This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
[patch perl@8102] dos/djgpp update
[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
IZ
90#define ANYOF_BITMAP_SIZE 32 /* 256 b/(8 b/B) */
91#define ANYOF_CLASSBITMAP_SIZE 4
92
93struct regnode_charclass {
94 U8 flags;
95 U8 type;
96 U16 next_off;
97 char bitmap[ANYOF_BITMAP_SIZE];
98};
99
100struct regnode_charclass_class {
101 U8 flags;
102 U8 type;
103 U16 next_off;
104 char bitmap[ANYOF_BITMAP_SIZE];
105 char classflags[ANYOF_CLASSBITMAP_SIZE];
106};
107
9c7e81e8
AD
108/* XXX fix this description.
109 Impose a limit of REG_INFTY on various pattern matching operations
110 to limit stack growth and to avoid "infinite" recursions.
111*/
112/* The default size for REG_INFTY is I16_MAX, which is the same as
113 SHORT_MAX (see perl.h). Unfortunately I16 isn't necessarily 16 bits
114 (see handy.h). On the Cray C90, sizeof(short)==4 and hence I16_MAX is
115 ((1<<31)-1), while on the Cray T90, sizeof(short)==8 and I16_MAX is
116 ((1<<63)-1). To limit stack growth to reasonable sizes, supply a
117 smaller default.
118 --Andy Dougherty 11 June 1998
119*/
120#if SHORTSIZE > 2
121# ifndef REG_INFTY
122# define REG_INFTY ((1<<15)-1)
123# endif
124#endif
125
126#ifndef REG_INFTY
83cfe48c
IZ
127# define REG_INFTY I16_MAX
128#endif
a687059c 129
e91177ed
IZ
130#define ARG_VALUE(arg) (arg)
131#define ARG__SET(arg,val) ((arg) = (val))
c277df42
IZ
132
133#define ARG(p) ARG_VALUE(ARG_LOC(p))
134#define ARG1(p) ARG_VALUE(ARG1_LOC(p))
135#define ARG2(p) ARG_VALUE(ARG2_LOC(p))
136#define ARG_SET(p, val) ARG__SET(ARG_LOC(p), (val))
137#define ARG1_SET(p, val) ARG__SET(ARG1_LOC(p), (val))
138#define ARG2_SET(p, val) ARG__SET(ARG2_LOC(p), (val))
139
140#ifndef lint
e91177ed
IZ
141# define NEXT_OFF(p) ((p)->next_off)
142# define NODE_ALIGN(node)
143# define NODE_ALIGN_FILL(node) ((node)->flags = 0xde) /* deadbeef */
a687059c 144#else /* lint */
c277df42 145# define NEXT_OFF(p) 0
e91177ed
IZ
146# define NODE_ALIGN(node)
147# define NODE_ALIGN_FILL(node)
a687059c
LW
148#endif /* lint */
149
c277df42
IZ
150#define SIZE_ALIGN NODE_ALIGN
151
e91177ed
IZ
152#define OP(p) ((p)->type)
153#define OPERAND(p) (((struct regnode_string *)p)->string)
cd439c50
IZ
154#define MASK(p) ((char*)OPERAND(p))
155#define STR_LEN(p) (((struct regnode_string *)p)->str_len)
156#define STRING(p) (((struct regnode_string *)p)->string)
157#define STR_SZ(l) ((l + sizeof(regnode) - 1) / sizeof(regnode))
158#define NODE_SZ_STR(p) (STR_SZ(STR_LEN(p))+1)
159
e91177ed
IZ
160#define NODE_ALIGN(node)
161#define ARG_LOC(p) (((struct regnode_1 *)p)->arg1)
162#define ARG1_LOC(p) (((struct regnode_2 *)p)->arg1)
163#define ARG2_LOC(p) (((struct regnode_2 *)p)->arg2)
164#define NODE_STEP_REGNODE 1 /* sizeof(regnode)/sizeof(regnode) */
165#define EXTRA_STEP_2ARGS EXTRA_SIZE(struct regnode_2)
166
167#define NODE_STEP_B 4
c277df42
IZ
168
169#define NEXTOPER(p) ((p) + NODE_STEP_REGNODE)
170#define PREVOPER(p) ((p) - NODE_STEP_REGNODE)
171
e91177ed 172#define FILL_ADVANCE_NODE(ptr, op) STMT_START { \
c277df42 173 (ptr)->type = op; (ptr)->next_off = 0; (ptr)++; } STMT_END
e91177ed 174#define FILL_ADVANCE_NODE_ARG(ptr, op, arg) STMT_START { \
c277df42 175 ARG_SET(ptr, arg); FILL_ADVANCE_NODE(ptr, op); (ptr) += 1; } STMT_END
a687059c 176
22c35a8c 177#define REG_MAGIC 0234
a687059c 178
830247a4 179#define SIZE_ONLY (RExC_emit == &PL_regdummy)
c277df42 180
936ed897 181/* Flags for node->flags of ANYOF */
b8c5462f
JH
182
183#define ANYOF_CLASS 0x08
184#define ANYOF_INVERT 0x04
185#define ANYOF_FOLD 0x02
186#define ANYOF_LOCALE 0x01
187
653099ff
GS
188/* Used for regstclass only */
189#define ANYOF_EOS 0x10 /* Can match an empty string too */
190
936ed897 191/* Character classes for node->classflags of ANYOF */
653099ff
GS
192/* Should be synchronized with a table in regprop() */
193/* 2n should pair with 2n+1 */
b8c5462f 194
289d4f09 195#define ANYOF_ALNUM 0 /* \w, PL_utf8_alnum, utf8::IsWord, ALNUM */
b8c5462f 196#define ANYOF_NALNUM 1
aaa51d5e 197#define ANYOF_SPACE 2 /* \s */
b8c5462f
JH
198#define ANYOF_NSPACE 3
199#define ANYOF_DIGIT 4
200#define ANYOF_NDIGIT 5
289d4f09 201#define ANYOF_ALNUMC 6 /* isalnum(3), utf8::IsAlnum, ALNUMC */
b8c5462f
JH
202#define ANYOF_NALNUMC 7
203#define ANYOF_ALPHA 8
204#define ANYOF_NALPHA 9
205#define ANYOF_ASCII 10
206#define ANYOF_NASCII 11
207#define ANYOF_CNTRL 12
208#define ANYOF_NCNTRL 13
209#define ANYOF_GRAPH 14
210#define ANYOF_NGRAPH 15
211#define ANYOF_LOWER 16
212#define ANYOF_NLOWER 17
213#define ANYOF_PRINT 18
214#define ANYOF_NPRINT 19
215#define ANYOF_PUNCT 20
216#define ANYOF_NPUNCT 21
217#define ANYOF_UPPER 22
218#define ANYOF_NUPPER 23
219#define ANYOF_XDIGIT 24
220#define ANYOF_NXDIGIT 25
aaa51d5e
JF
221#define ANYOF_PSXSPC 26 /* POSIX space: \s plus the vertical tab */
222#define ANYOF_NPSXSPC 27
289d4f09
ML
223#define ANYOF_BLANK 28 /* GNU extension: space and tab */
224#define ANYOF_NBLANK 29
b8c5462f 225
aaa51d5e 226#define ANYOF_MAX 32
b8c5462f
JH
227
228/* Backward source code compatibility. */
229
230#define ANYOF_ALNUML ANYOF_ALNUM
231#define ANYOF_NALNUML ANYOF_NALNUM
232#define ANYOF_SPACEL ANYOF_SPACE
233#define ANYOF_NSPACEL ANYOF_NSPACE
234
235/* Utility macros for the bitmap and classes of ANYOF */
236
936ed897
IZ
237#define ANYOF_SIZE (sizeof(struct regnode_charclass))
238#define ANYOF_CLASS_SIZE (sizeof(struct regnode_charclass_class))
b8c5462f 239
936ed897 240#define ANYOF_FLAGS(p) ((p)->flags)
b8c5462f
JH
241#define ANYOF_FLAGS_ALL 0xff
242
243#define ANYOF_BIT(c) (1 << ((c) & 7))
244
936ed897 245#define ANYOF_CLASS_BYTE(p, c) (((struct regnode_charclass_class*)(p))->classflags[((c) >> 3) & 3])
b8c5462f
JH
246#define ANYOF_CLASS_SET(p, c) (ANYOF_CLASS_BYTE(p, c) |= ANYOF_BIT(c))
247#define ANYOF_CLASS_CLEAR(p, c) (ANYOF_CLASS_BYTE(p, c) &= ~ANYOF_BIT(c))
248#define ANYOF_CLASS_TEST(p, c) (ANYOF_CLASS_BYTE(p, c) & ANYOF_BIT(c))
249
936ed897
IZ
250#define ANYOF_CLASS_ZERO(ret) Zero(((struct regnode_charclass_class*)(ret))->classflags, ANYOF_CLASSBITMAP_SIZE, char)
251#define ANYOF_BITMAP_ZERO(ret) Zero(((struct regnode_charclass*)(ret))->bitmap, ANYOF_BITMAP_SIZE, char)
252
253#define ANYOF_BITMAP(p) (((struct regnode_charclass*)(p))->bitmap)
254#define ANYOF_BITMAP_BYTE(p, c) (ANYOF_BITMAP(p)[((c) >> 3) & 31])
b8c5462f
JH
255#define ANYOF_BITMAP_SET(p, c) (ANYOF_BITMAP_BYTE(p, c) |= ANYOF_BIT(c))
256#define ANYOF_BITMAP_CLEAR(p,c) (ANYOF_BITMAP_BYTE(p, c) &= ~ANYOF_BIT(c))
257#define ANYOF_BITMAP_TEST(p, c) (ANYOF_BITMAP_BYTE(p, c) & ANYOF_BIT(c))
258
936ed897
IZ
259#define ANYOF_SKIP ((ANYOF_SIZE - 1)/sizeof(regnode))
260#define ANYOF_CLASS_SKIP ((ANYOF_CLASS_SIZE - 1)/sizeof(regnode))
261#define ANYOF_CLASS_ADD_SKIP (ANYOF_CLASS_SKIP - ANYOF_SKIP)
c277df42 262
a687059c
LW
263/*
264 * Utility definitions.
265 */
266#ifndef lint
2304df62 267#ifndef CHARMASK
4d61ec05 268#define UCHARAT(p) ((int)*(U8*)(p))
a687059c 269#else
2304df62 270#define UCHARAT(p) ((int)*(p)&CHARMASK)
a687059c
LW
271#endif
272#else /* lint */
6b88bc9c 273#define UCHARAT(p) PL_regdummy
a687059c
LW
274#endif /* lint */
275
c277df42
IZ
276#define EXTRA_SIZE(guy) ((sizeof(guy)-1)/sizeof(struct regnode))
277
d09b2d29
IZ
278#define REG_SEEN_ZERO_LEN 1
279#define REG_SEEN_LOOKBEHIND 2
280#define REG_SEEN_GPOS 4
281#define REG_SEEN_EVAL 8
282
73c4f7a1
GS
283START_EXTERN_C
284
d09b2d29
IZ
285#include "regnodes.h"
286
4d61ec05 287/* The following have no fixed length. U8 so we can do strchr() on it. */
d09b2d29 288#ifndef DOINIT
4d61ec05 289EXTCONST U8 PL_varies[];
d09b2d29 290#else
4d61ec05 291EXTCONST U8 PL_varies[] = {
d09b2d29 292 BRANCH, BACK, STAR, PLUS, CURLY, CURLYX, REF, REFF, REFFL,
a0ed51b3 293 WHILEM, CURLYM, CURLYN, BRANCHJ, IFTHEN, SUSPEND, CLUMP, 0
c277df42 294};
d09b2d29 295#endif
c277df42 296
4d61ec05
GS
297/* The following always have a length of 1. U8 we can do strchr() on it. */
298/* (Note that length 1 means "one character" under UTF8, not "one octet".) */
d09b2d29 299#ifndef DOINIT
4d61ec05 300EXTCONST U8 PL_simple[];
d09b2d29 301#else
4d61ec05 302EXTCONST U8 PL_simple[] = {
22c35a8c 303 REG_ANY, ANYUTF8, SANY, SANYUTF8, ANYOF, ANYOFUTF8,
a0ed51b3
LW
304 ALNUM, ALNUMUTF8, ALNUML, ALNUMLUTF8,
305 NALNUM, NALNUMUTF8, NALNUML, NALNUMLUTF8,
306 SPACE, SPACEUTF8, SPACEL, SPACELUTF8,
307 NSPACE, NSPACEUTF8, NSPACEL, NSPACELUTF8,
308 DIGIT, DIGITUTF8, NDIGIT, NDIGITUTF8, 0
c277df42
IZ
309};
310#endif
311
73c4f7a1 312END_EXTERN_C
cad2e5aa
JH
313
314typedef struct re_scream_pos_data_s
315{
316 char **scream_olds; /* match pos */
317 I32 *scream_pos; /* Internal iterator of scream. */
318} re_scream_pos_data;
319
320struct reg_data {
321 U32 count;
322 U8 *what;
323 void* data[1];
324};
325
326struct reg_substr_datum {
327 I32 min_offset;
328 I32 max_offset;
329 SV *substr;
330};
331
332struct reg_substr_data {
333 struct reg_substr_datum data[3]; /* Actual array */
334};
335
336#define anchored_substr substrs->data[0].substr
337#define anchored_offset substrs->data[0].min_offset
338#define float_substr substrs->data[1].substr
339#define float_min_offset substrs->data[1].min_offset
340#define float_max_offset substrs->data[1].max_offset
341#define check_substr substrs->data[2].substr
342#define check_offset_min substrs->data[2].min_offset
343#define check_offset_max substrs->data[2].max_offset