void
reinit(Regex *r, char *regx) {
refree(r);
if(regx[0] != '\0') {
r->regex = estrdup(regx);
r->regc = regcomp(regx);
}
}
/*
* Implementation
*/
static re_cod *
LispRecomp(LispBuiltin *builtin, char *pattern, int cflags)
{
int code;
re_cod *regex = LispMalloc(sizeof(re_cod));
if ((code = recomp(regex, pattern, cflags)) != 0) {
char buffer[256];
reerror(code, regex, buffer, sizeof(buffer));
refree(regex);
LispFree(regex);
LispDestroy("%s: recomp(\"%s\"): %s", STRFUN(builtin), pattern, buffer);
}
return (regex);
}
Regexp *
re_update(NODE *t)
{
NODE *t1;
if ((t->re_flags & CASE) == IGNORECASE) {
if ((t->re_flags & CONST) != 0) {
assert(t->type == Node_regex);
return t->re_reg;
}
t1 = force_string(tree_eval(t->re_exp));
if (t->re_text != NULL) {
if (cmp_nodes(t->re_text, t1) == 0) {
free_temp(t1);
return t->re_reg;
}
unref(t->re_text);
}
t->re_text = dupnode(t1);
free_temp(t1);
}
if (t->re_reg != NULL)
refree(t->re_reg);
if (t->re_cnt > 0)
t->re_cnt++;
if (t->re_cnt > 10)
t->re_cnt = 0;
if (t->re_text == NULL || (t->re_flags & CASE) != IGNORECASE) {
t1 = force_string(tree_eval(t->re_exp));
unref(t->re_text);
t->re_text = dupnode(t1);
free_temp(t1);
}
t->re_reg = make_regexp(t->re_text->stptr, t->re_text->stlen,
IGNORECASE, t->re_cnt);
t->re_flags &= ~CASE;
t->re_flags |= IGNORECASE;
return t->re_reg;
}
int
main(int argc, char *argv[])
{
re_cod cod;
re_mat mat[10];
int line, ecode, i, len, group, failed;
long eo, so;
char buf[8192];
char str[8192];
FILE *fp = fopen("tests.txt", "r");
if (fp == NULL) {
fprintf(stderr, "failed to open tests.txt\n");
exit(1);
}
ecode = line = group = failed = 0;
cod.cod = NULL;
while (fgets(buf, sizeof(buf), fp)) {
++line;
if (buf[0] == '#' || buf[0] == '\n')
continue;
else if (buf[0] == '/') {
char *ptr = strrchr(buf, '/');
if (ptr == buf) {
fprintf(stderr, "syntax error at line %d\n", line);
break;
}
else {
int flags = 0;
refree(&cod);
for (*ptr++ = '\0'; *ptr; ptr++) {
if (*ptr == 'i')
flags |= RE_ICASE;
else if (*ptr == 'n')
flags |= RE_NEWLINE;
}
ecode = recomp(&cod, buf + 1, flags);
failed = ecode;
}
}
else if (buf[0] == '>') {
if (cod.cod == NULL) {
fprintf(stderr, "no previous pattern at line %d\n", line);
break;
}
len = strlen(buf) - 1;
buf[len] = '\0';
strcpy(str, buf + 1);
for (i = 0, --len; i < len - 1; i++) {
if (str[i] == '\\') {
memmove(str + i, str + i + 1, len);
--len;
switch (str[i]) {
case 'a':
str[i] = '\a';
break;
case 'b':
str[i] = '\b';
break;
case 'f':
str[i] = '\f';
break;
case 'n':
str[i] = '\n';
break;
case 'r':
str[i] = '\r';
break;
case 't':
str[i] = '\t';
break;
case 'v':
str[i] = '\v';
break;
default:
break;
}
}
}
group = 0;
ecode = reexec(&cod, str, 10, &mat[0], 0);
if (ecode && ecode != RE_NOMATCH) {
reerror(failed, &cod, buf, sizeof(buf));
fprintf(stderr, "%s, at line %d\n", buf, line);
break;
}
}
else if (buf[0] == ':') {
if (failed) {
len = strlen(buf) - 1;
buf[len] = '\0';
if (failed == RE_EESCAPE && strcmp(buf, ":EESCAPE") == 0)
continue;
if (failed == RE_ESUBREG && strcmp(buf, ":ESUBREG") == 0)
continue;
if (failed == RE_EBRACK && strcmp(buf, ":EBRACK") == 0)
continue;
if (failed == RE_EPAREN && strcmp(buf, ":EPAREN") == 0)
continue;
if (failed == RE_EBRACE && strcmp(buf, ":EBRACE") == 0)
continue;
if (failed == RE_EBADBR && strcmp(buf, ":EBADBR") == 0)
continue;
if (failed == RE_ERANGE && strcmp(buf, ":ERANGE") == 0)
continue;
if (failed == RE_ESPACE && strcmp(buf, ":ESPACE") == 0)
continue;
if (failed == RE_BADRPT && strcmp(buf, ":BADRPT") == 0)
continue;
if (failed == RE_EMPTY && strcmp(buf, ":EMPTY") == 0)
continue;
reerror(failed, &cod, buf, sizeof(buf));
fprintf(stderr, "Error value %d doesn't match: %s, at line %d\n",
failed, buf, line);
break;
}
else if (!ecode) {
fprintf(stderr, "found match when shoudn't, at line %d\n", line);
break;
}
}
else {
if (failed) {
reerror(failed, &cod, buf, sizeof(buf));
fprintf(stderr, "%s, at line %d\n", buf, line);
break;
}
if (sscanf(buf, "%ld,%ld:", &so, &eo) != 2) {
fprintf(stderr, "expecting match offsets at line %d\n", line);
break;
}
else if (ecode) {
fprintf(stderr, "didn't match, at line %d\n", line);
break;
}
else if (group >= 10) {
fprintf(stderr, "syntax error at line %d (too many groups)\n",
line);
break;
}
else if (so != mat[group].rm_so || eo != mat[group].rm_eo) {
fprintf(stderr, "match failed at line %d, got %ld,%ld: ",
line, mat[group].rm_so, mat[group].rm_eo);
if (mat[group].rm_so < mat[group].rm_eo)
fwrite(str + mat[group].rm_so,
mat[group].rm_eo - mat[group].rm_so, 1, stderr);
fputc('\n', stderr);
break;
}
++group;
}
}
fclose(fp);
return (ecode);
}
LispObj *
Lisp_Reexec(LispBuiltin *builtin)
/*
re-exec regex string &key count start end notbol noteol
*/
{
size_t nmatch;
re_mat match[10];
long start, end, length;
int code, cflags, eflags;
char *string;
LispObj *result;
re_cod *regexp;
LispObj *regex, *ostring, *count, *ostart, *oend, *notbol, *noteol;
noteol = ARGUMENT(6);
notbol = ARGUMENT(5);
oend = ARGUMENT(4);
ostart = ARGUMENT(3);
count = ARGUMENT(2);
ostring = ARGUMENT(1);
regex = ARGUMENT(0);
if (STRINGP(regex))
regexp = LispRecomp(builtin, THESTR(regex), cflags = 0);
else {
CHECK_REGEX(regex);
regexp = regex->data.regex.regex;
cflags = regex->data.regex.options;
}
CHECK_STRING(ostring);
if (count == UNSPEC)
nmatch = 1;
else {
CHECK_INDEX(count);
nmatch = FIXNUM_VALUE(count);
if (nmatch > 10)
LispDestroy("%s: COUNT cannot be larger than 10", STRFUN(builtin));
}
if (nmatch && (cflags & RE_NOSUB))
nmatch = 1;
eflags = RE_STARTEND;
if (notbol != UNSPEC && notbol != NIL)
eflags |= RE_NOTBOL;
if (noteol != UNSPEC && noteol != NIL)
eflags |= RE_NOTEOL;
string = THESTR(ostring);
LispCheckSequenceStartEnd(builtin, ostring, ostart, oend,
&start, &end, &length);
match[0].rm_so = start;
match[0].rm_eo = end;
code = reexec(regexp, string, nmatch, &match[0], eflags);
if (code == 0) {
if (nmatch && match[0].rm_eo >= match[0].rm_so) {
result = CONS(CONS(FIXNUM(match[0].rm_so),
FIXNUM(match[0].rm_eo)), NIL);
if (nmatch > 1 && match[1].rm_eo >= match[1].rm_so) {
int i;
GC_ENTER();
LispObj *cons = result;
GC_PROTECT(result);
for (i = 1;
i < nmatch && match[i].rm_eo >= match[i].rm_so;
i++) {
RPLACD(cons, CONS(CONS(FIXNUM(match[i].rm_so),
FIXNUM(match[i].rm_eo)), NIL));
cons = CDR(cons);
}
GC_LEAVE();
}
}
else
result = NIL;
}
else
result = Knomatch;
/* Maybe shoud cache compiled regex, but better the caller do it */
if (!XREGEXP(regex)) {
refree(regexp);
LispFree(regexp);
}
return (result);
}
