projects / m6w6 / ext-http / blob
? search:
summary | shortlog | log | commit | commitdiff | tree | github
history | raw | HEAD
Merge branch 'R_2_3'
[m6w6/ext-http] / php_http_url.c
1 /*
2 +--------------------------------------------------------------------+
3 | PECL :: http |
4 +--------------------------------------------------------------------+
5 | Redistribution and use in source and binary forms, with or without |
6 | modification, are permitted provided that the conditions mentioned |
7 | in the accompanying LICENSE file are met. |
8 +--------------------------------------------------------------------+
9 | Copyright (c) 2004-2014, Michael Wallner <mike@php.net> |
10 +--------------------------------------------------------------------+
11 */
12
13 #include "php_http_api.h"
14
15 #ifdef PHP_HTTP_HAVE_IDN
16 # include <idna.h>
17 #endif
18
19 #ifdef PHP_HTTP_HAVE_WCHAR
20 # include <wchar.h>
21 # include <wctype.h>
22 #endif
23
24 #ifdef HAVE_ARPA_INET_H
25 # include <arpa/inet.h>
26 #endif
27
28 #include "php_http_utf8.h"
29
30 static inline char *localhostname(void)
31 {
32 char hostname[1024] = {0};
33
34 #ifdef PHP_WIN32
35 if (SUCCESS == gethostname(hostname, lenof(hostname))) {
36 return estrdup(hostname);
37 }
38 #elif defined(HAVE_GETHOSTNAME)
39 if (SUCCESS == gethostname(hostname, lenof(hostname))) {
40 # if defined(HAVE_GETDOMAINNAME)
41 size_t hlen = strlen(hostname);
42 if (hlen <= lenof(hostname) - lenof("(none)")) {
43 hostname[hlen++] = '.';
44 if (SUCCESS == getdomainname(&hostname[hlen], lenof(hostname) - hlen)) {
45 if (!strcmp(&hostname[hlen], "(none)")) {
46 hostname[hlen - 1] = '\0';
47 }
48 return estrdup(hostname);
49 }
50 }
51 # endif
52 if (strcmp(hostname, "(none)")) {
53 return estrdup(hostname);
54 }
55 }
56 #endif
57 return estrndup("localhost", lenof("localhost"));
58 }
59
60 #define url(buf) ((php_http_url_t *) (buf).data)
61
62 static php_http_url_t *php_http_url_from_env(TSRMLS_D)
63 {
64 zval *https, *zhost, *zport;
65 long port;
66 php_http_buffer_t buf;
67
68 php_http_buffer_init_ex(&buf, MAX(PHP_HTTP_BUFFER_DEFAULT_SIZE, sizeof(php_http_url_t)<<2), PHP_HTTP_BUFFER_INIT_PREALLOC);
69 php_http_buffer_account(&buf, sizeof(php_http_url_t));
70 memset(buf.data, 0, buf.used);
71
72 /* scheme */
73 url(buf)->scheme = &buf.data[buf.used];
74 https = php_http_env_get_server_var(ZEND_STRL("HTTPS"), 1 TSRMLS_CC);
75 if (https && !strcasecmp(Z_STRVAL_P(https), "ON")) {
76 php_http_buffer_append(&buf, "https", sizeof("https"));
77 } else {
78 php_http_buffer_append(&buf, "http", sizeof("http"));
79 }
80
81 /* host */
82 url(buf)->host = &buf.data[buf.used];
83 if ((((zhost = php_http_env_get_server_var(ZEND_STRL("HTTP_HOST"), 1 TSRMLS_CC)) ||
84 (zhost = php_http_env_get_server_var(ZEND_STRL("SERVER_NAME"), 1 TSRMLS_CC)) ||
85 (zhost = php_http_env_get_server_var(ZEND_STRL("SERVER_ADDR"), 1 TSRMLS_CC)))) && Z_STRLEN_P(zhost)) {
86 size_t stop_at = strspn(Z_STRVAL_P(zhost), "0123456789abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ-.");
87
88 php_http_buffer_append(&buf, Z_STRVAL_P(zhost), stop_at);
89 php_http_buffer_append(&buf, "", 1);
90 } else {
91 char *host_str = localhostname();
92
93 php_http_buffer_append(&buf, host_str, strlen(host_str) + 1);
94 efree(host_str);
95 }
96
97 /* port */
98 zport = php_http_env_get_server_var(ZEND_STRL("SERVER_PORT"), 1 TSRMLS_CC);
99 if (zport && IS_LONG == is_numeric_string(Z_STRVAL_P(zport), Z_STRLEN_P(zport), &port, NULL, 0)) {
100 url(buf)->port = port;
101 }
102
103 /* path */
104 if (SG(request_info).request_uri && SG(request_info).request_uri[0]) {
105 const char *q = strchr(SG(request_info).request_uri, '?');
106
107 url(buf)->path = &buf.data[buf.used];
108
109 if (q) {
110 php_http_buffer_append(&buf, SG(request_info).request_uri, q - SG(request_info).request_uri);
111 php_http_buffer_append(&buf, "", 1);
112 } else {
113 php_http_buffer_append(&buf, SG(request_info).request_uri, strlen(SG(request_info).request_uri) + 1);
114 }
115 }
116
117 /* query */
118 if (SG(request_info).query_string && SG(request_info).query_string[0]) {
119 url(buf)->query = &buf.data[buf.used];
120 php_http_buffer_append(&buf, SG(request_info).query_string, strlen(SG(request_info).query_string) + 1);
121 }
122
123 return url(buf);
124 }
125
126 #define url_isset(u,n) \
127 ((u)&&(u)->n)
128 #define url_append(buf, append) do { \
129 char *_ptr = (buf)->data; \
130 php_http_url_t *_url = (php_http_url_t *) _ptr, _mem = *_url; \
131 append; \
132 /* relocate */ \
133 if (_ptr != (buf)->data) { \
134 ptrdiff_t diff = (buf)->data - _ptr; \
135 _url = (php_http_url_t *) (buf)->data; \
136 if (_mem.scheme) _url->scheme += diff; \
137 if (_mem.user) _url->user += diff; \
138 if (_mem.pass) _url->pass += diff; \
139 if (_mem.host) _url->host += diff; \
140 if (_mem.path) _url->path += diff; \
141 if (_mem.query) _url->query += diff; \
142 if (_mem.fragment) _url->fragment += diff; \
143 } \
144 } while (0)
145 #define url_copy(n) do { \
146 if (url_isset(new_url, n)) { \
147 url(buf)->n = &buf.data[buf.used]; \
148 url_append(&buf, php_http_buffer_append(&buf, new_url->n, strlen(new_url->n) + 1)); \
149 } else if (url_isset(old_url, n)) { \
150 url(buf)->n = &buf.data[buf.used]; \
151 url_append(&buf, php_http_buffer_append(&buf, old_url->n, strlen(old_url->n) + 1)); \
152 } \
153 } while (0)
154
155 php_http_url_t *php_http_url_mod(const php_http_url_t *old_url, const php_http_url_t *new_url, unsigned flags TSRMLS_DC)
156 {
157 php_http_url_t *tmp_url = NULL;
158 php_http_buffer_t buf;
159
160 php_http_buffer_init_ex(&buf, MAX(PHP_HTTP_BUFFER_DEFAULT_SIZE, sizeof(php_http_url_t)<<2), PHP_HTTP_BUFFER_INIT_PREALLOC);
161 php_http_buffer_account(&buf, sizeof(php_http_url_t));
162 memset(buf.data, 0, buf.used);
163
164 /* set from env if requested */
165 if (flags & PHP_HTTP_URL_FROM_ENV) {
166 php_http_url_t *env_url = php_http_url_from_env(TSRMLS_C);
167
168 old_url = tmp_url = php_http_url_mod(env_url, old_url, flags ^ PHP_HTTP_URL_FROM_ENV TSRMLS_CC);
169 php_http_url_free(&env_url);
170 }
171
172 url_copy(scheme);
173
174 if (!(flags & PHP_HTTP_URL_STRIP_USER)) {
175 url_copy(user);
176 }
177
178 if (!(flags & PHP_HTTP_URL_STRIP_PASS)) {
179 url_copy(pass);
180 }
181
182 url_copy(host);
183
184 if (!(flags & PHP_HTTP_URL_STRIP_PORT)) {
185 url(buf)->port = url_isset(new_url, port) ? new_url->port : ((old_url) ? old_url->port : 0);
186 }
187
188 if (!(flags & PHP_HTTP_URL_STRIP_PATH)) {
189 if ((flags & PHP_HTTP_URL_JOIN_PATH) && url_isset(old_url, path) && url_isset(new_url, path) && *new_url->path != '/') {
190 size_t old_path_len = strlen(old_url->path), new_path_len = strlen(new_url->path);
191 char *path = ecalloc(1, old_path_len + new_path_len + 1 + 1);
192
193 strcat(path, old_url->path);
194 if (path[old_path_len - 1] != '/') {
195 php_dirname(path, old_path_len);
196 strcat(path, "/");
197 }
198 strcat(path, new_url->path);
199
200 url(buf)->path = &buf.data[buf.used];
201 if (path[0] != '/') {
202 url_append(&buf, php_http_buffer_append(&buf, "/", 1));
203 }
204 url_append(&buf, php_http_buffer_append(&buf, path, strlen(path) + 1));
205 efree(path);
206 } else {
207 const char *path = NULL;
208
209 if (url_isset(new_url, path)) {
210 path = new_url->path;
211 } else if (url_isset(old_url, path)) {
212 path = old_url->path;
213 }
214
215 if (path) {
216 url(buf)->path = &buf.data[buf.used];
217
218 url_append(&buf, php_http_buffer_append(&buf, path, strlen(path) + 1));
219 }
220
221
222 }
223 }
224
225 if (!(flags & PHP_HTTP_URL_STRIP_QUERY)) {
226 if ((flags & PHP_HTTP_URL_JOIN_QUERY) && url_isset(new_url, query) && url_isset(old_url, query)) {
227 zval qarr, qstr;
228
229 INIT_PZVAL(&qstr);
230 INIT_PZVAL(&qarr);
231 array_init(&qarr);
232
233 ZVAL_STRING(&qstr, old_url->query, 0);
234 php_http_querystring_update(&qarr, &qstr, NULL TSRMLS_CC);
235 ZVAL_STRING(&qstr, new_url->query, 0);
236 php_http_querystring_update(&qarr, &qstr, NULL TSRMLS_CC);
237
238 ZVAL_NULL(&qstr);
239 php_http_querystring_update(&qarr, NULL, &qstr TSRMLS_CC);
240
241 url(buf)->query = &buf.data[buf.used];
242 url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL(qstr), Z_STRLEN(qstr) + 1));
243
244 zval_dtor(&qstr);
245 zval_dtor(&qarr);
246 } else {
247 url_copy(query);
248 }
249 }
250
251 if (!(flags & PHP_HTTP_URL_STRIP_FRAGMENT)) {
252 url_copy(fragment);
253 }
254
255 /* done with copy & combine & strip */
256
257 if (flags & PHP_HTTP_URL_FROM_ENV) {
258 /* free old_url we tainted above */
259 php_http_url_free(&tmp_url);
260 }
261
262 /* replace directory references if path is not a single slash */
263 if ((flags & PHP_HTTP_URL_SANITIZE_PATH)
264 && url(buf)->path[0] && url(buf)->path[1]) {
265 char *ptr, *end = url(buf)->path + strlen(url(buf)->path) + 1;
266
267 for (ptr = strchr(url(buf)->path, '/'); ptr; ptr = strchr(ptr, '/')) {
268 switch (ptr[1]) {
269 case '/':
270 memmove(&ptr[1], &ptr[2], end - &ptr[2]);
271 break;
272
273 case '.':
274 switch (ptr[2]) {
275 case '\0':
276 ptr[1] = '\0';
277 break;
278
279 case '/':
280 memmove(&ptr[1], &ptr[3], end - &ptr[3]);
281 break;
282
283 case '.':
284 if (ptr[3] == '/') {
285 char *pos = &ptr[4];
286 while (ptr != url(buf)->path) {
287 if (*--ptr == '/') {
288 break;
289 }
290 }
291 memmove(&ptr[1], pos, end - pos);
292 break;
293 } else if (!ptr[3]) {
294 /* .. at the end */
295 ptr[1] = '\0';
296 }
297 /* no break */
298
299 default:
300 /* something else */
301 ++ptr;
302 break;
303 }
304 break;
305
306 default:
307 ++ptr;
308 break;
309 }
310 }
311 }
312 /* unset default ports */
313 if (url(buf)->port) {
314 if ( ((url(buf)->port == 80) && url(buf)->scheme && !strcmp(url(buf)->scheme, "http"))
315 || ((url(buf)->port ==443) && url(buf)->scheme && !strcmp(url(buf)->scheme, "https"))
316 ) {
317 url(buf)->port = 0;
318 }
319 }
320
321 return url(buf);
322 }
323
324 char *php_http_url_to_string(const php_http_url_t *url, char **url_str, size_t *url_len, zend_bool persistent)
325 {
326 php_http_buffer_t buf;
327
328 php_http_buffer_init_ex(&buf, PHP_HTTP_BUFFER_DEFAULT_SIZE, persistent ?
329 PHP_HTTP_BUFFER_INIT_PERSISTENT : 0);
330
331 if (url->scheme && *url->scheme) {
332 php_http_buffer_appendl(&buf, url->scheme);
333 php_http_buffer_appends(&buf, "://");
334 } else if ((url->user && *url->user) || (url->host && *url->host)) {
335 php_http_buffer_appends(&buf, "//");
336 }
337
338 if (url->user && *url->user) {
339 php_http_buffer_appendl(&buf, url->user);
340 if (url->pass && *url->pass) {
341 php_http_buffer_appends(&buf, ":");
342 php_http_buffer_appendl(&buf, url->pass);
343 }
344 php_http_buffer_appends(&buf, "@");
345 }
346
347 if (url->host && *url->host) {
348 php_http_buffer_appendl(&buf, url->host);
349 if (url->port) {
350 php_http_buffer_appendf(&buf, ":%hu", url->port);
351 }
352 }
353
354 if (url->path && *url->path) {
355 if (*url->path != '/') {
356 php_http_buffer_appends(&buf, "/");
357 }
358 php_http_buffer_appendl(&buf, url->path);
359 } else if (buf.used) {
360 php_http_buffer_appends(&buf, "/");
361 }
362
363 if (url->query && *url->query) {
364 php_http_buffer_appends(&buf, "?");
365 php_http_buffer_appendl(&buf, url->query);
366 }
367
368 if (url->fragment && *url->fragment) {
369 php_http_buffer_appends(&buf, "#");
370 php_http_buffer_appendl(&buf, url->fragment);
371 }
372
373 php_http_buffer_shrink(&buf);
374 php_http_buffer_fix(&buf);
375
376 if (url_len) {
377 *url_len = buf.used;
378 }
379
380 if (url_str) {
381 *url_str = buf.data;
382 }
383
384 return buf.data;
385 }
386
387 char *php_http_url_authority_to_string(const php_http_url_t *url, char **url_str, size_t *url_len)
388 {
389 php_http_buffer_t buf;
390
391 php_http_buffer_init(&buf);
392
393 if (url->user && *url->user) {
394 php_http_buffer_appendl(&buf, url->user);
395 if (url->pass && *url->pass) {
396 php_http_buffer_appends(&buf, ":");
397 php_http_buffer_appendl(&buf, url->pass);
398 }
399 php_http_buffer_appends(&buf, "@");
400 }
401
402 if (url->host && *url->host) {
403 php_http_buffer_appendl(&buf, url->host);
404 if (url->port) {
405 php_http_buffer_appendf(&buf, ":%hu", url->port);
406 }
407 }
408
409 php_http_buffer_shrink(&buf);
410 php_http_buffer_fix(&buf);
411
412 if (url_len) {
413 *url_len = buf.used;
414 }
415
416 if (url_str) {
417 *url_str = buf.data;
418 }
419
420 return buf.data;
421 }
422
423 php_http_url_t *php_http_url_from_zval(zval *value, unsigned flags TSRMLS_DC)
424 {
425 zval *zcpy;
426 php_http_url_t *purl;
427
428 switch (Z_TYPE_P(value)) {
429 case IS_ARRAY:
430 case IS_OBJECT:
431 purl = php_http_url_from_struct(HASH_OF(value));
432 break;
433
434 default:
435 zcpy = php_http_ztyp(IS_STRING, value);
436 purl = php_http_url_parse(Z_STRVAL_P(zcpy), Z_STRLEN_P(zcpy), flags TSRMLS_CC);
437 zval_ptr_dtor(&zcpy);
438 }
439
440 return purl;
441 }
442
443 php_http_url_t *php_http_url_from_struct(HashTable *ht)
444 {
445 zval **e;
446 php_http_buffer_t buf;
447
448 php_http_buffer_init_ex(&buf, MAX(PHP_HTTP_BUFFER_DEFAULT_SIZE, sizeof(php_http_url_t)<<2), PHP_HTTP_BUFFER_INIT_PREALLOC);
449 php_http_buffer_account(&buf, sizeof(php_http_url_t));
450 memset(buf.data, 0, buf.used);
451
452 if (SUCCESS == zend_hash_find(ht, "scheme", sizeof("scheme"), (void *) &e)) {
453 zval *cpy = php_http_ztyp(IS_STRING, *e);
454 url(buf)->scheme = &buf.data[buf.used];
455 url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
456 zval_ptr_dtor(&cpy);
457 }
458 if (SUCCESS == zend_hash_find(ht, "user", sizeof("user"), (void *) &e)) {
459 zval *cpy = php_http_ztyp(IS_STRING, *e);
460 url(buf)->user = &buf.data[buf.used];
461 url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
462 zval_ptr_dtor(&cpy);
463 }
464 if (SUCCESS == zend_hash_find(ht, "pass", sizeof("pass"), (void *) &e)) {
465 zval *cpy = php_http_ztyp(IS_STRING, *e);
466 url(buf)->pass = &buf.data[buf.used];
467 url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
468 zval_ptr_dtor(&cpy);
469 }
470 if (SUCCESS == zend_hash_find(ht, "host", sizeof("host"), (void *) &e)) {
471 zval *cpy = php_http_ztyp(IS_STRING, *e);
472 url(buf)->host = &buf.data[buf.used];
473 url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
474 zval_ptr_dtor(&cpy);
475 }
476 if (SUCCESS == zend_hash_find(ht, "port", sizeof("port"), (void *) &e)) {
477 zval *cpy = php_http_ztyp(IS_LONG, *e);
478 url(buf)->port = (unsigned short) Z_LVAL_P(cpy);
479 zval_ptr_dtor(&cpy);
480 }
481 if (SUCCESS == zend_hash_find(ht, "path", sizeof("path"), (void *) &e)) {
482 zval *cpy = php_http_ztyp(IS_STRING, *e);
483 url(buf)->path = &buf.data[buf.used];
484 url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
485 zval_ptr_dtor(&cpy);
486 }
487 if (SUCCESS == zend_hash_find(ht, "query", sizeof("query"), (void *) &e)) {
488 zval *cpy = php_http_ztyp(IS_STRING, *e);
489 url(buf)->query = &buf.data[buf.used];
490 url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
491 zval_ptr_dtor(&cpy);
492 }
493 if (SUCCESS == zend_hash_find(ht, "fragment", sizeof("fragment"), (void *) &e)) {
494 zval *cpy = php_http_ztyp(IS_STRING, *e);
495 url(buf)->fragment = &buf.data[buf.used];
496 url_append(&buf, php_http_buffer_append(&buf, Z_STRVAL_P(cpy), Z_STRLEN_P(cpy) + 1));
497 zval_ptr_dtor(&cpy);
498 }
499
500 return url(buf);
501 }
502
503 HashTable *php_http_url_to_struct(const php_http_url_t *url, zval *strct TSRMLS_DC)
504 {
505 zval arr;
506
507 if (strct) {
508 switch (Z_TYPE_P(strct)) {
509 default:
510 zval_dtor(strct);
511 array_init(strct);
512 /* no break */
513 case IS_ARRAY:
514 case IS_OBJECT:
515 INIT_PZVAL_ARRAY((&arr), HASH_OF(strct));
516 break;
517 }
518 } else {
519 INIT_PZVAL(&arr);
520 array_init(&arr);
521 }
522
523 if (url) {
524 if (url->scheme) {
525 add_assoc_string(&arr, "scheme", url->scheme, 1);
526 }
527 if (url->user) {
528 add_assoc_string(&arr, "user", url->user, 1);
529 }
530 if (url->pass) {
531 add_assoc_string(&arr, "pass", url->pass, 1);
532 }
533 if (url->host) {
534 add_assoc_string(&arr, "host", url->host, 1);
535 }
536 if (url->port) {
537 add_assoc_long(&arr, "port", (long) url->port);
538 }
539 if (url->path) {
540 add_assoc_string(&arr, "path", url->path, 1);
541 }
542 if (url->query) {
543 add_assoc_string(&arr, "query", url->query, 1);
544 }
545 if (url->fragment) {
546 add_assoc_string(&arr, "fragment", url->fragment, 1);
547 }
548 }
549
550 return Z_ARRVAL(arr);
551 }
552
553 STATUS php_http_url_encode_hash(HashTable *hash, const char *pre_encoded_str, size_t pre_encoded_len, char **encoded_str, size_t *encoded_len TSRMLS_DC)
554 {
555 const char *arg_sep_str;
556 size_t arg_sep_len;
557 php_http_buffer_t *qstr = php_http_buffer_new();
558
559 php_http_url_argsep(&arg_sep_str, &arg_sep_len TSRMLS_CC);
560
561 if (SUCCESS != php_http_url_encode_hash_ex(hash, qstr, arg_sep_str, arg_sep_len, "=", 1, pre_encoded_str, pre_encoded_len TSRMLS_CC)) {
562 php_http_buffer_free(&qstr);
563 return FAILURE;
564 }
565
566 php_http_buffer_data(qstr, encoded_str, encoded_len);
567 php_http_buffer_free(&qstr);
568
569 return SUCCESS;
570 }
571
572 STATUS php_http_url_encode_hash_ex(HashTable *hash, php_http_buffer_t *qstr, const char *arg_sep_str, size_t arg_sep_len, const char *val_sep_str, size_t val_sep_len, const char *pre_encoded_str, size_t pre_encoded_len TSRMLS_DC)
573 {
574 if (pre_encoded_len && pre_encoded_str) {
575 php_http_buffer_append(qstr, pre_encoded_str, pre_encoded_len);
576 }
577
578 if (!php_http_params_to_string(qstr, hash, arg_sep_str, arg_sep_len, "", 0, val_sep_str, val_sep_len, PHP_HTTP_PARAMS_QUERY TSRMLS_CC)) {
579 return FAILURE;
580 }
581
582 return SUCCESS;
583 }
584
585 struct parse_state {
586 php_http_url_t url;
587 #ifdef ZTS
588 void ***ts;
589 #endif
590 const char *ptr;
591 const char *end;
592 size_t maxlen;
593 off_t offset;
594 unsigned flags;
595 char buffer[1]; /* last member */
596 };
597
598 void php_http_url_free(php_http_url_t **url)
599 {
600 if (*url) {
601 efree(*url);
602 *url = NULL;
603 }
604 }
605
606 php_http_url_t *php_http_url_copy(const php_http_url_t *url, zend_bool persistent)
607 {
608 php_http_url_t *cpy;
609 const char *end = NULL, *url_ptr = (const char *) url;
610 char *cpy_ptr;
611
612 end = MAX(url->scheme, end);
613 end = MAX(url->pass, end);
614 end = MAX(url->user, end);
615 end = MAX(url->host, end);
616 end = MAX(url->path, end);
617 end = MAX(url->query, end);
618 end = MAX(url->fragment, end);
619
620 if (end) {
621 end += strlen(end) + 1;
622 cpy_ptr = pecalloc(1, end - url_ptr, persistent);
623 cpy = (php_http_url_t *) cpy_ptr;
624
625 memcpy(cpy_ptr + sizeof(*cpy), url_ptr + sizeof(*url), end - url_ptr - sizeof(*url));
626
627 cpy->scheme = url->scheme ? cpy_ptr + (url->scheme - url_ptr) : NULL;
628 cpy->pass = url->pass ? cpy_ptr + (url->pass - url_ptr) : NULL;
629 cpy->user = url->user ? cpy_ptr + (url->user - url_ptr) : NULL;
630 cpy->host = url->host ? cpy_ptr + (url->host - url_ptr) : NULL;
631 cpy->path = url->path ? cpy_ptr + (url->path - url_ptr) : NULL;
632 cpy->query = url->query ? cpy_ptr + (url->query - url_ptr) : NULL;
633 cpy->fragment = url->fragment ? cpy_ptr + (url->fragment - url_ptr) : NULL;
634 } else {
635 cpy = ecalloc(1, sizeof(*url));
636 }
637
638 cpy->port = url->port;
639
640 return cpy;
641 }
642
643 static size_t parse_mb_utf8(unsigned *wc, const char *ptr, const char *end)
644 {
645 unsigned wchar;
646 size_t consumed = utf8towc(&wchar, (const unsigned char *) ptr, end - ptr);
647
648 if (!consumed || consumed == (size_t) -1) {
649 return 0;
650 }
651
652 if (wc) {
653 *wc = wchar;
654 }
655 return consumed;
656 }
657
658 #ifdef PHP_HTTP_HAVE_WCHAR
659 static size_t parse_mb_loc(unsigned *wc, const char *ptr, const char *end)
660 {
661 wchar_t wchar;
662 size_t consumed = 0;
663 #if defined(HAVE_MBRTOWC)
664 mbstate_t ps;
665
666 memset(&ps, 0, sizeof(ps));
667 consumed = mbrtowc(&wchar, ptr, end - ptr, &ps);
668 #elif defined(HAVE_MBTOWC)
669 consumed = mbtowc(&wchar, ptr, end - ptr);
670 #endif
671
672 if (!consumed || consumed == (size_t) -1) {
673 return 0;
674 }
675
676 if (wc) {
677 *wc = wchar;
678 }
679 return consumed;
680 }
681 #endif
682
683 typedef enum parse_mb_what {
684 PARSE_SCHEME,
685 PARSE_USERINFO,
686 PARSE_HOSTINFO,
687 PARSE_PATH,
688 PARSE_QUERY,
689 PARSE_FRAGMENT
690 } parse_mb_what_t;
691
692 static const char * const parse_what[] = {
693 "scheme",
694 "userinfo",
695 "hostinfo",
696 "path",
697 "query",
698 "fragment"
699 };
700
701 static const char parse_xdigits[] = "0123456789ABCDEF";
702
703 static size_t parse_mb(struct parse_state *state, parse_mb_what_t what, const char *ptr, const char *end, const char *begin, zend_bool silent)
704 {
705 unsigned wchar;
706 size_t consumed = 0;
707
708 if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
709 consumed = parse_mb_utf8(&wchar, ptr, end);
710 }
711 #ifdef PHP_HTTP_HAVE_WCHAR
712 else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
713 consumed = parse_mb_loc(&wchar, ptr, end);
714 }
715 #endif
716
717 while (consumed) {
718 if (!(state->flags & PHP_HTTP_URL_PARSE_TOPCT) || what == PARSE_HOSTINFO || what == PARSE_SCHEME) {
719 if (what == PARSE_HOSTINFO && (state->flags & PHP_HTTP_URL_PARSE_TOIDN)) {
720 /* idna */
721 } else if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
722 if (!isualnum(wchar)) {
723 break;
724 }
725 #ifdef PHP_HTTP_HAVE_WCHAR
726 } else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
727 if (!iswalnum(wchar)) {
728 break;
729 }
730 #endif
731 }
732 PHP_HTTP_DUFF(consumed, state->buffer[state->offset++] = *ptr++);
733 } else {
734 int i = 0;
735
736 PHP_HTTP_DUFF(consumed,
737 state->buffer[state->offset++] = '%';
738 state->buffer[state->offset++] = parse_xdigits[((unsigned char) ptr[i]) >> 4];
739 state->buffer[state->offset++] = parse_xdigits[((unsigned char) ptr[i]) & 0xf];
740 ++i;
741 );
742 }
743
744 return consumed;
745 }
746
747 if (!silent) {
748 TSRMLS_FETCH_FROM_CTX(state->ts);
749 if (consumed) {
750 php_error_docref(NULL TSRMLS_CC, E_WARNING,
751 "Failed to parse %s; unexpected multibyte sequence 0x%x at pos %u in '%s'",
752 parse_what[what], wchar, (unsigned) (ptr - begin), begin);
753 } else {
754 php_error_docref(NULL TSRMLS_CC, E_WARNING,
755 "Failed to parse %s; unexpected byte 0x%02x at pos %u in '%s'",
756 parse_what[what], (unsigned char) *ptr, (unsigned) (ptr - begin), begin);
757 }
758 }
759
760 return 0;
761 }
762
763 static STATUS parse_userinfo(struct parse_state *state, const char *ptr)
764 {
765 size_t mb;
766 const char *password = NULL, *end = state->ptr, *tmp = ptr;
767 TSRMLS_FETCH_FROM_CTX(state->ts);
768
769 state->url.user = &state->buffer[state->offset];
770
771 do {
772 switch (*ptr) {
773 case ':':
774 if (password) {
775 php_error_docref(NULL TSRMLS_CC, E_WARNING,
776 "Failed to parse password; duplicate ':' at pos %u in '%s'",
777 (unsigned) (ptr - tmp), tmp);
778 return FAILURE;
779 }
780 password = ptr + 1;
781 state->buffer[state->offset++] = 0;
782 state->url.pass = &state->buffer[state->offset];
783 break;
784
785 case '%':
786 if (ptr[1] != '%' && (end - ptr <= 2 || !isxdigit(*(ptr+1)) || !isxdigit(*(ptr+2)))) {
787 php_error_docref(NULL TSRMLS_CC, E_WARNING,
788 "Failed to parse userinfo; invalid percent encoding at pos %u in '%s'",
789 (unsigned) (ptr - tmp), tmp);
790 return FAILURE;
791 }
792 state->buffer[state->offset++] = *ptr++;
793 state->buffer[state->offset++] = *ptr++;
794 state->buffer[state->offset++] = *ptr;
795 break;
796
797 case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
798 case '+': case ',': case ';': case '=': /* sub-delims */
799 case '-': case '.': case '_': case '~': /* unreserved */
800 case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
801 case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
802 case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
803 case 'V': case 'W': case 'X': case 'Y': case 'Z':
804 case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
805 case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
806 case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
807 case 'v': case 'w': case 'x': case 'y': case 'z':
808 case '0': case '1': case '2': case '3': case '4': case '5': case '6':
809 case '7': case '8': case '9':
810 /* allowed */
811 state->buffer[state->offset++] = *ptr;
812 break;
813
814 default:
815 if (!(mb = parse_mb(state, PARSE_USERINFO, ptr, end, tmp, 0))) {
816 return FAILURE;
817 }
818 ptr += mb - 1;
819 }
820 } while(++ptr != end);
821
822
823 state->buffer[state->offset++] = 0;
824
825 return SUCCESS;
826 }
827
828 #if defined(PHP_WIN32) || defined(HAVE_UIDNA_IDNTOASCII)
829 typedef size_t (*parse_mb_func)(unsigned *wc, const char *ptr, const char *end);
830 static STATUS to_utf16(parse_mb_func fn, const char *u8, uint16_t **u16, size_t *len)
831 {
832 size_t offset = 0, u8_len = strlen(u8);
833
834 *u16 = ecalloc(4 * sizeof(uint16_t), u8_len + 1);
835 *len = 0;
836
837 while (offset < u8_len) {
838 unsigned wc;
839 uint16_t buf[2], *ptr = buf;
840 size_t consumed = fn(&wc, &u8[offset], &u8[u8_len]);
841
842 if (!consumed) {
843 efree(*u16);
844 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse UTF-8 at pos %zu of '%s'", offset, u8);
845 return FAILURE;
846 } else {
847 offset += consumed;
848 }
849
850 switch (wctoutf16(buf, wc)) {
851 case 2:
852 (*u16)[(*len)++] = *ptr++;
853 /* no break */
854 case 1:
855 (*u16)[(*len)++] = *ptr++;
856 break;
857 case 0:
858 default:
859 efree(*u16);
860 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to convert UTF-32 'U+%X' to UTF-16", wc);
861 return FAILURE;
862 }
863 }
864
865 return SUCCESS;
866 }
867 #endif
868
869 #ifndef MAXHOSTNAMELEN
870 # define MAXHOSTNAMELEN 256
871 #endif
872
873 #ifdef PHP_HTTP_HAVE_IDN
874 static STATUS parse_idn(struct parse_state *state)
875 {
876 char *idn = NULL;
877 int rv = -1;
878
879 if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
880 rv = idna_to_ascii_8z(state->url.host, &idn, IDNA_ALLOW_UNASSIGNED|IDNA_USE_STD3_ASCII_RULES);
881 }
882 # ifdef PHP_HTTP_HAVE_WCHAR
883 else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
884 rv = idna_to_ascii_lz(state->url.host, &idn, IDNA_ALLOW_UNASSIGNED|IDNA_USE_STD3_ASCII_RULES);
885 }
886 # endif
887 if (rv != IDNA_SUCCESS) {
888 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse IDN; %s", idna_strerror(rv));
889 return FAILURE;
890 } else {
891 size_t idnlen = strlen(idn);
892 memcpy(state->url.host, idn, idnlen + 1);
893 free(idn);
894 state->offset += idnlen - len;
895 return SUCCESS;
896 }
897 }
898 #endif
899
900 #ifdef HAVE_UIDNA_IDNTOASCII
901 # if HAVE_UNICODE_UIDNA_H
902 # include <unicode/uidna.h>
903 # else
904 typedef uint16_t UChar;
905 typedef enum { U_ZERO_ERROR = 0 } UErrorCode;
906 int32_t uidna_IDNToASCII(const UChar *src, int32_t srcLength, UChar *dest, int32_t destCapacity, int32_t options, void *parseError, UErrorCode *status);
907 # endif
908 static STATUS parse_uidn(struct parse_state *state)
909 {
910 char *host_ptr;
911 uint16_t *uhost_str, ahost_str[MAXHOSTNAMELEN], *ahost_ptr;
912 size_t uhost_len, ahost_len;
913 UErrorCode error = U_ZERO_ERROR;
914
915 if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
916 if (SUCCESS != to_utf16(parse_mb_utf8, state->url.host, &uhost_str, &uhost_len)) {
917 return FAILURE;
918 }
919 #ifdef PHP_HTTP_HAVE_WCHAR
920 } else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
921 if (SUCCESS != to_utf16(parse_mb_loc, state->url.host, &uhost_str, &uhost_len)) {
922 return FAILURE;
923 }
924 #endif
925 } else {
926 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse IDN; codepage not specified");
927 return FAILURE;
928 }
929
930 ahost_len = uidna_IDNToASCII(uhost_str, uhost_len, ahost_str, MAXHOSTNAMELEN, 3, NULL, &error);
931 efree(uhost_str);
932
933 if (error != U_ZERO_ERROR) {
934 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse IDN; ICU error %d", error);
935 return FAILURE;
936 }
937
938 host_ptr = state->url.host;
939 ahost_ptr = ahost_str;
940 PHP_HTTP_DUFF(ahost_len, *host_ptr++ = *ahost_ptr++);
941
942 *host_ptr = '\0';
943 state->offset += host_ptr - state->url.host;
944
945 return SUCCESS;
946 }
947 #endif
948
949 #if 0 && defined(PHP_WIN32)
950 static STATUS parse_widn(struct parse_state *state)
951 {
952 char *host_ptr;
953 uint16_t *uhost_str, ahost_str[MAXHOSTNAMELEN], *ahost_ptr;
954 size_t uhost_len;
955
956 if (state->flags & PHP_HTTP_URL_PARSE_MBUTF8) {
957 if (SUCCESS != to_utf16(parse_mb_utf8, state->url.host, &uhost_str, &uhost_len)) {
958 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse IDN");
959 return FAILURE;
960 }
961 #ifdef PHP_HTTP_HAVE_WCHAR
962 } else if (state->flags & PHP_HTTP_URL_PARSE_MBLOC) {
963 if (SUCCESS != to_utf16(parse_mb_loc, state->url.host, &uhost_str, &uhost_len)) {
964 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse IDN");
965 return FAILURE;
966 }
967 #endif
968 } else {
969 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse IDN");
970 return FAILURE;
971 }
972
973 if (!IdnToAscii(IDN_ALLOW_UNASSIGNED|IDN_USE_STD3_ASCII_RULES, uhost_str, uhost_len, ahost_str, MAXHOSTNAMELEN)) {
974 efree(uhost_str);
975 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse IDN");
976 return FAILURE;
977 }
978
979 efree(uhost_str);
980 host_ptr = state->url.host;
981 ahost_ptr = ahost_str;
982 PHP_HTTP_DUFF(wcslen(ahost_str), *host_ptr++ = *ahost_ptr++);
983 efree(ahost_str);
984
985 *host_ptr = '\0';
986 state->offset += host_ptr - state->url.host;
987
988 return SUCCESS;
989 }
990 #endif
991
992 static STATUS parse_hostinfo(struct parse_state *state, const char *ptr)
993 {
994 size_t mb, len;
995 const char *end = state->ptr, *tmp = ptr, *port = NULL;
996 TSRMLS_FETCH_FROM_CTX(state->ts);
997
998 #ifdef HAVE_INET_PTON
999 if (*ptr == '[') {
1000 char *error = NULL, *tmp = memchr(ptr, ']', end - ptr);
1001
1002 if (tmp) {
1003 size_t addrlen = tmp - ptr + 1;
1004 char buf[16], *addr = estrndup(ptr + 1, addrlen - 2);
1005 int rv = inet_pton(AF_INET6, addr, buf);
1006
1007 efree(addr);
1008 if (rv == 1) {
1009 state->buffer[state->offset] = '[';
1010 state->url.host = &state->buffer[state->offset];
1011 inet_ntop(AF_INET6, buf, state->url.host + 1, state->maxlen - state->offset);
1012 state->offset += strlen(state->url.host);
1013 state->buffer[state->offset++] = ']';
1014 state->buffer[state->offset++] = 0;
1015 ptr = tmp + 1;
1016 } else if (rv == -1) {
1017 error = strerror(errno);
1018 } else {
1019 error = "unexpected '['";
1020 }
1021 } else {
1022 error = "expected ']'";
1023 }
1024
1025 if (error) {
1026 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse hostinfo; %s", error);
1027 return FAILURE;
1028 }
1029 }
1030 #endif
1031 if (ptr != end) do {
1032 switch (*ptr) {
1033 case ':':
1034 if (port) {
1035 php_error_docref(NULL TSRMLS_CC, E_WARNING,
1036 "Failed to parse port; unexpected ':' at pos %u in '%s'",
1037 (unsigned) (ptr - tmp), tmp);
1038 return FAILURE;
1039 }
1040 port = ptr + 1;
1041 break;
1042
1043 case '%':
1044 if (ptr[1] != '%' && (end - ptr <= 2 || !isxdigit(*(ptr+1)) || !isxdigit(*(ptr+2)))) {
1045 php_error_docref(NULL TSRMLS_CC, E_WARNING,
1046 "Failed to parse hostinfo; invalid percent encoding at pos %u in '%s'",
1047 (unsigned) (ptr - tmp), tmp);
1048 return FAILURE;
1049 }
1050 state->buffer[state->offset++] = *ptr++;
1051 state->buffer[state->offset++] = *ptr++;
1052 state->buffer[state->offset++] = *ptr;
1053 break;
1054
1055 case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
1056 case '+': case ',': case ';': case '=': /* sub-delims */
1057 case '-': case '.': case '_': case '~': /* unreserved */
1058 case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
1059 case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
1060 case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
1061 case 'V': case 'W': case 'X': case 'Y': case 'Z':
1062 case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
1063 case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
1064 case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
1065 case 'v': case 'w': case 'x': case 'y': case 'z':
1066 if (port) {
1067 php_error_docref(NULL TSRMLS_CC, E_WARNING,
1068 "Failed to parse port; unexpected char '%c' at pos %u in '%s'",
1069 (unsigned char) *ptr, (unsigned) (ptr - tmp), tmp);
1070 return FAILURE;
1071 }
1072 /* no break */
1073 case '0': case '1': case '2': case '3': case '4': case '5': case '6':
1074 case '7': case '8': case '9':
1075 /* allowed */
1076 if (port) {
1077 state->url.port *= 10;
1078 state->url.port += *ptr - '0';
1079 } else {
1080 state->buffer[state->offset++] = *ptr;
1081 }
1082 break;
1083
1084 default:
1085 if (ptr == end) {
1086 break;
1087 } else if (port) {
1088 php_error_docref(NULL TSRMLS_CC, E_WARNING,
1089 "Failed to parse port; unexpected byte 0x%02x at pos %u in '%s'",
1090 (unsigned char) *ptr, (unsigned) (ptr - tmp), tmp);
1091 return FAILURE;
1092 } else if (!(mb = parse_mb(state, PARSE_HOSTINFO, ptr, end, tmp, 0))) {
1093 return FAILURE;
1094 }
1095 ptr += mb - 1;
1096 }
1097 } while (++ptr != end);
1098
1099 if (!state->url.host) {
1100 len = (port ? port - tmp - 1 : end - tmp);
1101 state->url.host = &state->buffer[state->offset - len];
1102 state->buffer[state->offset++] = 0;
1103 }
1104
1105 if (state->flags & PHP_HTTP_URL_PARSE_TOIDN) {
1106 #ifdef PHP_HTTP_HAVE_IDN
1107 return parse_idn(state);
1108 #endif
1109 #ifdef HAVE_UIDNA_IDNTOASCII
1110 return parse_uidn(state);
1111 #endif
1112 }
1113
1114 return SUCCESS;
1115 }
1116
1117 static const char *parse_authority(struct parse_state *state)
1118 {
1119 const char *tmp = state->ptr, *host = NULL;
1120
1121 do {
1122 switch (*state->ptr) {
1123 case '@':
1124 /* userinfo delimiter */
1125 if (host) {
1126 TSRMLS_FETCH_FROM_CTX(state->ts);
1127 php_error_docref(NULL TSRMLS_CC, E_WARNING,
1128 "Failed to parse userinfo; unexpected '@'");
1129 return NULL;
1130 }
1131 host = state->ptr + 1;
1132 if (tmp != state->ptr && SUCCESS != parse_userinfo(state, tmp)) {
1133 return NULL;
1134 }
1135 tmp = state->ptr + 1;
1136 break;
1137
1138 case '/':
1139 case '?':
1140 case '#':
1141 case '\0':
1142 EOD:
1143 /* host delimiter */
1144 if (tmp != state->ptr && SUCCESS != parse_hostinfo(state, tmp)) {
1145 return NULL;
1146 }
1147 return state->ptr;
1148 }
1149 } while (++state->ptr <= state->end);
1150
1151 --state->ptr;
1152 goto EOD;
1153 }
1154
1155 static const char *parse_path(struct parse_state *state)
1156 {
1157 size_t mb;
1158 const char *tmp;
1159 TSRMLS_FETCH_FROM_CTX(state->ts);
1160
1161 /* is there actually a path to parse? */
1162 if (!*state->ptr) {
1163 return state->ptr;
1164 }
1165 tmp = state->ptr;
1166 state->url.path = &state->buffer[state->offset];
1167
1168 do {
1169 switch (*state->ptr) {
1170 case '#':
1171 case '?':
1172 goto done;
1173
1174 case '%':
1175 if (state->ptr[1] != '%' && (state->end - state->ptr <= 2 || !isxdigit(*(state->ptr+1)) || !isxdigit(*(state->ptr+2)))) {
1176 php_error_docref(NULL TSRMLS_CC, E_WARNING,
1177 "Failed to parse path; invalid percent encoding at pos %u in '%s'",
1178 (unsigned) (state->ptr - tmp), tmp);
1179 return NULL;
1180 }
1181 state->buffer[state->offset++] = *state->ptr++;
1182 state->buffer[state->offset++] = *state->ptr++;
1183 state->buffer[state->offset++] = *state->ptr;
1184 break;
1185
1186 case '/': /* yeah, well */
1187 case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
1188 case '+': case ',': case ';': case '=': /* sub-delims */
1189 case '-': case '.': case '_': case '~': /* unreserved */
1190 case ':': case '@': /* pchar */
1191 case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
1192 case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
1193 case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
1194 case 'V': case 'W': case 'X': case 'Y': case 'Z':
1195 case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
1196 case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
1197 case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
1198 case 'v': case 'w': case 'x': case 'y': case 'z':
1199 case '0': case '1': case '2': case '3': case '4': case '5': case '6':
1200 case '7': case '8': case '9':
1201 /* allowed */
1202 state->buffer[state->offset++] = *state->ptr;
1203 break;
1204
1205 default:
1206 if (!(mb = parse_mb(state, PARSE_PATH, state->ptr, state->end, tmp, 0))) {
1207 return NULL;
1208 }
1209 state->ptr += mb - 1;
1210 }
1211 } while (++state->ptr < state->end);
1212
1213 done:
1214 /* did we have any path component ? */
1215 if (tmp != state->ptr) {
1216 state->buffer[state->offset++] = 0;
1217 } else {
1218 state->url.path = NULL;
1219 }
1220 return state->ptr;
1221 }
1222
1223 static const char *parse_query(struct parse_state *state)
1224 {
1225 size_t mb;
1226 const char *tmp = state->ptr + !!*state->ptr;
1227 TSRMLS_FETCH_FROM_CTX(state->ts);
1228
1229 /* is there actually a query to parse? */
1230 if (*state->ptr != '?') {
1231 return state->ptr;
1232 }
1233
1234 /* skip initial '?' */
1235 tmp = ++state->ptr;
1236 state->url.query = &state->buffer[state->offset];
1237
1238 do {
1239 switch (*state->ptr) {
1240 case '#':
1241 goto done;
1242
1243 case '%':
1244 if (state->ptr[1] != '%' && (state->end - state->ptr <= 2 || !isxdigit(*(state->ptr+1)) || !isxdigit(*(state->ptr+2)))) {
1245 php_error_docref(NULL TSRMLS_CC, E_WARNING,
1246 "Failed to parse query; invalid percent encoding at pos %u in '%s'",
1247 (unsigned) (state->ptr - tmp), tmp);
1248 return NULL;
1249 }
1250 state->buffer[state->offset++] = *state->ptr++;
1251 state->buffer[state->offset++] = *state->ptr++;
1252 state->buffer[state->offset++] = *state->ptr;
1253 break;
1254
1255 case ']':
1256 case '[':
1257 if (state->flags & PHP_HTTP_URL_PARSE_TOPCT) {
1258 state->buffer[state->offset++] = '%';
1259 state->buffer[state->offset++] = parse_xdigits[((unsigned char) *state->ptr) >> 4];
1260 state->buffer[state->offset++] = parse_xdigits[((unsigned char) *state->ptr) & 0xf];
1261 break;
1262 }
1263 /* no break */
1264
1265 case '?': case '/': /* yeah, well */
1266 case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
1267 case '+': case ',': case ';': case '=': /* sub-delims */
1268 case '-': case '.': case '_': case '~': /* unreserved */
1269 case ':': case '@': /* pchar */
1270 case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
1271 case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
1272 case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
1273 case 'V': case 'W': case 'X': case 'Y': case 'Z':
1274 case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
1275 case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
1276 case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
1277 case 'v': case 'w': case 'x': case 'y': case 'z':
1278 case '0': case '1': case '2': case '3': case '4': case '5': case '6':
1279 case '7': case '8': case '9':
1280 /* allowed */
1281 state->buffer[state->offset++] = *state->ptr;
1282 break;
1283
1284 default:
1285 if (!(mb = parse_mb(state, PARSE_QUERY, state->ptr, state->end, tmp, 0))) {
1286 return NULL;
1287 }
1288 state->ptr += mb - 1;
1289 }
1290 } while (++state->ptr < state->end);
1291
1292 done:
1293 state->buffer[state->offset++] = 0;
1294 return state->ptr;
1295 }
1296
1297 static const char *parse_fragment(struct parse_state *state)
1298 {
1299 size_t mb;
1300 const char *tmp;
1301 TSRMLS_FETCH_FROM_CTX(state->ts);
1302
1303 /* is there actually a fragment to parse? */
1304 if (*state->ptr != '#') {
1305 return state->ptr;
1306 }
1307
1308 /* skip initial '#' */
1309 tmp = ++state->ptr;
1310 state->url.fragment = &state->buffer[state->offset];
1311
1312 do {
1313 switch (*state->ptr) {
1314 case '%':
1315 if (state->ptr[1] != '%' && (state->end - state->ptr <= 2 || !isxdigit(*(state->ptr+1)) || !isxdigit(*(state->ptr+2)))) {
1316 php_error_docref(NULL TSRMLS_CC, E_WARNING,
1317 "Failed to parse fragment; invalid percent encoding at pos %u in '%s'",
1318 (unsigned) (state->ptr - tmp), tmp);
1319 return NULL;
1320 }
1321 state->buffer[state->offset++] = *state->ptr++;
1322 state->buffer[state->offset++] = *state->ptr++;
1323 state->buffer[state->offset++] = *state->ptr;
1324 break;
1325
1326 case '?': case '/':
1327 case '!': case '$': case '&': case '\'': case '(': case ')': case '*':
1328 case '+': case ',': case ';': case '=': /* sub-delims */
1329 case '-': case '.': case '_': case '~': /* unreserved */
1330 case ':': case '@': /* pchar */
1331 case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
1332 case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
1333 case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
1334 case 'V': case 'W': case 'X': case 'Y': case 'Z':
1335 case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
1336 case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
1337 case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
1338 case 'v': case 'w': case 'x': case 'y': case 'z':
1339 case '0': case '1': case '2': case '3': case '4': case '5': case '6':
1340 case '7': case '8': case '9':
1341 /* allowed */
1342 state->buffer[state->offset++] = *state->ptr;
1343 break;
1344
1345 default:
1346 if (!(mb = parse_mb(state, PARSE_FRAGMENT, state->ptr, state->end, tmp, 0))) {
1347 return NULL;
1348 }
1349 state->ptr += mb - 1;
1350 }
1351 } while (++state->ptr < state->end);
1352
1353 state->buffer[state->offset++] = 0;
1354 return state->ptr;
1355 }
1356
1357 static const char *parse_hier(struct parse_state *state)
1358 {
1359 if (*state->ptr == '/') {
1360 if (state->end - state->ptr > 1) {
1361 if (*(state->ptr + 1) == '/') {
1362 state->ptr += 2;
1363 if (!(state->ptr = parse_authority(state))) {
1364 return NULL;
1365 }
1366 }
1367 }
1368 }
1369 return parse_path(state);
1370 }
1371
1372 static const char *parse_scheme(struct parse_state *state)
1373 {
1374 size_t mb;
1375 const char *tmp = state->ptr;
1376
1377 do {
1378 switch (*state->ptr) {
1379 case ':':
1380 /* scheme delimiter */
1381 state->url.scheme = &state->buffer[0];
1382 state->buffer[state->offset++] = 0;
1383 return ++state->ptr;
1384
1385 case '0': case '1': case '2': case '3': case '4': case '5': case '6':
1386 case '7': case '8': case '9':
1387 case '+': case '-': case '.':
1388 if (state->ptr == tmp) {
1389 return tmp;
1390 }
1391 /* no break */
1392 case 'A': case 'B': case 'C': case 'D': case 'E': case 'F': case 'G':
1393 case 'H': case 'I': case 'J': case 'K': case 'L': case 'M': case 'N':
1394 case 'O': case 'P': case 'Q': case 'R': case 'S': case 'T': case 'U':
1395 case 'V': case 'W': case 'X': case 'Y': case 'Z':
1396 case 'a': case 'b': case 'c': case 'd': case 'e': case 'f': case 'g':
1397 case 'h': case 'i': case 'j': case 'k': case 'l': case 'm': case 'n':
1398 case 'o': case 'p': case 'q': case 'r': case 's': case 't': case 'u':
1399 case 'v': case 'w': case 'x': case 'y': case 'z':
1400 /* scheme part */
1401 state->buffer[state->offset++] = *state->ptr;
1402 break;
1403
1404 default:
1405 if (!(mb = parse_mb(state, PARSE_SCHEME, state->ptr, state->end, tmp, 1))) {
1406 /* soft fail; parse path next */
1407 return tmp;
1408 }
1409 state->ptr += mb - 1;
1410 }
1411 } while (++state->ptr != state->end);
1412
1413 return state->ptr = tmp;
1414 }
1415
1416 php_http_url_t *php_http_url_parse(const char *str, size_t len, unsigned flags TSRMLS_DC)
1417 {
1418 size_t maxlen = 3 * len;
1419 struct parse_state *state = ecalloc(1, sizeof(*state) + maxlen);
1420
1421 state->end = str + len;
1422 state->ptr = str;
1423 state->flags = flags;
1424 state->maxlen = maxlen;
1425 TSRMLS_SET_CTX(state->ts);
1426
1427 if (!parse_scheme(state)) {
1428 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse URL scheme: '%s'", state->ptr);
1429 efree(state);
1430 return NULL;
1431 }
1432
1433 if (!parse_hier(state)) {
1434 efree(state);
1435 return NULL;
1436 }
1437
1438 if (!parse_query(state)) {
1439 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse URL query: '%s'", state->ptr);
1440 efree(state);
1441 return NULL;
1442 }
1443
1444 if (!parse_fragment(state)) {
1445 php_error_docref(NULL TSRMLS_CC, E_WARNING, "Failed to parse URL fragment: '%s'", state->ptr);
1446 efree(state);
1447 return NULL;
1448 }
1449
1450 return (php_http_url_t *) state;
1451 }
1452
1453 php_http_url_t *php_http_url_parse_authority(const char *str, size_t len, unsigned flags TSRMLS_DC)
1454 {
1455 size_t maxlen = 3 * len;
1456 struct parse_state *state = ecalloc(1, sizeof(*state) + maxlen);
1457
1458 state->end = str + len;
1459 state->ptr = str;
1460 state->flags = flags;
1461 state->maxlen = maxlen;
1462 TSRMLS_SET_CTX(state->ts);
1463
1464 if (!(state->ptr = parse_authority(state))) {
1465 efree(state);
1466 return NULL;
1467 }
1468
1469 if (state->ptr != state->end) {
1470 php_error_docref(NULL TSRMLS_CC, E_WARNING,
1471 "Failed to parse URL authority, unexpected character at pos %u in '%s'",
1472 (unsigned) (state->ptr - str), str);
1473 efree(state);
1474 return NULL;
1475 }
1476
1477 return (php_http_url_t *) state;
1478 }
1479
1480 ZEND_BEGIN_ARG_INFO_EX(ai_HttpUrl___construct, 0, 0, 0)
1481 ZEND_ARG_INFO(0, old_url)
1482 ZEND_ARG_INFO(0, new_url)
1483 ZEND_ARG_INFO(0, flags)
1484 ZEND_END_ARG_INFO();
1485 PHP_METHOD(HttpUrl, __construct)
1486 {
1487 zval *new_url = NULL, *old_url = NULL;
1488 long flags = PHP_HTTP_URL_FROM_ENV;
1489 zend_error_handling zeh;
1490
1491 php_http_expect(SUCCESS == zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "|z!z!l", &old_url, &new_url, &flags), invalid_arg, return);
1492
1493 zend_replace_error_handling(EH_THROW, php_http_exception_bad_url_class_entry, &zeh TSRMLS_CC);
1494 {
1495 php_http_url_t *res_purl, *new_purl = NULL, *old_purl = NULL;
1496
1497 if (new_url) {
1498 new_purl = php_http_url_from_zval(new_url, flags TSRMLS_CC);
1499 if (!new_purl) {
1500 zend_restore_error_handling(&zeh TSRMLS_CC);
1501 return;
1502 }
1503 }
1504 if (old_url) {
1505 old_purl = php_http_url_from_zval(old_url, flags TSRMLS_CC);
1506 if (!old_purl) {
1507 if (new_purl) {
1508 php_http_url_free(&new_purl);
1509 }
1510 zend_restore_error_handling(&zeh TSRMLS_CC);
1511 return;
1512 }
1513 }
1514
1515 res_purl = php_http_url_mod(old_purl, new_purl, flags TSRMLS_CC);
1516 php_http_url_to_struct(res_purl, getThis() TSRMLS_CC);
1517
1518 php_http_url_free(&res_purl);
1519 if (old_purl) {
1520 php_http_url_free(&old_purl);
1521 }
1522 if (new_purl) {
1523 php_http_url_free(&new_purl);
1524 }
1525 }
1526 zend_restore_error_handling(&zeh TSRMLS_CC);
1527 }
1528
1529 ZEND_BEGIN_ARG_INFO_EX(ai_HttpUrl_mod, 0, 0, 1)
1530 ZEND_ARG_INFO(0, more_url_parts)
1531 ZEND_ARG_INFO(0, flags)
1532 ZEND_END_ARG_INFO();
1533 PHP_METHOD(HttpUrl, mod)
1534 {
1535 zval *new_url = NULL;
1536 long flags = PHP_HTTP_URL_JOIN_PATH | PHP_HTTP_URL_JOIN_QUERY;
1537 zend_error_handling zeh;
1538
1539 php_http_expect(SUCCESS == zend_parse_parameters(ZEND_NUM_ARGS() TSRMLS_CC, "z!|l", &new_url, &flags), invalid_arg, return);
1540
1541 zend_replace_error_handling(EH_THROW, php_http_exception_bad_url_class_entry, &zeh TSRMLS_CC);
1542 {
1543 php_http_url_t *new_purl = NULL, *old_purl = NULL;
1544
1545 if (new_url) {
1546 new_purl = php_http_url_from_zval(new_url, flags TSRMLS_CC);
1547 if (!new_purl) {
1548 zend_restore_error_handling(&zeh TSRMLS_CC);
1549 return;
1550 }
1551 }
1552
1553 if ((old_purl = php_http_url_from_struct(HASH_OF(getThis())))) {
1554 php_http_url_t *res_purl;
1555
1556 ZVAL_OBJVAL(return_value, zend_objects_clone_obj(getThis() TSRMLS_CC), 0);
1557
1558 res_purl = php_http_url_mod(old_purl, new_purl, flags TSRMLS_CC);
1559 php_http_url_to_struct(res_purl, return_value TSRMLS_CC);
1560
1561 php_http_url_free(&res_purl);
1562 php_http_url_free(&old_purl);
1563 }
1564 if (new_purl) {
1565 php_http_url_free(&new_purl);
1566 }
1567 }
1568 zend_restore_error_handling(&zeh TSRMLS_CC);
1569 }
1570
1571 ZEND_BEGIN_ARG_INFO_EX(ai_HttpUrl_toString, 0, 0, 0)
1572 ZEND_END_ARG_INFO();
1573 PHP_METHOD(HttpUrl, toString)
1574 {
1575 if (SUCCESS == zend_parse_parameters_none()) {
1576 php_http_url_t *purl;
1577
1578 if ((purl = php_http_url_from_struct(HASH_OF(getThis())))) {
1579 char *str;
1580 size_t len;
1581
1582 php_http_url_to_string(purl, &str, &len, 0);
1583 php_http_url_free(&purl);
1584 RETURN_STRINGL(str, len, 0);
1585 }
1586 }
1587 RETURN_EMPTY_STRING();
1588 }
1589
1590 ZEND_BEGIN_ARG_INFO_EX(ai_HttpUrl_toArray, 0, 0, 0)
1591 ZEND_END_ARG_INFO();
1592 PHP_METHOD(HttpUrl, toArray)
1593 {
1594 php_http_url_t *purl;
1595
1596 if (SUCCESS != zend_parse_parameters_none()) {
1597 return;
1598 }
1599
1600 /* strip any non-URL properties */
1601 purl = php_http_url_from_struct(HASH_OF(getThis()));
1602 php_http_url_to_struct(purl, return_value TSRMLS_CC);
1603 php_http_url_free(&purl);
1604 }
1605
1606 static zend_function_entry php_http_url_methods[] = {
1607 PHP_ME(HttpUrl, __construct, ai_HttpUrl___construct, ZEND_ACC_PUBLIC|ZEND_ACC_CTOR)
1608 PHP_ME(HttpUrl, mod, ai_HttpUrl_mod, ZEND_ACC_PUBLIC)
1609 PHP_ME(HttpUrl, toString, ai_HttpUrl_toString, ZEND_ACC_PUBLIC)
1610 ZEND_MALIAS(HttpUrl, __toString, toString, ai_HttpUrl_toString, ZEND_ACC_PUBLIC)
1611 PHP_ME(HttpUrl, toArray, ai_HttpUrl_toArray, ZEND_ACC_PUBLIC)
1612 EMPTY_FUNCTION_ENTRY
1613 };
1614
1615 zend_class_entry *php_http_url_class_entry;
1616
1617 PHP_MINIT_FUNCTION(http_url)
1618 {
1619 zend_class_entry ce = {0};
1620
1621 INIT_NS_CLASS_ENTRY(ce, "http", "Url", php_http_url_methods);
1622 php_http_url_class_entry = zend_register_internal_class(&ce TSRMLS_CC);
1623
1624 zend_declare_property_null(php_http_url_class_entry, ZEND_STRL("scheme"), ZEND_ACC_PUBLIC TSRMLS_CC);
1625 zend_declare_property_null(php_http_url_class_entry, ZEND_STRL("user"), ZEND_ACC_PUBLIC TSRMLS_CC);
1626 zend_declare_property_null(php_http_url_class_entry, ZEND_STRL("pass"), ZEND_ACC_PUBLIC TSRMLS_CC);
1627 zend_declare_property_null(php_http_url_class_entry, ZEND_STRL("host"), ZEND_ACC_PUBLIC TSRMLS_CC);
1628 zend_declare_property_null(php_http_url_class_entry, ZEND_STRL("port"), ZEND_ACC_PUBLIC TSRMLS_CC);
1629 zend_declare_property_null(php_http_url_class_entry, ZEND_STRL("path"), ZEND_ACC_PUBLIC TSRMLS_CC);
1630 zend_declare_property_null(php_http_url_class_entry, ZEND_STRL("query"), ZEND_ACC_PUBLIC TSRMLS_CC);
1631 zend_declare_property_null(php_http_url_class_entry, ZEND_STRL("fragment"), ZEND_ACC_PUBLIC TSRMLS_CC);
1632
1633 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("REPLACE"), PHP_HTTP_URL_REPLACE TSRMLS_CC);
1634 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("JOIN_PATH"), PHP_HTTP_URL_JOIN_PATH TSRMLS_CC);
1635 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("JOIN_QUERY"), PHP_HTTP_URL_JOIN_QUERY TSRMLS_CC);
1636 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("STRIP_USER"), PHP_HTTP_URL_STRIP_USER TSRMLS_CC);
1637 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("STRIP_PASS"), PHP_HTTP_URL_STRIP_PASS TSRMLS_CC);
1638 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("STRIP_AUTH"), PHP_HTTP_URL_STRIP_AUTH TSRMLS_CC);
1639 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("STRIP_PORT"), PHP_HTTP_URL_STRIP_PORT TSRMLS_CC);
1640 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("STRIP_PATH"), PHP_HTTP_URL_STRIP_PATH TSRMLS_CC);
1641 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("STRIP_QUERY"), PHP_HTTP_URL_STRIP_QUERY TSRMLS_CC);
1642 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("STRIP_FRAGMENT"), PHP_HTTP_URL_STRIP_FRAGMENT TSRMLS_CC);
1643 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("STRIP_ALL"), PHP_HTTP_URL_STRIP_ALL TSRMLS_CC);
1644 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("FROM_ENV"), PHP_HTTP_URL_FROM_ENV TSRMLS_CC);
1645 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("SANITIZE_PATH"), PHP_HTTP_URL_SANITIZE_PATH TSRMLS_CC);
1646
1647 #ifdef PHP_HTTP_HAVE_WCHAR
1648 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("PARSE_MBLOC"), PHP_HTTP_URL_PARSE_MBLOC TSRMLS_CC);
1649 #endif
1650 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("PARSE_MBUTF8"), PHP_HTTP_URL_PARSE_MBUTF8 TSRMLS_CC);
1651 #if defined(PHP_HTTP_HAVE_IDN) || defined(HAVE_UIDNA_IDNTOASCII)
1652 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("PARSE_TOIDN"), PHP_HTTP_URL_PARSE_TOIDN TSRMLS_CC);
1653 #endif
1654 zend_declare_class_constant_long(php_http_url_class_entry, ZEND_STRL("PARSE_TOPCT"), PHP_HTTP_URL_PARSE_TOPCT TSRMLS_CC);
1655
1656 return SUCCESS;
1657 }
1658
1659
1660 /*
1661 * Local variables:
1662 * tab-width: 4
1663 * c-basic-offset: 4
1664 * End:
1665 * vim600: noet sw=4 ts=4 fdm=marker
1666 * vim<600: noet sw=4 ts=4
1667 */
1668
Extended HTTP Support