Commit | Line | Data |
a0d0e21e |
1 | /* regcomp.h |
d6376244 |
2 | * |
4bb101f2 |
3 | * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999, |
1d325971 |
4 | * 2000, 2001, 2002, 2003, 2005 by Larry Wall and others |
d6376244 |
5 | * |
6 | * You may distribute under the terms of either the GNU General Public |
7 | * License or the Artistic License, as specified in the README file. |
8 | * |
a687059c |
9 | */ |
10 | |
c277df42 |
11 | typedef OP OP_4tree; /* Will be redefined later. */ |
12 | |
a687059c |
13 | /* |
14 | * The "internal use only" fields in regexp.h are present to pass info from |
15 | * compile to execute that permits the execute phase to run lots faster on |
16 | * simple cases. They are: |
17 | * |
79072805 |
18 | * regstart sv that must begin a match; Nullch if none obvious |
a687059c |
19 | * reganch is the match anchored (at beginning-of-line only)? |
20 | * regmust string (pointer into program) that match must include, or NULL |
79072805 |
21 | * [regmust changed to SV* for bminstr()--law] |
a687059c |
22 | * regmlen length of regmust string |
23 | * [regmlen not used currently] |
24 | * |
25 | * Regstart and reganch permit very fast decisions on suitable starting points |
26 | * for a match, cutting down the work a lot. Regmust permits fast rejection |
27 | * of lines that cannot possibly match. The regmust tests are costly enough |
e50aee73 |
28 | * that pregcomp() supplies a regmust only if the r.e. contains something |
a687059c |
29 | * potentially expensive (at present, the only such thing detected is * or + |
30 | * at the start of the r.e., which can involve a lot of backup). Regmlen is |
e50aee73 |
31 | * supplied because the test in pregexec() needs it and pregcomp() is computing |
a687059c |
32 | * it anyway. |
33 | * [regmust is now supplied always. The tests that use regmust have a |
34 | * heuristic that disables the test if it usually matches.] |
35 | * |
36 | * [In fact, we now use regmust in many cases to locate where the search |
37 | * starts in the string, so if regback is >= 0, the regmust search is never |
38 | * wasted effort. The regback variable says how many characters back from |
39 | * where regmust matched is the earliest possible start of the match. |
40 | * For instance, /[a-z].foo/ has a regmust of 'foo' and a regback of 2.] |
41 | */ |
42 | |
43 | /* |
44 | * Structure for regexp "program". This is essentially a linear encoding |
45 | * of a nondeterministic finite-state machine (aka syntax charts or |
46 | * "railroad normal form" in parsing technology). Each node is an opcode |
47 | * plus a "next" pointer, possibly plus an operand. "Next" pointers of |
48 | * all nodes except BRANCH implement concatenation; a "next" pointer with |
49 | * a BRANCH on both ends of it is connecting two alternatives. (Here we |
50 | * have one of the subtle syntax dependencies: an individual BRANCH (as |
51 | * opposed to a collection of them) is never concatenated with anything |
52 | * because of operator precedence.) The operand of some types of node is |
53 | * a literal string; for others, it is a node leading into a sub-FSM. In |
54 | * particular, the operand of a BRANCH node is the first node of the branch. |
55 | * (NB this is *not* a tree structure: the tail of the branch connects |
a3621e74 |
56 | * to the thing following the set of BRANCHes.) The opcodes are defined |
57 | * in regnodes.h which is generated from regcomp.sym by regcomp.pl. |
a687059c |
58 | */ |
59 | |
a687059c |
60 | /* |
61 | * A node is one char of opcode followed by two chars of "next" pointer. |
62 | * "Next" pointers are stored as two 8-bit pieces, high order first. The |
63 | * value is a positive offset from the opcode of the node containing it. |
64 | * An operand, if any, simply follows the node. (Note that much of the |
65 | * code generation knows about this implicit relationship.) |
66 | * |
67 | * Using two bytes for the "next" pointer is vast overkill for most things, |
68 | * but allows patterns to get big without disasters. |
69 | * |
e91177ed |
70 | * [The "next" pointer is always aligned on an even |
a687059c |
71 | * boundary, and reads the offset directly as a short. Also, there is no |
72 | * special test to reverse the sign of BACK pointers since the offset is |
73 | * stored negative.] |
74 | */ |
75 | |
c277df42 |
76 | struct regnode_string { |
cd439c50 |
77 | U8 str_len; |
c277df42 |
78 | U8 type; |
79 | U16 next_off; |
cd439c50 |
80 | char string[1]; |
c277df42 |
81 | }; |
a687059c |
82 | |
c277df42 |
83 | struct regnode_1 { |
84 | U8 flags; |
85 | U8 type; |
86 | U16 next_off; |
87 | U32 arg1; |
88 | }; |
89 | |
90 | struct regnode_2 { |
91 | U8 flags; |
92 | U8 type; |
93 | U16 next_off; |
94 | U16 arg1; |
95 | U16 arg2; |
96 | }; |
97 | |
936ed897 |
98 | #define ANYOF_BITMAP_SIZE 32 /* 256 b/(8 b/B) */ |
ffc61ed2 |
99 | #define ANYOF_CLASSBITMAP_SIZE 4 /* up to 32 (8*4) named classes */ |
936ed897 |
100 | |
101 | struct regnode_charclass { |
102 | U8 flags; |
103 | U8 type; |
104 | U16 next_off; |
ffc61ed2 |
105 | U32 arg1; |
19103b91 |
106 | char bitmap[ANYOF_BITMAP_SIZE]; /* only compile-time */ |
936ed897 |
107 | }; |
108 | |
19103b91 |
109 | struct regnode_charclass_class { /* has [[:blah:]] classes */ |
110 | U8 flags; /* should have ANYOF_CLASS here */ |
936ed897 |
111 | U8 type; |
112 | U16 next_off; |
ffc61ed2 |
113 | U32 arg1; |
19103b91 |
114 | char bitmap[ANYOF_BITMAP_SIZE]; /* both compile-time */ |
115 | char classflags[ANYOF_CLASSBITMAP_SIZE]; /* and run-time */ |
936ed897 |
116 | }; |
117 | |
9c7e81e8 |
118 | /* XXX fix this description. |
119 | Impose a limit of REG_INFTY on various pattern matching operations |
120 | to limit stack growth and to avoid "infinite" recursions. |
121 | */ |
122 | /* The default size for REG_INFTY is I16_MAX, which is the same as |
123 | SHORT_MAX (see perl.h). Unfortunately I16 isn't necessarily 16 bits |
124 | (see handy.h). On the Cray C90, sizeof(short)==4 and hence I16_MAX is |
125 | ((1<<31)-1), while on the Cray T90, sizeof(short)==8 and I16_MAX is |
126 | ((1<<63)-1). To limit stack growth to reasonable sizes, supply a |
127 | smaller default. |
128 | --Andy Dougherty 11 June 1998 |
129 | */ |
130 | #if SHORTSIZE > 2 |
131 | # ifndef REG_INFTY |
132 | # define REG_INFTY ((1<<15)-1) |
133 | # endif |
134 | #endif |
135 | |
136 | #ifndef REG_INFTY |
83cfe48c |
137 | # define REG_INFTY I16_MAX |
138 | #endif |
a687059c |
139 | |
e91177ed |
140 | #define ARG_VALUE(arg) (arg) |
141 | #define ARG__SET(arg,val) ((arg) = (val)) |
c277df42 |
142 | |
e8e3b0fb |
143 | #undef ARG |
144 | #undef ARG1 |
145 | #undef ARG2 |
146 | |
c277df42 |
147 | #define ARG(p) ARG_VALUE(ARG_LOC(p)) |
148 | #define ARG1(p) ARG_VALUE(ARG1_LOC(p)) |
149 | #define ARG2(p) ARG_VALUE(ARG2_LOC(p)) |
150 | #define ARG_SET(p, val) ARG__SET(ARG_LOC(p), (val)) |
151 | #define ARG1_SET(p, val) ARG__SET(ARG1_LOC(p), (val)) |
152 | #define ARG2_SET(p, val) ARG__SET(ARG2_LOC(p), (val)) |
153 | |
e8e3b0fb |
154 | #undef NEXT_OFF |
155 | #undef NODE_ALIGN |
156 | |
c277df42 |
157 | #ifndef lint |
e91177ed |
158 | # define NEXT_OFF(p) ((p)->next_off) |
159 | # define NODE_ALIGN(node) |
160 | # define NODE_ALIGN_FILL(node) ((node)->flags = 0xde) /* deadbeef */ |
a687059c |
161 | #else /* lint */ |
c277df42 |
162 | # define NEXT_OFF(p) 0 |
e91177ed |
163 | # define NODE_ALIGN(node) |
164 | # define NODE_ALIGN_FILL(node) |
a687059c |
165 | #endif /* lint */ |
166 | |
c277df42 |
167 | #define SIZE_ALIGN NODE_ALIGN |
168 | |
e8e3b0fb |
169 | #undef OP |
170 | #undef OPERAND |
171 | #undef MASK |
172 | #undef STRING |
173 | |
e91177ed |
174 | #define OP(p) ((p)->type) |
175 | #define OPERAND(p) (((struct regnode_string *)p)->string) |
cd439c50 |
176 | #define MASK(p) ((char*)OPERAND(p)) |
177 | #define STR_LEN(p) (((struct regnode_string *)p)->str_len) |
178 | #define STRING(p) (((struct regnode_string *)p)->string) |
179 | #define STR_SZ(l) ((l + sizeof(regnode) - 1) / sizeof(regnode)) |
180 | #define NODE_SZ_STR(p) (STR_SZ(STR_LEN(p))+1) |
181 | |
e8e3b0fb |
182 | #undef NODE_ALIGN |
183 | #undef ARG_LOC |
184 | #undef NEXTOPER |
185 | #undef PREVOPER |
186 | |
e91177ed |
187 | #define NODE_ALIGN(node) |
188 | #define ARG_LOC(p) (((struct regnode_1 *)p)->arg1) |
189 | #define ARG1_LOC(p) (((struct regnode_2 *)p)->arg1) |
190 | #define ARG2_LOC(p) (((struct regnode_2 *)p)->arg2) |
191 | #define NODE_STEP_REGNODE 1 /* sizeof(regnode)/sizeof(regnode) */ |
192 | #define EXTRA_STEP_2ARGS EXTRA_SIZE(struct regnode_2) |
193 | |
194 | #define NODE_STEP_B 4 |
c277df42 |
195 | |
196 | #define NEXTOPER(p) ((p) + NODE_STEP_REGNODE) |
197 | #define PREVOPER(p) ((p) - NODE_STEP_REGNODE) |
198 | |
e91177ed |
199 | #define FILL_ADVANCE_NODE(ptr, op) STMT_START { \ |
c277df42 |
200 | (ptr)->type = op; (ptr)->next_off = 0; (ptr)++; } STMT_END |
e91177ed |
201 | #define FILL_ADVANCE_NODE_ARG(ptr, op, arg) STMT_START { \ |
c277df42 |
202 | ARG_SET(ptr, arg); FILL_ADVANCE_NODE(ptr, op); (ptr) += 1; } STMT_END |
a687059c |
203 | |
22c35a8c |
204 | #define REG_MAGIC 0234 |
a687059c |
205 | |
830247a4 |
206 | #define SIZE_ONLY (RExC_emit == &PL_regdummy) |
c277df42 |
207 | |
936ed897 |
208 | /* Flags for node->flags of ANYOF */ |
b8c5462f |
209 | |
19103b91 |
210 | #define ANYOF_CLASS 0x08 /* has [[:blah:]] classes */ |
ffc61ed2 |
211 | #define ANYOF_INVERT 0x04 |
212 | #define ANYOF_FOLD 0x02 |
213 | #define ANYOF_LOCALE 0x01 |
b8c5462f |
214 | |
653099ff |
215 | /* Used for regstclass only */ |
ffc61ed2 |
216 | #define ANYOF_EOS 0x10 /* Can match an empty string too */ |
217 | |
218 | /* There is a character or a range past 0xff */ |
219 | #define ANYOF_UNICODE 0x20 |
1aa99e6b |
220 | #define ANYOF_UNICODE_ALL 0x40 /* Can match any char past 0xff */ |
ffc61ed2 |
221 | |
4f66b38d |
222 | /* size of node is large (includes class pointer) */ |
223 | #define ANYOF_LARGE 0x80 |
224 | |
ffc61ed2 |
225 | /* Are there any runtime flags on in this node? */ |
226 | #define ANYOF_RUNTIME(s) (ANYOF_FLAGS(s) & 0x0f) |
227 | |
228 | #define ANYOF_FLAGS_ALL 0xff |
653099ff |
229 | |
936ed897 |
230 | /* Character classes for node->classflags of ANYOF */ |
653099ff |
231 | /* Should be synchronized with a table in regprop() */ |
232 | /* 2n should pair with 2n+1 */ |
b8c5462f |
233 | |
289d4f09 |
234 | #define ANYOF_ALNUM 0 /* \w, PL_utf8_alnum, utf8::IsWord, ALNUM */ |
b8c5462f |
235 | #define ANYOF_NALNUM 1 |
aaa51d5e |
236 | #define ANYOF_SPACE 2 /* \s */ |
b8c5462f |
237 | #define ANYOF_NSPACE 3 |
238 | #define ANYOF_DIGIT 4 |
239 | #define ANYOF_NDIGIT 5 |
289d4f09 |
240 | #define ANYOF_ALNUMC 6 /* isalnum(3), utf8::IsAlnum, ALNUMC */ |
b8c5462f |
241 | #define ANYOF_NALNUMC 7 |
242 | #define ANYOF_ALPHA 8 |
243 | #define ANYOF_NALPHA 9 |
244 | #define ANYOF_ASCII 10 |
245 | #define ANYOF_NASCII 11 |
246 | #define ANYOF_CNTRL 12 |
247 | #define ANYOF_NCNTRL 13 |
248 | #define ANYOF_GRAPH 14 |
249 | #define ANYOF_NGRAPH 15 |
250 | #define ANYOF_LOWER 16 |
251 | #define ANYOF_NLOWER 17 |
252 | #define ANYOF_PRINT 18 |
253 | #define ANYOF_NPRINT 19 |
254 | #define ANYOF_PUNCT 20 |
255 | #define ANYOF_NPUNCT 21 |
256 | #define ANYOF_UPPER 22 |
257 | #define ANYOF_NUPPER 23 |
258 | #define ANYOF_XDIGIT 24 |
259 | #define ANYOF_NXDIGIT 25 |
aaa51d5e |
260 | #define ANYOF_PSXSPC 26 /* POSIX space: \s plus the vertical tab */ |
261 | #define ANYOF_NPSXSPC 27 |
ffc61ed2 |
262 | #define ANYOF_BLANK 28 /* GNU extension: space and tab: non-vertical space */ |
289d4f09 |
263 | #define ANYOF_NBLANK 29 |
b8c5462f |
264 | |
aaa51d5e |
265 | #define ANYOF_MAX 32 |
b8c5462f |
266 | |
267 | /* Backward source code compatibility. */ |
268 | |
269 | #define ANYOF_ALNUML ANYOF_ALNUM |
270 | #define ANYOF_NALNUML ANYOF_NALNUM |
271 | #define ANYOF_SPACEL ANYOF_SPACE |
272 | #define ANYOF_NSPACEL ANYOF_NSPACE |
273 | |
274 | /* Utility macros for the bitmap and classes of ANYOF */ |
275 | |
936ed897 |
276 | #define ANYOF_SIZE (sizeof(struct regnode_charclass)) |
277 | #define ANYOF_CLASS_SIZE (sizeof(struct regnode_charclass_class)) |
b8c5462f |
278 | |
936ed897 |
279 | #define ANYOF_FLAGS(p) ((p)->flags) |
b8c5462f |
280 | |
281 | #define ANYOF_BIT(c) (1 << ((c) & 7)) |
282 | |
936ed897 |
283 | #define ANYOF_CLASS_BYTE(p, c) (((struct regnode_charclass_class*)(p))->classflags[((c) >> 3) & 3]) |
b8c5462f |
284 | #define ANYOF_CLASS_SET(p, c) (ANYOF_CLASS_BYTE(p, c) |= ANYOF_BIT(c)) |
285 | #define ANYOF_CLASS_CLEAR(p, c) (ANYOF_CLASS_BYTE(p, c) &= ~ANYOF_BIT(c)) |
286 | #define ANYOF_CLASS_TEST(p, c) (ANYOF_CLASS_BYTE(p, c) & ANYOF_BIT(c)) |
287 | |
936ed897 |
288 | #define ANYOF_CLASS_ZERO(ret) Zero(((struct regnode_charclass_class*)(ret))->classflags, ANYOF_CLASSBITMAP_SIZE, char) |
289 | #define ANYOF_BITMAP_ZERO(ret) Zero(((struct regnode_charclass*)(ret))->bitmap, ANYOF_BITMAP_SIZE, char) |
290 | |
291 | #define ANYOF_BITMAP(p) (((struct regnode_charclass*)(p))->bitmap) |
292 | #define ANYOF_BITMAP_BYTE(p, c) (ANYOF_BITMAP(p)[((c) >> 3) & 31]) |
b8c5462f |
293 | #define ANYOF_BITMAP_SET(p, c) (ANYOF_BITMAP_BYTE(p, c) |= ANYOF_BIT(c)) |
294 | #define ANYOF_BITMAP_CLEAR(p,c) (ANYOF_BITMAP_BYTE(p, c) &= ~ANYOF_BIT(c)) |
295 | #define ANYOF_BITMAP_TEST(p, c) (ANYOF_BITMAP_BYTE(p, c) & ANYOF_BIT(c)) |
296 | |
f8bef550 |
297 | #define ANYOF_BITMAP_SETALL(p) \ |
298 | memset (ANYOF_BITMAP(p), 255, ANYOF_BITMAP_SIZE) |
299 | #define ANYOF_BITMAP_CLEARALL(p) \ |
300 | Zero (ANYOF_BITMAP(p), ANYOF_BITMAP_SIZE) |
301 | /* Check that all 256 bits are all set. Used in S_cl_is_anything() */ |
302 | #define ANYOF_BITMAP_TESTALLSET(p) \ |
303 | memEQ (ANYOF_BITMAP(p), "\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377\377", ANYOF_BITMAP_SIZE) |
304 | |
936ed897 |
305 | #define ANYOF_SKIP ((ANYOF_SIZE - 1)/sizeof(regnode)) |
306 | #define ANYOF_CLASS_SKIP ((ANYOF_CLASS_SIZE - 1)/sizeof(regnode)) |
307 | #define ANYOF_CLASS_ADD_SKIP (ANYOF_CLASS_SKIP - ANYOF_SKIP) |
c277df42 |
308 | |
a687059c |
309 | /* |
310 | * Utility definitions. |
311 | */ |
312 | #ifndef lint |
2304df62 |
313 | #ifndef CHARMASK |
e1ec3a88 |
314 | #define UCHARAT(p) ((int)*(const U8*)(p)) |
a687059c |
315 | #else |
2304df62 |
316 | #define UCHARAT(p) ((int)*(p)&CHARMASK) |
a687059c |
317 | #endif |
318 | #else /* lint */ |
6b88bc9c |
319 | #define UCHARAT(p) PL_regdummy |
a687059c |
320 | #endif /* lint */ |
321 | |
c277df42 |
322 | #define EXTRA_SIZE(guy) ((sizeof(guy)-1)/sizeof(struct regnode)) |
323 | |
3baa4c62 |
324 | #define REG_SEEN_ZERO_LEN 1 |
325 | #define REG_SEEN_LOOKBEHIND 2 |
326 | #define REG_SEEN_GPOS 4 |
327 | #define REG_SEEN_EVAL 8 |
cce850e4 |
328 | #define REG_SEEN_CANY 16 |
329 | #define REG_SEEN_SANY REG_SEEN_CANY /* src bckwrd cmpt */ |
d09b2d29 |
330 | |
73c4f7a1 |
331 | START_EXTERN_C |
332 | |
d09b2d29 |
333 | #include "regnodes.h" |
334 | |
4d61ec05 |
335 | /* The following have no fixed length. U8 so we can do strchr() on it. */ |
d09b2d29 |
336 | #ifndef DOINIT |
4d61ec05 |
337 | EXTCONST U8 PL_varies[]; |
d09b2d29 |
338 | #else |
4d61ec05 |
339 | EXTCONST U8 PL_varies[] = { |
0111c4fd |
340 | BRANCH, BACK, STAR, PLUS, CURLY, CURLYX, REF, REFF, REFFL, |
a0ed51b3 |
341 | WHILEM, CURLYM, CURLYN, BRANCHJ, IFTHEN, SUSPEND, CLUMP, 0 |
c277df42 |
342 | }; |
d09b2d29 |
343 | #endif |
c277df42 |
344 | |
4d61ec05 |
345 | /* The following always have a length of 1. U8 we can do strchr() on it. */ |
346 | /* (Note that length 1 means "one character" under UTF8, not "one octet".) */ |
d09b2d29 |
347 | #ifndef DOINIT |
4d61ec05 |
348 | EXTCONST U8 PL_simple[]; |
d09b2d29 |
349 | #else |
4d61ec05 |
350 | EXTCONST U8 PL_simple[] = { |
f33976b4 |
351 | REG_ANY, SANY, CANY, |
ffc61ed2 |
352 | ANYOF, |
353 | ALNUM, ALNUML, |
354 | NALNUM, NALNUML, |
355 | SPACE, SPACEL, |
356 | NSPACE, NSPACEL, |
357 | DIGIT, NDIGIT, |
358 | 0 |
c277df42 |
359 | }; |
360 | #endif |
361 | |
73c4f7a1 |
362 | END_EXTERN_C |
cad2e5aa |
363 | |
364 | typedef struct re_scream_pos_data_s |
365 | { |
366 | char **scream_olds; /* match pos */ |
367 | I32 *scream_pos; /* Internal iterator of scream. */ |
368 | } re_scream_pos_data; |
369 | |
261faec3 |
370 | /* .what is a character array with one character for each member of .data |
371 | * The character describes the function of the corresponding .data item: |
0111c4fd |
372 | * f - start-class data for regstclass optimization |
261faec3 |
373 | * n - Root of op tree for (?{EVAL}) item |
374 | * o - Start op for (?{EVAL}) item |
375 | * p - Pad for (?{EVAL} item |
9e55ce06 |
376 | * s - swash for unicode-style character class, and the multicharacter |
377 | * strings resulting from casefolding the single-character entries |
378 | * in the character class |
a3621e74 |
379 | * t - trie struct |
261faec3 |
380 | * 20010712 mjd@plover.com |
c3613ed7 |
381 | * (Remember to update re_dup() and pregfree() if you add any items.) |
261faec3 |
382 | */ |
cad2e5aa |
383 | struct reg_data { |
384 | U32 count; |
385 | U8 *what; |
386 | void* data[1]; |
387 | }; |
388 | |
389 | struct reg_substr_datum { |
390 | I32 min_offset; |
391 | I32 max_offset; |
33b8afdf |
392 | SV *substr; /* non-utf8 variant */ |
393 | SV *utf8_substr; /* utf8 variant */ |
cad2e5aa |
394 | }; |
395 | |
396 | struct reg_substr_data { |
397 | struct reg_substr_datum data[3]; /* Actual array */ |
398 | }; |
399 | |
400 | #define anchored_substr substrs->data[0].substr |
33b8afdf |
401 | #define anchored_utf8 substrs->data[0].utf8_substr |
cad2e5aa |
402 | #define anchored_offset substrs->data[0].min_offset |
403 | #define float_substr substrs->data[1].substr |
33b8afdf |
404 | #define float_utf8 substrs->data[1].utf8_substr |
cad2e5aa |
405 | #define float_min_offset substrs->data[1].min_offset |
406 | #define float_max_offset substrs->data[1].max_offset |
407 | #define check_substr substrs->data[2].substr |
33b8afdf |
408 | #define check_utf8 substrs->data[2].utf8_substr |
cad2e5aa |
409 | #define check_offset_min substrs->data[2].min_offset |
410 | #define check_offset_max substrs->data[2].max_offset |
a3621e74 |
411 | |
412 | |
413 | |
414 | /* trie related stuff */ |
415 | /* an accepting state/position*/ |
416 | struct _reg_trie_accepted { |
417 | U8 *endpos; |
418 | U16 wordnum; |
419 | }; |
420 | /* a transition record for the state machine. the |
421 | check field determines which state "owns" the |
422 | transition. the char the transition is for is |
423 | determined by offset from the owning states base |
424 | field. the next field determines which state |
425 | is to be transitioned to if any. |
426 | */ |
427 | struct _reg_trie_trans { |
428 | U32 next; |
429 | U32 check; |
430 | }; |
431 | |
432 | /* a transition list element for the list based representation */ |
433 | struct _reg_trie_trans_list_elem { |
434 | U16 forid; |
435 | U32 newstate; |
436 | }; |
437 | typedef struct _reg_trie_trans_list_elem reg_trie_trans_le; |
438 | |
439 | /* a state for compressed nodes. base is an offset |
440 | into an array of reg_trie_trans array. If wordnum is |
441 | nonzero the state is accepting. if base is zero then |
442 | the state has no children (and will be accepting) |
443 | */ |
444 | struct _reg_trie_state { |
445 | U16 wordnum; |
446 | union { |
447 | U32 base; |
448 | reg_trie_trans_le* list; |
449 | } trans; |
450 | }; |
451 | |
452 | |
453 | |
454 | typedef struct _reg_trie_accepted reg_trie_accepted; |
455 | typedef struct _reg_trie_state reg_trie_state; |
456 | typedef struct _reg_trie_trans reg_trie_trans; |
457 | |
458 | |
459 | /* anything in here that needs to be freed later |
460 | should be dealt with in pregfree */ |
461 | struct _reg_trie_data { |
462 | U16 uniquecharcount; |
463 | U16 wordcount; |
464 | STRLEN charcount; |
465 | U32 laststate; |
cc601c31 |
466 | U32 lasttrans; |
a3621e74 |
467 | U16 *charmap; |
468 | HV *widecharmap; |
469 | reg_trie_state *states; |
470 | reg_trie_trans *trans; |
471 | U32 refcount; |
472 | #ifdef DEBUGGING |
473 | AV *words; |
474 | AV *revcharmap; |
475 | #endif |
476 | }; |
477 | |
478 | typedef struct _reg_trie_data reg_trie_data; |
479 | |
480 | /* these defines assume uniquecharcount is the correct variable, and state may be evaluated twice */ |
481 | #define TRIE_NODENUM(state) (((state)-1)/(trie->uniquecharcount)+1) |
482 | #define SAFE_TRIE_NODENUM(state) ((state) ? (((state)-1)/(trie->uniquecharcount)+1) : (state)) |
483 | #define TRIE_NODEIDX(state) ((state) ? (((state)-1)*(trie->uniquecharcount)+1) : (state)) |
484 | |
485 | #define DO_TRIE 1 |
486 | #define TRIE_DEBUG 1 |
487 | |
0111c4fd |
488 | #define RE_TRIE_MAXBUF_INIT 65536 |
489 | #define RE_TRIE_MAXBUF_NAME "\022E_TRIE_MAXBUF" |
a3621e74 |
490 | #define RE_DEBUG_FLAGS "\022E_DEBUG_FLAGS" |
491 | |
492 | /* If you change these be sure to update ext/re/re.pm as well */ |
493 | #define RE_DEBUG_COMPILE 1 |
494 | #define RE_DEBUG_EXECUTE 2 |
495 | #define RE_DEBUG_TRIE_COMPILE 4 |
496 | #define RE_DEBUG_TRIE_EXECUTE 8 |
497 | #define RE_DEBUG_TRIE_MORE 16 |
498 | #define RE_DEBUG_OPTIMISE 32 |
499 | #define RE_DEBUG_OFFSETS 64 |
500 | |
501 | #define DEBUG_OPTIMISE_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_OPTIMISE) x ) |
502 | #define DEBUG_EXECUTE_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_EXECUTE) x ) |
503 | #define DEBUG_COMPILE_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_COMPILE) x ) |
504 | #define DEBUG_OFFSETS_r(x) DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_OFFSETS) x ) |
505 | #define DEBUG_TRIE_r(x) DEBUG_r( \ |
506 | if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_COMPILE \ |
507 | || SvIV(re_debug_flags) & RE_DEBUG_TRIE_EXECUTE ) \ |
508 | x \ |
509 | ) |
510 | #define DEBUG_TRIE_EXECUTE_r(x) \ |
511 | DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_EXECUTE) x ) |
512 | |
513 | #define DEBUG_TRIE_COMPILE_r(x) \ |
514 | DEBUG_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_COMPILE) x ) |
515 | |
516 | #define DEBUG_TRIE_EXECUTE_MORE_r(x) \ |
517 | DEBUG_TRIE_EXECUTE_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_MORE) x ) |
518 | |
519 | #define DEBUG_TRIE_COMPILE_MORE_r(x) \ |
520 | DEBUG_TRIE_COMPILE_r( if (SvIV(re_debug_flags) & RE_DEBUG_TRIE_MORE) x ) |
521 | |
522 | #define GET_RE_DEBUG_FLAGS DEBUG_r( \ |
523 | re_debug_flags=get_sv(RE_DEBUG_FLAGS, 1); \ |
524 | if (!SvIOK(re_debug_flags)) { \ |
525 | sv_setiv(re_debug_flags, RE_DEBUG_COMPILE | RE_DEBUG_EXECUTE | RE_DEBUG_OFFSETS); \ |
526 | } \ |
527 | ) |
528 | |
529 | |
530 | #ifdef DEBUGGING |
e1ec3a88 |
531 | #define GET_RE_DEBUG_FLAGS_DECL SV *re_debug_flags = NULL; GET_RE_DEBUG_FLAGS; |
a3621e74 |
532 | #else |
533 | #define GET_RE_DEBUG_FLAGS_DECL |
534 | #endif |
535 | |
536 | |