gboolean
camel_url_file_end (const gchar *in,
const gchar *pos,
const gchar *inend,
urlmatch_t *match)
{
register const gchar *inptr = pos;
gchar close_brace;
inptr += strlen (match->pattern);
if (*inptr == '/')
inptr++;
close_brace = url_stop_at_brace (in, match->um_so, NULL);
while (inptr < inend && is_urlsafe (*inptr) && *inptr != close_brace)
inptr++;
if (inptr == pos)
return FALSE;
match->um_eo = (inptr - in);
return TRUE;
}
gboolean
camel_url_web_end (const gchar *in,
const gchar *pos,
const gchar *inend,
urlmatch_t *match)
{
register const gchar *inptr = pos;
gboolean passwd = FALSE;
const gchar *save;
gchar close_brace, open_brace;
gint brace_stack = 0;
gint port;
inptr += strlen (match->pattern);
close_brace = url_stop_at_brace (in, match->um_so, &open_brace);
/* find the end of the domain */
if (is_atom (*inptr)) {
/* might be a domain or user@domain */
save = inptr;
while (inptr < inend) {
if (!is_atom (*inptr))
break;
inptr++;
while (inptr < inend && is_atom (*inptr))
inptr++;
if ((inptr + 1) < inend && *inptr == '.' && (is_atom (inptr[1]) || inptr[1] == '/'))
inptr++;
}
if (*inptr != '@')
inptr = save;
else
inptr++;
goto domain;
} else if (is_domain (*inptr)) {
domain:
while (inptr < inend) {
if (!is_domain (*inptr))
break;
inptr++;
while (inptr < inend && is_domain (*inptr))
inptr++;
if ((inptr + 1) < inend && *inptr == '.' && (is_domain (inptr[1]) || inptr[1] == '/'))
inptr++;
}
} else {
return FALSE;
}
if (inptr < inend) {
switch (*inptr) {
case ':': /* we either have a port or a password */
inptr++;
if (is_digit (*inptr) || passwd) {
port = (*inptr++ - '0');
while (inptr < inend && is_digit (*inptr) && port < 65536)
port = (port * 10) + (*inptr++ - '0');
if (!passwd && (port >= 65536 || *inptr == '@')) {
if (inptr < inend) {
/* this must be a password? */
goto passwd;
}
inptr--;
}
} else {
passwd:
passwd = TRUE;
save = inptr;
while (inptr < inend && is_atom (*inptr))
inptr++;
if ((inptr + 2) < inend) {
if (*inptr == '@') {
inptr++;
if (is_domain (*inptr))
goto domain;
}
return FALSE;
}
}
if (inptr >= inend || *inptr != '/')
break;
/* we have a '/' so there could be a path - fall through */
case '/': /* we've detected a path component to our url */
inptr++;
case '?':
while (inptr < inend && is_urlsafe (*inptr)) {
if (*inptr == open_brace) {
brace_stack++;
} else if (*inptr == close_brace) {
brace_stack--;
if (brace_stack == -1)
break;
}
inptr++;
}
break;
default:
break;
}
}
/* urls are extremely unlikely to end with any
* punctuation, so strip any trailing
* punctuation off. Also strip off any closing
* double-quotes. */
while (inptr > pos && strchr (",.:;?!-|}])\"", inptr[-1]))
inptr--;
match->um_eo = (inptr - in);
return TRUE;
}
gboolean
url_web_end (const char *in, const char *pos, const char *inend, urlmatch_t *match)
{
register const char *inptr = pos;
gboolean openbracket = FALSE;
gboolean passwd = FALSE;
const char *save;
char close_brace;
int port, val, n;
char *end;
inptr += strlen (match->pattern);
close_brace = url_stop_at_brace (in, match->um_so);
/* find the end of the domain */
if (is_digit (*inptr)) {
goto ip_literal2;
} else if (is_atom (*inptr)) {
/* might be a domain or user@domain */
save = inptr;
while (inptr < inend) {
if (!is_atom (*inptr))
break;
inptr++;
while (inptr < inend && is_atom (*inptr))
inptr++;
if ((inptr + 1) < inend && *inptr == '.' && is_atom (inptr[1]))
inptr++;
}
if (*inptr != '@')
inptr = save;
else
inptr++;
if (*inptr == '[') {
/* IPv6 (or possibly IPv4) address literal */
goto ip_literal;
}
if (is_domain (*inptr)) {
/* domain name or IPv4 address */
goto domain;
}
} else if (*inptr == '[') {
ip_literal:
openbracket = TRUE;
inptr++;
if (is_digit (*inptr)) {
ip_literal2:
/* could be IPv4 or IPv6 */
if ((val = strtol (inptr, &end, 10)) < 0)
return FALSE;
} else if ((*inptr >= 'A' && *inptr <= 'F') || (*inptr >= 'a' && *inptr <= 'f')) {
/* IPv6 address literals are in hex */
if ((val = strtol (inptr, &end, 16)) < 0 || *end != ':')
return FALSE;
} else if (*inptr == ':') {
/* IPv6 can start with a ':' */
end = (char *) inptr;
val = 256; /* invalid value */
} else {
return FALSE;
}
switch (*end) {
case '.': /* IPv4 address literal */
n = 1;
do {
if (val > 255 || *end != '.')
return FALSE;
inptr = end + 1;
if ((val = strtol (inptr, &end, 10)) < 0)
return FALSE;
n++;
} while (n < 4);
if (val > 255 || n < 4 || (openbracket && *end != ']'))
return FALSE;
inptr = end + 1;
break;
case ':': /* IPv6 address literal */
if (!openbracket)
return FALSE;
do {
if (end[1] != ':') {
inptr = end + 1;
if ((val = strtol (inptr, &end, 16)) < 0)
return FALSE;
} else {
inptr = end;
end++;
}
} while (end > inptr && *end == ':');
if (*end != ']')
return FALSE;
inptr = end + 1;
break;
default:
return FALSE;
}
} else if (is_domain (*inptr)) {
domain:
while (inptr < inend) {
if (!is_domain (*inptr))
break;
inptr++;
while (inptr < inend && is_domain (*inptr))
inptr++;
if ((inptr + 1) < inend && *inptr == '.' && (is_domain (inptr[1]) || inptr[1] == '/'))
inptr++;
}
} else {
return FALSE;
}
if (inptr < inend) {
switch (*inptr) {
case ':': /* we either have a port or a password */
inptr++;
if (is_digit (*inptr) || passwd) {
port = (*inptr++ - '0');
while (inptr < inend && is_digit (*inptr) && port < 65536)
port = (port * 10) + (*inptr++ - '0');
if (!passwd && (port >= 65536 || *inptr == '@')) {
if (inptr < inend) {
/* this must be a password? */
goto passwd;
}
inptr--;
}
} else {
passwd:
passwd = TRUE;
save = inptr;
while (inptr < inend && is_atom (*inptr))
inptr++;
if ((inptr + 2) < inend) {
if (*inptr == '@') {
inptr++;
if (is_domain (*inptr))
goto domain;
}
return FALSE;
}
}
if (inptr >= inend || *inptr != '/')
break;
/* we have a '/' so there could be a path - fall through */
case '/': /* we've detected a path component to our url */
inptr++;
while (inptr < inend && is_urlsafe (*inptr) && *inptr != close_brace)
inptr++;
break;
default:
break;
}
}
/* urls are extremely unlikely to end with any
* punctuation, so strip any trailing
* punctuation off. Also strip off any closing
* braces or quotes. */
while (inptr > pos && strchr (",.:;?!-|)}]'\"", inptr[-1]))
inptr--;
match->um_eo = (inptr - in);
return TRUE;
}
