Mercurial > repo
comparison perl-5.22.2/regnodes.h @ 8045:a16537d2fe07
<xfix> tar xf perl-5.22.2.tar.gz # Ah, whatever, I\'m doing it anyway
author | HackBot |
---|---|
date | Sat, 14 May 2016 14:54:38 +0000 |
parents | |
children |
comparison
equal
deleted
inserted
replaced
8044:711c038a7dce | 8045:a16537d2fe07 |
---|---|
1 /* -*- buffer-read-only: t -*- | |
2 !!!!!!! DO NOT EDIT THIS FILE !!!!!!! | |
3 This file is built by regen/regcomp.pl from regcomp.sym. | |
4 Any changes made here will be lost! | |
5 */ | |
6 | |
7 /* Regops and State definitions */ | |
8 | |
9 #define REGNODE_MAX 93 | |
10 #define REGMATCH_STATE_MAX 133 | |
11 | |
12 #define END 0 /* 0000 End of program. */ | |
13 #define SUCCEED 1 /* 0x01 Return from a subroutine, basically. */ | |
14 #define SBOL 2 /* 0x02 Match "" at beginning of line: /^/, /\A/ */ | |
15 #define BOL 2 /* 0x02 type alias */ | |
16 #define MBOL 3 /* 0x03 Same, assuming multiline: /^/m */ | |
17 #define SEOL 4 /* 0x04 Match "" at end of line: /$/ */ | |
18 #define EOL 4 /* 0x04 type alias */ | |
19 #define MEOL 5 /* 0x05 Same, assuming multiline: /$/m */ | |
20 #define EOS 6 /* 0x06 Match "" at end of string: /\z/ */ | |
21 #define GPOS 7 /* 0x07 Matches where last m//g left off. */ | |
22 #define BOUND 8 /* 0x08 Like BOUNDA for non-utf8, otherwise 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 */ | |
24 #define BOUNDU 10 /* 0x0a Match "" at any boundary of a given type using Unicode rules */ | |
25 #define BOUNDA 11 /* 0x0b Match "" at any boundary between \w\W or \W\w, where \w is [_a-zA-Z0-9] */ | |
26 #define NBOUND 12 /* 0x0c Like NBOUNDA for non-utf8, otherwise 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] */ | |
30 #define REG_ANY 16 /* 0x10 Match any one character (except newline). */ | |
31 #define SANY 17 /* 0x11 Match any one character. */ | |
32 #define CANY 18 /* 0x12 Match any one byte. */ | |
33 #define ANYOF 19 /* 0x13 Match character in (or not in) this class, single char match only */ | |
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. */ | |
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 */ | |
96 #define OPFAIL 82 /* 0x52 Same as (?!) */ | |
97 #define ACCEPT 83 /* 0x53 Accepts the current matched string. */ | |
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. */ | |
108 /* ------------ States ------------- */ | |
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 */ | |
141 #define MARKPOINT_next (REGNODE_MAX + 33) /* state for MARKPOINT */ | |
142 #define MARKPOINT_next_fail (REGNODE_MAX + 34) /* state for MARKPOINT */ | |
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 */ | |
147 #define KEEPS_next (REGNODE_MAX + 39) /* state for KEEPS */ | |
148 #define KEEPS_next_fail (REGNODE_MAX + 40) /* state for KEEPS */ | |
149 | |
150 /* PL_regkind[] What type of regop or state is this. */ | |
151 | |
152 #ifndef DOINIT | |
153 EXTCONST U8 PL_regkind[]; | |
154 #else | |
155 EXTCONST U8 PL_regkind[] = { | |
156 END, /* END */ | |
157 END, /* SUCCEED */ | |
158 BOL, /* SBOL */ | |
159 BOL, /* MBOL */ | |
160 EOL, /* SEOL */ | |
161 EOL, /* MEOL */ | |
162 EOL, /* EOS */ | |
163 GPOS, /* GPOS */ | |
164 BOUND, /* BOUND */ | |
165 BOUND, /* BOUNDL */ | |
166 BOUND, /* BOUNDU */ | |
167 BOUND, /* BOUNDA */ | |
168 NBOUND, /* NBOUND */ | |
169 NBOUND, /* NBOUNDL */ | |
170 NBOUND, /* NBOUNDU */ | |
171 NBOUND, /* NBOUNDA */ | |
172 REG_ANY, /* REG_ANY */ | |
173 REG_ANY, /* SANY */ | |
174 REG_ANY, /* CANY */ | |
175 ANYOF, /* ANYOF */ | |
176 ANYOF, /* ANYOFL */ | |
177 POSIXD, /* POSIXD */ | |
178 POSIXD, /* POSIXL */ | |
179 POSIXD, /* POSIXU */ | |
180 POSIXD, /* POSIXA */ | |
181 NPOSIXD, /* NPOSIXD */ | |
182 NPOSIXD, /* NPOSIXL */ | |
183 NPOSIXD, /* NPOSIXU */ | |
184 NPOSIXD, /* NPOSIXA */ | |
185 CLUMP, /* CLUMP */ | |
186 BRANCH, /* BRANCH */ | |
187 EXACT, /* EXACT */ | |
188 EXACT, /* EXACTL */ | |
189 EXACT, /* EXACTF */ | |
190 EXACT, /* EXACTFL */ | |
191 EXACT, /* EXACTFU */ | |
192 EXACT, /* EXACTFA */ | |
193 EXACT, /* EXACTFU_SS */ | |
194 EXACT, /* EXACTFLU8 */ | |
195 EXACT, /* EXACTFA_NO_TRIE */ | |
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 */ | |
210 REF, /* REFFU */ | |
211 REF, /* REFFA */ | |
212 REF, /* NREF */ | |
213 REF, /* NREFF */ | |
214 REF, /* NREFFL */ | |
215 REF, /* NREFFU */ | |
216 REF, /* NREFFA */ | |
217 LONGJMP, /* LONGJMP */ | |
218 BRANCHJ, /* BRANCHJ */ | |
219 BRANCHJ, /* IFMATCH */ | |
220 BRANCHJ, /* UNLESSM */ | |
221 BRANCHJ, /* SUSPEND */ | |
222 BRANCHJ, /* IFTHEN */ | |
223 GROUPP, /* GROUPP */ | |
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 */ | |
234 NGROUPP, /* NGROUPP */ | |
235 INSUBP, /* INSUBP */ | |
236 DEFINEP, /* DEFINEP */ | |
237 ENDLIKE, /* ENDLIKE */ | |
238 ENDLIKE, /* OPFAIL */ | |
239 ENDLIKE, /* ACCEPT */ | |
240 VERB, /* VERB */ | |
241 VERB, /* PRUNE */ | |
242 VERB, /* MARKPOINT */ | |
243 VERB, /* SKIP */ | |
244 VERB, /* COMMIT */ | |
245 VERB, /* CUTGROUP */ | |
246 KEEPS, /* KEEPS */ | |
247 LNBREAK, /* LNBREAK */ | |
248 NOTHING, /* OPTIMIZED */ | |
249 PSEUDO, /* PSEUDO */ | |
250 /* ------------ States ------------- */ | |
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 */ | |
285 SKIP, /* SKIP_next */ | |
286 SKIP, /* SKIP_next_fail */ | |
287 CUTGROUP, /* CUTGROUP_next */ | |
288 CUTGROUP, /* CUTGROUP_next_fail */ | |
289 KEEPS, /* KEEPS_next */ | |
290 KEEPS, /* KEEPS_next_fail */ | |
291 }; | |
292 #endif | |
293 | |
294 /* regarglen[] - How large is the argument part of the node (in regnodes) */ | |
295 | |
296 #ifdef REG_COMP_C | |
297 static const U8 regarglen[] = { | |
298 0, /* END */ | |
299 0, /* SUCCEED */ | |
300 0, /* SBOL */ | |
301 0, /* MBOL */ | |
302 0, /* SEOL */ | |
303 0, /* MEOL */ | |
304 0, /* EOS */ | |
305 0, /* GPOS */ | |
306 0, /* BOUND */ | |
307 0, /* BOUNDL */ | |
308 0, /* BOUNDU */ | |
309 0, /* BOUNDA */ | |
310 0, /* NBOUND */ | |
311 0, /* NBOUNDL */ | |
312 0, /* NBOUNDU */ | |
313 0, /* NBOUNDA */ | |
314 0, /* REG_ANY */ | |
315 0, /* SANY */ | |
316 0, /* CANY */ | |
317 EXTRA_SIZE(struct regnode_1), /* ANYOF */ | |
318 EXTRA_SIZE(struct regnode_1), /* ANYOFL */ | |
319 0, /* POSIXD */ | |
320 0, /* POSIXL */ | |
321 0, /* POSIXU */ | |
322 0, /* POSIXA */ | |
323 0, /* NPOSIXD */ | |
324 0, /* NPOSIXL */ | |
325 0, /* NPOSIXU */ | |
326 0, /* NPOSIXA */ | |
327 0, /* CLUMP */ | |
328 0, /* BRANCH */ | |
329 0, /* EXACT */ | |
330 0, /* EXACTL */ | |
331 0, /* EXACTF */ | |
332 0, /* EXACTFL */ | |
333 0, /* EXACTFU */ | |
334 0, /* EXACTFA */ | |
335 0, /* EXACTFU_SS */ | |
336 0, /* EXACTFLU8 */ | |
337 0, /* EXACTFA_NO_TRIE */ | |
338 0, /* NOTHING */ | |
339 0, /* TAIL */ | |
340 0, /* STAR */ | |
341 0, /* PLUS */ | |
342 EXTRA_SIZE(struct regnode_2), /* CURLY */ | |
343 EXTRA_SIZE(struct regnode_2), /* CURLYN */ | |
344 EXTRA_SIZE(struct regnode_2), /* CURLYM */ | |
345 EXTRA_SIZE(struct regnode_2), /* CURLYX */ | |
346 0, /* WHILEM */ | |
347 EXTRA_SIZE(struct regnode_1), /* OPEN */ | |
348 EXTRA_SIZE(struct regnode_1), /* CLOSE */ | |
349 EXTRA_SIZE(struct regnode_1), /* REF */ | |
350 EXTRA_SIZE(struct regnode_1), /* REFF */ | |
351 EXTRA_SIZE(struct regnode_1), /* REFFL */ | |
352 EXTRA_SIZE(struct regnode_1), /* REFFU */ | |
353 EXTRA_SIZE(struct regnode_1), /* REFFA */ | |
354 EXTRA_SIZE(struct regnode_1), /* NREF */ | |
355 EXTRA_SIZE(struct regnode_1), /* NREFF */ | |
356 EXTRA_SIZE(struct regnode_1), /* NREFFL */ | |
357 EXTRA_SIZE(struct regnode_1), /* NREFFU */ | |
358 EXTRA_SIZE(struct regnode_1), /* NREFFA */ | |
359 EXTRA_SIZE(struct regnode_1), /* LONGJMP */ | |
360 EXTRA_SIZE(struct regnode_1), /* BRANCHJ */ | |
361 EXTRA_SIZE(struct regnode_1), /* IFMATCH */ | |
362 EXTRA_SIZE(struct regnode_1), /* UNLESSM */ | |
363 EXTRA_SIZE(struct regnode_1), /* SUSPEND */ | |
364 EXTRA_SIZE(struct regnode_1), /* IFTHEN */ | |
365 EXTRA_SIZE(struct regnode_1), /* GROUPP */ | |
366 EXTRA_SIZE(struct regnode_2L), /* EVAL */ | |
367 0, /* MINMOD */ | |
368 0, /* LOGICAL */ | |
369 EXTRA_SIZE(struct regnode_1), /* RENUM */ | |
370 EXTRA_SIZE(struct regnode_1), /* TRIE */ | |
371 EXTRA_SIZE(struct regnode_charclass), /* TRIEC */ | |
372 EXTRA_SIZE(struct regnode_1), /* AHOCORASICK */ | |
373 EXTRA_SIZE(struct regnode_charclass), /* AHOCORASICKC */ | |
374 EXTRA_SIZE(struct regnode_2L), /* GOSUB */ | |
375 0, /* GOSTART */ | |
376 EXTRA_SIZE(struct regnode_1), /* NGROUPP */ | |
377 EXTRA_SIZE(struct regnode_1), /* INSUBP */ | |
378 EXTRA_SIZE(struct regnode_1), /* DEFINEP */ | |
379 0, /* ENDLIKE */ | |
380 0, /* OPFAIL */ | |
381 EXTRA_SIZE(struct regnode_1), /* ACCEPT */ | |
382 EXTRA_SIZE(struct regnode_1), /* VERB */ | |
383 EXTRA_SIZE(struct regnode_1), /* PRUNE */ | |
384 EXTRA_SIZE(struct regnode_1), /* MARKPOINT */ | |
385 EXTRA_SIZE(struct regnode_1), /* SKIP */ | |
386 EXTRA_SIZE(struct regnode_1), /* COMMIT */ | |
387 EXTRA_SIZE(struct regnode_1), /* CUTGROUP */ | |
388 0, /* KEEPS */ | |
389 0, /* LNBREAK */ | |
390 0, /* OPTIMIZED */ | |
391 0, /* PSEUDO */ | |
392 }; | |
393 | |
394 /* reg_off_by_arg[] - Which argument holds the offset to the next node */ | |
395 | |
396 static const char reg_off_by_arg[] = { | |
397 0, /* END */ | |
398 0, /* SUCCEED */ | |
399 0, /* SBOL */ | |
400 0, /* MBOL */ | |
401 0, /* SEOL */ | |
402 0, /* MEOL */ | |
403 0, /* EOS */ | |
404 0, /* GPOS */ | |
405 0, /* BOUND */ | |
406 0, /* BOUNDL */ | |
407 0, /* BOUNDU */ | |
408 0, /* BOUNDA */ | |
409 0, /* NBOUND */ | |
410 0, /* NBOUNDL */ | |
411 0, /* NBOUNDU */ | |
412 0, /* NBOUNDA */ | |
413 0, /* REG_ANY */ | |
414 0, /* SANY */ | |
415 0, /* CANY */ | |
416 0, /* ANYOF */ | |
417 0, /* ANYOFL */ | |
418 0, /* POSIXD */ | |
419 0, /* POSIXL */ | |
420 0, /* POSIXU */ | |
421 0, /* POSIXA */ | |
422 0, /* NPOSIXD */ | |
423 0, /* NPOSIXL */ | |
424 0, /* NPOSIXU */ | |
425 0, /* NPOSIXA */ | |
426 0, /* CLUMP */ | |
427 0, /* BRANCH */ | |
428 0, /* EXACT */ | |
429 0, /* EXACTL */ | |
430 0, /* EXACTF */ | |
431 0, /* EXACTFL */ | |
432 0, /* EXACTFU */ | |
433 0, /* EXACTFA */ | |
434 0, /* EXACTFU_SS */ | |
435 0, /* EXACTFLU8 */ | |
436 0, /* EXACTFA_NO_TRIE */ | |
437 0, /* NOTHING */ | |
438 0, /* TAIL */ | |
439 0, /* STAR */ | |
440 0, /* PLUS */ | |
441 0, /* CURLY */ | |
442 0, /* CURLYN */ | |
443 0, /* CURLYM */ | |
444 0, /* CURLYX */ | |
445 0, /* WHILEM */ | |
446 0, /* OPEN */ | |
447 0, /* CLOSE */ | |
448 0, /* REF */ | |
449 0, /* REFF */ | |
450 0, /* REFFL */ | |
451 0, /* REFFU */ | |
452 0, /* REFFA */ | |
453 0, /* NREF */ | |
454 0, /* NREFF */ | |
455 0, /* NREFFL */ | |
456 0, /* NREFFU */ | |
457 0, /* NREFFA */ | |
458 1, /* LONGJMP */ | |
459 1, /* BRANCHJ */ | |
460 2, /* IFMATCH */ | |
461 2, /* UNLESSM */ | |
462 1, /* SUSPEND */ | |
463 1, /* IFTHEN */ | |
464 0, /* GROUPP */ | |
465 0, /* EVAL */ | |
466 0, /* MINMOD */ | |
467 0, /* LOGICAL */ | |
468 1, /* RENUM */ | |
469 0, /* TRIE */ | |
470 0, /* TRIEC */ | |
471 0, /* AHOCORASICK */ | |
472 0, /* AHOCORASICKC */ | |
473 0, /* GOSUB */ | |
474 0, /* GOSTART */ | |
475 0, /* NGROUPP */ | |
476 0, /* INSUBP */ | |
477 0, /* DEFINEP */ | |
478 0, /* ENDLIKE */ | |
479 0, /* OPFAIL */ | |
480 0, /* ACCEPT */ | |
481 0, /* VERB */ | |
482 0, /* PRUNE */ | |
483 0, /* MARKPOINT */ | |
484 0, /* SKIP */ | |
485 0, /* COMMIT */ | |
486 0, /* CUTGROUP */ | |
487 0, /* KEEPS */ | |
488 0, /* LNBREAK */ | |
489 0, /* OPTIMIZED */ | |
490 0, /* PSEUDO */ | |
491 }; | |
492 | |
493 #endif /* REG_COMP_C */ | |
494 | |
495 /* reg_name[] - Opcode/state names in string form, for debugging */ | |
496 | |
497 #ifndef DOINIT | |
498 EXTCONST char * PL_reg_name[]; | |
499 #else | |
500 EXTCONST char * const PL_reg_name[] = { | |
501 "END", /* 0000 */ | |
502 "SUCCEED", /* 0x01 */ | |
503 "SBOL", /* 0x02 */ | |
504 "MBOL", /* 0x03 */ | |
505 "SEOL", /* 0x04 */ | |
506 "MEOL", /* 0x05 */ | |
507 "EOS", /* 0x06 */ | |
508 "GPOS", /* 0x07 */ | |
509 "BOUND", /* 0x08 */ | |
510 "BOUNDL", /* 0x09 */ | |
511 "BOUNDU", /* 0x0a */ | |
512 "BOUNDA", /* 0x0b */ | |
513 "NBOUND", /* 0x0c */ | |
514 "NBOUNDL", /* 0x0d */ | |
515 "NBOUNDU", /* 0x0e */ | |
516 "NBOUNDA", /* 0x0f */ | |
517 "REG_ANY", /* 0x10 */ | |
518 "SANY", /* 0x11 */ | |
519 "CANY", /* 0x12 */ | |
520 "ANYOF", /* 0x13 */ | |
521 "ANYOFL", /* 0x14 */ | |
522 "POSIXD", /* 0x15 */ | |
523 "POSIXL", /* 0x16 */ | |
524 "POSIXU", /* 0x17 */ | |
525 "POSIXA", /* 0x18 */ | |
526 "NPOSIXD", /* 0x19 */ | |
527 "NPOSIXL", /* 0x1a */ | |
528 "NPOSIXU", /* 0x1b */ | |
529 "NPOSIXA", /* 0x1c */ | |
530 "CLUMP", /* 0x1d */ | |
531 "BRANCH", /* 0x1e */ | |
532 "EXACT", /* 0x1f */ | |
533 "EXACTL", /* 0x20 */ | |
534 "EXACTF", /* 0x21 */ | |
535 "EXACTFL", /* 0x22 */ | |
536 "EXACTFU", /* 0x23 */ | |
537 "EXACTFA", /* 0x24 */ | |
538 "EXACTFU_SS", /* 0x25 */ | |
539 "EXACTFLU8", /* 0x26 */ | |
540 "EXACTFA_NO_TRIE", /* 0x27 */ | |
541 "NOTHING", /* 0x28 */ | |
542 "TAIL", /* 0x29 */ | |
543 "STAR", /* 0x2a */ | |
544 "PLUS", /* 0x2b */ | |
545 "CURLY", /* 0x2c */ | |
546 "CURLYN", /* 0x2d */ | |
547 "CURLYM", /* 0x2e */ | |
548 "CURLYX", /* 0x2f */ | |
549 "WHILEM", /* 0x30 */ | |
550 "OPEN", /* 0x31 */ | |
551 "CLOSE", /* 0x32 */ | |
552 "REF", /* 0x33 */ | |
553 "REFF", /* 0x34 */ | |
554 "REFFL", /* 0x35 */ | |
555 "REFFU", /* 0x36 */ | |
556 "REFFA", /* 0x37 */ | |
557 "NREF", /* 0x38 */ | |
558 "NREFF", /* 0x39 */ | |
559 "NREFFL", /* 0x3a */ | |
560 "NREFFU", /* 0x3b */ | |
561 "NREFFA", /* 0x3c */ | |
562 "LONGJMP", /* 0x3d */ | |
563 "BRANCHJ", /* 0x3e */ | |
564 "IFMATCH", /* 0x3f */ | |
565 "UNLESSM", /* 0x40 */ | |
566 "SUSPEND", /* 0x41 */ | |
567 "IFTHEN", /* 0x42 */ | |
568 "GROUPP", /* 0x43 */ | |
569 "EVAL", /* 0x44 */ | |
570 "MINMOD", /* 0x45 */ | |
571 "LOGICAL", /* 0x46 */ | |
572 "RENUM", /* 0x47 */ | |
573 "TRIE", /* 0x48 */ | |
574 "TRIEC", /* 0x49 */ | |
575 "AHOCORASICK", /* 0x4a */ | |
576 "AHOCORASICKC", /* 0x4b */ | |
577 "GOSUB", /* 0x4c */ | |
578 "GOSTART", /* 0x4d */ | |
579 "NGROUPP", /* 0x4e */ | |
580 "INSUBP", /* 0x4f */ | |
581 "DEFINEP", /* 0x50 */ | |
582 "ENDLIKE", /* 0x51 */ | |
583 "OPFAIL", /* 0x52 */ | |
584 "ACCEPT", /* 0x53 */ | |
585 "VERB", /* 0x54 */ | |
586 "PRUNE", /* 0x55 */ | |
587 "MARKPOINT", /* 0x56 */ | |
588 "SKIP", /* 0x57 */ | |
589 "COMMIT", /* 0x58 */ | |
590 "CUTGROUP", /* 0x59 */ | |
591 "KEEPS", /* 0x5a */ | |
592 "LNBREAK", /* 0x5b */ | |
593 "OPTIMIZED", /* 0x5c */ | |
594 "PSEUDO", /* 0x5d */ | |
595 /* ------------ States ------------- */ | |
596 "TRIE_next", /* REGNODE_MAX +0x01 */ | |
597 "TRIE_next_fail", /* REGNODE_MAX +0x02 */ | |
598 "EVAL_AB", /* REGNODE_MAX +0x03 */ | |
599 "EVAL_AB_fail", /* REGNODE_MAX +0x04 */ | |
600 "CURLYX_end", /* REGNODE_MAX +0x05 */ | |
601 "CURLYX_end_fail", /* REGNODE_MAX +0x06 */ | |
602 "WHILEM_A_pre", /* REGNODE_MAX +0x07 */ | |
603 "WHILEM_A_pre_fail", /* REGNODE_MAX +0x08 */ | |
604 "WHILEM_A_min", /* REGNODE_MAX +0x09 */ | |
605 "WHILEM_A_min_fail", /* REGNODE_MAX +0x0a */ | |
606 "WHILEM_A_max", /* REGNODE_MAX +0x0b */ | |
607 "WHILEM_A_max_fail", /* REGNODE_MAX +0x0c */ | |
608 "WHILEM_B_min", /* REGNODE_MAX +0x0d */ | |
609 "WHILEM_B_min_fail", /* REGNODE_MAX +0x0e */ | |
610 "WHILEM_B_max", /* REGNODE_MAX +0x0f */ | |
611 "WHILEM_B_max_fail", /* REGNODE_MAX +0x10 */ | |
612 "BRANCH_next", /* REGNODE_MAX +0x11 */ | |
613 "BRANCH_next_fail", /* REGNODE_MAX +0x12 */ | |
614 "CURLYM_A", /* REGNODE_MAX +0x13 */ | |
615 "CURLYM_A_fail", /* REGNODE_MAX +0x14 */ | |
616 "CURLYM_B", /* REGNODE_MAX +0x15 */ | |
617 "CURLYM_B_fail", /* REGNODE_MAX +0x16 */ | |
618 "IFMATCH_A", /* REGNODE_MAX +0x17 */ | |
619 "IFMATCH_A_fail", /* REGNODE_MAX +0x18 */ | |
620 "CURLY_B_min_known", /* REGNODE_MAX +0x19 */ | |
621 "CURLY_B_min_known_fail", /* REGNODE_MAX +0x1a */ | |
622 "CURLY_B_min", /* REGNODE_MAX +0x1b */ | |
623 "CURLY_B_min_fail", /* REGNODE_MAX +0x1c */ | |
624 "CURLY_B_max", /* REGNODE_MAX +0x1d */ | |
625 "CURLY_B_max_fail", /* REGNODE_MAX +0x1e */ | |
626 "COMMIT_next", /* REGNODE_MAX +0x1f */ | |
627 "COMMIT_next_fail", /* REGNODE_MAX +0x20 */ | |
628 "MARKPOINT_next", /* REGNODE_MAX +0x21 */ | |
629 "MARKPOINT_next_fail", /* REGNODE_MAX +0x22 */ | |
630 "SKIP_next", /* REGNODE_MAX +0x23 */ | |
631 "SKIP_next_fail", /* REGNODE_MAX +0x24 */ | |
632 "CUTGROUP_next", /* REGNODE_MAX +0x25 */ | |
633 "CUTGROUP_next_fail", /* REGNODE_MAX +0x26 */ | |
634 "KEEPS_next", /* REGNODE_MAX +0x27 */ | |
635 "KEEPS_next_fail", /* REGNODE_MAX +0x28 */ | |
636 }; | |
637 #endif /* DOINIT */ | |
638 | |
639 /* PL_reg_extflags_name[] - Opcode/state names in string form, for debugging */ | |
640 | |
641 #ifndef DOINIT | |
642 EXTCONST char * PL_reg_extflags_name[]; | |
643 #else | |
644 EXTCONST char * const PL_reg_extflags_name[] = { | |
645 /* Bits in extflags defined: 11111111111111110000111111111111 */ | |
646 "MULTILINE", /* 0x00000001 */ | |
647 "SINGLELINE", /* 0x00000002 */ | |
648 "FOLD", /* 0x00000004 */ | |
649 "EXTENDED", /* 0x00000008 */ | |
650 "EXTENDED_MORE", /* 0x00000010 */ | |
651 "NOCAPTURE", /* 0x00000020 */ | |
652 "KEEPCOPY", /* 0x00000040 */ | |
653 "CHARSET0", /* 0x00000080 : "CHARSET" - 0x00000380 */ | |
654 "CHARSET1", /* 0x00000100 : "CHARSET" - 0x00000380 */ | |
655 "CHARSET2", /* 0x00000200 : "CHARSET" - 0x00000380 */ | |
656 "STRICT", /* 0x00000400 */ | |
657 "SPLIT", /* 0x00000800 */ | |
658 "UNUSED_BIT_12", /* 0x00001000 */ | |
659 "UNUSED_BIT_13", /* 0x00002000 */ | |
660 "UNUSED_BIT_14", /* 0x00004000 */ | |
661 "UNUSED_BIT_15", /* 0x00008000 */ | |
662 "NO_INPLACE_SUBST", /* 0x00010000 */ | |
663 "EVAL_SEEN", /* 0x00020000 */ | |
664 "UNBOUNDED_QUANTIFIER_SEEN",/* 0x00040000 */ | |
665 "CHECK_ALL", /* 0x00080000 */ | |
666 "MATCH_UTF8", /* 0x00100000 */ | |
667 "USE_INTUIT_NOML", /* 0x00200000 */ | |
668 "USE_INTUIT_ML", /* 0x00400000 */ | |
669 "INTUIT_TAIL", /* 0x00800000 */ | |
670 "IS_ANCHORED", /* 0x01000000 */ | |
671 "COPY_DONE", /* 0x02000000 */ | |
672 "TAINTED_SEEN", /* 0x04000000 */ | |
673 "TAINTED", /* 0x08000000 */ | |
674 "START_ONLY", /* 0x10000000 */ | |
675 "SKIPWHITE", /* 0x20000000 */ | |
676 "WHITE", /* 0x40000000 */ | |
677 "NULL", /* 0x80000000 */ | |
678 }; | |
679 #endif /* DOINIT */ | |
680 | |
681 #ifdef DEBUGGING | |
682 # define REG_EXTFLAGS_NAME_SIZE 32 | |
683 #endif | |
684 | |
685 /* PL_reg_intflags_name[] - Opcode/state names in string form, for debugging */ | |
686 | |
687 #ifndef DOINIT | |
688 EXTCONST char * PL_reg_intflags_name[]; | |
689 #else | |
690 EXTCONST char * const PL_reg_intflags_name[] = { | |
691 "SKIP", /* 0x00000001 - PREGf_SKIP */ | |
692 "IMPLICIT", /* 0x00000002 - PREGf_IMPLICIT - Converted .* to ^.* */ | |
693 "NAUGHTY", /* 0x00000004 - PREGf_NAUGHTY - how exponential is this pattern? */ | |
694 "VERBARG_SEEN", /* 0x00000008 - PREGf_VERBARG_SEEN */ | |
695 "CUTGROUP_SEEN", /* 0x00000010 - PREGf_CUTGROUP_SEEN */ | |
696 "USE_RE_EVAL", /* 0x00000020 - PREGf_USE_RE_EVAL - compiled with "use re 'eval'" */ | |
697 "NOSCAN", /* 0x00000040 - PREGf_NOSCAN */ | |
698 "CANY_SEEN", /* 0x00000080 - PREGf_CANY_SEEN */ | |
699 "GPOS_SEEN", /* 0x00000100 - PREGf_GPOS_SEEN */ | |
700 "GPOS_FLOAT", /* 0x00000200 - PREGf_GPOS_FLOAT */ | |
701 "ANCH_MBOL", /* 0x00000400 - PREGf_ANCH_MBOL */ | |
702 "ANCH_SBOL", /* 0x00000800 - PREGf_ANCH_SBOL */ | |
703 "ANCH_GPOS", /* 0x00001000 - PREGf_ANCH_GPOS */ | |
704 }; | |
705 #endif /* DOINIT */ | |
706 | |
707 #ifdef DEBUGGING | |
708 # define REG_INTFLAGS_NAME_SIZE 13 | |
709 #endif | |
710 | |
711 /* The following have no fixed length. U8 so we can do strchr() on it. */ | |
712 #define REGNODE_VARIES(node) (PL_varies_bitmask[(node) >> 3] & (1 << ((node) & 7))) | |
713 | |
714 #ifndef DOINIT | |
715 EXTCONST U8 PL_varies[] __attribute__deprecated__; | |
716 #else | |
717 EXTCONST U8 PL_varies[] __attribute__deprecated__ = { | |
718 CLUMP, BRANCH, STAR, PLUS, CURLY, CURLYN, CURLYM, CURLYX, WHILEM, REF, | |
719 REFF, REFFL, REFFU, REFFA, NREF, NREFF, NREFFL, NREFFU, NREFFA, | |
720 BRANCHJ, SUSPEND, IFTHEN, | |
721 0 | |
722 }; | |
723 #endif /* DOINIT */ | |
724 | |
725 #ifndef DOINIT | |
726 EXTCONST U8 PL_varies_bitmask[]; | |
727 #else | |
728 EXTCONST U8 PL_varies_bitmask[] = { | |
729 0x00, 0x00, 0x00, 0x60, 0x00, 0xFC, 0xF9, 0x5F, 0x06, 0x00, 0x00, 0x00 | |
730 }; | |
731 #endif /* DOINIT */ | |
732 | |
733 /* The following always have a length of 1. U8 we can do strchr() on it. */ | |
734 /* (Note that length 1 means "one character" under UTF8, not "one octet".) */ | |
735 #define REGNODE_SIMPLE(node) (PL_simple_bitmask[(node) >> 3] & (1 << ((node) & 7))) | |
736 | |
737 #ifndef DOINIT | |
738 EXTCONST U8 PL_simple[] __attribute__deprecated__; | |
739 #else | |
740 EXTCONST U8 PL_simple[] __attribute__deprecated__ = { | |
741 REG_ANY, SANY, CANY, ANYOF, ANYOFL, POSIXD, POSIXL, POSIXU, POSIXA, | |
742 NPOSIXD, NPOSIXL, NPOSIXU, NPOSIXA, | |
743 0 | |
744 }; | |
745 #endif /* DOINIT */ | |
746 | |
747 #ifndef DOINIT | |
748 EXTCONST U8 PL_simple_bitmask[]; | |
749 #else | |
750 EXTCONST U8 PL_simple_bitmask[] = { | |
751 0x00, 0x00, 0xFF, 0x1F, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00 | |
752 }; | |
753 #endif /* DOINIT */ | |
754 | |
755 /* ex: set ro: */ |