This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
regex matching: Don't do unnecessary work
[perl5.git] / regcomp.h
CommitLineData
a0d0e21e 1/* regcomp.h
d6376244 2 *
4bb101f2 3 * Copyright (C) 1991, 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
663f364b 4 * 2000, 2001, 2002, 2003, 2005, 2006, 2007, by Larry Wall and others
d6376244
JH
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 */
e1d1eefb 10#include "regcharclass.h"
a687059c 11
58e23c8d 12/* Convert branch sequences to more efficient trie ops? */
07be1b83 13#define PERL_ENABLE_TRIE_OPTIMISATION 1
58e23c8d 14
486ec47a 15/* Be really aggressive about optimising patterns with trie sequences? */
07be1b83 16#define PERL_ENABLE_EXTENDED_TRIE_OPTIMISATION 1
58e23c8d
YO
17
18/* Should the optimiser take positive assertions into account? */
68ba3a3f 19#define PERL_ENABLE_POSITIVE_ASSERTION_STUDY 0
58e23c8d
YO
20
21/* Not for production use: */
07be1b83 22#define PERL_ENABLE_EXPERIMENTAL_REGEX_OPTIMISATIONS 0
58e23c8d 23
7122b237
YO
24/* Activate offsets code - set to if 1 to enable */
25#ifdef DEBUGGING
26#define RE_TRACK_PATTERN_OFFSETS
27#endif
28
a687059c
LW
29/*
30 * The "internal use only" fields in regexp.h are present to pass info from
31 * compile to execute that permits the execute phase to run lots faster on
32 * simple cases. They are:
33 *
bd61b366 34 * regstart sv that must begin a match; NULL if none obvious
a687059c
LW
35 * reganch is the match anchored (at beginning-of-line only)?
36 * regmust string (pointer into program) that match must include, or NULL
79072805 37 * [regmust changed to SV* for bminstr()--law]
a687059c
LW
38 * regmlen length of regmust string
39 * [regmlen not used currently]
40 *
41 * Regstart and reganch permit very fast decisions on suitable starting points
42 * for a match, cutting down the work a lot. Regmust permits fast rejection
43 * of lines that cannot possibly match. The regmust tests are costly enough
e50aee73 44 * that pregcomp() supplies a regmust only if the r.e. contains something
a687059c
LW
45 * potentially expensive (at present, the only such thing detected is * or +
46 * at the start of the r.e., which can involve a lot of backup). Regmlen is
e50aee73 47 * supplied because the test in pregexec() needs it and pregcomp() is computing
a687059c
LW
48 * it anyway.
49 * [regmust is now supplied always. The tests that use regmust have a
50 * heuristic that disables the test if it usually matches.]
51 *
52 * [In fact, we now use regmust in many cases to locate where the search
53 * starts in the string, so if regback is >= 0, the regmust search is never
54 * wasted effort. The regback variable says how many characters back from
55 * where regmust matched is the earliest possible start of the match.
56 * For instance, /[a-z].foo/ has a regmust of 'foo' and a regback of 2.]
57 */
58
59/*
60 * Structure for regexp "program". This is essentially a linear encoding
61 * of a nondeterministic finite-state machine (aka syntax charts or
62 * "railroad normal form" in parsing technology). Each node is an opcode
63 * plus a "next" pointer, possibly plus an operand. "Next" pointers of
64 * all nodes except BRANCH implement concatenation; a "next" pointer with
65 * a BRANCH on both ends of it is connecting two alternatives. (Here we
66 * have one of the subtle syntax dependencies: an individual BRANCH (as
67 * opposed to a collection of them) is never concatenated with anything
68 * because of operator precedence.) The operand of some types of node is
69 * a literal string; for others, it is a node leading into a sub-FSM. In
70 * particular, the operand of a BRANCH node is the first node of the branch.
71 * (NB this is *not* a tree structure: the tail of the branch connects
a3621e74
YO
72 * to the thing following the set of BRANCHes.) The opcodes are defined
73 * in regnodes.h which is generated from regcomp.sym by regcomp.pl.
a687059c
LW
74 */
75
a687059c
LW
76/*
77 * A node is one char of opcode followed by two chars of "next" pointer.
78 * "Next" pointers are stored as two 8-bit pieces, high order first. The
79 * value is a positive offset from the opcode of the node containing it.
80 * An operand, if any, simply follows the node. (Note that much of the
81 * code generation knows about this implicit relationship.)
82 *
83 * Using two bytes for the "next" pointer is vast overkill for most things,
84 * but allows patterns to get big without disasters.
85 *
e91177ed 86 * [The "next" pointer is always aligned on an even
62e6ef33 87 * boundary, and reads the offset directly as a short.]
a687059c 88 */
785a26d5
YO
89
90/* This is the stuff that used to live in regexp.h that was truly
91 private to the engine itself. It now lives here. */
92
28d8d7f4 93
785a26d5
YO
94
95 typedef struct regexp_internal {
1f1031fe 96 int name_list_idx; /* Optional data index of an array of paren names */
7122b237
YO
97 union {
98 U32 *offsets; /* offset annotations 20001228 MJD
99 data about mapping the program to the
100 string -
101 offsets[0] is proglen when this is used
102 */
103 U32 proglen;
104 } u;
105
6d5c990f
RGS
106 regnode *regstclass; /* Optional startclass as identified or constructed
107 by the optimiser */
108 struct reg_data *data; /* Additional miscellaneous data used by the program.
109 Used to make it easier to clone and free arbitrary
110 data that the regops need. Often the ARG field of
111 a regop is an index into this structure */
3d2bd50a
DM
112 struct reg_code_block *code_blocks;/* positions of literal (?{}) */
113 int num_code_blocks; /* size of code_blocks[] */
6d5c990f
RGS
114 regnode program[1]; /* Unwarranted chumminess with compiler. */
115} regexp_internal;
116
117#define RXi_SET(x,y) (x)->pprivate = (void*)(y)
118#define RXi_GET(x) ((regexp_internal *)((x)->pprivate))
119#define RXi_GET_DECL(r,ri) regexp_internal *ri = RXi_GET(r)
785a26d5
YO
120/*
121 * Flags stored in regexp->intflags
122 * These are used only internally to the regexp engine
123 *
124 * See regexp.h for flags used externally to the regexp engine
125 */
e3e400ec
YO
126#define RXp_INTFLAGS(rx) ((rx)->intflags)
127#define RX_INTFLAGS(prog) RXp_INTFLAGS(ReANY(prog))
128
785a26d5
YO
129#define PREGf_SKIP 0x00000001
130#define PREGf_IMPLICIT 0x00000002 /* Converted .* to ^.* */
131#define PREGf_NAUGHTY 0x00000004 /* how exponential is this pattern? */
132#define PREGf_VERBARG_SEEN 0x00000008
133#define PREGf_CUTGROUP_SEEN 0x00000010
732caac7 134#define PREGf_USE_RE_EVAL 0x00000020 /* compiled with "use re 'eval'" */
e3e400ec
YO
135/* these used to be extflags, but are now intflags */
136#define PREGf_NOSCAN 0x00000040
33c28ab2 137 /* spare */
8e1490ee
YO
138#define PREGf_GPOS_SEEN 0x00000100
139#define PREGf_GPOS_FLOAT 0x00000200
785a26d5 140
d3d47aac
YO
141#define PREGf_ANCH_MBOL 0x00000400
142#define PREGf_ANCH_SBOL 0x00000800
143#define PREGf_ANCH_GPOS 0x00001000
8e1490ee 144
d3d47aac
YO
145#define PREGf_ANCH \
146 ( PREGf_ANCH_SBOL | PREGf_ANCH_GPOS | PREGf_ANCH_MBOL )
785a26d5
YO
147
148/* this is where the old regcomp.h started */
a687059c 149
c277df42 150struct regnode_string {
cd439c50 151 U8 str_len;
c277df42
IZ
152 U8 type;
153 U16 next_off;
cd439c50 154 char string[1];
c277df42 155};
a687059c 156
6bda09f9
YO
157/* Argument bearing node - workhorse,
158 arg1 is often for the data field */
c277df42
IZ
159struct regnode_1 {
160 U8 flags;
161 U8 type;
162 U16 next_off;
163 U32 arg1;
164};
165
6bda09f9
YO
166/* Similar to a regnode_1 but with an extra signed argument */
167struct regnode_2L {
168 U8 flags;
169 U8 type;
170 U16 next_off;
171 U32 arg1;
172 I32 arg2;
173};
174
175/* 'Two field' -- Two 16 bit unsigned args */
c277df42
IZ
176struct regnode_2 {
177 U8 flags;
178 U8 type;
179 U16 next_off;
180 U16 arg1;
181 U16 arg2;
182};
183
e0a1ff7a
KW
184/* This give the number of code points that can be in the bitmap of an ANYOF
185 * node. The shift number must currently be one of: 8..12. It can't be less
186 * than 8 (256) because some code relies on it being at least that. Above 12
187 * (4096), and you start running into warnings that some data structure widths
188 * have been exceeded, though the test suite as of this writing still passes
189 * for up through 16, which is as high as anyone would ever want to go,
190 * encompassing all of the Unicode BMP, and thus including all the economically
191 * important world scripts. At 12 most of them are: including Arabic,
192 * Cyrillic, Greek, Hebrew, Indian subcontinent, Latin, and Thai; but not Han,
193 * Japanese, nor Korean. (The regarglen structure in regnodes.h is a U8, and
194 * the trie types TRIEC and AHOCORASICKC are larger than U8 for shift values
195 * below above 12.) Be sure to benchmark before changing, as larger sizes do
196 * significantly slow down the test suite */
197#define NUM_ANYOF_CODE_POINTS (1 << 8)
6bda09f9 198
dcb20b36 199#define ANYOF_BITMAP_SIZE (NUM_ANYOF_CODE_POINTS / 8) /* 8 bits/Byte */
936ed897 200
1ee208c4
KW
201/* Note that these form structs which are supersets of the next smaller one, by
202 * appending fields. Alignment problems can occur if one of those optional
203 * fields requires stricter alignment than the base struct. And formal
204 * parameters that can really be two or more of the structs should be
205 * declared as the smallest one it could be. See commit message for
206 * 7dcac5f6a5195002b55c935ee1d67f67e1df280b. Regnode allocation is done
207 * without regard to alignment, and changing it to would also require changing
208 * the code that inserts and deletes regnodes. The basic single-argument
209 * regnode has a U32, which is what reganode() allocates as a unit. Therefore
210 * no field can require stricter alignment than U32. */
211
786e8c11 212/* also used by trie */
936ed897
IZ
213struct regnode_charclass {
214 U8 flags;
215 U8 type;
216 U16 next_off;
9da99c5e 217 U32 arg1; /* set by set_ANYOF_arg() */
19103b91 218 char bitmap[ANYOF_BITMAP_SIZE]; /* only compile-time */
936ed897
IZ
219};
220
bf3f2d85
KW
221/* has runtime (locale) \d, \w, ..., [:posix:] classes */
222struct regnode_charclass_class {
93e92956 223 U8 flags; /* ANYOF_MATCHES_POSIXL bit must go here */
936ed897
IZ
224 U8 type;
225 U16 next_off;
7e817188 226 U32 arg1;
03fa83ba 227 char bitmap[ANYOF_BITMAP_SIZE]; /* both compile-time ... */
9cf78fa0 228 U32 classflags; /* and run-time */
936ed897
IZ
229};
230
85c8e306
KW
231/* A synthetic start class (SSC); is a regnode_charclass_posixl_fold, plus an
232 * extra SV*, used only during its construction and which is not used by
233 * regexec.c. Note that the 'next_off' field is unused, as the SSC stands
234 * alone, so there is never a next node. Also, there is no alignment issue,
b8fda935 235 * because these are declared or allocated as a complete unit so the compiler
85c8e306
KW
236 * takes care of alignment. This is unlike the other regnodes which are
237 * allocated in terms of multiples of a single-argument regnode. SSC nodes can
238 * have a pointer field because there is no alignment issue, and because it is
239 * set to NULL after construction, before any cloning of the pattern */
b8f7bb16 240struct regnode_ssc {
93e92956 241 U8 flags; /* ANYOF_MATCHES_POSIXL bit must go here */
b8f7bb16
KW
242 U8 type;
243 U16 next_off;
7e817188 244 U32 arg1;
03fa83ba 245 char bitmap[ANYOF_BITMAP_SIZE]; /* both compile-time ... */
631e9161 246 U32 classflags; /* ... and run-time */
85c8e306
KW
247
248 /* Auxiliary, only used during construction; NULL afterwards: list of code
249 * points matched */
250 SV* invlist;
b8f7bb16
KW
251};
252
71068078
KW
253/* We take advantage of 'next_off' not otherwise being used in the SSC by
254 * actually using it: by setting it to 1. This allows us to test and
255 * distinguish between an SSC and other ANYOF node types, as 'next_off' cannot
256 * otherwise be 1, because it is the offset to the next regnode expressed in
257 * units of regnodes. Since an ANYOF node contains extra fields, it adds up
258 * to 12 regnode units on 32-bit systems, (hence the minimum this can be (if
259 * not 0) is 11 there. Even if things get tightly packed on a 64-bit system,
260 * it still would be more than 1. */
261#define set_ANYOF_SYNTHETIC(n) STMT_START{ OP(n) = ANYOF; \
262 NEXT_OFF(n) = 1; \
263 } STMT_END
c1071334 264#define is_ANYOF_SYNTHETIC(n) (PL_regkind[OP(n)] == ANYOF && NEXT_OFF(n) == 1)
71068078 265
9c7e81e8
AD
266/* XXX fix this description.
267 Impose a limit of REG_INFTY on various pattern matching operations
268 to limit stack growth and to avoid "infinite" recursions.
269*/
270/* The default size for REG_INFTY is I16_MAX, which is the same as
271 SHORT_MAX (see perl.h). Unfortunately I16 isn't necessarily 16 bits
272 (see handy.h). On the Cray C90, sizeof(short)==4 and hence I16_MAX is
273 ((1<<31)-1), while on the Cray T90, sizeof(short)==8 and I16_MAX is
274 ((1<<63)-1). To limit stack growth to reasonable sizes, supply a
275 smaller default.
276 --Andy Dougherty 11 June 1998
277*/
278#if SHORTSIZE > 2
279# ifndef REG_INFTY
280# define REG_INFTY ((1<<15)-1)
281# endif
282#endif
283
284#ifndef REG_INFTY
83cfe48c
IZ
285# define REG_INFTY I16_MAX
286#endif
a687059c 287
e91177ed
IZ
288#define ARG_VALUE(arg) (arg)
289#define ARG__SET(arg,val) ((arg) = (val))
c277df42 290
e8e3b0fb
JH
291#undef ARG
292#undef ARG1
293#undef ARG2
294
c277df42
IZ
295#define ARG(p) ARG_VALUE(ARG_LOC(p))
296#define ARG1(p) ARG_VALUE(ARG1_LOC(p))
297#define ARG2(p) ARG_VALUE(ARG2_LOC(p))
6bda09f9 298#define ARG2L(p) ARG_VALUE(ARG2L_LOC(p))
786e8c11 299
c277df42
IZ
300#define ARG_SET(p, val) ARG__SET(ARG_LOC(p), (val))
301#define ARG1_SET(p, val) ARG__SET(ARG1_LOC(p), (val))
302#define ARG2_SET(p, val) ARG__SET(ARG2_LOC(p), (val))
6bda09f9 303#define ARG2L_SET(p, val) ARG__SET(ARG2L_LOC(p), (val))
c277df42 304
e8e3b0fb
JH
305#undef NEXT_OFF
306#undef NODE_ALIGN
307
075abff3
AL
308#define NEXT_OFF(p) ((p)->next_off)
309#define NODE_ALIGN(node)
aa48e906
YO
310/* the following define was set to 0xde in 075abff3
311 * as part of some linting logic. I have set it to 0
312 * as otherwise in every place where we /might/ set flags
313 * we have to set it 0 explicitly, which duplicates
314 * assignments and IMO adds an unacceptable level of
315 * surprise to working in the regex engine. If this
316 * is changed from 0 then at the very least make sure
317 * that SBOL for /^/ sets the flags to 0 explicitly.
318 * -- Yves */
319#define NODE_ALIGN_FILL(node) ((node)->flags = 0)
a687059c 320
c277df42
IZ
321#define SIZE_ALIGN NODE_ALIGN
322
e8e3b0fb
JH
323#undef OP
324#undef OPERAND
325#undef MASK
326#undef STRING
327
e91177ed 328#define OP(p) ((p)->type)
b988e673
KW
329#define FLAGS(p) ((p)->flags) /* Caution: Doesn't apply to all \
330 regnode types. For some, it's the \
331 character set of the regnode */
e91177ed 332#define OPERAND(p) (((struct regnode_string *)p)->string)
cd439c50
IZ
333#define MASK(p) ((char*)OPERAND(p))
334#define STR_LEN(p) (((struct regnode_string *)p)->str_len)
335#define STRING(p) (((struct regnode_string *)p)->string)
336#define STR_SZ(l) ((l + sizeof(regnode) - 1) / sizeof(regnode))
337#define NODE_SZ_STR(p) (STR_SZ(STR_LEN(p))+1)
338
e8e3b0fb
JH
339#undef NODE_ALIGN
340#undef ARG_LOC
341#undef NEXTOPER
342#undef PREVOPER
343
e91177ed
IZ
344#define NODE_ALIGN(node)
345#define ARG_LOC(p) (((struct regnode_1 *)p)->arg1)
346#define ARG1_LOC(p) (((struct regnode_2 *)p)->arg1)
347#define ARG2_LOC(p) (((struct regnode_2 *)p)->arg2)
6bda09f9 348#define ARG2L_LOC(p) (((struct regnode_2L *)p)->arg2)
786e8c11 349
e91177ed
IZ
350#define NODE_STEP_REGNODE 1 /* sizeof(regnode)/sizeof(regnode) */
351#define EXTRA_STEP_2ARGS EXTRA_SIZE(struct regnode_2)
352
353#define NODE_STEP_B 4
c277df42
IZ
354
355#define NEXTOPER(p) ((p) + NODE_STEP_REGNODE)
356#define PREVOPER(p) ((p) - NODE_STEP_REGNODE)
357
e91177ed 358#define FILL_ADVANCE_NODE(ptr, op) STMT_START { \
c277df42 359 (ptr)->type = op; (ptr)->next_off = 0; (ptr)++; } STMT_END
e91177ed 360#define FILL_ADVANCE_NODE_ARG(ptr, op, arg) STMT_START { \
c277df42 361 ARG_SET(ptr, arg); FILL_ADVANCE_NODE(ptr, op); (ptr) += 1; } STMT_END
9990d5ab
KW
362#define FILL_ADVANCE_NODE_2L_ARG(ptr, op, arg1, arg2) \
363 STMT_START { \
364 ARG_SET(ptr, arg1); \
365 ARG2L_SET(ptr, arg2); \
366 FILL_ADVANCE_NODE(ptr, op); \
367 (ptr) += 2; \
368 } STMT_END
a687059c 369
22c35a8c 370#define REG_MAGIC 0234
a687059c 371
9a81a976 372#define SIZE_ONLY (RExC_emit == (regnode *) & RExC_emit_dummy)
6d15ebf3
KW
373#define PASS1 SIZE_ONLY
374#define PASS2 (! SIZE_ONLY)
c277df42 375
6f16c8da 376/* If the bitmap fully represents what this ANYOF node can match, the
137165a6
KW
377 * ARG is set to this special value (since 0, 1, ... are legal, but will never
378 * reach this high). */
93e92956 379#define ANYOF_ONLY_HAS_BITMAP ((U32) -1)
137165a6 380
d2f465e9
KW
381/* Below are the flags for node->flags of ANYOF. These are in short supply,
382 * with none currently available. The ABOVE_BITMAP_ALL bit could be freed up
7bc66b18 383 * by resorting to creating a swash containing everything above 255. This
d2f465e9
KW
384 * seems likely to introduce a performance penalty (but actual numbers haven't
385 * been done), so its probably better do some of the other possibilities below
386 * in preference to this.
387 *
388 * If just one bit is required, it seems to me (khw) that the best option would
389 * be to turn the ANYOF_LOC_REQ_UTF8 bit into a separate node type: a
390 * specialization of the ANYOFL type, freeing up the currently occupied bit.
391 * When turning a bit into a node type, one has to take into consideration that
392 * a SSC may use that bit -- not just a regular ANYOF[DL]?. In the case of
393 * ANYOF_LOC_REQ_UTF8, the only likely problem is accurately settting the SSC
394 * node-type to the new one, which would likely involve S_ssc_or and S_ssc_and,
395 * and not how the SSC currently gets set to ANYOFL. This bit is a natural
396 * candidate for being a separate node type because it is a specialization of
397 * the current ANYOFL, and because no other ANYOFL-only bits are set when it
398 * is; also most of its uses are actually outside the reginclass() function, so
399 * this could be done with no performance penalty. The other potential bits
400 * seem to me to have a potential issue with a combinatorial explosion of node
401 * types, because of not having that mutual exclusivity, where you may end up
402 * having to have a node type for bitX being set, one for bitY, and one for
403 * both bitXY.
404 *
405 * If you don't want to do this, or two bits are required, one could instead
406 * rename the ANYOF_POSIXL bit to be ANYOFL_LARGE, to mean that the ANYOF node
407 * has an extra 32 bits beyond what a regular one does. That's what it
408 * effectively means now, with the extra space all for the POSIX class bits.
409 * But those classes actually only occupy 30 bits, so the ANYOF_LOC_REQ_BIT (if
410 * an extra node type for it hasn't been created) and/or the ANYOF_LOC_FOLD
411 * bits could be moved there. The downside of this is that ANYOFL nodes with
412 * whichever of the bits get moved would have to have the extra space always
413 * allocated.
414 *
415 * If three bits are required, one could additionally make a node type for
416 * ANYOFL_LARGE, removing that as a bit, and move both the above bits to that
417 * extra word. There isn't an SSC problem as all SSCs are this large anyway,
418 * and the SSC could be set to this node type. REGINCLASS would have to be
419 * modified so that if the node type were this, it would call reginclass().
420 * as the flag bit that does this now would be gone.
a4525e78 421 *
34fdef84 422 * Several flags are not used in synthetic start class (SSC) nodes, so could be
93e92956
KW
423 * shared should new flags be needed for SSCs, like SSC_MATCHES_EMPTY_STRING
424 * now. */
b8c5462f 425
d2f465e9
KW
426/* If this is set, the result of the match should be complemented. regexec.c
427 * is expecting this to be in the low bit. Never in an SSC */
93e92956 428#define ANYOF_INVERT 0x01
1984f6a0 429
a0dd4231 430/* For the SSC node only, which cannot be inverted, so is shared with that bit.
93e92956
KW
431 * This is used only during regex compilation. */
432#define SSC_MATCHES_EMPTY_STRING ANYOF_INVERT
a0dd4231 433
509d4de0
KW
434/* Set if this is a regnode_charclass_posixl vs a regnode_charclass. This
435 * is used for runtime \d, \w, [:posix:], ..., which are used only in locale
436 * and the optimizer's synthetic start class. Non-locale \d, etc are resolved
437 * at compile-time. Only set under /l; can be in SSC */
438#define ANYOF_MATCHES_POSIXL 0x02
b63f73da
KW
439
440/* The fold is calculated and stored in the bitmap where possible at compile
441 * time. However under locale, the actual folding varies depending on
442 * what the locale is at the time of execution, so it has to be deferred until
d2f465e9 443 * then. Only set under /l; never in an SSC */
93e92956 444#define ANYOF_LOC_FOLD 0x04
b63f73da 445
8c9b4e63
KW
446/* If set, ANYOF_LOC_FOLD is also set, and there are potential matches that
447 * will be valid only if the locale is a UTF-8 one. */
448#define ANYOF_ONLY_UTF8_LOC_FOLD_MATCHES 0x08
b3b1cf17 449
509d4de0 450/* If set, means to warn if runtime locale isn't a UTF-8 one. Only under /l.
108316fb
KW
451 * If set, none of INVERT, LOC_FOLD, POSIXL,
452 * ANYOF_SHARED_d_UPPER_LATIN1_UTF8_STRING_MATCHES_non_d_RUNTIME_USER_PROP can
509d4de0 453 * be set. Can be in an SSC */
b3b1cf17 454#define ANYOF_LOC_REQ_UTF8 0x10
c8d3cd88 455
509d4de0
KW
456/* If set, the node matches every code point NUM_ANYOF_CODE_POINTS and above.
457 * Can be in an SSC */
b3b1cf17 458#define ANYOF_MATCHES_ALL_ABOVE_BITMAP 0x20
ef87b810 459
108316fb
KW
460/* Shared bit:
461 * Under /d it means the ANYOFD node matches more things if the target
462 * string is encoded in UTF-8; any such things will be non-ASCII,
463 * characters that are < 256, and can be accessed via the swash.
464 * When not under /d, it means the ANYOF node contains a user-defined
465 * property that wasn't yet defined at the time the regex was compiled,
466 * and so must be looked up at runtime, by creating a swash
467 * (These uses are mutually exclusive because a user-defined property is
468 * specified by \p{}, and \p{} implies /u which deselects /d). The long macro
469 * name is to make sure that you are cautioned about its shared nature. Only
470 * the non-/d meaning can be in an SSC */
471#define ANYOF_SHARED_d_UPPER_LATIN1_UTF8_STRING_MATCHES_non_d_RUNTIME_USER_PROP 0x40
f240c685
KW
472
473/* Shared bit:
d2f465e9 474 * Under /d it means the ANYOFD node matches all non-ASCII Latin1
f240c685
KW
475 * characters when the target string is not in utf8.
476 * When not under /d, it means the ANYOF node should raise a warning if
477 * matching against an above-Unicode code point.
478 * (These uses are mutually exclusive because the warning requires a \p{}, and
479 * \p{} implies /u which deselects /d). An SSC node only has this bit set if
480 * what is meant is the warning. The long macro name is to make sure that you
481 * are cautioned about its shared nature */
482#define ANYOF_SHARED_d_MATCHES_ALL_NON_UTF8_NON_ASCII_non_d_WARN_SUPER 0x80
7bc66b18 483
3b04b210 484#define ANYOF_FLAGS_ALL (0xff)
653099ff 485
93e92956 486#define ANYOF_LOCALE_FLAGS (ANYOF_LOC_FOLD | ANYOF_MATCHES_POSIXL)
ace6b0e4 487
eff8b7dc
KW
488/* These are the flags that apply to both regular ANYOF nodes and synthetic
489 * start class nodes during construction of the SSC. During finalization of
d2f465e9 490 * the SSC, other of the flags may get added to it */
108316fb 491#define ANYOF_COMMON_FLAGS ANYOF_LOC_REQ_UTF8
eff8b7dc 492
936ed897 493/* Character classes for node->classflags of ANYOF */
653099ff 494/* Should be synchronized with a table in regprop() */
575a8a29 495/* 2n should be the normal one, paired with its complement at 2n+1 */
b8c5462f 496
08ada29e
KW
497#define ANYOF_ALPHA ((_CC_ALPHA) * 2)
498#define ANYOF_NALPHA ((ANYOF_ALPHA) + 1)
e8d596e0
KW
499#define ANYOF_ALPHANUMERIC ((_CC_ALPHANUMERIC) * 2) /* [[:alnum:]] isalnum(3), utf8::IsAlnum */
500#define ANYOF_NALPHANUMERIC ((ANYOF_ALPHANUMERIC) + 1)
08ada29e
KW
501#define ANYOF_ASCII ((_CC_ASCII) * 2)
502#define ANYOF_NASCII ((ANYOF_ASCII) + 1)
e8d596e0
KW
503#define ANYOF_BLANK ((_CC_BLANK) * 2) /* GNU extension: space and tab: non-vertical space */
504#define ANYOF_NBLANK ((ANYOF_BLANK) + 1)
b0d691b2
KW
505#define ANYOF_CASED ((_CC_CASED) * 2) /* Pseudo class for [:lower:] or
506 [:upper:] under /i */
507#define ANYOF_NCASED ((ANYOF_CASED) + 1)
08ada29e
KW
508#define ANYOF_CNTRL ((_CC_CNTRL) * 2)
509#define ANYOF_NCNTRL ((ANYOF_CNTRL) + 1)
e8d596e0
KW
510#define ANYOF_DIGIT ((_CC_DIGIT) * 2) /* \d */
511#define ANYOF_NDIGIT ((ANYOF_DIGIT) + 1)
08ada29e
KW
512#define ANYOF_GRAPH ((_CC_GRAPH) * 2)
513#define ANYOF_NGRAPH ((ANYOF_GRAPH) + 1)
514#define ANYOF_LOWER ((_CC_LOWER) * 2)
515#define ANYOF_NLOWER ((ANYOF_LOWER) + 1)
516#define ANYOF_PRINT ((_CC_PRINT) * 2)
517#define ANYOF_NPRINT ((ANYOF_PRINT) + 1)
518#define ANYOF_PUNCT ((_CC_PUNCT) * 2)
519#define ANYOF_NPUNCT ((ANYOF_PUNCT) + 1)
e8d596e0
KW
520#define ANYOF_SPACE ((_CC_SPACE) * 2) /* \s */
521#define ANYOF_NSPACE ((ANYOF_SPACE) + 1)
08ada29e
KW
522#define ANYOF_UPPER ((_CC_UPPER) * 2)
523#define ANYOF_NUPPER ((ANYOF_UPPER) + 1)
e8d596e0
KW
524#define ANYOF_WORDCHAR ((_CC_WORDCHAR) * 2) /* \w, PL_utf8_alnum, utf8::IsWord, ALNUM */
525#define ANYOF_NWORDCHAR ((ANYOF_WORDCHAR) + 1)
08ada29e
KW
526#define ANYOF_XDIGIT ((_CC_XDIGIT) * 2)
527#define ANYOF_NXDIGIT ((ANYOF_XDIGIT) + 1)
b8c5462f 528
87cad4c0 529/* pseudo classes below this, not stored in the class bitmap, but used as flags
e1d1eefb
YO
530 during compilation of char classes */
531
b2ef2e34
KW
532#define ANYOF_VERTWS ((_CC_VERTSPACE) * 2)
533#define ANYOF_NVERTWS ((ANYOF_VERTWS)+1)
e1d1eefb 534
b2ef2e34
KW
535/* It is best if this is the last one, as all above it are stored as bits in a
536 * bitmap, and it isn't part of that bitmap */
537#if _CC_VERTSPACE != _HIGHEST_REGCOMP_DOT_H_SYNC
538# error Problem with handy.h _HIGHEST_REGCOMP_DOT_H_SYNC #define
539#endif
540
8efd3f97
KW
541#define ANYOF_POSIXL_MAX (ANYOF_VERTWS) /* So upper loop limit is written:
542 * '< ANYOF_MAX'
543 * Hence doesn't include VERTWS, as that
544 * is a pseudo class */
545#define ANYOF_MAX ANYOF_POSIXL_MAX
546
547#if (ANYOF_POSIXL_MAX > 32) /* Must fit in 32-bit word */
b2ef2e34 548# error Problem with handy.h _CC_foo #defines
a0947d7b
KW
549#endif
550
8efd3f97
KW
551#define ANYOF_HORIZWS ((ANYOF_POSIXL_MAX)+2) /* = (ANYOF_NVERTWS + 1) */
552#define ANYOF_NHORIZWS ((ANYOF_POSIXL_MAX)+3)
87cad4c0 553
8efd3f97
KW
554#define ANYOF_UNIPROP ((ANYOF_POSIXL_MAX)+4) /* Used to indicate a Unicode
555 property: \p{} or \P{} */
f39a4dc2 556
b8c5462f
JH
557/* Backward source code compatibility. */
558
559#define ANYOF_ALNUML ANYOF_ALNUM
560#define ANYOF_NALNUML ANYOF_NALNUM
561#define ANYOF_SPACEL ANYOF_SPACE
562#define ANYOF_NSPACEL ANYOF_NSPACE
8ebfc29a
KW
563#define ANYOF_ALNUM ANYOF_WORDCHAR
564#define ANYOF_NALNUM ANYOF_NWORDCHAR
b8c5462f
JH
565
566/* Utility macros for the bitmap and classes of ANYOF */
567
936ed897 568#define ANYOF_FLAGS(p) ((p)->flags)
b8c5462f 569
bc51fd78 570#define ANYOF_BIT(c) (1U << ((c) & 7))
b8c5462f 571
d1d040e5 572#define ANYOF_POSIXL_SET(p, c) (((regnode_charclass_posixl*) (p))->classflags |= (1U << (c)))
8efd3f97
KW
573#define ANYOF_CLASS_SET(p, c) ANYOF_POSIXL_SET((p), (c))
574
d1d040e5 575#define ANYOF_POSIXL_CLEAR(p, c) (((regnode_charclass_posixl*) (p))->classflags &= ~ (1U <<(c)))
8efd3f97
KW
576#define ANYOF_CLASS_CLEAR(p, c) ANYOF_POSIXL_CLEAR((p), (c))
577
d1d040e5 578#define ANYOF_POSIXL_TEST(p, c) (((regnode_charclass_posixl*) (p))->classflags & (1U << (c)))
8efd3f97 579#define ANYOF_CLASS_TEST(p, c) ANYOF_POSIXL_TEST((p), (c))
b8c5462f 580
d1d040e5 581#define ANYOF_POSIXL_ZERO(ret) STMT_START { ((regnode_charclass_posixl*) (ret))->classflags = 0; } STMT_END
8efd3f97 582#define ANYOF_CLASS_ZERO(ret) ANYOF_POSIXL_ZERO(ret)
9cf78fa0
KW
583
584/* Shifts a bit to get, eg. 0x4000_0000, then subtracts 1 to get 0x3FFF_FFFF */
d1d040e5 585#define ANYOF_POSIXL_SETALL(ret) STMT_START { ((regnode_charclass_posixl*) (ret))->classflags = ((1U << ((ANYOF_POSIXL_MAX) - 1))) - 1; } STMT_END
8efd3f97 586#define ANYOF_CLASS_SETALL(ret) ANYOF_POSIXL_SETALL(ret)
9cf78fa0 587
8efd3f97 588#define ANYOF_POSIXL_TEST_ANY_SET(p) \
93e92956 589 ((ANYOF_FLAGS(p) & ANYOF_MATCHES_POSIXL) \
d1d040e5 590 && (((regnode_charclass_posixl*)(p))->classflags))
8efd3f97 591#define ANYOF_CLASS_TEST_ANY_SET(p) ANYOF_POSIXL_TEST_ANY_SET(p)
e867ce45 592
e0e1be5f
KW
593/* Since an SSC always has this field, we don't have to test for that; nor do
594 * we want to because the bit isn't set for SSC during its construction */
595#define ANYOF_POSIXL_SSC_TEST_ANY_SET(p) \
596 cBOOL(((regnode_ssc*)(p))->classflags)
597#define ANYOF_POSIXL_SSC_TEST_ALL_SET(p) /* Are all bits set? */ \
598 (((regnode_ssc*) (p))->classflags \
599 == ((1U << ((ANYOF_POSIXL_MAX) - 1))) - 1)
600
601#define ANYOF_POSIXL_TEST_ALL_SET(p) \
93e92956 602 ((ANYOF_FLAGS(p) & ANYOF_MATCHES_POSIXL) \
e0e1be5f
KW
603 && ((regnode_charclass_posixl*) (p))->classflags \
604 == ((1U << ((ANYOF_POSIXL_MAX) - 1))) - 1)
fb38762f 605
b27fd97a 606#define ANYOF_POSIXL_OR(source, dest) STMT_START { (dest)->classflags |= (source)->classflags ; } STMT_END
8efd3f97 607#define ANYOF_CLASS_OR(source, dest) ANYOF_POSIXL_OR((source), (dest))
936ed897 608
cdd87c1d
KW
609#define ANYOF_POSIXL_AND(source, dest) STMT_START { (dest)->classflags &= (source)->classflags ; } STMT_END
610
9cf78fa0 611#define ANYOF_BITMAP_ZERO(ret) Zero(((struct regnode_charclass*)(ret))->bitmap, ANYOF_BITMAP_SIZE, char)
936ed897 612#define ANYOF_BITMAP(p) (((struct regnode_charclass*)(p))->bitmap)
786e8c11 613#define ANYOF_BITMAP_BYTE(p, c) (ANYOF_BITMAP(p)[(((U8)(c)) >> 3) & 31])
b8c5462f
JH
614#define ANYOF_BITMAP_SET(p, c) (ANYOF_BITMAP_BYTE(p, c) |= ANYOF_BIT(c))
615#define ANYOF_BITMAP_CLEAR(p,c) (ANYOF_BITMAP_BYTE(p, c) &= ~ANYOF_BIT(c))
bc51fd78 616#define ANYOF_BITMAP_TEST(p, c) cBOOL(ANYOF_BITMAP_BYTE(p, c) & ANYOF_BIT(c))
b8c5462f 617
f8bef550
NC
618#define ANYOF_BITMAP_SETALL(p) \
619 memset (ANYOF_BITMAP(p), 255, ANYOF_BITMAP_SIZE)
620#define ANYOF_BITMAP_CLEARALL(p) \
621 Zero (ANYOF_BITMAP(p), ANYOF_BITMAP_SIZE)
f8bef550 622
888ee784
KW
623#define ANYOF_SKIP EXTRA_SIZE(struct regnode_charclass)
624#define ANYOF_POSIXL_SKIP EXTRA_SIZE(regnode_charclass_posixl)
3a15e693 625
a687059c
LW
626/*
627 * Utility definitions.
628 */
2304df62 629#ifndef CHARMASK
075abff3 630# define UCHARAT(p) ((int)*(const U8*)(p))
a687059c 631#else
075abff3 632# define UCHARAT(p) ((int)*(p)&CHARMASK)
a687059c 633#endif
a687059c 634
c277df42
IZ
635#define EXTRA_SIZE(guy) ((sizeof(guy)-1)/sizeof(struct regnode))
636
e384d5c1
YO
637#define REG_ZERO_LEN_SEEN 0x00000001
638#define REG_LOOKBEHIND_SEEN 0x00000002
639#define REG_GPOS_SEEN 0x00000004
471f5387 640/* spare */
e384d5c1
YO
641#define REG_RECURSE_SEEN 0x00000020
642#define REG_TOP_LEVEL_BRANCHES_SEEN 0x00000040
643#define REG_VERBARG_SEEN 0x00000080
644#define REG_CUTGROUP_SEEN 0x00000100
645#define REG_RUN_ON_COMMENT_SEEN 0x00000200
646#define REG_UNFOLDED_MULTI_SEEN 0x00000400
647#define REG_GOSTART_SEEN 0x00000800
ee273784
YO
648#define REG_UNBOUNDED_QUANTIFIER_SEEN 0x00001000
649
d09b2d29 650
73c4f7a1
GS
651START_EXTERN_C
652
be8e71aa
YO
653#ifdef PLUGGABLE_RE_EXTENSION
654#include "re_nodes.h"
655#else
d09b2d29 656#include "regnodes.h"
be8e71aa 657#endif
d09b2d29 658
f9f4320a
YO
659#ifndef PLUGGABLE_RE_EXTENSION
660#ifndef DOINIT
661EXTCONST regexp_engine PL_core_reg_engine;
662#else /* DOINIT */
663EXTCONST regexp_engine PL_core_reg_engine = {
fe578d7f 664 Perl_re_compile,
2fdbfb4d 665 Perl_regexec_flags,
f9f4320a
YO
666 Perl_re_intuit_start,
667 Perl_re_intuit_string,
2fdbfb4d
AB
668 Perl_regfree_internal,
669 Perl_reg_numbered_buff_fetch,
670 Perl_reg_numbered_buff_store,
671 Perl_reg_numbered_buff_length,
192b9cd1
AB
672 Perl_reg_named_buff,
673 Perl_reg_named_buff_iter,
49d7dfbc 674 Perl_reg_qr_package,
f9f4320a 675#if defined(USE_ITHREADS)
3c13cae6 676 Perl_regdupe_internal,
f9f4320a 677#endif
3c13cae6 678 Perl_re_op_compile
f9f4320a
YO
679};
680#endif /* DOINIT */
681#endif /* PLUGGABLE_RE_EXTENSION */
682
683
73c4f7a1 684END_EXTERN_C
cad2e5aa 685
cad2e5aa 686
261faec3
MJD
687/* .what is a character array with one character for each member of .data
688 * The character describes the function of the corresponding .data item:
af534a04 689 * a - AV for paren_name_list under DEBUGGING
0111c4fd 690 * f - start-class data for regstclass optimization
68e2671b 691 * l - start op for literal (?{EVAL}) item
d63c20f2 692 * L - start op for literal (?{EVAL}) item, with separate CV (qr//)
b30fcab9 693 * r - pointer to an embedded code-containing qr, e.g. /ab$qr/
38a44b82 694 * s - swash for Unicode-style character class, and the multicharacter
9e55ce06
JH
695 * strings resulting from casefolding the single-character entries
696 * in the character class
a3621e74 697 * t - trie struct
55eed653 698 * u - trie struct's widecharmap (a HV, so can't share, must dup)
2b8b4781 699 * also used for revcharmap and words under DEBUGGING
07be1b83 700 * T - aho-trie struct
81714fb9 701 * S - sv for named capture lookup
261faec3 702 * 20010712 mjd@plover.com
c3613ed7 703 * (Remember to update re_dup() and pregfree() if you add any items.)
261faec3 704 */
cad2e5aa
JH
705struct reg_data {
706 U32 count;
707 U8 *what;
708 void* data[1];
709};
710
a1cac82e
NC
711/* Code in S_to_utf8_substr() and S_to_byte_substr() in regexec.c accesses
712 anchored* and float* via array indexes 0 and 1. */
cad2e5aa 713#define anchored_substr substrs->data[0].substr
33b8afdf 714#define anchored_utf8 substrs->data[0].utf8_substr
cad2e5aa 715#define anchored_offset substrs->data[0].min_offset
1de06328
YO
716#define anchored_end_shift substrs->data[0].end_shift
717
cad2e5aa 718#define float_substr substrs->data[1].substr
33b8afdf 719#define float_utf8 substrs->data[1].utf8_substr
cad2e5aa
JH
720#define float_min_offset substrs->data[1].min_offset
721#define float_max_offset substrs->data[1].max_offset
1de06328
YO
722#define float_end_shift substrs->data[1].end_shift
723
cad2e5aa 724#define check_substr substrs->data[2].substr
33b8afdf 725#define check_utf8 substrs->data[2].utf8_substr
cad2e5aa
JH
726#define check_offset_min substrs->data[2].min_offset
727#define check_offset_max substrs->data[2].max_offset
1de06328 728#define check_end_shift substrs->data[2].end_shift
a3621e74 729
8d919b0a
FC
730#define RX_ANCHORED_SUBSTR(rx) (ReANY(rx)->anchored_substr)
731#define RX_ANCHORED_UTF8(rx) (ReANY(rx)->anchored_utf8)
732#define RX_FLOAT_SUBSTR(rx) (ReANY(rx)->float_substr)
733#define RX_FLOAT_UTF8(rx) (ReANY(rx)->float_utf8)
a3621e74
YO
734
735/* trie related stuff */
5d9a96ca 736
a3621e74
YO
737/* a transition record for the state machine. the
738 check field determines which state "owns" the
739 transition. the char the transition is for is
740 determined by offset from the owning states base
741 field. the next field determines which state
742 is to be transitioned to if any.
743*/
744struct _reg_trie_trans {
745 U32 next;
746 U32 check;
747};
748
749/* a transition list element for the list based representation */
750struct _reg_trie_trans_list_elem {
751 U16 forid;
752 U32 newstate;
753};
754typedef struct _reg_trie_trans_list_elem reg_trie_trans_le;
755
756/* a state for compressed nodes. base is an offset
757 into an array of reg_trie_trans array. If wordnum is
758 nonzero the state is accepting. if base is zero then
759 the state has no children (and will be accepting)
760*/
761struct _reg_trie_state {
762 U16 wordnum;
763 union {
764 U32 base;
765 reg_trie_trans_le* list;
766 } trans;
767};
768
2e64971a
DM
769/* info per word; indexed by wordnum */
770typedef struct {
771 U16 prev; /* previous word in acceptance chain; eg in
772 * zzz|abc|ab/ after matching the chars abc, the
773 * accepted word is #2, and the previous accepted
774 * word is #3 */
775 U32 len; /* how many chars long is this word? */
776 U32 accept; /* accept state for this word */
777} reg_trie_wordinfo;
a3621e74
YO
778
779
a3621e74
YO
780typedef struct _reg_trie_state reg_trie_state;
781typedef struct _reg_trie_trans reg_trie_trans;
782
783
784/* anything in here that needs to be freed later
23eab42c
NC
785 should be dealt with in pregfree.
786 refcount is first in both this and _reg_ac_data to allow a space
787 optimisation in Perl_regdupe. */
a3621e74 788struct _reg_trie_data {
23eab42c 789 U32 refcount; /* number of times this trie is referenced */
786e8c11
YO
790 U32 lasttrans; /* last valid transition element */
791 U16 *charmap; /* byte to charid lookup array */
786e8c11
YO
792 reg_trie_state *states; /* state data */
793 reg_trie_trans *trans; /* array of transition elements */
794 char *bitmap; /* stclass bitmap */
786e8c11
YO
795 U16 *jump; /* optional 1 indexed array of offsets before tail
796 for the node following a given word. */
2e64971a 797 reg_trie_wordinfo *wordinfo; /* array of info per word */
1b5c067c
NC
798 U16 uniquecharcount; /* unique chars in trie (width of trans table) */
799 U32 startstate; /* initial state - used for common prefix optimisation */
800 STRLEN minlen; /* minimum length of words in trie - build/opt only? */
801 STRLEN maxlen; /* maximum length of words in trie - build/opt only? */
2e64971a 802 U32 prefixlen; /* #chars in common prefix */
1e2e3d02
YO
803 U32 statecount; /* Build only - number of states in the states array
804 (including the unused zero state) */
786e8c11 805 U32 wordcount; /* Build only */
a3621e74 806#ifdef DEBUGGING
786e8c11 807 STRLEN charcount; /* Build only */
a3621e74
YO
808#endif
809};
2b8b4781
NC
810/* There is one (3 under DEBUGGING) pointers that logically belong in this
811 structure, but are held outside as they need duplication on thread cloning,
812 whereas the rest of the structure can be read only:
813 HV *widecharmap; code points > 255 to charid
814#ifdef DEBUGGING
815 AV *words; Array of words contained in trie, for dumping
816 AV *revcharmap; Map of each charid back to its character representation
817#endif
818*/
819
820#define TRIE_WORDS_OFFSET 2
821
a3621e74
YO
822typedef struct _reg_trie_data reg_trie_data;
823
23eab42c
NC
824/* refcount is first in both this and _reg_trie_data to allow a space
825 optimisation in Perl_regdupe. */
07be1b83 826struct _reg_ac_data {
23eab42c 827 U32 refcount;
1b5c067c 828 U32 trie;
07be1b83
YO
829 U32 *fail;
830 reg_trie_state *states;
07be1b83
YO
831};
832typedef struct _reg_ac_data reg_ac_data;
833
486ec47a 834/* ANY_BIT doesn't use the structure, so we can borrow it here.
3dab1dad
YO
835 This is simpler than refactoring all of it as wed end up with
836 three different sets... */
837
838#define TRIE_BITMAP(p) (((reg_trie_data *)(p))->bitmap)
786e8c11 839#define TRIE_BITMAP_BYTE(p, c) (TRIE_BITMAP(p)[(((U8)(c)) >> 3) & 31])
5dfe062f
YO
840#define TRIE_BITMAP_SET(p, c) (TRIE_BITMAP_BYTE(p, c) |= ANYOF_BIT((U8)c))
841#define TRIE_BITMAP_CLEAR(p,c) (TRIE_BITMAP_BYTE(p, c) &= ~ANYOF_BIT((U8)c))
842#define TRIE_BITMAP_TEST(p, c) (TRIE_BITMAP_BYTE(p, c) & ANYOF_BIT((U8)c))
3dab1dad 843
1de06328
YO
844#define IS_ANYOF_TRIE(op) ((op)==TRIEC || (op)==AHOCORASICKC)
845#define IS_TRIE_AC(op) ((op)>=AHOCORASICK)
846
847
786e8c11
YO
848#define BITMAP_BYTE(p, c) (((U8*)p)[(((U8)(c)) >> 3) & 31])
849#define BITMAP_TEST(p, c) (BITMAP_BYTE(p, c) & ANYOF_BIT((U8)c))
3dab1dad 850
a3621e74
YO
851/* these defines assume uniquecharcount is the correct variable, and state may be evaluated twice */
852#define TRIE_NODENUM(state) (((state)-1)/(trie->uniquecharcount)+1)
853#define SAFE_TRIE_NODENUM(state) ((state) ? (((state)-1)/(trie->uniquecharcount)+1) : (state))
854#define TRIE_NODEIDX(state) ((state) ? (((state)-1)*(trie->uniquecharcount)+1) : (state))
855
3dab1dad 856#ifdef DEBUGGING
3dab1dad 857#define TRIE_CHARCOUNT(trie) ((trie)->charcount)
3dab1dad 858#else
3dab1dad 859#define TRIE_CHARCOUNT(trie) (trie_charcount)
3dab1dad 860#endif
a3621e74 861
0111c4fd
RGS
862#define RE_TRIE_MAXBUF_INIT 65536
863#define RE_TRIE_MAXBUF_NAME "\022E_TRIE_MAXBUF"
a3621e74
YO
864#define RE_DEBUG_FLAGS "\022E_DEBUG_FLAGS"
865
be8e71aa
YO
866/*
867
868RE_DEBUG_FLAGS is used to control what debug output is emitted
869its divided into three groups of options, some of which interact.
870The three groups are: Compile, Execute, Extra. There is room for a
871further group, as currently only the low three bytes are used.
872
873 Compile Options:
07be1b83 874
be8e71aa
YO
875 PARSE
876 PEEP
877 TRIE
878 PROGRAM
879 OFFSETS
a3621e74 880
be8e71aa 881 Execute Options:
a3621e74 882
be8e71aa
YO
883 INTUIT
884 MATCH
885 TRIE
a3621e74 886
be8e71aa 887 Extra Options
a3621e74 888
be8e71aa
YO
889 TRIE
890 OFFSETS
891
892If you modify any of these make sure you make corresponding changes to
893re.pm, especially to the documentation.
a3621e74 894
be8e71aa
YO
895*/
896
897
898/* Compile */
899#define RE_DEBUG_COMPILE_MASK 0x0000FF
900#define RE_DEBUG_COMPILE_PARSE 0x000001
901#define RE_DEBUG_COMPILE_OPTIMISE 0x000002
902#define RE_DEBUG_COMPILE_TRIE 0x000004
903#define RE_DEBUG_COMPILE_DUMP 0x000008
f7819f85 904#define RE_DEBUG_COMPILE_FLAGS 0x000010
d9a72fcc 905#define RE_DEBUG_COMPILE_TEST 0x000020
be8e71aa
YO
906
907/* Execute */
908#define RE_DEBUG_EXECUTE_MASK 0x00FF00
909#define RE_DEBUG_EXECUTE_INTUIT 0x000100
910#define RE_DEBUG_EXECUTE_MATCH 0x000200
911#define RE_DEBUG_EXECUTE_TRIE 0x000400
912
913/* Extra */
914#define RE_DEBUG_EXTRA_MASK 0xFF0000
915#define RE_DEBUG_EXTRA_TRIE 0x010000
916#define RE_DEBUG_EXTRA_OFFSETS 0x020000
a5ca303d
YO
917#define RE_DEBUG_EXTRA_OFFDEBUG 0x040000
918#define RE_DEBUG_EXTRA_STATE 0x080000
919#define RE_DEBUG_EXTRA_OPTIMISE 0x100000
e7707071 920#define RE_DEBUG_EXTRA_BUFFERS 0x400000
2c296965 921#define RE_DEBUG_EXTRA_GPOS 0x800000
24b23f37
YO
922/* combined */
923#define RE_DEBUG_EXTRA_STACK 0x280000
be8e71aa 924
e68ec53f 925#define RE_DEBUG_FLAG(x) (re_debug_flags & x)
be8e71aa
YO
926/* Compile */
927#define DEBUG_COMPILE_r(x) DEBUG_r( \
e68ec53f 928 if (re_debug_flags & RE_DEBUG_COMPILE_MASK) x )
be8e71aa 929#define DEBUG_PARSE_r(x) DEBUG_r( \
e68ec53f 930 if (re_debug_flags & RE_DEBUG_COMPILE_PARSE) x )
be8e71aa 931#define DEBUG_OPTIMISE_r(x) DEBUG_r( \
e68ec53f 932 if (re_debug_flags & RE_DEBUG_COMPILE_OPTIMISE) x )
be8e71aa 933#define DEBUG_PARSE_r(x) DEBUG_r( \
e68ec53f 934 if (re_debug_flags & RE_DEBUG_COMPILE_PARSE) x )
be8e71aa 935#define DEBUG_DUMP_r(x) DEBUG_r( \
e68ec53f 936 if (re_debug_flags & RE_DEBUG_COMPILE_DUMP) x )
be8e71aa 937#define DEBUG_TRIE_COMPILE_r(x) DEBUG_r( \
e68ec53f 938 if (re_debug_flags & RE_DEBUG_COMPILE_TRIE) x )
f7819f85
A
939#define DEBUG_FLAGS_r(x) DEBUG_r( \
940 if (re_debug_flags & RE_DEBUG_COMPILE_FLAGS) x )
d9a72fcc
YO
941#define DEBUG_TEST_r(x) DEBUG_r( \
942 if (re_debug_flags & RE_DEBUG_COMPILE_TEST) x )
be8e71aa
YO
943/* Execute */
944#define DEBUG_EXECUTE_r(x) DEBUG_r( \
e68ec53f 945 if (re_debug_flags & RE_DEBUG_EXECUTE_MASK) x )
be8e71aa 946#define DEBUG_INTUIT_r(x) DEBUG_r( \
e68ec53f 947 if (re_debug_flags & RE_DEBUG_EXECUTE_INTUIT) x )
be8e71aa 948#define DEBUG_MATCH_r(x) DEBUG_r( \
e68ec53f 949 if (re_debug_flags & RE_DEBUG_EXECUTE_MATCH) x )
be8e71aa 950#define DEBUG_TRIE_EXECUTE_r(x) DEBUG_r( \
e68ec53f 951 if (re_debug_flags & RE_DEBUG_EXECUTE_TRIE) x )
be8e71aa
YO
952
953/* Extra */
954#define DEBUG_EXTRA_r(x) DEBUG_r( \
e68ec53f 955 if (re_debug_flags & RE_DEBUG_EXTRA_MASK) x )
a5ca303d
YO
956#define DEBUG_OFFSETS_r(x) DEBUG_r( \
957 if (re_debug_flags & RE_DEBUG_EXTRA_OFFSETS) x )
ab3bbdeb
YO
958#define DEBUG_STATE_r(x) DEBUG_r( \
959 if (re_debug_flags & RE_DEBUG_EXTRA_STATE) x )
24b23f37
YO
960#define DEBUG_STACK_r(x) DEBUG_r( \
961 if (re_debug_flags & RE_DEBUG_EXTRA_STACK) x )
e7707071
YO
962#define DEBUG_BUFFERS_r(x) DEBUG_r( \
963 if (re_debug_flags & RE_DEBUG_EXTRA_BUFFERS) x )
964
a5ca303d
YO
965#define DEBUG_OPTIMISE_MORE_r(x) DEBUG_r( \
966 if ((RE_DEBUG_EXTRA_OPTIMISE|RE_DEBUG_COMPILE_OPTIMISE) == \
967 (re_debug_flags & (RE_DEBUG_EXTRA_OPTIMISE|RE_DEBUG_COMPILE_OPTIMISE)) ) x )
07be1b83 968#define MJD_OFFSET_DEBUG(x) DEBUG_r( \
a5ca303d 969 if (re_debug_flags & RE_DEBUG_EXTRA_OFFDEBUG) \
be8e71aa
YO
970 Perl_warn_nocontext x )
971#define DEBUG_TRIE_COMPILE_MORE_r(x) DEBUG_TRIE_COMPILE_r( \
e68ec53f 972 if (re_debug_flags & RE_DEBUG_EXTRA_TRIE) x )
be8e71aa 973#define DEBUG_TRIE_EXECUTE_MORE_r(x) DEBUG_TRIE_EXECUTE_r( \
e68ec53f 974 if (re_debug_flags & RE_DEBUG_EXTRA_TRIE) x )
be8e71aa
YO
975
976#define DEBUG_TRIE_r(x) DEBUG_r( \
e68ec53f 977 if (re_debug_flags & (RE_DEBUG_COMPILE_TRIE \
be8e71aa 978 | RE_DEBUG_EXECUTE_TRIE )) x )
2c296965
YO
979#define DEBUG_GPOS_r(x) DEBUG_r( \
980 if (re_debug_flags & RE_DEBUG_EXTRA_GPOS) x )
be8e71aa
YO
981
982/* initialization */
d2d8c711 983/* get_sv() can return NULL during global destruction. */
e68ec53f
YO
984#define GET_RE_DEBUG_FLAGS DEBUG_r({ \
985 SV * re_debug_flags_sv = NULL; \
986 re_debug_flags_sv = get_sv(RE_DEBUG_FLAGS, 1); \
987 if (re_debug_flags_sv) { \
8e9a8a48 988 if (!SvIOK(re_debug_flags_sv)) \
e68ec53f 989 sv_setuv(re_debug_flags_sv, RE_DEBUG_COMPILE_DUMP | RE_DEBUG_EXECUTE_MASK ); \
8e9a8a48 990 re_debug_flags=SvIV(re_debug_flags_sv); \
e68ec53f
YO
991 }\
992})
a3621e74
YO
993
994#ifdef DEBUGGING
ab3bbdeb 995
cb65b687
DM
996#define GET_RE_DEBUG_FLAGS_DECL VOL IV re_debug_flags = 0; \
997 PERL_UNUSED_VAR(re_debug_flags); GET_RE_DEBUG_FLAGS;
ab3bbdeb
YO
998
999#define RE_PV_COLOR_DECL(rpv,rlen,isuni,dsv,pv,l,m,c1,c2) \
1000 const char * const rpv = \
ddc5bc0f 1001 pv_pretty((dsv), (pv), (l), (m), \
ab3bbdeb 1002 PL_colors[(c1)],PL_colors[(c2)], \
c89df6cf 1003 PERL_PV_ESCAPE_RE|PERL_PV_ESCAPE_NONASCII |((isuni) ? PERL_PV_ESCAPE_UNI : 0) ); \
0df25f3d 1004 const int rlen = SvCUR(dsv)
ab3bbdeb
YO
1005
1006#define RE_SV_ESCAPE(rpv,isuni,dsv,sv,m) \
1007 const char * const rpv = \
ddc5bc0f 1008 pv_pretty((dsv), (SvPV_nolen_const(sv)), (SvCUR(sv)), (m), \
ab3bbdeb 1009 PL_colors[(c1)],PL_colors[(c2)], \
c89df6cf 1010 PERL_PV_ESCAPE_RE|PERL_PV_ESCAPE_NONASCII |((isuni) ? PERL_PV_ESCAPE_UNI : 0) )
ab3bbdeb
YO
1011
1012#define RE_PV_QUOTED_DECL(rpv,isuni,dsv,pv,l,m) \
1013 const char * const rpv = \
ddc5bc0f 1014 pv_pretty((dsv), (pv), (l), (m), \
ab3bbdeb 1015 PL_colors[0], PL_colors[1], \
c89df6cf 1016 ( PERL_PV_PRETTY_QUOTE | PERL_PV_ESCAPE_RE | PERL_PV_ESCAPE_NONASCII | PERL_PV_PRETTY_ELLIPSES | \
ab3bbdeb 1017 ((isuni) ? PERL_PV_ESCAPE_UNI : 0)) \
0a0e9afc 1018 )
ab3bbdeb
YO
1019
1020#define RE_SV_DUMPLEN(ItEm) (SvCUR(ItEm) - (SvTAIL(ItEm)!=0))
1021#define RE_SV_TAIL(ItEm) (SvTAIL(ItEm) ? "$" : "")
1022
1023#else /* if not DEBUGGING */
1024
a3621e74 1025#define GET_RE_DEBUG_FLAGS_DECL
ab3bbdeb
YO
1026#define RE_PV_COLOR_DECL(rpv,rlen,isuni,dsv,pv,l,m,c1,c2)
1027#define RE_SV_ESCAPE(rpv,isuni,dsv,sv,m)
1028#define RE_PV_QUOTED_DECL(rpv,isuni,dsv,pv,l,m)
1029#define RE_SV_DUMPLEN(ItEm)
1030#define RE_SV_TAIL(ItEm)
1031
1032#endif /* DEBUG RELATED DEFINES */
a3621e74 1033
64935bc6
KW
1034typedef enum {
1035 TRADITIONAL_BOUND = _CC_WORDCHAR,
ae3bb8ea 1036 GCB_BOUND,
06ae2722 1037 SB_BOUND,
ae3bb8ea 1038 WB_BOUND
64935bc6
KW
1039} bound_type;
1040
e9a8c099 1041/*
14d04a33 1042 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 1043 */