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