block: Move request tracking to children in copy offloading
[qemu/ar7.git] / util / uri.c
blob8624a7ac23d9317a6a26abfcdb05a46bd3748c40
1 /**
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.
29 * daniel@veillard.com
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
49 * Authors:
50 * Richard W.M. Jones <rjones@redhat.com>
54 #include "qemu/osdep.h"
56 #include "qemu/uri.h"
58 static void uri_clean(URI *uri);
61 * Old rule from 2396 used in legacy handling code
62 * alpha = lowalpha | upalpha
64 #define IS_ALPHA(x) (IS_LOWALPHA(x) || IS_UPALPHA(x))
67 * lowalpha = "a" | "b" | "c" | "d" | "e" | "f" | "g" | "h" | "i" | "j" |
68 * "k" | "l" | "m" | "n" | "o" | "p" | "q" | "r" | "s" | "t" |
69 * "u" | "v" | "w" | "x" | "y" | "z"
72 #define IS_LOWALPHA(x) (((x) >= 'a') && ((x) <= 'z'))
75 * upalpha = "A" | "B" | "C" | "D" | "E" | "F" | "G" | "H" | "I" | "J" |
76 * "K" | "L" | "M" | "N" | "O" | "P" | "Q" | "R" | "S" | "T" |
77 * "U" | "V" | "W" | "X" | "Y" | "Z"
79 #define IS_UPALPHA(x) (((x) >= 'A') && ((x) <= 'Z'))
81 #ifdef IS_DIGIT
82 #undef IS_DIGIT
83 #endif
85 * digit = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9"
87 #define IS_DIGIT(x) (((x) >= '0') && ((x) <= '9'))
90 * alphanum = alpha | digit
93 #define IS_ALPHANUM(x) (IS_ALPHA(x) || IS_DIGIT(x))
96 * mark = "-" | "_" | "." | "!" | "~" | "*" | "'" | "(" | ")"
99 #define IS_MARK(x) (((x) == '-') || ((x) == '_') || ((x) == '.') || \
100 ((x) == '!') || ((x) == '~') || ((x) == '*') || ((x) == '\'') || \
101 ((x) == '(') || ((x) == ')'))
104 * unwise = "{" | "}" | "|" | "\" | "^" | "`"
107 #define IS_UNWISE(p) \
108 (((*(p) == '{')) || ((*(p) == '}')) || ((*(p) == '|')) || \
109 ((*(p) == '\\')) || ((*(p) == '^')) || ((*(p) == '[')) || \
110 ((*(p) == ']')) || ((*(p) == '`')))
112 * reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" | "$" | "," |
113 * "[" | "]"
116 #define IS_RESERVED(x) (((x) == ';') || ((x) == '/') || ((x) == '?') || \
117 ((x) == ':') || ((x) == '@') || ((x) == '&') || ((x) == '=') || \
118 ((x) == '+') || ((x) == '$') || ((x) == ',') || ((x) == '[') || \
119 ((x) == ']'))
122 * unreserved = alphanum | mark
125 #define IS_UNRESERVED(x) (IS_ALPHANUM(x) || IS_MARK(x))
128 * Skip to next pointer char, handle escaped sequences
131 #define NEXT(p) ((*p == '%') ? p += 3 : p++)
134 * Productions from the spec.
136 * authority = server | reg_name
137 * reg_name = 1*( unreserved | escaped | "$" | "," |
138 * ";" | ":" | "@" | "&" | "=" | "+" )
140 * path = [ abs_path | opaque_part ]
143 /************************************************************************
145 * RFC 3986 parser *
147 ************************************************************************/
149 #define ISA_DIGIT(p) ((*(p) >= '0') && (*(p) <= '9'))
150 #define ISA_ALPHA(p) (((*(p) >= 'a') && (*(p) <= 'z')) || \
151 ((*(p) >= 'A') && (*(p) <= 'Z')))
152 #define ISA_HEXDIG(p) \
153 (ISA_DIGIT(p) || ((*(p) >= 'a') && (*(p) <= 'f')) || \
154 ((*(p) >= 'A') && (*(p) <= 'F')))
157 * sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
158 * / "*" / "+" / "," / ";" / "="
160 #define ISA_SUB_DELIM(p) \
161 (((*(p) == '!')) || ((*(p) == '$')) || ((*(p) == '&')) || \
162 ((*(p) == '(')) || ((*(p) == ')')) || ((*(p) == '*')) || \
163 ((*(p) == '+')) || ((*(p) == ',')) || ((*(p) == ';')) || \
164 ((*(p) == '=')) || ((*(p) == '\'')))
167 * gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@"
169 #define ISA_GEN_DELIM(p) \
170 (((*(p) == ':')) || ((*(p) == '/')) || ((*(p) == '?')) || \
171 ((*(p) == '#')) || ((*(p) == '[')) || ((*(p) == ']')) || \
172 ((*(p) == '@')))
175 * reserved = gen-delims / sub-delims
177 #define ISA_RESERVED(p) (ISA_GEN_DELIM(p) || (ISA_SUB_DELIM(p)))
180 * unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
182 #define ISA_UNRESERVED(p) \
183 ((ISA_ALPHA(p)) || (ISA_DIGIT(p)) || ((*(p) == '-')) || \
184 ((*(p) == '.')) || ((*(p) == '_')) || ((*(p) == '~')))
187 * pct-encoded = "%" HEXDIG HEXDIG
189 #define ISA_PCT_ENCODED(p) \
190 ((*(p) == '%') && (ISA_HEXDIG(p + 1)) && (ISA_HEXDIG(p + 2)))
193 * pchar = unreserved / pct-encoded / sub-delims / ":" / "@"
195 #define ISA_PCHAR(p) \
196 (ISA_UNRESERVED(p) || ISA_PCT_ENCODED(p) || ISA_SUB_DELIM(p) || \
197 ((*(p) == ':')) || ((*(p) == '@')))
200 * rfc3986_parse_scheme:
201 * @uri: pointer to an URI structure
202 * @str: pointer to the string to analyze
204 * Parse an URI scheme
206 * ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
208 * Returns 0 or the error code
210 static int rfc3986_parse_scheme(URI *uri, const char **str)
212 const char *cur;
214 if (str == NULL) {
215 return -1;
218 cur = *str;
219 if (!ISA_ALPHA(cur)) {
220 return 2;
222 cur++;
223 while (ISA_ALPHA(cur) || ISA_DIGIT(cur) || (*cur == '+') || (*cur == '-') ||
224 (*cur == '.')) {
225 cur++;
227 if (uri != NULL) {
228 g_free(uri->scheme);
229 uri->scheme = g_strndup(*str, cur - *str);
231 *str = cur;
232 return 0;
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
250 static int rfc3986_parse_fragment(URI *uri, const char **str)
252 const char *cur;
254 if (str == NULL) {
255 return -1;
258 cur = *str;
260 while ((ISA_PCHAR(cur)) || (*cur == '/') || (*cur == '?') ||
261 (*cur == '[') || (*cur == ']') ||
262 ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur)))) {
263 NEXT(cur);
265 if (uri != NULL) {
266 g_free(uri->fragment);
267 if (uri->cleanup & 2) {
268 uri->fragment = g_strndup(*str, cur - *str);
269 } else {
270 uri->fragment = uri_string_unescape(*str, cur - *str, NULL);
273 *str = cur;
274 return 0;
278 * rfc3986_parse_query:
279 * @uri: pointer to an URI structure
280 * @str: pointer to the string to analyze
282 * Parse the query part of an URI
284 * query = *uric
286 * Returns 0 or the error code
288 static int rfc3986_parse_query(URI *uri, const char **str)
290 const char *cur;
292 if (str == NULL) {
293 return -1;
296 cur = *str;
298 while ((ISA_PCHAR(cur)) || (*cur == '/') || (*cur == '?') ||
299 ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur)))) {
300 NEXT(cur);
302 if (uri != NULL) {
303 g_free(uri->query);
304 uri->query = g_strndup(*str, cur - *str);
306 *str = cur;
307 return 0;
311 * rfc3986_parse_port:
312 * @uri: pointer to an URI structure
313 * @str: the string to analyze
315 * Parse a port part and fills in the appropriate fields
316 * of the @uri structure
318 * port = *DIGIT
320 * Returns 0 or the error code
322 static int rfc3986_parse_port(URI *uri, const char **str)
324 const char *cur = *str;
325 int port = 0;
327 if (ISA_DIGIT(cur)) {
328 while (ISA_DIGIT(cur)) {
329 port = port * 10 + (*cur - '0');
330 if (port > 65535) {
331 return 1;
333 cur++;
335 if (uri) {
336 uri->port = port;
338 *str = cur;
339 return 0;
341 return 1;
345 * rfc3986_parse_user_info:
346 * @uri: pointer to an URI structure
347 * @str: the string to analyze
349 * Parse a user information part and fill in the appropriate fields
350 * of the @uri structure
352 * userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
354 * Returns 0 or the error code
356 static int rfc3986_parse_user_info(URI *uri, const char **str)
358 const char *cur;
360 cur = *str;
361 while (ISA_UNRESERVED(cur) || ISA_PCT_ENCODED(cur) || ISA_SUB_DELIM(cur) ||
362 (*cur == ':')) {
363 NEXT(cur);
365 if (*cur == '@') {
366 if (uri != NULL) {
367 g_free(uri->user);
368 if (uri->cleanup & 2) {
369 uri->user = g_strndup(*str, cur - *str);
370 } else {
371 uri->user = uri_string_unescape(*str, cur - *str, NULL);
374 *str = cur;
375 return 0;
377 return 1;
381 * rfc3986_parse_dec_octet:
382 * @str: the string to analyze
384 * dec-octet = DIGIT ; 0-9
385 * / %x31-39 DIGIT ; 10-99
386 * / "1" 2DIGIT ; 100-199
387 * / "2" %x30-34 DIGIT ; 200-249
388 * / "25" %x30-35 ; 250-255
390 * Skip a dec-octet.
392 * Returns 0 if found and skipped, 1 otherwise
394 static int rfc3986_parse_dec_octet(const char **str)
396 const char *cur = *str;
398 if (!(ISA_DIGIT(cur))) {
399 return 1;
401 if (!ISA_DIGIT(cur + 1)) {
402 cur++;
403 } else if ((*cur != '0') && (ISA_DIGIT(cur + 1)) && (!ISA_DIGIT(cur + 2))) {
404 cur += 2;
405 } else if ((*cur == '1') && (ISA_DIGIT(cur + 1)) && (ISA_DIGIT(cur + 2))) {
406 cur += 3;
407 } else if ((*cur == '2') && (*(cur + 1) >= '0') && (*(cur + 1) <= '4') &&
408 (ISA_DIGIT(cur + 2))) {
409 cur += 3;
410 } else if ((*cur == '2') && (*(cur + 1) == '5') && (*(cur + 2) >= '0') &&
411 (*(cur + 1) <= '5')) {
412 cur += 3;
413 } else {
414 return 1;
416 *str = cur;
417 return 0;
420 * rfc3986_parse_host:
421 * @uri: pointer to an URI structure
422 * @str: the string to analyze
424 * Parse an host part and fills in the appropriate fields
425 * of the @uri structure
427 * host = IP-literal / IPv4address / reg-name
428 * IP-literal = "[" ( IPv6address / IPvFuture ) "]"
429 * IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
430 * reg-name = *( unreserved / pct-encoded / sub-delims )
432 * Returns 0 or the error code
434 static int rfc3986_parse_host(URI *uri, const char **str)
436 const char *cur = *str;
437 const char *host;
439 host = cur;
441 * IPv6 and future addressing scheme are enclosed between brackets
443 if (*cur == '[') {
444 cur++;
445 while ((*cur != ']') && (*cur != 0)) {
446 cur++;
448 if (*cur != ']') {
449 return 1;
451 cur++;
452 goto found;
455 * try to parse an IPv4
457 if (ISA_DIGIT(cur)) {
458 if (rfc3986_parse_dec_octet(&cur) != 0) {
459 goto not_ipv4;
461 if (*cur != '.') {
462 goto not_ipv4;
464 cur++;
465 if (rfc3986_parse_dec_octet(&cur) != 0) {
466 goto not_ipv4;
468 if (*cur != '.') {
469 goto not_ipv4;
471 if (rfc3986_parse_dec_octet(&cur) != 0) {
472 goto not_ipv4;
474 if (*cur != '.') {
475 goto not_ipv4;
477 if (rfc3986_parse_dec_octet(&cur) != 0) {
478 goto not_ipv4;
480 goto found;
481 not_ipv4:
482 cur = *str;
485 * then this should be a hostname which can be empty
487 while (ISA_UNRESERVED(cur) || ISA_PCT_ENCODED(cur) || ISA_SUB_DELIM(cur)) {
488 NEXT(cur);
490 found:
491 if (uri != NULL) {
492 g_free(uri->authority);
493 uri->authority = NULL;
494 g_free(uri->server);
495 if (cur != host) {
496 if (uri->cleanup & 2) {
497 uri->server = g_strndup(host, cur - host);
498 } else {
499 uri->server = uri_string_unescape(host, cur - host, NULL);
501 } else {
502 uri->server = NULL;
505 *str = cur;
506 return 0;
510 * rfc3986_parse_authority:
511 * @uri: pointer to an URI structure
512 * @str: the string to analyze
514 * Parse an authority part and fills in the appropriate fields
515 * of the @uri structure
517 * authority = [ userinfo "@" ] host [ ":" port ]
519 * Returns 0 or the error code
521 static int rfc3986_parse_authority(URI *uri, const char **str)
523 const char *cur;
524 int ret;
526 cur = *str;
528 * try to parse a userinfo and check for the trailing @
530 ret = rfc3986_parse_user_info(uri, &cur);
531 if ((ret != 0) || (*cur != '@')) {
532 cur = *str;
533 } else {
534 cur++;
536 ret = rfc3986_parse_host(uri, &cur);
537 if (ret != 0) {
538 return ret;
540 if (*cur == ':') {
541 cur++;
542 ret = rfc3986_parse_port(uri, &cur);
543 if (ret != 0) {
544 return ret;
547 *str = cur;
548 return 0;
552 * rfc3986_parse_segment:
553 * @str: the string to analyze
554 * @forbid: an optional forbidden character
555 * @empty: allow an empty segment
557 * Parse a segment and fills in the appropriate fields
558 * of the @uri structure
560 * segment = *pchar
561 * segment-nz = 1*pchar
562 * segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" )
563 * ; non-zero-length segment without any colon ":"
565 * Returns 0 or the error code
567 static int rfc3986_parse_segment(const char **str, char forbid, int empty)
569 const char *cur;
571 cur = *str;
572 if (!ISA_PCHAR(cur)) {
573 if (empty) {
574 return 0;
576 return 1;
578 while (ISA_PCHAR(cur) && (*cur != forbid)) {
579 NEXT(cur);
581 *str = cur;
582 return 0;
586 * rfc3986_parse_path_ab_empty:
587 * @uri: pointer to an URI structure
588 * @str: the string to analyze
590 * Parse an path absolute or empty and fills in the appropriate fields
591 * of the @uri structure
593 * path-abempty = *( "/" segment )
595 * Returns 0 or the error code
597 static int rfc3986_parse_path_ab_empty(URI *uri, const char **str)
599 const char *cur;
600 int ret;
602 cur = *str;
604 while (*cur == '/') {
605 cur++;
606 ret = rfc3986_parse_segment(&cur, 0, 1);
607 if (ret != 0) {
608 return ret;
611 if (uri != NULL) {
612 g_free(uri->path);
613 if (*str != cur) {
614 if (uri->cleanup & 2) {
615 uri->path = g_strndup(*str, cur - *str);
616 } else {
617 uri->path = uri_string_unescape(*str, cur - *str, NULL);
619 } else {
620 uri->path = NULL;
623 *str = cur;
624 return 0;
628 * rfc3986_parse_path_absolute:
629 * @uri: pointer to an URI structure
630 * @str: the string to analyze
632 * Parse an path absolute and fills in the appropriate fields
633 * of the @uri structure
635 * path-absolute = "/" [ segment-nz *( "/" segment ) ]
637 * Returns 0 or the error code
639 static int rfc3986_parse_path_absolute(URI *uri, const char **str)
641 const char *cur;
642 int ret;
644 cur = *str;
646 if (*cur != '/') {
647 return 1;
649 cur++;
650 ret = rfc3986_parse_segment(&cur, 0, 0);
651 if (ret == 0) {
652 while (*cur == '/') {
653 cur++;
654 ret = rfc3986_parse_segment(&cur, 0, 1);
655 if (ret != 0) {
656 return ret;
660 if (uri != NULL) {
661 g_free(uri->path);
662 if (cur != *str) {
663 if (uri->cleanup & 2) {
664 uri->path = g_strndup(*str, cur - *str);
665 } else {
666 uri->path = uri_string_unescape(*str, cur - *str, NULL);
668 } else {
669 uri->path = NULL;
672 *str = cur;
673 return 0;
677 * rfc3986_parse_path_rootless:
678 * @uri: pointer to an URI structure
679 * @str: the string to analyze
681 * Parse an path without root and fills in the appropriate fields
682 * of the @uri structure
684 * path-rootless = segment-nz *( "/" segment )
686 * Returns 0 or the error code
688 static int rfc3986_parse_path_rootless(URI *uri, const char **str)
690 const char *cur;
691 int ret;
693 cur = *str;
695 ret = rfc3986_parse_segment(&cur, 0, 0);
696 if (ret != 0) {
697 return ret;
699 while (*cur == '/') {
700 cur++;
701 ret = rfc3986_parse_segment(&cur, 0, 1);
702 if (ret != 0) {
703 return ret;
706 if (uri != NULL) {
707 g_free(uri->path);
708 if (cur != *str) {
709 if (uri->cleanup & 2) {
710 uri->path = g_strndup(*str, cur - *str);
711 } else {
712 uri->path = uri_string_unescape(*str, cur - *str, NULL);
714 } else {
715 uri->path = NULL;
718 *str = cur;
719 return 0;
723 * rfc3986_parse_path_no_scheme:
724 * @uri: pointer to an URI structure
725 * @str: the string to analyze
727 * Parse an path which is not a scheme and fills in the appropriate fields
728 * of the @uri structure
730 * path-noscheme = segment-nz-nc *( "/" segment )
732 * Returns 0 or the error code
734 static int rfc3986_parse_path_no_scheme(URI *uri, const char **str)
736 const char *cur;
737 int ret;
739 cur = *str;
741 ret = rfc3986_parse_segment(&cur, ':', 0);
742 if (ret != 0) {
743 return ret;
745 while (*cur == '/') {
746 cur++;
747 ret = rfc3986_parse_segment(&cur, 0, 1);
748 if (ret != 0) {
749 return ret;
752 if (uri != NULL) {
753 g_free(uri->path);
754 if (cur != *str) {
755 if (uri->cleanup & 2) {
756 uri->path = g_strndup(*str, cur - *str);
757 } else {
758 uri->path = uri_string_unescape(*str, cur - *str, NULL);
760 } else {
761 uri->path = NULL;
764 *str = cur;
765 return 0;
769 * rfc3986_parse_hier_part:
770 * @uri: pointer to an URI structure
771 * @str: the string to analyze
773 * Parse an hierarchical part and fills in the appropriate fields
774 * of the @uri structure
776 * hier-part = "//" authority path-abempty
777 * / path-absolute
778 * / path-rootless
779 * / path-empty
781 * Returns 0 or the error code
783 static int rfc3986_parse_hier_part(URI *uri, const char **str)
785 const char *cur;
786 int ret;
788 cur = *str;
790 if ((*cur == '/') && (*(cur + 1) == '/')) {
791 cur += 2;
792 ret = rfc3986_parse_authority(uri, &cur);
793 if (ret != 0) {
794 return ret;
796 ret = rfc3986_parse_path_ab_empty(uri, &cur);
797 if (ret != 0) {
798 return ret;
800 *str = cur;
801 return 0;
802 } else if (*cur == '/') {
803 ret = rfc3986_parse_path_absolute(uri, &cur);
804 if (ret != 0) {
805 return ret;
807 } else if (ISA_PCHAR(cur)) {
808 ret = rfc3986_parse_path_rootless(uri, &cur);
809 if (ret != 0) {
810 return ret;
812 } else {
813 /* path-empty is effectively empty */
814 if (uri != NULL) {
815 g_free(uri->path);
816 uri->path = NULL;
819 *str = cur;
820 return 0;
824 * rfc3986_parse_relative_ref:
825 * @uri: pointer to an URI structure
826 * @str: the string to analyze
828 * Parse an URI string and fills in the appropriate fields
829 * of the @uri structure
831 * relative-ref = relative-part [ "?" query ] [ "#" fragment ]
832 * relative-part = "//" authority path-abempty
833 * / path-absolute
834 * / path-noscheme
835 * / path-empty
837 * Returns 0 or the error code
839 static int rfc3986_parse_relative_ref(URI *uri, const char *str)
841 int ret;
843 if ((*str == '/') && (*(str + 1) == '/')) {
844 str += 2;
845 ret = rfc3986_parse_authority(uri, &str);
846 if (ret != 0) {
847 return ret;
849 ret = rfc3986_parse_path_ab_empty(uri, &str);
850 if (ret != 0) {
851 return ret;
853 } else if (*str == '/') {
854 ret = rfc3986_parse_path_absolute(uri, &str);
855 if (ret != 0) {
856 return ret;
858 } else if (ISA_PCHAR(str)) {
859 ret = rfc3986_parse_path_no_scheme(uri, &str);
860 if (ret != 0) {
861 return ret;
863 } else {
864 /* path-empty is effectively empty */
865 if (uri != NULL) {
866 g_free(uri->path);
867 uri->path = NULL;
871 if (*str == '?') {
872 str++;
873 ret = rfc3986_parse_query(uri, &str);
874 if (ret != 0) {
875 return ret;
878 if (*str == '#') {
879 str++;
880 ret = rfc3986_parse_fragment(uri, &str);
881 if (ret != 0) {
882 return ret;
885 if (*str != 0) {
886 uri_clean(uri);
887 return 1;
889 return 0;
893 * rfc3986_parse:
894 * @uri: pointer to an URI structure
895 * @str: the string to analyze
897 * Parse an URI string and fills in the appropriate fields
898 * of the @uri structure
900 * scheme ":" hier-part [ "?" query ] [ "#" fragment ]
902 * Returns 0 or the error code
904 static int rfc3986_parse(URI *uri, const char *str)
906 int ret;
908 ret = rfc3986_parse_scheme(uri, &str);
909 if (ret != 0) {
910 return ret;
912 if (*str != ':') {
913 return 1;
915 str++;
916 ret = rfc3986_parse_hier_part(uri, &str);
917 if (ret != 0) {
918 return ret;
920 if (*str == '?') {
921 str++;
922 ret = rfc3986_parse_query(uri, &str);
923 if (ret != 0) {
924 return ret;
927 if (*str == '#') {
928 str++;
929 ret = rfc3986_parse_fragment(uri, &str);
930 if (ret != 0) {
931 return ret;
934 if (*str != 0) {
935 uri_clean(uri);
936 return 1;
938 return 0;
942 * rfc3986_parse_uri_reference:
943 * @uri: pointer to an URI structure
944 * @str: the string to analyze
946 * Parse an URI reference string and fills in the appropriate fields
947 * of the @uri structure
949 * URI-reference = URI / relative-ref
951 * Returns 0 or the error code
953 static int rfc3986_parse_uri_reference(URI *uri, const char *str)
955 int ret;
957 if (str == NULL) {
958 return -1;
960 uri_clean(uri);
963 * Try first to parse absolute refs, then fallback to relative if
964 * it fails.
966 ret = rfc3986_parse(uri, str);
967 if (ret != 0) {
968 uri_clean(uri);
969 ret = rfc3986_parse_relative_ref(uri, str);
970 if (ret != 0) {
971 uri_clean(uri);
972 return ret;
975 return 0;
979 * uri_parse:
980 * @str: the URI string to analyze
982 * Parse an URI based on RFC 3986
984 * URI-reference = [ absoluteURI | relativeURI ] [ "#" fragment ]
986 * Returns a newly built URI or NULL in case of error
988 URI *uri_parse(const char *str)
990 URI *uri;
991 int ret;
993 if (str == NULL) {
994 return NULL;
996 uri = uri_new();
997 ret = rfc3986_parse_uri_reference(uri, str);
998 if (ret) {
999 uri_free(uri);
1000 return NULL;
1002 return uri;
1006 * uri_parse_into:
1007 * @uri: pointer to an URI structure
1008 * @str: the string to analyze
1010 * Parse an URI reference string based on RFC 3986 and fills in the
1011 * appropriate fields of the @uri structure
1013 * URI-reference = URI / relative-ref
1015 * Returns 0 or the error code
1017 int uri_parse_into(URI *uri, const char *str)
1019 return rfc3986_parse_uri_reference(uri, str);
1023 * uri_parse_raw:
1024 * @str: the URI string to analyze
1025 * @raw: if 1 unescaping of URI pieces are disabled
1027 * Parse an URI but allows to keep intact the original fragments.
1029 * URI-reference = URI / relative-ref
1031 * Returns a newly built URI or NULL in case of error
1033 URI *uri_parse_raw(const char *str, int raw)
1035 URI *uri;
1036 int ret;
1038 if (str == NULL) {
1039 return NULL;
1041 uri = uri_new();
1042 if (raw) {
1043 uri->cleanup |= 2;
1045 ret = uri_parse_into(uri, str);
1046 if (ret) {
1047 uri_free(uri);
1048 return NULL;
1050 return uri;
1053 /************************************************************************
1055 * Generic URI structure functions *
1057 ************************************************************************/
1060 * uri_new:
1062 * Simply creates an empty URI
1064 * Returns the new structure or NULL in case of error
1066 URI *uri_new(void)
1068 return g_new0(URI, 1);
1072 * realloc2n:
1074 * Function to handle properly a reallocation when saving an URI
1075 * Also imposes some limit on the length of an URI string output
1077 static char *realloc2n(char *ret, int *max)
1079 char *temp;
1080 int tmp;
1082 tmp = *max * 2;
1083 temp = g_realloc(ret, (tmp + 1));
1084 *max = tmp;
1085 return temp;
1089 * uri_to_string:
1090 * @uri: pointer to an URI
1092 * Save the URI as an escaped string
1094 * Returns a new string (to be deallocated by caller)
1096 char *uri_to_string(URI *uri)
1098 char *ret = NULL;
1099 char *temp;
1100 const char *p;
1101 int len;
1102 int max;
1104 if (uri == NULL) {
1105 return NULL;
1108 max = 80;
1109 ret = g_malloc(max + 1);
1110 len = 0;
1112 if (uri->scheme != NULL) {
1113 p = uri->scheme;
1114 while (*p != 0) {
1115 if (len >= max) {
1116 temp = realloc2n(ret, &max);
1117 ret = temp;
1119 ret[len++] = *p++;
1121 if (len >= max) {
1122 temp = realloc2n(ret, &max);
1123 ret = temp;
1125 ret[len++] = ':';
1127 if (uri->opaque != NULL) {
1128 p = uri->opaque;
1129 while (*p != 0) {
1130 if (len + 3 >= max) {
1131 temp = realloc2n(ret, &max);
1132 ret = temp;
1134 if (IS_RESERVED(*(p)) || IS_UNRESERVED(*(p))) {
1135 ret[len++] = *p++;
1136 } else {
1137 int val = *(unsigned char *)p++;
1138 int hi = val / 0x10, lo = val % 0x10;
1139 ret[len++] = '%';
1140 ret[len++] = hi + (hi > 9 ? 'A' - 10 : '0');
1141 ret[len++] = lo + (lo > 9 ? 'A' - 10 : '0');
1144 } else {
1145 if (uri->server != NULL) {
1146 if (len + 3 >= max) {
1147 temp = realloc2n(ret, &max);
1148 ret = temp;
1150 ret[len++] = '/';
1151 ret[len++] = '/';
1152 if (uri->user != NULL) {
1153 p = uri->user;
1154 while (*p != 0) {
1155 if (len + 3 >= max) {
1156 temp = realloc2n(ret, &max);
1157 ret = temp;
1159 if ((IS_UNRESERVED(*(p))) || ((*(p) == ';')) ||
1160 ((*(p) == ':')) || ((*(p) == '&')) || ((*(p) == '=')) ||
1161 ((*(p) == '+')) || ((*(p) == '$')) || ((*(p) == ','))) {
1162 ret[len++] = *p++;
1163 } else {
1164 int val = *(unsigned char *)p++;
1165 int hi = val / 0x10, lo = val % 0x10;
1166 ret[len++] = '%';
1167 ret[len++] = hi + (hi > 9 ? 'A' - 10 : '0');
1168 ret[len++] = lo + (lo > 9 ? 'A' - 10 : '0');
1171 if (len + 3 >= max) {
1172 temp = realloc2n(ret, &max);
1173 ret = temp;
1175 ret[len++] = '@';
1177 p = uri->server;
1178 while (*p != 0) {
1179 if (len >= max) {
1180 temp = realloc2n(ret, &max);
1181 ret = temp;
1183 ret[len++] = *p++;
1185 if (uri->port > 0) {
1186 if (len + 10 >= max) {
1187 temp = realloc2n(ret, &max);
1188 ret = temp;
1190 len += snprintf(&ret[len], max - len, ":%d", uri->port);
1192 } else if (uri->authority != NULL) {
1193 if (len + 3 >= max) {
1194 temp = realloc2n(ret, &max);
1195 ret = temp;
1197 ret[len++] = '/';
1198 ret[len++] = '/';
1199 p = uri->authority;
1200 while (*p != 0) {
1201 if (len + 3 >= max) {
1202 temp = realloc2n(ret, &max);
1203 ret = temp;
1205 if ((IS_UNRESERVED(*(p))) || ((*(p) == '$')) ||
1206 ((*(p) == ',')) || ((*(p) == ';')) || ((*(p) == ':')) ||
1207 ((*(p) == '@')) || ((*(p) == '&')) || ((*(p) == '=')) ||
1208 ((*(p) == '+'))) {
1209 ret[len++] = *p++;
1210 } else {
1211 int val = *(unsigned char *)p++;
1212 int hi = val / 0x10, lo = val % 0x10;
1213 ret[len++] = '%';
1214 ret[len++] = hi + (hi > 9 ? 'A' - 10 : '0');
1215 ret[len++] = lo + (lo > 9 ? 'A' - 10 : '0');
1218 } else if (uri->scheme != NULL) {
1219 if (len + 3 >= max) {
1220 temp = realloc2n(ret, &max);
1221 ret = temp;
1223 ret[len++] = '/';
1224 ret[len++] = '/';
1226 if (uri->path != NULL) {
1227 p = uri->path;
1229 * the colon in file:///d: should not be escaped or
1230 * Windows accesses fail later.
1232 if ((uri->scheme != NULL) && (p[0] == '/') &&
1233 (((p[1] >= 'a') && (p[1] <= 'z')) ||
1234 ((p[1] >= 'A') && (p[1] <= 'Z'))) &&
1235 (p[2] == ':') && (!strcmp(uri->scheme, "file"))) {
1236 if (len + 3 >= max) {
1237 temp = realloc2n(ret, &max);
1238 ret = temp;
1240 ret[len++] = *p++;
1241 ret[len++] = *p++;
1242 ret[len++] = *p++;
1244 while (*p != 0) {
1245 if (len + 3 >= max) {
1246 temp = realloc2n(ret, &max);
1247 ret = temp;
1249 if ((IS_UNRESERVED(*(p))) || ((*(p) == '/')) ||
1250 ((*(p) == ';')) || ((*(p) == '@')) || ((*(p) == '&')) ||
1251 ((*(p) == '=')) || ((*(p) == '+')) || ((*(p) == '$')) ||
1252 ((*(p) == ','))) {
1253 ret[len++] = *p++;
1254 } else {
1255 int val = *(unsigned char *)p++;
1256 int hi = val / 0x10, lo = val % 0x10;
1257 ret[len++] = '%';
1258 ret[len++] = hi + (hi > 9 ? 'A' - 10 : '0');
1259 ret[len++] = lo + (lo > 9 ? 'A' - 10 : '0');
1263 if (uri->query != NULL) {
1264 if (len + 1 >= max) {
1265 temp = realloc2n(ret, &max);
1266 ret = temp;
1268 ret[len++] = '?';
1269 p = uri->query;
1270 while (*p != 0) {
1271 if (len + 1 >= max) {
1272 temp = realloc2n(ret, &max);
1273 ret = temp;
1275 ret[len++] = *p++;
1279 if (uri->fragment != NULL) {
1280 if (len + 3 >= max) {
1281 temp = realloc2n(ret, &max);
1282 ret = temp;
1284 ret[len++] = '#';
1285 p = uri->fragment;
1286 while (*p != 0) {
1287 if (len + 3 >= max) {
1288 temp = realloc2n(ret, &max);
1289 ret = temp;
1291 if ((IS_UNRESERVED(*(p))) || (IS_RESERVED(*(p)))) {
1292 ret[len++] = *p++;
1293 } else {
1294 int val = *(unsigned char *)p++;
1295 int hi = val / 0x10, lo = val % 0x10;
1296 ret[len++] = '%';
1297 ret[len++] = hi + (hi > 9 ? 'A' - 10 : '0');
1298 ret[len++] = lo + (lo > 9 ? 'A' - 10 : '0');
1302 if (len >= max) {
1303 temp = realloc2n(ret, &max);
1304 ret = temp;
1306 ret[len] = 0;
1307 return ret;
1311 * uri_clean:
1312 * @uri: pointer to an URI
1314 * Make sure the URI struct is free of content
1316 static void uri_clean(URI *uri)
1318 if (uri == NULL) {
1319 return;
1322 g_free(uri->scheme);
1323 uri->scheme = NULL;
1324 g_free(uri->server);
1325 uri->server = NULL;
1326 g_free(uri->user);
1327 uri->user = NULL;
1328 g_free(uri->path);
1329 uri->path = NULL;
1330 g_free(uri->fragment);
1331 uri->fragment = NULL;
1332 g_free(uri->opaque);
1333 uri->opaque = NULL;
1334 g_free(uri->authority);
1335 uri->authority = NULL;
1336 g_free(uri->query);
1337 uri->query = NULL;
1341 * uri_free:
1342 * @uri: pointer to an URI
1344 * Free up the URI struct
1346 void uri_free(URI *uri)
1348 uri_clean(uri);
1349 g_free(uri);
1352 /************************************************************************
1354 * Helper functions *
1356 ************************************************************************/
1359 * normalize_uri_path:
1360 * @path: pointer to the path string
1362 * Applies the 5 normalization steps to a path string--that is, RFC 2396
1363 * Section 5.2, steps 6.c through 6.g.
1365 * Normalization occurs directly on the string, no new allocation is done
1367 * Returns 0 or an error code
1369 static int normalize_uri_path(char *path)
1371 char *cur, *out;
1373 if (path == NULL) {
1374 return -1;
1377 /* Skip all initial "/" chars. We want to get to the beginning of the
1378 * first non-empty segment.
1380 cur = path;
1381 while (cur[0] == '/') {
1382 ++cur;
1384 if (cur[0] == '\0') {
1385 return 0;
1388 /* Keep everything we've seen so far. */
1389 out = cur;
1392 * Analyze each segment in sequence for cases (c) and (d).
1394 while (cur[0] != '\0') {
1396 * c) All occurrences of "./", where "." is a complete path segment,
1397 * are removed from the buffer string.
1399 if ((cur[0] == '.') && (cur[1] == '/')) {
1400 cur += 2;
1401 /* '//' normalization should be done at this point too */
1402 while (cur[0] == '/') {
1403 cur++;
1405 continue;
1409 * d) If the buffer string ends with "." as a complete path segment,
1410 * that "." is removed.
1412 if ((cur[0] == '.') && (cur[1] == '\0')) {
1413 break;
1416 /* Otherwise keep the segment. */
1417 while (cur[0] != '/') {
1418 if (cur[0] == '\0') {
1419 goto done_cd;
1421 (out++)[0] = (cur++)[0];
1423 /* nomalize // */
1424 while ((cur[0] == '/') && (cur[1] == '/')) {
1425 cur++;
1428 (out++)[0] = (cur++)[0];
1430 done_cd:
1431 out[0] = '\0';
1433 /* Reset to the beginning of the first segment for the next sequence. */
1434 cur = path;
1435 while (cur[0] == '/') {
1436 ++cur;
1438 if (cur[0] == '\0') {
1439 return 0;
1443 * Analyze each segment in sequence for cases (e) and (f).
1445 * e) All occurrences of "<segment>/../", where <segment> is a
1446 * complete path segment not equal to "..", are removed from the
1447 * buffer string. Removal of these path segments is performed
1448 * iteratively, removing the leftmost matching pattern on each
1449 * iteration, until no matching pattern remains.
1451 * f) If the buffer string ends with "<segment>/..", where <segment>
1452 * is a complete path segment not equal to "..", that
1453 * "<segment>/.." is removed.
1455 * To satisfy the "iterative" clause in (e), we need to collapse the
1456 * string every time we find something that needs to be removed. Thus,
1457 * we don't need to keep two pointers into the string: we only need a
1458 * "current position" pointer.
1460 while (1) {
1461 char *segp, *tmp;
1463 /* At the beginning of each iteration of this loop, "cur" points to
1464 * the first character of the segment we want to examine.
1467 /* Find the end of the current segment. */
1468 segp = cur;
1469 while ((segp[0] != '/') && (segp[0] != '\0')) {
1470 ++segp;
1473 /* If this is the last segment, we're done (we need at least two
1474 * segments to meet the criteria for the (e) and (f) cases).
1476 if (segp[0] == '\0') {
1477 break;
1480 /* If the first segment is "..", or if the next segment _isn't_ "..",
1481 * keep this segment and try the next one.
1483 ++segp;
1484 if (((cur[0] == '.') && (cur[1] == '.') && (segp == cur + 3)) ||
1485 ((segp[0] != '.') || (segp[1] != '.') ||
1486 ((segp[2] != '/') && (segp[2] != '\0')))) {
1487 cur = segp;
1488 continue;
1491 /* If we get here, remove this segment and the next one and back up
1492 * to the previous segment (if there is one), to implement the
1493 * "iteratively" clause. It's pretty much impossible to back up
1494 * while maintaining two pointers into the buffer, so just compact
1495 * the whole buffer now.
1498 /* If this is the end of the buffer, we're done. */
1499 if (segp[2] == '\0') {
1500 cur[0] = '\0';
1501 break;
1503 /* Valgrind complained, strcpy(cur, segp + 3); */
1504 /* string will overlap, do not use strcpy */
1505 tmp = cur;
1506 segp += 3;
1507 while ((*tmp++ = *segp++) != 0) {
1508 /* No further work */
1511 /* If there are no previous segments, then keep going from here. */
1512 segp = cur;
1513 while ((segp > path) && ((--segp)[0] == '/')) {
1514 /* No further work */
1516 if (segp == path) {
1517 continue;
1520 /* "segp" is pointing to the end of a previous segment; find it's
1521 * start. We need to back up to the previous segment and start
1522 * over with that to handle things like "foo/bar/../..". If we
1523 * don't do this, then on the first pass we'll remove the "bar/..",
1524 * but be pointing at the second ".." so we won't realize we can also
1525 * remove the "foo/..".
1527 cur = segp;
1528 while ((cur > path) && (cur[-1] != '/')) {
1529 --cur;
1532 out[0] = '\0';
1535 * g) If the resulting buffer string still begins with one or more
1536 * complete path segments of "..", then the reference is
1537 * considered to be in error. Implementations may handle this
1538 * error by retaining these components in the resolved path (i.e.,
1539 * treating them as part of the final URI), by removing them from
1540 * the resolved path (i.e., discarding relative levels above the
1541 * root), or by avoiding traversal of the reference.
1543 * We discard them from the final path.
1545 if (path[0] == '/') {
1546 cur = path;
1547 while ((cur[0] == '/') && (cur[1] == '.') && (cur[2] == '.') &&
1548 ((cur[3] == '/') || (cur[3] == '\0'))) {
1549 cur += 3;
1552 if (cur != path) {
1553 out = path;
1554 while (cur[0] != '\0') {
1555 (out++)[0] = (cur++)[0];
1557 out[0] = 0;
1561 return 0;
1564 static int is_hex(char c)
1566 if (((c >= '0') && (c <= '9')) || ((c >= 'a') && (c <= 'f')) ||
1567 ((c >= 'A') && (c <= 'F'))) {
1568 return 1;
1570 return 0;
1574 * uri_string_unescape:
1575 * @str: the string to unescape
1576 * @len: the length in bytes to unescape (or <= 0 to indicate full string)
1577 * @target: optional destination buffer
1579 * Unescaping routine, but does not check that the string is an URI. The
1580 * output is a direct unsigned char translation of %XX values (no encoding)
1581 * Note that the length of the result can only be smaller or same size as
1582 * the input string.
1584 * Returns a copy of the string, but unescaped, will return NULL only in case
1585 * of error
1587 char *uri_string_unescape(const char *str, int len, char *target)
1589 char *ret, *out;
1590 const char *in;
1592 if (str == NULL) {
1593 return NULL;
1595 if (len <= 0) {
1596 len = strlen(str);
1598 if (len < 0) {
1599 return NULL;
1602 if (target == NULL) {
1603 ret = g_malloc(len + 1);
1604 } else {
1605 ret = target;
1607 in = str;
1608 out = ret;
1609 while (len > 0) {
1610 if ((len > 2) && (*in == '%') && (is_hex(in[1])) && (is_hex(in[2]))) {
1611 in++;
1612 if ((*in >= '0') && (*in <= '9')) {
1613 *out = (*in - '0');
1614 } else if ((*in >= 'a') && (*in <= 'f')) {
1615 *out = (*in - 'a') + 10;
1616 } else if ((*in >= 'A') && (*in <= 'F')) {
1617 *out = (*in - 'A') + 10;
1619 in++;
1620 if ((*in >= '0') && (*in <= '9')) {
1621 *out = *out * 16 + (*in - '0');
1622 } else if ((*in >= 'a') && (*in <= 'f')) {
1623 *out = *out * 16 + (*in - 'a') + 10;
1624 } else if ((*in >= 'A') && (*in <= 'F')) {
1625 *out = *out * 16 + (*in - 'A') + 10;
1627 in++;
1628 len -= 3;
1629 out++;
1630 } else {
1631 *out++ = *in++;
1632 len--;
1635 *out = 0;
1636 return ret;
1640 * uri_string_escape:
1641 * @str: string to escape
1642 * @list: exception list string of chars not to escape
1644 * This routine escapes a string to hex, ignoring reserved characters (a-z)
1645 * and the characters in the exception list.
1647 * Returns a new escaped string or NULL in case of error.
1649 char *uri_string_escape(const char *str, const char *list)
1651 char *ret, ch;
1652 char *temp;
1653 const char *in;
1654 int len, out;
1656 if (str == NULL) {
1657 return NULL;
1659 if (str[0] == 0) {
1660 return g_strdup(str);
1662 len = strlen(str);
1663 if (!(len > 0)) {
1664 return NULL;
1667 len += 20;
1668 ret = g_malloc(len);
1669 in = str;
1670 out = 0;
1671 while (*in != 0) {
1672 if (len - out <= 3) {
1673 temp = realloc2n(ret, &len);
1674 ret = temp;
1677 ch = *in;
1679 if ((ch != '@') && (!IS_UNRESERVED(ch)) && (!strchr(list, ch))) {
1680 unsigned char val;
1681 ret[out++] = '%';
1682 val = ch >> 4;
1683 if (val <= 9) {
1684 ret[out++] = '0' + val;
1685 } else {
1686 ret[out++] = 'A' + val - 0xA;
1688 val = ch & 0xF;
1689 if (val <= 9) {
1690 ret[out++] = '0' + val;
1691 } else {
1692 ret[out++] = 'A' + val - 0xA;
1694 in++;
1695 } else {
1696 ret[out++] = *in++;
1699 ret[out] = 0;
1700 return ret;
1703 /************************************************************************
1705 * Public functions *
1707 ************************************************************************/
1710 * uri_resolve:
1711 * @URI: the URI instance found in the document
1712 * @base: the base value
1714 * Computes he final URI of the reference done by checking that
1715 * the given URI is valid, and building the final URI using the
1716 * base URI. This is processed according to section 5.2 of the
1717 * RFC 2396
1719 * 5.2. Resolving Relative References to Absolute Form
1721 * Returns a new URI string (to be freed by the caller) or NULL in case
1722 * of error.
1724 char *uri_resolve(const char *uri, const char *base)
1726 char *val = NULL;
1727 int ret, len, indx, cur, out;
1728 URI *ref = NULL;
1729 URI *bas = NULL;
1730 URI *res = NULL;
1733 * 1) The URI reference is parsed into the potential four components and
1734 * fragment identifier, as described in Section 4.3.
1736 * NOTE that a completely empty URI is treated by modern browsers
1737 * as a reference to "." rather than as a synonym for the current
1738 * URI. Should we do that here?
1740 if (uri == NULL) {
1741 ret = -1;
1742 } else {
1743 if (*uri) {
1744 ref = uri_new();
1745 ret = uri_parse_into(ref, uri);
1746 } else {
1747 ret = 0;
1750 if (ret != 0) {
1751 goto done;
1753 if ((ref != NULL) && (ref->scheme != NULL)) {
1755 * The URI is absolute don't modify.
1757 val = g_strdup(uri);
1758 goto done;
1760 if (base == NULL) {
1761 ret = -1;
1762 } else {
1763 bas = uri_new();
1764 ret = uri_parse_into(bas, base);
1766 if (ret != 0) {
1767 if (ref) {
1768 val = uri_to_string(ref);
1770 goto done;
1772 if (ref == NULL) {
1774 * the base fragment must be ignored
1776 g_free(bas->fragment);
1777 bas->fragment = NULL;
1778 val = uri_to_string(bas);
1779 goto done;
1783 * 2) If the path component is empty and the scheme, authority, and
1784 * query components are undefined, then it is a reference to the
1785 * current document and we are done. Otherwise, the reference URI's
1786 * query and fragment components are defined as found (or not found)
1787 * within the URI reference and not inherited from the base URI.
1789 * NOTE that in modern browsers, the parsing differs from the above
1790 * in the following aspect: the query component is allowed to be
1791 * defined while still treating this as a reference to the current
1792 * document.
1794 res = uri_new();
1795 if ((ref->scheme == NULL) && (ref->path == NULL) &&
1796 ((ref->authority == NULL) && (ref->server == NULL))) {
1797 res->scheme = g_strdup(bas->scheme);
1798 if (bas->authority != NULL) {
1799 res->authority = g_strdup(bas->authority);
1800 } else if (bas->server != NULL) {
1801 res->server = g_strdup(bas->server);
1802 res->user = g_strdup(bas->user);
1803 res->port = bas->port;
1805 res->path = g_strdup(bas->path);
1806 if (ref->query != NULL) {
1807 res->query = g_strdup(ref->query);
1808 } else {
1809 res->query = g_strdup(bas->query);
1811 res->fragment = g_strdup(ref->fragment);
1812 goto step_7;
1816 * 3) If the scheme component is defined, indicating that the reference
1817 * starts with a scheme name, then the reference is interpreted as an
1818 * absolute URI and we are done. Otherwise, the reference URI's
1819 * scheme is inherited from the base URI's scheme component.
1821 if (ref->scheme != NULL) {
1822 val = uri_to_string(ref);
1823 goto done;
1825 res->scheme = g_strdup(bas->scheme);
1827 res->query = g_strdup(ref->query);
1828 res->fragment = g_strdup(ref->fragment);
1831 * 4) If the authority component is defined, then the reference is a
1832 * network-path and we skip to step 7. Otherwise, the reference
1833 * URI's authority is inherited from the base URI's authority
1834 * component, which will also be undefined if the URI scheme does not
1835 * use an authority component.
1837 if ((ref->authority != NULL) || (ref->server != NULL)) {
1838 if (ref->authority != NULL) {
1839 res->authority = g_strdup(ref->authority);
1840 } else {
1841 res->server = g_strdup(ref->server);
1842 res->user = g_strdup(ref->user);
1843 res->port = ref->port;
1845 res->path = g_strdup(ref->path);
1846 goto step_7;
1848 if (bas->authority != NULL) {
1849 res->authority = g_strdup(bas->authority);
1850 } else if (bas->server != NULL) {
1851 res->server = g_strdup(bas->server);
1852 res->user = g_strdup(bas->user);
1853 res->port = bas->port;
1857 * 5) If the path component begins with a slash character ("/"), then
1858 * the reference is an absolute-path and we skip to step 7.
1860 if ((ref->path != NULL) && (ref->path[0] == '/')) {
1861 res->path = g_strdup(ref->path);
1862 goto step_7;
1866 * 6) If this step is reached, then we are resolving a relative-path
1867 * reference. The relative path needs to be merged with the base
1868 * URI's path. Although there are many ways to do this, we will
1869 * describe a simple method using a separate string buffer.
1871 * Allocate a buffer large enough for the result string.
1873 len = 2; /* extra / and 0 */
1874 if (ref->path != NULL) {
1875 len += strlen(ref->path);
1877 if (bas->path != NULL) {
1878 len += strlen(bas->path);
1880 res->path = g_malloc(len);
1881 res->path[0] = 0;
1884 * a) All but the last segment of the base URI's path component is
1885 * copied to the buffer. In other words, any characters after the
1886 * last (right-most) slash character, if any, are excluded.
1888 cur = 0;
1889 out = 0;
1890 if (bas->path != NULL) {
1891 while (bas->path[cur] != 0) {
1892 while ((bas->path[cur] != 0) && (bas->path[cur] != '/')) {
1893 cur++;
1895 if (bas->path[cur] == 0) {
1896 break;
1899 cur++;
1900 while (out < cur) {
1901 res->path[out] = bas->path[out];
1902 out++;
1906 res->path[out] = 0;
1909 * b) The reference's path component is appended to the buffer
1910 * string.
1912 if (ref->path != NULL && ref->path[0] != 0) {
1913 indx = 0;
1915 * Ensure the path includes a '/'
1917 if ((out == 0) && (bas->server != NULL)) {
1918 res->path[out++] = '/';
1920 while (ref->path[indx] != 0) {
1921 res->path[out++] = ref->path[indx++];
1924 res->path[out] = 0;
1927 * Steps c) to h) are really path normalization steps
1929 normalize_uri_path(res->path);
1931 step_7:
1934 * 7) The resulting URI components, including any inherited from the
1935 * base URI, are recombined to give the absolute form of the URI
1936 * reference.
1938 val = uri_to_string(res);
1940 done:
1941 if (ref != NULL) {
1942 uri_free(ref);
1944 if (bas != NULL) {
1945 uri_free(bas);
1947 if (res != NULL) {
1948 uri_free(res);
1950 return val;
1954 * uri_resolve_relative:
1955 * @URI: the URI reference under consideration
1956 * @base: the base value
1958 * Expresses the URI of the reference in terms relative to the
1959 * base. Some examples of this operation include:
1960 * base = "http://site1.com/docs/book1.html"
1961 * URI input URI returned
1962 * docs/pic1.gif pic1.gif
1963 * docs/img/pic1.gif img/pic1.gif
1964 * img/pic1.gif ../img/pic1.gif
1965 * http://site1.com/docs/pic1.gif pic1.gif
1966 * http://site2.com/docs/pic1.gif http://site2.com/docs/pic1.gif
1968 * base = "docs/book1.html"
1969 * URI input URI returned
1970 * docs/pic1.gif pic1.gif
1971 * docs/img/pic1.gif img/pic1.gif
1972 * img/pic1.gif ../img/pic1.gif
1973 * http://site1.com/docs/pic1.gif http://site1.com/docs/pic1.gif
1976 * Note: if the URI reference is really weird or complicated, it may be
1977 * worthwhile to first convert it into a "nice" one by calling
1978 * uri_resolve (using 'base') before calling this routine,
1979 * since this routine (for reasonable efficiency) assumes URI has
1980 * already been through some validation.
1982 * Returns a new URI string (to be freed by the caller) or NULL in case
1983 * error.
1985 char *uri_resolve_relative(const char *uri, const char *base)
1987 char *val = NULL;
1988 int ret;
1989 int ix;
1990 int pos = 0;
1991 int nbslash = 0;
1992 int len;
1993 URI *ref = NULL;
1994 URI *bas = NULL;
1995 char *bptr, *uptr, *vptr;
1996 int remove_path = 0;
1998 if ((uri == NULL) || (*uri == 0)) {
1999 return NULL;
2003 * First parse URI into a standard form
2005 ref = uri_new();
2006 /* If URI not already in "relative" form */
2007 if (uri[0] != '.') {
2008 ret = uri_parse_into(ref, uri);
2009 if (ret != 0) {
2010 goto done; /* Error in URI, return NULL */
2012 } else {
2013 ref->path = g_strdup(uri);
2017 * Next parse base into the same standard form
2019 if ((base == NULL) || (*base == 0)) {
2020 val = g_strdup(uri);
2021 goto done;
2023 bas = uri_new();
2024 if (base[0] != '.') {
2025 ret = uri_parse_into(bas, base);
2026 if (ret != 0) {
2027 goto done; /* Error in base, return NULL */
2029 } else {
2030 bas->path = g_strdup(base);
2034 * If the scheme / server on the URI differs from the base,
2035 * just return the URI
2037 if ((ref->scheme != NULL) &&
2038 ((bas->scheme == NULL) || (strcmp(bas->scheme, ref->scheme)) ||
2039 (strcmp(bas->server, ref->server)))) {
2040 val = g_strdup(uri);
2041 goto done;
2043 if (bas->path == ref->path ||
2044 (bas->path && ref->path && !strcmp(bas->path, ref->path))) {
2045 val = g_strdup("");
2046 goto done;
2048 if (bas->path == NULL) {
2049 val = g_strdup(ref->path);
2050 goto done;
2052 if (ref->path == NULL) {
2053 ref->path = (char *)"/";
2054 remove_path = 1;
2058 * At this point (at last!) we can compare the two paths
2060 * First we take care of the special case where either of the
2061 * two path components may be missing (bug 316224)
2063 if (bas->path == NULL) {
2064 if (ref->path != NULL) {
2065 uptr = ref->path;
2066 if (*uptr == '/') {
2067 uptr++;
2069 /* exception characters from uri_to_string */
2070 val = uri_string_escape(uptr, "/;&=+$,");
2072 goto done;
2074 bptr = bas->path;
2075 if (ref->path == NULL) {
2076 for (ix = 0; bptr[ix] != 0; ix++) {
2077 if (bptr[ix] == '/') {
2078 nbslash++;
2081 uptr = NULL;
2082 len = 1; /* this is for a string terminator only */
2083 } else {
2085 * Next we compare the two strings and find where they first differ
2087 if ((ref->path[pos] == '.') && (ref->path[pos + 1] == '/')) {
2088 pos += 2;
2090 if ((*bptr == '.') && (bptr[1] == '/')) {
2091 bptr += 2;
2092 } else if ((*bptr == '/') && (ref->path[pos] != '/')) {
2093 bptr++;
2095 while ((bptr[pos] == ref->path[pos]) && (bptr[pos] != 0)) {
2096 pos++;
2099 if (bptr[pos] == ref->path[pos]) {
2100 val = g_strdup("");
2101 goto done; /* (I can't imagine why anyone would do this) */
2105 * In URI, "back up" to the last '/' encountered. This will be the
2106 * beginning of the "unique" suffix of URI
2108 ix = pos;
2109 if ((ref->path[ix] == '/') && (ix > 0)) {
2110 ix--;
2111 } else if ((ref->path[ix] == 0) && (ix > 1)
2112 && (ref->path[ix - 1] == '/')) {
2113 ix -= 2;
2115 for (; ix > 0; ix--) {
2116 if (ref->path[ix] == '/') {
2117 break;
2120 if (ix == 0) {
2121 uptr = ref->path;
2122 } else {
2123 ix++;
2124 uptr = &ref->path[ix];
2128 * In base, count the number of '/' from the differing point
2130 if (bptr[pos] != ref->path[pos]) { /* check for trivial URI == base */
2131 for (; bptr[ix] != 0; ix++) {
2132 if (bptr[ix] == '/') {
2133 nbslash++;
2137 len = strlen(uptr) + 1;
2140 if (nbslash == 0) {
2141 if (uptr != NULL) {
2142 /* exception characters from uri_to_string */
2143 val = uri_string_escape(uptr, "/;&=+$,");
2145 goto done;
2149 * Allocate just enough space for the returned string -
2150 * length of the remainder of the URI, plus enough space
2151 * for the "../" groups, plus one for the terminator
2153 val = g_malloc(len + 3 * nbslash);
2154 vptr = val;
2156 * Put in as many "../" as needed
2158 for (; nbslash > 0; nbslash--) {
2159 *vptr++ = '.';
2160 *vptr++ = '.';
2161 *vptr++ = '/';
2164 * Finish up with the end of the URI
2166 if (uptr != NULL) {
2167 if ((vptr > val) && (len > 0) && (uptr[0] == '/') &&
2168 (vptr[-1] == '/')) {
2169 memcpy(vptr, uptr + 1, len - 1);
2170 vptr[len - 2] = 0;
2171 } else {
2172 memcpy(vptr, uptr, len);
2173 vptr[len - 1] = 0;
2175 } else {
2176 vptr[len - 1] = 0;
2179 /* escape the freshly-built path */
2180 vptr = val;
2181 /* exception characters from uri_to_string */
2182 val = uri_string_escape(vptr, "/;&=+$,");
2183 g_free(vptr);
2185 done:
2187 * Free the working variables
2189 if (remove_path != 0) {
2190 ref->path = NULL;
2192 if (ref != NULL) {
2193 uri_free(ref);
2195 if (bas != NULL) {
2196 uri_free(bas);
2199 return val;
2203 * Utility functions to help parse and assemble query strings.
2206 struct QueryParams *query_params_new(int init_alloc)
2208 struct QueryParams *ps;
2210 if (init_alloc <= 0) {
2211 init_alloc = 1;
2214 ps = g_new(QueryParams, 1);
2215 ps->n = 0;
2216 ps->alloc = init_alloc;
2217 ps->p = g_new(QueryParam, ps->alloc);
2219 return ps;
2222 /* Ensure there is space to store at least one more parameter
2223 * at the end of the set.
2225 static int query_params_append(struct QueryParams *ps, const char *name,
2226 const char *value)
2228 if (ps->n >= ps->alloc) {
2229 ps->p = g_renew(QueryParam, ps->p, ps->alloc * 2);
2230 ps->alloc *= 2;
2233 ps->p[ps->n].name = g_strdup(name);
2234 ps->p[ps->n].value = g_strdup(value);
2235 ps->p[ps->n].ignore = 0;
2236 ps->n++;
2238 return 0;
2241 void query_params_free(struct QueryParams *ps)
2243 int i;
2245 for (i = 0; i < ps->n; ++i) {
2246 g_free(ps->p[i].name);
2247 g_free(ps->p[i].value);
2249 g_free(ps->p);
2250 g_free(ps);
2253 struct QueryParams *query_params_parse(const char *query)
2255 struct QueryParams *ps;
2256 const char *end, *eq;
2258 ps = query_params_new(0);
2259 if (!query || query[0] == '\0') {
2260 return ps;
2263 while (*query) {
2264 char *name = NULL, *value = NULL;
2266 /* Find the next separator, or end of the string. */
2267 end = strchr(query, '&');
2268 if (!end) {
2269 end = strchr(query, ';');
2271 if (!end) {
2272 end = query + strlen(query);
2275 /* Find the first '=' character between here and end. */
2276 eq = strchr(query, '=');
2277 if (eq && eq >= end) {
2278 eq = NULL;
2281 /* Empty section (eg. "&&"). */
2282 if (end == query) {
2283 goto next;
2286 /* If there is no '=' character, then we have just "name"
2287 * and consistent with CGI.pm we assume value is "".
2289 else if (!eq) {
2290 name = uri_string_unescape(query, end - query, NULL);
2291 value = NULL;
2293 /* Or if we have "name=" here (works around annoying
2294 * problem when calling uri_string_unescape with len = 0).
2296 else if (eq + 1 == end) {
2297 name = uri_string_unescape(query, eq - query, NULL);
2298 value = g_new0(char, 1);
2300 /* If the '=' character is at the beginning then we have
2301 * "=value" and consistent with CGI.pm we _ignore_ this.
2303 else if (query == eq) {
2304 goto next;
2307 /* Otherwise it's "name=value". */
2308 else {
2309 name = uri_string_unescape(query, eq - query, NULL);
2310 value = uri_string_unescape(eq + 1, end - (eq + 1), NULL);
2313 /* Append to the parameter set. */
2314 query_params_append(ps, name, value);
2315 g_free(name);
2316 g_free(value);
2318 next:
2319 query = end;
2320 if (*query) {
2321 query++; /* skip '&' separator */
2325 return ps;