/*
* UTF-32 string version of strncasecmp().
*/
int
idn__utf32_strncasecmp(const unsigned long *str1, const unsigned long *str2,
size_t n) {
unsigned long c1, c2;
while (n > 0 && *str1 != '\0') {
c1 = ASCII_TOLOWER(*str1);
c2 = ASCII_TOLOWER(*str2);
if (c1 > c2)
return (1);
else if (c1 < c2)
return (-1);
str1++;
str2++;
n--;
}
if (n > 0) {
c1 = ASCII_TOLOWER(*str1);
c2 = ASCII_TOLOWER(*str2);
if (c1 > c2)
return (1);
else if (c1 < c2)
return (-1);
}
return (0);
}
int
idn__util_asciihaveaceprefix(const char *str, const char *prefix) {
assert(str != NULL && prefix != NULL);
while (*prefix != '\0') {
if (ASCII_TOLOWER(*str) != ASCII_TOLOWER(*prefix))
return 0;
str++;
prefix++;
}
return (1);
}
int
idn_ucs4_strcasecmp(const unsigned long *str1, const unsigned long *str2) {
unsigned long c1, c2;
while (*str1 != '\0') {
c1 = ASCII_TOLOWER(*str1);
c2 = ASCII_TOLOWER(*str2);
if (c1 > c2)
return (1);
else if (c1 < c2)
return (-1);
str1++;
str2++;
}
c1 = ASCII_TOLOWER(*str1);
c2 = ASCII_TOLOWER(*str2);
if (c1 > c2)
return (1);
else if (c1 < c2)
return (-1);
return (0);
}
enum queryparse_ret queryparse_parse(struct queryparse *parser,
char *word, unsigned int *len) {
char c;
*len = 0;
JUMP(parser->state);
/* not in any entities yet */
toplevel_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
/* note: AND and OR could be more efficiently detected by special
* casing their first values, but that would make them harder to
* change */
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = ASCII_TOLOWER(c);
} else {
word[*len] = '\0';
parser->state = TOPLEVEL;
return QUERYPARSE_WORD;
}
if (c == AND_OP[0]) {
/* it could be an AND op */
goto inand_label;
} else if (c == OR_OP[0]) {
/* it could be an OR op */
goto inor_label;
} else {
/* start of a normal word */
goto inword_label;
}
break;
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
word[*len] = '\0';
parser->state = TOPLEVEL;
return QUERYPARSE_WORD;
}
/* start of a normal word */
goto inword_label;
break;
case NOSTOP_OP:
/* start of a non-stopping word */
goto inword_nostop_label;
break;
case EXCLUDE_OP:
/* start of an excluded word */
goto inword_exclude_label;
break;
case PHRASE_DELIM:
parser->state = INPHRASE;
return QUERYPARSE_START_PHRASE;
break;
case MOD_START:
goto inmod_mod_label;
break;
case ASCII_CASE_SPACE:
/* silently eat whitespace */
break;
case '(':
parser->warn |= QUERYPARSE_WARN_PARENS_BOOLEAN; /*warn,fallthrough*/
default:
/* anything else we don't record in the word, but go to inword
* anyway (so we know when a string of junk characters occurred) */
goto inword_label;
break;
}
}
/* if we end in toplevel we just EOF */
goto endfile_label;
/* in a normal word, we may or may not have actually pushed any characters onto
* the return word */
inword_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
}
break;
case '-':
default:
/* break across two punctuation marks in a row */
goto punc_label;
case ASCII_CASE_CONTROL:
/* ignore junk characters */
break;
case ASCII_CASE_SPACE:
/* word ended */
if (*len) {
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto toplevel_label;
}
break;
}
}
/* if we end in a word, we have to return it */
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE; /* transition to eof */
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
}
goto endfile_label;
punc_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
}
goto inword_label;
case ASCII_CASE_PUNCTUATION:
/* break across two punctuation marks in a row */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
default:
case ASCII_CASE_CONTROL:
/* ignore junk characters */
break;
case ASCII_CASE_SPACE:
/* word ended */
if (*len) {
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto toplevel_label;
}
break;
}
}
/* if we end in a word, we have to return it */
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE; /* transition to eof */
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
}
goto endfile_label;
/* in a word that shouldn't be stopped (pretty much the same as normal word) */
inword_nostop_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD_NOSTOP;
}
break;
case ASCII_CASE_CONTROL:
/* ignore */
break;
default:
goto inword_nostop_punc_label;
case ASCII_CASE_SPACE:
/* word ended */
if (*len) {
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD_NOSTOP;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto toplevel_label;
}
break;
}
}
/* if we end in a word, we have to return it */
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE; /* transition to eof */
return QUERYPARSE_WORD_NOSTOP;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
}
goto endfile_label;
/* in a word that shouldn't be stopped (pretty much the same as normal word) */
inword_nostop_punc_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD_NOSTOP;
}
goto inword_nostop_label;
case ASCII_CASE_CONTROL:
/* ignore */
break;
case '-':
default: /* break across punctuation */
case ASCII_CASE_SPACE:
/* word ended */
if (*len) {
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD_NOSTOP;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto toplevel_label;
}
break;
}
}
/* if we end in a word, we have to return it */
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE; /* transition to eof */
return QUERYPARSE_WORD_NOSTOP;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
}
goto endfile_label;
/* in a word that should be excluded (pretty much the same as normal word) */
inword_exclude_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD_EXCLUDE;
}
break;
case ASCII_CASE_CONTROL:
/* ignore */
break;
default:
goto inword_exclude_punc_label;
case ASCII_CASE_SPACE:
/* word ended */
if (*len) {
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD_EXCLUDE;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto toplevel_label;
}
break;
}
}
/* if we end in a word, we have to return it */
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE; /* transition to eof */
return QUERYPARSE_WORD_EXCLUDE;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
}
goto endfile_label;
/* in a word that should be excluded (pretty much the same as normal word) */
inword_exclude_punc_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD_EXCLUDE;
}
break;
case ASCII_CASE_CONTROL:
/* ignore */
break;
case '-':
default: /* break across junk characters */
case ASCII_CASE_SPACE:
/* word ended */
if (*len) {
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD_EXCLUDE;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto toplevel_label;
}
break;
}
}
/* if we end in a word, we have to return it */
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE; /* transition to eof */
return QUERYPARSE_WORD_EXCLUDE;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
}
goto endfile_label;
/* in what may be an OR operator */
inor_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = ASCII_TOLOWER(c);
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
}
if (c != OR_OP[*len - 1]) {
/* it turned out not to be OR, go back to regular word
* parsing */
goto inword_label;
}
break;
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
}
/* it turned out not to be OR, go back to regular word parsing */
goto inword_label;
break;
case ASCII_CASE_CONTROL:
goto inword_label;
case '-':
default:
goto punc_label;
case ASCII_CASE_SPACE:
/* word ended */
if (!OR_OP[*len]) {
/* whitespace ended OR operator */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_OR;
} else if (*len) {
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto toplevel_label;
}
break;
}
}
/* if we end in an unfinished OR, return it as word */
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE;
return QUERYPARSE_WORD;
} else goto inword_label;
/* in what may be an AND operator */
inand_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = ASCII_TOLOWER(c);
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
}
if (c != AND_OP[*len - 1]) {
/* it turned out not to be AND, go back to regular word
* parsing */
goto inword_label;
}
break;
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
}
/* it turned out not to be AND, go back to regular word parsing */
goto inword_label;
break;
case ASCII_CASE_CONTROL:
goto inword_label;
case '-':
default:
goto punc_label;
case ASCII_CASE_SPACE:
/* word ended */
if (!AND_OP[*len]) {
/* whitespace ended OR operator */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_AND;
} else if (*len) {
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto toplevel_label;
}
break;
}
}
/* if we end in an unfinished AND, return it as word */
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE;
return QUERYPARSE_WORD;
} else goto inword_label;
/* inside a phrase */
inphrase_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = INPHRASE; /* transition back to phrase */
return QUERYPARSE_WORD;
}
/* phrase word started */
goto inphrase_word_label;
break;
case PHRASE_DELIM:
/* phrase ended */
goto inphrase_end_label;
break;
case ASCII_CASE_SPACE:
/* silently eat whitespace characters */
break;
case ASCII_CASE_CONTROL:
default:
/* phrase word started */
goto inphrase_word_label;
break;
}
}
/* if we end in an unfinished phrase, warn them */
parser->warn |= QUERYPARSE_WARN_QUOTES_UNMATCHED;
goto endfile_label;
/* in a word, inside a phrase */
inphrase_word_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = INPHRASE; /* transition back to phrase */
return QUERYPARSE_WORD;
}
break;
case PHRASE_DELIM:
/* phrase ended */
if (*len) {
word[*len] = '\0';
parser->state = INPHRASE_END;
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto inphrase_end_label;
}
break;
case '-':
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = INPHRASE; /* transition back to phrase */
return QUERYPARSE_WORD;
}
break;
case ASCII_CASE_CONTROL:
/* ignore */
break;
default:
goto inphrase_word_punc_label;
case ASCII_CASE_SPACE:
/* phrase word ended */
if (*len) {
word[*len] = '\0';
parser->state = INPHRASE;
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto toplevel_label;
}
break;
case CASE_END_SENTENCE:
/* watch out for the end of sentences */
goto endsentence_label;
break;
}
}
/* if we end in a word in an unfinished phrase, warn them and return word */
parser->warn |= QUERYPARSE_WARN_QUOTES_UNMATCHED;
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE;
return QUERYPARSE_WORD;
} else goto inphrase_end_label;
inphrase_word_punc_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = INPHRASE; /* transition back to phrase */
return QUERYPARSE_WORD;
}
/* phrase word started */
goto inphrase_word_label;
break;
case PHRASE_DELIM:
/* phrase ended */
if (*len) {
word[*len] = '\0';
parser->state = INPHRASE_END;
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto inphrase_end_label;
}
break;
case ASCII_CASE_CONTROL:
/* ignore */
break;
default: /* break across junk characters */
case ASCII_CASE_SPACE:
/* phrase word ended */
if (*len) {
word[*len] = '\0';
parser->state = INPHRASE;
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto toplevel_label;
}
break;
case CASE_END_SENTENCE:
/* watch out for the end of sentences */
goto endsentence_label;
break;
}
}
/* if we end in a word in an unfinished phrase, warn them and return word */
parser->warn |= QUERYPARSE_WARN_QUOTES_UNMATCHED;
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE;
return QUERYPARSE_WORD;
} else goto inphrase_end_label;
/* need to return phrase end and then continue */
inphrase_end_label:
parser->state = TOPLEVEL;
return QUERYPARSE_END_PHRASE;
/* need to return sentence end and then continue */
endsentence_end_label:
parser->state = INPHRASE;
return QUERYPARSE_END_SENTENCE;
/* in what might be the end of a sentence */
endsentence_label:
while (parser->buf < parser->end) {
/* don't consume this character (yet) */
c = *parser->buf;
switch (c) {
case ASCII_CASE_SPACE:
/* it was the end of a sentence */
parser->buf++; /* consume the byte */
parser->bytes++;
assert(*len);
word[*len] = '\0';
parser->state = ENDSENTENCE_END;
return QUERYPARSE_WORD;
default:
/* back to a normal phrase word */
goto inphrase_word_label;
break;
}
}
/* finishing in endsentence means we need to return the phrase word */
if (*len) {
word[*len] = '\0';
parser->state = INPHRASE;
return QUERYPARSE_WORD;
} else goto inphrase_word_label;
/* parsing the modifier of a modifier */
inmod_mod_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case '-':
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
}
break;
case MOD_STRING_END:
/* end of the modifier in the mod */
word[*len] = '\0';
parser->state = INMOD; /* transition to inmod */
return QUERYPARSE_START_MODIFIER;
break;
case MOD_START:
parser->warn |= QUERYPARSE_WARN_PARENS_NESTED; /* warn,fallthrough*/
default:
case ASCII_CASE_SPACE:
/* it wasn't a mod after all (return word) */
word[*len] = '\0';
parser->state = TOPLEVEL; /* transition back to toplevel */
return QUERYPARSE_WORD;
break;
}
}
/* if we end in a mod, return as word */
parser->warn |= QUERYPARSE_WARN_PARENS_UNMATCHED;
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE;
return QUERYPARSE_WORD;
} else goto inmod_end_label;
/* in the parameters of a modifier */
inmod_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
case ASCII_CASE_EXTENDED:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = INMOD; /* transition to this state */
return QUERYPARSE_WORD;
}
/* start parsing word */
goto inmod_word_label;
break;
case MOD_END:
goto inmod_end_label;
break;
case ASCII_CASE_SPACE:
/* silently eat whitespace */
break;
case MOD_START:
parser->warn |= QUERYPARSE_WARN_PARENS_NESTED; /* warn,fallthrough*/
default:
/* anything else starts a word, but isn't recorded */
goto inmod_word_label;
break;
}
}
/* if we end in a mod, return as word */
parser->warn |= QUERYPARSE_WARN_PARENS_UNMATCHED;
goto endfile_label;
/* in a word, in a modifier */
inmod_word_label:
while (parser->buf < parser->end) {
c = *parser->buf++;
parser->bytes++;
switch (c) {
case ASCII_CASE_UPPER:
c = ASCII_TOLOWER(c);
case '-':
case ASCII_CASE_LOWER:
case ASCII_CASE_DIGIT:
/* push character onto word */
if (*len < parser->maxwordlen) {
word[(*len)++] = c;
} else {
/* word finished due to length constraint */
word[*len] = '\0';
parser->state = INMOD; /* transition back to inmod */
return QUERYPARSE_WORD;
}
break;
case MOD_END:
if (*len) {
/* return recorded word */
word[*len] = '\0';
parser->state = INMOD_END;
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto inmod_end_label;
}
break;
case '\'':
/* don't break across apostrophe */
break;
case MOD_START:
parser->warn |= QUERYPARSE_WARN_PARENS_NESTED; /* warn,fallthrough*/
default: /* break across other crap */
case ASCII_CASE_SPACE:
if (*len) {
/* return recorded word */
word[*len] = '\0';
parser->state = INMOD;
return QUERYPARSE_WORD;
} else {
/* it was empty, so warn them of that */
parser->warn |= QUERYPARSE_WARN_NONWORD;
goto inmod_label;
}
break;
}
}
/* if we end in a mod, return as word */
parser->warn |= QUERYPARSE_WARN_PARENS_UNMATCHED;
if (*len) {
word[*len] = '\0';
parser->state = ENDFILE;
return QUERYPARSE_WORD;
} else goto inmod_end_label;
/* modifier has ended */
inmod_end_label:
parser->state = TOPLEVEL;
return QUERYPARSE_END_MODIFIER;
endfile_label:
parser->state = ENDFILE;
return QUERYPARSE_EOF;
err_label:
if (!parser->err) {
parser->err = EINVAL;
}
return ERR;
}
