2 * uri.c: set of generic URI related routines
4 * Reference: RFCs 3986, 2732 and 2373
6 * Copyright (C) 1998-2003 Daniel Veillard. All Rights Reserved.
8 * Permission is hereby granted, free of charge, to any person obtaining a copy
9 * of this software and associated documentation files (the "Software"), to deal
10 * in the Software without restriction, including without limitation the rights
11 * to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
12 * copies of the Software, and to permit persons to whom the Software is
13 * furnished to do so, subject to the following conditions:
15 * The above copyright notice and this permission notice shall be included in
16 * all copies or substantial portions of the Software.
18 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
19 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
20 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
21 * DANIEL VEILLARD BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER
22 * IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
23 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
25 * Except as contained in this notice, the name of Daniel Veillard shall not
26 * be used in advertising or otherwise to promote the sale, use or other
27 * dealings in this Software without prior written authorization from him.
33 * Copyright (C) 2007, 2009-2010 Red Hat, Inc.
35 * This library is free software; you can redistribute it and/or
36 * modify it under the terms of the GNU Lesser General Public
37 * License as published by the Free Software Foundation; either
38 * version 2.1 of the License, or (at your option) any later version.
40 * This library is distributed in the hope that it will be useful,
41 * but WITHOUT ANY WARRANTY; without even the implied warranty of
42 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
43 * Lesser General Public License for more details.
45 * You should have received a copy of the GNU Lesser General Public
46 * License along with this library; if not, write to the Free Software
47 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
50 * Richard W.M. Jones <rjones@redhat.com>
60 static void uri_clean(URI *uri);
63 * Old rule from 2396 used in legacy handling code
64 * alpha = lowalpha | upalpha
66 #define IS_ALPHA(x) (IS_LOWALPHA(x) || IS_UPALPHA(x))
70 * lowalpha = "a" | "b" | "c" | "d" | "e" | "f" | "g" | "h" | "i" | "j" |
71 * "k" | "l" | "m" | "n" | "o" | "p" | "q" | "r" | "s" | "t" |
72 * "u" | "v" | "w" | "x" | "y" | "z"
75 #define IS_LOWALPHA(x) (((x) >= 'a') && ((x) <= 'z'))
78 * upalpha = "A" | "B" | "C" | "D" | "E" | "F" | "G" | "H" | "I" | "J" |
79 * "K" | "L" | "M" | "N" | "O" | "P" | "Q" | "R" | "S" | "T" |
80 * "U" | "V" | "W" | "X" | "Y" | "Z"
82 #define IS_UPALPHA(x) (((x) >= 'A') && ((x) <= 'Z'))
88 * digit = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9"
90 #define IS_DIGIT(x) (((x) >= '0') && ((x) <= '9'))
93 * alphanum = alpha | digit
96 #define IS_ALPHANUM(x) (IS_ALPHA(x) || IS_DIGIT(x))
99 * mark = "-" | "_" | "." | "!" | "~" | "*" | "'" | "(" | ")"
102 #define IS_MARK(x) (((x) == '-') || ((x) == '_') || ((x) == '.') || \
103 ((x) == '!') || ((x) == '~') || ((x) == '*') || ((x) == '\'') || \
104 ((x) == '(') || ((x) == ')'))
107 * unwise = "{" | "}" | "|" | "\" | "^" | "`"
110 #define IS_UNWISE(p) \
111 (((*(p) == '{')) || ((*(p) == '}')) || ((*(p) == '|')) || \
112 ((*(p) == '\\')) || ((*(p) == '^')) || ((*(p) == '[')) || \
113 ((*(p) == ']')) || ((*(p) == '`')))
115 * reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" | "$" | "," |
119 #define IS_RESERVED(x) (((x) == ';') || ((x) == '/') || ((x) == '?') || \
120 ((x) == ':') || ((x) == '@') || ((x) == '&') || ((x) == '=') || \
121 ((x) == '+') || ((x) == '$') || ((x) == ',') || ((x) == '[') || \
125 * unreserved = alphanum | mark
128 #define IS_UNRESERVED(x) (IS_ALPHANUM(x) || IS_MARK(x))
131 * Skip to next pointer char, handle escaped sequences
134 #define NEXT(p) ((*p == '%')? p += 3 : p++)
137 * Productions from the spec.
139 * authority = server | reg_name
140 * reg_name = 1*( unreserved | escaped | "$" | "," |
141 * ";" | ":" | "@" | "&" | "=" | "+" )
143 * path = [ abs_path | opaque_part ]
147 /************************************************************************
151 ************************************************************************/
153 #define ISA_DIGIT(p) ((*(p) >= '0') && (*(p) <= '9'))
154 #define ISA_ALPHA(p) (((*(p) >= 'a') && (*(p) <= 'z')) || \
155 ((*(p) >= 'A') && (*(p) <= 'Z')))
156 #define ISA_HEXDIG(p) \
157 (ISA_DIGIT(p) || ((*(p) >= 'a') && (*(p) <= 'f')) || \
158 ((*(p) >= 'A') && (*(p) <= 'F')))
161 * sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
162 * / "*" / "+" / "," / ";" / "="
164 #define ISA_SUB_DELIM(p) \
165 (((*(p) == '!')) || ((*(p) == '$')) || ((*(p) == '&')) || \
166 ((*(p) == '(')) || ((*(p) == ')')) || ((*(p) == '*')) || \
167 ((*(p) == '+')) || ((*(p) == ',')) || ((*(p) == ';')) || \
168 ((*(p) == '=')) || ((*(p) == '\'')))
171 * gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@"
173 #define ISA_GEN_DELIM(p) \
174 (((*(p) == ':')) || ((*(p) == '/')) || ((*(p) == '?')) || \
175 ((*(p) == '#')) || ((*(p) == '[')) || ((*(p) == ']')) || \
179 * reserved = gen-delims / sub-delims
181 #define ISA_RESERVED(p) (ISA_GEN_DELIM(p) || (ISA_SUB_DELIM(p)))
184 * unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
186 #define ISA_UNRESERVED(p) \
187 ((ISA_ALPHA(p)) || (ISA_DIGIT(p)) || ((*(p) == '-')) || \
188 ((*(p) == '.')) || ((*(p) == '_')) || ((*(p) == '~')))
191 * pct-encoded = "%" HEXDIG HEXDIG
193 #define ISA_PCT_ENCODED(p) \
194 ((*(p) == '%') && (ISA_HEXDIG(p + 1)) && (ISA_HEXDIG(p + 2)))
197 * pchar = unreserved / pct-encoded / sub-delims / ":" / "@"
199 #define ISA_PCHAR(p) \
200 (ISA_UNRESERVED(p) || ISA_PCT_ENCODED(p) || ISA_SUB_DELIM(p) || \
201 ((*(p) == ':')) || ((*(p) == '@')))
204 * rfc3986_parse_scheme:
205 * @uri: pointer to an URI structure
206 * @str: pointer to the string to analyze
208 * Parse an URI scheme
210 * ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
212 * Returns 0 or the error code
215 rfc3986_parse_scheme(URI *uri, const char **str) {
225 while (ISA_ALPHA(cur) || ISA_DIGIT(cur) ||
226 (*cur == '+') || (*cur == '-') || (*cur == '.')) cur++;
229 uri->scheme = g_strndup(*str, cur - *str);
236 * rfc3986_parse_fragment:
237 * @uri: pointer to an URI structure
238 * @str: pointer to the string to analyze
240 * Parse the query part of an URI
242 * fragment = *( pchar / "/" / "?" )
243 * NOTE: the strict syntax as defined by 3986 does not allow '[' and ']'
244 * in the fragment identifier but this is used very broadly for
245 * xpointer scheme selection, so we are allowing it here to not break
246 * for example all the DocBook processing chains.
248 * Returns 0 or the error code
251 rfc3986_parse_fragment(URI *uri, const char **str)
260 while ((ISA_PCHAR(cur)) || (*cur == '/') || (*cur == '?') ||
261 (*cur == '[') || (*cur == ']') ||
262 ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))
265 g_free(uri->fragment);
266 if (uri->cleanup & 2)
267 uri->fragment = g_strndup(*str, cur - *str);
269 uri->fragment = uri_string_unescape(*str, cur - *str, NULL);
276 * rfc3986_parse_query:
277 * @uri: pointer to an URI structure
278 * @str: pointer to the string to analyze
280 * Parse the query part of an URI
284 * Returns 0 or the error code
287 rfc3986_parse_query(URI *uri, const char **str)
296 while ((ISA_PCHAR(cur)) || (*cur == '/') || (*cur == '?') ||
297 ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))
301 uri->query = g_strndup (*str, cur - *str);
308 * rfc3986_parse_port:
309 * @uri: pointer to an URI structure
310 * @str: the string to analyze
312 * Parse a port part and fills in the appropriate fields
313 * of the @uri structure
317 * Returns 0 or the error code
320 rfc3986_parse_port(URI *uri, const char **str)
322 const char *cur = *str;
325 if (ISA_DIGIT(cur)) {
326 while (ISA_DIGIT(cur)) {
327 port = port * 10 + (*cur - '0');
343 * rfc3986_parse_user_info:
344 * @uri: pointer to an URI structure
345 * @str: the string to analyze
347 * Parse an user informations part and fills in the appropriate fields
348 * of the @uri structure
350 * userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
352 * Returns 0 or the error code
355 rfc3986_parse_user_info(URI *uri, const char **str)
360 while (ISA_UNRESERVED(cur) || ISA_PCT_ENCODED(cur) ||
361 ISA_SUB_DELIM(cur) || (*cur == ':'))
366 if (uri->cleanup & 2)
367 uri->user = g_strndup(*str, cur - *str);
369 uri->user = uri_string_unescape(*str, cur - *str, NULL);
378 * rfc3986_parse_dec_octet:
379 * @str: the string to analyze
381 * dec-octet = DIGIT ; 0-9
382 * / %x31-39 DIGIT ; 10-99
383 * / "1" 2DIGIT ; 100-199
384 * / "2" %x30-34 DIGIT ; 200-249
385 * / "25" %x30-35 ; 250-255
389 * Returns 0 if found and skipped, 1 otherwise
392 rfc3986_parse_dec_octet(const char **str) {
393 const char *cur = *str;
395 if (!(ISA_DIGIT(cur)))
397 if (!ISA_DIGIT(cur+1))
399 else if ((*cur != '0') && (ISA_DIGIT(cur + 1)) && (!ISA_DIGIT(cur+2)))
401 else if ((*cur == '1') && (ISA_DIGIT(cur + 1)) && (ISA_DIGIT(cur + 2)))
403 else if ((*cur == '2') && (*(cur + 1) >= '0') &&
404 (*(cur + 1) <= '4') && (ISA_DIGIT(cur + 2)))
406 else if ((*cur == '2') && (*(cur + 1) == '5') &&
407 (*(cur + 2) >= '0') && (*(cur + 1) <= '5'))
415 * rfc3986_parse_host:
416 * @uri: pointer to an URI structure
417 * @str: the string to analyze
419 * Parse an host part and fills in the appropriate fields
420 * of the @uri structure
422 * host = IP-literal / IPv4address / reg-name
423 * IP-literal = "[" ( IPv6address / IPvFuture ) "]"
424 * IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
425 * reg-name = *( unreserved / pct-encoded / sub-delims )
427 * Returns 0 or the error code
430 rfc3986_parse_host(URI *uri, const char **str)
432 const char *cur = *str;
437 * IPv6 and future addressing scheme are enclosed between brackets
441 while ((*cur != ']') && (*cur != 0))
449 * try to parse an IPv4
451 if (ISA_DIGIT(cur)) {
452 if (rfc3986_parse_dec_octet(&cur) != 0)
457 if (rfc3986_parse_dec_octet(&cur) != 0)
461 if (rfc3986_parse_dec_octet(&cur) != 0)
465 if (rfc3986_parse_dec_octet(&cur) != 0)
472 * then this should be a hostname which can be empty
474 while (ISA_UNRESERVED(cur) || ISA_PCT_ENCODED(cur) || ISA_SUB_DELIM(cur))
478 g_free(uri->authority);
479 uri->authority = NULL;
482 if (uri->cleanup & 2)
483 uri->server = g_strndup(host, cur - host);
485 uri->server = uri_string_unescape(host, cur - host, NULL);
494 * rfc3986_parse_authority:
495 * @uri: pointer to an URI structure
496 * @str: the string to analyze
498 * Parse an authority part and fills in the appropriate fields
499 * of the @uri structure
501 * authority = [ userinfo "@" ] host [ ":" port ]
503 * Returns 0 or the error code
506 rfc3986_parse_authority(URI *uri, const char **str)
513 * try to parse an userinfo and check for the trailing @
515 ret = rfc3986_parse_user_info(uri, &cur);
516 if ((ret != 0) || (*cur != '@'))
520 ret = rfc3986_parse_host(uri, &cur);
521 if (ret != 0) return(ret);
524 ret = rfc3986_parse_port(uri, &cur);
525 if (ret != 0) return(ret);
532 * rfc3986_parse_segment:
533 * @str: the string to analyze
534 * @forbid: an optional forbidden character
535 * @empty: allow an empty segment
537 * Parse a segment and fills in the appropriate fields
538 * of the @uri structure
541 * segment-nz = 1*pchar
542 * segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" )
543 * ; non-zero-length segment without any colon ":"
545 * Returns 0 or the error code
548 rfc3986_parse_segment(const char **str, char forbid, int empty)
553 if (!ISA_PCHAR(cur)) {
558 while (ISA_PCHAR(cur) && (*cur != forbid))
565 * rfc3986_parse_path_ab_empty:
566 * @uri: pointer to an URI structure
567 * @str: the string to analyze
569 * Parse an path absolute or empty and fills in the appropriate fields
570 * of the @uri structure
572 * path-abempty = *( "/" segment )
574 * Returns 0 or the error code
577 rfc3986_parse_path_ab_empty(URI *uri, const char **str)
584 while (*cur == '/') {
586 ret = rfc3986_parse_segment(&cur, 0, 1);
587 if (ret != 0) return(ret);
592 if (uri->cleanup & 2)
593 uri->path = g_strndup(*str, cur - *str);
595 uri->path = uri_string_unescape(*str, cur - *str, NULL);
605 * rfc3986_parse_path_absolute:
606 * @uri: pointer to an URI structure
607 * @str: the string to analyze
609 * Parse an path absolute and fills in the appropriate fields
610 * of the @uri structure
612 * path-absolute = "/" [ segment-nz *( "/" segment ) ]
614 * Returns 0 or the error code
617 rfc3986_parse_path_absolute(URI *uri, const char **str)
627 ret = rfc3986_parse_segment(&cur, 0, 0);
629 while (*cur == '/') {
631 ret = rfc3986_parse_segment(&cur, 0, 1);
632 if (ret != 0) return(ret);
638 if (uri->cleanup & 2)
639 uri->path = g_strndup(*str, cur - *str);
641 uri->path = uri_string_unescape(*str, cur - *str, NULL);
651 * rfc3986_parse_path_rootless:
652 * @uri: pointer to an URI structure
653 * @str: the string to analyze
655 * Parse an path without root and fills in the appropriate fields
656 * of the @uri structure
658 * path-rootless = segment-nz *( "/" segment )
660 * Returns 0 or the error code
663 rfc3986_parse_path_rootless(URI *uri, const char **str)
670 ret = rfc3986_parse_segment(&cur, 0, 0);
671 if (ret != 0) return(ret);
672 while (*cur == '/') {
674 ret = rfc3986_parse_segment(&cur, 0, 1);
675 if (ret != 0) return(ret);
680 if (uri->cleanup & 2)
681 uri->path = g_strndup(*str, cur - *str);
683 uri->path = uri_string_unescape(*str, cur - *str, NULL);
693 * rfc3986_parse_path_no_scheme:
694 * @uri: pointer to an URI structure
695 * @str: the string to analyze
697 * Parse an path which is not a scheme and fills in the appropriate fields
698 * of the @uri structure
700 * path-noscheme = segment-nz-nc *( "/" segment )
702 * Returns 0 or the error code
705 rfc3986_parse_path_no_scheme(URI *uri, const char **str)
712 ret = rfc3986_parse_segment(&cur, ':', 0);
713 if (ret != 0) return(ret);
714 while (*cur == '/') {
716 ret = rfc3986_parse_segment(&cur, 0, 1);
717 if (ret != 0) return(ret);
722 if (uri->cleanup & 2)
723 uri->path = g_strndup(*str, cur - *str);
725 uri->path = uri_string_unescape(*str, cur - *str, NULL);
735 * rfc3986_parse_hier_part:
736 * @uri: pointer to an URI structure
737 * @str: the string to analyze
739 * Parse an hierarchical part and fills in the appropriate fields
740 * of the @uri structure
742 * hier-part = "//" authority path-abempty
747 * Returns 0 or the error code
750 rfc3986_parse_hier_part(URI *uri, const char **str)
757 if ((*cur == '/') && (*(cur + 1) == '/')) {
759 ret = rfc3986_parse_authority(uri, &cur);
760 if (ret != 0) return(ret);
761 ret = rfc3986_parse_path_ab_empty(uri, &cur);
762 if (ret != 0) return(ret);
765 } else if (*cur == '/') {
766 ret = rfc3986_parse_path_absolute(uri, &cur);
767 if (ret != 0) return(ret);
768 } else if (ISA_PCHAR(cur)) {
769 ret = rfc3986_parse_path_rootless(uri, &cur);
770 if (ret != 0) return(ret);
772 /* path-empty is effectively empty */
783 * rfc3986_parse_relative_ref:
784 * @uri: pointer to an URI structure
785 * @str: the string to analyze
787 * Parse an URI string and fills in the appropriate fields
788 * of the @uri structure
790 * relative-ref = relative-part [ "?" query ] [ "#" fragment ]
791 * relative-part = "//" authority path-abempty
796 * Returns 0 or the error code
799 rfc3986_parse_relative_ref(URI *uri, const char *str) {
802 if ((*str == '/') && (*(str + 1) == '/')) {
804 ret = rfc3986_parse_authority(uri, &str);
805 if (ret != 0) return(ret);
806 ret = rfc3986_parse_path_ab_empty(uri, &str);
807 if (ret != 0) return(ret);
808 } else if (*str == '/') {
809 ret = rfc3986_parse_path_absolute(uri, &str);
810 if (ret != 0) return(ret);
811 } else if (ISA_PCHAR(str)) {
812 ret = rfc3986_parse_path_no_scheme(uri, &str);
813 if (ret != 0) return(ret);
815 /* path-empty is effectively empty */
824 ret = rfc3986_parse_query(uri, &str);
825 if (ret != 0) return(ret);
829 ret = rfc3986_parse_fragment(uri, &str);
830 if (ret != 0) return(ret);
842 * @uri: pointer to an URI structure
843 * @str: the string to analyze
845 * Parse an URI string and fills in the appropriate fields
846 * of the @uri structure
848 * scheme ":" hier-part [ "?" query ] [ "#" fragment ]
850 * Returns 0 or the error code
853 rfc3986_parse(URI *uri, const char *str) {
856 ret = rfc3986_parse_scheme(uri, &str);
857 if (ret != 0) return(ret);
862 ret = rfc3986_parse_hier_part(uri, &str);
863 if (ret != 0) return(ret);
866 ret = rfc3986_parse_query(uri, &str);
867 if (ret != 0) return(ret);
871 ret = rfc3986_parse_fragment(uri, &str);
872 if (ret != 0) return(ret);
882 * rfc3986_parse_uri_reference:
883 * @uri: pointer to an URI structure
884 * @str: the string to analyze
886 * Parse an URI reference string and fills in the appropriate fields
887 * of the @uri structure
889 * URI-reference = URI / relative-ref
891 * Returns 0 or the error code
894 rfc3986_parse_uri_reference(URI *uri, const char *str) {
902 * Try first to parse absolute refs, then fallback to relative if
905 ret = rfc3986_parse(uri, str);
908 ret = rfc3986_parse_relative_ref(uri, str);
919 * @str: the URI string to analyze
921 * Parse an URI based on RFC 3986
923 * URI-reference = [ absoluteURI | relativeURI ] [ "#" fragment ]
925 * Returns a newly built URI or NULL in case of error
928 uri_parse(const char *str) {
935 ret = rfc3986_parse_uri_reference(uri, str);
945 * @uri: pointer to an URI structure
946 * @str: the string to analyze
948 * Parse an URI reference string based on RFC 3986 and fills in the
949 * appropriate fields of the @uri structure
951 * URI-reference = URI / relative-ref
953 * Returns 0 or the error code
956 uri_parse_into(URI *uri, const char *str) {
957 return(rfc3986_parse_uri_reference(uri, str));
962 * @str: the URI string to analyze
963 * @raw: if 1 unescaping of URI pieces are disabled
965 * Parse an URI but allows to keep intact the original fragments.
967 * URI-reference = URI / relative-ref
969 * Returns a newly built URI or NULL in case of error
972 uri_parse_raw(const char *str, int raw) {
982 ret = uri_parse_into(uri, str);
990 /************************************************************************
992 * Generic URI structure functions *
994 ************************************************************************/
999 * Simply creates an empty URI
1001 * Returns the new structure or NULL in case of error
1007 ret = g_new0(URI, 1);
1014 * Function to handle properly a reallocation when saving an URI
1015 * Also imposes some limit on the length of an URI string output
1018 realloc2n(char *ret, int *max) {
1023 temp = g_realloc(ret, (tmp + 1));
1030 * @uri: pointer to an URI
1032 * Save the URI as an escaped string
1034 * Returns a new string (to be deallocated by caller)
1037 uri_to_string(URI *uri) {
1044 if (uri == NULL) return(NULL);
1048 ret = g_malloc(max + 1);
1051 if (uri->scheme != NULL) {
1055 temp = realloc2n(ret, &max);
1061 temp = realloc2n(ret, &max);
1066 if (uri->opaque != NULL) {
1069 if (len + 3 >= max) {
1070 temp = realloc2n(ret, &max);
1073 if (IS_RESERVED(*(p)) || IS_UNRESERVED(*(p)))
1076 int val = *(unsigned char *)p++;
1077 int hi = val / 0x10, lo = val % 0x10;
1079 ret[len++] = hi + (hi > 9? 'A'-10 : '0');
1080 ret[len++] = lo + (lo > 9? 'A'-10 : '0');
1084 if (uri->server != NULL) {
1085 if (len + 3 >= max) {
1086 temp = realloc2n(ret, &max);
1091 if (uri->user != NULL) {
1094 if (len + 3 >= max) {
1095 temp = realloc2n(ret, &max);
1098 if ((IS_UNRESERVED(*(p))) ||
1099 ((*(p) == ';')) || ((*(p) == ':')) ||
1100 ((*(p) == '&')) || ((*(p) == '=')) ||
1101 ((*(p) == '+')) || ((*(p) == '$')) ||
1105 int val = *(unsigned char *)p++;
1106 int hi = val / 0x10, lo = val % 0x10;
1108 ret[len++] = hi + (hi > 9? 'A'-10 : '0');
1109 ret[len++] = lo + (lo > 9? 'A'-10 : '0');
1112 if (len + 3 >= max) {
1113 temp = realloc2n(ret, &max);
1121 temp = realloc2n(ret, &max);
1126 if (uri->port > 0) {
1127 if (len + 10 >= max) {
1128 temp = realloc2n(ret, &max);
1131 len += snprintf(&ret[len], max - len, ":%d", uri->port);
1133 } else if (uri->authority != NULL) {
1134 if (len + 3 >= max) {
1135 temp = realloc2n(ret, &max);
1142 if (len + 3 >= max) {
1143 temp = realloc2n(ret, &max);
1146 if ((IS_UNRESERVED(*(p))) ||
1147 ((*(p) == '$')) || ((*(p) == ',')) || ((*(p) == ';')) ||
1148 ((*(p) == ':')) || ((*(p) == '@')) || ((*(p) == '&')) ||
1149 ((*(p) == '=')) || ((*(p) == '+')))
1152 int val = *(unsigned char *)p++;
1153 int hi = val / 0x10, lo = val % 0x10;
1155 ret[len++] = hi + (hi > 9? 'A'-10 : '0');
1156 ret[len++] = lo + (lo > 9? 'A'-10 : '0');
1159 } else if (uri->scheme != NULL) {
1160 if (len + 3 >= max) {
1161 temp = realloc2n(ret, &max);
1167 if (uri->path != NULL) {
1170 * the colon in file:///d: should not be escaped or
1171 * Windows accesses fail later.
1173 if ((uri->scheme != NULL) &&
1175 (((p[1] >= 'a') && (p[1] <= 'z')) ||
1176 ((p[1] >= 'A') && (p[1] <= 'Z'))) &&
1178 (!strcmp(uri->scheme, "file"))) {
1179 if (len + 3 >= max) {
1180 temp = realloc2n(ret, &max);
1188 if (len + 3 >= max) {
1189 temp = realloc2n(ret, &max);
1192 if ((IS_UNRESERVED(*(p))) || ((*(p) == '/')) ||
1193 ((*(p) == ';')) || ((*(p) == '@')) || ((*(p) == '&')) ||
1194 ((*(p) == '=')) || ((*(p) == '+')) || ((*(p) == '$')) ||
1198 int val = *(unsigned char *)p++;
1199 int hi = val / 0x10, lo = val % 0x10;
1201 ret[len++] = hi + (hi > 9? 'A'-10 : '0');
1202 ret[len++] = lo + (lo > 9? 'A'-10 : '0');
1206 if (uri->query != NULL) {
1207 if (len + 1 >= max) {
1208 temp = realloc2n(ret, &max);
1214 if (len + 1 >= max) {
1215 temp = realloc2n(ret, &max);
1222 if (uri->fragment != NULL) {
1223 if (len + 3 >= max) {
1224 temp = realloc2n(ret, &max);
1230 if (len + 3 >= max) {
1231 temp = realloc2n(ret, &max);
1234 if ((IS_UNRESERVED(*(p))) || (IS_RESERVED(*(p))))
1237 int val = *(unsigned char *)p++;
1238 int hi = val / 0x10, lo = val % 0x10;
1240 ret[len++] = hi + (hi > 9? 'A'-10 : '0');
1241 ret[len++] = lo + (lo > 9? 'A'-10 : '0');
1246 temp = realloc2n(ret, &max);
1255 * @uri: pointer to an URI
1257 * Make sure the URI struct is free of content
1260 uri_clean(URI *uri) {
1261 if (uri == NULL) return;
1263 g_free(uri->scheme);
1265 g_free(uri->server);
1271 g_free(uri->fragment);
1272 uri->fragment = NULL;
1273 g_free(uri->opaque);
1275 g_free(uri->authority);
1276 uri->authority = NULL;
1283 * @uri: pointer to an URI
1285 * Free up the URI struct
1288 uri_free(URI *uri) {
1293 /************************************************************************
1295 * Helper functions *
1297 ************************************************************************/
1300 * normalize_uri_path:
1301 * @path: pointer to the path string
1303 * Applies the 5 normalization steps to a path string--that is, RFC 2396
1304 * Section 5.2, steps 6.c through 6.g.
1306 * Normalization occurs directly on the string, no new allocation is done
1308 * Returns 0 or an error code
1311 normalize_uri_path(char *path) {
1317 /* Skip all initial "/" chars. We want to get to the beginning of the
1318 * first non-empty segment.
1321 while (cur[0] == '/')
1326 /* Keep everything we've seen so far. */
1330 * Analyze each segment in sequence for cases (c) and (d).
1332 while (cur[0] != '\0') {
1334 * c) All occurrences of "./", where "." is a complete path segment,
1335 * are removed from the buffer string.
1337 if ((cur[0] == '.') && (cur[1] == '/')) {
1339 /* '//' normalization should be done at this point too */
1340 while (cur[0] == '/')
1346 * d) If the buffer string ends with "." as a complete path segment,
1347 * that "." is removed.
1349 if ((cur[0] == '.') && (cur[1] == '\0'))
1352 /* Otherwise keep the segment. */
1353 while (cur[0] != '/') {
1356 (out++)[0] = (cur++)[0];
1359 while ((cur[0] == '/') && (cur[1] == '/'))
1362 (out++)[0] = (cur++)[0];
1367 /* Reset to the beginning of the first segment for the next sequence. */
1369 while (cur[0] == '/')
1375 * Analyze each segment in sequence for cases (e) and (f).
1377 * e) All occurrences of "<segment>/../", where <segment> is a
1378 * complete path segment not equal to "..", are removed from the
1379 * buffer string. Removal of these path segments is performed
1380 * iteratively, removing the leftmost matching pattern on each
1381 * iteration, until no matching pattern remains.
1383 * f) If the buffer string ends with "<segment>/..", where <segment>
1384 * is a complete path segment not equal to "..", that
1385 * "<segment>/.." is removed.
1387 * To satisfy the "iterative" clause in (e), we need to collapse the
1388 * string every time we find something that needs to be removed. Thus,
1389 * we don't need to keep two pointers into the string: we only need a
1390 * "current position" pointer.
1395 /* At the beginning of each iteration of this loop, "cur" points to
1396 * the first character of the segment we want to examine.
1399 /* Find the end of the current segment. */
1401 while ((segp[0] != '/') && (segp[0] != '\0'))
1404 /* If this is the last segment, we're done (we need at least two
1405 * segments to meet the criteria for the (e) and (f) cases).
1407 if (segp[0] == '\0')
1410 /* If the first segment is "..", or if the next segment _isn't_ "..",
1411 * keep this segment and try the next one.
1414 if (((cur[0] == '.') && (cur[1] == '.') && (segp == cur+3))
1415 || ((segp[0] != '.') || (segp[1] != '.')
1416 || ((segp[2] != '/') && (segp[2] != '\0')))) {
1421 /* If we get here, remove this segment and the next one and back up
1422 * to the previous segment (if there is one), to implement the
1423 * "iteratively" clause. It's pretty much impossible to back up
1424 * while maintaining two pointers into the buffer, so just compact
1425 * the whole buffer now.
1428 /* If this is the end of the buffer, we're done. */
1429 if (segp[2] == '\0') {
1433 /* Valgrind complained, strcpy(cur, segp + 3); */
1434 /* string will overlap, do not use strcpy */
1437 while ((*tmp++ = *segp++) != 0)
1440 /* If there are no previous segments, then keep going from here. */
1442 while ((segp > path) && ((--segp)[0] == '/'))
1447 /* "segp" is pointing to the end of a previous segment; find it's
1448 * start. We need to back up to the previous segment and start
1449 * over with that to handle things like "foo/bar/../..". If we
1450 * don't do this, then on the first pass we'll remove the "bar/..",
1451 * but be pointing at the second ".." so we won't realize we can also
1452 * remove the "foo/..".
1455 while ((cur > path) && (cur[-1] != '/'))
1461 * g) If the resulting buffer string still begins with one or more
1462 * complete path segments of "..", then the reference is
1463 * considered to be in error. Implementations may handle this
1464 * error by retaining these components in the resolved path (i.e.,
1465 * treating them as part of the final URI), by removing them from
1466 * the resolved path (i.e., discarding relative levels above the
1467 * root), or by avoiding traversal of the reference.
1469 * We discard them from the final path.
1471 if (path[0] == '/') {
1473 while ((cur[0] == '/') && (cur[1] == '.') && (cur[2] == '.')
1474 && ((cur[3] == '/') || (cur[3] == '\0')))
1479 while (cur[0] != '\0')
1480 (out++)[0] = (cur++)[0];
1488 static int is_hex(char c) {
1489 if (((c >= '0') && (c <= '9')) ||
1490 ((c >= 'a') && (c <= 'f')) ||
1491 ((c >= 'A') && (c <= 'F')))
1498 * uri_string_unescape:
1499 * @str: the string to unescape
1500 * @len: the length in bytes to unescape (or <= 0 to indicate full string)
1501 * @target: optional destination buffer
1503 * Unescaping routine, but does not check that the string is an URI. The
1504 * output is a direct unsigned char translation of %XX values (no encoding)
1505 * Note that the length of the result can only be smaller or same size as
1508 * Returns a copy of the string, but unescaped, will return NULL only in case
1512 uri_string_unescape(const char *str, int len, char *target) {
1518 if (len <= 0) len = strlen(str);
1519 if (len < 0) return(NULL);
1521 if (target == NULL) {
1522 ret = g_malloc(len + 1);
1528 if ((len > 2) && (*in == '%') && (is_hex(in[1])) && (is_hex(in[2]))) {
1530 if ((*in >= '0') && (*in <= '9'))
1532 else if ((*in >= 'a') && (*in <= 'f'))
1533 *out = (*in - 'a') + 10;
1534 else if ((*in >= 'A') && (*in <= 'F'))
1535 *out = (*in - 'A') + 10;
1537 if ((*in >= '0') && (*in <= '9'))
1538 *out = *out * 16 + (*in - '0');
1539 else if ((*in >= 'a') && (*in <= 'f'))
1540 *out = *out * 16 + (*in - 'a') + 10;
1541 else if ((*in >= 'A') && (*in <= 'F'))
1542 *out = *out * 16 + (*in - 'A') + 10;
1556 * uri_string_escape:
1557 * @str: string to escape
1558 * @list: exception list string of chars not to escape
1560 * This routine escapes a string to hex, ignoring reserved characters (a-z)
1561 * and the characters in the exception list.
1563 * Returns a new escaped string or NULL in case of error.
1566 uri_string_escape(const char *str, const char *list) {
1575 return(g_strdup(str));
1577 if (!(len > 0)) return(NULL);
1580 ret = g_malloc(len);
1584 if (len - out <= 3) {
1585 temp = realloc2n(ret, &len);
1591 if ((ch != '@') && (!IS_UNRESERVED(ch)) && (!strchr(list, ch))) {
1596 ret[out++] = '0' + val;
1598 ret[out++] = 'A' + val - 0xA;
1601 ret[out++] = '0' + val;
1603 ret[out++] = 'A' + val - 0xA;
1614 /************************************************************************
1616 * Public functions *
1618 ************************************************************************/
1622 * @URI: the URI instance found in the document
1623 * @base: the base value
1625 * Computes he final URI of the reference done by checking that
1626 * the given URI is valid, and building the final URI using the
1627 * base URI. This is processed according to section 5.2 of the
1630 * 5.2. Resolving Relative References to Absolute Form
1632 * Returns a new URI string (to be freed by the caller) or NULL in case
1636 uri_resolve(const char *uri, const char *base) {
1638 int ret, len, indx, cur, out;
1644 * 1) The URI reference is parsed into the potential four components and
1645 * fragment identifier, as described in Section 4.3.
1647 * NOTE that a completely empty URI is treated by modern browsers
1648 * as a reference to "." rather than as a synonym for the current
1649 * URI. Should we do that here?
1656 ret = uri_parse_into(ref, uri);
1663 if ((ref != NULL) && (ref->scheme != NULL)) {
1665 * The URI is absolute don't modify.
1667 val = g_strdup(uri);
1674 ret = uri_parse_into(bas, base);
1678 val = uri_to_string(ref);
1683 * the base fragment must be ignored
1685 g_free(bas->fragment);
1686 bas->fragment = NULL;
1687 val = uri_to_string(bas);
1692 * 2) If the path component is empty and the scheme, authority, and
1693 * query components are undefined, then it is a reference to the
1694 * current document and we are done. Otherwise, the reference URI's
1695 * query and fragment components are defined as found (or not found)
1696 * within the URI reference and not inherited from the base URI.
1698 * NOTE that in modern browsers, the parsing differs from the above
1699 * in the following aspect: the query component is allowed to be
1700 * defined while still treating this as a reference to the current
1704 if ((ref->scheme == NULL) && (ref->path == NULL) &&
1705 ((ref->authority == NULL) && (ref->server == NULL))) {
1706 res->scheme = g_strdup(bas->scheme);
1707 if (bas->authority != NULL)
1708 res->authority = g_strdup(bas->authority);
1709 else if (bas->server != NULL) {
1710 res->server = g_strdup(bas->server);
1711 res->user = g_strdup(bas->user);
1712 res->port = bas->port;
1714 res->path = g_strdup(bas->path);
1715 if (ref->query != NULL) {
1716 res->query = g_strdup (ref->query);
1718 res->query = g_strdup(bas->query);
1720 res->fragment = g_strdup(ref->fragment);
1725 * 3) If the scheme component is defined, indicating that the reference
1726 * starts with a scheme name, then the reference is interpreted as an
1727 * absolute URI and we are done. Otherwise, the reference URI's
1728 * scheme is inherited from the base URI's scheme component.
1730 if (ref->scheme != NULL) {
1731 val = uri_to_string(ref);
1734 res->scheme = g_strdup(bas->scheme);
1736 res->query = g_strdup(ref->query);
1737 res->fragment = g_strdup(ref->fragment);
1740 * 4) If the authority component is defined, then the reference is a
1741 * network-path and we skip to step 7. Otherwise, the reference
1742 * URI's authority is inherited from the base URI's authority
1743 * component, which will also be undefined if the URI scheme does not
1744 * use an authority component.
1746 if ((ref->authority != NULL) || (ref->server != NULL)) {
1747 if (ref->authority != NULL)
1748 res->authority = g_strdup(ref->authority);
1750 res->server = g_strdup(ref->server);
1751 res->user = g_strdup(ref->user);
1752 res->port = ref->port;
1754 res->path = g_strdup(ref->path);
1757 if (bas->authority != NULL)
1758 res->authority = g_strdup(bas->authority);
1759 else if (bas->server != NULL) {
1760 res->server = g_strdup(bas->server);
1761 res->user = g_strdup(bas->user);
1762 res->port = bas->port;
1766 * 5) If the path component begins with a slash character ("/"), then
1767 * the reference is an absolute-path and we skip to step 7.
1769 if ((ref->path != NULL) && (ref->path[0] == '/')) {
1770 res->path = g_strdup(ref->path);
1776 * 6) If this step is reached, then we are resolving a relative-path
1777 * reference. The relative path needs to be merged with the base
1778 * URI's path. Although there are many ways to do this, we will
1779 * describe a simple method using a separate string buffer.
1781 * Allocate a buffer large enough for the result string.
1783 len = 2; /* extra / and 0 */
1784 if (ref->path != NULL)
1785 len += strlen(ref->path);
1786 if (bas->path != NULL)
1787 len += strlen(bas->path);
1788 res->path = g_malloc(len);
1792 * a) All but the last segment of the base URI's path component is
1793 * copied to the buffer. In other words, any characters after the
1794 * last (right-most) slash character, if any, are excluded.
1798 if (bas->path != NULL) {
1799 while (bas->path[cur] != 0) {
1800 while ((bas->path[cur] != 0) && (bas->path[cur] != '/'))
1802 if (bas->path[cur] == 0)
1807 res->path[out] = bas->path[out];
1815 * b) The reference's path component is appended to the buffer
1818 if (ref->path != NULL && ref->path[0] != 0) {
1821 * Ensure the path includes a '/'
1823 if ((out == 0) && (bas->server != NULL))
1824 res->path[out++] = '/';
1825 while (ref->path[indx] != 0) {
1826 res->path[out++] = ref->path[indx++];
1832 * Steps c) to h) are really path normalization steps
1834 normalize_uri_path(res->path);
1839 * 7) The resulting URI components, including any inherited from the
1840 * base URI, are recombined to give the absolute form of the URI
1843 val = uri_to_string(res);
1856 * uri_resolve_relative:
1857 * @URI: the URI reference under consideration
1858 * @base: the base value
1860 * Expresses the URI of the reference in terms relative to the
1861 * base. Some examples of this operation include:
1862 * base = "http://site1.com/docs/book1.html"
1863 * URI input URI returned
1864 * docs/pic1.gif pic1.gif
1865 * docs/img/pic1.gif img/pic1.gif
1866 * img/pic1.gif ../img/pic1.gif
1867 * http://site1.com/docs/pic1.gif pic1.gif
1868 * http://site2.com/docs/pic1.gif http://site2.com/docs/pic1.gif
1870 * base = "docs/book1.html"
1871 * URI input URI returned
1872 * docs/pic1.gif pic1.gif
1873 * docs/img/pic1.gif img/pic1.gif
1874 * img/pic1.gif ../img/pic1.gif
1875 * http://site1.com/docs/pic1.gif http://site1.com/docs/pic1.gif
1878 * Note: if the URI reference is really weird or complicated, it may be
1879 * worthwhile to first convert it into a "nice" one by calling
1880 * uri_resolve (using 'base') before calling this routine,
1881 * since this routine (for reasonable efficiency) assumes URI has
1882 * already been through some validation.
1884 * Returns a new URI string (to be freed by the caller) or NULL in case
1888 uri_resolve_relative (const char *uri, const char * base)
1898 char *bptr, *uptr, *vptr;
1899 int remove_path = 0;
1901 if ((uri == NULL) || (*uri == 0))
1905 * First parse URI into a standard form
1908 /* If URI not already in "relative" form */
1909 if (uri[0] != '.') {
1910 ret = uri_parse_into (ref, uri);
1912 goto done; /* Error in URI, return NULL */
1914 ref->path = g_strdup(uri);
1917 * Next parse base into the same standard form
1919 if ((base == NULL) || (*base == 0)) {
1920 val = g_strdup (uri);
1924 if (base[0] != '.') {
1925 ret = uri_parse_into (bas, base);
1927 goto done; /* Error in base, return NULL */
1929 bas->path = g_strdup(base);
1932 * If the scheme / server on the URI differs from the base,
1933 * just return the URI
1935 if ((ref->scheme != NULL) &&
1936 ((bas->scheme == NULL) ||
1937 (strcmp (bas->scheme, ref->scheme)) ||
1938 (strcmp (bas->server, ref->server)))) {
1939 val = g_strdup (uri);
1942 if (bas->path == ref->path ||
1943 (bas->path && ref->path && !strcmp(bas->path, ref->path))) {
1947 if (bas->path == NULL) {
1948 val = g_strdup(ref->path);
1951 if (ref->path == NULL) {
1952 ref->path = (char *) "/";
1957 * At this point (at last!) we can compare the two paths
1959 * First we take care of the special case where either of the
1960 * two path components may be missing (bug 316224)
1962 if (bas->path == NULL) {
1963 if (ref->path != NULL) {
1967 /* exception characters from uri_to_string */
1968 val = uri_string_escape(uptr, "/;&=+$,");
1973 if (ref->path == NULL) {
1974 for (ix = 0; bptr[ix] != 0; ix++) {
1975 if (bptr[ix] == '/')
1979 len = 1; /* this is for a string terminator only */
1982 * Next we compare the two strings and find where they first differ
1984 if ((ref->path[pos] == '.') && (ref->path[pos+1] == '/'))
1986 if ((*bptr == '.') && (bptr[1] == '/'))
1988 else if ((*bptr == '/') && (ref->path[pos] != '/'))
1990 while ((bptr[pos] == ref->path[pos]) && (bptr[pos] != 0))
1993 if (bptr[pos] == ref->path[pos]) {
1995 goto done; /* (I can't imagine why anyone would do this) */
1999 * In URI, "back up" to the last '/' encountered. This will be the
2000 * beginning of the "unique" suffix of URI
2003 if ((ref->path[ix] == '/') && (ix > 0))
2005 else if ((ref->path[ix] == 0) && (ix > 1) && (ref->path[ix - 1] == '/'))
2007 for (; ix > 0; ix--) {
2008 if (ref->path[ix] == '/')
2015 uptr = &ref->path[ix];
2019 * In base, count the number of '/' from the differing point
2021 if (bptr[pos] != ref->path[pos]) {/* check for trivial URI == base */
2022 for (; bptr[ix] != 0; ix++) {
2023 if (bptr[ix] == '/')
2027 len = strlen (uptr) + 1;
2032 /* exception characters from uri_to_string */
2033 val = uri_string_escape(uptr, "/;&=+$,");
2038 * Allocate just enough space for the returned string -
2039 * length of the remainder of the URI, plus enough space
2040 * for the "../" groups, plus one for the terminator
2042 val = g_malloc (len + 3 * nbslash);
2045 * Put in as many "../" as needed
2047 for (; nbslash>0; nbslash--) {
2053 * Finish up with the end of the URI
2056 if ((vptr > val) && (len > 0) &&
2057 (uptr[0] == '/') && (vptr[-1] == '/')) {
2058 memcpy (vptr, uptr + 1, len - 1);
2061 memcpy (vptr, uptr, len);
2068 /* escape the freshly-built path */
2070 /* exception characters from uri_to_string */
2071 val = uri_string_escape(vptr, "/;&=+$,");
2076 * Free the working variables
2078 if (remove_path != 0)
2089 * Utility functions to help parse and assemble query strings.
2092 struct QueryParams *
2093 query_params_new (int init_alloc)
2095 struct QueryParams *ps;
2097 if (init_alloc <= 0) init_alloc = 1;
2099 ps = g_new(QueryParams, 1);
2101 ps->alloc = init_alloc;
2102 ps->p = g_new(QueryParam, ps->alloc);
2107 /* Ensure there is space to store at least one more parameter
2108 * at the end of the set.
2111 query_params_append (struct QueryParams *ps,
2112 const char *name, const char *value)
2114 if (ps->n >= ps->alloc) {
2115 ps->p = g_renew(QueryParam, ps->p, ps->alloc * 2);
2119 ps->p[ps->n].name = g_strdup(name);
2120 ps->p[ps->n].value = g_strdup(value);
2121 ps->p[ps->n].ignore = 0;
2128 query_params_free (struct QueryParams *ps)
2132 for (i = 0; i < ps->n; ++i) {
2133 g_free (ps->p[i].name);
2134 g_free (ps->p[i].value);
2140 struct QueryParams *
2141 query_params_parse (const char *query)
2143 struct QueryParams *ps;
2144 const char *end, *eq;
2146 ps = query_params_new (0);
2147 if (!query || query[0] == '\0') return ps;
2150 char *name = NULL, *value = NULL;
2152 /* Find the next separator, or end of the string. */
2153 end = strchr (query, '&');
2155 end = strchr (query, ';');
2157 end = query + strlen (query);
2159 /* Find the first '=' character between here and end. */
2160 eq = strchr (query, '=');
2161 if (eq && eq >= end) eq = NULL;
2163 /* Empty section (eg. "&&"). */
2167 /* If there is no '=' character, then we have just "name"
2168 * and consistent with CGI.pm we assume value is "".
2171 name = uri_string_unescape (query, end - query, NULL);
2174 /* Or if we have "name=" here (works around annoying
2175 * problem when calling uri_string_unescape with len = 0).
2177 else if (eq+1 == end) {
2178 name = uri_string_unescape (query, eq - query, NULL);
2179 value = g_new0(char, 1);
2181 /* If the '=' character is at the beginning then we have
2182 * "=value" and consistent with CGI.pm we _ignore_ this.
2184 else if (query == eq)
2187 /* Otherwise it's "name=value". */
2189 name = uri_string_unescape (query, eq - query, NULL);
2190 value = uri_string_unescape (eq+1, end - (eq+1), NULL);
2193 /* Append to the parameter set. */
2194 query_params_append (ps, name, value);
2200 if (*query) query ++; /* skip '&' separator */