This is a live mirror of the Perl 5 development currently hosted at https://github.com/perl/perl5
fixes for logical bugs in the lexwarn patch; other tweaks to avoid
[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
IZ
68struct regnode_string {
69 U8 flags;
70 U8 type;
71 U16 next_off;
72 U8 string[1];
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
9c7e81e8
AD
90/* XXX fix this description.
91 Impose a limit of REG_INFTY on various pattern matching operations
92 to limit stack growth and to avoid "infinite" recursions.
93*/
94/* The default size for REG_INFTY is I16_MAX, which is the same as
95 SHORT_MAX (see perl.h). Unfortunately I16 isn't necessarily 16 bits
96 (see handy.h). On the Cray C90, sizeof(short)==4 and hence I16_MAX is
97 ((1<<31)-1), while on the Cray T90, sizeof(short)==8 and I16_MAX is
98 ((1<<63)-1). To limit stack growth to reasonable sizes, supply a
99 smaller default.
100 --Andy Dougherty 11 June 1998
101*/
102#if SHORTSIZE > 2
103# ifndef REG_INFTY
104# define REG_INFTY ((1<<15)-1)
105# endif
106#endif
107
108#ifndef REG_INFTY
83cfe48c
IZ
109# define REG_INFTY I16_MAX
110#endif
a687059c 111
e91177ed
IZ
112#define ARG_VALUE(arg) (arg)
113#define ARG__SET(arg,val) ((arg) = (val))
c277df42
IZ
114
115#define ARG(p) ARG_VALUE(ARG_LOC(p))
116#define ARG1(p) ARG_VALUE(ARG1_LOC(p))
117#define ARG2(p) ARG_VALUE(ARG2_LOC(p))
118#define ARG_SET(p, val) ARG__SET(ARG_LOC(p), (val))
119#define ARG1_SET(p, val) ARG__SET(ARG1_LOC(p), (val))
120#define ARG2_SET(p, val) ARG__SET(ARG2_LOC(p), (val))
121
122#ifndef lint
e91177ed
IZ
123# define NEXT_OFF(p) ((p)->next_off)
124# define NODE_ALIGN(node)
125# define NODE_ALIGN_FILL(node) ((node)->flags = 0xde) /* deadbeef */
a687059c 126#else /* lint */
c277df42 127# define NEXT_OFF(p) 0
e91177ed
IZ
128# define NODE_ALIGN(node)
129# define NODE_ALIGN_FILL(node)
a687059c
LW
130#endif /* lint */
131
c277df42
IZ
132#define SIZE_ALIGN NODE_ALIGN
133
e91177ed
IZ
134#define OP(p) ((p)->type)
135#define OPERAND(p) (((struct regnode_string *)p)->string)
136#define NODE_ALIGN(node)
137#define ARG_LOC(p) (((struct regnode_1 *)p)->arg1)
138#define ARG1_LOC(p) (((struct regnode_2 *)p)->arg1)
139#define ARG2_LOC(p) (((struct regnode_2 *)p)->arg2)
140#define NODE_STEP_REGNODE 1 /* sizeof(regnode)/sizeof(regnode) */
141#define EXTRA_STEP_2ARGS EXTRA_SIZE(struct regnode_2)
142
143#define NODE_STEP_B 4
c277df42
IZ
144
145#define NEXTOPER(p) ((p) + NODE_STEP_REGNODE)
146#define PREVOPER(p) ((p) - NODE_STEP_REGNODE)
147
e91177ed 148#define FILL_ADVANCE_NODE(ptr, op) STMT_START { \
c277df42 149 (ptr)->type = op; (ptr)->next_off = 0; (ptr)++; } STMT_END
e91177ed 150#define FILL_ADVANCE_NODE_ARG(ptr, op, arg) STMT_START { \
c277df42 151 ARG_SET(ptr, arg); FILL_ADVANCE_NODE(ptr, op); (ptr) += 1; } STMT_END
a687059c 152
22c35a8c 153#define REG_MAGIC 0234
a687059c 154
3280af22 155#define SIZE_ONLY (PL_regcode == &PL_regdummy)
c277df42 156
b8c5462f
JH
157/* Flags for first parameter byte [0] of ANYOF */
158
159#define ANYOF_CLASS 0x08
160#define ANYOF_INVERT 0x04
161#define ANYOF_FOLD 0x02
162#define ANYOF_LOCALE 0x01
163
164/* Character classes for bytes [1..4] of ANYOF */
165
166#define ANYOF_ALNUM 0 /* \w, utf8::IsWord, isALNUM() */
167#define ANYOF_NALNUM 1
168#define ANYOF_SPACE 2
169#define ANYOF_NSPACE 3
170#define ANYOF_DIGIT 4
171#define ANYOF_NDIGIT 5
172#define ANYOF_ALNUMC 6 /* isalnum(3), utf8::IsAlnum, isALNUMC() */
173#define ANYOF_NALNUMC 7
174#define ANYOF_ALPHA 8
175#define ANYOF_NALPHA 9
176#define ANYOF_ASCII 10
177#define ANYOF_NASCII 11
178#define ANYOF_CNTRL 12
179#define ANYOF_NCNTRL 13
180#define ANYOF_GRAPH 14
181#define ANYOF_NGRAPH 15
182#define ANYOF_LOWER 16
183#define ANYOF_NLOWER 17
184#define ANYOF_PRINT 18
185#define ANYOF_NPRINT 19
186#define ANYOF_PUNCT 20
187#define ANYOF_NPUNCT 21
188#define ANYOF_UPPER 22
189#define ANYOF_NUPPER 23
190#define ANYOF_XDIGIT 24
191#define ANYOF_NXDIGIT 25
192
193#define ANYOF_MAX 31
194
195/* Backward source code compatibility. */
196
197#define ANYOF_ALNUML ANYOF_ALNUM
198#define ANYOF_NALNUML ANYOF_NALNUM
199#define ANYOF_SPACEL ANYOF_SPACE
200#define ANYOF_NSPACEL ANYOF_NSPACE
201
202/* Utility macros for the bitmap and classes of ANYOF */
203
204#define ANYOF_OPND_SIZE 1
205#define ANYOF_CLASS_SIZE 4
206#define ANYOF_BITMAP_SIZE 32 /* 256 b/(8 b/B) */
207#define ANYOF_SIZE (ANYOF_OPND_SIZE+ANYOF_CLASS_SIZE+ANYOF_BITMAP_SIZE)
208
209#define ANYOF_FLAGS(p) ((p)[0])
210#define ANYOF_FLAGS_ALL 0xff
211
212#define ANYOF_BIT(c) (1 << ((c) & 7))
213
214#define ANYOF_CLASS_OFFSET ANYOF_OPND_SIZE
215#define ANYOF_CLASS_BYTE(p, c) ((p)[ANYOF_CLASS_OFFSET + (((c) >> 3) & 3)])
216#define ANYOF_CLASS_SET(p, c) (ANYOF_CLASS_BYTE(p, c) |= ANYOF_BIT(c))
217#define ANYOF_CLASS_CLEAR(p, c) (ANYOF_CLASS_BYTE(p, c) &= ~ANYOF_BIT(c))
218#define ANYOF_CLASS_TEST(p, c) (ANYOF_CLASS_BYTE(p, c) & ANYOF_BIT(c))
219
220#define ANYOF_BITMAP_OFFSET (ANYOF_CLASS_OFFSET+ANYOF_CLASS_SIZE)
221#define ANYOF_BITMAP_BYTE(p, c) ((p)[ANYOF_BITMAP_OFFSET + (((c) >> 3) & 31)])
222#define ANYOF_BITMAP_SET(p, c) (ANYOF_BITMAP_BYTE(p, c) |= ANYOF_BIT(c))
223#define ANYOF_BITMAP_CLEAR(p,c) (ANYOF_BITMAP_BYTE(p, c) &= ~ANYOF_BIT(c))
224#define ANYOF_BITMAP_TEST(p, c) (ANYOF_BITMAP_BYTE(p, c) & ANYOF_BIT(c))
225
226#define ANY_SKIP ((ANYOF_SIZE - 1)/sizeof(regnode) + 1)
c277df42 227
a687059c
LW
228/*
229 * Utility definitions.
230 */
231#ifndef lint
2304df62 232#ifndef CHARMASK
a687059c
LW
233#define UCHARAT(p) ((int)*(unsigned char *)(p))
234#else
2304df62 235#define UCHARAT(p) ((int)*(p)&CHARMASK)
a687059c
LW
236#endif
237#else /* lint */
6b88bc9c 238#define UCHARAT(p) PL_regdummy
a687059c
LW
239#endif /* lint */
240
34184a49 241#define FAIL(m) \
cea2e8a9
GS
242 STMT_START { \
243 if (!SIZE_ONLY) \
244 SAVEDESTRUCTOR(S_clear_re,(void*)PL_regcomp_rx); \
245 Perl_croak(aTHX_ "/%.127s/: %s", PL_regprecomp,m); \
34184a49
GS
246 } STMT_END
247
248#define FAIL2(pat,m) \
cea2e8a9
GS
249 STMT_START { \
250 if (!SIZE_ONLY) \
251 SAVEDESTRUCTOR(S_clear_re,(void*)PL_regcomp_rx); \
252 S_re_croak2(aTHX_ "/%.127s/: ",pat,PL_regprecomp,m); \
34184a49 253 } STMT_END
c277df42
IZ
254
255#define EXTRA_SIZE(guy) ((sizeof(guy)-1)/sizeof(struct regnode))
256
d09b2d29
IZ
257#define REG_SEEN_ZERO_LEN 1
258#define REG_SEEN_LOOKBEHIND 2
259#define REG_SEEN_GPOS 4
260#define REG_SEEN_EVAL 8
261
73c4f7a1
GS
262START_EXTERN_C
263
d09b2d29
IZ
264#include "regnodes.h"
265
266/* The following have no fixed length. char* since we do strchr on it. */
267#ifndef DOINIT
22c35a8c 268EXTCONST char PL_varies[];
d09b2d29 269#else
22c35a8c 270EXTCONST char PL_varies[] = {
d09b2d29 271 BRANCH, BACK, STAR, PLUS, CURLY, CURLYX, REF, REFF, REFFL,
a0ed51b3 272 WHILEM, CURLYM, CURLYN, BRANCHJ, IFTHEN, SUSPEND, CLUMP, 0
c277df42 273};
d09b2d29 274#endif
c277df42 275
d09b2d29 276/* The following always have a length of 1. char* since we do strchr on it. */
a0ed51b3 277/* (Note that lenght 1 means "one character" under UTF8, not "one octet".) */
d09b2d29 278#ifndef DOINIT
22c35a8c 279EXTCONST char PL_simple[];
d09b2d29 280#else
22c35a8c
GS
281EXTCONST char PL_simple[] = {
282 REG_ANY, ANYUTF8, SANY, SANYUTF8, ANYOF, ANYOFUTF8,
a0ed51b3
LW
283 ALNUM, ALNUMUTF8, ALNUML, ALNUMLUTF8,
284 NALNUM, NALNUMUTF8, NALNUML, NALNUMLUTF8,
285 SPACE, SPACEUTF8, SPACEL, SPACELUTF8,
286 NSPACE, NSPACEUTF8, NSPACEL, NSPACELUTF8,
287 DIGIT, DIGITUTF8, NDIGIT, NDIGITUTF8, 0
c277df42
IZ
288};
289#endif
290
73c4f7a1 291END_EXTERN_C
cad2e5aa
JH
292
293typedef struct re_scream_pos_data_s
294{
295 char **scream_olds; /* match pos */
296 I32 *scream_pos; /* Internal iterator of scream. */
297} re_scream_pos_data;
298
299struct reg_data {
300 U32 count;
301 U8 *what;
302 void* data[1];
303};
304
305struct reg_substr_datum {
306 I32 min_offset;
307 I32 max_offset;
308 SV *substr;
309};
310
311struct reg_substr_data {
312 struct reg_substr_datum data[3]; /* Actual array */
313};
314
315#define anchored_substr substrs->data[0].substr
316#define anchored_offset substrs->data[0].min_offset
317#define float_substr substrs->data[1].substr
318#define float_min_offset substrs->data[1].min_offset
319#define float_max_offset substrs->data[1].max_offset
320#define check_substr substrs->data[2].substr
321#define check_offset_min substrs->data[2].min_offset
322#define check_offset_max substrs->data[2].max_offset