This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
regcomp.h: Add some macros
[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
67a1b5f9 9#define REGNODE_MAX 97
21cbe009 10#define REGMATCH_STATE_MAX 137
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 */
67a1b5f9
KW
35#define ANYOFM 21 /* 0x15 Like ANYOF, but matches an invariant byte as determined by the mask and arg */
36#define POSIXD 22 /* 0x16 Some [[:class:]] under /d; the FLAGS field gives which one */
37#define POSIXL 23 /* 0x17 Some [[:class:]] under /l; the FLAGS field gives which one */
38#define POSIXU 24 /* 0x18 Some [[:class:]] under /u; the FLAGS field gives which one */
39#define POSIXA 25 /* 0x19 Some [[:class:]] under /a; the FLAGS field gives which one */
40#define NPOSIXD 26 /* 0x1a complement of POSIXD, [[:^class:]] */
41#define NPOSIXL 27 /* 0x1b complement of POSIXL, [[:^class:]] */
42#define NPOSIXU 28 /* 0x1c complement of POSIXU, [[:^class:]] */
43#define NPOSIXA 29 /* 0x1d complement of POSIXA, [[:^class:]] */
44#define ASCII 30 /* 0x1e [[:ascii:]] */
45#define NASCII 31 /* 0x1f [[:^ascii:]] */
46#define CLUMP 32 /* 0x20 Match any extended grapheme cluster sequence */
47#define BRANCH 33 /* 0x21 Match this alternative, or the next... */
48#define EXACT 34 /* 0x22 Match this string (preceded by length). */
49#define EXACTL 35 /* 0x23 Like EXACT, but /l is in effect (used so locale-related warnings can be checked for). */
50#define EXACTF 36 /* 0x24 Match this non-UTF-8 string (not guaranteed to be folded) using /id rules (w/len). */
51#define EXACTFL 37 /* 0x25 Match this string (not guaranteed to be folded) using /il rules (w/len). */
52#define EXACTFU 38 /* 0x26 Match this string (folded iff in UTF-8, length in folding doesn't change if not in UTF-8) using /iu rules (w/len). */
89829bb5 53#define EXACTFAA 39 /* 0x27 Match this string (not guaranteed to be folded) using /iaa rules (w/len). */
67a1b5f9 54#define EXACTFU_SS 40 /* 0x28 Match this string (folded iff in UTF-8, length in folding may change even if not in UTF-8) using /iu rules (w/len). */
5baf6663 55#define EXACTFLU8 41 /* 0x29 Rare circumstances: like EXACTFU, but is under /l, UTF-8, folded, and everything in it is above 255. */
89829bb5 56#define EXACTFAA_NO_TRIE 42 /* 0x2a Match this string (which is not trie-able; not guaranteed to be folded) using /iaa rules (w/len). */
67a1b5f9
KW
57#define NOTHING 43 /* 0x2b Match empty string. */
58#define TAIL 44 /* 0x2c Match empty string. Can jump here from outside. */
59#define STAR 45 /* 0x2d Match this (simple) thing 0 or more times. */
60#define PLUS 46 /* 0x2e Match this (simple) thing 1 or more times. */
61#define CURLY 47 /* 0x2f Match this simple thing {n,m} times. */
62#define CURLYN 48 /* 0x30 Capture next-after-this simple thing */
63#define CURLYM 49 /* 0x31 Capture this medium-complex thing {n,m} times. */
64#define CURLYX 50 /* 0x32 Match this complex thing {n,m} times. */
65#define WHILEM 51 /* 0x33 Do curly processing and see if rest matches. */
66#define OPEN 52 /* 0x34 Mark this point in input as start of #n. */
67#define CLOSE 53 /* 0x35 Close corresponding OPEN of #n. */
68#define SROPEN 54 /* 0x36 Same as OPEN, but for script run */
69#define SRCLOSE 55 /* 0x37 Close preceding SROPEN */
70#define REF 56 /* 0x38 Match some already matched string */
71#define REFF 57 /* 0x39 Match already matched string, folded using native charset rules for non-utf8 */
72#define REFFL 58 /* 0x3a Match already matched string, folded in loc. */
73#define REFFU 59 /* 0x3b Match already matched string, folded using unicode rules for non-utf8 */
74#define REFFA 60 /* 0x3c Match already matched string, folded using unicode rules for non-utf8, no mixing ASCII, non-ASCII */
75#define NREF 61 /* 0x3d Match some already matched string */
76#define NREFF 62 /* 0x3e Match already matched string, folded using native charset rules for non-utf8 */
77#define NREFFL 63 /* 0x3f Match already matched string, folded in loc. */
78#define NREFFU 64 /* 0x40 Match already matched string, folded using unicode rules for non-utf8 */
79#define NREFFA 65 /* 0x41 Match already matched string, folded using unicode rules for non-utf8, no mixing ASCII, non-ASCII */
80#define LONGJMP 66 /* 0x42 Jump far away. */
81#define BRANCHJ 67 /* 0x43 BRANCH with long offset. */
82#define IFMATCH 68 /* 0x44 Succeeds if the following matches. */
83#define UNLESSM 69 /* 0x45 Fails if the following matches. */
84#define SUSPEND 70 /* 0x46 "Independent" sub-RE. */
85#define IFTHEN 71 /* 0x47 Switch, should be preceded by switcher. */
86#define GROUPP 72 /* 0x48 Whether the group matched. */
87#define EVAL 73 /* 0x49 Execute some Perl code. */
88#define MINMOD 74 /* 0x4a Next operator is not greedy. */
89#define LOGICAL 75 /* 0x4b Next opcode should set the flag only. */
90#define RENUM 76 /* 0x4c Group with independently numbered parens. */
91#define TRIE 77 /* 0x4d Match many EXACT(F[ALU]?)? at once. flags==type */
92#define TRIEC 78 /* 0x4e Same as TRIE, but with embedded charclass data */
93#define AHOCORASICK 79 /* 0x4f Aho Corasick stclass. flags==type */
94#define AHOCORASICKC 80 /* 0x50 Same as AHOCORASICK, but with embedded charclass data */
95#define GOSUB 81 /* 0x51 recurse to paren arg1 at (signed) ofs arg2 */
96#define NGROUPP 82 /* 0x52 Whether the group matched. */
97#define INSUBP 83 /* 0x53 Whether we are in a specific recurse. */
98#define DEFINEP 84 /* 0x54 Never execute directly. */
99#define ENDLIKE 85 /* 0x55 Used only for the type field of verbs */
100#define OPFAIL 86 /* 0x56 Same as (?!), but with verb arg */
101#define ACCEPT 87 /* 0x57 Accepts the current matched string, with verbar */
102#define VERB 88 /* 0x58 Used only for the type field of verbs */
103#define PRUNE 89 /* 0x59 Pattern fails at this startpoint if no-backtracking through this */
104#define MARKPOINT 90 /* 0x5a Push the current location for rollback by cut. */
105#define SKIP 91 /* 0x5b On failure skip forward (to the mark) before retrying */
106#define COMMIT 92 /* 0x5c Pattern fails outright if backtracking through this */
107#define CUTGROUP 93 /* 0x5d On failure go to the next alternation in the group */
108#define KEEPS 94 /* 0x5e $& begins here. */
109#define LNBREAK 95 /* 0x5f generic newline pattern */
110#define OPTIMIZED 96 /* 0x60 Placeholder for dump. */
111#define PSEUDO 97 /* 0x61 Pseudo opcode for internal use. */
03363afd 112 /* ------------ States ------------- */
24b23f37
YO
113#define TRIE_next (REGNODE_MAX + 1) /* state for TRIE */
114#define TRIE_next_fail (REGNODE_MAX + 2) /* state for TRIE */
4ee16520
DM
115#define EVAL_B (REGNODE_MAX + 3) /* state for EVAL */
116#define EVAL_B_fail (REGNODE_MAX + 4) /* state for EVAL */
117#define EVAL_postponed_AB (REGNODE_MAX + 5) /* state for EVAL */
118#define EVAL_postponed_AB_fail (REGNODE_MAX + 6) /* state for EVAL */
119#define CURLYX_end (REGNODE_MAX + 7) /* state for CURLYX */
120#define CURLYX_end_fail (REGNODE_MAX + 8) /* state for CURLYX */
121#define WHILEM_A_pre (REGNODE_MAX + 9) /* state for WHILEM */
122#define WHILEM_A_pre_fail (REGNODE_MAX + 10) /* state for WHILEM */
123#define WHILEM_A_min (REGNODE_MAX + 11) /* state for WHILEM */
124#define WHILEM_A_min_fail (REGNODE_MAX + 12) /* state for WHILEM */
125#define WHILEM_A_max (REGNODE_MAX + 13) /* state for WHILEM */
126#define WHILEM_A_max_fail (REGNODE_MAX + 14) /* state for WHILEM */
127#define WHILEM_B_min (REGNODE_MAX + 15) /* state for WHILEM */
128#define WHILEM_B_min_fail (REGNODE_MAX + 16) /* state for WHILEM */
129#define WHILEM_B_max (REGNODE_MAX + 17) /* state for WHILEM */
130#define WHILEM_B_max_fail (REGNODE_MAX + 18) /* state for WHILEM */
131#define BRANCH_next (REGNODE_MAX + 19) /* state for BRANCH */
132#define BRANCH_next_fail (REGNODE_MAX + 20) /* state for BRANCH */
133#define CURLYM_A (REGNODE_MAX + 21) /* state for CURLYM */
134#define CURLYM_A_fail (REGNODE_MAX + 22) /* state for CURLYM */
135#define CURLYM_B (REGNODE_MAX + 23) /* state for CURLYM */
136#define CURLYM_B_fail (REGNODE_MAX + 24) /* state for CURLYM */
137#define IFMATCH_A (REGNODE_MAX + 25) /* state for IFMATCH */
138#define IFMATCH_A_fail (REGNODE_MAX + 26) /* state for IFMATCH */
21cbe009
DM
139#define CURLY_B_min (REGNODE_MAX + 27) /* state for CURLY */
140#define CURLY_B_min_fail (REGNODE_MAX + 28) /* state for CURLY */
141#define CURLY_B_max (REGNODE_MAX + 29) /* state for CURLY */
142#define CURLY_B_max_fail (REGNODE_MAX + 30) /* state for CURLY */
143#define COMMIT_next (REGNODE_MAX + 31) /* state for COMMIT */
144#define COMMIT_next_fail (REGNODE_MAX + 32) /* state for COMMIT */
145#define MARKPOINT_next (REGNODE_MAX + 33) /* state for MARKPOINT */
146#define MARKPOINT_next_fail (REGNODE_MAX + 34) /* state for MARKPOINT */
147#define SKIP_next (REGNODE_MAX + 35) /* state for SKIP */
148#define SKIP_next_fail (REGNODE_MAX + 36) /* state for SKIP */
149#define CUTGROUP_next (REGNODE_MAX + 37) /* state for CUTGROUP */
150#define CUTGROUP_next_fail (REGNODE_MAX + 38) /* state for CUTGROUP */
151#define KEEPS_next (REGNODE_MAX + 39) /* state for KEEPS */
152#define KEEPS_next_fail (REGNODE_MAX + 40) /* state for KEEPS */
03363afd 153
6bda09f9 154/* PL_regkind[] What type of regop or state is this. */
d09b2d29
IZ
155
156#ifndef DOINIT
22c35a8c 157EXTCONST U8 PL_regkind[];
d09b2d29 158#else
22c35a8c 159EXTCONST U8 PL_regkind[] = {
e2e6a0f1
YO
160 END, /* END */
161 END, /* SUCCEED */
e2e6a0f1 162 BOL, /* SBOL */
d3d47aac 163 BOL, /* MBOL */
e2e6a0f1 164 EOL, /* SEOL */
d3d47aac
YO
165 EOL, /* MEOL */
166 EOL, /* EOS */
167 GPOS, /* GPOS */
e2e6a0f1
YO
168 BOUND, /* BOUND */
169 BOUND, /* BOUNDL */
1e355c70 170 BOUND, /* BOUNDU */
0c6e81eb 171 BOUND, /* BOUNDA */
e2e6a0f1
YO
172 NBOUND, /* NBOUND */
173 NBOUND, /* NBOUNDL */
1e355c70 174 NBOUND, /* NBOUNDU */
0c6e81eb 175 NBOUND, /* NBOUNDA */
e2e6a0f1
YO
176 REG_ANY, /* REG_ANY */
177 REG_ANY, /* SANY */
e2e6a0f1 178 ANYOF, /* ANYOF */
ac44c12e 179 ANYOF, /* ANYOFD */
a4525e78 180 ANYOF, /* ANYOFL */
67a1b5f9 181 ANYOFM, /* ANYOFM */
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 */
89829bb5 199 EXACT, /* EXACTFAA */
3c760661 200 EXACT, /* EXACTFU_SS */
a4525e78 201 EXACT, /* EXACTFLU8 */
89829bb5 202 EXACT, /* EXACTFAA_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 */
e2e6a0f1
YO
285 CURLY, /* CURLY_B_min */
286 CURLY, /* CURLY_B_min_fail */
287 CURLY, /* CURLY_B_max */
288 CURLY, /* CURLY_B_max_fail */
289 COMMIT, /* COMMIT_next */
290 COMMIT, /* COMMIT_next_fail */
291 MARKPOINT, /* MARKPOINT_next */
292 MARKPOINT, /* MARKPOINT_next_fail */
5d458dd8
YO
293 SKIP, /* SKIP_next */
294 SKIP, /* SKIP_next_fail */
295 CUTGROUP, /* CUTGROUP_next */
296 CUTGROUP, /* CUTGROUP_next_fail */
ee9b8eae
YO
297 KEEPS, /* KEEPS_next */
298 KEEPS, /* KEEPS_next_fail */
d09b2d29
IZ
299};
300#endif
301
f83e001e
YO
302#ifdef REG_COMP_C
303
6bda09f9 304/* regarglen[] - How large is the argument part of the node (in regnodes) */
d09b2d29 305
29de9391 306static const U8 regarglen[] = {
03363afd
YO
307 0, /* END */
308 0, /* SUCCEED */
03363afd 309 0, /* SBOL */
d3d47aac 310 0, /* MBOL */
03363afd 311 0, /* SEOL */
d3d47aac
YO
312 0, /* MEOL */
313 0, /* EOS */
314 0, /* GPOS */
03363afd
YO
315 0, /* BOUND */
316 0, /* BOUNDL */
1e355c70 317 0, /* BOUNDU */
0c6e81eb 318 0, /* BOUNDA */
03363afd
YO
319 0, /* NBOUND */
320 0, /* NBOUNDL */
1e355c70 321 0, /* NBOUNDU */
0c6e81eb 322 0, /* NBOUNDA */
03363afd
YO
323 0, /* REG_ANY */
324 0, /* SANY */
975a06f7 325 EXTRA_SIZE(struct regnode_1), /* ANYOF */
ac44c12e 326 EXTRA_SIZE(struct regnode_1), /* ANYOFD */
a4525e78 327 EXTRA_SIZE(struct regnode_1), /* ANYOFL */
67a1b5f9 328 EXTRA_SIZE(struct regnode_1), /* ANYOFM */
3615ea58
KW
329 0, /* POSIXD */
330 0, /* POSIXL */
331 0, /* POSIXU */
332 0, /* POSIXA */
333 0, /* NPOSIXD */
334 0, /* NPOSIXL */
335 0, /* NPOSIXU */
336 0, /* NPOSIXA */
39d24220
KW
337 0, /* ASCII */
338 0, /* NASCII */
03363afd
YO
339 0, /* CLUMP */
340 0, /* BRANCH */
03363afd 341 0, /* EXACT */
a4525e78 342 0, /* EXACTL */
03363afd
YO
343 0, /* EXACTF */
344 0, /* EXACTFL */
01f98ec2 345 0, /* EXACTFU */
89829bb5 346 0, /* EXACTFAA */
3c760661 347 0, /* EXACTFU_SS */
a4525e78 348 0, /* EXACTFLU8 */
89829bb5 349 0, /* EXACTFAA_NO_TRIE */
03363afd
YO
350 0, /* NOTHING */
351 0, /* TAIL */
352 0, /* STAR */
353 0, /* PLUS */
354 EXTRA_SIZE(struct regnode_2), /* CURLY */
355 EXTRA_SIZE(struct regnode_2), /* CURLYN */
356 EXTRA_SIZE(struct regnode_2), /* CURLYM */
357 EXTRA_SIZE(struct regnode_2), /* CURLYX */
358 0, /* WHILEM */
359 EXTRA_SIZE(struct regnode_1), /* OPEN */
360 EXTRA_SIZE(struct regnode_1), /* CLOSE */
07093db4
KW
361 0, /* SROPEN */
362 0, /* SRCLOSE */
03363afd
YO
363 EXTRA_SIZE(struct regnode_1), /* REF */
364 EXTRA_SIZE(struct regnode_1), /* REFF */
365 EXTRA_SIZE(struct regnode_1), /* REFFL */
01f98ec2 366 EXTRA_SIZE(struct regnode_1), /* REFFU */
781aab5c 367 EXTRA_SIZE(struct regnode_1), /* REFFA */
01f98ec2
KW
368 EXTRA_SIZE(struct regnode_1), /* NREF */
369 EXTRA_SIZE(struct regnode_1), /* NREFF */
370 EXTRA_SIZE(struct regnode_1), /* NREFFL */
371 EXTRA_SIZE(struct regnode_1), /* NREFFU */
781aab5c 372 EXTRA_SIZE(struct regnode_1), /* NREFFA */
d3d47aac
YO
373 EXTRA_SIZE(struct regnode_1), /* LONGJMP */
374 EXTRA_SIZE(struct regnode_1), /* BRANCHJ */
03363afd
YO
375 EXTRA_SIZE(struct regnode_1), /* IFMATCH */
376 EXTRA_SIZE(struct regnode_1), /* UNLESSM */
377 EXTRA_SIZE(struct regnode_1), /* SUSPEND */
378 EXTRA_SIZE(struct regnode_1), /* IFTHEN */
379 EXTRA_SIZE(struct regnode_1), /* GROUPP */
13f27704 380 EXTRA_SIZE(struct regnode_2L), /* EVAL */
03363afd
YO
381 0, /* MINMOD */
382 0, /* LOGICAL */
383 EXTRA_SIZE(struct regnode_1), /* RENUM */
384 EXTRA_SIZE(struct regnode_1), /* TRIE */
385 EXTRA_SIZE(struct regnode_charclass), /* TRIEC */
386 EXTRA_SIZE(struct regnode_1), /* AHOCORASICK */
387 EXTRA_SIZE(struct regnode_charclass), /* AHOCORASICKC */
1a147d38 388 EXTRA_SIZE(struct regnode_2L), /* GOSUB */
0a4db386 389 EXTRA_SIZE(struct regnode_1), /* NGROUPP */
1a147d38 390 EXTRA_SIZE(struct regnode_1), /* INSUBP */
0a4db386 391 EXTRA_SIZE(struct regnode_1), /* DEFINEP */
e2e6a0f1 392 0, /* ENDLIKE */
fee50582
YO
393 EXTRA_SIZE(struct regnode_1), /* OPFAIL */
394 EXTRA_SIZE(struct regnode_2L), /* ACCEPT */
20832bc5 395 EXTRA_SIZE(struct regnode_1), /* VERB */
5d458dd8 396 EXTRA_SIZE(struct regnode_1), /* PRUNE */
e2e6a0f1 397 EXTRA_SIZE(struct regnode_1), /* MARKPOINT */
5d458dd8 398 EXTRA_SIZE(struct regnode_1), /* SKIP */
e2e6a0f1 399 EXTRA_SIZE(struct regnode_1), /* COMMIT */
5d458dd8 400 EXTRA_SIZE(struct regnode_1), /* CUTGROUP */
ee9b8eae 401 0, /* KEEPS */
e1d1eefb 402 0, /* LNBREAK */
03363afd
YO
403 0, /* OPTIMIZED */
404 0, /* PSEUDO */
d09b2d29
IZ
405};
406
6bda09f9
YO
407/* reg_off_by_arg[] - Which argument holds the offset to the next node */
408
29de9391 409static const char reg_off_by_arg[] = {
03363afd
YO
410 0, /* END */
411 0, /* SUCCEED */
03363afd 412 0, /* SBOL */
d3d47aac 413 0, /* MBOL */
03363afd 414 0, /* SEOL */
d3d47aac
YO
415 0, /* MEOL */
416 0, /* EOS */
417 0, /* GPOS */
03363afd
YO
418 0, /* BOUND */
419 0, /* BOUNDL */
1e355c70 420 0, /* BOUNDU */
0c6e81eb 421 0, /* BOUNDA */
03363afd
YO
422 0, /* NBOUND */
423 0, /* NBOUNDL */
1e355c70 424 0, /* NBOUNDU */
0c6e81eb 425 0, /* NBOUNDA */
03363afd
YO
426 0, /* REG_ANY */
427 0, /* SANY */
03363afd 428 0, /* ANYOF */
ac44c12e 429 0, /* ANYOFD */
a4525e78 430 0, /* ANYOFL */
67a1b5f9 431 0, /* ANYOFM */
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 */
89829bb5 449 0, /* EXACTFAA */
3c760661 450 0, /* EXACTFU_SS */
a4525e78 451 0, /* EXACTFLU8 */
89829bb5 452 0, /* EXACTFAA_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 */
67a1b5f9
KW
540 "ANYOFM", /* 0x15 */
541 "POSIXD", /* 0x16 */
542 "POSIXL", /* 0x17 */
543 "POSIXU", /* 0x18 */
544 "POSIXA", /* 0x19 */
545 "NPOSIXD", /* 0x1a */
546 "NPOSIXL", /* 0x1b */
547 "NPOSIXU", /* 0x1c */
548 "NPOSIXA", /* 0x1d */
549 "ASCII", /* 0x1e */
550 "NASCII", /* 0x1f */
551 "CLUMP", /* 0x20 */
552 "BRANCH", /* 0x21 */
553 "EXACT", /* 0x22 */
554 "EXACTL", /* 0x23 */
555 "EXACTF", /* 0x24 */
556 "EXACTFL", /* 0x25 */
557 "EXACTFU", /* 0x26 */
89829bb5 558 "EXACTFAA", /* 0x27 */
67a1b5f9
KW
559 "EXACTFU_SS", /* 0x28 */
560 "EXACTFLU8", /* 0x29 */
89829bb5 561 "EXACTFAA_NO_TRIE", /* 0x2a */
67a1b5f9
KW
562 "NOTHING", /* 0x2b */
563 "TAIL", /* 0x2c */
564 "STAR", /* 0x2d */
565 "PLUS", /* 0x2e */
566 "CURLY", /* 0x2f */
567 "CURLYN", /* 0x30 */
568 "CURLYM", /* 0x31 */
569 "CURLYX", /* 0x32 */
570 "WHILEM", /* 0x33 */
571 "OPEN", /* 0x34 */
572 "CLOSE", /* 0x35 */
573 "SROPEN", /* 0x36 */
574 "SRCLOSE", /* 0x37 */
575 "REF", /* 0x38 */
576 "REFF", /* 0x39 */
577 "REFFL", /* 0x3a */
578 "REFFU", /* 0x3b */
579 "REFFA", /* 0x3c */
580 "NREF", /* 0x3d */
581 "NREFF", /* 0x3e */
582 "NREFFL", /* 0x3f */
583 "NREFFU", /* 0x40 */
584 "NREFFA", /* 0x41 */
585 "LONGJMP", /* 0x42 */
586 "BRANCHJ", /* 0x43 */
587 "IFMATCH", /* 0x44 */
588 "UNLESSM", /* 0x45 */
589 "SUSPEND", /* 0x46 */
590 "IFTHEN", /* 0x47 */
591 "GROUPP", /* 0x48 */
592 "EVAL", /* 0x49 */
593 "MINMOD", /* 0x4a */
594 "LOGICAL", /* 0x4b */
595 "RENUM", /* 0x4c */
596 "TRIE", /* 0x4d */
597 "TRIEC", /* 0x4e */
598 "AHOCORASICK", /* 0x4f */
599 "AHOCORASICKC", /* 0x50 */
600 "GOSUB", /* 0x51 */
601 "NGROUPP", /* 0x52 */
602 "INSUBP", /* 0x53 */
603 "DEFINEP", /* 0x54 */
604 "ENDLIKE", /* 0x55 */
605 "OPFAIL", /* 0x56 */
606 "ACCEPT", /* 0x57 */
607 "VERB", /* 0x58 */
608 "PRUNE", /* 0x59 */
609 "MARKPOINT", /* 0x5a */
610 "SKIP", /* 0x5b */
611 "COMMIT", /* 0x5c */
612 "CUTGROUP", /* 0x5d */
613 "KEEPS", /* 0x5e */
614 "LNBREAK", /* 0x5f */
615 "OPTIMIZED", /* 0x60 */
616 "PSEUDO", /* 0x61 */
03363afd 617 /* ------------ States ------------- */
24b23f37
YO
618 "TRIE_next", /* REGNODE_MAX +0x01 */
619 "TRIE_next_fail", /* REGNODE_MAX +0x02 */
4ee16520
DM
620 "EVAL_B", /* REGNODE_MAX +0x03 */
621 "EVAL_B_fail", /* REGNODE_MAX +0x04 */
622 "EVAL_postponed_AB", /* REGNODE_MAX +0x05 */
623 "EVAL_postponed_AB_fail", /* REGNODE_MAX +0x06 */
624 "CURLYX_end", /* REGNODE_MAX +0x07 */
625 "CURLYX_end_fail", /* REGNODE_MAX +0x08 */
626 "WHILEM_A_pre", /* REGNODE_MAX +0x09 */
627 "WHILEM_A_pre_fail", /* REGNODE_MAX +0x0a */
628 "WHILEM_A_min", /* REGNODE_MAX +0x0b */
629 "WHILEM_A_min_fail", /* REGNODE_MAX +0x0c */
630 "WHILEM_A_max", /* REGNODE_MAX +0x0d */
631 "WHILEM_A_max_fail", /* REGNODE_MAX +0x0e */
632 "WHILEM_B_min", /* REGNODE_MAX +0x0f */
633 "WHILEM_B_min_fail", /* REGNODE_MAX +0x10 */
634 "WHILEM_B_max", /* REGNODE_MAX +0x11 */
635 "WHILEM_B_max_fail", /* REGNODE_MAX +0x12 */
636 "BRANCH_next", /* REGNODE_MAX +0x13 */
637 "BRANCH_next_fail", /* REGNODE_MAX +0x14 */
638 "CURLYM_A", /* REGNODE_MAX +0x15 */
639 "CURLYM_A_fail", /* REGNODE_MAX +0x16 */
640 "CURLYM_B", /* REGNODE_MAX +0x17 */
641 "CURLYM_B_fail", /* REGNODE_MAX +0x18 */
642 "IFMATCH_A", /* REGNODE_MAX +0x19 */
643 "IFMATCH_A_fail", /* REGNODE_MAX +0x1a */
21cbe009
DM
644 "CURLY_B_min", /* REGNODE_MAX +0x1b */
645 "CURLY_B_min_fail", /* REGNODE_MAX +0x1c */
646 "CURLY_B_max", /* REGNODE_MAX +0x1d */
647 "CURLY_B_max_fail", /* REGNODE_MAX +0x1e */
648 "COMMIT_next", /* REGNODE_MAX +0x1f */
649 "COMMIT_next_fail", /* REGNODE_MAX +0x20 */
650 "MARKPOINT_next", /* REGNODE_MAX +0x21 */
651 "MARKPOINT_next_fail", /* REGNODE_MAX +0x22 */
652 "SKIP_next", /* REGNODE_MAX +0x23 */
653 "SKIP_next_fail", /* REGNODE_MAX +0x24 */
654 "CUTGROUP_next", /* REGNODE_MAX +0x25 */
655 "CUTGROUP_next_fail", /* REGNODE_MAX +0x26 */
656 "KEEPS_next", /* REGNODE_MAX +0x27 */
657 "KEEPS_next_fail", /* REGNODE_MAX +0x28 */
885f9e59 658};
22429478 659#endif /* DOINIT */
d09b2d29 660
f7819f85
A
661/* PL_reg_extflags_name[] - Opcode/state names in string form, for debugging */
662
663#ifndef DOINIT
664EXTCONST char * PL_reg_extflags_name[];
665#else
666EXTCONST char * const PL_reg_extflags_name[] = {
d262c0c7 667 /* Bits in extflags defined: 11111111111111110000111111111111 */
52d81aa8
NC
668 "MULTILINE", /* 0x00000001 */
669 "SINGLELINE", /* 0x00000002 */
670 "FOLD", /* 0x00000004 */
671 "EXTENDED", /* 0x00000008 */
334afb3e 672 "EXTENDED_MORE", /* 0x00000010 */
e3b64d84
KW
673 "NOCAPTURE", /* 0x00000020 */
674 "KEEPCOPY", /* 0x00000040 */
675 "CHARSET0", /* 0x00000080 : "CHARSET" - 0x00000380 */
676 "CHARSET1", /* 0x00000100 : "CHARSET" - 0x00000380 */
677 "CHARSET2", /* 0x00000200 : "CHARSET" - 0x00000380 */
d262c0c7
KW
678 "STRICT", /* 0x00000400 */
679 "SPLIT", /* 0x00000800 */
1d32d911
KW
680 "UNUSED_BIT_12", /* 0x00001000 */
681 "UNUSED_BIT_13", /* 0x00002000 */
682 "UNUSED_BIT_14", /* 0x00004000 */
a3b51d37
KW
683 "UNUSED_BIT_15", /* 0x00008000 */
684 "NO_INPLACE_SUBST", /* 0x00010000 */
685 "EVAL_SEEN", /* 0x00020000 */
ee273784 686 "UNBOUNDED_QUANTIFIER_SEEN",/* 0x00040000 */
e795e964 687 "CHECK_ALL", /* 0x00080000 */
52d81aa8
NC
688 "MATCH_UTF8", /* 0x00100000 */
689 "USE_INTUIT_NOML", /* 0x00200000 */
690 "USE_INTUIT_ML", /* 0x00400000 */
691 "INTUIT_TAIL", /* 0x00800000 */
a3b51d37 692 "IS_ANCHORED", /* 0x01000000 */
52d81aa8
NC
693 "COPY_DONE", /* 0x02000000 */
694 "TAINTED_SEEN", /* 0x04000000 */
695 "TAINTED", /* 0x08000000 */
696 "START_ONLY", /* 0x10000000 */
dbc200c5 697 "SKIPWHITE", /* 0x20000000 */
52d81aa8
NC
698 "WHITE", /* 0x40000000 */
699 "NULL", /* 0x80000000 */
f7819f85
A
700};
701#endif /* DOINIT */
702
adc2d0c9
JH
703#ifdef DEBUGGING
704# define REG_EXTFLAGS_NAME_SIZE 32
705#endif
706
337ff307
YO
707/* PL_reg_intflags_name[] - Opcode/state names in string form, for debugging */
708
709#ifndef DOINIT
710EXTCONST char * PL_reg_intflags_name[];
711#else
712EXTCONST char * const PL_reg_intflags_name[] = {
b8f6efdd
YO
713 "SKIP", /* 0x00000001 - PREGf_SKIP */
714 "IMPLICIT", /* 0x00000002 - PREGf_IMPLICIT - Converted .* to ^.* */
715 "NAUGHTY", /* 0x00000004 - PREGf_NAUGHTY - how exponential is this pattern? */
716 "VERBARG_SEEN", /* 0x00000008 - PREGf_VERBARG_SEEN */
717 "CUTGROUP_SEEN", /* 0x00000010 - PREGf_CUTGROUP_SEEN */
718 "USE_RE_EVAL", /* 0x00000020 - PREGf_USE_RE_EVAL - compiled with "use re 'eval'" */
58430ea8 719 "NOSCAN", /* 0x00000040 - PREGf_NOSCAN */
58430ea8
YO
720 "GPOS_SEEN", /* 0x00000100 - PREGf_GPOS_SEEN */
721 "GPOS_FLOAT", /* 0x00000200 - PREGf_GPOS_FLOAT */
d3d47aac
YO
722 "ANCH_MBOL", /* 0x00000400 - PREGf_ANCH_MBOL */
723 "ANCH_SBOL", /* 0x00000800 - PREGf_ANCH_SBOL */
724 "ANCH_GPOS", /* 0x00001000 - PREGf_ANCH_GPOS */
d5a00e4a 725 "RECURSE_SEEN", /* 0x00002000 - PREGf_RECURSE_SEEN */
337ff307
YO
726};
727#endif /* DOINIT */
728
adc2d0c9 729#ifdef DEBUGGING
d5a00e4a 730# define REG_INTFLAGS_NAME_SIZE 13
adc2d0c9
JH
731#endif
732
f9ef50a7 733/* The following have no fixed length. U8 so we can do strchr() on it. */
ded4dd2a 734#define REGNODE_VARIES(node) (PL_varies_bitmask[(node) >> 3] & (1 << ((node) & 7)))
e52fc539 735
f9ef50a7 736#ifndef DOINIT
ded4dd2a 737EXTCONST U8 PL_varies[] __attribute__deprecated__;
f9ef50a7 738#else
ded4dd2a 739EXTCONST U8 PL_varies[] __attribute__deprecated__ = {
62e6ef33
AC
740 CLUMP, BRANCH, STAR, PLUS, CURLY, CURLYN, CURLYM, CURLYX, WHILEM, REF,
741 REFF, REFFL, REFFU, REFFA, NREF, NREFF, NREFFL, NREFFU, NREFFA,
d3d47aac 742 BRANCHJ, SUSPEND, IFTHEN,
f9ef50a7
NC
743 0
744};
745#endif /* DOINIT */
746
ded4dd2a
NC
747#ifndef DOINIT
748EXTCONST U8 PL_varies_bitmask[];
749#else
750EXTCONST U8 PL_varies_bitmask[] = {
67a1b5f9 751 0x00, 0x00, 0x00, 0x00, 0x03, 0xE0, 0x0F, 0xFF, 0xCB, 0x00, 0x00, 0x00, 0x00
ded4dd2a
NC
752};
753#endif /* DOINIT */
754
f9ef50a7
NC
755/* The following always have a length of 1. U8 we can do strchr() on it. */
756/* (Note that length 1 means "one character" under UTF8, not "one octet".) */
ded4dd2a 757#define REGNODE_SIMPLE(node) (PL_simple_bitmask[(node) >> 3] & (1 << ((node) & 7)))
e52fc539 758
f9ef50a7 759#ifndef DOINIT
ded4dd2a 760EXTCONST U8 PL_simple[] __attribute__deprecated__;
f9ef50a7 761#else
ded4dd2a 762EXTCONST U8 PL_simple[] __attribute__deprecated__ = {
67a1b5f9
KW
763 REG_ANY, SANY, ANYOF, ANYOFD, ANYOFL, ANYOFM, POSIXD, POSIXL, POSIXU,
764 POSIXA, NPOSIXD, NPOSIXL, NPOSIXU, NPOSIXA, ASCII, NASCII,
f9ef50a7
NC
765 0
766};
767#endif /* DOINIT */
768
ded4dd2a
NC
769#ifndef DOINIT
770EXTCONST U8 PL_simple_bitmask[];
771#else
772EXTCONST U8 PL_simple_bitmask[] = {
67a1b5f9 773 0x00, 0x00, 0xFF, 0xFF, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
ded4dd2a
NC
774};
775#endif /* DOINIT */
776
37442d52 777/* ex: set ro: */