/*
- p_ere - ERE parser top level, concatenation and alternation
== static void p_ere(struct parse *p, int stop, size_t reclimit);
*/
static void
p_ere(
struct parse *p,
int stop, /* character this ERE should end at */
size_t reclimit)
{
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);
if (reclimit++ > RECLIMIT || p->error == REG_ESPACE) {
p->error = REG_ESPACE;
return;
}
for (;;) {
/* do a bunch of concatenated expressions */
conc = HERE();
while (MORE() && (c = PEEK()) != '|' && c != stop)
p_ere_exp(p, reclimit);
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 - ERE parser top level, concatenation and alternation
*/
static void
p_ere(struct parse *p, int stop) /* character this ERE should end at */
{
char c;
sopno prevback;
sopno prevfwd;
sopno conc;
int first = 1; /* is this the first alternative? */
for (;;) {
/* do a bunch of concatenated expressions */
conc = HERE();
while (MORE() && (c = PEEK()) != '|' && c != stop)
p_ere_exp(p);
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));
}
/*
- repeat - generate code for a bounded repetition, recursively if needed
*/
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) */
{
sopno finish = HERE();
# 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;
if (p->error != 0) /* head off possible runaway recursion */
return;
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);
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);
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);
break;
case REP(N, INF): /* as xx{n-1,INF} */
copy = dupl(p, start, finish);
repeat(p, copy, from-1, to);
break;
default: /* "can't happen" */
SETERROR(REG_ASSERT); /* just in case */
break;
}
}
/*
- p_simp_re - parse a simple RE, an atom possibly followed by a repetition
*/
static int /* was the simple RE an unbackslashed $? */
p_simp_re(struct parse *p,
int starordinary) /* is a leading * an ordinary character? */
{
int c;
int count;
int count2;
sopno pos;
int i;
sopno subno;
# define BACKSL (1<<CHAR_BIT)
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 | GETNEXT();
}
switch (c) {
case '.':
if (p->g->cflags®_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, '\\', ')');
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, (char)c);
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((uch)PEEK())) {
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);
if (!EATTWO('\\', '}')) { /* error heuristics */
while (MORE() && !SEETWO('\\', '}'))
NEXT();
REQUIRE(MORE(), REG_EBRACE);
SETERROR(REG_BADBR);
}
} else if (c == '$') /* $ (but not \$) ends it */
return(1);
return(0);
}
/*
- p_ere_exp - parse one subERE, an atom possibly followed by a repetition op
*/
static void
p_ere_exp(struct parse *p)
{
char c;
sopno pos;
int count;
int count2;
sopno subno;
int wascaret = 0;
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, ')');
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®_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((uch)PEEK()), REG_BADRPT);
/* 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((uch)PEEK2())) ))
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((uch)PEEK())) {
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);
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((uch)PEEK2())) ) )
return;
SETERROR(REG_BADRPT);
}