVAL_AB_fail (REGNODE_MAX + 4) /* state for EVAL */
113#define CURLYX_end (REGNODE_MAX + 5) /* state for CURLYX */
114#define CURLYX_end_fail (REGNODE_MAX + 6) /* state for CURLYX */
115#define WHILEM_A_pre (REGNODE_MAX + 7) /* state for WHILEM */
116#define WHILEM_A_pre_fail (REGNODE_MAX + 8) /* state for WHILEM */
117#define WHILEM_A_min (REGNODE_MAX + 9) /* state for WHILEM */
118#define WHILEM_A_min_fail (REGNODE_MAX + 10) /* state for WHILEM */
119#define WHILEM_A_max (REGNODE_MAX + 11) /* state for WHILEM */
120#define WHILEM_A_max_fail (REGNODE_MAX + 12) /* state for WHILEM */
121#define WHILEM_B_min (REGNODE_MAX + 13) /* state for WHILEM */
122#define WHILEM_B_min_fail (REGNODE_MAX + 14) /* state for WHILEM */
123#define WHILEM_B_max (REGNODE_MAX + 15) /* state for WHILEM */
124#define WHILEM_B_max_fail (REGNODE_MAX + 16) /* state for WHILEM */
125#define BRANCH_next (REGNODE_MAX + 17) /* state for BRANCH */
126#define BRANCH_next_fail (REGNODE_MAX + 18) /* state for BRANCH */
127#define CURLYM_A (REGNODE_MAX + 19) /* state for CURLYM */
128#define CURLYM_A_fail (REGNODE_MAX + 20) /* state for CURLYM */
129#define CURLYM_B (REGNODE_MAX + 21) /* state for CURLYM */
130#define CURLYM_B_fail (REGNODE_MAX + 22) /* state for CURLYM */
131#define IFMATCH_A (REGNODE_MAX + 23) /* state for IFMATCH */
132#define IFMATCH_A_fail (REGNODE_MAX + 24) /* state for IFMATCH */
133#define CURLY_B_min_known (REGNODE_MAX + 25) /* state for CURLY */
134#define CURLY_B_min_known_fail (REGNODE_MAX + 26) /* state for CURLY */
135#define CURLY_B_min (REGNODE_MAX + 27) /* state for CURLY */
136#define CURLY_B_min_fail (REGNODE_MAX + 28) /* state for CURLY */
137#define CURLY_B_max (REGNODE_MAX + 29) /* state for CURLY */
138#define CURLY_B_max_fail (REGNODE_MAX + 30) /* state for CURLY */
139#define COMMIT_next (REGNODE_MAX + 31) /* state for COMMIT */
140#define COMMIT_next_fail (REGNODE_MAX + 32) /* state for COMMIT */
e2e6a0f1
YO
141#define MARKPOINT_next (REGNODE_MAX + 33) /* state for MARKPOINT */
142#define MARKPOINT_next_fail (REGNODE_MAX + 34) /* state for MARKPOINT */
5d458dd8
YO
143#define SKIP_next (REGNODE_MAX + 35) /* state for SKIP */
144#define SKIP_next_fail (REGNODE_MAX + 36) /* state for SKIP */
145#define CUTGROUP_next (REGNODE_MAX + 37) /* state for CUTGROUP */
146#define CUTGROUP_next_fail (REGNODE_MAX + 38) /* state for CUTGROUP */
ee9b8eae
YO
147#define KEEPS_next (REGNODE_MAX + 39) /* state for KEEPS */
148#define KEEPS_next_fail (REGNODE_MAX + 40) /* state for KEEPS */
03363afd 149
6bda09f9 150/* PL_regkind[] What type of regop or state is this. */
d09b2d29
IZ
151
152#ifndef DOINIT
22c35a8c 153EXTCONST U8 PL_regkind[];
d09b2d29 154#else
22c35a8c 155EXTCONST U8 PL_regkind[] = {
e2e6a0f1
YO
156 END, /* END */
157 END, /* SUCCEED */
e2e6a0f1 158 BOL, /* SBOL */
d3d47aac 159 BOL, /* MBOL */
e2e6a0f1 160 EOL, /* SEOL */
d3d47aac
YO
161 EOL, /* MEOL */
162 EOL, /* EOS */
163 GPOS, /* GPOS */
e2e6a0f1
YO
164 BOUND, /* BOUND */
165 BOUND, /* BOUNDL */
1e355c70 166 BOUND, /* BOUNDU */
0c6e81eb 167 BOUND, /* BOUNDA */
e2e6a0f1
YO
168 NBOUND, /* NBOUND */
169 NBOUND, /* NBOUNDL */
1e355c70 170 NBOUND, /* NBOUNDU */
0c6e81eb 171 NBOUND, /* NBOUNDA */
e2e6a0f1
YO
172 REG_ANY, /* REG_ANY */
173 REG_ANY, /* SANY */
e2e6a0f1 174 ANYOF, /* ANYOF */
ac44c12e 175 ANYOF, /* ANYOFD */
a4525e78 176 ANYOF, /* ANYOFL */
3615ea58
KW
177 POSIXD, /* POSIXD */
178 POSIXD, /* POSIXL */
179 POSIXD, /* POSIXU */
180 POSIXD, /* POSIXA */
9e84774b
KW
181 NPOSIXD, /* NPOSIXD */
182 NPOSIXD, /* NPOSIXL */
183 NPOSIXD, /* NPOSIXU */
184 NPOSIXD, /* NPOSIXA */
e2e6a0f1
YO
185 CLUMP, /* CLUMP */
186 BRANCH, /* BRANCH */
e2e6a0f1 187 EXACT, /* EXACT */
a4525e78 188 EXACT, /* EXACTL */
e2e6a0f1
YO
189 EXACT, /* EXACTF */
190 EXACT, /* EXACTFL */
01f98ec2 191 EXACT, /* EXACTFU */
8c1182fd 192 EXACT, /* EXACTFA */
3c760661 193 EXACT, /* EXACTFU_SS */
a4525e78 194 EXACT, /* EXACTFLU8 */
098b07d5 195 EXACT, /* EXACTFA_NO_TRIE */
e2e6a0f1
YO
196 NOTHING, /* NOTHING */
197 NOTHING, /* TAIL */
198 STAR, /* STAR */
199 PLUS, /* PLUS */
200 CURLY, /* CURLY */
201 CURLY, /* CURLYN */
202 CURLY, /* CURLYM */
203 CURLY, /* CURLYX */
204 WHILEM, /* WHILEM */
205 OPEN, /* OPEN */
206 CLOSE, /* CLOSE */
207 REF, /* REF */
208 REF, /* REFF */
209 REF, /* REFFL */
01f98ec2 210 REF, /* REFFU */
781aab5c 211 REF, /* REFFA */
01f98ec2
KW
212 REF, /* NREF */
213 REF, /* NREFF */
214 REF, /* NREFFL */
215 REF, /* NREFFU */
781aab5c 216 REF, /* NREFFA */
d3d47aac
YO
217 LONGJMP, /* LONGJMP */
218 BRANCHJ, /* BRANCHJ */
e2e6a0f1
YO
219 BRANCHJ, /* IFMATCH */
220 BRANCHJ, /* UNLESSM */
221 BRANCHJ, /* SUSPEND */
222 BRANCHJ, /* IFTHEN */
223 GROUPP, /* GROUPP */
e2e6a0f1
YO
224 EVAL, /* EVAL */
225 MINMOD, /* MINMOD */
226 LOGICAL, /* LOGICAL */
227 BRANCHJ, /* RENUM */
228 TRIE, /* TRIE */
229 TRIE, /* TRIEC */
230 TRIE, /* AHOCORASICK */
231 TRIE, /* AHOCORASICKC */
232 GOSUB, /* GOSUB */
233 GOSTART, /* GOSTART */
e2e6a0f1
YO
234 NGROUPP, /* NGROUPP */
235 INSUBP, /* INSUBP */
236 DEFINEP, /* DEFINEP */
237 ENDLIKE, /* ENDLIKE */
238 ENDLIKE, /* OPFAIL */
239 ENDLIKE, /* ACCEPT */
240 VERB, /* VERB */
5d458dd8 241 VERB, /* PRUNE */
e2e6a0f1 242 VERB, /* MARKPOINT */
5d458dd8 243 VERB, /* SKIP */
e2e6a0f1 244 VERB, /* COMMIT */
5d458dd8 245 VERB, /* CUTGROUP */
ee9b8eae 246 KEEPS, /* KEEPS */
e1d1eefb 247 LNBREAK, /* LNBREAK */
e2e6a0f1
YO
248 NOTHING, /* OPTIMIZED */
249 PSEUDO, /* PSEUDO */
03363afd 250 /* ------------ States ------------- */
e2e6a0f1
YO
251 TRIE, /* TRIE_next */
252 TRIE, /* TRIE_next_fail */
253 EVAL, /* EVAL_AB */
254 EVAL, /* EVAL_AB_fail */
255 CURLYX, /* CURLYX_end */
256 CURLYX, /* CURLYX_end_fail */
257 WHILEM, /* WHILEM_A_pre */
258 WHILEM, /* WHILEM_A_pre_fail */
259 WHILEM, /* WHILEM_A_min */
260 WHILEM, /* WHILEM_A_min_fail */
261 WHILEM, /* WHILEM_A_max */
262 WHILEM, /* WHILEM_A_max_fail */
263 WHILEM, /* WHILEM_B_min */
264 WHILEM, /* WHILEM_B_min_fail */
265 WHILEM, /* WHILEM_B_max */
266 WHILEM, /* WHILEM_B_max_fail */
267 BRANCH, /* BRANCH_next */
268 BRANCH, /* BRANCH_next_fail */
269 CURLYM, /* CURLYM_A */
270 CURLYM, /* CURLYM_A_fail */
271 CURLYM, /* CURLYM_B */
272 CURLYM, /* CURLYM_B_fail */
273 IFMATCH, /* IFMATCH_A */
274 IFMATCH, /* IFMATCH_A_fail */
275 CURLY, /* CURLY_B_min_known */
276 CURLY, /* CURLY_B_min_known_fail */
277 CURLY, /* CURLY_B_min */
278 CURLY, /* CURLY_B_min_fail */
279 CURLY, /* CURLY_B_max */
280 CURLY, /* CURLY_B_max_fail */
281 COMMIT, /* COMMIT_next */
282 COMMIT, /* COMMIT_next_fail */
283 MARKPOINT, /* MARKPOINT_next */
284 MARKPOINT, /* MARKPOINT_next_fail */
5d458dd8
YO
285 SKIP, /* SKIP_next */
286 SKIP, /* SKIP_next_fail */
287 CUTGROUP, /* CUTGROUP_next */
288 CUTGROUP, /* CUTGROUP_next_fail */
ee9b8eae
YO
289 KEEPS, /* KEEPS_next */
290 KEEPS, /* KEEPS_next_fail */
d09b2d29
IZ
291};
292#endif
293
f83e001e
YO
294#ifdef REG_COMP_C
295
6bda09f9 296/* regarglen[] - How large is the argument part of the node (in regnodes) */
d09b2d29 297
29de9391 298static const U8 regarglen[] = {
03363afd
YO
299 0, /* END */
300 0, /* SUCCEED */
03363afd 301 0, /* SBOL */
d3d47aac 302 0, /* MBOL */
03363afd 303 0, /* SEOL */
d3d47aac
YO
304 0, /* MEOL */
305 0, /* EOS */
306 0, /* GPOS */
03363afd
YO
307 0, /* BOUND */
308 0, /* BOUNDL */
1e355c70 309 0, /* BOUNDU */
0c6e81eb 310 0, /* BOUNDA */
03363afd
YO
311 0, /* NBOUND */
312 0, /* NBOUNDL */
1e355c70 313 0, /* NBOUNDU */
0c6e81eb 314 0, /* NBOUNDA */
03363afd
YO
315 0, /* REG_ANY */
316 0, /* SANY */
975a06f7 317 EXTRA_SIZE(struct regnode_1), /* ANYOF */
ac44c12e 318 EXTRA_SIZE(struct regnode_1), /* ANYOFD */
a4525e78 319 EXTRA_SIZE(struct regnode_1), /* ANYOFL */
3615ea58
KW
320 0, /* POSIXD */
321 0, /* POSIXL */
322 0, /* POSIXU */
323 0, /* POSIXA */
324 0, /* NPOSIXD */
325 0, /* NPOSIXL */
326 0, /* NPOSIXU */
327 0, /* NPOSIXA */
03363afd
YO
328 0, /* CLUMP */
329 0, /* BRANCH */
03363afd 330 0, /* EXACT */
a4525e78 331 0, /* EXACTL */
03363afd
YO
332 0, /* EXACTF */
333 0, /* EXACTFL */
01f98ec2 334 0, /* EXACTFU */
8c1182fd 335 0, /* EXACTFA */
3c760661 336 0, /* EXACTFU_SS */
a4525e78 337 0, /* EXACTFLU8 */
098b07d5 338 0, /* EXACTFA_NO_TRIE */
03363afd
YO
339 0, /* NOTHING */
340 0, /* TAIL */
341 0, /* STAR */
342 0, /* PLUS */
343 EXTRA_SIZE(struct regnode_2), /* CURLY */
344 EXTRA_SIZE(struct regnode_2), /* CURLYN */
345 EXTRA_SIZE(struct regnode_2), /* CURLYM */
346 EXTRA_SIZE(struct regnode_2), /* CURLYX */
347 0, /* WHILEM */
348 EXTRA_SIZE(struct regnode_1), /* OPEN */
349 EXTRA_SIZE(struct regnode_1), /* CLOSE */
350 EXTRA_SIZE(struct regnode_1), /* REF */
351 EXTRA_SIZE(struct regnode_1), /* REFF */
352 EXTRA_SIZE(struct regnode_1), /* REFFL */
01f98ec2 353 EXTRA_SIZE(struct regnode_1), /* REFFU */
781aab5c 354 EXTRA_SIZE(struct regnode_1), /* REFFA */
01f98ec2
KW
355 EXTRA_SIZE(struct regnode_1), /* NREF */
356 EXTRA_SIZE(struct regnode_1), /* NREFF */
357 EXTRA_SIZE(struct regnode_1), /* NREFFL */
358 EXTRA_SIZE(struct regnode_1), /* NREFFU */
781aab5c 359 EXTRA_SIZE(struct regnode_1), /* NREFFA */
d3d47aac
YO
360 EXTRA_SIZE(struct regnode_1), /* LONGJMP */
361 EXTRA_SIZE(struct regnode_1), /* BRANCHJ */
03363afd
YO
362 EXTRA_SIZE(struct regnode_1), /* IFMATCH */
363 EXTRA_SIZE(struct regnode_1), /* UNLESSM */
364 EXTRA_SIZE(struct regnode_1), /* SUSPEND */
365 EXTRA_SIZE(struct regnode_1), /* IFTHEN */
366 EXTRA_SIZE(struct regnode_1), /* GROUPP */
13f27704 367 EXTRA_SIZE(struct regnode_2L), /* EVAL */
03363afd
YO
368 0, /* MINMOD */
369 0, /* LOGICAL */
370 EXTRA_SIZE(struct regnode_1), /* RENUM */
371 EXTRA_SIZE(struct regnode_1), /* TRIE */
372 EXTRA_SIZE(struct regnode_charclass), /* TRIEC */
373 EXTRA_SIZE(struct regnode_1), /* AHOCORASICK */
374 EXTRA_SIZE(struct regnode_charclass), /* AHOCORASICKC */
1a147d38
YO
375 EXTRA_SIZE(struct regnode_2L), /* GOSUB */
376 0, /* GOSTART */
0a4db386 377 EXTRA_SIZE(struct regnode_1), /* NGROUPP */
1a147d38 378 EXTRA_SIZE(struct regnode_1), /* INSUBP */
0a4db386 379 EXTRA_SIZE(struct regnode_1), /* DEFINEP */
e2e6a0f1 380 0, /* ENDLIKE */
fee50582
YO
381 EXTRA_SIZE(struct regnode_1), /* OPFAIL */
382 EXTRA_SIZE(struct regnode_2L), /* ACCEPT */
20832bc5 383 EXTRA_SIZE(struct regnode_1), /* VERB */
5d458dd8 384 EXTRA_SIZE(struct regnode_1), /* PRUNE */
e2e6a0f1 385 EXTRA_SIZE(struct regnode_1), /* MARKPOINT */
5d458dd8 386 EXTRA_SIZE(struct regnode_1), /* SKIP */
e2e6a0f1 387 EXTRA_SIZE(struct regnode_1), /* COMMIT */
5d458dd8 388 EXTRA_SIZE(struct regnode_1), /* CUTGROUP */
ee9b8eae 389 0, /* KEEPS */
e1d1eefb 390 0, /* LNBREAK */
03363afd
YO
391 0, /* OPTIMIZED */
392 0, /* PSEUDO */
d09b2d29
IZ
393};
394
6bda09f9
YO
395/* reg_off_by_arg[] - Which argument holds the offset to the next node */
396
29de9391 397static const char reg_off_by_arg[] = {
03363afd
YO
398 0, /* END */
399 0, /* SUCCEED */
03363afd 400 0, /* SBOL */
d3d47aac 401 0, /* MBOL */
03363afd 402 0, /* SEOL */
d3d47aac
YO
403 0, /* MEOL */
404 0, /* EOS */
405 0, /* GPOS */
03363afd
YO
406 0, /* BOUND */
407 0, /* BOUNDL */
1e355c70 408 0, /* BOUNDU */
0c6e81eb 409 0, /* BOUNDA */
03363afd
YO
410 0, /* NBOUND */
411 0, /* NBOUNDL */
1e355c70 412 0, /* NBOUNDU */
0c6e81eb 413 0, /* NBOUNDA */
03363afd
YO
414 0, /* REG_ANY */
415 0, /* SANY */
03363afd 416 0, /* ANYOF */
ac44c12e 417 0, /* ANYOFD */
a4525e78 418 0, /* ANYOFL */
3615ea58
KW
419 0, /* POSIXD */
420 0, /* POSIXL */
421 0, /* POSIXU */
422 0, /* POSIXA */
423 0, /* NPOSIXD */
424 0, /* NPOSIXL */
425 0, /* NPOSIXU */
426 0, /* NPOSIXA */
03363afd
YO
427 0, /* CLUMP */
428 0, /* BRANCH */
03363afd 429 0, /* EXACT */
a4525e78 430 0, /* EXACTL */
03363afd
YO
431 0, /* EXACTF */
432 0, /* EXACTFL */
01f98ec2 433 0, /* EXACTFU */
8c1182fd 434 0, /* EXACTFA */
3c760661 435 0, /* EXACTFU_SS */
a4525e78 436 0, /* EXACTFLU8 */
098b07d5 437 0, /* EXACTFA_NO_TRIE */
03363afd
YO
438 0, /* NOTHING */
439 0, /* TAIL */
440 0, /* STAR */
441 0, /* PLUS */
442 0, /* CURLY */
443 0, /* CURLYN */
444 0, /* CURLYM */
445 0, /* CURLYX */
446 0, /* WHILEM */
447 0, /* OPEN */
448 0, /* CLOSE */
449 0, /* REF */
450 0, /* REFF */
451 0, /* REFFL */
01f98ec2 452 0, /* REFFU */
781aab5c 453 0, /* REFFA */
01f98ec2
KW
454 0, /* NREF */
455 0, /* NREFF */
456 0, /* NREFFL */
457 0, /* NREFFU */
781aab5c 458 0, /* NREFFA */
d3d47aac
YO
459 1, /* LONGJMP */
460 1, /* BRANCHJ */
03363afd
YO
461 2, /* IFMATCH */
462 2, /* UNLESSM */
463 1, /* SUSPEND */
464 1, /* IFTHEN */
465 0, /* GROUPP */
03363afd
YO
466 0, /* EVAL */
467 0, /* MINMOD */
468 0, /* LOGICAL */
469 1, /* RENUM */
470 0, /* TRIE */
471 0, /* TRIEC */
472 0, /* AHOCORASICK */
473 0, /* AHOCORASICKC */
1a147d38
YO
474 0, /* GOSUB */
475 0, /* GOSTART */
0a4db386 476 0, /* NGROUPP */
1a147d38 477 0, /* INSUBP */
0a4db386 478 0, /* DEFINEP */
e2e6a0f1 479 0, /* ENDLIKE */
7f69552c 480 0, /* OPFAIL */
e2e6a0f1
YO
481 0, /* ACCEPT */
482 0, /* VERB */
5d458dd8 483 0, /* PRUNE */
e2e6a0f1 484 0, /* MARKPOINT */
5d458dd8 485 0, /* SKIP */
e2e6a0f1 486 0, /* COMMIT */
5d458dd8 487 0, /* CUTGROUP */
ee9b8eae 488 0, /* KEEPS */
e1d1eefb 489 0, /* LNBREAK */
03363afd
YO
490 0, /* OPTIMIZED */
491 0, /* PSEUDO */
d09b2d29 492};
885f9e59 493
13d6edb4
NC
494#endif /* REG_COMP_C */
495
f83e001e 496
6bda09f9
YO
497/* reg_name[] - Opcode/state names in string form, for debugging */
498
22429478 499#ifndef DOINIT
13d6edb4 500EXTCONST char * PL_reg_name[];
22429478 501#else
4764e399 502EXTCONST char * const PL_reg_name[] = {
03363afd
YO
503 "END", /* 0000 */
504 "SUCCEED", /* 0x01 */
d3d47aac 505 "SBOL", /* 0x02 */
03363afd 506 "MBOL", /* 0x03 */
d3d47aac
YO
507 "SEOL", /* 0x04 */
508 "MEOL", /* 0x05 */
509 "EOS", /* 0x06 */
510 "GPOS", /* 0x07 */
511 "BOUND", /* 0x08 */
512 "BOUNDL", /* 0x09 */
513 "BOUNDU", /* 0x0a */
514 "BOUNDA", /* 0x0b */
515 "NBOUND", /* 0x0c */
516 "NBOUNDL", /* 0x0d */
517 "NBOUNDU", /* 0x0e */
518 "NBOUNDA", /* 0x0f */
519 "REG_ANY", /* 0x10 */
520 "SANY", /* 0x11 */
33c28ab2 521 "ANYOF", /* 0x12 */
ac44c12e
KW
522 "ANYOFD", /* 0x13 */
523 "ANYOFL", /* 0x14 */
524 "POSIXD", /* 0x15 */
525 "POSIXL", /* 0x16 */
526 "POSIXU", /* 0x17 */
527 "POSIXA", /* 0x18 */
528 "NPOSIXD", /* 0x19 */
529 "NPOSIXL", /* 0x1a */
530 "NPOSIXU", /* 0x1b */
531 "NPOSIXA", /* 0x1c */
532 "CLUMP", /* 0x1d */
533 "BRANCH", /* 0x1e */
534 "EXACT", /* 0x1f */
535 "EXACTL", /* 0x20 */
536 "EXACTF", /* 0x21 */
537 "EXACTFL", /* 0x22 */
538 "EXACTFU", /* 0x23 */
539 "EXACTFA", /* 0x24 */
540 "EXACTFU_SS", /* 0x25 */
541 "EXACTFLU8", /* 0x26 */
542 "EXACTFA_NO_TRIE", /* 0x27 */
543 "NOTHING", /* 0x28 */
544 "TAIL", /* 0x29 */
545 "STAR", /* 0x2a */
546 "PLUS", /* 0x2b */
547 "CURLY", /* 0x2c */
548 "CURLYN", /* 0x2d */
549 "CURLYM", /* 0x2e */
550 "CURLYX", /* 0x2f */
551 "WHILEM", /* 0x30 */
552 "OPEN", /* 0x31 */
553 "CLOSE", /* 0x32 */
554 "REF", /* 0x33 */
555 "REFF", /* 0x34 */
556 "REFFL", /* 0x35 */
557 "REFFU", /* 0x36 */
558 "REFFA", /* 0x37 */
559 "NREF", /* 0x38 */
560 "NREFF", /* 0x39 */
561 "NREFFL", /* 0x3a */
562 "NREFFU", /* 0x3b */
563 "NREFFA", /* 0x3c */
564 "LONGJMP", /* 0x3d */
565 "BRANCHJ", /* 0x3e */
566 "IFMATCH", /* 0x3f */
567 "UNLESSM", /* 0x40 */
568 "SUSPEND", /* 0x41 */
569 "IFTHEN", /* 0x42 */
570 "GROUPP", /* 0x43 */
571 "EVAL", /* 0x44 */
572 "MINMOD", /* 0x45 */
573 "LOGICAL", /* 0x46 */
574 "RENUM", /* 0x47 */
575 "TRIE", /* 0x48 */
576 "TRIEC", /* 0x49 */
577 "AHOCORASICK", /* 0x4a */
578 "AHOCORASICKC", /* 0x4b */
579 "GOSUB", /* 0x4c */
580 "GOSTART", /* 0x4d */
581 "NGROUPP", /* 0x4e */
582 "INSUBP", /* 0x4f */
583 "DEFINEP", /* 0x50 */
584 "ENDLIKE", /* 0x51 */
585 "OPFAIL", /* 0x52 */
586 "ACCEPT", /* 0x53 */
587 "VERB", /* 0x54 */
588 "PRUNE", /* 0x55 */
589 "MARKPOINT", /* 0x56 */
590 "SKIP", /* 0x57 */
591 "COMMIT", /* 0x58 */
592 "CUTGROUP", /* 0x59 */
593 "KEEPS", /* 0x5a */
594 "LNBREAK", /* 0x5b */
595 "OPTIMIZED", /* 0x5c */
596 "PSEUDO", /* 0x5d */
03363afd 597 /* ------------ States ------------- */
24b23f37
YO
598 "TRIE_next", /* REGNODE_MAX +0x01 */
599 "TRIE_next_fail", /* REGNODE_MAX +0x02 */
600 "EVAL_AB", /* REGNODE_MAX +0x03 */
601 "EVAL_AB_fail", /* REGNODE_MAX +0x04 */
602 "CURLYX_end", /* REGNODE_MAX +0x05 */
603 "CURLYX_end_fail", /* REGNODE_MAX +0x06 */
604 "WHILEM_A_pre", /* REGNODE_MAX +0x07 */
605 "WHILEM_A_pre_fail", /* REGNODE_MAX +0x08 */
606 "WHILEM_A_min", /* REGNODE_MAX +0x09 */
607 "WHILEM_A_min_fail", /* REGNODE_MAX +0x0a */
608 "WHILEM_A_max", /* REGNODE_MAX +0x0b */
609 "WHILEM_A_max_fail", /* REGNODE_MAX +0x0c */
610 "WHILEM_B_min", /* REGNODE_MAX +0x0d */
611 "WHILEM_B_min_fail", /* REGNODE_MAX +0x0e */
612 "WHILEM_B_max", /* REGNODE_MAX +0x0f */
613 "WHILEM_B_max_fail", /* REGNODE_MAX +0x10 */
614 "BRANCH_next", /* REGNODE_MAX +0x11 */
615 "BRANCH_next_fail", /* REGNODE_MAX +0x12 */
616 "CURLYM_A", /* REGNODE_MAX +0x13 */
617 "CURLYM_A_fail", /* REGNODE_MAX +0x14 */
618 "CURLYM_B", /* REGNODE_MAX +0x15 */
619 "CURLYM_B_fail", /* REGNODE_MAX +0x16 */
620 "IFMATCH_A", /* REGNODE_MAX +0x17 */
621 "IFMATCH_A_fail", /* REGNODE_MAX +0x18 */
622 "CURLY_B_min_known", /* REGNODE_MAX +0x19 */
623 "CURLY_B_min_known_fail", /* REGNODE_MAX +0x1a */
624 "CURLY_B_min", /* REGNODE_MAX +0x1b */
625 "CURLY_B_min_fail", /* REGNODE_MAX +0x1c */
626 "CURLY_B_max", /* REGNODE_MAX +0x1d */
627 "CURLY_B_max_fail", /* REGNODE_MAX +0x1e */
628 "COMMIT_next", /* REGNODE_MAX +0x1f */
629 "COMMIT_next_fail", /* REGNODE_MAX +0x20 */
e2e6a0f1
YO
630 "MARKPOINT_next", /* REGNODE_MAX +0x21 */
631 "MARKPOINT_next_fail", /* REGNODE_MAX +0x22 */
5d458dd8
YO
632 "SKIP_next", /* REGNODE_MAX +0x23 */
633 "SKIP_next_fail", /* REGNODE_MAX +0x24 */
634 "CUTGROUP_next", /* REGNODE_MAX +0x25 */
635 "CUTGROUP_next_fail", /* REGNODE_MAX +0x26 */
ee9b8eae
YO
636 "KEEPS_next", /* REGNODE_MAX +0x27 */
637 "KEEPS_next_fail", /* REGNODE_MAX +0x28 */
885f9e59 638};
22429478 639#endif /* DOINIT */
d09b2d29 640
f7819f85
A
641/* PL_reg_extflags_name[] - Opcode/state names in string form, for debugging */
642
643#ifndef DOINIT
644EXTCONST char * PL_reg_extflags_name[];
645#else
646EXTCONST char * const PL_reg_extflags_name[] = {
d262c0c7 647 /* Bits in extflags defined: 11111111111111110000111111111111 */
52d81aa8
NC
648 "MULTILINE", /* 0x00000001 */
649 "SINGLELINE", /* 0x00000002 */
650 "FOLD", /* 0x00000004 */
651 "EXTENDED", /* 0x00000008 */
334afb3e 652 "EXTENDED_MORE", /* 0x00000010 */
e3b64d84
KW
653 "NOCAPTURE", /* 0x00000020 */
654 "KEEPCOPY", /* 0x00000040 */
655 "CHARSET0", /* 0x00000080 : "CHARSET" - 0x00000380 */
656 "CHARSET1", /* 0x00000100 : "CHARSET" - 0x00000380 */
657 "CHARSET2", /* 0x00000200 : "CHARSET" - 0x00000380 */
d262c0c7
KW
658 "STRICT", /* 0x00000400 */
659 "SPLIT", /* 0x00000800 */
1d32d911
KW
660 "UNUSED_BIT_12", /* 0x00001000 */
661 "UNUSED_BIT_13", /* 0x00002000 */
662 "UNUSED_BIT_14", /* 0x00004000 */
a3b51d37
KW
663 "UNUSED_BIT_15", /* 0x00008000 */
664 "NO_INPLACE_SUBST", /* 0x00010000 */
665 "EVAL_SEEN", /* 0x00020000 */
ee273784 666 "UNBOUNDED_QUANTIFIER_SEEN",/* 0x00040000 */
e795e964 667 "CHECK_ALL", /* 0x00080000 */
52d81aa8
NC
668 "MATCH_UTF8", /* 0x00100000 */
669 "USE_INTUIT_NOML", /* 0x00200000 */
670 "USE_INTUIT_ML", /* 0x00400000 */
671 "INTUIT_TAIL", /* 0x00800000 */
a3b51d37 672 "IS_ANCHORED", /* 0x01000000 */
52d81aa8
NC
673 "COPY_DONE", /* 0x02000000 */
674 "TAINTED_SEEN", /* 0x04000000 */
675 "TAINTED", /* 0x08000000 */
676 "START_ONLY", /* 0x10000000 */
dbc200c5 677 "SKIPWHITE", /* 0x20000000 */
52d81aa8
NC
678 "WHITE", /* 0x40000000 */
679 "NULL", /* 0x80000000 */
f7819f85
A
680};
681#endif /* DOINIT */
682
adc2d0c9
JH
683#ifdef DEBUGGING
684# define REG_EXTFLAGS_NAME_SIZE 32
685#endif
686
337ff307
YO
687/* PL_reg_intflags_name[] - Opcode/state names in string form, for debugging */
688
689#ifndef DOINIT
690EXTCONST char * PL_reg_intflags_name[];
691#else
692EXTCONST char * const PL_reg_intflags_name[] = {
b8f6efdd
YO
693 "SKIP", /* 0x00000001 - PREGf_SKIP */
694 "IMPLICIT", /* 0x00000002 - PREGf_IMPLICIT - Converted .* to ^.* */
695 "NAUGHTY", /* 0x00000004 - PREGf_NAUGHTY - how exponential is this pattern? */
696 "VERBARG_SEEN", /* 0x00000008 - PREGf_VERBARG_SEEN */
697 "CUTGROUP_SEEN", /* 0x00000010 - PREGf_CUTGROUP_SEEN */
698 "USE_RE_EVAL", /* 0x00000020 - PREGf_USE_RE_EVAL - compiled with "use re 'eval'" */
58430ea8 699 "NOSCAN", /* 0x00000040 - PREGf_NOSCAN */
58430ea8
YO
700 "GPOS_SEEN", /* 0x00000100 - PREGf_GPOS_SEEN */
701 "GPOS_FLOAT", /* 0x00000200 - PREGf_GPOS_FLOAT */
d3d47aac
YO
702 "ANCH_MBOL", /* 0x00000400 - PREGf_ANCH_MBOL */
703 "ANCH_SBOL", /* 0x00000800 - PREGf_ANCH_SBOL */
704 "ANCH_GPOS", /* 0x00001000 - PREGf_ANCH_GPOS */
337ff307
YO
705};
706#endif /* DOINIT */
707
adc2d0c9 708#ifdef DEBUGGING
33c28ab2 709# define REG_INTFLAGS_NAME_SIZE 12
adc2d0c9
JH
710#endif
711
f9ef50a7 712/* The following have no fixed length. U8 so we can do strchr() on it. */
ded4dd2a 713#define REGNODE_VARIES(node) (PL_varies_bitmask[(node) >> 3] & (1 << ((node) & 7)))
e52fc539 714
f9ef50a7 715#ifndef DOINIT
ded4dd2a 716EXTCONST U8 PL_varies[] __attribute__deprecated__;
f9ef50a7 717#else
ded4dd2a 718EXTCONST U8 PL_varies[] __attribute__deprecated__ = {
62e6ef33
AC
719 CLUMP, BRANCH, STAR, PLUS, CURLY, CURLYN, CURLYM, CURLYX, WHILEM, REF,
720 REFF, REFFL, REFFU, REFFA, NREF, NREFF, NREFFL, NREFFU, NREFFA,
d3d47aac 721 BRANCHJ, SUSPEND, IFTHEN,
f9ef50a7
NC
722 0
723};
724#endif /* DOINIT */
725
ded4dd2a
NC
726#ifndef DOINIT
727EXTCONST U8 PL_varies_bitmask[];
728#else
729EXTCONST U8 PL_varies_bitmask[] = {
ac44c12e 730 0x00, 0x00, 0x00, 0x60, 0x00, 0xFC, 0xF9, 0x5F, 0x06, 0x00, 0x00, 0x00
ded4dd2a
NC
731};
732#endif /* DOINIT */
733
f9ef50a7
NC
734/* The following always have a length of 1. U8 we can do strchr() on it. */
735/* (Note that length 1 means "one character" under UTF8, not "one octet".) */
ded4dd2a 736#define REGNODE_SIMPLE(node) (PL_simple_bitmask[(node) >> 3] & (1 << ((node) & 7)))
e52fc539 737
f9ef50a7 738#ifndef DOINIT
ded4dd2a 739EXTCONST U8 PL_simple[] __attribute__deprecated__;
f9ef50a7 740#else
ded4dd2a 741EXTCONST U8 PL_simple[] __attribute__deprecated__ = {
ac44c12e
KW
742 REG_ANY, SANY, ANYOF, ANYOFD, ANYOFL, POSIXD, POSIXL, POSIXU, POSIXA,
743 NPOSIXD, NPOSIXL, NPOSIXU, NPOSIXA,
f9ef50a7
NC
744 0
745};
746#endif /* DOINIT */
747
ded4dd2a
NC
748#ifndef DOINIT
749EXTCONST U8 PL_simple_bitmask[];
750#else
751EXTCONST U8 PL_simple_bitmask[] = {
ac44c12e 752 0x00, 0x00, 0xFF, 0x1F, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
ded4dd2a
NC
753};
754#endif /* DOINIT */
755
37442d52 756/* ex: set ro: */