/*
- p_str - string (no metacharacters) "parser"
*/
static void
p_str(struct parse *p)
{
REQUIRE(MORE(), REG_EMPTY);
while (MORE())
ordinary(p, GETNEXT());
}
/*
- p_str - string (no metacharacters) "parser"
== static void p_str(struct parse *p);
*/
static void
p_str(
struct parse *p)
{
_DIAGASSERT(p != NULL);
REQUIRE(MORE(), REG_EMPTY);
while (MORE())
ordinary(p, GETNEXT());
}
/*
- p_bracket - parse a bracketed character list
*
* Note a significant property of this code: if the allocset() did SETERROR,
* no set operations are done.
*/
static void
p_bracket(struct parse *p)
{
cset *cs;
int invert = 0;
/* Dept of Truly Sickening Special-Case Kludges */
if (p->next + 5 < p->end && strncmp(p->next, "[:<:]]", 6) == 0) {
EMIT(OBOW, 0);
NEXTn(6);
return;
}
if (p->next + 5 < p->end && strncmp(p->next, "[:>:]]", 6) == 0) {
EMIT(OEOW, 0);
NEXTn(6);
return;
}
if ((cs = allocset(p)) == NULL) {
/* allocset did set error status in p */
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 */
freeset(p, cs);
return;
}
if (p->g->cflags®_ICASE) {
int i;
int ci;
for (i = p->g->csetsize - 1; i >= 0; i--)
if (CHIN(cs, i) && isalpha(i)) {
ci = othercase(i);
if (ci != i)
CHadd(cs, ci);
}
if (cs->multis != NULL)
mccase(p, cs);
}
if (invert) {
int i;
for (i = p->g->csetsize - 1; i >= 0; i--)
if (CHIN(cs, i))
CHsub(cs, i);
else
CHadd(cs, i);
if (p->g->cflags®_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_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);
}