This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
Use the perl.org news gw instead of the perl.com one
[perl5.git] / regcomp.h
CommitLineData
a0d0e21e 1/* regcomp.h
a687059c
LW
2 */
3
c277df42
IZ
4typedef OP OP_4tree; /* Will be redefined later. */
5
a687059c
LW
6/*
7 * The "internal use only" fields in regexp.h are present to pass info from
8 * compile to execute that permits the execute phase to run lots faster on
9 * simple cases. They are:
10 *
79072805 11 * regstart sv that must begin a match; Nullch if none obvious
a687059c
LW
12 * reganch is the match anchored (at beginning-of-line only)?
13 * regmust string (pointer into program) that match must include, or NULL
79072805 14 * [regmust changed to SV* for bminstr()--law]
a687059c
LW
15 * regmlen length of regmust string
16 * [regmlen not used currently]
17 *
18 * Regstart and reganch permit very fast decisions on suitable starting points
19 * for a match, cutting down the work a lot. Regmust permits fast rejection
20 * of lines that cannot possibly match. The regmust tests are costly enough
e50aee73 21 * that pregcomp() supplies a regmust only if the r.e. contains something
a687059c
LW
22 * potentially expensive (at present, the only such thing detected is * or +
23 * at the start of the r.e., which can involve a lot of backup). Regmlen is
e50aee73 24 * supplied because the test in pregexec() needs it and pregcomp() is computing
a687059c
LW
25 * it anyway.
26 * [regmust is now supplied always. The tests that use regmust have a
27 * heuristic that disables the test if it usually matches.]
28 *
29 * [In fact, we now use regmust in many cases to locate where the search
30 * starts in the string, so if regback is >= 0, the regmust search is never
31 * wasted effort. The regback variable says how many characters back from
32 * where regmust matched is the earliest possible start of the match.
33 * For instance, /[a-z].foo/ has a regmust of 'foo' and a regback of 2.]
34 */
35
36/*
37 * Structure for regexp "program". This is essentially a linear encoding
38 * of a nondeterministic finite-state machine (aka syntax charts or
39 * "railroad normal form" in parsing technology). Each node is an opcode
40 * plus a "next" pointer, possibly plus an operand. "Next" pointers of
41 * all nodes except BRANCH implement concatenation; a "next" pointer with
42 * a BRANCH on both ends of it is connecting two alternatives. (Here we
43 * have one of the subtle syntax dependencies: an individual BRANCH (as
44 * opposed to a collection of them) is never concatenated with anything
45 * because of operator precedence.) The operand of some types of node is
46 * a literal string; for others, it is a node leading into a sub-FSM. In
47 * particular, the operand of a BRANCH node is the first node of the branch.
48 * (NB this is *not* a tree structure: the tail of the branch connects
49 * to the thing following the set of BRANCHes.) The opcodes are:
50 */
51
a687059c
LW
52/*
53 * A node is one char of opcode followed by two chars of "next" pointer.
54 * "Next" pointers are stored as two 8-bit pieces, high order first. The
55 * value is a positive offset from the opcode of the node containing it.
56 * An operand, if any, simply follows the node. (Note that much of the
57 * code generation knows about this implicit relationship.)
58 *
59 * Using two bytes for the "next" pointer is vast overkill for most things,
60 * but allows patterns to get big without disasters.
61 *
e91177ed 62 * [The "next" pointer is always aligned on an even
a687059c
LW
63 * boundary, and reads the offset directly as a short. Also, there is no
64 * special test to reverse the sign of BACK pointers since the offset is
65 * stored negative.]
66 */
67
c277df42 68struct regnode_string {
cd439c50 69 U8 str_len;
c277df42
IZ
70 U8 type;
71 U16 next_off;
cd439c50 72 char string[1];
c277df42 73};
a687059c 74
c277df42
IZ
75struct regnode_1 {
76 U8 flags;
77 U8 type;
78 U16 next_off;
79 U32 arg1;
80};
81
82struct regnode_2 {
83 U8 flags;
84 U8 type;
85 U16 next_off;
86 U16 arg1;
87 U16 arg2;
88};
89
936ed897 90#define ANYOF_BITMAP_SIZE 32 /* 256 b/(8 b/B) */
ffc61ed2 91#define ANYOF_CLASSBITMAP_SIZE 4 /* up to 32 (8*4) named classes */
936ed897
IZ
92
93struct regnode_charclass {
94 U8 flags;
95 U8 type;
96 U16 next_off;
ffc61ed2 97 U32 arg1;
19103b91 98 char bitmap[ANYOF_BITMAP_SIZE]; /* only compile-time */
936ed897
IZ
99};
100
19103b91
JH
101struct regnode_charclass_class { /* has [[:blah:]] classes */
102 U8 flags; /* should have ANYOF_CLASS here */
936ed897
IZ
103 U8 type;
104 U16 next_off;
ffc61ed2 105 U32 arg1;
19103b91
JH
106 char bitmap[ANYOF_BITMAP_SIZE]; /* both compile-time */
107 char classflags[ANYOF_CLASSBITMAP_SIZE]; /* and run-time */
936ed897
IZ
108};
109
9c7e81e8
AD
110/* XXX fix this description.
111 Impose a limit of REG_INFTY on various pattern matching operations
112 to limit stack growth and to avoid "infinite" recursions.
113*/
114/* The default size for REG_INFTY is I16_MAX, which is the same as
115 SHORT_MAX (see perl.h). Unfortunately I16 isn't necessarily 16 bits
116 (see handy.h). On the Cray C90, sizeof(short)==4 and hence I16_MAX is
117 ((1<<31)-1), while on the Cray T90, sizeof(short)==8 and I16_MAX is
118 ((1<<63)-1). To limit stack growth to reasonable sizes, supply a
119 smaller default.
120 --Andy Dougherty 11 June 1998
121*/
122#if SHORTSIZE > 2
123# ifndef REG_INFTY
124# define REG_INFTY ((1<<15)-1)
125# endif
126#endif
127
128#ifndef REG_INFTY
83cfe48c
IZ
129# define REG_INFTY I16_MAX
130#endif
a687059c 131
e91177ed
IZ
132#define ARG_VALUE(arg) (arg)
133#define ARG__SET(arg,val) ((arg) = (val))
c277df42 134
e8e3b0fb
JH
135#undef ARG
136#undef ARG1
137#undef ARG2
138
c277df42
IZ
139#define ARG(p) ARG_VALUE(ARG_LOC(p))
140#define ARG1(p) ARG_VALUE(ARG1_LOC(p))
141#define ARG2(p) ARG_VALUE(ARG2_LOC(p))
142#define ARG_SET(p, val) ARG__SET(ARG_LOC(p), (val))
143#define ARG1_SET(p, val) ARG__SET(ARG1_LOC(p), (val))
144#define ARG2_SET(p, val) ARG__SET(ARG2_LOC(p), (val))
145
e8e3b0fb
JH
146#undef NEXT_OFF
147#undef NODE_ALIGN
148
c277df42 149#ifndef lint
e91177ed
IZ
150# define NEXT_OFF(p) ((p)->next_off)
151# define NODE_ALIGN(node)
152# define NODE_ALIGN_FILL(node) ((node)->flags = 0xde) /* deadbeef */
a687059c 153#else /* lint */
c277df42 154# define NEXT_OFF(p) 0
e91177ed
IZ
155# define NODE_ALIGN(node)
156# define NODE_ALIGN_FILL(node)
a687059c
LW
157#endif /* lint */
158
c277df42
IZ
159#define SIZE_ALIGN NODE_ALIGN
160
e8e3b0fb
JH
161#undef OP
162#undef OPERAND
163#undef MASK
164#undef STRING
165
e91177ed
IZ
166#define OP(p) ((p)->type)
167#define OPERAND(p) (((struct regnode_string *)p)->string)
cd439c50
IZ
168#define MASK(p) ((char*)OPERAND(p))
169#define STR_LEN(p) (((struct regnode_string *)p)->str_len)
170#define STRING(p) (((struct regnode_string *)p)->string)
171#define STR_SZ(l) ((l + sizeof(regnode) - 1) / sizeof(regnode))
172#define NODE_SZ_STR(p) (STR_SZ(STR_LEN(p))+1)
173
e8e3b0fb
JH
174#undef NODE_ALIGN
175#undef ARG_LOC
176#undef NEXTOPER
177#undef PREVOPER
178
e91177ed
IZ
179#define NODE_ALIGN(node)
180#define ARG_LOC(p) (((struct regnode_1 *)p)->arg1)
181#define ARG1_LOC(p) (((struct regnode_2 *)p)->arg1)
182#define ARG2_LOC(p) (((struct regnode_2 *)p)->arg2)
183#define NODE_STEP_REGNODE 1 /* sizeof(regnode)/sizeof(regnode) */
184#define EXTRA_STEP_2ARGS EXTRA_SIZE(struct regnode_2)
185
186#define NODE_STEP_B 4
c277df42
IZ
187
188#define NEXTOPER(p) ((p) + NODE_STEP_REGNODE)
189#define PREVOPER(p) ((p) - NODE_STEP_REGNODE)
190
e91177ed 191#define FILL_ADVANCE_NODE(ptr, op) STMT_START { \
c277df42 192 (ptr)->type = op; (ptr)->next_off = 0; (ptr)++; } STMT_END
e91177ed 193#define FILL_ADVANCE_NODE_ARG(ptr, op, arg) STMT_START { \
c277df42 194 ARG_SET(ptr, arg); FILL_ADVANCE_NODE(ptr, op); (ptr) += 1; } STMT_END
a687059c 195
22c35a8c 196#define REG_MAGIC 0234
a687059c 197
830247a4 198#define SIZE_ONLY (RExC_emit == &PL_regdummy)
c277df42 199
936ed897 200/* Flags for node->flags of ANYOF */
b8c5462f 201
19103b91 202#define ANYOF_CLASS 0x08 /* has [[:blah:]] classes */
ffc61ed2
JH
203#define ANYOF_INVERT 0x04
204#define ANYOF_FOLD 0x02
205#define ANYOF_LOCALE 0x01
b8c5462f 206
653099ff 207/* Used for regstclass only */
ffc61ed2
JH
208#define ANYOF_EOS 0x10 /* Can match an empty string too */
209
210/* There is a character or a range past 0xff */
211#define ANYOF_UNICODE 0x20
1aa99e6b 212#define ANYOF_UNICODE_ALL 0x40 /* Can match any char past 0xff */
ffc61ed2 213
4f66b38d
HS
214/* size of node is large (includes class pointer) */
215#define ANYOF_LARGE 0x80
216
ffc61ed2
JH
217/* Are there any runtime flags on in this node? */
218#define ANYOF_RUNTIME(s) (ANYOF_FLAGS(s) & 0x0f)
219
220#define ANYOF_FLAGS_ALL 0xff
653099ff 221
936ed897 222/* Character classes for node->classflags of ANYOF */
653099ff
GS
223/* Should be synchronized with a table in regprop() */
224/* 2n should pair with 2n+1 */
b8c5462f 225
289d4f09 226#define ANYOF_ALNUM 0 /* \w, PL_utf8_alnum, utf8::IsWord, ALNUM */
b8c5462f 227#define ANYOF_NALNUM 1
aaa51d5e 228#define ANYOF_SPACE 2 /* \s */
b8c5462f
JH
229#define ANYOF_NSPACE 3
230#define ANYOF_DIGIT 4
231#define ANYOF_NDIGIT 5
289d4f09 232#define ANYOF_ALNUMC 6 /* isalnum(3), utf8::IsAlnum, ALNUMC */
b8c5462f
JH
233#define ANYOF_NALNUMC 7
234#define ANYOF_ALPHA 8
235#define ANYOF_NALPHA 9
236#define ANYOF_ASCII 10
237#define ANYOF_NASCII 11
238#define ANYOF_CNTRL 12
239#define ANYOF_NCNTRL 13
240#define ANYOF_GRAPH 14
241#define ANYOF_NGRAPH 15
242#define ANYOF_LOWER 16
243#define ANYOF_NLOWER 17
244#define ANYOF_PRINT 18
245#define ANYOF_NPRINT 19
246#define ANYOF_PUNCT 20
247#define ANYOF_NPUNCT 21
248#define ANYOF_UPPER 22
249#define ANYOF_NUPPER 23
250#define ANYOF_XDIGIT 24
251#define ANYOF_NXDIGIT 25
aaa51d5e
JF
252#define ANYOF_PSXSPC 26 /* POSIX space: \s plus the vertical tab */
253#define ANYOF_NPSXSPC 27
ffc61ed2 254#define ANYOF_BLANK 28 /* GNU extension: space and tab: non-vertical space */
289d4f09 255#define ANYOF_NBLANK 29
b8c5462f 256
aaa51d5e 257#define ANYOF_MAX 32
b8c5462f
JH
258
259/* Backward source code compatibility. */
260
261#define ANYOF_ALNUML ANYOF_ALNUM
262#define ANYOF_NALNUML ANYOF_NALNUM
263#define ANYOF_SPACEL ANYOF_SPACE
264#define ANYOF_NSPACEL ANYOF_NSPACE
265
266/* Utility macros for the bitmap and classes of ANYOF */
267
936ed897
IZ
268#define ANYOF_SIZE (sizeof(struct regnode_charclass))
269#define ANYOF_CLASS_SIZE (sizeof(struct regnode_charclass_class))
b8c5462f 270
936ed897 271#define ANYOF_FLAGS(p) ((p)->flags)
b8c5462f
JH
272
273#define ANYOF_BIT(c) (1 << ((c) & 7))
274
936ed897 275#define ANYOF_CLASS_BYTE(p, c) (((struct regnode_charclass_class*)(p))->classflags[((c) >> 3) & 3])
b8c5462f
JH
276#define ANYOF_CLASS_SET(p, c) (ANYOF_CLASS_BYTE(p, c) |= ANYOF_BIT(c))
277#define ANYOF_CLASS_CLEAR(p, c) (ANYOF_CLASS_BYTE(p, c) &= ~ANYOF_BIT(c))
278#define ANYOF_CLASS_TEST(p, c) (ANYOF_CLASS_BYTE(p, c) & ANYOF_BIT(c))
279
936ed897
IZ
280#define ANYOF_CLASS_ZERO(ret) Zero(((struct regnode_charclass_class*)(ret))->classflags, ANYOF_CLASSBITMAP_SIZE, char)
281#define ANYOF_BITMAP_ZERO(ret) Zero(((struct regnode_charclass*)(ret))->bitmap, ANYOF_BITMAP_SIZE, char)
282
283#define ANYOF_BITMAP(p) (((struct regnode_charclass*)(p))->bitmap)
284#define ANYOF_BITMAP_BYTE(p, c) (ANYOF_BITMAP(p)[((c) >> 3) & 31])
b8c5462f
JH
285#define ANYOF_BITMAP_SET(p, c) (ANYOF_BITMAP_BYTE(p, c) |= ANYOF_BIT(c))
286#define ANYOF_BITMAP_CLEAR(p,c) (ANYOF_BITMAP_BYTE(p, c) &= ~ANYOF_BIT(c))
287#define ANYOF_BITMAP_TEST(p, c) (ANYOF_BITMAP_BYTE(p, c) & ANYOF_BIT(c))
288
f8bef550
NC
289#define ANYOF_BITMAP_SETALL(p) \
290 memset (ANYOF_BITMAP(p), 255, ANYOF_BITMAP_SIZE)
291#define ANYOF_BITMAP_CLEARALL(p) \
292 Zero (ANYOF_BITMAP(p), ANYOF_BITMAP_SIZE)
293/* Check that all 256 bits are all set. Used in S_cl_is_anything() */
294#define ANYOF_BITMAP_TESTALLSET(p) \
295 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)
296
936ed897
IZ
297#define ANYOF_SKIP ((ANYOF_SIZE - 1)/sizeof(regnode))
298#define ANYOF_CLASS_SKIP ((ANYOF_CLASS_SIZE - 1)/sizeof(regnode))
299#define ANYOF_CLASS_ADD_SKIP (ANYOF_CLASS_SKIP - ANYOF_SKIP)
c277df42 300
a687059c
LW
301/*
302 * Utility definitions.
303 */
304#ifndef lint
2304df62 305#ifndef CHARMASK
4d61ec05 306#define UCHARAT(p) ((int)*(U8*)(p))
a687059c 307#else
2304df62 308#define UCHARAT(p) ((int)*(p)&CHARMASK)
a687059c
LW
309#endif
310#else /* lint */
6b88bc9c 311#define UCHARAT(p) PL_regdummy
a687059c
LW
312#endif /* lint */
313
c277df42
IZ
314#define EXTRA_SIZE(guy) ((sizeof(guy)-1)/sizeof(struct regnode))
315
3baa4c62
JH
316#define REG_SEEN_ZERO_LEN 1
317#define REG_SEEN_LOOKBEHIND 2
318#define REG_SEEN_GPOS 4
319#define REG_SEEN_EVAL 8
cce850e4
JH
320#define REG_SEEN_CANY 16
321#define REG_SEEN_SANY REG_SEEN_CANY /* src bckwrd cmpt */
d09b2d29 322
73c4f7a1
GS
323START_EXTERN_C
324
d09b2d29
IZ
325#include "regnodes.h"
326
4d61ec05 327/* The following have no fixed length. U8 so we can do strchr() on it. */
d09b2d29 328#ifndef DOINIT
4d61ec05 329EXTCONST U8 PL_varies[];
d09b2d29 330#else
4d61ec05 331EXTCONST U8 PL_varies[] = {
d09b2d29 332 BRANCH, BACK, STAR, PLUS, CURLY, CURLYX, REF, REFF, REFFL,
a0ed51b3 333 WHILEM, CURLYM, CURLYN, BRANCHJ, IFTHEN, SUSPEND, CLUMP, 0
c277df42 334};
d09b2d29 335#endif
c277df42 336
4d61ec05
GS
337/* The following always have a length of 1. U8 we can do strchr() on it. */
338/* (Note that length 1 means "one character" under UTF8, not "one octet".) */
d09b2d29 339#ifndef DOINIT
4d61ec05 340EXTCONST U8 PL_simple[];
d09b2d29 341#else
4d61ec05 342EXTCONST U8 PL_simple[] = {
f33976b4 343 REG_ANY, SANY, CANY,
ffc61ed2
JH
344 ANYOF,
345 ALNUM, ALNUML,
346 NALNUM, NALNUML,
347 SPACE, SPACEL,
348 NSPACE, NSPACEL,
349 DIGIT, NDIGIT,
350 0
c277df42
IZ
351};
352#endif
353
73c4f7a1 354END_EXTERN_C
cad2e5aa
JH
355
356typedef struct re_scream_pos_data_s
357{
358 char **scream_olds; /* match pos */
359 I32 *scream_pos; /* Internal iterator of scream. */
360} re_scream_pos_data;
361
261faec3
MJD
362/* .what is a character array with one character for each member of .data
363 * The character describes the function of the corresponding .data item:
364 * f - start-class data for regstclass optimization
365 * n - Root of op tree for (?{EVAL}) item
366 * o - Start op for (?{EVAL}) item
367 * p - Pad for (?{EVAL} item
e2e799e5 368 * s - swash for unicode-style character class
261faec3 369 * 20010712 mjd@plover.com
c3613ed7 370 * (Remember to update re_dup() and pregfree() if you add any items.)
261faec3 371 */
cad2e5aa
JH
372struct reg_data {
373 U32 count;
374 U8 *what;
375 void* data[1];
376};
377
378struct reg_substr_datum {
379 I32 min_offset;
380 I32 max_offset;
381 SV *substr;
382};
383
384struct reg_substr_data {
385 struct reg_substr_datum data[3]; /* Actual array */
386};
387
388#define anchored_substr substrs->data[0].substr
389#define anchored_offset substrs->data[0].min_offset
390#define float_substr substrs->data[1].substr
391#define float_min_offset substrs->data[1].min_offset
392#define float_max_offset substrs->data[1].max_offset
393#define check_substr substrs->data[2].substr
394#define check_offset_min substrs->data[2].min_offset
395#define check_offset_max substrs->data[2].max_offset