tags/pcre-1.08/internal.h

/*************************************************
*      Perl-Compatible Regular Expressions       *
*************************************************/


#define PCRE_VERSION       "1.08 27-Mar-1998"


/* This is a library of functions to support regular expressions whose syntax
and semantics are as close as possible to those of the Perl 5 language. See
the file Tech.Notes for some information on the internals.

Written by: Philip Hazel <ph10@cam.ac.uk>

           Copyright (c) 1998 University of Cambridge

-----------------------------------------------------------------------------
Permission is granted to anyone to use this software for any purpose on any
computer system, and to redistribute it freely, subject to the following
restrictions:

1. This software is distributed in the hope that it will be useful,
   but WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

2. The origin of this software must not be misrepresented, either by
   explicit claim or by omission.

3. Altered versions must be plainly marked as such, and must not be
   misrepresented as being the original software.
-----------------------------------------------------------------------------
*/

/* This header contains definitions that are shared between the different
modules, but which are not relevant to the outside. */

/* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
define a macro for memmove() if USE_BCOPY is defined. */

#ifdef USE_BCOPY
#undef  memmove        /* some systems may have a macro */
#define memmove(a, b, c) bcopy(b, a, c)
#endif

/* Standard C headers plus the external interface definition */

#include <ctype.h>
#include <limits.h>
#include <setjmp.h>
#include <stddef.h>
#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include "pcre.h"

/* In case there is no definition of offsetof() provided - though any proper
Standard C system should have one. */

#ifndef offsetof
#define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
#endif

/* Private options flags start at the most significant end of the two bytes.
The public options defined in pcre.h start at the least significant end. Make
sure they don't overlap! */

#define PCRE_FIRSTSET           0x8000  /* first_char is set */
#define PCRE_STARTLINE          0x4000  /* start after \n for multiline */
#define PCRE_COMPILED_CASELESS  0x2000  /* like it says */

/* Options for the "extra" block produced by pcre_study(). */

#define PCRE_STUDY_CASELESS 0x01     /* study was caseless */
#define PCRE_STUDY_MAPPED   0x02     /* a map of starting chars exists */

/* Masks for identifying the public options: all permitted at compile time,
only some permitted at run or study time. */

#define PUBLIC_OPTIONS \
  (PCRE_CASELESS|PCRE_EXTENDED|PCRE_ANCHORED|PCRE_MULTILINE| \
   PCRE_DOTALL|PCRE_DOLLAR_ENDONLY|PCRE_EXTRA|PCRE_UNGREEDY)

#define PUBLIC_EXEC_OPTIONS \
  (PCRE_CASELESS|PCRE_ANCHORED|PCRE_MULTILINE|PCRE_NOTBOL|PCRE_NOTEOL| \
   PCRE_DOTALL|PCRE_DOLLAR_ENDONLY)

#define PUBLIC_STUDY_OPTIONS (PCRE_CASELESS)

/* Magic number to provide a small check against being handed junk. */

#define MAGIC_NUMBER  0x50435245   /* 'PCRE' */

/* Miscellaneous definitions */

typedef int BOOL;

#define FALSE   0
#define TRUE    1

/* These are escaped items that aren't just an encoding of a particular data
value such as \n. They must have non-zero values, as check_escape() returns
their negation. Also, they must appear in the same order as in the opcode
definitions below, up to ESC_Z. The final one must be ESC_REF as subsequent
values are used for \1, \2, \3, etc. There is a test in the code for an escape
greater than ESC_b and less than ESC_X to detect the types that may be
repeated. If any new escapes are put in-between that don't consume a character,
that code will have to change. */

enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,

                    /* These are not Perl escapes, so can't appear in the */
       ESC_X,       /* simple table-lookup because they must be conditional */
                    /* on PCRE_EXTRA. */
       ESC_Z,
       ESC_REF };

/* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
OP_EOD must correspond in order to the list of escapes immediately above. */

enum {
  OP_END,            /* End of pattern */

  /* Values corresponding to backslashed metacharacters */

  OP_SOD,            /* Start of data: \A */
  OP_NOT_WORD_BOUNDARY,  /* \B */
  OP_WORD_BOUNDARY,      /* \b */
  OP_NOT_DIGIT,          /* \D */
  OP_DIGIT,              /* \d */
  OP_NOT_WHITESPACE,     /* \S */
  OP_WHITESPACE,         /* \s */
  OP_NOT_WORDCHAR,       /* \W */
  OP_WORDCHAR,           /* \w */
  OP_CUT,            /* The analogue of Prolog's "cut" operation (extension) */
  OP_EOD,            /* End of data: \Z. */

  OP_CIRC,           /* Start of line - varies with multiline switch */
  OP_DOLL,           /* End of line - varies with multiline switch */
  OP_ANY,            /* Match any character */
  OP_CHARS,          /* Match string of characters */
  OP_NOT,            /* Match anything but the following char */

  OP_STAR,           /* The maximizing and minimizing versions of */
  OP_MINSTAR,        /* all these opcodes must come in pairs, with */
  OP_PLUS,           /* the minimizing one second. */
  OP_MINPLUS,        /* This first set applies to single characters */
  OP_QUERY,
  OP_MINQUERY,
  OP_UPTO,           /* From 0 to n matches */
  OP_MINUPTO,
  OP_EXACT,          /* Exactly n matches */

  OP_NOTSTAR,        /* The maximizing and minimizing versions of */
  OP_NOTMINSTAR,     /* all these opcodes must come in pairs, with */
  OP_NOTPLUS,        /* the minimizing one second. */
  OP_NOTMINPLUS,     /* This first set applies to "not" single characters */
  OP_NOTQUERY,
  OP_NOTMINQUERY,
  OP_NOTUPTO,        /* From 0 to n matches */
  OP_NOTMINUPTO,
  OP_NOTEXACT,       /* Exactly n matches */

  OP_TYPESTAR,       /* The maximizing and minimizing versions of */
  OP_TYPEMINSTAR,    /* all these opcodes must come in pairs, with */
  OP_TYPEPLUS,       /* the minimizing one second. These codes must */
  OP_TYPEMINPLUS,    /* be in exactly the same order as those above. */
  OP_TYPEQUERY,      /* This set applies to character types such as \d */
  OP_TYPEMINQUERY,
  OP_TYPEUPTO,       /* From 0 to n matches */
  OP_TYPEMINUPTO,
  OP_TYPEEXACT,      /* Exactly n matches */

  OP_CRSTAR,         /* The maximizing and minimizing versions of */
  OP_CRMINSTAR,      /* all these opcodes must come in pairs, with */
  OP_CRPLUS,         /* the minimizing one second. These codes must */
  OP_CRMINPLUS,      /* be in exactly the same order as those above. */
  OP_CRQUERY,        /* These are for character classes and back refs */
  OP_CRMINQUERY,
  OP_CRRANGE,        /* These are different to the three seta above. */
  OP_CRMINRANGE,

  OP_CLASS,          /* Match a character class */
  OP_NEGCLASS,       /* Match a character class, specified negatively */
  OP_REF,            /* Match a back reference */

  OP_ALT,            /* Start of alternation */
  OP_KET,            /* End of group that doesn't have an unbounded repeat */
  OP_KETRMAX,        /* These two must remain together and in this */
  OP_KETRMIN,        /* order. They are for groups the repeat for ever. */

  OP_ASSERT,
  OP_ASSERT_NOT,
  OP_ONCE,           /* Once matched, don't back up into the subpattern */

  OP_BRAZERO,        /* These two must remain together and in this */
  OP_BRAMINZERO,     /* order. */

  OP_BRA             /* This and greater values are used for brackets that
                        extract substrings. */
};

/* The highest extraction number. This is limited by the number of opcodes
left after OP_BRA, i.e. 255 - OP_BRA. We actually set it somewhat lower. */

#define EXTRACT_MAX  99

/* The texts of compile-time error messages are defined as macros here so that
they can be accessed by the POSIX wrapper and converted into error codes.  Yes,
I could have used error codes in the first place, but didn't feel like changing
just to accommodate the POSIX wrapper. */

#define ERR1  "\\ at end of pattern"
#define ERR2  "\\c at end of pattern"
#define ERR3  "unrecognized character follows \\"
#define ERR4  "numbers out of order in {} quantifier"
#define ERR5  "number too big in {} quantifier"
#define ERR6  "missing terminating ] for character class"
#define ERR7  "invalid escape sequence in character class"
#define ERR8  "range out of order in character class"
#define ERR9  "nothing to repeat"
#define ERR10 "operand of unlimited repeat could match the empty string"
#define ERR11 "internal error: unexpected repeat"
#define ERR12 "unrecognized character after (?"
#define ERR13 "too many capturing parenthesized sub-patterns"
#define ERR14 "missing )"
#define ERR15 "back reference to non-existent subpattern"
#define ERR16 "erroffset passed as NULL"
#define ERR17 "unknown option bit(s) set"
#define ERR18 "missing ) after comment"
#define ERR19 "too many sets of parentheses"
#define ERR20 "regular expression too large"
#define ERR21 "failed to get memory"
#define ERR22 "unmatched brackets"
#define ERR23 "internal error: code overflow"

/* All character handling must be done as unsigned characters. Otherwise there
are problems with top-bit-set characters and functions such as isspace().
However, we leave the interface to the outside world as char *, because that
should make things easier for callers. We define a short type for unsigned char
to save lots of typing. I tried "uchar", but it causes problems on Digital
Unix, where it is defined in sys/types, so use "uschar" instead. */

typedef unsigned char uschar;

/* The real format of the start of the pcre block; the actual code vector
runs on as long as necessary after the end. */

typedef struct real_pcre {
  unsigned int  magic_number;
  unsigned short int options;
  unsigned char top_bracket;
  unsigned char top_backref;
  unsigned char first_char;
  unsigned char code[1];
} real_pcre;

/* The real format of the extra block returned by pcre_study(). */

typedef struct real_pcre_extra {
  unsigned char options;
  unsigned char start_bits[32];
} real_pcre_extra;

/* Global tables from chartables.c */

extern uschar pcre_lcc[];
extern uschar pcre_fcc[];
extern uschar pcre_cbits[];
extern uschar pcre_ctypes[];

/* Bit definitions for entries in pcre_ctypes[]. */

#define ctype_space   0x01
#define ctype_letter  0x02
#define ctype_digit   0x04
#define ctype_xdigit  0x08
#define ctype_word    0x10   /* alphameric or '_' */
#define ctype_meta    0x80   /* regexp meta char or zero (end pattern) */

/* Offsets for the bitmap tables */

#define cbit_digit    0
#define cbit_letter  32
#define cbit_word    64
#define cbit_space   96
#define cbit_length 128      /* Length of the cbits table */

/* End of internal.h */
1	/*************************************************
2	* Perl-Compatible Regular Expressions *
3	*************************************************/
4
5
6	#define PCRE_VERSION "1.08 27-Mar-1998"
7
8
9	/* This is a library of functions to support regular expressions whose syntax
10	and semantics are as close as possible to those of the Perl 5 language. See
11	the file Tech.Notes for some information on the internals.
12
13	Written by: Philip Hazel <ph10@cam.ac.uk>
14
15	Copyright (c) 1998 University of Cambridge
16
17	-----------------------------------------------------------------------------
18	Permission is granted to anyone to use this software for any purpose on any
19	computer system, and to redistribute it freely, subject to the following
20	restrictions:
21
22	1. This software is distributed in the hope that it will be useful,
23	but WITHOUT ANY WARRANTY; without even the implied warranty of
24	MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
25
26	2. The origin of this software must not be misrepresented, either by
27	explicit claim or by omission.
28
29	3. Altered versions must be plainly marked as such, and must not be
30	misrepresented as being the original software.
31	-----------------------------------------------------------------------------
32	*/
33
34	/* This header contains definitions that are shared between the different
35	modules, but which are not relevant to the outside. */
36
37	/* To cope with SunOS4 and other systems that lack memmove() but have bcopy(),
38	define a macro for memmove() if USE_BCOPY is defined. */
39
40	#ifdef USE_BCOPY
41	#undef memmove /* some systems may have a macro */
42	#define memmove(a, b, c) bcopy(b, a, c)
43	#endif
44
45	/* Standard C headers plus the external interface definition */
46
47	#include <ctype.h>
48	#include <limits.h>
49	#include <setjmp.h>
50	#include <stddef.h>
51	#include <stdio.h>
52	#include <stdlib.h>
53	#include <string.h>
54	#include "pcre.h"
55
56	/* In case there is no definition of offsetof() provided - though any proper
57	Standard C system should have one. */
58
59	#ifndef offsetof
60	#define offsetof(p_type,field) ((size_t)&(((p_type *)0)->field))
61	#endif
62
63	/* Private options flags start at the most significant end of the two bytes.
64	The public options defined in pcre.h start at the least significant end. Make
65	sure they don't overlap! */
66
67	#define PCRE_FIRSTSET 0x8000 /* first_char is set */
68	#define PCRE_STARTLINE 0x4000 /* start after \n for multiline */
69	#define PCRE_COMPILED_CASELESS 0x2000 /* like it says */
70
71	/* Options for the "extra" block produced by pcre_study(). */
72
73	#define PCRE_STUDY_CASELESS 0x01 /* study was caseless */
74	#define PCRE_STUDY_MAPPED 0x02 /* a map of starting chars exists */
75
76	/* Masks for identifying the public options: all permitted at compile time,
77	only some permitted at run or study time. */
78
79	#define PUBLIC_OPTIONS \
80	(PCRE_CASELESS\|PCRE_EXTENDED\|PCRE_ANCHORED\|PCRE_MULTILINE\| \
81	PCRE_DOTALL\|PCRE_DOLLAR_ENDONLY\|PCRE_EXTRA\|PCRE_UNGREEDY)
82
83	#define PUBLIC_EXEC_OPTIONS \
84	(PCRE_CASELESS\|PCRE_ANCHORED\|PCRE_MULTILINE\|PCRE_NOTBOL\|PCRE_NOTEOL\| \
85	PCRE_DOTALL\|PCRE_DOLLAR_ENDONLY)
86
87	#define PUBLIC_STUDY_OPTIONS (PCRE_CASELESS)
88
89	/* Magic number to provide a small check against being handed junk. */
90
91	#define MAGIC_NUMBER 0x50435245 /* 'PCRE' */
92
93	/* Miscellaneous definitions */
94
95	typedef int BOOL;
96
97	#define FALSE 0
98	#define TRUE 1
99
100	/* These are escaped items that aren't just an encoding of a particular data
101	value such as \n. They must have non-zero values, as check_escape() returns
102	their negation. Also, they must appear in the same order as in the opcode
103	definitions below, up to ESC_Z. The final one must be ESC_REF as subsequent
104	values are used for \1, \2, \3, etc. There is a test in the code for an escape
105	greater than ESC_b and less than ESC_X to detect the types that may be
106	repeated. If any new escapes are put in-between that don't consume a character,
107	that code will have to change. */
108
109	enum { ESC_A = 1, ESC_B, ESC_b, ESC_D, ESC_d, ESC_S, ESC_s, ESC_W, ESC_w,
110
111	/* These are not Perl escapes, so can't appear in the */
112	ESC_X, /* simple table-lookup because they must be conditional */
113	/* on PCRE_EXTRA. */
114	ESC_Z,
115	ESC_REF };
116
117	/* Opcode table: OP_BRA must be last, as all values >= it are used for brackets
118	that extract substrings. Starting from 1 (i.e. after OP_END), the values up to
119	OP_EOD must correspond in order to the list of escapes immediately above. */
120
121	enum {
122	OP_END, /* End of pattern */
123
124	/* Values corresponding to backslashed metacharacters */
125
126	OP_SOD, /* Start of data: \A */
127	OP_NOT_WORD_BOUNDARY, /* \B */
128	OP_WORD_BOUNDARY, /* \b */
129	OP_NOT_DIGIT, /* \D */
130	OP_DIGIT, /* \d */
131	OP_NOT_WHITESPACE, /* \S */
132	OP_WHITESPACE, /* \s */
133	OP_NOT_WORDCHAR, /* \W */
134	OP_WORDCHAR, /* \w */
135	OP_CUT, /* The analogue of Prolog's "cut" operation (extension) */
136	OP_EOD, /* End of data: \Z. */
137
138	OP_CIRC, /* Start of line - varies with multiline switch */
139	OP_DOLL, /* End of line - varies with multiline switch */
140	OP_ANY, /* Match any character */
141	OP_CHARS, /* Match string of characters */
142	OP_NOT, /* Match anything but the following char */
143
144	OP_STAR, /* The maximizing and minimizing versions of */
145	OP_MINSTAR, /* all these opcodes must come in pairs, with */
146	OP_PLUS, /* the minimizing one second. */
147	OP_MINPLUS, /* This first set applies to single characters */
148	OP_QUERY,
149	OP_MINQUERY,
150	OP_UPTO, /* From 0 to n matches */
151	OP_MINUPTO,
152	OP_EXACT, /* Exactly n matches */
153
154	OP_NOTSTAR, /* The maximizing and minimizing versions of */
155	OP_NOTMINSTAR, /* all these opcodes must come in pairs, with */
156	OP_NOTPLUS, /* the minimizing one second. */
157	OP_NOTMINPLUS, /* This first set applies to "not" single characters */
158	OP_NOTQUERY,
159	OP_NOTMINQUERY,
160	OP_NOTUPTO, /* From 0 to n matches */
161	OP_NOTMINUPTO,
162	OP_NOTEXACT, /* Exactly n matches */
163
164	OP_TYPESTAR, /* The maximizing and minimizing versions of */
165	OP_TYPEMINSTAR, /* all these opcodes must come in pairs, with */
166	OP_TYPEPLUS, /* the minimizing one second. These codes must */
167	OP_TYPEMINPLUS, /* be in exactly the same order as those above. */
168	OP_TYPEQUERY, /* This set applies to character types such as \d */
169	OP_TYPEMINQUERY,
170	OP_TYPEUPTO, /* From 0 to n matches */
171	OP_TYPEMINUPTO,
172	OP_TYPEEXACT, /* Exactly n matches */
173
174	OP_CRSTAR, /* The maximizing and minimizing versions of */
175	OP_CRMINSTAR, /* all these opcodes must come in pairs, with */
176	OP_CRPLUS, /* the minimizing one second. These codes must */
177	OP_CRMINPLUS, /* be in exactly the same order as those above. */
178	OP_CRQUERY, /* These are for character classes and back refs */
179	OP_CRMINQUERY,
180	OP_CRRANGE, /* These are different to the three seta above. */
181	OP_CRMINRANGE,
182
183	OP_CLASS, /* Match a character class */
184	OP_NEGCLASS, /* Match a character class, specified negatively */
185	OP_REF, /* Match a back reference */
186
187	OP_ALT, /* Start of alternation */
188	OP_KET, /* End of group that doesn't have an unbounded repeat */
189	OP_KETRMAX, /* These two must remain together and in this */
190	OP_KETRMIN, /* order. They are for groups the repeat for ever. */
191
192	OP_ASSERT,
193	OP_ASSERT_NOT,
194	OP_ONCE, /* Once matched, don't back up into the subpattern */
195
196	OP_BRAZERO, /* These two must remain together and in this */
197	OP_BRAMINZERO, /* order. */
198
199	OP_BRA /* This and greater values are used for brackets that
200	extract substrings. */
201	};
202
203	/* The highest extraction number. This is limited by the number of opcodes
204	left after OP_BRA, i.e. 255 - OP_BRA. We actually set it somewhat lower. */
205
206	#define EXTRACT_MAX 99
207
208	/* The texts of compile-time error messages are defined as macros here so that
209	they can be accessed by the POSIX wrapper and converted into error codes. Yes,
210	I could have used error codes in the first place, but didn't feel like changing
211	just to accommodate the POSIX wrapper. */
212
213	#define ERR1 "\\ at end of pattern"
214	#define ERR2 "\\c at end of pattern"
215	#define ERR3 "unrecognized character follows \\"
216	#define ERR4 "numbers out of order in {} quantifier"
217	#define ERR5 "number too big in {} quantifier"
218	#define ERR6 "missing terminating ] for character class"
219	#define ERR7 "invalid escape sequence in character class"
220	#define ERR8 "range out of order in character class"
221	#define ERR9 "nothing to repeat"
222	#define ERR10 "operand of unlimited repeat could match the empty string"
223	#define ERR11 "internal error: unexpected repeat"
224	#define ERR12 "unrecognized character after (?"
225	#define ERR13 "too many capturing parenthesized sub-patterns"
226	#define ERR14 "missing )"
227	#define ERR15 "back reference to non-existent subpattern"
228	#define ERR16 "erroffset passed as NULL"
229	#define ERR17 "unknown option bit(s) set"
230	#define ERR18 "missing ) after comment"
231	#define ERR19 "too many sets of parentheses"
232	#define ERR20 "regular expression too large"
233	#define ERR21 "failed to get memory"
234	#define ERR22 "unmatched brackets"
235	#define ERR23 "internal error: code overflow"
236
237	/* All character handling must be done as unsigned characters. Otherwise there
238	are problems with top-bit-set characters and functions such as isspace().
239	However, we leave the interface to the outside world as char *, because that
240	should make things easier for callers. We define a short type for unsigned char
241	to save lots of typing. I tried "uchar", but it causes problems on Digital
242	Unix, where it is defined in sys/types, so use "uschar" instead. */
243
244	typedef unsigned char uschar;
245
246	/* The real format of the start of the pcre block; the actual code vector
247	runs on as long as necessary after the end. */
248
249	typedef struct real_pcre {
250	unsigned int magic_number;
251	unsigned short int options;
252	unsigned char top_bracket;
253	unsigned char top_backref;
254	unsigned char first_char;
255	unsigned char code[1];
256	} real_pcre;
257
258	/* The real format of the extra block returned by pcre_study(). */
259
260	typedef struct real_pcre_extra {
261	unsigned char options;
262	unsigned char start_bits[32];
263	} real_pcre_extra;
264
265	/* Global tables from chartables.c */
266
267	extern uschar pcre_lcc[];
268	extern uschar pcre_fcc[];
269	extern uschar pcre_cbits[];
270	extern uschar pcre_ctypes[];
271
272	/* Bit definitions for entries in pcre_ctypes[]. */
273
274	#define ctype_space 0x01
275	#define ctype_letter 0x02
276	#define ctype_digit 0x04
277	#define ctype_xdigit 0x08
278	#define ctype_word 0x10 /* alphameric or '_' */
279	#define ctype_meta 0x80 /* regexp meta char or zero (end pattern) */
280
281	/* Offsets for the bitmap tables */
282
283	#define cbit_digit 0
284	#define cbit_letter 32
285	#define cbit_word 64
286	#define cbit_space 96
287	#define cbit_length 128 /* Length of the cbits table */
288
289	/* End of internal.h */