minix/lib/libc/regex/regcomp.c

1945 lines
44 KiB
C
Raw Normal View History

/* $NetBSD: regcomp.c,v 1.33 2012/03/13 21:13:43 christos Exp $ */
2005-04-21 16:53:53 +02:00
/*-
* Copyright (c) 1992, 1993, 1994
* The Regents of the University of California. All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Henry Spencer.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)regcomp.c 8.5 (Berkeley) 3/20/94
*/
/*-
* Copyright (c) 1992, 1993, 1994 Henry Spencer.
*
* This code is derived from software contributed to Berkeley by
* Henry Spencer.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
2005-04-21 16:53:53 +02:00
* 3. All advertising materials mentioning features or use of this software
* must display the following acknowledgement:
* This product includes software developed by the University of
* California, Berkeley and its contributors.
* 4. Neither the name of the University nor the names of its contributors
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* @(#)regcomp.c 8.5 (Berkeley) 3/20/94
*/
#include <sys/cdefs.h>
2005-04-21 16:53:53 +02:00
#if defined(LIBC_SCCS) && !defined(lint)
#if 0
2005-04-21 16:53:53 +02:00
static char sccsid[] = "@(#)regcomp.c 8.5 (Berkeley) 3/20/94";
#else
__RCSID("$NetBSD: regcomp.c,v 1.33 2012/03/13 21:13:43 christos Exp $");
#endif
2005-04-21 16:53:53 +02:00
#endif /* LIBC_SCCS and not lint */
#include "namespace.h"
2005-04-21 16:53:53 +02:00
#include <sys/types.h>
#include <assert.h>
2005-04-21 16:53:53 +02:00
#include <ctype.h>
#include <limits.h>
#include <stdio.h>
2005-04-21 16:53:53 +02:00
#include <stdlib.h>
#include <string.h>
2005-04-21 16:53:53 +02:00
#include <regex.h>
#ifdef __weak_alias
__weak_alias(regcomp,_regcomp)
#endif
2005-04-21 16:53:53 +02:00
#include "utils.h"
#include "regex2.h"
#include "cclass.h"
#include "cname.h"
/*
* parse structure, passed up and down to avoid global variables and
* other clumsinesses
*/
struct parse {
const char *next; /* next character in RE */
const char *end; /* end of string (-> NUL normally) */
2005-04-21 16:53:53 +02:00
int error; /* has an error been seen? */
sop *strip; /* malloced strip */
sopno ssize; /* malloced strip size (allocated) */
sopno slen; /* malloced strip length (used) */
size_t ncsalloc; /* number of csets allocated */
2005-04-21 16:53:53 +02:00
struct re_guts *g;
# define NPAREN 10 /* we need to remember () 1-9 for back refs */
sopno pbegin[NPAREN]; /* -> ( ([0] unused) */
sopno pend[NPAREN]; /* -> ) ([0] unused) */
};
/* ========= begin header generated by ./mkh ========= */
#ifdef __cplusplus
extern "C" {
#endif
/* === regcomp.c === */
static void p_ere(struct parse *p, int stop, size_t reclimit);
static void p_ere_exp(struct parse *p, size_t reclimit);
2005-04-21 16:53:53 +02:00
static void p_str(struct parse *p);
static void p_bre(struct parse *p, int end1, int end2, size_t reclimit);
static int p_simp_re(struct parse *p, int starordinary, size_t reclimit);
2005-04-21 16:53:53 +02:00
static int p_count(struct parse *p);
static void p_bracket(struct parse *p);
static void p_b_term(struct parse *p, cset *cs);
static void p_b_cclass(struct parse *p, cset *cs);
static void p_b_eclass(struct parse *p, cset *cs);
static char p_b_symbol(struct parse *p);
static char p_b_coll_elem(struct parse *p, int endc);
static int othercase(int ch);
2005-04-21 16:53:53 +02:00
static void bothcases(struct parse *p, int ch);
static void ordinary(struct parse *p, int ch);
static void nonnewline(struct parse *p);
static void repeat(struct parse *p, sopno start, int from, int to, size_t reclimit);
2005-04-21 16:53:53 +02:00
static int seterr(struct parse *p, int e);
static cset *allocset(struct parse *p);
static void freeset(struct parse *p, cset *cs);
static sopno freezeset(struct parse *p, cset *cs);
2005-04-21 16:53:53 +02:00
static int firstch(struct parse *p, cset *cs);
static int nch(struct parse *p, cset *cs);
static void mcadd(struct parse *p, cset *cs, const char *cp);
#if 0
static void mcsub(cset *cs, char *cp);
static int mcin(cset *cs, char *cp);
static char *mcfind(cset *cs, char *cp);
#endif
2005-04-21 16:53:53 +02:00
static void mcinvert(struct parse *p, cset *cs);
static void mccase(struct parse *p, cset *cs);
static int isinsets(struct re_guts *g, int c);
static int samesets(struct re_guts *g, int c1, int c2);
static void categorize(struct parse *p, struct re_guts *g);
static sopno dupl(struct parse *p, sopno start, sopno finish);
static void doemit(struct parse *p, sop op, sopno opnd);
static void doinsert(struct parse *p, sop op, sopno opnd, sopno pos);
static void dofwd(struct parse *p, sopno pos, sopno value);
static int enlarge(struct parse *p, sopno size);
2005-04-21 16:53:53 +02:00
static void stripsnug(struct parse *p, struct re_guts *g);
static void findmust(struct parse *p, struct re_guts *g);
static sopno pluscount(struct parse *p, struct re_guts *g);
#ifdef __cplusplus
}
#endif
/* ========= end header generated by ./mkh ========= */
static char nuls[10]; /* place to point scanner in event of error */
/*
* macros for use with parse structure
* BEWARE: these know that the parse structure is named `p' !!!
*/
#define PEEK() (*p->next)
#define PEEK2() (*(p->next+1))
#define MORE() (p->next < p->end)
#define MORE2() (p->next+1 < p->end)
#define SEE(c) (MORE() && PEEK() == (c))
#define SEETWO(a, b) (MORE() && MORE2() && PEEK() == (a) && PEEK2() == (b))
#define EAT(c) ((SEE(c)) ? (NEXT(), 1) : 0)
#define EATTWO(a, b) ((SEETWO(a, b)) ? (NEXT2(), 1) : 0)
#define NEXT() (p->next++)
#define NEXT2() (p->next += 2)
#define NEXTn(n) (p->next += (n))
#define GETNEXT() (*p->next++)
#define SETERROR(e) seterr(p, (e))
#define REQUIRE(co, e) (void) ((co) || SETERROR(e))
2005-04-21 16:53:53 +02:00
#define MUSTSEE(c, e) (REQUIRE(MORE() && PEEK() == (c), e))
#define MUSTEAT(c, e) (void) (REQUIRE(MORE() && GETNEXT() == (c), e))
2005-04-21 16:53:53 +02:00
#define MUSTNOTSEE(c, e) (REQUIRE(!MORE() || PEEK() != (c), e))
#define EMIT(op, sopnd) doemit(p, (sop)(op), sopnd)
2005-04-21 16:53:53 +02:00
#define INSERT(op, pos) doinsert(p, (sop)(op), HERE()-(pos)+1, pos)
#define AHEAD(pos) dofwd(p, pos, HERE()-(pos))
#define ASTERN(sop, pos) EMIT(sop, HERE()-pos)
#define HERE() (p->slen)
#define THERE() (p->slen - 1)
#define THERETHERE() (p->slen - 2)
#define DROP(n) (p->slen -= (n))
#ifndef NDEBUG
static int never = 0; /* for use in asserts; shuts lint up */
#else
#define never 0 /* some <assert.h>s have bugs too */
#endif
#define MEMLIMIT 0x8000000
#define MEMSIZE(p) \
((p)->ncsalloc / CHAR_BIT * (p)->g->csetsize + \
(p)->ncsalloc * sizeof(cset) + \
(p)->ssize * sizeof(sop))
#define RECLIMIT 256
2005-04-21 16:53:53 +02:00
/*
- regcomp - interface for parser and compilation
= extern int regcomp(regex_t *, const char *, int);
= #define REG_BASIC 0000
= #define REG_EXTENDED 0001
= #define REG_ICASE 0002
= #define REG_NOSUB 0004
= #define REG_NEWLINE 0010
= #define REG_NOSPEC 0020
= #define REG_PEND 0040
= #define REG_DUMP 0200
*/
int /* 0 success, otherwise REG_something */
regcomp(
regex_t *preg,
const char *pattern,
int cflags)
2005-04-21 16:53:53 +02:00
{
struct parse pa;
struct re_guts *g;
struct parse *p = &pa;
int i;
size_t len;
2005-04-21 16:53:53 +02:00
#ifdef REDEBUG
# define GOODFLAGS(f) (f)
#else
# define GOODFLAGS(f) ((f)&~REG_DUMP)
#endif
_DIAGASSERT(preg != NULL);
_DIAGASSERT(pattern != NULL);
2005-04-21 16:53:53 +02:00
cflags = GOODFLAGS(cflags);
if ((cflags&REG_EXTENDED) && (cflags&REG_NOSPEC))
return(REG_INVARG);
if (cflags&REG_PEND) {
if (preg->re_endp < pattern)
return(REG_INVARG);
len = preg->re_endp - pattern;
} else
len = strlen(pattern);
2005-04-21 16:53:53 +02:00
/* do the mallocs early so failure handling is easy */
g = (struct re_guts *)malloc(sizeof(struct re_guts) +
(NC-1)*sizeof(cat_t));
if (g == NULL)
return(REG_ESPACE);
p->ssize = len/(size_t)2*(size_t)3 + (size_t)1; /* ugh */
p->strip = malloc(p->ssize * sizeof(sop));
2005-04-21 16:53:53 +02:00
p->slen = 0;
if (p->strip == NULL) {
free(g);
2005-04-21 16:53:53 +02:00
return(REG_ESPACE);
}
/* set things up */
p->g = g;
p->next = pattern;
2005-04-21 16:53:53 +02:00
p->end = p->next + len;
p->error = 0;
p->ncsalloc = 0;
for (i = 0; i < NPAREN; i++) {
p->pbegin[i] = 0;
p->pend[i] = 0;
}
g->csetsize = NC;
g->sets = NULL;
g->setbits = NULL;
g->ncsets = 0;
g->cflags = cflags;
g->iflags = 0;
g->nbol = 0;
g->neol = 0;
g->must = NULL;
g->mlen = 0;
g->nsub = 0;
g->ncategories = 1; /* category 0 is "everything else" */
g->categories = &g->catspace[-(CHAR_MIN)];
(void) memset((char *)g->catspace, 0, NC*sizeof(cat_t));
g->backrefs = 0;
/* do it */
EMIT(OEND, 0);
g->firststate = THERE();
if (cflags&REG_EXTENDED)
p_ere(p, OUT, 0);
2005-04-21 16:53:53 +02:00
else if (cflags&REG_NOSPEC)
p_str(p);
else
p_bre(p, OUT, OUT, 0);
2005-04-21 16:53:53 +02:00
EMIT(OEND, 0);
g->laststate = THERE();
/* tidy up loose ends and fill things in */
categorize(p, g);
stripsnug(p, g);
findmust(p, g);
g->nplus = pluscount(p, g);
g->magic = MAGIC2;
preg->re_nsub = g->nsub;
preg->re_g = g;
preg->re_magic = MAGIC1;
#ifndef REDEBUG
/* not debugging, so can't rely on the assert() in regexec() */
if (g->iflags&BAD)
SETERROR(REG_ASSERT);
#endif
/* win or lose, we're done */
if (p->error != 0) /* lose */
regfree(preg);
return(p->error);
}
/*
- p_ere - ERE parser top level, concatenation and alternation
== static void p_ere(struct parse *p, int stop, size_t reclimit);
2005-04-21 16:53:53 +02:00
*/
static void
p_ere(
struct parse *p,
int stop, /* character this ERE should end at */
size_t reclimit)
2005-04-21 16:53:53 +02:00
{
char c;
sopno prevback = 0; /* pacify gcc */
sopno prevfwd = 0; /* pacify gcc */
sopno conc;
int first = 1; /* is this the first alternative? */
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
if (reclimit++ > RECLIMIT || p->error == REG_ESPACE) {
p->error = REG_ESPACE;
return;
}
2005-04-21 16:53:53 +02:00
for (;;) {
/* do a bunch of concatenated expressions */
conc = HERE();
while (MORE() && (c = PEEK()) != '|' && c != stop)
p_ere_exp(p, reclimit);
2005-04-21 16:53:53 +02:00
REQUIRE(HERE() != conc, REG_EMPTY); /* require nonempty */
if (!EAT('|'))
break; /* NOTE BREAK OUT */
if (first) {
INSERT(OCH_, conc); /* offset is wrong */
prevfwd = conc;
prevback = conc;
first = 0;
}
ASTERN(OOR1, prevback);
prevback = THERE();
AHEAD(prevfwd); /* fix previous offset */
prevfwd = HERE();
EMIT(OOR2, 0); /* offset is very wrong */
}
if (!first) { /* tail-end fixups */
AHEAD(prevfwd);
ASTERN(O_CH, prevback);
}
assert(!MORE() || SEE(stop));
}
/*
- p_ere_exp - parse one subERE, an atom possibly followed by a repetition op
== static void p_ere_exp(struct parse *p, size_t reclimit);
2005-04-21 16:53:53 +02:00
*/
static void
p_ere_exp(
struct parse *p,
size_t reclimit)
2005-04-21 16:53:53 +02:00
{
char c;
sopno pos;
int count;
int count2;
sopno subno;
2005-04-21 16:53:53 +02:00
int wascaret = 0;
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
assert(MORE()); /* caller should have ensured this */
c = GETNEXT();
pos = HERE();
switch (c) {
case '(':
REQUIRE(MORE(), REG_EPAREN);
p->g->nsub++;
subno = p->g->nsub;
if (subno < NPAREN)
p->pbegin[subno] = HERE();
EMIT(OLPAREN, subno);
if (!SEE(')'))
p_ere(p, ')', reclimit);
2005-04-21 16:53:53 +02:00
if (subno < NPAREN) {
p->pend[subno] = HERE();
assert(p->pend[subno] != 0);
}
EMIT(ORPAREN, subno);
MUSTEAT(')', REG_EPAREN);
break;
#ifndef POSIX_MISTAKE
case ')': /* happens only if no current unmatched ( */
/*
* You may ask, why the ifndef? Because I didn't notice
* this until slightly too late for 1003.2, and none of the
* other 1003.2 regular-expression reviewers noticed it at
* all. So an unmatched ) is legal POSIX, at least until
* we can get it fixed.
*/
SETERROR(REG_EPAREN);
break;
#endif
case '^':
EMIT(OBOL, 0);
p->g->iflags |= USEBOL;
p->g->nbol++;
wascaret = 1;
break;
case '$':
EMIT(OEOL, 0);
p->g->iflags |= USEEOL;
p->g->neol++;
break;
case '|':
SETERROR(REG_EMPTY);
break;
case '*':
case '+':
case '?':
SETERROR(REG_BADRPT);
break;
case '.':
if (p->g->cflags&REG_NEWLINE)
nonnewline(p);
else
EMIT(OANY, 0);
break;
case '[':
p_bracket(p);
break;
case '\\':
REQUIRE(MORE(), REG_EESCAPE);
c = GETNEXT();
ordinary(p, c);
break;
case '{': /* okay as ordinary except if digit follows */
REQUIRE(!MORE() || !isdigit((unsigned char)PEEK()), REG_BADRPT);
2005-04-21 16:53:53 +02:00
/* FALLTHROUGH */
default:
ordinary(p, c);
break;
}
if (!MORE())
return;
c = PEEK();
/* we call { a repetition if followed by a digit */
if (!( c == '*' || c == '+' || c == '?' ||
(c == '{' && MORE2() && isdigit((unsigned char)PEEK2())) ))
2005-04-21 16:53:53 +02:00
return; /* no repetition, we're done */
NEXT();
REQUIRE(!wascaret, REG_BADRPT);
switch (c) {
case '*': /* implemented as +? */
/* this case does not require the (y|) trick, noKLUDGE */
INSERT(OPLUS_, pos);
ASTERN(O_PLUS, pos);
INSERT(OQUEST_, pos);
ASTERN(O_QUEST, pos);
break;
case '+':
INSERT(OPLUS_, pos);
ASTERN(O_PLUS, pos);
break;
case '?':
/* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
INSERT(OCH_, pos); /* offset slightly wrong */
ASTERN(OOR1, pos); /* this one's right */
AHEAD(pos); /* fix the OCH_ */
EMIT(OOR2, 0); /* offset very wrong... */
AHEAD(THERE()); /* ...so fix it */
ASTERN(O_CH, THERETHERE());
break;
case '{':
count = p_count(p);
if (EAT(',')) {
if (isdigit((unsigned char)PEEK())) {
2005-04-21 16:53:53 +02:00
count2 = p_count(p);
REQUIRE(count <= count2, REG_BADBR);
} else /* single number with comma */
count2 = INFINITY;
} else /* just a single number */
count2 = count;
repeat(p, pos, count, count2, 0);
2005-04-21 16:53:53 +02:00
if (!EAT('}')) { /* error heuristics */
while (MORE() && PEEK() != '}')
NEXT();
REQUIRE(MORE(), REG_EBRACE);
SETERROR(REG_BADBR);
}
break;
}
if (!MORE())
return;
c = PEEK();
if (!( c == '*' || c == '+' || c == '?' ||
(c == '{' && MORE2() && isdigit((unsigned char)PEEK2())) ) )
2005-04-21 16:53:53 +02:00
return;
SETERROR(REG_BADRPT);
}
/*
- p_str - string (no metacharacters) "parser"
== static void p_str(struct parse *p);
2005-04-21 16:53:53 +02:00
*/
static void
p_str(
struct parse *p)
2005-04-21 16:53:53 +02:00
{
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
REQUIRE(MORE(), REG_EMPTY);
while (MORE())
ordinary(p, GETNEXT());
}
/*
- p_bre - BRE parser top level, anchoring and concatenation
== static void p_bre(struct parse *p, int end1, \
== int end2, size_t reclimit);
2005-04-21 16:53:53 +02:00
* Giving end1 as OUT essentially eliminates the end1/end2 check.
*
* This implementation is a bit of a kludge, in that a trailing $ is first
* taken as an ordinary character and then revised to be an anchor. The
* only undesirable side effect is that '$' gets included as a character
* category in such cases. This is fairly harmless; not worth fixing.
* The amount of lookahead needed to avoid this kludge is excessive.
*/
static void
p_bre(
struct parse *p,
int end1, /* first terminating character */
int end2, /* second terminating character */
size_t reclimit)
2005-04-21 16:53:53 +02:00
{
sopno start;
int first = 1; /* first subexpression? */
int wasdollar = 0;
_DIAGASSERT(p != NULL);
if (reclimit++ > RECLIMIT || p->error == REG_ESPACE) {
p->error = REG_ESPACE;
return;
}
start = HERE();
2005-04-21 16:53:53 +02:00
if (EAT('^')) {
EMIT(OBOL, 0);
p->g->iflags |= USEBOL;
p->g->nbol++;
}
while (MORE() && !SEETWO(end1, end2)) {
wasdollar = p_simp_re(p, first, reclimit);
2005-04-21 16:53:53 +02:00
first = 0;
}
if (wasdollar) { /* oops, that was a trailing anchor */
DROP(1);
EMIT(OEOL, 0);
p->g->iflags |= USEEOL;
p->g->neol++;
}
REQUIRE(HERE() != start, REG_EMPTY); /* require nonempty */
}
/*
- p_simp_re - parse a simple RE, an atom possibly followed by a repetition
== static int p_simp_re(struct parse *p, int starordinary, size_t reclimit);
2005-04-21 16:53:53 +02:00
*/
static int /* was the simple RE an unbackslashed $? */
p_simp_re(
struct parse *p,
int starordinary, /* is a leading * an ordinary character? */
size_t reclimit)
2005-04-21 16:53:53 +02:00
{
int c;
int count;
int count2;
sopno pos, i;
sopno subno;
2005-04-21 16:53:53 +02:00
# define BACKSL (1<<CHAR_BIT)
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
pos = HERE(); /* repetion op, if any, covers from here */
assert(MORE()); /* caller should have ensured this */
c = GETNEXT();
if (c == '\\') {
REQUIRE(MORE(), REG_EESCAPE);
c = BACKSL | (unsigned char)GETNEXT();
}
switch (c) {
case '.':
if (p->g->cflags&REG_NEWLINE)
nonnewline(p);
else
EMIT(OANY, 0);
break;
case '[':
p_bracket(p);
break;
case BACKSL|'{':
SETERROR(REG_BADRPT);
break;
case BACKSL|'(':
p->g->nsub++;
subno = p->g->nsub;
if (subno < NPAREN)
p->pbegin[subno] = HERE();
EMIT(OLPAREN, subno);
/* the MORE here is an error heuristic */
if (MORE() && !SEETWO('\\', ')'))
p_bre(p, '\\', ')', reclimit);
2005-04-21 16:53:53 +02:00
if (subno < NPAREN) {
p->pend[subno] = HERE();
assert(p->pend[subno] != 0);
}
EMIT(ORPAREN, subno);
REQUIRE(EATTWO('\\', ')'), REG_EPAREN);
break;
case BACKSL|')': /* should not get here -- must be user */
case BACKSL|'}':
SETERROR(REG_EPAREN);
break;
case BACKSL|'1':
case BACKSL|'2':
case BACKSL|'3':
case BACKSL|'4':
case BACKSL|'5':
case BACKSL|'6':
case BACKSL|'7':
case BACKSL|'8':
case BACKSL|'9':
i = (c&~BACKSL) - '0';
assert(i < NPAREN);
if (p->pend[i] != 0) {
assert(i <= p->g->nsub);
EMIT(OBACK_, i);
assert(p->pbegin[i] != 0);
assert(OP(p->strip[p->pbegin[i]]) == OLPAREN);
assert(OP(p->strip[p->pend[i]]) == ORPAREN);
(void) dupl(p, p->pbegin[i]+1, p->pend[i]);
EMIT(O_BACK, i);
} else
SETERROR(REG_ESUBREG);
p->g->backrefs = 1;
break;
case '*':
REQUIRE(starordinary, REG_BADRPT);
/* FALLTHROUGH */
default:
ordinary(p, c &~ BACKSL);
break;
}
if (EAT('*')) { /* implemented as +? */
/* this case does not require the (y|) trick, noKLUDGE */
INSERT(OPLUS_, pos);
ASTERN(O_PLUS, pos);
INSERT(OQUEST_, pos);
ASTERN(O_QUEST, pos);
} else if (EATTWO('\\', '{')) {
count = p_count(p);
if (EAT(',')) {
if (MORE() && isdigit((unsigned char)PEEK())) {
2005-04-21 16:53:53 +02:00
count2 = p_count(p);
REQUIRE(count <= count2, REG_BADBR);
} else /* single number with comma */
count2 = INFINITY;
} else /* just a single number */
count2 = count;
repeat(p, pos, count, count2, 0);
2005-04-21 16:53:53 +02:00
if (!EATTWO('\\', '}')) { /* error heuristics */
while (MORE() && !SEETWO('\\', '}'))
NEXT();
REQUIRE(MORE(), REG_EBRACE);
SETERROR(REG_BADBR);
}
} else if (c == (unsigned char)'$') /* $ (but not \$) ends it */
return(1);
return(0);
}
/*
- p_count - parse a repetition count
== static int p_count(struct parse *p);
2005-04-21 16:53:53 +02:00
*/
static int /* the value */
p_count(
struct parse *p)
2005-04-21 16:53:53 +02:00
{
int count = 0;
int ndigits = 0;
2005-04-21 16:53:53 +02:00
_DIAGASSERT(p != NULL);
while (MORE() && isdigit((unsigned char)PEEK()) && count <= DUPMAX) {
2005-04-21 16:53:53 +02:00
count = count*10 + (GETNEXT() - '0');
ndigits++;
}
REQUIRE(ndigits > 0 && count <= DUPMAX, REG_BADBR);
return(count);
}
/*
- p_bracket - parse a bracketed character list
== static void p_bracket(struct parse *p);
2005-04-21 16:53:53 +02:00
*
* Note a significant property of this code: if the allocset() did SETERROR,
* no set operations are done.
*/
static void
p_bracket(
struct parse *p)
2005-04-21 16:53:53 +02:00
{
cset *cs;
int invert = 0;
_DIAGASSERT(p != NULL);
cs = allocset(p);
if (cs == NULL)
return;
2005-04-21 16:53:53 +02:00
/* Dept of Truly Sickening Special-Case Kludges */
if (p->next + 5 < p->end && strncmp(p->next, "[:<:]]",
(size_t)6) == 0) {
2005-04-21 16:53:53 +02:00
EMIT(OBOW, 0);
NEXTn(6);
return;
}
if (p->next + 5 < p->end && strncmp(p->next, "[:>:]]",
(size_t)6) == 0) {
2005-04-21 16:53:53 +02:00
EMIT(OEOW, 0);
NEXTn(6);
return;
}
if (EAT('^'))
invert++; /* make note to invert set at end */
if (EAT(']'))
CHadd(cs, ']');
else if (EAT('-'))
CHadd(cs, '-');
while (MORE() && PEEK() != ']' && !SEETWO('-', ']'))
p_b_term(p, cs);
if (EAT('-'))
CHadd(cs, '-');
MUSTEAT(']', REG_EBRACK);
if (p->error != 0) /* don't mess things up further */
return;
if (p->g->cflags&REG_ICASE) {
ssize_t i;
int ci;
2005-04-21 16:53:53 +02:00
for (i = p->g->csetsize - 1; i >= 0; i--)
if (CHIN(cs, i) && isalpha(i)) {
ci = othercase((int)i);
2005-04-21 16:53:53 +02:00
if (ci != i)
CHadd(cs, ci);
}
if (cs->multis != NULL)
mccase(p, cs);
}
if (invert) {
ssize_t i;
2005-04-21 16:53:53 +02:00
for (i = p->g->csetsize - 1; i >= 0; i--)
if (CHIN(cs, i))
CHsub(cs, (int)i);
2005-04-21 16:53:53 +02:00
else
CHadd(cs, (int)i);
2005-04-21 16:53:53 +02:00
if (p->g->cflags&REG_NEWLINE)
CHsub(cs, '\n');
if (cs->multis != NULL)
mcinvert(p, cs);
}
assert(cs->multis == NULL); /* xxx */
if (nch(p, cs) == 1) { /* optimize singleton sets */
ordinary(p, firstch(p, cs));
freeset(p, cs);
} else
EMIT(OANYOF, freezeset(p, cs));
}
/*
- p_b_term - parse one term of a bracketed character list
== static void p_b_term(struct parse *p, cset *cs);
2005-04-21 16:53:53 +02:00
*/
static void
p_b_term(
struct parse *p,
cset *cs)
2005-04-21 16:53:53 +02:00
{
char c;
char start, finish;
int i;
_DIAGASSERT(p != NULL);
_DIAGASSERT(cs != NULL);
2005-04-21 16:53:53 +02:00
/* classify what we've got */
switch ((MORE()) ? PEEK() : '\0') {
case '[':
c = (MORE2()) ? PEEK2() : '\0';
break;
2005-04-21 16:53:53 +02:00
case '-':
SETERROR(REG_ERANGE);
return; /* NOTE RETURN */
2005-04-21 16:53:53 +02:00
default:
c = '\0';
break;
}
switch (c) {
case ':': /* character class */
NEXT2();
REQUIRE(MORE(), REG_EBRACK);
c = PEEK();
REQUIRE(c != '-' && c != ']', REG_ECTYPE);
p_b_cclass(p, cs);
REQUIRE(MORE(), REG_EBRACK);
REQUIRE(EATTWO(':', ']'), REG_ECTYPE);
break;
case '=': /* equivalence class */
NEXT2();
REQUIRE(MORE(), REG_EBRACK);
c = PEEK();
REQUIRE(c != '-' && c != ']', REG_ECOLLATE);
p_b_eclass(p, cs);
REQUIRE(MORE(), REG_EBRACK);
REQUIRE(EATTWO('=', ']'), REG_ECOLLATE);
break;
default: /* symbol, ordinary character, or range */
/* xxx revision needed for multichar stuff */
start = p_b_symbol(p);
if (SEE('-') && MORE2() && PEEK2() != ']') {
/* range */
NEXT();
if (EAT('-'))
finish = '-';
else
finish = p_b_symbol(p);
} else
finish = start;
/* xxx what about signed chars here... */
REQUIRE(start <= finish, REG_ERANGE);
for (i = start; i <= finish; i++)
CHadd(cs, i);
break;
}
}
/*
- p_b_cclass - parse a character-class name and deal with it
== static void p_b_cclass(struct parse *p, cset *cs);
2005-04-21 16:53:53 +02:00
*/
static void
p_b_cclass(
struct parse *p,
cset *cs)
2005-04-21 16:53:53 +02:00
{
const char *sp;
const struct cclass *cp;
size_t len;
const char *u;
char c;
_DIAGASSERT(p != NULL);
_DIAGASSERT(cs != NULL);
sp = p->next;
2005-04-21 16:53:53 +02:00
while (MORE() && isalpha((unsigned char)PEEK()))
2005-04-21 16:53:53 +02:00
NEXT();
len = p->next - sp;
for (cp = cclasses; cp->name != NULL; cp++)
if (strncmp(cp->name, sp, len) == 0 && cp->name[len] == '\0')
break;
if (cp->name == NULL) {
/* oops, didn't find it */
SETERROR(REG_ECTYPE);
return;
}
u = cp->chars;
while ((c = *u++) != '\0')
CHadd(cs, c);
for (u = cp->multis; *u != '\0'; u += strlen(u) + 1)
MCadd(p, cs, u);
}
/*
- p_b_eclass - parse an equivalence-class name and deal with it
== static void p_b_eclass(struct parse *p, cset *cs);
2005-04-21 16:53:53 +02:00
*
* This implementation is incomplete. xxx
*/
static void
p_b_eclass(
struct parse *p,
cset *cs)
2005-04-21 16:53:53 +02:00
{
char c;
_DIAGASSERT(p != NULL);
_DIAGASSERT(cs != NULL);
2005-04-21 16:53:53 +02:00
c = p_b_coll_elem(p, '=');
CHadd(cs, c);
}
/*
- p_b_symbol - parse a character or [..]ed multicharacter collating symbol
== static char p_b_symbol(struct parse *p);
2005-04-21 16:53:53 +02:00
*/
static char /* value of symbol */
p_b_symbol(
struct parse *p)
2005-04-21 16:53:53 +02:00
{
char value;
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
REQUIRE(MORE(), REG_EBRACK);
if (!EATTWO('[', '.'))
return(GETNEXT());
/* collating symbol */
value = p_b_coll_elem(p, '.');
REQUIRE(EATTWO('.', ']'), REG_ECOLLATE);
return(value);
}
/*
- p_b_coll_elem - parse a collating-element name and look it up
== static char p_b_coll_elem(struct parse *p, int endc);
2005-04-21 16:53:53 +02:00
*/
static char /* value of collating element */
p_b_coll_elem(
struct parse *p,
int endc) /* name ended by endc,']' */
2005-04-21 16:53:53 +02:00
{
const char *sp;
const struct cname *cp;
size_t len;
_DIAGASSERT(p != NULL);
sp = p->next;
2005-04-21 16:53:53 +02:00
while (MORE() && !SEETWO(endc, ']'))
NEXT();
if (!MORE()) {
SETERROR(REG_EBRACK);
return(0);
}
len = p->next - sp;
for (cp = cnames; cp->name != NULL; cp++)
if (strncmp(cp->name, sp, len) == 0 && cp->name[len] == '\0')
return(cp->code); /* known name */
if (len == 1)
return(*sp); /* single character */
SETERROR(REG_ECOLLATE); /* neither */
return(0);
}
/*
- othercase - return the case counterpart of an alphabetic
== static int othercase(int ch);
2005-04-21 16:53:53 +02:00
*/
static int /* if no counterpart, return ch */
othercase(
int ch)
2005-04-21 16:53:53 +02:00
{
assert(isalpha(ch));
if (isupper(ch))
return(tolower(ch));
else if (islower(ch))
return(toupper(ch));
else /* peculiar, but could happen */
return(ch);
}
/*
- bothcases - emit a dualcase version of a two-case character
== static void bothcases(struct parse *p, int ch);
2005-04-21 16:53:53 +02:00
*
* Boy, is this implementation ever a kludge...
*/
static void
bothcases(
struct parse *p,
int ch)
2005-04-21 16:53:53 +02:00
{
const char *oldnext;
const char *oldend;
2005-04-21 16:53:53 +02:00
char bracket[3];
_DIAGASSERT(p != NULL);
oldnext = p->next;
oldend = p->end;
2005-04-21 16:53:53 +02:00
assert(othercase(ch) != ch); /* p_bracket() would recurse */
p->next = bracket;
p->end = bracket+2;
bracket[0] = ch;
bracket[1] = ']';
bracket[2] = '\0';
p_bracket(p);
assert(p->next == bracket+2);
p->next = oldnext;
p->end = oldend;
}
/*
- ordinary - emit an ordinary character
== static void ordinary(struct parse *p, int ch);
2005-04-21 16:53:53 +02:00
*/
static void
ordinary(
struct parse *p,
int ch)
2005-04-21 16:53:53 +02:00
{
cat_t *cap;
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
cap = p->g->categories;
if ((p->g->cflags&REG_ICASE) && isalpha((unsigned char) ch)
&& othercase((unsigned char) ch) != (unsigned char) ch)
bothcases(p, (unsigned char) ch);
2005-04-21 16:53:53 +02:00
else {
EMIT(OCHAR, (sopno)(unsigned char)ch);
if (cap[ch] == 0) {
_DIAGASSERT(__type_fit(unsigned char,
p->g->ncategories + 1));
cap[ch] = (unsigned char)p->g->ncategories++;
}
2005-04-21 16:53:53 +02:00
}
}
/*
- nonnewline - emit REG_NEWLINE version of OANY
== static void nonnewline(struct parse *p);
2005-04-21 16:53:53 +02:00
*
* Boy, is this implementation ever a kludge...
*/
static void
nonnewline(
struct parse *p)
2005-04-21 16:53:53 +02:00
{
const char *oldnext;
const char *oldend;
2005-04-21 16:53:53 +02:00
char bracket[4];
_DIAGASSERT(p != NULL);
oldnext = p->next;
oldend = p->end;
2005-04-21 16:53:53 +02:00
p->next = bracket;
p->end = bracket+3;
bracket[0] = '^';
bracket[1] = '\n';
bracket[2] = ']';
bracket[3] = '\0';
p_bracket(p);
assert(p->next == bracket+3);
p->next = oldnext;
p->end = oldend;
}
/*
- repeat - generate code for a bounded repetition, recursively if needed
== static void repeat(struct parse *p, sopno start, int from, int to,
== size_t reclimit);
2005-04-21 16:53:53 +02:00
*/
static void
repeat(
struct parse *p,
sopno start, /* operand from here to end of strip */
int from, /* repeated from this number */
int to, /* to this number of times (maybe INFINITY) */
size_t reclimit)
2005-04-21 16:53:53 +02:00
{
sopno finish;
2005-04-21 16:53:53 +02:00
# define N 2
# define INF 3
# define REP(f, t) ((f)*8 + (t))
# define MAP(n) (((n) <= 1) ? (n) : ((n) == INFINITY) ? INF : N)
sopno copy;
_DIAGASSERT(p != NULL);
if (reclimit++ > RECLIMIT)
p->error = REG_ESPACE;
if (p->error)
2005-04-21 16:53:53 +02:00
return;
finish = HERE();
2005-04-21 16:53:53 +02:00
assert(from <= to);
switch (REP(MAP(from), MAP(to))) {
case REP(0, 0): /* must be user doing this */
DROP(finish-start); /* drop the operand */
break;
case REP(0, 1): /* as x{1,1}? */
case REP(0, N): /* as x{1,n}? */
case REP(0, INF): /* as x{1,}? */
/* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
INSERT(OCH_, start); /* offset is wrong... */
repeat(p, start+1, 1, to, reclimit);
2005-04-21 16:53:53 +02:00
ASTERN(OOR1, start);
AHEAD(start); /* ... fix it */
EMIT(OOR2, 0);
AHEAD(THERE());
ASTERN(O_CH, THERETHERE());
break;
case REP(1, 1): /* trivial case */
/* done */
break;
case REP(1, N): /* as x?x{1,n-1} */
/* KLUDGE: emit y? as (y|) until subtle bug gets fixed */
INSERT(OCH_, start);
ASTERN(OOR1, start);
AHEAD(start);
EMIT(OOR2, 0); /* offset very wrong... */
AHEAD(THERE()); /* ...so fix it */
ASTERN(O_CH, THERETHERE());
copy = dupl(p, start+1, finish+1);
assert(copy == finish+4);
repeat(p, copy, 1, to-1, reclimit);
2005-04-21 16:53:53 +02:00
break;
case REP(1, INF): /* as x+ */
INSERT(OPLUS_, start);
ASTERN(O_PLUS, start);
break;
case REP(N, N): /* as xx{m-1,n-1} */
copy = dupl(p, start, finish);
repeat(p, copy, from-1, to-1, reclimit);
2005-04-21 16:53:53 +02:00
break;
case REP(N, INF): /* as xx{n-1,INF} */
copy = dupl(p, start, finish);
repeat(p, copy, from-1, to, reclimit);
2005-04-21 16:53:53 +02:00
break;
default: /* "can't happen" */
SETERROR(REG_ASSERT); /* just in case */
break;
}
}
/*
- seterr - set an error condition
== static int seterr(struct parse *p, int e);
2005-04-21 16:53:53 +02:00
*/
static int /* useless but makes type checking happy */
seterr(
struct parse *p,
int e)
2005-04-21 16:53:53 +02:00
{
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
if (p->error == 0) /* keep earliest error condition */
p->error = e;
p->next = nuls; /* try to bring things to a halt */
p->end = nuls;
return(0); /* make the return value well-defined */
}
/*
- allocset - allocate a set of characters for []
== static cset *allocset(struct parse *p);
2005-04-21 16:53:53 +02:00
*/
static cset *
allocset(
struct parse *p)
2005-04-21 16:53:53 +02:00
{
size_t no;
size_t nc;
size_t nbytes;
cset *cs;
size_t css;
size_t i;
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
no = p->g->ncsets++;
css = (size_t)p->g->csetsize;
2005-04-21 16:53:53 +02:00
if (no >= p->ncsalloc) { /* need another column of space */
p->ncsalloc += CHAR_BIT;
nc = p->ncsalloc;
assert(nc % CHAR_BIT == 0);
nbytes = nc / CHAR_BIT * css;
if (MEMSIZE(p) > MEMLIMIT)
goto oomem;
2005-04-21 16:53:53 +02:00
if (p->g->sets == NULL)
p->g->sets = malloc(nc * sizeof(cset));
2005-04-21 16:53:53 +02:00
else
p->g->sets = realloc(p->g->sets, nc * sizeof(cset));
2005-04-21 16:53:53 +02:00
if (p->g->setbits == NULL)
p->g->setbits = malloc(nbytes);
2005-04-21 16:53:53 +02:00
else {
p->g->setbits = realloc(p->g->setbits, nbytes);
2005-04-21 16:53:53 +02:00
/* xxx this isn't right if setbits is now NULL */
for (i = 0; i < no; i++)
p->g->sets[i].ptr = p->g->setbits + css*(i/CHAR_BIT);
}
if (p->g->sets != NULL && p->g->setbits != NULL)
(void) memset((char *)p->g->setbits + (nbytes - css),
0, css);
else {
oomem:
2005-04-21 16:53:53 +02:00
no = 0;
SETERROR(REG_ESPACE);
/* caller's responsibility not to do set ops */
return NULL;
2005-04-21 16:53:53 +02:00
}
}
cs = &p->g->sets[no];
cs->ptr = p->g->setbits + css*((no)/CHAR_BIT);
cs->mask = 1 << (unsigned int)((no) % CHAR_BIT);
2005-04-21 16:53:53 +02:00
cs->hash = 0;
cs->smultis = 0;
cs->multis = NULL;
return(cs);
}
/*
- freeset - free a now-unused set
== static void freeset(struct parse *p, cset *cs);
2005-04-21 16:53:53 +02:00
*/
static void
freeset(
struct parse *p,
cset *cs)
2005-04-21 16:53:53 +02:00
{
size_t i;
cset *top;
size_t css;
_DIAGASSERT(p != NULL);
_DIAGASSERT(cs != NULL);
top = &p->g->sets[p->g->ncsets];
css = (size_t)p->g->csetsize;
2005-04-21 16:53:53 +02:00
for (i = 0; i < css; i++)
CHsub(cs, (int)i);
2005-04-21 16:53:53 +02:00
if (cs == top-1) /* recover only the easy case */
p->g->ncsets--;
}
/*
- freezeset - final processing on a set of characters
== static int freezeset(struct parse *p, cset *cs);
2005-04-21 16:53:53 +02:00
*
* The main task here is merging identical sets. This is usually a waste
* of time (although the hash code minimizes the overhead), but can win
* big if REG_ICASE is being used. REG_ICASE, by the way, is why the hash
* is done using addition rather than xor -- all ASCII [aA] sets xor to
* the same value!
*/
static sopno /* set number */
freezeset(
struct parse *p,
cset *cs)
2005-04-21 16:53:53 +02:00
{
uch h;
size_t i;
cset *top;
cset *cs2;
size_t css;
_DIAGASSERT(p != NULL);
_DIAGASSERT(cs != NULL);
h = cs->hash;
top = &p->g->sets[p->g->ncsets];
css = (size_t)p->g->csetsize;
2005-04-21 16:53:53 +02:00
/* look for an earlier one which is the same */
for (cs2 = &p->g->sets[0]; cs2 < top; cs2++)
if (cs2->hash == h && cs2 != cs) {
/* maybe */
for (i = 0; i < css; i++)
if (!!CHIN(cs2, i) != !!CHIN(cs, i))
break; /* no */
if (i == css)
break; /* yes */
}
if (cs2 < top) { /* found one */
freeset(p, cs);
cs = cs2;
}
return (sopno)(cs - p->g->sets);
2005-04-21 16:53:53 +02:00
}
/*
- firstch - return first character in a set (which must have at least one)
== static int firstch(struct parse *p, cset *cs);
2005-04-21 16:53:53 +02:00
*/
static int /* character; there is no "none" value */
firstch(
struct parse *p,
cset *cs)
2005-04-21 16:53:53 +02:00
{
size_t i;
size_t css;
_DIAGASSERT(p != NULL);
_DIAGASSERT(cs != NULL);
css = (size_t)p->g->csetsize;
2005-04-21 16:53:53 +02:00
for (i = 0; i < css; i++)
if (CHIN(cs, i))
return((char)i);
assert(never);
return(0); /* arbitrary */
}
/*
- nch - number of characters in a set
== static int nch(struct parse *p, cset *cs);
2005-04-21 16:53:53 +02:00
*/
static int
nch(
struct parse *p,
cset *cs)
2005-04-21 16:53:53 +02:00
{
size_t i;
size_t css;
int n = 0;
_DIAGASSERT(p != NULL);
_DIAGASSERT(cs != NULL);
css = (size_t)p->g->csetsize;
2005-04-21 16:53:53 +02:00
for (i = 0; i < css; i++)
if (CHIN(cs, i))
n++;
return(n);
}
/*
- mcadd - add a collating element to a cset
== static void mcadd(struct parse *p, cset *cs, \
== char *cp);
2005-04-21 16:53:53 +02:00
*/
static void
mcadd(
struct parse *p,
cset *cs,
const char *cp)
2005-04-21 16:53:53 +02:00
{
size_t oldend;
_DIAGASSERT(p != NULL);
_DIAGASSERT(cs != NULL);
_DIAGASSERT(cp != NULL);
oldend = cs->smultis;
2005-04-21 16:53:53 +02:00
cs->smultis += strlen(cp) + 1;
if (cs->multis == NULL)
cs->multis = malloc(cs->smultis);
else
cs->multis = realloc(cs->multis, cs->smultis);
if (cs->multis == NULL) {
SETERROR(REG_ESPACE);
return;
}
(void) strcpy(cs->multis + oldend - 1, cp);
cs->multis[cs->smultis - 1] = '\0';
}
#if 0
/*
- mcsub - subtract a collating element from a cset
== static void mcsub(cset *cs, char *cp);
*/
static void
mcsub(
cset *cs,
char *cp)
{
char *fp;
size_t len;
_DIAGASSERT(cs != NULL);
_DIAGASSERT(cp != NULL);
fp = mcfind(cs, cp);
len = strlen(fp);
assert(fp != NULL);
(void) memmove(fp, fp + len + 1,
cs->smultis - (fp + len + 1 - cs->multis));
cs->smultis -= len;
if (cs->smultis == 0) {
free(cs->multis);
cs->multis = NULL;
return;
}
cs->multis = realloc(cs->multis, cs->smultis);
assert(cs->multis != NULL);
}
/*
- mcin - is a collating element in a cset?
== static int mcin(cset *cs, char *cp);
*/
static int
mcin(
cset *cs,
char *cp)
{
_DIAGASSERT(cs != NULL);
_DIAGASSERT(cp != NULL);
return(mcfind(cs, cp) != NULL);
}
/*
- mcfind - find a collating element in a cset
== static char *mcfind(cset *cs, char *cp);
*/
static char *
mcfind(
cset *cs,
char *cp)
{
char *p;
_DIAGASSERT(cs != NULL);
_DIAGASSERT(cp != NULL);
if (cs->multis == NULL)
return(NULL);
for (p = cs->multis; *p != '\0'; p += strlen(p) + 1)
if (strcmp(cp, p) == 0)
return(p);
return(NULL);
}
#endif
2005-04-21 16:53:53 +02:00
/*
- mcinvert - invert the list of collating elements in a cset
== static void mcinvert(struct parse *p, cset *cs);
2005-04-21 16:53:53 +02:00
*
* This would have to know the set of possibilities. Implementation
* is deferred.
*/
/* ARGSUSED */
2005-04-21 16:53:53 +02:00
static void
mcinvert(
struct parse *p,
cset *cs)
2005-04-21 16:53:53 +02:00
{
_DIAGASSERT(p != NULL);
_DIAGASSERT(cs != NULL);
2005-04-21 16:53:53 +02:00
assert(cs->multis == NULL); /* xxx */
}
/*
- mccase - add case counterparts of the list of collating elements in a cset
== static void mccase(struct parse *p, cset *cs);
2005-04-21 16:53:53 +02:00
*
* This would have to know the set of possibilities. Implementation
* is deferred.
*/
/* ARGSUSED */
2005-04-21 16:53:53 +02:00
static void
mccase(
struct parse *p,
cset *cs)
2005-04-21 16:53:53 +02:00
{
_DIAGASSERT(p != NULL);
_DIAGASSERT(cs != NULL);
2005-04-21 16:53:53 +02:00
assert(cs->multis == NULL); /* xxx */
}
/*
- isinsets - is this character in any sets?
== static int isinsets(struct re_guts *g, int c);
2005-04-21 16:53:53 +02:00
*/
static int /* predicate */
isinsets(
struct re_guts *g,
int c)
2005-04-21 16:53:53 +02:00
{
uch *col;
size_t i;
size_t ncols;
unsigned uc = (unsigned char)c;
_DIAGASSERT(g != NULL);
if (g->setbits == NULL)
return 0;
ncols = (g->ncsets+(CHAR_BIT-1)) / CHAR_BIT;
2005-04-21 16:53:53 +02:00
for (i = 0, col = g->setbits; i < ncols; i++, col += g->csetsize)
if (col[uc] != 0)
return(1);
return(0);
}
/*
- samesets - are these two characters in exactly the same sets?
== static int samesets(struct re_guts *g, int c1, int c2);
2005-04-21 16:53:53 +02:00
*/
static int /* predicate */
samesets(
struct re_guts *g,
int c1,
int c2)
2005-04-21 16:53:53 +02:00
{
uch *col;
size_t i;
size_t ncols;
unsigned uc1 = (unsigned char)c1;
unsigned uc2 = (unsigned char)c2;
_DIAGASSERT(g != NULL);
ncols = (g->ncsets+(CHAR_BIT-1)) / CHAR_BIT;
2005-04-21 16:53:53 +02:00
for (i = 0, col = g->setbits; i < ncols; i++, col += g->csetsize)
if (col[uc1] != col[uc2])
return(0);
return(1);
}
/*
- categorize - sort out character categories
== static void categorize(struct parse *p, struct re_guts *g);
2005-04-21 16:53:53 +02:00
*/
static void
categorize(
struct parse *p,
struct re_guts *g)
2005-04-21 16:53:53 +02:00
{
cat_t *cats;
int c;
int c2;
cat_t cat;
_DIAGASSERT(p != NULL);
_DIAGASSERT(g != NULL);
cats = g->categories;
2005-04-21 16:53:53 +02:00
/* avoid making error situations worse */
if (p->error != 0)
return;
for (c = CHAR_MIN; c <= CHAR_MAX; c++)
if (cats[c] == 0 && isinsets(g, c)) {
_DIAGASSERT(__type_fit(unsigned char,
g->ncategories + 1));
2005-04-21 16:53:53 +02:00
cat = g->ncategories++;
cats[c] = cat;
for (c2 = c+1; c2 <= CHAR_MAX; c2++)
if (cats[c2] == 0 && samesets(g, c, c2))
cats[c2] = cat;
}
}
/*
- dupl - emit a duplicate of a bunch of sops
== static sopno dupl(struct parse *p, sopno start, sopno finish);
2005-04-21 16:53:53 +02:00
*/
static sopno /* start of duplicate */
dupl(
struct parse *p,
sopno start, /* from here */
sopno finish) /* to this less one */
2005-04-21 16:53:53 +02:00
{
sopno ret;
sopno len = finish - start;
_DIAGASSERT(p != NULL);
ret = HERE();
2005-04-21 16:53:53 +02:00
assert(finish >= start);
if (len == 0)
return(ret);
if (!enlarge(p, p->ssize + len))/* this many unexpected additions */
return ret;
(void)memcpy(p->strip + p->slen, p->strip + start,
(size_t)len * sizeof(sop));
2005-04-21 16:53:53 +02:00
p->slen += len;
return(ret);
}
/*
- doemit - emit a strip operator
== static void doemit(struct parse *p, sop op, size_t opnd);
2005-04-21 16:53:53 +02:00
*
* It might seem better to implement this as a macro with a function as
* hard-case backup, but it's just too big and messy unless there are
* some changes to the data structures. Maybe later.
*/
static void
doemit(
struct parse *p,
sop op,
sopno opnd)
2005-04-21 16:53:53 +02:00
{
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
/* avoid making error situations worse */
if (p->error != 0)
return;
/* deal with oversize operands ("can't happen", more or less) */
assert(opnd < 1<<OPSHIFT);
/* deal with undersized strip */
if (p->slen >= p->ssize)
if (!enlarge(p, (p->ssize+1) / 2 * 3)) /* +50% */
return;
2005-04-21 16:53:53 +02:00
/* finally, it's all reduced to the easy case */
p->strip[p->slen++] = (sop)SOP(op, opnd);
2005-04-21 16:53:53 +02:00
}
/*
- doinsert - insert a sop into the strip
== static void doinsert(struct parse *p, sop op, size_t opnd, sopno pos);
2005-04-21 16:53:53 +02:00
*/
static void
doinsert(
struct parse *p,
sop op,
sopno opnd,
sopno pos)
2005-04-21 16:53:53 +02:00
{
sopno sn;
sop s;
int i;
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
/* avoid making error situations worse */
if (p->error != 0)
return;
sn = HERE();
EMIT(op, opnd); /* do checks, ensure space */
assert(HERE() == sn+1);
s = p->strip[sn];
/* adjust paren pointers */
assert(pos > 0);
for (i = 1; i < NPAREN; i++) {
if (p->pbegin[i] >= pos) {
p->pbegin[i]++;
}
if (p->pend[i] >= pos) {
p->pend[i]++;
}
}
memmove(&p->strip[pos+1], &p->strip[pos], (HERE()-pos-1)*sizeof(sop));
2005-04-21 16:53:53 +02:00
p->strip[pos] = s;
}
/*
- dofwd - complete a forward reference
== static void dofwd(struct parse *p, sopno pos, sop value);
2005-04-21 16:53:53 +02:00
*/
static void
dofwd(
struct parse *p,
sopno pos,
sopno value)
2005-04-21 16:53:53 +02:00
{
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
/* avoid making error situations worse */
if (p->error != 0)
return;
assert(value < 1<<OPSHIFT);
p->strip[pos] = (sop)(OP(p->strip[pos]) | value);
2005-04-21 16:53:53 +02:00
}
/*
- enlarge - enlarge the strip
== static void enlarge(struct parse *p, sopno size);
2005-04-21 16:53:53 +02:00
*/
static int
enlarge(
struct parse *p,
sopno size)
2005-04-21 16:53:53 +02:00
{
sop *sp;
sopno osize;
_DIAGASSERT(p != NULL);
2005-04-21 16:53:53 +02:00
if (p->ssize >= size)
return 1;
2005-04-21 16:53:53 +02:00
osize = p->ssize;
p->ssize = size;
if (MEMSIZE(p) > MEMLIMIT)
goto oomem;
sp = realloc(p->strip, p->ssize * sizeof(sop));
2005-04-21 16:53:53 +02:00
if (sp == NULL) {
oomem:
p->ssize = osize;
2005-04-21 16:53:53 +02:00
SETERROR(REG_ESPACE);
return 0;
2005-04-21 16:53:53 +02:00
}
p->strip = sp;
return 1;
2005-04-21 16:53:53 +02:00
}
/*
- stripsnug - compact the strip
== static void stripsnug(struct parse *p, struct re_guts *g);
2005-04-21 16:53:53 +02:00
*/
static void
stripsnug(
struct parse *p,
struct re_guts *g)
2005-04-21 16:53:53 +02:00
{
_DIAGASSERT(p != NULL);
_DIAGASSERT(g != NULL);
2005-04-21 16:53:53 +02:00
g->nstates = p->slen;
g->strip = realloc(p->strip, p->slen * sizeof(sop));
2005-04-21 16:53:53 +02:00
if (g->strip == NULL) {
SETERROR(REG_ESPACE);
g->strip = p->strip;
}
}
/*
- findmust - fill in must and mlen with longest mandatory literal string
== static void findmust(struct parse *p, struct re_guts *g);
2005-04-21 16:53:53 +02:00
*
* This algorithm could do fancy things like analyzing the operands of |
* for common subsequences. Someday. This code is simple and finds most
* of the interesting cases.
*
* Note that must and mlen got initialized during setup.
*/
static void
findmust(
struct parse *p,
struct re_guts *g)
2005-04-21 16:53:53 +02:00
{
sop *scan;
sop *start = NULL;
sop *newstart = NULL;
sopno newlen;
sop s;
char *cp;
sopno i;
_DIAGASSERT(p != NULL);
_DIAGASSERT(g != NULL);
2005-04-21 16:53:53 +02:00
/* avoid making error situations worse */
if (p->error != 0)
return;
/* find the longest OCHAR sequence in strip */
newlen = 0;
scan = g->strip + 1;
do {
s = *scan++;
switch (OP(s)) {
case OCHAR: /* sequence member */
if (newlen == 0) /* new sequence */
newstart = scan - 1;
newlen++;
break;
case OPLUS_: /* things that don't break one */
case OLPAREN:
case ORPAREN:
break;
case OQUEST_: /* things that must be skipped */
case OCH_:
scan--;
do {
scan += OPND(s);
s = *scan;
/* assert() interferes w debug printouts */
if (OP(s) != O_QUEST && OP(s) != O_CH &&
OP(s) != OOR2) {
g->iflags |= BAD;
return;
}
} while (OP(s) != O_QUEST && OP(s) != O_CH);
/* FALLTHROUGH */
2005-04-21 16:53:53 +02:00
default: /* things that break a sequence */
if (newlen > g->mlen) { /* ends one */
start = newstart;
g->mlen = newlen;
}
newlen = 0;
break;
}
} while (OP(s) != OEND);
if (start == NULL)
g->mlen = 0;
if (g->mlen == 0) /* there isn't one */
2005-04-21 16:53:53 +02:00
return;
/* turn it into a character string */
g->must = malloc((size_t)g->mlen + 1);
if (g->must == NULL) { /* argh; just forget it */
g->mlen = 0;
return;
}
cp = g->must;
scan = start;
for (i = g->mlen; i > 0; i--) {
while (OP(s = *scan++) != OCHAR)
continue;
assert(cp < g->must + g->mlen);
*cp++ = (char)OPND(s);
}
assert(cp == g->must + g->mlen);
*cp++ = '\0'; /* just on general principles */
}
/*
- pluscount - count + nesting
== static sopno pluscount(struct parse *p, struct re_guts *g);
2005-04-21 16:53:53 +02:00
*/
static sopno /* nesting depth */
pluscount(
struct parse *p,
struct re_guts *g)
2005-04-21 16:53:53 +02:00
{
sop *scan;
sop s;
sopno plusnest = 0;
sopno maxnest = 0;
_DIAGASSERT(p != NULL);
_DIAGASSERT(g != NULL);
2005-04-21 16:53:53 +02:00
if (p->error != 0)
return(0); /* there may not be an OEND */
scan = g->strip + 1;
do {
s = *scan++;
switch (OP(s)) {
case OPLUS_:
plusnest++;
break;
case O_PLUS:
if (plusnest > maxnest)
maxnest = plusnest;
plusnest--;
break;
}
} while (OP(s) != OEND);
if (plusnest != 0)
g->iflags |= BAD;
return(maxnest);
}