1
|
<?php
|
2
|
|
3
|
/**
|
4
|
* @file
|
5
|
* Download via HTTP.
|
6
|
*
|
7
|
* Support caching, HTTP Basic Authentication, detection of RSS/Atom feeds,
|
8
|
* redirects.
|
9
|
*/
|
10
|
|
11
|
/**
|
12
|
* PCRE for finding the link tags in html.
|
13
|
*/
|
14
|
define('HTTP_REQUEST_PCRE_LINK_TAG', '/<link((?:[\x09\x0A\x0B\x0C\x0D\x20]+[^\x09\x0A\x0B\x0C\x0D\x20\x2F\x3E][^\x09\x0A\x0B\x0C\x0D\x20\x2F\x3D\x3E]*(?:[\x09\x0A\x0B\x0C\x0D\x20]*=[\x09\x0A\x0B\x0C\x0D\x20]*(?:"(?:[^"]*)"|\'(?:[^\']*)\'|(?:[^\x09\x0A\x0B\x0C\x0D\x20\x22\x27\x3E][^\x09\x0A\x0B\x0C\x0D\x20\x3E]*)?))?)*)[\x09\x0A\x0B\x0C\x0D\x20]*(>(.*)<\/link>|(\/)?>)/si');
|
15
|
|
16
|
/**
|
17
|
* PCRE for matching all the attributes in a tag.
|
18
|
*/
|
19
|
define('HTTP_REQUEST_PCRE_TAG_ATTRIBUTES', '/[\x09\x0A\x0B\x0C\x0D\x20]+([^\x09\x0A\x0B\x0C\x0D\x20\x2F\x3E][^\x09\x0A\x0B\x0C\x0D\x20\x2F\x3D\x3E]*)(?:[\x09\x0A\x0B\x0C\x0D\x20]*=[\x09\x0A\x0B\x0C\x0D\x20]*(?:"([^"]*)"|\'([^\']*)\'|([^\x09\x0A\x0B\x0C\x0D\x20\x22\x27\x3E][^\x09\x0A\x0B\x0C\x0D\x20\x3E]*)?))?/');
|
20
|
|
21
|
/**
|
22
|
* For cUrl specific errors.
|
23
|
*/
|
24
|
class HRCurlException extends Exception {}
|
25
|
|
26
|
/**
|
27
|
* Discovers RSS or atom feeds at the given URL.
|
28
|
*
|
29
|
* If document in given URL is an HTML document, function attempts to discover
|
30
|
* RSS or Atom feeds.
|
31
|
*
|
32
|
* @param string $url
|
33
|
* The url of the feed to retrieve.
|
34
|
* @param array $settings
|
35
|
* An optional array of settings. Valid options are: accept_invalid_cert.
|
36
|
*
|
37
|
* @return bool|string
|
38
|
* The discovered feed, or FALSE if the URL is not reachable or there was an
|
39
|
* error.
|
40
|
*/
|
41
|
function http_request_get_common_syndication($url, $settings = array()) {
|
42
|
|
43
|
$accept_invalid_cert = isset($settings['accept_invalid_cert']) ? $settings['accept_invalid_cert'] : FALSE;
|
44
|
$download = http_request_get($url, NULL, NULL, $accept_invalid_cert);
|
45
|
|
46
|
// Cannot get the feed, return.
|
47
|
// http_request_get() always returns 200 even if its 304.
|
48
|
if ($download->code != 200) {
|
49
|
return FALSE;
|
50
|
}
|
51
|
|
52
|
// Drop the data into a seperate variable so all manipulations of the html
|
53
|
// will not effect the actual object that exists in the static cache.
|
54
|
// @see http_request_get.
|
55
|
$downloaded_string = $download->data;
|
56
|
// If this happens to be a feed then just return the url.
|
57
|
if (isset($download->headers['content-type']) && http_request_is_feed($download->headers['content-type'], $downloaded_string)) {
|
58
|
return $url;
|
59
|
}
|
60
|
|
61
|
$discovered_feeds = http_request_find_feeds($downloaded_string);
|
62
|
foreach ($discovered_feeds as $feed_url) {
|
63
|
$absolute = http_request_create_absolute_url($feed_url, $url);
|
64
|
if (!empty($absolute)) {
|
65
|
// @TODO: something more intelligent?
|
66
|
return $absolute;
|
67
|
}
|
68
|
}
|
69
|
}
|
70
|
|
71
|
/**
|
72
|
* Get the content from the given URL.
|
73
|
*
|
74
|
* @param string $url
|
75
|
* A valid URL (not only web URLs).
|
76
|
* @param string $username
|
77
|
* If the URL uses authentication, supply the username.
|
78
|
* @param string $password
|
79
|
* If the URL uses authentication, supply the password.
|
80
|
* @param bool $accept_invalid_cert
|
81
|
* Whether to accept invalid certificates.
|
82
|
* @param integer $timeout
|
83
|
* Timeout in seconds to wait for an HTTP get request to finish.
|
84
|
*
|
85
|
* @return stdClass
|
86
|
* An object that describes the data downloaded from $url.
|
87
|
*/
|
88
|
function http_request_get($url, $username = NULL, $password = NULL, $accept_invalid_cert = FALSE, $timeout = NULL) {
|
89
|
// Intra-pagedownload cache, avoid to download the same content twice within
|
90
|
// one page download (it's possible, compatible and parse calls).
|
91
|
static $download_cache = array();
|
92
|
if (isset($download_cache[$url])) {
|
93
|
return $download_cache[$url];
|
94
|
}
|
95
|
|
96
|
// Determine request timeout.
|
97
|
$request_timeout = !empty($timeout) ? $timeout : variable_get('http_request_timeout', 30);
|
98
|
|
99
|
if (!$username && valid_url($url, TRUE)) {
|
100
|
// Handle password protected feeds.
|
101
|
$url_parts = parse_url($url);
|
102
|
if (!empty($url_parts['user'])) {
|
103
|
$password = urldecode($url_parts['pass']);
|
104
|
$username = urldecode($url_parts['user']);
|
105
|
}
|
106
|
}
|
107
|
|
108
|
$curl = http_request_use_curl();
|
109
|
|
110
|
// Only download and parse data if really needs refresh.
|
111
|
// Based on "Last-Modified" and "If-Modified-Since".
|
112
|
$headers = array();
|
113
|
if ($cache = http_request_get_cache($url)) {
|
114
|
$last_result = $cache->data;
|
115
|
$last_headers = array_change_key_case($last_result->headers);
|
116
|
|
117
|
if (!empty($last_headers['etag'])) {
|
118
|
if ($curl) {
|
119
|
$headers[] = 'If-None-Match: ' . $last_headers['etag'];
|
120
|
}
|
121
|
else {
|
122
|
$headers['If-None-Match'] = $last_headers['etag'];
|
123
|
}
|
124
|
}
|
125
|
if (!empty($last_headers['last-modified'])) {
|
126
|
if ($curl) {
|
127
|
$headers[] = 'If-Modified-Since: ' . $last_headers['last-modified'];
|
128
|
}
|
129
|
else {
|
130
|
$headers['If-Modified-Since'] = $last_headers['last-modified'];
|
131
|
}
|
132
|
}
|
133
|
if (!empty($username) && !$curl) {
|
134
|
$headers['Authorization'] = 'Basic ' . base64_encode("$username:$password");
|
135
|
}
|
136
|
}
|
137
|
|
138
|
// Support the 'feed' and 'webcal' schemes by converting them into 'http'.
|
139
|
$url = strtr($url, array('feed://' => 'http://', 'webcal://' => 'http://'));
|
140
|
|
141
|
if ($curl) {
|
142
|
$headers[] = 'User-Agent: Drupal (+http://drupal.org/)';
|
143
|
$result = new stdClass();
|
144
|
$result->headers = array();
|
145
|
|
146
|
// Parse the URL and make sure we can handle the schema.
|
147
|
// cURL can only support either http:// or https://.
|
148
|
// CURLOPT_PROTOCOLS is only supported with cURL 7.19.4
|
149
|
$uri = parse_url($url);
|
150
|
if (!isset($uri['scheme'])) {
|
151
|
$result->error = 'missing schema';
|
152
|
$result->code = -1002;
|
153
|
}
|
154
|
else {
|
155
|
switch ($uri['scheme']) {
|
156
|
case 'http':
|
157
|
case 'https':
|
158
|
// Valid scheme.
|
159
|
break;
|
160
|
|
161
|
default:
|
162
|
$result->error = 'invalid schema ' . $uri['scheme'];
|
163
|
$result->code = -1003;
|
164
|
break;
|
165
|
}
|
166
|
}
|
167
|
|
168
|
// If the scheme was valid, continue to request the feed using cURL.
|
169
|
if (empty($result->error)) {
|
170
|
$download = curl_init($url);
|
171
|
curl_setopt($download, CURLOPT_FOLLOWLOCATION, TRUE);
|
172
|
if (!empty($username)) {
|
173
|
curl_setopt($download, CURLOPT_USERPWD, "{$username}:{$password}");
|
174
|
curl_setopt($download, CURLOPT_HTTPAUTH, CURLAUTH_ANY);
|
175
|
}
|
176
|
curl_setopt($download, CURLOPT_HTTPHEADER, $headers);
|
177
|
curl_setopt($download, CURLOPT_HEADER, TRUE);
|
178
|
curl_setopt($download, CURLOPT_RETURNTRANSFER, TRUE);
|
179
|
curl_setopt($download, CURLOPT_ENCODING, '');
|
180
|
curl_setopt($download, CURLOPT_TIMEOUT, $request_timeout);
|
181
|
|
182
|
$proxy_server = variable_get('proxy_server');
|
183
|
|
184
|
if ($proxy_server && _drupal_http_use_proxy($uri['host'])) {
|
185
|
curl_setopt($download, CURLOPT_PROXY, $proxy_server);
|
186
|
curl_setopt($download, CURLOPT_PROXYPORT, variable_get('proxy_port', 8080));
|
187
|
|
188
|
// Proxy user/password.
|
189
|
if ($proxy_username = variable_get('proxy_username')) {
|
190
|
$username_password = $proxy_username . ':' . variable_get('proxy_password', '');
|
191
|
|
192
|
curl_setopt($download, CURLOPT_PROXYUSERPWD, $username_password);
|
193
|
curl_setopt($download, CURLOPT_PROXYAUTH, variable_get('proxy_auth_method', CURLAUTH_BASIC));
|
194
|
}
|
195
|
}
|
196
|
|
197
|
if ($accept_invalid_cert) {
|
198
|
curl_setopt($download, CURLOPT_SSL_VERIFYPEER, 0);
|
199
|
}
|
200
|
$header = '';
|
201
|
$data = curl_exec($download);
|
202
|
if (curl_error($download)) {
|
203
|
throw new HRCurlException(
|
204
|
t('cURL error (@code) @error for @url', array(
|
205
|
'@code' => curl_errno($download),
|
206
|
'@error' => curl_error($download),
|
207
|
'@url' => $url,
|
208
|
)), curl_errno($download)
|
209
|
);
|
210
|
}
|
211
|
|
212
|
// When using a proxy, remove extra data from the header which is not
|
213
|
// considered by CURLINFO_HEADER_SIZE (possibly cURL bug).
|
214
|
// This data is only added when to HTTP header when working with a proxy.
|
215
|
// Example string added: <HTTP/1.0 200 Connection established\r\n\r\n>
|
216
|
// This was fixed in libcurl version 7.30.0 (0x71e00) (April 12, 2013),
|
217
|
// so this workaround only removes the proxy-added headers if we are using
|
218
|
// an older version of libcurl.
|
219
|
$curl_ver = curl_version();
|
220
|
|
221
|
if ($proxy_server && $curl_ver['version_number'] < 0x71e00 && _drupal_http_use_proxy($uri['host'])) {
|
222
|
$http_header_break = "\r\n\r\n";
|
223
|
$response = explode($http_header_break, $data);
|
224
|
if (count($response) > 2) {
|
225
|
$data = substr($data, strlen($response[0] . $http_header_break), strlen($data));
|
226
|
}
|
227
|
}
|
228
|
|
229
|
$header_size = curl_getinfo($download, CURLINFO_HEADER_SIZE);
|
230
|
$header = substr($data, 0, $header_size - 1);
|
231
|
$result->data = substr($data, $header_size);
|
232
|
$headers = preg_split("/(\r\n){2}/", $header);
|
233
|
$header_lines = preg_split("/\r\n|\n|\r/", end($headers));
|
234
|
// Skip HTTP response status.
|
235
|
array_shift($header_lines);
|
236
|
|
237
|
while ($line = trim(array_shift($header_lines))) {
|
238
|
list($header, $value) = explode(':', $line, 2);
|
239
|
// Normalize the headers.
|
240
|
$header = strtolower($header);
|
241
|
|
242
|
if (isset($result->headers[$header]) && $header == 'set-cookie') {
|
243
|
// RFC 2109: the Set-Cookie response header comprises the token Set-
|
244
|
// Cookie:, followed by a comma-separated list of one or more cookies.
|
245
|
$result->headers[$header] .= ',' . trim($value);
|
246
|
}
|
247
|
else {
|
248
|
$result->headers[$header] = trim($value);
|
249
|
}
|
250
|
}
|
251
|
$result->code = curl_getinfo($download, CURLINFO_HTTP_CODE);
|
252
|
|
253
|
curl_close($download);
|
254
|
}
|
255
|
}
|
256
|
else {
|
257
|
$result = drupal_http_request($url, array('headers' => $headers, 'timeout' => $request_timeout));
|
258
|
$result->headers = isset($result->headers) ? $result->headers : array();
|
259
|
}
|
260
|
|
261
|
$result->code = isset($result->code) ? $result->code : 200;
|
262
|
|
263
|
// In case of 304 Not Modified try to return cached data.
|
264
|
if ($result->code == 304) {
|
265
|
|
266
|
if (isset($last_result)) {
|
267
|
$last_result->from_cache = TRUE;
|
268
|
return $last_result;
|
269
|
}
|
270
|
else {
|
271
|
// It's a tragedy, this file must exist and contain good data.
|
272
|
// In this case, clear cache and repeat.
|
273
|
http_request_clear_cache($url);
|
274
|
return http_request_get($url, $username, $password, $accept_invalid_cert, $request_timeout);
|
275
|
}
|
276
|
}
|
277
|
|
278
|
// Set caches.
|
279
|
http_request_set_cache($url, $result);
|
280
|
$download_cache[$url] = $result;
|
281
|
|
282
|
return $result;
|
283
|
}
|
284
|
|
285
|
/**
|
286
|
* Decides if it's possible to use cURL or not.
|
287
|
*
|
288
|
* @return bool
|
289
|
* TRUE if cURL may be used, FALSE otherwise.
|
290
|
*/
|
291
|
function http_request_use_curl() {
|
292
|
// Allow site administrators to choose to not use cURL.
|
293
|
if (variable_get('feeds_never_use_curl', FALSE)) {
|
294
|
return FALSE;
|
295
|
}
|
296
|
|
297
|
// Check that the PHP cURL extension has been enabled.
|
298
|
if (!extension_loaded('curl')) {
|
299
|
return FALSE;
|
300
|
}
|
301
|
|
302
|
// cURL below PHP 5.6.0 must not have open_basedir or safe_mode enabled.
|
303
|
if (version_compare(PHP_VERSION, '5.6.0', '<')) {
|
304
|
return !ini_get('safe_mode') && !ini_get('open_basedir');
|
305
|
}
|
306
|
|
307
|
// cURL in PHP 5.6.0 and above re-enables CURLOPT_FOLLOWLOCATION with
|
308
|
// open_basedir so there is no need to check for this.
|
309
|
return TRUE;
|
310
|
}
|
311
|
|
312
|
/**
|
313
|
* Clear cache for a specific URL.
|
314
|
*
|
315
|
* @param string $url
|
316
|
* The URL to clear.
|
317
|
*/
|
318
|
function http_request_clear_cache($url) {
|
319
|
cache_clear_all(hash('sha256', $url), 'cache_feeds_http');
|
320
|
}
|
321
|
|
322
|
/**
|
323
|
* Gets the cache for a specific URL.
|
324
|
*
|
325
|
* @param string $url
|
326
|
* The URL to find the cached item.
|
327
|
*
|
328
|
* @return object|false
|
329
|
* The cache or FALSE on failure.
|
330
|
*/
|
331
|
function http_request_get_cache($url) {
|
332
|
return cache_get(hash('sha256', $url), 'cache_feeds_http');
|
333
|
}
|
334
|
|
335
|
/**
|
336
|
* Sets the cache for a specific URL.
|
337
|
*
|
338
|
* @param string $url
|
339
|
* The URL to cache.
|
340
|
* @param stdClass $result
|
341
|
* The result of the HTTP request.
|
342
|
*/
|
343
|
function http_request_set_cache($url, stdClass $result) {
|
344
|
cache_set(hash('sha256', $url), $result, 'cache_feeds_http');
|
345
|
}
|
346
|
|
347
|
/**
|
348
|
* Returns if the provided $content_type is a feed.
|
349
|
*
|
350
|
* @param string $content_type
|
351
|
* The Content-Type header.
|
352
|
*
|
353
|
* @param string $data
|
354
|
* The actual data from the http request.
|
355
|
*
|
356
|
* @return bool
|
357
|
* Returns TRUE if this is a parsable feed.
|
358
|
*/
|
359
|
function http_request_is_feed($content_type, $data) {
|
360
|
$pos = strpos($content_type, ';');
|
361
|
if ($pos !== FALSE) {
|
362
|
$content_type = substr($content_type, 0, $pos);
|
363
|
}
|
364
|
$content_type = strtolower($content_type);
|
365
|
if (strpos($content_type, 'xml') !== FALSE) {
|
366
|
return TRUE;
|
367
|
}
|
368
|
|
369
|
// @TODO: Sometimes the content-type can be text/html but still be a valid
|
370
|
// feed.
|
371
|
return FALSE;
|
372
|
}
|
373
|
|
374
|
/**
|
375
|
* Finds potential feed tags in the HTML document.
|
376
|
*
|
377
|
* @param string $html
|
378
|
* The html string to search.
|
379
|
*
|
380
|
* @return array
|
381
|
* An array of href to feeds.
|
382
|
*/
|
383
|
function http_request_find_feeds($html) {
|
384
|
$matches = array();
|
385
|
preg_match_all(HTTP_REQUEST_PCRE_LINK_TAG, $html, $matches);
|
386
|
$links = $matches[1];
|
387
|
$valid_links = array();
|
388
|
|
389
|
// Build up all the links information.
|
390
|
foreach ($links as $link_tag) {
|
391
|
$attributes = array();
|
392
|
$candidate = array();
|
393
|
|
394
|
preg_match_all(HTTP_REQUEST_PCRE_TAG_ATTRIBUTES, $link_tag, $attributes, PREG_SET_ORDER);
|
395
|
foreach ($attributes as $attribute) {
|
396
|
// Find the key value pairs, attribute[1] is key and attribute[2] is the
|
397
|
// value. However, if the link tag used single quotes, the value might
|
398
|
// be in attribute[3] instead.
|
399
|
if (empty($attribute[2])) {
|
400
|
$attribute[2] = $attribute[3];
|
401
|
}
|
402
|
if (!empty($attribute[1]) && !empty($attribute[2])) {
|
403
|
$candidate[drupal_strtolower($attribute[1])] = drupal_strtolower(decode_entities($attribute[2]));
|
404
|
}
|
405
|
}
|
406
|
|
407
|
// Examine candidate to see if it s a feed.
|
408
|
// @TODO: could/should use http_request_is_feed ??
|
409
|
if (isset($candidate['rel']) && $candidate['rel'] == 'alternate') {
|
410
|
if (isset($candidate['href']) && isset($candidate['type']) && strpos($candidate['type'], 'xml') !== FALSE) {
|
411
|
// All tests pass, its a valid candidate.
|
412
|
$valid_links[] = $candidate['href'];
|
413
|
}
|
414
|
}
|
415
|
}
|
416
|
|
417
|
return $valid_links;
|
418
|
}
|
419
|
|
420
|
/**
|
421
|
* Create an absolute url.
|
422
|
*
|
423
|
* @param string $url
|
424
|
* The href to transform.
|
425
|
* @param string $base_url
|
426
|
* The url to be used as the base for a relative $url.
|
427
|
*
|
428
|
* @return string
|
429
|
* An absolute url
|
430
|
*/
|
431
|
function http_request_create_absolute_url($url, $base_url) {
|
432
|
$url = trim($url);
|
433
|
if (valid_url($url, TRUE)) {
|
434
|
// Valid absolute url already.
|
435
|
return $url;
|
436
|
}
|
437
|
|
438
|
// Turn relative url into absolute.
|
439
|
if (valid_url($url, FALSE)) {
|
440
|
// Produces variables $scheme, $host, $user, $pass, $path, $query and
|
441
|
// $fragment.
|
442
|
$parsed_url = parse_url($base_url);
|
443
|
|
444
|
$path = dirname($parsed_url['path']);
|
445
|
|
446
|
// Adding to the existing path.
|
447
|
if ($url{0} == '/') {
|
448
|
$cparts = array_filter(explode("/", $url));
|
449
|
}
|
450
|
else {
|
451
|
// Backtracking from the existing path.
|
452
|
$cparts = array_merge(array_filter(explode("/", $path)), array_filter(explode("/", $url)));
|
453
|
foreach ($cparts as $i => $part) {
|
454
|
if ($part == '.') {
|
455
|
$cparts[$i] = NULL;
|
456
|
}
|
457
|
if ($part == '..') {
|
458
|
$cparts[$i - 1] = NULL;
|
459
|
$cparts[$i] = NULL;
|
460
|
}
|
461
|
}
|
462
|
$cparts = array_filter($cparts);
|
463
|
}
|
464
|
$path = implode("/", $cparts);
|
465
|
|
466
|
// Build the prefix to the path.
|
467
|
$absolute_url = '';
|
468
|
if (isset($parsed_url['scheme'])) {
|
469
|
$absolute_url = $parsed_url['scheme'] . '://';
|
470
|
}
|
471
|
|
472
|
if (isset($parsed_url['user'])) {
|
473
|
$absolute_url .= $parsed_url['user'];
|
474
|
if (isset($pass)) {
|
475
|
$absolute_url .= ':' . $parsed_url['pass'];
|
476
|
}
|
477
|
$absolute_url .= '@';
|
478
|
}
|
479
|
if (isset($parsed_url['host'])) {
|
480
|
$absolute_url .= $parsed_url['host'] . '/';
|
481
|
}
|
482
|
|
483
|
$absolute_url .= $path;
|
484
|
|
485
|
if (valid_url($absolute_url, TRUE)) {
|
486
|
return $absolute_url;
|
487
|
}
|
488
|
}
|
489
|
return FALSE;
|
490
|
}
|