/*
* Process CDATA. When we get here, we've processed "<[CDATA"
*/
xml_token_type_t
xml_process_cdata(xml_reader_t *xr, string_t *res)
{
int cc, state = 0;
cc = xml_getc(xr);
if (cc == EOF)
goto unexpected_eof;
if (cc != '[') {
xml_parse_error(xr, "Unexpected '%c' after <[CDATA in XML stream", cc);
return None;
}
while (state != 3) {
cc = xml_getc(xr);
if (cc == EOF)
goto unexpected_eof;
if (cc == ']') {
if (state == 2) {
string_putc(res, ']');
} else {
++state;
}
} else
if (cc == '>' && state == 2) {
++state;
} else {
while (state) {
string_putc(res, ']');
state--;
}
string_putc(res, cc);
}
}
xr->state = Initial;
return CData;
unexpected_eof:
xml_parse_error(xr, "Unexpected EOF after <[CDATA in XML stream");
return None;
}
/*
* Expand an XML entity.
* For now, we support &<number>; as well as symbolic entities
* lt gt amp
*/
bool
xml_expand_entity(xml_reader_t *xr, string_t *res)
{
char entity[128];
unsigned int elen = 0;
int cc, expanded;
while ((cc = xml_getc(xr)) != ';') {
if (cc == EOF) {
xml_parse_error(xr, "Unexpenced EOF in entity");
return false;
}
if (isspace(cc))
continue;
if (elen + 1 >= sizeof(entity)) {
xml_parse_error(xr, "Entity string too long");
return false;
}
entity[elen++] = cc;
}
entity[elen] = '\0';
if (elen == 0) {
xml_parse_error(xr, "Empty entity &;");
return false;
}
if (!strcasecmp(entity, "lt"))
expanded = '<';
else if (!strcasecmp(entity, "gt"))
expanded = '>';
else if (!strcasecmp(entity, "amp"))
expanded = '&';
else {
const char *es = entity;
if (*es == '#') {
expanded = strtoul(es + 1, (char **) &es, 0);
if (*es == '\0')
goto good;
}
xml_parse_error(xr, "Cannot expand unknown entity &%s;", entity);
return false;
}
good:
string_putc(res, expanded);
return true;
}
/*
* Skip any space in the input stream, and copy if to @result
*/
void
xml_skip_space(xml_reader_t *xr, string_t *result)
{
int cc;
while ((cc = xml_getc(xr)) != EOF) {
if (!isspace(cc)) {
xml_ungetc(xr, cc);
break;
}
if (result)
string_putc(result, cc);
}
}
static int
handle_header(struct string* s, int* httpres)
{
char* b;
size_t l, i;
char* p, * p2;
b = string_get(s);
l = string_length(s);
if (l < 4) return HTTP_ENETERR;
for (i = 0; i < l; i++) if (b[i] == 0) return HTTP_ENETERR;
if (!string_putc(s, 0)) return HTTP_ENOMEM;
if (memcmp(b, "HTTP", 4)) return HTTP_ENETERR;
if ((p = strchr(b, ' ')) == NULL) return HTTP_ENETERR;
if ((p2 = strchr(p + 1, ' ')) == NULL) return HTTP_ENETERR;
*p2 = 0;
*httpres = atoi(p);
*p2 = 32;
return HTTP_EOK;
}
xml_token_type_t
xml_get_token_tag(xml_reader_t *xr, string_t *res)
{
int cc, oc;
xml_skip_space(xr, NULL);
cc = xml_getc(xr);
if (cc == EOF) {
xml_parse_error(xr, "Unexpected EOF while parsing tag");
return None;
}
string_putc(res, cc);
switch (cc) {
case '<':
goto error;
case '?':
if ((cc = xml_getc(xr)) != '>')
goto error;
string_putc(res, cc);
xr->state = Initial;
return RightAngleQ;
case '>':
xr->state = Initial;
return RightAngle;
case '/':
if ((cc = xml_getc(xr)) != '>')
goto error;
string_putc(res, cc);
xr->state = Initial;
return RightAngleSlash;
case '=':
return Equals;
case 'a' ... 'z':
case 'A' ... 'Z':
case '_':
case '!':
while ((cc = xml_getc(xr)) != EOF) {
if (!isalnum(cc) && cc != '_' && cc != '!' && cc != ':' && cc != '-') {
xml_ungetc(xr, cc);
break;
}
string_putc(res, cc);
}
return Identifier;
case '\'':
case '"':
string_destroy(res);
oc = cc;
while (1) {
cc = xml_getc(xr);
if (cc == EOF)
goto unexpected_eof;
if (cc == '\\' && oc == '"') {
cc = xml_getc(xr);
if (cc == EOF)
goto unexpected_eof;
} else
if (cc == oc)
break;
string_putc(res, cc);
}
return QuotedString;
default:
break;
}
error:
xml_parse_error(xr, "Unexpected character %c in XML document", cc);
return None;
unexpected_eof:
xml_parse_error(xr, "Unexpected EOF while parsing quoted string");
return None;
}
/*
* While in state Initial, obtain the next token
*/
xml_token_type_t
xml_get_token_initial(xml_reader_t *xr, string_t *res)
{
xml_token_type_t token;
int cc;
restart:
/* Eat initial white space and store it in @res */
xml_skip_space(xr, res);
cc = xml_getc(xr);
if (cc == EOF) {
string_destroy(res);
return EndOfDocument;
}
if (cc == '<') {
/* Discard the white space in @res - we're not interested in that. */
string_destroy(res);
string_putc(res, cc);
if (xr->state != Initial) {
xml_parse_error(xr, "Unexpected < in XML stream (state %s)",
xml_parser_state_name(xr->state));
return None;
}
/* tag is legal here */
xr->state = Tag;
cc = xml_getc(xr);
switch (cc) {
case '/':
string_putc(res, cc);
return LeftAngleSlash;
case '?':
string_putc(res, cc);
return LeftAngleQ;
case '!':
string_putc(res, cc);
/* If it's <!IDENTIFIER, return LeftAngleExclam */
cc = xml_getc(xr);
if (cc == '[') {
/* Looks like CDATA */
if (!xml_get_identifier(xr, res) || strcmp("CDATA", res->string)) {
xml_parse_error(xr, "Unexpected <[%s in XML stream", res->string);
return None;
}
string_destroy(res);
return xml_process_cdata(xr, res);
}
if (cc != '-') {
xml_ungetc(xr, cc);
return LeftAngleExclam;
}
token = xml_skip_comment(xr);
if (token == Comment) {
xr->state = Initial;
string_destroy(res);
goto restart;
}
return token;
default:
xml_ungetc(xr, cc);
break;
}
return LeftAngle;
}
// Looks like CDATA.
// Ignore initial newline, then scan to next <
do {
if (cc == '<') {
/* Looks like we're done.
* FIXME: handle comments within CDATA?
*/
xml_ungetc(xr, cc);
break;
} else
if (cc == '&') {
if (!xml_expand_entity(xr, res))
return None;
} else {
string_putc(res, cc);
}
cc = xml_getc(xr);
} while (cc != EOF);
string_trim_empty_lines(res);
return CData;
}
