This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
regcomp.h: Use handy.h constants
[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
b9b4dddf
YO
29/* Unless the next line is uncommented it is illegal to combine lazy
30 matching with possessive matching. Frankly it doesn't make much sense
31 to allow it as X*?+ matches nothing, X+?+ matches a single char only,
32 and X{min,max}?+ matches min times only.
33 */
34/* #define REG_ALLOW_MINMOD_SUSPEND */
07be1b83 35
a687059c
LW
36/*
37 * The "internal use only" fields in regexp.h are present to pass info from
38 * compile to execute that permits the execute phase to run lots faster on
39 * simple cases. They are:
40 *
bd61b366 41 * regstart sv that must begin a match; NULL if none obvious
a687059c
LW
42 * reganch is the match anchored (at beginning-of-line only)?
43 * regmust string (pointer into program) that match must include, or NULL
79072805 44 * [regmust changed to SV* for bminstr()--law]
a687059c
LW
45 * regmlen length of regmust string
46 * [regmlen not used currently]
47 *
48 * Regstart and reganch permit very fast decisions on suitable starting points
49 * for a match, cutting down the work a lot. Regmust permits fast rejection
50 * of lines that cannot possibly match. The regmust tests are costly enough
e50aee73 51 * that pregcomp() supplies a regmust only if the r.e. contains something
a687059c
LW
52 * potentially expensive (at present, the only such thing detected is * or +
53 * at the start of the r.e., which can involve a lot of backup). Regmlen is
e50aee73 54 * supplied because the test in pregexec() needs it and pregcomp() is computing
a687059c
LW
55 * it anyway.
56 * [regmust is now supplied always. The tests that use regmust have a
57 * heuristic that disables the test if it usually matches.]
58 *
59 * [In fact, we now use regmust in many cases to locate where the search
60 * starts in the string, so if regback is >= 0, the regmust search is never
61 * wasted effort. The regback variable says how many characters back from
62 * where regmust matched is the earliest possible start of the match.
63 * For instance, /[a-z].foo/ has a regmust of 'foo' and a regback of 2.]
64 */
65
66/*
67 * Structure for regexp "program". This is essentially a linear encoding
68 * of a nondeterministic finite-state machine (aka syntax charts or
69 * "railroad normal form" in parsing technology). Each node is an opcode
70 * plus a "next" pointer, possibly plus an operand. "Next" pointers of
71 * all nodes except BRANCH implement concatenation; a "next" pointer with
72 * a BRANCH on both ends of it is connecting two alternatives. (Here we
73 * have one of the subtle syntax dependencies: an individual BRANCH (as
74 * opposed to a collection of them) is never concatenated with anything
75 * because of operator precedence.) The operand of some types of node is
76 * a literal string; for others, it is a node leading into a sub-FSM. In
77 * particular, the operand of a BRANCH node is the first node of the branch.
78 * (NB this is *not* a tree structure: the tail of the branch connects
a3621e74
YO
79 * to the thing following the set of BRANCHes.) The opcodes are defined
80 * in regnodes.h which is generated from regcomp.sym by regcomp.pl.
a687059c
LW
81 */
82
a687059c
LW
83/*
84 * A node is one char of opcode followed by two chars of "next" pointer.
85 * "Next" pointers are stored as two 8-bit pieces, high order first. The
86 * value is a positive offset from the opcode of the node containing it.
87 * An operand, if any, simply follows the node. (Note that much of the
88 * code generation knows about this implicit relationship.)
89 *
90 * Using two bytes for the "next" pointer is vast overkill for most things,
91 * but allows patterns to get big without disasters.
92 *
e91177ed 93 * [The "next" pointer is always aligned on an even
a687059c
LW
94 * boundary, and reads the offset directly as a short. Also, there is no
95 * special test to reverse the sign of BACK pointers since the offset is
96 * stored negative.]
97 */
785a26d5
YO
98
99/* This is the stuff that used to live in regexp.h that was truly
100 private to the engine itself. It now lives here. */
101
28d8d7f4 102
785a26d5
YO
103
104 typedef struct regexp_internal {
1f1031fe 105 int name_list_idx; /* Optional data index of an array of paren names */
7122b237
YO
106 union {
107 U32 *offsets; /* offset annotations 20001228 MJD
108 data about mapping the program to the
109 string -
110 offsets[0] is proglen when this is used
111 */
112 U32 proglen;
113 } u;
114
6d5c990f
RGS
115 regnode *regstclass; /* Optional startclass as identified or constructed
116 by the optimiser */
117 struct reg_data *data; /* Additional miscellaneous data used by the program.
118 Used to make it easier to clone and free arbitrary
119 data that the regops need. Often the ARG field of
120 a regop is an index into this structure */
3d2bd50a
DM
121 struct reg_code_block *code_blocks;/* positions of literal (?{}) */
122 int num_code_blocks; /* size of code_blocks[] */
6d5c990f
RGS
123 regnode program[1]; /* Unwarranted chumminess with compiler. */
124} regexp_internal;
125
126#define RXi_SET(x,y) (x)->pprivate = (void*)(y)
127#define RXi_GET(x) ((regexp_internal *)((x)->pprivate))
128#define RXi_GET_DECL(r,ri) regexp_internal *ri = RXi_GET(r)
785a26d5
YO
129/*
130 * Flags stored in regexp->intflags
131 * These are used only internally to the regexp engine
132 *
133 * See regexp.h for flags used externally to the regexp engine
134 */
135#define PREGf_SKIP 0x00000001
136#define PREGf_IMPLICIT 0x00000002 /* Converted .* to ^.* */
137#define PREGf_NAUGHTY 0x00000004 /* how exponential is this pattern? */
138#define PREGf_VERBARG_SEEN 0x00000008
139#define PREGf_CUTGROUP_SEEN 0x00000010
732caac7 140#define PREGf_USE_RE_EVAL 0x00000020 /* compiled with "use re 'eval'" */
785a26d5
YO
141
142
143/* this is where the old regcomp.h started */
a687059c 144
c277df42 145struct regnode_string {
cd439c50 146 U8 str_len;
c277df42
IZ
147 U8 type;
148 U16 next_off;
cd439c50 149 char string[1];
c277df42 150};
a687059c 151
6bda09f9
YO
152/* Argument bearing node - workhorse,
153 arg1 is often for the data field */
c277df42
IZ
154struct regnode_1 {
155 U8 flags;
156 U8 type;
157 U16 next_off;
158 U32 arg1;
159};
160
6bda09f9
YO
161/* Similar to a regnode_1 but with an extra signed argument */
162struct regnode_2L {
163 U8 flags;
164 U8 type;
165 U16 next_off;
166 U32 arg1;
167 I32 arg2;
168};
169
170/* 'Two field' -- Two 16 bit unsigned args */
c277df42
IZ
171struct regnode_2 {
172 U8 flags;
173 U8 type;
174 U16 next_off;
175 U16 arg1;
176 U16 arg2;
177};
178
6bda09f9 179
936ed897 180#define ANYOF_BITMAP_SIZE 32 /* 256 b/(8 b/B) */
d1eb3177 181#define ANYOF_CLASSBITMAP_SIZE 4 /* up to 32 (8*4) named classes */
936ed897 182
786e8c11 183/* also used by trie */
936ed897
IZ
184struct regnode_charclass {
185 U8 flags;
186 U8 type;
187 U16 next_off;
bf3f2d85 188 U32 arg1; /* used as ptr in S_regclass */
19103b91 189 char bitmap[ANYOF_BITMAP_SIZE]; /* only compile-time */
936ed897
IZ
190};
191
bf3f2d85
KW
192/* has runtime (locale) \d, \w, ..., [:posix:] classes */
193struct regnode_charclass_class {
194 U8 flags; /* ANYOF_CLASS bit must go here */
936ed897
IZ
195 U8 type;
196 U16 next_off;
bf3f2d85 197 U32 arg1; /* used as ptr in S_regclass */
19103b91
JH
198 char bitmap[ANYOF_BITMAP_SIZE]; /* both compile-time */
199 char classflags[ANYOF_CLASSBITMAP_SIZE]; /* and run-time */
936ed897
IZ
200};
201
9c7e81e8
AD
202/* XXX fix this description.
203 Impose a limit of REG_INFTY on various pattern matching operations
204 to limit stack growth and to avoid "infinite" recursions.
205*/
206/* The default size for REG_INFTY is I16_MAX, which is the same as
207 SHORT_MAX (see perl.h). Unfortunately I16 isn't necessarily 16 bits
208 (see handy.h). On the Cray C90, sizeof(short)==4 and hence I16_MAX is
209 ((1<<31)-1), while on the Cray T90, sizeof(short)==8 and I16_MAX is
210 ((1<<63)-1). To limit stack growth to reasonable sizes, supply a
211 smaller default.
212 --Andy Dougherty 11 June 1998
213*/
214#if SHORTSIZE > 2
215# ifndef REG_INFTY
216# define REG_INFTY ((1<<15)-1)
217# endif
218#endif
219
220#ifndef REG_INFTY
83cfe48c
IZ
221# define REG_INFTY I16_MAX
222#endif
a687059c 223
e91177ed
IZ
224#define ARG_VALUE(arg) (arg)
225#define ARG__SET(arg,val) ((arg) = (val))
c277df42 226
e8e3b0fb
JH
227#undef ARG
228#undef ARG1
229#undef ARG2
230
c277df42
IZ
231#define ARG(p) ARG_VALUE(ARG_LOC(p))
232#define ARG1(p) ARG_VALUE(ARG1_LOC(p))
233#define ARG2(p) ARG_VALUE(ARG2_LOC(p))
6bda09f9 234#define ARG2L(p) ARG_VALUE(ARG2L_LOC(p))
786e8c11 235
c277df42
IZ
236#define ARG_SET(p, val) ARG__SET(ARG_LOC(p), (val))
237#define ARG1_SET(p, val) ARG__SET(ARG1_LOC(p), (val))
238#define ARG2_SET(p, val) ARG__SET(ARG2_LOC(p), (val))
6bda09f9 239#define ARG2L_SET(p, val) ARG__SET(ARG2L_LOC(p), (val))
c277df42 240
e8e3b0fb
JH
241#undef NEXT_OFF
242#undef NODE_ALIGN
243
075abff3
AL
244#define NEXT_OFF(p) ((p)->next_off)
245#define NODE_ALIGN(node)
246#define NODE_ALIGN_FILL(node) ((node)->flags = 0xde) /* deadbeef */
a687059c 247
c277df42
IZ
248#define SIZE_ALIGN NODE_ALIGN
249
e8e3b0fb
JH
250#undef OP
251#undef OPERAND
252#undef MASK
253#undef STRING
254
e91177ed 255#define OP(p) ((p)->type)
b988e673
KW
256#define FLAGS(p) ((p)->flags) /* Caution: Doesn't apply to all \
257 regnode types. For some, it's the \
258 character set of the regnode */
e91177ed 259#define OPERAND(p) (((struct regnode_string *)p)->string)
cd439c50
IZ
260#define MASK(p) ((char*)OPERAND(p))
261#define STR_LEN(p) (((struct regnode_string *)p)->str_len)
262#define STRING(p) (((struct regnode_string *)p)->string)
263#define STR_SZ(l) ((l + sizeof(regnode) - 1) / sizeof(regnode))
264#define NODE_SZ_STR(p) (STR_SZ(STR_LEN(p))+1)
265
e8e3b0fb
JH
266#undef NODE_ALIGN
267#undef ARG_LOC
268#undef NEXTOPER
269#undef PREVOPER
270
e91177ed
IZ
271#define NODE_ALIGN(node)
272#define ARG_LOC(p) (((struct regnode_1 *)p)->arg1)
273#define ARG1_LOC(p) (((struct regnode_2 *)p)->arg1)
274#define ARG2_LOC(p) (((struct regnode_2 *)p)->arg2)
6bda09f9 275#define ARG2L_LOC(p) (((struct regnode_2L *)p)->arg2)
786e8c11 276
e91177ed
IZ
277#define NODE_STEP_REGNODE 1 /* sizeof(regnode)/sizeof(regnode) */
278#define EXTRA_STEP_2ARGS EXTRA_SIZE(struct regnode_2)
279
280#define NODE_STEP_B 4
c277df42
IZ
281
282#define NEXTOPER(p) ((p) + NODE_STEP_REGNODE)
283#define PREVOPER(p) ((p) - NODE_STEP_REGNODE)
284
e91177ed 285#define FILL_ADVANCE_NODE(ptr, op) STMT_START { \
c277df42 286 (ptr)->type = op; (ptr)->next_off = 0; (ptr)++; } STMT_END
e91177ed 287#define FILL_ADVANCE_NODE_ARG(ptr, op, arg) STMT_START { \
c277df42 288 ARG_SET(ptr, arg); FILL_ADVANCE_NODE(ptr, op); (ptr) += 1; } STMT_END
a687059c 289
22c35a8c 290#define REG_MAGIC 0234
a687059c 291
830247a4 292#define SIZE_ONLY (RExC_emit == &PL_regdummy)
c277df42 293
137165a6
KW
294/* If the bitmap doesn't fully represent what this ANYOF node can match, the
295 * ARG is set to this special value (since 0, 1, ... are legal, but will never
296 * reach this high). */
297#define ANYOF_NONBITMAP_EMPTY ((U32) -1)
298
299/* The information used to be stored as as combination of the ANYOF_UTF8 and
300 * ANYOF_NONBITMAP_NON_UTF8 bits in the flags field, but was moved out of there
301 * to free up a bit for other uses. This tries to hide the change from
302 * existing code as much as possible. Now, the data structure that goes in ARG
303 * is not allocated unless it is needed, and that is what is used to determine
304 * if there is something outside the bitmap. The code now assumes that if
305 * that structure exists, that any UTF-8 encoded string should be tried against
306 * it, but a non-UTF8-encoded string will be tried only if the
307 * ANYOF_NONBITMAP_NON_UTF8 bit is also set. */
308#define ANYOF_NONBITMAP(node) (ARG(node) != ANYOF_NONBITMAP_EMPTY)
309
d5788240 310/* Flags for node->flags of ANYOF. These are in short supply, so some games
8b27d3db
KW
311 * are done to share them, as described below. Already, the ANYOF_LOCALE and
312 * ANYOF_CLASS bits are shared, making a space penalty for all locale nodes.
313 * An option would be to push them into new nodes. E.g. there could be an
314 * ANYOF_LOCALE node that would be in place of the flag of the same name. But
315 * there are better options. The UNICODE_ALL bit could be freed up by
316 * resorting to creating a swash containing everything above 255. This
317 * introduces a performance penalty. Better would be to split it off into a
318 * separate node, which actually would improve performance by allowing adding a
319 * case statement to regexec.c use the bit map for code points under 256, and
320 * to match everything above. If flags need to be added that are applicable to
321 * the synthetic start class only, with some work, they could be put in the
322 * next-node field, or in an unused bit of the classflags field. This could be
323 * done with the current EOS flag, and a new node type created that is just for
324 * the scc, freeing up that bit */
b8c5462f 325
a4e790c1 326#define ANYOF_LOCALE 0x01 /* /l modifier */
39065660
KW
327
328/* The fold is calculated and stored in the bitmap where possible at compile
329 * time. However there are two cases where it isn't possible. These share
330 * this bit: 1) under locale, where the actual folding varies depending on
331 * what the locale is at the time of execution; and 2) where the folding is
332 * specified in a swash, not the bitmap, such as characters which aren't
333 * specified in the bitmap, or properties that aren't looked at at compile time
334 */
335#define ANYOF_LOC_NONBITMAP_FOLD 0x02
336
337#define ANYOF_INVERT 0x04
1984f6a0 338
dd58aee1
KW
339/* Set if this is a struct regnode_charclass_class vs a regnode_charclass. This
340 * is used for runtime \d, \w, [:posix:], ..., which are used only in locale
341 * and the optimizer's synthetic start class. Non-locale \d, etc are resolved
8b27d3db
KW
342 * at compile-time. Now shared with ANYOF_LOCALE, forcing all locale nodes to
343 * be large */
344#define ANYOF_CLASS ANYOF_LOCALE
c8d3cd88
KW
345#define ANYOF_LARGE ANYOF_CLASS /* Same; name retained for back compat */
346
5073ffbd
KW
347/* Should this character class warn if matched against a character above
348 * Unicode */
349#define ANYOF_WARN_SUPER 0x08
350
f4244008 351/* EOS, meaning that it can match an empty string too, is used for the
c613755a 352 * synthetic start class only. */
84193928 353#define ANYOF_EOS 0x10
f4244008 354
c613755a
KW
355/* ? Is this node the synthetic start class (ssc). This bit is shared with
356 * ANYOF_EOS, as the latter is used only for the ssc, and then not used by
357 * regexec.c. And, the code is structured so that if it is set, the ssc is
358 * not used, so it is guaranteed to be 0 for the ssc by the time regexec.c
359 * gets executed, and 0 for a non-ssc ANYOF node, as it only ever gets set for
360 * a potential ssc candidate. Thus setting it to 1 after it has been
361 * determined that the ssc will be used is not ambiguous */
362#define ANYOF_IS_SYNTHETIC ANYOF_EOS
363
ef87b810
KW
364/* Can match something outside the bitmap that isn't in utf8 */
365#define ANYOF_NONBITMAP_NON_UTF8 0x20
366
61ef9872
KW
367/* Matches every code point 0x100 and above*/
368#define ANYOF_UNICODE_ALL 0x40
ffc61ed2 369
11454c59
KW
370/* Match all Latin1 characters that aren't ASCII when the target string is not
371 * in utf8. */
372#define ANYOF_NON_UTF8_LATIN1_ALL 0x80
373
ffc61ed2 374#define ANYOF_FLAGS_ALL 0xff
653099ff 375
ace6b0e4
KW
376/* These are the flags that ANYOF_INVERT being set or not doesn't affect
377 * whether they are operative or not. e.g., the node still has LOCALE
378 * regardless of being inverted; whereas ANYOF_UNICODE_ALL means something
379 * different if inverted */
380#define INVERSION_UNAFFECTED_FLAGS (ANYOF_LOCALE \
381 |ANYOF_LOC_NONBITMAP_FOLD \
382 |ANYOF_CLASS \
383 |ANYOF_EOS \
384 |ANYOF_NONBITMAP_NON_UTF8)
385
936ed897 386/* Character classes for node->classflags of ANYOF */
653099ff 387/* Should be synchronized with a table in regprop() */
575a8a29 388/* 2n should be the normal one, paired with its complement at 2n+1 */
b8c5462f 389
08ada29e
KW
390#define ANYOF_ALNUM ((_CC_WORDCHAR) * 2) /* \w, PL_utf8_alnum, utf8::IsWord, ALNUM */
391#define ANYOF_NALNUM ((ANYOF_ALNUM) + 1)
392#define ANYOF_SPACE ((_CC_SPACE) * 2) /* \s */
393#define ANYOF_NSPACE ((ANYOF_SPACE) + 1)
394#define ANYOF_DIGIT ((_CC_DIGIT) * 2) /* \d */
395#define ANYOF_NDIGIT ((ANYOF_DIGIT) + 1)
396#define ANYOF_ALNUMC ((_CC_ALNUMC) * 2) /* [[:alnum:]] isalnum(3), utf8::IsAlnum, ALNUMC */
397#define ANYOF_NALNUMC ((ANYOF_ALNUMC) + 1)
398#define ANYOF_ALPHA ((_CC_ALPHA) * 2)
399#define ANYOF_NALPHA ((ANYOF_ALPHA) + 1)
400#define ANYOF_ASCII ((_CC_ASCII) * 2)
401#define ANYOF_NASCII ((ANYOF_ASCII) + 1)
402#define ANYOF_CNTRL ((_CC_CNTRL) * 2)
403#define ANYOF_NCNTRL ((ANYOF_CNTRL) + 1)
404#define ANYOF_GRAPH ((_CC_GRAPH) * 2)
405#define ANYOF_NGRAPH ((ANYOF_GRAPH) + 1)
406#define ANYOF_LOWER ((_CC_LOWER) * 2)
407#define ANYOF_NLOWER ((ANYOF_LOWER) + 1)
408#define ANYOF_PRINT ((_CC_PRINT) * 2)
409#define ANYOF_NPRINT ((ANYOF_PRINT) + 1)
410#define ANYOF_PUNCT ((_CC_PUNCT) * 2)
411#define ANYOF_NPUNCT ((ANYOF_PUNCT) + 1)
412#define ANYOF_UPPER ((_CC_UPPER) * 2)
413#define ANYOF_NUPPER ((ANYOF_UPPER) + 1)
414#define ANYOF_XDIGIT ((_CC_XDIGIT) * 2)
415#define ANYOF_NXDIGIT ((ANYOF_XDIGIT) + 1)
416#define ANYOF_PSXSPC ((_CC_PSXSPC) * 2) /* POSIX space: \s plus the vertical tab */
417#define ANYOF_NPSXSPC ((ANYOF_PSXSPC) + 1)
418#define ANYOF_BLANK ((_CC_BLANK) * 2) /* GNU extension: space and tab: non-vertical space */
419#define ANYOF_NBLANK ((ANYOF_BLANK) + 1)
b8c5462f 420
aaa51d5e 421#define ANYOF_MAX 32
08ada29e
KW
422#if (ANYOF_MAX <= _HIGHEST_REGCOMP_DOT_H_SYNC * 2 + 1)
423# error Problem with handy.h _CC_foo #defines
424#endif
b8c5462f 425
e1d1eefb
YO
426/* pseudo classes, not stored in the class bitmap, but used as flags
427 during compilation of char classes */
428
429#define ANYOF_VERTWS (ANYOF_MAX+1)
430#define ANYOF_NVERTWS (ANYOF_MAX+2)
431#define ANYOF_HORIZWS (ANYOF_MAX+3)
432#define ANYOF_NHORIZWS (ANYOF_MAX+4)
433
b8c5462f
JH
434/* Backward source code compatibility. */
435
436#define ANYOF_ALNUML ANYOF_ALNUM
437#define ANYOF_NALNUML ANYOF_NALNUM
438#define ANYOF_SPACEL ANYOF_SPACE
439#define ANYOF_NSPACEL ANYOF_NSPACE
440
441/* Utility macros for the bitmap and classes of ANYOF */
442
936ed897
IZ
443#define ANYOF_SIZE (sizeof(struct regnode_charclass))
444#define ANYOF_CLASS_SIZE (sizeof(struct regnode_charclass_class))
b8c5462f 445
936ed897 446#define ANYOF_FLAGS(p) ((p)->flags)
b8c5462f
JH
447
448#define ANYOF_BIT(c) (1 << ((c) & 7))
449
936ed897 450#define ANYOF_CLASS_BYTE(p, c) (((struct regnode_charclass_class*)(p))->classflags[((c) >> 3) & 3])
b8c5462f
JH
451#define ANYOF_CLASS_SET(p, c) (ANYOF_CLASS_BYTE(p, c) |= ANYOF_BIT(c))
452#define ANYOF_CLASS_CLEAR(p, c) (ANYOF_CLASS_BYTE(p, c) &= ~ANYOF_BIT(c))
453#define ANYOF_CLASS_TEST(p, c) (ANYOF_CLASS_BYTE(p, c) & ANYOF_BIT(c))
454
936ed897 455#define ANYOF_CLASS_ZERO(ret) Zero(((struct regnode_charclass_class*)(ret))->classflags, ANYOF_CLASSBITMAP_SIZE, char)
ab753c70
KW
456#define ANYOF_CLASS_SETALL(ret) \
457 memset (((struct regnode_charclass_class*)(ret))->classflags, 255, ANYOF_CLASSBITMAP_SIZE)
936ed897
IZ
458#define ANYOF_BITMAP_ZERO(ret) Zero(((struct regnode_charclass*)(ret))->bitmap, ANYOF_BITMAP_SIZE, char)
459
460#define ANYOF_BITMAP(p) (((struct regnode_charclass*)(p))->bitmap)
786e8c11 461#define ANYOF_BITMAP_BYTE(p, c) (ANYOF_BITMAP(p)[(((U8)(c)) >> 3) & 31])
b8c5462f
JH
462#define ANYOF_BITMAP_SET(p, c) (ANYOF_BITMAP_BYTE(p, c) |= ANYOF_BIT(c))
463#define ANYOF_BITMAP_CLEAR(p,c) (ANYOF_BITMAP_BYTE(p, c) &= ~ANYOF_BIT(c))
464#define ANYOF_BITMAP_TEST(p, c) (ANYOF_BITMAP_BYTE(p, c) & ANYOF_BIT(c))
465
f8bef550
NC
466#define ANYOF_BITMAP_SETALL(p) \
467 memset (ANYOF_BITMAP(p), 255, ANYOF_BITMAP_SIZE)
468#define ANYOF_BITMAP_CLEARALL(p) \
469 Zero (ANYOF_BITMAP(p), ANYOF_BITMAP_SIZE)
470/* Check that all 256 bits are all set. Used in S_cl_is_anything() */
c355c09a 471#define ANYOF_BITMAP_TESTALLSET(p) /* Assumes sizeof(p) == 32 */ \
f8bef550
NC
472 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)
473
936ed897
IZ
474#define ANYOF_SKIP ((ANYOF_SIZE - 1)/sizeof(regnode))
475#define ANYOF_CLASS_SKIP ((ANYOF_CLASS_SIZE - 1)/sizeof(regnode))
3a15e693 476
dd58aee1
KW
477#if ANYOF_CLASSBITMAP_SIZE != 4
478# error ANYOF_CLASSBITMAP_SIZE is expected to be 4
3a15e693 479#endif
dd58aee1
KW
480#define ANYOF_CLASS_TEST_ANY_SET(p) ((ANYOF_FLAGS(p) & ANYOF_CLASS) \
481 && memNE (((struct regnode_charclass_class*)(p))->classflags, \
482 "\0\0\0\0", ANYOF_CLASSBITMAP_SIZE))
483/*#define ANYOF_CLASS_ADD_SKIP (ANYOF_CLASS_SKIP - ANYOF_SKIP)
484 * */
c277df42 485
786e8c11 486
a687059c
LW
487/*
488 * Utility definitions.
489 */
2304df62 490#ifndef CHARMASK
075abff3 491# define UCHARAT(p) ((int)*(const U8*)(p))
a687059c 492#else
075abff3 493# define UCHARAT(p) ((int)*(p)&CHARMASK)
a687059c 494#endif
a687059c 495
c277df42
IZ
496#define EXTRA_SIZE(guy) ((sizeof(guy)-1)/sizeof(struct regnode))
497
3dab1dad
YO
498#define REG_SEEN_ZERO_LEN 0x00000001
499#define REG_SEEN_LOOKBEHIND 0x00000002
500#define REG_SEEN_GPOS 0x00000004
471f5387 501/* spare */
3dab1dad 502#define REG_SEEN_CANY 0x00000010
cce850e4 503#define REG_SEEN_SANY REG_SEEN_CANY /* src bckwrd cmpt */
6bda09f9 504#define REG_SEEN_RECURSE 0x00000020
eaf3ca90 505#define REG_TOP_LEVEL_BRANCHES 0x00000040
e2e6a0f1 506#define REG_SEEN_VERBARG 0x00000080
5d458dd8 507#define REG_SEEN_CUTGROUP 0x00000100
bcdf7404 508#define REG_SEEN_RUN_ON_COMMENT 0x00000200
bb914485 509#define REG_SEEN_EXACTF_SHARP_S 0x00000400
d09b2d29 510
73c4f7a1
GS
511START_EXTERN_C
512
be8e71aa
YO
513#ifdef PLUGGABLE_RE_EXTENSION
514#include "re_nodes.h"
515#else
d09b2d29 516#include "regnodes.h"
be8e71aa 517#endif
d09b2d29 518
f9f4320a
YO
519#ifndef PLUGGABLE_RE_EXTENSION
520#ifndef DOINIT
521EXTCONST regexp_engine PL_core_reg_engine;
522#else /* DOINIT */
523EXTCONST regexp_engine PL_core_reg_engine = {
fe578d7f 524 Perl_re_compile,
2fdbfb4d 525 Perl_regexec_flags,
f9f4320a
YO
526 Perl_re_intuit_start,
527 Perl_re_intuit_string,
2fdbfb4d
AB
528 Perl_regfree_internal,
529 Perl_reg_numbered_buff_fetch,
530 Perl_reg_numbered_buff_store,
531 Perl_reg_numbered_buff_length,
192b9cd1
AB
532 Perl_reg_named_buff,
533 Perl_reg_named_buff_iter,
49d7dfbc 534 Perl_reg_qr_package,
f9f4320a 535#if defined(USE_ITHREADS)
3c13cae6 536 Perl_regdupe_internal,
f9f4320a 537#endif
3c13cae6 538 Perl_re_op_compile
f9f4320a
YO
539};
540#endif /* DOINIT */
541#endif /* PLUGGABLE_RE_EXTENSION */
542
543
73c4f7a1 544END_EXTERN_C
cad2e5aa 545
cad2e5aa 546
261faec3
MJD
547/* .what is a character array with one character for each member of .data
548 * The character describes the function of the corresponding .data item:
af534a04 549 * a - AV for paren_name_list under DEBUGGING
0111c4fd 550 * f - start-class data for regstclass optimization
68e2671b 551 * l - start op for literal (?{EVAL}) item
d63c20f2 552 * L - start op for literal (?{EVAL}) item, with separate CV (qr//)
b30fcab9 553 * r - pointer to an embedded code-containing qr, e.g. /ab$qr/
38a44b82 554 * s - swash for Unicode-style character class, and the multicharacter
9e55ce06
JH
555 * strings resulting from casefolding the single-character entries
556 * in the character class
a3621e74 557 * t - trie struct
55eed653 558 * u - trie struct's widecharmap (a HV, so can't share, must dup)
2b8b4781 559 * also used for revcharmap and words under DEBUGGING
07be1b83 560 * T - aho-trie struct
81714fb9 561 * S - sv for named capture lookup
261faec3 562 * 20010712 mjd@plover.com
c3613ed7 563 * (Remember to update re_dup() and pregfree() if you add any items.)
261faec3 564 */
cad2e5aa
JH
565struct reg_data {
566 U32 count;
567 U8 *what;
568 void* data[1];
569};
570
a1cac82e
NC
571/* Code in S_to_utf8_substr() and S_to_byte_substr() in regexec.c accesses
572 anchored* and float* via array indexes 0 and 1. */
cad2e5aa 573#define anchored_substr substrs->data[0].substr
33b8afdf 574#define anchored_utf8 substrs->data[0].utf8_substr
cad2e5aa 575#define anchored_offset substrs->data[0].min_offset
1de06328
YO
576#define anchored_end_shift substrs->data[0].end_shift
577
cad2e5aa 578#define float_substr substrs->data[1].substr
33b8afdf 579#define float_utf8 substrs->data[1].utf8_substr
cad2e5aa
JH
580#define float_min_offset substrs->data[1].min_offset
581#define float_max_offset substrs->data[1].max_offset
1de06328
YO
582#define float_end_shift substrs->data[1].end_shift
583
cad2e5aa 584#define check_substr substrs->data[2].substr
33b8afdf 585#define check_utf8 substrs->data[2].utf8_substr
cad2e5aa
JH
586#define check_offset_min substrs->data[2].min_offset
587#define check_offset_max substrs->data[2].max_offset
1de06328 588#define check_end_shift substrs->data[2].end_shift
a3621e74 589
288b8c02
NC
590#define RX_ANCHORED_SUBSTR(rx) (((struct regexp *)SvANY(rx))->anchored_substr)
591#define RX_ANCHORED_UTF8(rx) (((struct regexp *)SvANY(rx))->anchored_utf8)
592#define RX_FLOAT_SUBSTR(rx) (((struct regexp *)SvANY(rx))->float_substr)
593#define RX_FLOAT_UTF8(rx) (((struct regexp *)SvANY(rx))->float_utf8)
a3621e74
YO
594
595/* trie related stuff */
5d9a96ca 596
a3621e74
YO
597/* a transition record for the state machine. the
598 check field determines which state "owns" the
599 transition. the char the transition is for is
600 determined by offset from the owning states base
601 field. the next field determines which state
602 is to be transitioned to if any.
603*/
604struct _reg_trie_trans {
605 U32 next;
606 U32 check;
607};
608
609/* a transition list element for the list based representation */
610struct _reg_trie_trans_list_elem {
611 U16 forid;
612 U32 newstate;
613};
614typedef struct _reg_trie_trans_list_elem reg_trie_trans_le;
615
616/* a state for compressed nodes. base is an offset
617 into an array of reg_trie_trans array. If wordnum is
618 nonzero the state is accepting. if base is zero then
619 the state has no children (and will be accepting)
620*/
621struct _reg_trie_state {
622 U16 wordnum;
623 union {
624 U32 base;
625 reg_trie_trans_le* list;
626 } trans;
627};
628
2e64971a
DM
629/* info per word; indexed by wordnum */
630typedef struct {
631 U16 prev; /* previous word in acceptance chain; eg in
632 * zzz|abc|ab/ after matching the chars abc, the
633 * accepted word is #2, and the previous accepted
634 * word is #3 */
635 U32 len; /* how many chars long is this word? */
636 U32 accept; /* accept state for this word */
637} reg_trie_wordinfo;
a3621e74
YO
638
639
a3621e74
YO
640typedef struct _reg_trie_state reg_trie_state;
641typedef struct _reg_trie_trans reg_trie_trans;
642
643
644/* anything in here that needs to be freed later
23eab42c
NC
645 should be dealt with in pregfree.
646 refcount is first in both this and _reg_ac_data to allow a space
647 optimisation in Perl_regdupe. */
a3621e74 648struct _reg_trie_data {
23eab42c 649 U32 refcount; /* number of times this trie is referenced */
786e8c11
YO
650 U32 lasttrans; /* last valid transition element */
651 U16 *charmap; /* byte to charid lookup array */
786e8c11
YO
652 reg_trie_state *states; /* state data */
653 reg_trie_trans *trans; /* array of transition elements */
654 char *bitmap; /* stclass bitmap */
786e8c11
YO
655 U16 *jump; /* optional 1 indexed array of offsets before tail
656 for the node following a given word. */
2e64971a 657 reg_trie_wordinfo *wordinfo; /* array of info per word */
1b5c067c
NC
658 U16 uniquecharcount; /* unique chars in trie (width of trans table) */
659 U32 startstate; /* initial state - used for common prefix optimisation */
660 STRLEN minlen; /* minimum length of words in trie - build/opt only? */
661 STRLEN maxlen; /* maximum length of words in trie - build/opt only? */
2e64971a 662 U32 prefixlen; /* #chars in common prefix */
1e2e3d02
YO
663 U32 statecount; /* Build only - number of states in the states array
664 (including the unused zero state) */
786e8c11 665 U32 wordcount; /* Build only */
a3621e74 666#ifdef DEBUGGING
786e8c11 667 STRLEN charcount; /* Build only */
a3621e74
YO
668#endif
669};
2b8b4781
NC
670/* There is one (3 under DEBUGGING) pointers that logically belong in this
671 structure, but are held outside as they need duplication on thread cloning,
672 whereas the rest of the structure can be read only:
673 HV *widecharmap; code points > 255 to charid
674#ifdef DEBUGGING
675 AV *words; Array of words contained in trie, for dumping
676 AV *revcharmap; Map of each charid back to its character representation
677#endif
678*/
679
680#define TRIE_WORDS_OFFSET 2
681
a3621e74
YO
682typedef struct _reg_trie_data reg_trie_data;
683
23eab42c
NC
684/* refcount is first in both this and _reg_trie_data to allow a space
685 optimisation in Perl_regdupe. */
07be1b83 686struct _reg_ac_data {
23eab42c 687 U32 refcount;
1b5c067c 688 U32 trie;
07be1b83
YO
689 U32 *fail;
690 reg_trie_state *states;
07be1b83
YO
691};
692typedef struct _reg_ac_data reg_ac_data;
693
486ec47a 694/* ANY_BIT doesn't use the structure, so we can borrow it here.
3dab1dad
YO
695 This is simpler than refactoring all of it as wed end up with
696 three different sets... */
697
698#define TRIE_BITMAP(p) (((reg_trie_data *)(p))->bitmap)
786e8c11 699#define TRIE_BITMAP_BYTE(p, c) (TRIE_BITMAP(p)[(((U8)(c)) >> 3) & 31])
5dfe062f
YO
700#define TRIE_BITMAP_SET(p, c) (TRIE_BITMAP_BYTE(p, c) |= ANYOF_BIT((U8)c))
701#define TRIE_BITMAP_CLEAR(p,c) (TRIE_BITMAP_BYTE(p, c) &= ~ANYOF_BIT((U8)c))
702#define TRIE_BITMAP_TEST(p, c) (TRIE_BITMAP_BYTE(p, c) & ANYOF_BIT((U8)c))
3dab1dad 703
1de06328
YO
704#define IS_ANYOF_TRIE(op) ((op)==TRIEC || (op)==AHOCORASICKC)
705#define IS_TRIE_AC(op) ((op)>=AHOCORASICK)
706
707
786e8c11
YO
708#define BITMAP_BYTE(p, c) (((U8*)p)[(((U8)(c)) >> 3) & 31])
709#define BITMAP_TEST(p, c) (BITMAP_BYTE(p, c) & ANYOF_BIT((U8)c))
3dab1dad 710
a3621e74
YO
711/* these defines assume uniquecharcount is the correct variable, and state may be evaluated twice */
712#define TRIE_NODENUM(state) (((state)-1)/(trie->uniquecharcount)+1)
713#define SAFE_TRIE_NODENUM(state) ((state) ? (((state)-1)/(trie->uniquecharcount)+1) : (state))
714#define TRIE_NODEIDX(state) ((state) ? (((state)-1)*(trie->uniquecharcount)+1) : (state))
715
3dab1dad 716#ifdef DEBUGGING
3dab1dad 717#define TRIE_CHARCOUNT(trie) ((trie)->charcount)
3dab1dad 718#else
3dab1dad 719#define TRIE_CHARCOUNT(trie) (trie_charcount)
3dab1dad 720#endif
a3621e74 721
0111c4fd
RGS
722#define RE_TRIE_MAXBUF_INIT 65536
723#define RE_TRIE_MAXBUF_NAME "\022E_TRIE_MAXBUF"
a3621e74
YO
724#define RE_DEBUG_FLAGS "\022E_DEBUG_FLAGS"
725
be8e71aa
YO
726/*
727
728RE_DEBUG_FLAGS is used to control what debug output is emitted
729its divided into three groups of options, some of which interact.
730The three groups are: Compile, Execute, Extra. There is room for a
731further group, as currently only the low three bytes are used.
732
733 Compile Options:
07be1b83 734
be8e71aa
YO
735 PARSE
736 PEEP
737 TRIE
738 PROGRAM
739 OFFSETS
a3621e74 740
be8e71aa 741 Execute Options:
a3621e74 742
be8e71aa
YO
743 INTUIT
744 MATCH
745 TRIE
a3621e74 746
be8e71aa 747 Extra Options
a3621e74 748
be8e71aa
YO
749 TRIE
750 OFFSETS
751
752If you modify any of these make sure you make corresponding changes to
753re.pm, especially to the documentation.
a3621e74 754
be8e71aa
YO
755*/
756
757
758/* Compile */
759#define RE_DEBUG_COMPILE_MASK 0x0000FF
760#define RE_DEBUG_COMPILE_PARSE 0x000001
761#define RE_DEBUG_COMPILE_OPTIMISE 0x000002
762#define RE_DEBUG_COMPILE_TRIE 0x000004
763#define RE_DEBUG_COMPILE_DUMP 0x000008
f7819f85 764#define RE_DEBUG_COMPILE_FLAGS 0x000010
be8e71aa
YO
765
766/* Execute */
767#define RE_DEBUG_EXECUTE_MASK 0x00FF00
768#define RE_DEBUG_EXECUTE_INTUIT 0x000100
769#define RE_DEBUG_EXECUTE_MATCH 0x000200
770#define RE_DEBUG_EXECUTE_TRIE 0x000400
771
772/* Extra */
773#define RE_DEBUG_EXTRA_MASK 0xFF0000
774#define RE_DEBUG_EXTRA_TRIE 0x010000
775#define RE_DEBUG_EXTRA_OFFSETS 0x020000
a5ca303d
YO
776#define RE_DEBUG_EXTRA_OFFDEBUG 0x040000
777#define RE_DEBUG_EXTRA_STATE 0x080000
778#define RE_DEBUG_EXTRA_OPTIMISE 0x100000
e7707071 779#define RE_DEBUG_EXTRA_BUFFERS 0x400000
2c296965 780#define RE_DEBUG_EXTRA_GPOS 0x800000
24b23f37
YO
781/* combined */
782#define RE_DEBUG_EXTRA_STACK 0x280000
be8e71aa 783
e68ec53f 784#define RE_DEBUG_FLAG(x) (re_debug_flags & x)
be8e71aa
YO
785/* Compile */
786#define DEBUG_COMPILE_r(x) DEBUG_r( \
e68ec53f 787 if (re_debug_flags & RE_DEBUG_COMPILE_MASK) x )
be8e71aa 788#define DEBUG_PARSE_r(x) DEBUG_r( \
e68ec53f 789 if (re_debug_flags & RE_DEBUG_COMPILE_PARSE) x )
be8e71aa 790#define DEBUG_OPTIMISE_r(x) DEBUG_r( \
e68ec53f 791 if (re_debug_flags & RE_DEBUG_COMPILE_OPTIMISE) x )
be8e71aa 792#define DEBUG_PARSE_r(x) DEBUG_r( \
e68ec53f 793 if (re_debug_flags & RE_DEBUG_COMPILE_PARSE) x )
be8e71aa 794#define DEBUG_DUMP_r(x) DEBUG_r( \
e68ec53f 795 if (re_debug_flags & RE_DEBUG_COMPILE_DUMP) x )
be8e71aa 796#define DEBUG_TRIE_COMPILE_r(x) DEBUG_r( \
e68ec53f 797 if (re_debug_flags & RE_DEBUG_COMPILE_TRIE) x )
f7819f85
A
798#define DEBUG_FLAGS_r(x) DEBUG_r( \
799 if (re_debug_flags & RE_DEBUG_COMPILE_FLAGS) x )
be8e71aa
YO
800/* Execute */
801#define DEBUG_EXECUTE_r(x) DEBUG_r( \
e68ec53f 802 if (re_debug_flags & RE_DEBUG_EXECUTE_MASK) x )
be8e71aa 803#define DEBUG_INTUIT_r(x) DEBUG_r( \
e68ec53f 804 if (re_debug_flags & RE_DEBUG_EXECUTE_INTUIT) x )
be8e71aa 805#define DEBUG_MATCH_r(x) DEBUG_r( \
e68ec53f 806 if (re_debug_flags & RE_DEBUG_EXECUTE_MATCH) x )
be8e71aa 807#define DEBUG_TRIE_EXECUTE_r(x) DEBUG_r( \
e68ec53f 808 if (re_debug_flags & RE_DEBUG_EXECUTE_TRIE) x )
be8e71aa
YO
809
810/* Extra */
811#define DEBUG_EXTRA_r(x) DEBUG_r( \
e68ec53f 812 if (re_debug_flags & RE_DEBUG_EXTRA_MASK) x )
a5ca303d
YO
813#define DEBUG_OFFSETS_r(x) DEBUG_r( \
814 if (re_debug_flags & RE_DEBUG_EXTRA_OFFSETS) x )
ab3bbdeb
YO
815#define DEBUG_STATE_r(x) DEBUG_r( \
816 if (re_debug_flags & RE_DEBUG_EXTRA_STATE) x )
24b23f37
YO
817#define DEBUG_STACK_r(x) DEBUG_r( \
818 if (re_debug_flags & RE_DEBUG_EXTRA_STACK) x )
e7707071
YO
819#define DEBUG_BUFFERS_r(x) DEBUG_r( \
820 if (re_debug_flags & RE_DEBUG_EXTRA_BUFFERS) x )
821
a5ca303d
YO
822#define DEBUG_OPTIMISE_MORE_r(x) DEBUG_r( \
823 if ((RE_DEBUG_EXTRA_OPTIMISE|RE_DEBUG_COMPILE_OPTIMISE) == \
824 (re_debug_flags & (RE_DEBUG_EXTRA_OPTIMISE|RE_DEBUG_COMPILE_OPTIMISE)) ) x )
07be1b83 825#define MJD_OFFSET_DEBUG(x) DEBUG_r( \
a5ca303d 826 if (re_debug_flags & RE_DEBUG_EXTRA_OFFDEBUG) \
be8e71aa
YO
827 Perl_warn_nocontext x )
828#define DEBUG_TRIE_COMPILE_MORE_r(x) DEBUG_TRIE_COMPILE_r( \
e68ec53f 829 if (re_debug_flags & RE_DEBUG_EXTRA_TRIE) x )
be8e71aa 830#define DEBUG_TRIE_EXECUTE_MORE_r(x) DEBUG_TRIE_EXECUTE_r( \
e68ec53f 831 if (re_debug_flags & RE_DEBUG_EXTRA_TRIE) x )
be8e71aa
YO
832
833#define DEBUG_TRIE_r(x) DEBUG_r( \
e68ec53f 834 if (re_debug_flags & (RE_DEBUG_COMPILE_TRIE \
be8e71aa 835 | RE_DEBUG_EXECUTE_TRIE )) x )
2c296965
YO
836#define DEBUG_GPOS_r(x) DEBUG_r( \
837 if (re_debug_flags & RE_DEBUG_EXTRA_GPOS) x )
be8e71aa
YO
838
839/* initialization */
595598ee
KW
840/* get_sv() can return NULL during global destruction. re_debug_flags can get
841 * clobbered by a longjmp, so must be initialized */
e68ec53f
YO
842#define GET_RE_DEBUG_FLAGS DEBUG_r({ \
843 SV * re_debug_flags_sv = NULL; \
595598ee 844 re_debug_flags = 0; \
e68ec53f
YO
845 re_debug_flags_sv = get_sv(RE_DEBUG_FLAGS, 1); \
846 if (re_debug_flags_sv) { \
8e9a8a48 847 if (!SvIOK(re_debug_flags_sv)) \
e68ec53f 848 sv_setuv(re_debug_flags_sv, RE_DEBUG_COMPILE_DUMP | RE_DEBUG_EXECUTE_MASK ); \
8e9a8a48 849 re_debug_flags=SvIV(re_debug_flags_sv); \
e68ec53f
YO
850 }\
851})
a3621e74
YO
852
853#ifdef DEBUGGING
ab3bbdeb 854
9c4fdda1 855#define GET_RE_DEBUG_FLAGS_DECL VOL IV re_debug_flags \
9d63fa07 856 PERL_UNUSED_DECL = 0; GET_RE_DEBUG_FLAGS;
ab3bbdeb
YO
857
858#define RE_PV_COLOR_DECL(rpv,rlen,isuni,dsv,pv,l,m,c1,c2) \
859 const char * const rpv = \
ddc5bc0f 860 pv_pretty((dsv), (pv), (l), (m), \
ab3bbdeb 861 PL_colors[(c1)],PL_colors[(c2)], \
c89df6cf 862 PERL_PV_ESCAPE_RE|PERL_PV_ESCAPE_NONASCII |((isuni) ? PERL_PV_ESCAPE_UNI : 0) ); \
0df25f3d 863 const int rlen = SvCUR(dsv)
ab3bbdeb
YO
864
865#define RE_SV_ESCAPE(rpv,isuni,dsv,sv,m) \
866 const char * const rpv = \
ddc5bc0f 867 pv_pretty((dsv), (SvPV_nolen_const(sv)), (SvCUR(sv)), (m), \
ab3bbdeb 868 PL_colors[(c1)],PL_colors[(c2)], \
c89df6cf 869 PERL_PV_ESCAPE_RE|PERL_PV_ESCAPE_NONASCII |((isuni) ? PERL_PV_ESCAPE_UNI : 0) )
ab3bbdeb
YO
870
871#define RE_PV_QUOTED_DECL(rpv,isuni,dsv,pv,l,m) \
872 const char * const rpv = \
ddc5bc0f 873 pv_pretty((dsv), (pv), (l), (m), \
ab3bbdeb 874 PL_colors[0], PL_colors[1], \
c89df6cf 875 ( PERL_PV_PRETTY_QUOTE | PERL_PV_ESCAPE_RE | PERL_PV_ESCAPE_NONASCII | PERL_PV_PRETTY_ELLIPSES | \
ab3bbdeb 876 ((isuni) ? PERL_PV_ESCAPE_UNI : 0)) \
0a0e9afc 877 )
ab3bbdeb
YO
878
879#define RE_SV_DUMPLEN(ItEm) (SvCUR(ItEm) - (SvTAIL(ItEm)!=0))
880#define RE_SV_TAIL(ItEm) (SvTAIL(ItEm) ? "$" : "")
881
882#else /* if not DEBUGGING */
883
a3621e74 884#define GET_RE_DEBUG_FLAGS_DECL
ab3bbdeb
YO
885#define RE_PV_COLOR_DECL(rpv,rlen,isuni,dsv,pv,l,m,c1,c2)
886#define RE_SV_ESCAPE(rpv,isuni,dsv,sv,m)
887#define RE_PV_QUOTED_DECL(rpv,isuni,dsv,pv,l,m)
888#define RE_SV_DUMPLEN(ItEm)
889#define RE_SV_TAIL(ItEm)
890
891#endif /* DEBUG RELATED DEFINES */
a3621e74 892
e9a8c099
MHM
893/*
894 * Local variables:
895 * c-indentation-style: bsd
896 * c-basic-offset: 4
14d04a33 897 * indent-tabs-mode: nil
e9a8c099
MHM
898 * End:
899 *
14d04a33 900 * ex: set ts=8 sts=4 sw=4 et:
e9a8c099 901 */