blob: e348c1768c4adbd3a7bbdf1a40da2a3632306de8 [file] [log] [blame]
Paolo Bonzinica0defb2012-09-24 14:42:02 +05301/**
2 * uri.c: set of generic URI related routines
3 *
4 * Reference: RFCs 3986, 2732 and 2373
5 *
6 * Copyright (C) 1998-2003 Daniel Veillard. All Rights Reserved.
7 *
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:
14 *
15 * The above copyright notice and this permission notice shall be included in
16 * all copies or substantial portions of the Software.
17 *
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.
24 *
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.
28 *
29 * daniel@veillard.com
30 *
31 **
32 *
33 * Copyright (C) 2007, 2009-2010 Red Hat, Inc.
34 *
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.
39 *
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.
44 *
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
48 *
49 * Authors:
50 * Richard W.M. Jones <rjones@redhat.com>
51 *
52 */
53
54#include <glib.h>
55#include <string.h>
56#include <stdio.h>
57
Paolo Bonzini1de7afc2012-12-17 18:20:00 +010058#include "qemu/uri.h"
Paolo Bonzinica0defb2012-09-24 14:42:02 +053059
60static void uri_clean(URI *uri);
61
62/*
63 * Old rule from 2396 used in legacy handling code
64 * alpha = lowalpha | upalpha
65 */
66#define IS_ALPHA(x) (IS_LOWALPHA(x) || IS_UPALPHA(x))
67
68
69/*
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"
73 */
74
75#define IS_LOWALPHA(x) (((x) >= 'a') && ((x) <= 'z'))
76
77/*
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"
81 */
82#define IS_UPALPHA(x) (((x) >= 'A') && ((x) <= 'Z'))
83
84#ifdef IS_DIGIT
85#undef IS_DIGIT
86#endif
87/*
88 * digit = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7" | "8" | "9"
89 */
90#define IS_DIGIT(x) (((x) >= '0') && ((x) <= '9'))
91
92/*
93 * alphanum = alpha | digit
94 */
95
96#define IS_ALPHANUM(x) (IS_ALPHA(x) || IS_DIGIT(x))
97
98/*
99 * mark = "-" | "_" | "." | "!" | "~" | "*" | "'" | "(" | ")"
100 */
101
102#define IS_MARK(x) (((x) == '-') || ((x) == '_') || ((x) == '.') || \
103 ((x) == '!') || ((x) == '~') || ((x) == '*') || ((x) == '\'') || \
104 ((x) == '(') || ((x) == ')'))
105
106/*
107 * unwise = "{" | "}" | "|" | "\" | "^" | "`"
108 */
109
110#define IS_UNWISE(p) \
111 (((*(p) == '{')) || ((*(p) == '}')) || ((*(p) == '|')) || \
112 ((*(p) == '\\')) || ((*(p) == '^')) || ((*(p) == '[')) || \
113 ((*(p) == ']')) || ((*(p) == '`')))
114/*
115 * reserved = ";" | "/" | "?" | ":" | "@" | "&" | "=" | "+" | "$" | "," |
116 * "[" | "]"
117 */
118
119#define IS_RESERVED(x) (((x) == ';') || ((x) == '/') || ((x) == '?') || \
120 ((x) == ':') || ((x) == '@') || ((x) == '&') || ((x) == '=') || \
121 ((x) == '+') || ((x) == '$') || ((x) == ',') || ((x) == '[') || \
122 ((x) == ']'))
123
124/*
125 * unreserved = alphanum | mark
126 */
127
128#define IS_UNRESERVED(x) (IS_ALPHANUM(x) || IS_MARK(x))
129
130/*
131 * Skip to next pointer char, handle escaped sequences
132 */
133
134#define NEXT(p) ((*p == '%')? p += 3 : p++)
135
136/*
137 * Productions from the spec.
138 *
139 * authority = server | reg_name
140 * reg_name = 1*( unreserved | escaped | "$" | "," |
141 * ";" | ":" | "@" | "&" | "=" | "+" )
142 *
143 * path = [ abs_path | opaque_part ]
144 */
145
146
147/************************************************************************
148 * *
149 * RFC 3986 parser *
150 * *
151 ************************************************************************/
152
153#define ISA_DIGIT(p) ((*(p) >= '0') && (*(p) <= '9'))
154#define ISA_ALPHA(p) (((*(p) >= 'a') && (*(p) <= 'z')) || \
155 ((*(p) >= 'A') && (*(p) <= 'Z')))
156#define ISA_HEXDIG(p) \
157 (ISA_DIGIT(p) || ((*(p) >= 'a') && (*(p) <= 'f')) || \
158 ((*(p) >= 'A') && (*(p) <= 'F')))
159
160/*
161 * sub-delims = "!" / "$" / "&" / "'" / "(" / ")"
162 * / "*" / "+" / "," / ";" / "="
163 */
164#define ISA_SUB_DELIM(p) \
165 (((*(p) == '!')) || ((*(p) == '$')) || ((*(p) == '&')) || \
166 ((*(p) == '(')) || ((*(p) == ')')) || ((*(p) == '*')) || \
167 ((*(p) == '+')) || ((*(p) == ',')) || ((*(p) == ';')) || \
168 ((*(p) == '=')) || ((*(p) == '\'')))
169
170/*
171 * gen-delims = ":" / "/" / "?" / "#" / "[" / "]" / "@"
172 */
173#define ISA_GEN_DELIM(p) \
174 (((*(p) == ':')) || ((*(p) == '/')) || ((*(p) == '?')) || \
175 ((*(p) == '#')) || ((*(p) == '[')) || ((*(p) == ']')) || \
176 ((*(p) == '@')))
177
178/*
179 * reserved = gen-delims / sub-delims
180 */
181#define ISA_RESERVED(p) (ISA_GEN_DELIM(p) || (ISA_SUB_DELIM(p)))
182
183/*
184 * unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
185 */
186#define ISA_UNRESERVED(p) \
187 ((ISA_ALPHA(p)) || (ISA_DIGIT(p)) || ((*(p) == '-')) || \
188 ((*(p) == '.')) || ((*(p) == '_')) || ((*(p) == '~')))
189
190/*
191 * pct-encoded = "%" HEXDIG HEXDIG
192 */
193#define ISA_PCT_ENCODED(p) \
194 ((*(p) == '%') && (ISA_HEXDIG(p + 1)) && (ISA_HEXDIG(p + 2)))
195
196/*
197 * pchar = unreserved / pct-encoded / sub-delims / ":" / "@"
198 */
199#define ISA_PCHAR(p) \
200 (ISA_UNRESERVED(p) || ISA_PCT_ENCODED(p) || ISA_SUB_DELIM(p) || \
201 ((*(p) == ':')) || ((*(p) == '@')))
202
203/**
204 * rfc3986_parse_scheme:
205 * @uri: pointer to an URI structure
206 * @str: pointer to the string to analyze
207 *
208 * Parse an URI scheme
209 *
210 * ALPHA *( ALPHA / DIGIT / "+" / "-" / "." )
211 *
212 * Returns 0 or the error code
213 */
214static int
215rfc3986_parse_scheme(URI *uri, const char **str) {
216 const char *cur;
217
218 if (str == NULL)
219 return(-1);
220
221 cur = *str;
222 if (!ISA_ALPHA(cur))
223 return(2);
224 cur++;
225 while (ISA_ALPHA(cur) || ISA_DIGIT(cur) ||
226 (*cur == '+') || (*cur == '-') || (*cur == '.')) cur++;
227 if (uri != NULL) {
228 if (uri->scheme != NULL) g_free(uri->scheme);
229 uri->scheme = g_strndup(*str, cur - *str);
230 }
231 *str = cur;
232 return(0);
233}
234
235/**
236 * rfc3986_parse_fragment:
237 * @uri: pointer to an URI structure
238 * @str: pointer to the string to analyze
239 *
240 * Parse the query part of an URI
241 *
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.
247 *
248 * Returns 0 or the error code
249 */
250static int
251rfc3986_parse_fragment(URI *uri, const char **str)
252{
253 const char *cur;
254
255 if (str == NULL)
256 return (-1);
257
258 cur = *str;
259
260 while ((ISA_PCHAR(cur)) || (*cur == '/') || (*cur == '?') ||
261 (*cur == '[') || (*cur == ']') ||
262 ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))
263 NEXT(cur);
264 if (uri != NULL) {
265 if (uri->fragment != 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);
271 }
272 *str = cur;
273 return (0);
274}
275
276/**
277 * rfc3986_parse_query:
278 * @uri: pointer to an URI structure
279 * @str: pointer to the string to analyze
280 *
281 * Parse the query part of an URI
282 *
283 * query = *uric
284 *
285 * Returns 0 or the error code
286 */
287static int
288rfc3986_parse_query(URI *uri, const char **str)
289{
290 const char *cur;
291
292 if (str == NULL)
293 return (-1);
294
295 cur = *str;
296
297 while ((ISA_PCHAR(cur)) || (*cur == '/') || (*cur == '?') ||
298 ((uri != NULL) && (uri->cleanup & 1) && (IS_UNWISE(cur))))
299 NEXT(cur);
300 if (uri != NULL) {
301 if (uri->query != NULL)
302 g_free (uri->query);
303 uri->query = g_strndup (*str, cur - *str);
304 }
305 *str = cur;
306 return (0);
307}
308
309/**
310 * rfc3986_parse_port:
311 * @uri: pointer to an URI structure
312 * @str: the string to analyze
313 *
314 * Parse a port part and fills in the appropriate fields
315 * of the @uri structure
316 *
317 * port = *DIGIT
318 *
319 * Returns 0 or the error code
320 */
321static int
322rfc3986_parse_port(URI *uri, const char **str)
323{
324 const char *cur = *str;
325
326 if (ISA_DIGIT(cur)) {
327 if (uri != NULL)
328 uri->port = 0;
329 while (ISA_DIGIT(cur)) {
330 if (uri != NULL)
331 uri->port = uri->port * 10 + (*cur - '0');
332 cur++;
333 }
334 *str = cur;
335 return(0);
336 }
337 return(1);
338}
339
340/**
341 * rfc3986_parse_user_info:
342 * @uri: pointer to an URI structure
343 * @str: the string to analyze
344 *
345 * Parse an user informations part and fills in the appropriate fields
346 * of the @uri structure
347 *
348 * userinfo = *( unreserved / pct-encoded / sub-delims / ":" )
349 *
350 * Returns 0 or the error code
351 */
352static int
353rfc3986_parse_user_info(URI *uri, const char **str)
354{
355 const char *cur;
356
357 cur = *str;
358 while (ISA_UNRESERVED(cur) || ISA_PCT_ENCODED(cur) ||
359 ISA_SUB_DELIM(cur) || (*cur == ':'))
360 NEXT(cur);
361 if (*cur == '@') {
362 if (uri != NULL) {
363 if (uri->user != NULL) g_free(uri->user);
364 if (uri->cleanup & 2)
365 uri->user = g_strndup(*str, cur - *str);
366 else
367 uri->user = uri_string_unescape(*str, cur - *str, NULL);
368 }
369 *str = cur;
370 return(0);
371 }
372 return(1);
373}
374
375/**
376 * rfc3986_parse_dec_octet:
377 * @str: the string to analyze
378 *
379 * dec-octet = DIGIT ; 0-9
380 * / %x31-39 DIGIT ; 10-99
381 * / "1" 2DIGIT ; 100-199
382 * / "2" %x30-34 DIGIT ; 200-249
383 * / "25" %x30-35 ; 250-255
384 *
385 * Skip a dec-octet.
386 *
387 * Returns 0 if found and skipped, 1 otherwise
388 */
389static int
390rfc3986_parse_dec_octet(const char **str) {
391 const char *cur = *str;
392
393 if (!(ISA_DIGIT(cur)))
394 return(1);
395 if (!ISA_DIGIT(cur+1))
396 cur++;
397 else if ((*cur != '0') && (ISA_DIGIT(cur + 1)) && (!ISA_DIGIT(cur+2)))
398 cur += 2;
399 else if ((*cur == '1') && (ISA_DIGIT(cur + 1)) && (ISA_DIGIT(cur + 2)))
400 cur += 3;
401 else if ((*cur == '2') && (*(cur + 1) >= '0') &&
402 (*(cur + 1) <= '4') && (ISA_DIGIT(cur + 2)))
403 cur += 3;
404 else if ((*cur == '2') && (*(cur + 1) == '5') &&
405 (*(cur + 2) >= '0') && (*(cur + 1) <= '5'))
406 cur += 3;
407 else
408 return(1);
409 *str = cur;
410 return(0);
411}
412/**
413 * rfc3986_parse_host:
414 * @uri: pointer to an URI structure
415 * @str: the string to analyze
416 *
417 * Parse an host part and fills in the appropriate fields
418 * of the @uri structure
419 *
420 * host = IP-literal / IPv4address / reg-name
421 * IP-literal = "[" ( IPv6address / IPvFuture ) "]"
422 * IPv4address = dec-octet "." dec-octet "." dec-octet "." dec-octet
423 * reg-name = *( unreserved / pct-encoded / sub-delims )
424 *
425 * Returns 0 or the error code
426 */
427static int
428rfc3986_parse_host(URI *uri, const char **str)
429{
430 const char *cur = *str;
431 const char *host;
432
433 host = cur;
434 /*
Stefan Weila93cf9d2012-11-02 08:29:53 +0100435 * IPv6 and future addressing scheme are enclosed between brackets
Paolo Bonzinica0defb2012-09-24 14:42:02 +0530436 */
437 if (*cur == '[') {
438 cur++;
439 while ((*cur != ']') && (*cur != 0))
440 cur++;
441 if (*cur != ']')
442 return(1);
443 cur++;
444 goto found;
445 }
446 /*
447 * try to parse an IPv4
448 */
449 if (ISA_DIGIT(cur)) {
450 if (rfc3986_parse_dec_octet(&cur) != 0)
451 goto not_ipv4;
452 if (*cur != '.')
453 goto not_ipv4;
454 cur++;
455 if (rfc3986_parse_dec_octet(&cur) != 0)
456 goto not_ipv4;
457 if (*cur != '.')
458 goto not_ipv4;
459 if (rfc3986_parse_dec_octet(&cur) != 0)
460 goto not_ipv4;
461 if (*cur != '.')
462 goto not_ipv4;
463 if (rfc3986_parse_dec_octet(&cur) != 0)
464 goto not_ipv4;
465 goto found;
466not_ipv4:
467 cur = *str;
468 }
469 /*
470 * then this should be a hostname which can be empty
471 */
472 while (ISA_UNRESERVED(cur) || ISA_PCT_ENCODED(cur) || ISA_SUB_DELIM(cur))
473 NEXT(cur);
474found:
475 if (uri != NULL) {
476 if (uri->authority != NULL) g_free(uri->authority);
477 uri->authority = NULL;
478 if (uri->server != NULL) g_free(uri->server);
479 if (cur != host) {
480 if (uri->cleanup & 2)
481 uri->server = g_strndup(host, cur - host);
482 else
483 uri->server = uri_string_unescape(host, cur - host, NULL);
484 } else
485 uri->server = NULL;
486 }
487 *str = cur;
488 return(0);
489}
490
491/**
492 * rfc3986_parse_authority:
493 * @uri: pointer to an URI structure
494 * @str: the string to analyze
495 *
496 * Parse an authority part and fills in the appropriate fields
497 * of the @uri structure
498 *
499 * authority = [ userinfo "@" ] host [ ":" port ]
500 *
501 * Returns 0 or the error code
502 */
503static int
504rfc3986_parse_authority(URI *uri, const char **str)
505{
506 const char *cur;
507 int ret;
508
509 cur = *str;
510 /*
511 * try to parse an userinfo and check for the trailing @
512 */
513 ret = rfc3986_parse_user_info(uri, &cur);
514 if ((ret != 0) || (*cur != '@'))
515 cur = *str;
516 else
517 cur++;
518 ret = rfc3986_parse_host(uri, &cur);
519 if (ret != 0) return(ret);
520 if (*cur == ':') {
521 cur++;
522 ret = rfc3986_parse_port(uri, &cur);
523 if (ret != 0) return(ret);
524 }
525 *str = cur;
526 return(0);
527}
528
529/**
530 * rfc3986_parse_segment:
531 * @str: the string to analyze
532 * @forbid: an optional forbidden character
533 * @empty: allow an empty segment
534 *
535 * Parse a segment and fills in the appropriate fields
536 * of the @uri structure
537 *
538 * segment = *pchar
539 * segment-nz = 1*pchar
540 * segment-nz-nc = 1*( unreserved / pct-encoded / sub-delims / "@" )
541 * ; non-zero-length segment without any colon ":"
542 *
543 * Returns 0 or the error code
544 */
545static int
546rfc3986_parse_segment(const char **str, char forbid, int empty)
547{
548 const char *cur;
549
550 cur = *str;
551 if (!ISA_PCHAR(cur)) {
552 if (empty)
553 return(0);
554 return(1);
555 }
556 while (ISA_PCHAR(cur) && (*cur != forbid))
557 NEXT(cur);
558 *str = cur;
559 return (0);
560}
561
562/**
563 * rfc3986_parse_path_ab_empty:
564 * @uri: pointer to an URI structure
565 * @str: the string to analyze
566 *
567 * Parse an path absolute or empty and fills in the appropriate fields
568 * of the @uri structure
569 *
570 * path-abempty = *( "/" segment )
571 *
572 * Returns 0 or the error code
573 */
574static int
575rfc3986_parse_path_ab_empty(URI *uri, const char **str)
576{
577 const char *cur;
578 int ret;
579
580 cur = *str;
581
582 while (*cur == '/') {
583 cur++;
584 ret = rfc3986_parse_segment(&cur, 0, 1);
585 if (ret != 0) return(ret);
586 }
587 if (uri != NULL) {
588 if (uri->path != NULL) g_free(uri->path);
589 if (*str != cur) {
590 if (uri->cleanup & 2)
591 uri->path = g_strndup(*str, cur - *str);
592 else
593 uri->path = uri_string_unescape(*str, cur - *str, NULL);
594 } else {
595 uri->path = NULL;
596 }
597 }
598 *str = cur;
599 return (0);
600}
601
602/**
603 * rfc3986_parse_path_absolute:
604 * @uri: pointer to an URI structure
605 * @str: the string to analyze
606 *
607 * Parse an path absolute and fills in the appropriate fields
608 * of the @uri structure
609 *
610 * path-absolute = "/" [ segment-nz *( "/" segment ) ]
611 *
612 * Returns 0 or the error code
613 */
614static int
615rfc3986_parse_path_absolute(URI *uri, const char **str)
616{
617 const char *cur;
618 int ret;
619
620 cur = *str;
621
622 if (*cur != '/')
623 return(1);
624 cur++;
625 ret = rfc3986_parse_segment(&cur, 0, 0);
626 if (ret == 0) {
627 while (*cur == '/') {
628 cur++;
629 ret = rfc3986_parse_segment(&cur, 0, 1);
630 if (ret != 0) return(ret);
631 }
632 }
633 if (uri != NULL) {
634 if (uri->path != NULL) g_free(uri->path);
635 if (cur != *str) {
636 if (uri->cleanup & 2)
637 uri->path = g_strndup(*str, cur - *str);
638 else
639 uri->path = uri_string_unescape(*str, cur - *str, NULL);
640 } else {
641 uri->path = NULL;
642 }
643 }
644 *str = cur;
645 return (0);
646}
647
648/**
649 * rfc3986_parse_path_rootless:
650 * @uri: pointer to an URI structure
651 * @str: the string to analyze
652 *
653 * Parse an path without root and fills in the appropriate fields
654 * of the @uri structure
655 *
656 * path-rootless = segment-nz *( "/" segment )
657 *
658 * Returns 0 or the error code
659 */
660static int
661rfc3986_parse_path_rootless(URI *uri, const char **str)
662{
663 const char *cur;
664 int ret;
665
666 cur = *str;
667
668 ret = rfc3986_parse_segment(&cur, 0, 0);
669 if (ret != 0) return(ret);
670 while (*cur == '/') {
671 cur++;
672 ret = rfc3986_parse_segment(&cur, 0, 1);
673 if (ret != 0) return(ret);
674 }
675 if (uri != NULL) {
676 if (uri->path != NULL) g_free(uri->path);
677 if (cur != *str) {
678 if (uri->cleanup & 2)
679 uri->path = g_strndup(*str, cur - *str);
680 else
681 uri->path = uri_string_unescape(*str, cur - *str, NULL);
682 } else {
683 uri->path = NULL;
684 }
685 }
686 *str = cur;
687 return (0);
688}
689
690/**
691 * rfc3986_parse_path_no_scheme:
692 * @uri: pointer to an URI structure
693 * @str: the string to analyze
694 *
695 * Parse an path which is not a scheme and fills in the appropriate fields
696 * of the @uri structure
697 *
698 * path-noscheme = segment-nz-nc *( "/" segment )
699 *
700 * Returns 0 or the error code
701 */
702static int
703rfc3986_parse_path_no_scheme(URI *uri, const char **str)
704{
705 const char *cur;
706 int ret;
707
708 cur = *str;
709
710 ret = rfc3986_parse_segment(&cur, ':', 0);
711 if (ret != 0) return(ret);
712 while (*cur == '/') {
713 cur++;
714 ret = rfc3986_parse_segment(&cur, 0, 1);
715 if (ret != 0) return(ret);
716 }
717 if (uri != NULL) {
718 if (uri->path != NULL) g_free(uri->path);
719 if (cur != *str) {
720 if (uri->cleanup & 2)
721 uri->path = g_strndup(*str, cur - *str);
722 else
723 uri->path = uri_string_unescape(*str, cur - *str, NULL);
724 } else {
725 uri->path = NULL;
726 }
727 }
728 *str = cur;
729 return (0);
730}
731
732/**
733 * rfc3986_parse_hier_part:
734 * @uri: pointer to an URI structure
735 * @str: the string to analyze
736 *
737 * Parse an hierarchical part and fills in the appropriate fields
738 * of the @uri structure
739 *
740 * hier-part = "//" authority path-abempty
741 * / path-absolute
742 * / path-rootless
743 * / path-empty
744 *
745 * Returns 0 or the error code
746 */
747static int
748rfc3986_parse_hier_part(URI *uri, const char **str)
749{
750 const char *cur;
751 int ret;
752
753 cur = *str;
754
755 if ((*cur == '/') && (*(cur + 1) == '/')) {
756 cur += 2;
757 ret = rfc3986_parse_authority(uri, &cur);
758 if (ret != 0) return(ret);
759 ret = rfc3986_parse_path_ab_empty(uri, &cur);
760 if (ret != 0) return(ret);
761 *str = cur;
762 return(0);
763 } else if (*cur == '/') {
764 ret = rfc3986_parse_path_absolute(uri, &cur);
765 if (ret != 0) return(ret);
766 } else if (ISA_PCHAR(cur)) {
767 ret = rfc3986_parse_path_rootless(uri, &cur);
768 if (ret != 0) return(ret);
769 } else {
770 /* path-empty is effectively empty */
771 if (uri != NULL) {
772 if (uri->path != NULL) g_free(uri->path);
773 uri->path = NULL;
774 }
775 }
776 *str = cur;
777 return (0);
778}
779
780/**
781 * rfc3986_parse_relative_ref:
782 * @uri: pointer to an URI structure
783 * @str: the string to analyze
784 *
785 * Parse an URI string and fills in the appropriate fields
786 * of the @uri structure
787 *
788 * relative-ref = relative-part [ "?" query ] [ "#" fragment ]
789 * relative-part = "//" authority path-abempty
790 * / path-absolute
791 * / path-noscheme
792 * / path-empty
793 *
794 * Returns 0 or the error code
795 */
796static int
797rfc3986_parse_relative_ref(URI *uri, const char *str) {
798 int ret;
799
800 if ((*str == '/') && (*(str + 1) == '/')) {
801 str += 2;
802 ret = rfc3986_parse_authority(uri, &str);
803 if (ret != 0) return(ret);
804 ret = rfc3986_parse_path_ab_empty(uri, &str);
805 if (ret != 0) return(ret);
806 } else if (*str == '/') {
807 ret = rfc3986_parse_path_absolute(uri, &str);
808 if (ret != 0) return(ret);
809 } else if (ISA_PCHAR(str)) {
810 ret = rfc3986_parse_path_no_scheme(uri, &str);
811 if (ret != 0) return(ret);
812 } else {
813 /* path-empty is effectively empty */
814 if (uri != NULL) {
815 if (uri->path != NULL) g_free(uri->path);
816 uri->path = NULL;
817 }
818 }
819
820 if (*str == '?') {
821 str++;
822 ret = rfc3986_parse_query(uri, &str);
823 if (ret != 0) return(ret);
824 }
825 if (*str == '#') {
826 str++;
827 ret = rfc3986_parse_fragment(uri, &str);
828 if (ret != 0) return(ret);
829 }
830 if (*str != 0) {
831 uri_clean(uri);
832 return(1);
833 }
834 return(0);
835}
836
837
838/**
839 * rfc3986_parse:
840 * @uri: pointer to an URI structure
841 * @str: the string to analyze
842 *
843 * Parse an URI string and fills in the appropriate fields
844 * of the @uri structure
845 *
846 * scheme ":" hier-part [ "?" query ] [ "#" fragment ]
847 *
848 * Returns 0 or the error code
849 */
850static int
851rfc3986_parse(URI *uri, const char *str) {
852 int ret;
853
854 ret = rfc3986_parse_scheme(uri, &str);
855 if (ret != 0) return(ret);
856 if (*str != ':') {
857 return(1);
858 }
859 str++;
860 ret = rfc3986_parse_hier_part(uri, &str);
861 if (ret != 0) return(ret);
862 if (*str == '?') {
863 str++;
864 ret = rfc3986_parse_query(uri, &str);
865 if (ret != 0) return(ret);
866 }
867 if (*str == '#') {
868 str++;
869 ret = rfc3986_parse_fragment(uri, &str);
870 if (ret != 0) return(ret);
871 }
872 if (*str != 0) {
873 uri_clean(uri);
874 return(1);
875 }
876 return(0);
877}
878
879/**
880 * rfc3986_parse_uri_reference:
881 * @uri: pointer to an URI structure
882 * @str: the string to analyze
883 *
884 * Parse an URI reference string and fills in the appropriate fields
885 * of the @uri structure
886 *
887 * URI-reference = URI / relative-ref
888 *
889 * Returns 0 or the error code
890 */
891static int
892rfc3986_parse_uri_reference(URI *uri, const char *str) {
893 int ret;
894
895 if (str == NULL)
896 return(-1);
897 uri_clean(uri);
898
899 /*
900 * Try first to parse absolute refs, then fallback to relative if
901 * it fails.
902 */
903 ret = rfc3986_parse(uri, str);
904 if (ret != 0) {
905 uri_clean(uri);
906 ret = rfc3986_parse_relative_ref(uri, str);
907 if (ret != 0) {
908 uri_clean(uri);
909 return(ret);
910 }
911 }
912 return(0);
913}
914
915/**
916 * uri_parse:
917 * @str: the URI string to analyze
918 *
919 * Parse an URI based on RFC 3986
920 *
921 * URI-reference = [ absoluteURI | relativeURI ] [ "#" fragment ]
922 *
923 * Returns a newly built URI or NULL in case of error
924 */
925URI *
926uri_parse(const char *str) {
927 URI *uri;
928 int ret;
929
930 if (str == NULL)
931 return(NULL);
932 uri = uri_new();
933 if (uri != NULL) {
934 ret = rfc3986_parse_uri_reference(uri, str);
935 if (ret) {
936 uri_free(uri);
937 return(NULL);
938 }
939 }
940 return(uri);
941}
942
943/**
944 * uri_parse_into:
945 * @uri: pointer to an URI structure
946 * @str: the string to analyze
947 *
948 * Parse an URI reference string based on RFC 3986 and fills in the
949 * appropriate fields of the @uri structure
950 *
951 * URI-reference = URI / relative-ref
952 *
953 * Returns 0 or the error code
954 */
955int
956uri_parse_into(URI *uri, const char *str) {
957 return(rfc3986_parse_uri_reference(uri, str));
958}
959
960/**
961 * uri_parse_raw:
962 * @str: the URI string to analyze
963 * @raw: if 1 unescaping of URI pieces are disabled
964 *
965 * Parse an URI but allows to keep intact the original fragments.
966 *
967 * URI-reference = URI / relative-ref
968 *
969 * Returns a newly built URI or NULL in case of error
970 */
971URI *
972uri_parse_raw(const char *str, int raw) {
973 URI *uri;
974 int ret;
975
976 if (str == NULL)
977 return(NULL);
978 uri = uri_new();
979 if (uri != NULL) {
980 if (raw) {
981 uri->cleanup |= 2;
982 }
983 ret = uri_parse_into(uri, str);
984 if (ret) {
985 uri_free(uri);
986 return(NULL);
987 }
988 }
989 return(uri);
990}
991
992/************************************************************************
993 * *
994 * Generic URI structure functions *
995 * *
996 ************************************************************************/
997
998/**
999 * uri_new:
1000 *
1001 * Simply creates an empty URI
1002 *
1003 * Returns the new structure or NULL in case of error
1004 */
1005URI *
1006uri_new(void) {
1007 URI *ret;
1008
1009 ret = (URI *) g_malloc(sizeof(URI));
1010 memset(ret, 0, sizeof(URI));
1011 return(ret);
1012}
1013
1014/**
1015 * realloc2n:
1016 *
1017 * Function to handle properly a reallocation when saving an URI
1018 * Also imposes some limit on the length of an URI string output
1019 */
1020static char *
1021realloc2n(char *ret, int *max) {
1022 char *temp;
1023 int tmp;
1024
1025 tmp = *max * 2;
1026 temp = g_realloc(ret, (tmp + 1));
1027 *max = tmp;
1028 return(temp);
1029}
1030
1031/**
1032 * uri_to_string:
1033 * @uri: pointer to an URI
1034 *
1035 * Save the URI as an escaped string
1036 *
1037 * Returns a new string (to be deallocated by caller)
1038 */
1039char *
1040uri_to_string(URI *uri) {
1041 char *ret = NULL;
1042 char *temp;
1043 const char *p;
1044 int len;
1045 int max;
1046
1047 if (uri == NULL) return(NULL);
1048
1049
1050 max = 80;
1051 ret = g_malloc(max + 1);
1052 len = 0;
1053
1054 if (uri->scheme != NULL) {
1055 p = uri->scheme;
1056 while (*p != 0) {
1057 if (len >= max) {
1058 temp = realloc2n(ret, &max);
1059 if (temp == NULL) goto mem_error;
1060 ret = temp;
1061 }
1062 ret[len++] = *p++;
1063 }
1064 if (len >= max) {
1065 temp = realloc2n(ret, &max);
1066 if (temp == NULL) goto mem_error;
1067 ret = temp;
1068 }
1069 ret[len++] = ':';
1070 }
1071 if (uri->opaque != NULL) {
1072 p = uri->opaque;
1073 while (*p != 0) {
1074 if (len + 3 >= max) {
1075 temp = realloc2n(ret, &max);
1076 if (temp == NULL) goto mem_error;
1077 ret = temp;
1078 }
1079 if (IS_RESERVED(*(p)) || IS_UNRESERVED(*(p)))
1080 ret[len++] = *p++;
1081 else {
1082 int val = *(unsigned char *)p++;
1083 int hi = val / 0x10, lo = val % 0x10;
1084 ret[len++] = '%';
1085 ret[len++] = hi + (hi > 9? 'A'-10 : '0');
1086 ret[len++] = lo + (lo > 9? 'A'-10 : '0');
1087 }
1088 }
1089 } else {
1090 if (uri->server != NULL) {
1091 if (len + 3 >= max) {
1092 temp = realloc2n(ret, &max);
1093 if (temp == NULL) goto mem_error;
1094 ret = temp;
1095 }
1096 ret[len++] = '/';
1097 ret[len++] = '/';
1098 if (uri->user != NULL) {
1099 p = uri->user;
1100 while (*p != 0) {
1101 if (len + 3 >= max) {
1102 temp = realloc2n(ret, &max);
1103 if (temp == NULL) goto mem_error;
1104 ret = temp;
1105 }
1106 if ((IS_UNRESERVED(*(p))) ||
1107 ((*(p) == ';')) || ((*(p) == ':')) ||
1108 ((*(p) == '&')) || ((*(p) == '=')) ||
1109 ((*(p) == '+')) || ((*(p) == '$')) ||
1110 ((*(p) == ',')))
1111 ret[len++] = *p++;
1112 else {
1113 int val = *(unsigned char *)p++;
1114 int hi = val / 0x10, lo = val % 0x10;
1115 ret[len++] = '%';
1116 ret[len++] = hi + (hi > 9? 'A'-10 : '0');
1117 ret[len++] = lo + (lo > 9? 'A'-10 : '0');
1118 }
1119 }
1120 if (len + 3 >= max) {
1121 temp = realloc2n(ret, &max);
1122 if (temp == NULL) goto mem_error;
1123 ret = temp;
1124 }
1125 ret[len++] = '@';
1126 }
1127 p = uri->server;
1128 while (*p != 0) {
1129 if (len >= max) {
1130 temp = realloc2n(ret, &max);
1131 if (temp == NULL) goto mem_error;
1132 ret = temp;
1133 }
1134 ret[len++] = *p++;
1135 }
1136 if (uri->port > 0) {
1137 if (len + 10 >= max) {
1138 temp = realloc2n(ret, &max);
1139 if (temp == NULL) goto mem_error;
1140 ret = temp;
1141 }
1142 len += snprintf(&ret[len], max - len, ":%d", uri->port);
1143 }
1144 } else if (uri->authority != NULL) {
1145 if (len + 3 >= max) {
1146 temp = realloc2n(ret, &max);
1147 if (temp == NULL) goto mem_error;
1148 ret = temp;
1149 }
1150 ret[len++] = '/';
1151 ret[len++] = '/';
1152 p = uri->authority;
1153 while (*p != 0) {
1154 if (len + 3 >= max) {
1155 temp = realloc2n(ret, &max);
1156 if (temp == NULL) goto mem_error;
1157 ret = temp;
1158 }
1159 if ((IS_UNRESERVED(*(p))) ||
1160 ((*(p) == '$')) || ((*(p) == ',')) || ((*(p) == ';')) ||
1161 ((*(p) == ':')) || ((*(p) == '@')) || ((*(p) == '&')) ||
1162 ((*(p) == '=')) || ((*(p) == '+')))
1163 ret[len++] = *p++;
1164 else {
1165 int val = *(unsigned char *)p++;
1166 int hi = val / 0x10, lo = val % 0x10;
1167 ret[len++] = '%';
1168 ret[len++] = hi + (hi > 9? 'A'-10 : '0');
1169 ret[len++] = lo + (lo > 9? 'A'-10 : '0');
1170 }
1171 }
1172 } else if (uri->scheme != NULL) {
1173 if (len + 3 >= max) {
1174 temp = realloc2n(ret, &max);
1175 if (temp == NULL) goto mem_error;
1176 ret = temp;
1177 }
1178 ret[len++] = '/';
1179 ret[len++] = '/';
1180 }
1181 if (uri->path != NULL) {
1182 p = uri->path;
1183 /*
1184 * the colon in file:///d: should not be escaped or
1185 * Windows accesses fail later.
1186 */
1187 if ((uri->scheme != NULL) &&
1188 (p[0] == '/') &&
1189 (((p[1] >= 'a') && (p[1] <= 'z')) ||
1190 ((p[1] >= 'A') && (p[1] <= 'Z'))) &&
1191 (p[2] == ':') &&
1192 (!strcmp(uri->scheme, "file"))) {
1193 if (len + 3 >= max) {
1194 temp = realloc2n(ret, &max);
1195 if (temp == NULL) goto mem_error;
1196 ret = temp;
1197 }
1198 ret[len++] = *p++;
1199 ret[len++] = *p++;
1200 ret[len++] = *p++;
1201 }
1202 while (*p != 0) {
1203 if (len + 3 >= max) {
1204 temp = realloc2n(ret, &max);
1205 if (temp == NULL) goto mem_error;
1206 ret = temp;
1207 }
1208 if ((IS_UNRESERVED(*(p))) || ((*(p) == '/')) ||
1209 ((*(p) == ';')) || ((*(p) == '@')) || ((*(p) == '&')) ||
1210 ((*(p) == '=')) || ((*(p) == '+')) || ((*(p) == '$')) ||
1211 ((*(p) == ',')))
1212 ret[len++] = *p++;
1213 else {
1214 int val = *(unsigned char *)p++;
1215 int hi = val / 0x10, lo = val % 0x10;
1216 ret[len++] = '%';
1217 ret[len++] = hi + (hi > 9? 'A'-10 : '0');
1218 ret[len++] = lo + (lo > 9? 'A'-10 : '0');
1219 }
1220 }
1221 }
1222 if (uri->query != NULL) {
1223 if (len + 1 >= max) {
1224 temp = realloc2n(ret, &max);
1225 if (temp == NULL) goto mem_error;
1226 ret = temp;
1227 }
1228 ret[len++] = '?';
1229 p = uri->query;
1230 while (*p != 0) {
1231 if (len + 1 >= max) {
1232 temp = realloc2n(ret, &max);
1233 if (temp == NULL) goto mem_error;
1234 ret = temp;
1235 }
1236 ret[len++] = *p++;
1237 }
1238 }
1239 }
1240 if (uri->fragment != NULL) {
1241 if (len + 3 >= max) {
1242 temp = realloc2n(ret, &max);
1243 if (temp == NULL) goto mem_error;
1244 ret = temp;
1245 }
1246 ret[len++] = '#';
1247 p = uri->fragment;
1248 while (*p != 0) {
1249 if (len + 3 >= max) {
1250 temp = realloc2n(ret, &max);
1251 if (temp == NULL) goto mem_error;
1252 ret = temp;
1253 }
1254 if ((IS_UNRESERVED(*(p))) || (IS_RESERVED(*(p))))
1255 ret[len++] = *p++;
1256 else {
1257 int val = *(unsigned char *)p++;
1258 int hi = val / 0x10, lo = val % 0x10;
1259 ret[len++] = '%';
1260 ret[len++] = hi + (hi > 9? 'A'-10 : '0');
1261 ret[len++] = lo + (lo > 9? 'A'-10 : '0');
1262 }
1263 }
1264 }
1265 if (len >= max) {
1266 temp = realloc2n(ret, &max);
1267 if (temp == NULL) goto mem_error;
1268 ret = temp;
1269 }
1270 ret[len] = 0;
1271 return(ret);
1272
1273mem_error:
1274 g_free(ret);
1275 return(NULL);
1276}
1277
1278/**
1279 * uri_clean:
1280 * @uri: pointer to an URI
1281 *
1282 * Make sure the URI struct is free of content
1283 */
1284static void
1285uri_clean(URI *uri) {
1286 if (uri == NULL) return;
1287
1288 if (uri->scheme != NULL) g_free(uri->scheme);
1289 uri->scheme = NULL;
1290 if (uri->server != NULL) g_free(uri->server);
1291 uri->server = NULL;
1292 if (uri->user != NULL) g_free(uri->user);
1293 uri->user = NULL;
1294 if (uri->path != NULL) g_free(uri->path);
1295 uri->path = NULL;
1296 if (uri->fragment != NULL) g_free(uri->fragment);
1297 uri->fragment = NULL;
1298 if (uri->opaque != NULL) g_free(uri->opaque);
1299 uri->opaque = NULL;
1300 if (uri->authority != NULL) g_free(uri->authority);
1301 uri->authority = NULL;
1302 if (uri->query != NULL) g_free(uri->query);
1303 uri->query = NULL;
1304}
1305
1306/**
1307 * uri_free:
1308 * @uri: pointer to an URI
1309 *
1310 * Free up the URI struct
1311 */
1312void
1313uri_free(URI *uri) {
1314 uri_clean(uri);
1315 g_free(uri);
1316}
1317
1318/************************************************************************
1319 * *
1320 * Helper functions *
1321 * *
1322 ************************************************************************/
1323
1324/**
1325 * normalize_uri_path:
1326 * @path: pointer to the path string
1327 *
1328 * Applies the 5 normalization steps to a path string--that is, RFC 2396
1329 * Section 5.2, steps 6.c through 6.g.
1330 *
1331 * Normalization occurs directly on the string, no new allocation is done
1332 *
1333 * Returns 0 or an error code
1334 */
1335static int
1336normalize_uri_path(char *path) {
1337 char *cur, *out;
1338
1339 if (path == NULL)
1340 return(-1);
1341
1342 /* Skip all initial "/" chars. We want to get to the beginning of the
1343 * first non-empty segment.
1344 */
1345 cur = path;
1346 while (cur[0] == '/')
1347 ++cur;
1348 if (cur[0] == '\0')
1349 return(0);
1350
1351 /* Keep everything we've seen so far. */
1352 out = cur;
1353
1354 /*
1355 * Analyze each segment in sequence for cases (c) and (d).
1356 */
1357 while (cur[0] != '\0') {
1358 /*
1359 * c) All occurrences of "./", where "." is a complete path segment,
1360 * are removed from the buffer string.
1361 */
1362 if ((cur[0] == '.') && (cur[1] == '/')) {
1363 cur += 2;
1364 /* '//' normalization should be done at this point too */
1365 while (cur[0] == '/')
1366 cur++;
1367 continue;
1368 }
1369
1370 /*
1371 * d) If the buffer string ends with "." as a complete path segment,
1372 * that "." is removed.
1373 */
1374 if ((cur[0] == '.') && (cur[1] == '\0'))
1375 break;
1376
1377 /* Otherwise keep the segment. */
1378 while (cur[0] != '/') {
1379 if (cur[0] == '\0')
1380 goto done_cd;
1381 (out++)[0] = (cur++)[0];
1382 }
1383 /* nomalize // */
1384 while ((cur[0] == '/') && (cur[1] == '/'))
1385 cur++;
1386
1387 (out++)[0] = (cur++)[0];
1388 }
1389 done_cd:
1390 out[0] = '\0';
1391
1392 /* Reset to the beginning of the first segment for the next sequence. */
1393 cur = path;
1394 while (cur[0] == '/')
1395 ++cur;
1396 if (cur[0] == '\0')
1397 return(0);
1398
1399 /*
1400 * Analyze each segment in sequence for cases (e) and (f).
1401 *
1402 * e) All occurrences of "<segment>/../", where <segment> is a
1403 * complete path segment not equal to "..", are removed from the
1404 * buffer string. Removal of these path segments is performed
1405 * iteratively, removing the leftmost matching pattern on each
1406 * iteration, until no matching pattern remains.
1407 *
1408 * f) If the buffer string ends with "<segment>/..", where <segment>
1409 * is a complete path segment not equal to "..", that
1410 * "<segment>/.." is removed.
1411 *
1412 * To satisfy the "iterative" clause in (e), we need to collapse the
1413 * string every time we find something that needs to be removed. Thus,
1414 * we don't need to keep two pointers into the string: we only need a
1415 * "current position" pointer.
1416 */
1417 while (1) {
1418 char *segp, *tmp;
1419
1420 /* At the beginning of each iteration of this loop, "cur" points to
1421 * the first character of the segment we want to examine.
1422 */
1423
1424 /* Find the end of the current segment. */
1425 segp = cur;
1426 while ((segp[0] != '/') && (segp[0] != '\0'))
1427 ++segp;
1428
1429 /* If this is the last segment, we're done (we need at least two
1430 * segments to meet the criteria for the (e) and (f) cases).
1431 */
1432 if (segp[0] == '\0')
1433 break;
1434
1435 /* If the first segment is "..", or if the next segment _isn't_ "..",
1436 * keep this segment and try the next one.
1437 */
1438 ++segp;
1439 if (((cur[0] == '.') && (cur[1] == '.') && (segp == cur+3))
1440 || ((segp[0] != '.') || (segp[1] != '.')
1441 || ((segp[2] != '/') && (segp[2] != '\0')))) {
1442 cur = segp;
1443 continue;
1444 }
1445
1446 /* If we get here, remove this segment and the next one and back up
1447 * to the previous segment (if there is one), to implement the
1448 * "iteratively" clause. It's pretty much impossible to back up
1449 * while maintaining two pointers into the buffer, so just compact
1450 * the whole buffer now.
1451 */
1452
1453 /* If this is the end of the buffer, we're done. */
1454 if (segp[2] == '\0') {
1455 cur[0] = '\0';
1456 break;
1457 }
1458 /* Valgrind complained, strcpy(cur, segp + 3); */
1459 /* string will overlap, do not use strcpy */
1460 tmp = cur;
1461 segp += 3;
1462 while ((*tmp++ = *segp++) != 0)
1463 ;
1464
1465 /* If there are no previous segments, then keep going from here. */
1466 segp = cur;
1467 while ((segp > path) && ((--segp)[0] == '/'))
1468 ;
1469 if (segp == path)
1470 continue;
1471
1472 /* "segp" is pointing to the end of a previous segment; find it's
1473 * start. We need to back up to the previous segment and start
1474 * over with that to handle things like "foo/bar/../..". If we
1475 * don't do this, then on the first pass we'll remove the "bar/..",
1476 * but be pointing at the second ".." so we won't realize we can also
1477 * remove the "foo/..".
1478 */
1479 cur = segp;
1480 while ((cur > path) && (cur[-1] != '/'))
1481 --cur;
1482 }
1483 out[0] = '\0';
1484
1485 /*
1486 * g) If the resulting buffer string still begins with one or more
1487 * complete path segments of "..", then the reference is
1488 * considered to be in error. Implementations may handle this
1489 * error by retaining these components in the resolved path (i.e.,
1490 * treating them as part of the final URI), by removing them from
1491 * the resolved path (i.e., discarding relative levels above the
1492 * root), or by avoiding traversal of the reference.
1493 *
1494 * We discard them from the final path.
1495 */
1496 if (path[0] == '/') {
1497 cur = path;
1498 while ((cur[0] == '/') && (cur[1] == '.') && (cur[2] == '.')
1499 && ((cur[3] == '/') || (cur[3] == '\0')))
1500 cur += 3;
1501
1502 if (cur != path) {
1503 out = path;
1504 while (cur[0] != '\0')
1505 (out++)[0] = (cur++)[0];
1506 out[0] = 0;
1507 }
1508 }
1509
1510 return(0);
1511}
1512
1513static int is_hex(char c) {
1514 if (((c >= '0') && (c <= '9')) ||
1515 ((c >= 'a') && (c <= 'f')) ||
1516 ((c >= 'A') && (c <= 'F')))
1517 return(1);
1518 return(0);
1519}
1520
1521
1522/**
1523 * uri_string_unescape:
1524 * @str: the string to unescape
1525 * @len: the length in bytes to unescape (or <= 0 to indicate full string)
1526 * @target: optional destination buffer
1527 *
1528 * Unescaping routine, but does not check that the string is an URI. The
1529 * output is a direct unsigned char translation of %XX values (no encoding)
1530 * Note that the length of the result can only be smaller or same size as
1531 * the input string.
1532 *
1533 * Returns a copy of the string, but unescaped, will return NULL only in case
1534 * of error
1535 */
1536char *
1537uri_string_unescape(const char *str, int len, char *target) {
1538 char *ret, *out;
1539 const char *in;
1540
1541 if (str == NULL)
1542 return(NULL);
1543 if (len <= 0) len = strlen(str);
1544 if (len < 0) return(NULL);
1545
1546 if (target == NULL) {
1547 ret = g_malloc(len + 1);
1548 } else
1549 ret = target;
1550 in = str;
1551 out = ret;
1552 while(len > 0) {
1553 if ((len > 2) && (*in == '%') && (is_hex(in[1])) && (is_hex(in[2]))) {
1554 in++;
1555 if ((*in >= '0') && (*in <= '9'))
1556 *out = (*in - '0');
1557 else if ((*in >= 'a') && (*in <= 'f'))
1558 *out = (*in - 'a') + 10;
1559 else if ((*in >= 'A') && (*in <= 'F'))
1560 *out = (*in - 'A') + 10;
1561 in++;
1562 if ((*in >= '0') && (*in <= '9'))
1563 *out = *out * 16 + (*in - '0');
1564 else if ((*in >= 'a') && (*in <= 'f'))
1565 *out = *out * 16 + (*in - 'a') + 10;
1566 else if ((*in >= 'A') && (*in <= 'F'))
1567 *out = *out * 16 + (*in - 'A') + 10;
1568 in++;
1569 len -= 3;
1570 out++;
1571 } else {
1572 *out++ = *in++;
1573 len--;
1574 }
1575 }
1576 *out = 0;
1577 return(ret);
1578}
1579
1580/**
1581 * uri_string_escape:
1582 * @str: string to escape
1583 * @list: exception list string of chars not to escape
1584 *
1585 * This routine escapes a string to hex, ignoring reserved characters (a-z)
1586 * and the characters in the exception list.
1587 *
1588 * Returns a new escaped string or NULL in case of error.
1589 */
1590char *
1591uri_string_escape(const char *str, const char *list) {
1592 char *ret, ch;
1593 char *temp;
1594 const char *in;
1595 int len, out;
1596
1597 if (str == NULL)
1598 return(NULL);
1599 if (str[0] == 0)
1600 return(g_strdup(str));
1601 len = strlen(str);
1602 if (!(len > 0)) return(NULL);
1603
1604 len += 20;
1605 ret = g_malloc(len);
1606 in = str;
1607 out = 0;
1608 while(*in != 0) {
1609 if (len - out <= 3) {
1610 temp = realloc2n(ret, &len);
1611 ret = temp;
1612 }
1613
1614 ch = *in;
1615
1616 if ((ch != '@') && (!IS_UNRESERVED(ch)) && (!strchr(list, ch))) {
1617 unsigned char val;
1618 ret[out++] = '%';
1619 val = ch >> 4;
1620 if (val <= 9)
1621 ret[out++] = '0' + val;
1622 else
1623 ret[out++] = 'A' + val - 0xA;
1624 val = ch & 0xF;
1625 if (val <= 9)
1626 ret[out++] = '0' + val;
1627 else
1628 ret[out++] = 'A' + val - 0xA;
1629 in++;
1630 } else {
1631 ret[out++] = *in++;
1632 }
1633
1634 }
1635 ret[out] = 0;
1636 return(ret);
1637}
1638
1639/************************************************************************
1640 * *
1641 * Public functions *
1642 * *
1643 ************************************************************************/
1644
1645/**
1646 * uri_resolve:
1647 * @URI: the URI instance found in the document
1648 * @base: the base value
1649 *
1650 * Computes he final URI of the reference done by checking that
1651 * the given URI is valid, and building the final URI using the
1652 * base URI. This is processed according to section 5.2 of the
1653 * RFC 2396
1654 *
1655 * 5.2. Resolving Relative References to Absolute Form
1656 *
1657 * Returns a new URI string (to be freed by the caller) or NULL in case
1658 * of error.
1659 */
1660char *
1661uri_resolve(const char *uri, const char *base) {
1662 char *val = NULL;
1663 int ret, len, indx, cur, out;
1664 URI *ref = NULL;
1665 URI *bas = NULL;
1666 URI *res = NULL;
1667
1668 /*
1669 * 1) The URI reference is parsed into the potential four components and
1670 * fragment identifier, as described in Section 4.3.
1671 *
1672 * NOTE that a completely empty URI is treated by modern browsers
1673 * as a reference to "." rather than as a synonym for the current
1674 * URI. Should we do that here?
1675 */
1676 if (uri == NULL)
1677 ret = -1;
1678 else {
1679 if (*uri) {
1680 ref = uri_new();
1681 if (ref == NULL)
1682 goto done;
1683 ret = uri_parse_into(ref, uri);
1684 }
1685 else
1686 ret = 0;
1687 }
1688 if (ret != 0)
1689 goto done;
1690 if ((ref != NULL) && (ref->scheme != NULL)) {
1691 /*
1692 * The URI is absolute don't modify.
1693 */
1694 val = g_strdup(uri);
1695 goto done;
1696 }
1697 if (base == NULL)
1698 ret = -1;
1699 else {
1700 bas = uri_new();
1701 if (bas == NULL)
1702 goto done;
1703 ret = uri_parse_into(bas, base);
1704 }
1705 if (ret != 0) {
1706 if (ref)
1707 val = uri_to_string(ref);
1708 goto done;
1709 }
1710 if (ref == NULL) {
1711 /*
1712 * the base fragment must be ignored
1713 */
1714 if (bas->fragment != NULL) {
1715 g_free(bas->fragment);
1716 bas->fragment = NULL;
1717 }
1718 val = uri_to_string(bas);
1719 goto done;
1720 }
1721
1722 /*
1723 * 2) If the path component is empty and the scheme, authority, and
1724 * query components are undefined, then it is a reference to the
1725 * current document and we are done. Otherwise, the reference URI's
1726 * query and fragment components are defined as found (or not found)
1727 * within the URI reference and not inherited from the base URI.
1728 *
1729 * NOTE that in modern browsers, the parsing differs from the above
1730 * in the following aspect: the query component is allowed to be
1731 * defined while still treating this as a reference to the current
1732 * document.
1733 */
1734 res = uri_new();
1735 if (res == NULL)
1736 goto done;
1737 if ((ref->scheme == NULL) && (ref->path == NULL) &&
1738 ((ref->authority == NULL) && (ref->server == NULL))) {
1739 if (bas->scheme != NULL)
1740 res->scheme = g_strdup(bas->scheme);
1741 if (bas->authority != NULL)
1742 res->authority = g_strdup(bas->authority);
1743 else if (bas->server != NULL) {
1744 res->server = g_strdup(bas->server);
1745 if (bas->user != NULL)
1746 res->user = g_strdup(bas->user);
1747 res->port = bas->port;
1748 }
1749 if (bas->path != NULL)
1750 res->path = g_strdup(bas->path);
1751 if (ref->query != NULL)
1752 res->query = g_strdup (ref->query);
1753 else if (bas->query != NULL)
1754 res->query = g_strdup(bas->query);
1755 if (ref->fragment != NULL)
1756 res->fragment = g_strdup(ref->fragment);
1757 goto step_7;
1758 }
1759
1760 /*
1761 * 3) If the scheme component is defined, indicating that the reference
1762 * starts with a scheme name, then the reference is interpreted as an
1763 * absolute URI and we are done. Otherwise, the reference URI's
1764 * scheme is inherited from the base URI's scheme component.
1765 */
1766 if (ref->scheme != NULL) {
1767 val = uri_to_string(ref);
1768 goto done;
1769 }
1770 if (bas->scheme != NULL)
1771 res->scheme = g_strdup(bas->scheme);
1772
1773 if (ref->query != NULL)
1774 res->query = g_strdup(ref->query);
1775 if (ref->fragment != NULL)
1776 res->fragment = g_strdup(ref->fragment);
1777
1778 /*
1779 * 4) If the authority component is defined, then the reference is a
1780 * network-path and we skip to step 7. Otherwise, the reference
1781 * URI's authority is inherited from the base URI's authority
1782 * component, which will also be undefined if the URI scheme does not
1783 * use an authority component.
1784 */
1785 if ((ref->authority != NULL) || (ref->server != NULL)) {
1786 if (ref->authority != NULL)
1787 res->authority = g_strdup(ref->authority);
1788 else {
1789 res->server = g_strdup(ref->server);
1790 if (ref->user != NULL)
1791 res->user = g_strdup(ref->user);
1792 res->port = ref->port;
1793 }
1794 if (ref->path != NULL)
1795 res->path = g_strdup(ref->path);
1796 goto step_7;
1797 }
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 if (bas->user != NULL)
1803 res->user = g_strdup(bas->user);
1804 res->port = bas->port;
1805 }
1806
1807 /*
1808 * 5) If the path component begins with a slash character ("/"), then
1809 * the reference is an absolute-path and we skip to step 7.
1810 */
1811 if ((ref->path != NULL) && (ref->path[0] == '/')) {
1812 res->path = g_strdup(ref->path);
1813 goto step_7;
1814 }
1815
1816
1817 /*
1818 * 6) If this step is reached, then we are resolving a relative-path
1819 * reference. The relative path needs to be merged with the base
1820 * URI's path. Although there are many ways to do this, we will
1821 * describe a simple method using a separate string buffer.
1822 *
1823 * Allocate a buffer large enough for the result string.
1824 */
1825 len = 2; /* extra / and 0 */
1826 if (ref->path != NULL)
1827 len += strlen(ref->path);
1828 if (bas->path != NULL)
1829 len += strlen(bas->path);
1830 res->path = g_malloc(len);
1831 res->path[0] = 0;
1832
1833 /*
1834 * a) All but the last segment of the base URI's path component is
1835 * copied to the buffer. In other words, any characters after the
1836 * last (right-most) slash character, if any, are excluded.
1837 */
1838 cur = 0;
1839 out = 0;
1840 if (bas->path != NULL) {
1841 while (bas->path[cur] != 0) {
1842 while ((bas->path[cur] != 0) && (bas->path[cur] != '/'))
1843 cur++;
1844 if (bas->path[cur] == 0)
1845 break;
1846
1847 cur++;
1848 while (out < cur) {
1849 res->path[out] = bas->path[out];
1850 out++;
1851 }
1852 }
1853 }
1854 res->path[out] = 0;
1855
1856 /*
1857 * b) The reference's path component is appended to the buffer
1858 * string.
1859 */
1860 if (ref->path != NULL && ref->path[0] != 0) {
1861 indx = 0;
1862 /*
1863 * Ensure the path includes a '/'
1864 */
1865 if ((out == 0) && (bas->server != NULL))
1866 res->path[out++] = '/';
1867 while (ref->path[indx] != 0) {
1868 res->path[out++] = ref->path[indx++];
1869 }
1870 }
1871 res->path[out] = 0;
1872
1873 /*
1874 * Steps c) to h) are really path normalization steps
1875 */
1876 normalize_uri_path(res->path);
1877
1878step_7:
1879
1880 /*
1881 * 7) The resulting URI components, including any inherited from the
1882 * base URI, are recombined to give the absolute form of the URI
1883 * reference.
1884 */
1885 val = uri_to_string(res);
1886
1887done:
1888 if (ref != NULL)
1889 uri_free(ref);
1890 if (bas != NULL)
1891 uri_free(bas);
1892 if (res != NULL)
1893 uri_free(res);
1894 return(val);
1895}
1896
1897/**
1898 * uri_resolve_relative:
1899 * @URI: the URI reference under consideration
1900 * @base: the base value
1901 *
1902 * Expresses the URI of the reference in terms relative to the
1903 * base. Some examples of this operation include:
1904 * base = "http://site1.com/docs/book1.html"
1905 * URI input URI returned
1906 * docs/pic1.gif pic1.gif
1907 * docs/img/pic1.gif img/pic1.gif
1908 * img/pic1.gif ../img/pic1.gif
1909 * http://site1.com/docs/pic1.gif pic1.gif
1910 * http://site2.com/docs/pic1.gif http://site2.com/docs/pic1.gif
1911 *
1912 * base = "docs/book1.html"
1913 * URI input URI returned
1914 * docs/pic1.gif pic1.gif
1915 * docs/img/pic1.gif img/pic1.gif
1916 * img/pic1.gif ../img/pic1.gif
1917 * http://site1.com/docs/pic1.gif http://site1.com/docs/pic1.gif
1918 *
1919 *
Stefan Weila93cf9d2012-11-02 08:29:53 +01001920 * Note: if the URI reference is really weird or complicated, it may be
Paolo Bonzinica0defb2012-09-24 14:42:02 +05301921 * worthwhile to first convert it into a "nice" one by calling
1922 * uri_resolve (using 'base') before calling this routine,
1923 * since this routine (for reasonable efficiency) assumes URI has
1924 * already been through some validation.
1925 *
1926 * Returns a new URI string (to be freed by the caller) or NULL in case
1927 * error.
1928 */
1929char *
1930uri_resolve_relative (const char *uri, const char * base)
1931{
1932 char *val = NULL;
1933 int ret;
1934 int ix;
1935 int pos = 0;
1936 int nbslash = 0;
1937 int len;
1938 URI *ref = NULL;
1939 URI *bas = NULL;
1940 char *bptr, *uptr, *vptr;
1941 int remove_path = 0;
1942
1943 if ((uri == NULL) || (*uri == 0))
1944 return NULL;
1945
1946 /*
1947 * First parse URI into a standard form
1948 */
1949 ref = uri_new ();
1950 if (ref == NULL)
1951 return NULL;
1952 /* If URI not already in "relative" form */
1953 if (uri[0] != '.') {
1954 ret = uri_parse_into (ref, uri);
1955 if (ret != 0)
1956 goto done; /* Error in URI, return NULL */
1957 } else
1958 ref->path = g_strdup(uri);
1959
1960 /*
1961 * Next parse base into the same standard form
1962 */
1963 if ((base == NULL) || (*base == 0)) {
1964 val = g_strdup (uri);
1965 goto done;
1966 }
1967 bas = uri_new ();
1968 if (bas == NULL)
1969 goto done;
1970 if (base[0] != '.') {
1971 ret = uri_parse_into (bas, base);
1972 if (ret != 0)
1973 goto done; /* Error in base, return NULL */
1974 } else
1975 bas->path = g_strdup(base);
1976
1977 /*
1978 * If the scheme / server on the URI differs from the base,
1979 * just return the URI
1980 */
1981 if ((ref->scheme != NULL) &&
1982 ((bas->scheme == NULL) ||
1983 (strcmp (bas->scheme, ref->scheme)) ||
1984 (strcmp (bas->server, ref->server)))) {
1985 val = g_strdup (uri);
1986 goto done;
1987 }
1988 if (!strcmp(bas->path, ref->path)) {
1989 val = g_strdup("");
1990 goto done;
1991 }
1992 if (bas->path == NULL) {
1993 val = g_strdup(ref->path);
1994 goto done;
1995 }
1996 if (ref->path == NULL) {
1997 ref->path = (char *) "/";
1998 remove_path = 1;
1999 }
2000
2001 /*
2002 * At this point (at last!) we can compare the two paths
2003 *
2004 * First we take care of the special case where either of the
2005 * two path components may be missing (bug 316224)
2006 */
2007 if (bas->path == NULL) {
2008 if (ref->path != NULL) {
2009 uptr = ref->path;
2010 if (*uptr == '/')
2011 uptr++;
2012 /* exception characters from uri_to_string */
2013 val = uri_string_escape(uptr, "/;&=+$,");
2014 }
2015 goto done;
2016 }
2017 bptr = bas->path;
2018 if (ref->path == NULL) {
2019 for (ix = 0; bptr[ix] != 0; ix++) {
2020 if (bptr[ix] == '/')
2021 nbslash++;
2022 }
2023 uptr = NULL;
2024 len = 1; /* this is for a string terminator only */
2025 } else {
2026 /*
2027 * Next we compare the two strings and find where they first differ
2028 */
2029 if ((ref->path[pos] == '.') && (ref->path[pos+1] == '/'))
2030 pos += 2;
2031 if ((*bptr == '.') && (bptr[1] == '/'))
2032 bptr += 2;
2033 else if ((*bptr == '/') && (ref->path[pos] != '/'))
2034 bptr++;
2035 while ((bptr[pos] == ref->path[pos]) && (bptr[pos] != 0))
2036 pos++;
2037
2038 if (bptr[pos] == ref->path[pos]) {
2039 val = g_strdup("");
2040 goto done; /* (I can't imagine why anyone would do this) */
2041 }
2042
2043 /*
2044 * In URI, "back up" to the last '/' encountered. This will be the
2045 * beginning of the "unique" suffix of URI
2046 */
2047 ix = pos;
2048 if ((ref->path[ix] == '/') && (ix > 0))
2049 ix--;
2050 else if ((ref->path[ix] == 0) && (ix > 1) && (ref->path[ix - 1] == '/'))
2051 ix -= 2;
2052 for (; ix > 0; ix--) {
2053 if (ref->path[ix] == '/')
2054 break;
2055 }
2056 if (ix == 0) {
2057 uptr = ref->path;
2058 } else {
2059 ix++;
2060 uptr = &ref->path[ix];
2061 }
2062
2063 /*
2064 * In base, count the number of '/' from the differing point
2065 */
2066 if (bptr[pos] != ref->path[pos]) {/* check for trivial URI == base */
2067 for (; bptr[ix] != 0; ix++) {
2068 if (bptr[ix] == '/')
2069 nbslash++;
2070 }
2071 }
2072 len = strlen (uptr) + 1;
2073 }
2074
2075 if (nbslash == 0) {
2076 if (uptr != NULL)
2077 /* exception characters from uri_to_string */
2078 val = uri_string_escape(uptr, "/;&=+$,");
2079 goto done;
2080 }
2081
2082 /*
2083 * Allocate just enough space for the returned string -
2084 * length of the remainder of the URI, plus enough space
2085 * for the "../" groups, plus one for the terminator
2086 */
2087 val = g_malloc (len + 3 * nbslash);
2088 vptr = val;
2089 /*
2090 * Put in as many "../" as needed
2091 */
2092 for (; nbslash>0; nbslash--) {
2093 *vptr++ = '.';
2094 *vptr++ = '.';
2095 *vptr++ = '/';
2096 }
2097 /*
2098 * Finish up with the end of the URI
2099 */
2100 if (uptr != NULL) {
2101 if ((vptr > val) && (len > 0) &&
2102 (uptr[0] == '/') && (vptr[-1] == '/')) {
2103 memcpy (vptr, uptr + 1, len - 1);
2104 vptr[len - 2] = 0;
2105 } else {
2106 memcpy (vptr, uptr, len);
2107 vptr[len - 1] = 0;
2108 }
2109 } else {
2110 vptr[len - 1] = 0;
2111 }
2112
2113 /* escape the freshly-built path */
2114 vptr = val;
2115 /* exception characters from uri_to_string */
2116 val = uri_string_escape(vptr, "/;&=+$,");
2117 g_free(vptr);
2118
2119done:
2120 /*
2121 * Free the working variables
2122 */
2123 if (remove_path != 0)
2124 ref->path = NULL;
2125 if (ref != NULL)
2126 uri_free (ref);
2127 if (bas != NULL)
2128 uri_free (bas);
2129
2130 return val;
2131}
2132
2133/*
2134 * Utility functions to help parse and assemble query strings.
2135 */
2136
2137struct QueryParams *
2138query_params_new (int init_alloc)
2139{
2140 struct QueryParams *ps;
2141
2142 if (init_alloc <= 0) init_alloc = 1;
2143
2144 ps = g_new(QueryParams, 1);
2145 ps->n = 0;
2146 ps->alloc = init_alloc;
2147 ps->p = g_new(QueryParam, ps->alloc);
2148
2149 return ps;
2150}
2151
2152/* Ensure there is space to store at least one more parameter
2153 * at the end of the set.
2154 */
2155static int
2156query_params_append (struct QueryParams *ps,
2157 const char *name, const char *value)
2158{
2159 if (ps->n >= ps->alloc) {
2160 ps->p = g_renew(QueryParam, ps->p, ps->alloc * 2);
2161 ps->alloc *= 2;
2162 }
2163
2164 ps->p[ps->n].name = g_strdup(name);
Dong Xu Wang7f303ad2013-05-09 15:53:49 +08002165 ps->p[ps->n].value = g_strdup(value);
Paolo Bonzinica0defb2012-09-24 14:42:02 +05302166 ps->p[ps->n].ignore = 0;
2167 ps->n++;
2168
2169 return 0;
2170}
2171
2172void
2173query_params_free (struct QueryParams *ps)
2174{
2175 int i;
2176
2177 for (i = 0; i < ps->n; ++i) {
2178 g_free (ps->p[i].name);
2179 g_free (ps->p[i].value);
2180 }
2181 g_free (ps->p);
2182 g_free (ps);
2183}
2184
2185struct QueryParams *
2186query_params_parse (const char *query)
2187{
2188 struct QueryParams *ps;
2189 const char *end, *eq;
2190
2191 ps = query_params_new (0);
2192 if (!query || query[0] == '\0') return ps;
2193
2194 while (*query) {
2195 char *name = NULL, *value = NULL;
2196
2197 /* Find the next separator, or end of the string. */
2198 end = strchr (query, '&');
2199 if (!end)
2200 end = strchr (query, ';');
2201 if (!end)
2202 end = query + strlen (query);
2203
2204 /* Find the first '=' character between here and end. */
2205 eq = strchr (query, '=');
2206 if (eq && eq >= end) eq = NULL;
2207
2208 /* Empty section (eg. "&&"). */
2209 if (end == query)
2210 goto next;
2211
2212 /* If there is no '=' character, then we have just "name"
2213 * and consistent with CGI.pm we assume value is "".
2214 */
2215 else if (!eq) {
2216 name = uri_string_unescape (query, end - query, NULL);
2217 value = NULL;
2218 }
2219 /* Or if we have "name=" here (works around annoying
2220 * problem when calling uri_string_unescape with len = 0).
2221 */
2222 else if (eq+1 == end) {
2223 name = uri_string_unescape (query, eq - query, NULL);
2224 value = g_new0(char, 1);
2225 }
2226 /* If the '=' character is at the beginning then we have
2227 * "=value" and consistent with CGI.pm we _ignore_ this.
2228 */
2229 else if (query == eq)
2230 goto next;
2231
2232 /* Otherwise it's "name=value". */
2233 else {
2234 name = uri_string_unescape (query, eq - query, NULL);
2235 value = uri_string_unescape (eq+1, end - (eq+1), NULL);
2236 }
2237
2238 /* Append to the parameter set. */
2239 query_params_append (ps, name, value);
2240 g_free(name);
2241 g_free(value);
2242
2243 next:
2244 query = end;
2245 if (*query) query ++; /* skip '&' separator */
2246 }
2247
2248 return ps;
2249}