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 = NULL) {
|
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 = $url_parts['pass'];
|
104
|
$username = $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 = cache_get('feeds_http_download_' . md5($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, $request_timeout);
|
181
|
if ($accept_invalid_cert) {
|
182
|
curl_setopt($download, CURLOPT_SSL_VERIFYPEER, 0);
|
183
|
}
|
184
|
$header = '';
|
185
|
$data = curl_exec($download);
|
186
|
if (curl_error($download)) {
|
187
|
throw new HRCurlException(
|
188
|
t('cURL error (@code) @error for @url', array(
|
189
|
'@code' => curl_errno($download),
|
190
|
'@error' => curl_error($download),
|
191
|
'@url' => $url,
|
192
|
)), curl_errno($download)
|
193
|
);
|
194
|
}
|
195
|
|
196
|
$header_size = curl_getinfo($download, CURLINFO_HEADER_SIZE);
|
197
|
$header = substr($data, 0, $header_size - 1);
|
198
|
$result->data = substr($data, $header_size);
|
199
|
$headers = preg_split("/(\r\n){2}/", $header);
|
200
|
$header_lines = preg_split("/\r\n|\n|\r/", end($headers));
|
201
|
// Skip HTTP response status.
|
202
|
array_shift($header_lines);
|
203
|
|
204
|
while ($line = trim(array_shift($header_lines))) {
|
205
|
list($header, $value) = explode(':', $line, 2);
|
206
|
// Normalize the headers.
|
207
|
$header = strtolower($header);
|
208
|
|
209
|
if (isset($result->headers[$header]) && $header == 'set-cookie') {
|
210
|
// RFC 2109: the Set-Cookie response header comprises the token Set-
|
211
|
// Cookie:, followed by a comma-separated list of one or more cookies.
|
212
|
$result->headers[$header] .= ',' . trim($value);
|
213
|
}
|
214
|
else {
|
215
|
$result->headers[$header] = trim($value);
|
216
|
}
|
217
|
}
|
218
|
$result->code = curl_getinfo($download, CURLINFO_HTTP_CODE);
|
219
|
|
220
|
curl_close($download);
|
221
|
}
|
222
|
}
|
223
|
else {
|
224
|
$result = drupal_http_request($url, array('headers' => $headers, 'timeout' => $request_timeout));
|
225
|
$result->headers = isset($result->headers) ? $result->headers : array();
|
226
|
}
|
227
|
|
228
|
$result->code = isset($result->code) ? $result->code : 200;
|
229
|
|
230
|
// In case of 304 Not Modified try to return cached data.
|
231
|
if ($result->code == 304) {
|
232
|
|
233
|
if (isset($last_result)) {
|
234
|
$last_result->from_cache = TRUE;
|
235
|
return $last_result;
|
236
|
}
|
237
|
else {
|
238
|
// It's a tragedy, this file must exist and contain good data.
|
239
|
// In this case, clear cache and repeat.
|
240
|
cache_clear_all('feeds_http_download_' . md5($url), 'cache');
|
241
|
return http_request_get($url, $username, $password, $accept_invalid_cert, $request_timeout);
|
242
|
}
|
243
|
}
|
244
|
|
245
|
// Set caches.
|
246
|
cache_set('feeds_http_download_' . md5($url), $result);
|
247
|
$download_cache[$url] = $result;
|
248
|
|
249
|
return $result;
|
250
|
}
|
251
|
|
252
|
/**
|
253
|
* Decides if it's possible to use cURL or not.
|
254
|
*
|
255
|
* @return bool
|
256
|
* TRUE if cURL is available, FALSE otherwise.
|
257
|
*/
|
258
|
function http_request_use_curl() {
|
259
|
// Allow site administrators to choose to not use cURL.
|
260
|
if (variable_get('feeds_never_use_curl', FALSE)) {
|
261
|
return FALSE;
|
262
|
}
|
263
|
|
264
|
// Check availability of cURL on the system.
|
265
|
$basedir = ini_get("open_basedir");
|
266
|
return function_exists('curl_init') && !ini_get('safe_mode') && empty($basedir);
|
267
|
}
|
268
|
|
269
|
/**
|
270
|
* Clear cache for a specific URL.
|
271
|
*/
|
272
|
function http_request_clear_cache($url) {
|
273
|
cache_clear_all('feeds_http_download_' . md5($url), 'cache');
|
274
|
}
|
275
|
|
276
|
/**
|
277
|
* Returns if the provided $content_type is a feed.
|
278
|
*
|
279
|
* @param string $content_type
|
280
|
* The Content-Type header.
|
281
|
*
|
282
|
* @param string $data
|
283
|
* The actual data from the http request.
|
284
|
*
|
285
|
* @return bool
|
286
|
* Returns TRUE if this is a parsable feed.
|
287
|
*/
|
288
|
function http_request_is_feed($content_type, $data) {
|
289
|
$pos = strpos($content_type, ';');
|
290
|
if ($pos !== FALSE) {
|
291
|
$content_type = substr($content_type, 0, $pos);
|
292
|
}
|
293
|
$content_type = strtolower($content_type);
|
294
|
if (strpos($content_type, 'xml') !== FALSE) {
|
295
|
return TRUE;
|
296
|
}
|
297
|
|
298
|
// @TODO: Sometimes the content-type can be text/html but still be a valid
|
299
|
// feed.
|
300
|
return FALSE;
|
301
|
}
|
302
|
|
303
|
/**
|
304
|
* Finds potential feed tags in the HTML document.
|
305
|
*
|
306
|
* @param string $html
|
307
|
* The html string to search.
|
308
|
*
|
309
|
* @return array
|
310
|
* An array of href to feeds.
|
311
|
*/
|
312
|
function http_request_find_feeds($html) {
|
313
|
$matches = array();
|
314
|
preg_match_all(HTTP_REQUEST_PCRE_LINK_TAG, $html, $matches);
|
315
|
$links = $matches[1];
|
316
|
$valid_links = array();
|
317
|
|
318
|
// Build up all the links information.
|
319
|
foreach ($links as $link_tag) {
|
320
|
$attributes = array();
|
321
|
$candidate = array();
|
322
|
|
323
|
preg_match_all(HTTP_REQUEST_PCRE_TAG_ATTRIBUTES, $link_tag, $attributes, PREG_SET_ORDER);
|
324
|
foreach ($attributes as $attribute) {
|
325
|
// Find the key value pairs, attribute[1] is key and attribute[2] is the
|
326
|
// value.
|
327
|
if (!empty($attribute[1]) && !empty($attribute[2])) {
|
328
|
$candidate[drupal_strtolower($attribute[1])] = drupal_strtolower(decode_entities($attribute[2]));
|
329
|
}
|
330
|
}
|
331
|
|
332
|
// Examine candidate to see if it s a feed.
|
333
|
// @TODO: could/should use http_request_is_feed ??
|
334
|
if (isset($candidate['rel']) && $candidate['rel'] == 'alternate') {
|
335
|
if (isset($candidate['href']) && isset($candidate['type']) && strpos($candidate['type'], 'xml') !== FALSE) {
|
336
|
// All tests pass, its a valid candidate.
|
337
|
$valid_links[] = $candidate['href'];
|
338
|
}
|
339
|
}
|
340
|
}
|
341
|
|
342
|
return $valid_links;
|
343
|
}
|
344
|
|
345
|
/**
|
346
|
* Create an absolute url.
|
347
|
*
|
348
|
* @param string $url
|
349
|
* The href to transform.
|
350
|
* @param string $base_url
|
351
|
* The url to be used as the base for a relative $url.
|
352
|
*
|
353
|
* @return string
|
354
|
* An absolute url
|
355
|
*/
|
356
|
function http_request_create_absolute_url($url, $base_url) {
|
357
|
$url = trim($url);
|
358
|
if (valid_url($url, TRUE)) {
|
359
|
// Valid absolute url already.
|
360
|
return $url;
|
361
|
}
|
362
|
|
363
|
// Turn relative url into absolute.
|
364
|
if (valid_url($url, FALSE)) {
|
365
|
// Produces variables $scheme, $host, $user, $pass, $path, $query and
|
366
|
// $fragment.
|
367
|
$parsed_url = parse_url($base_url);
|
368
|
|
369
|
$path = dirname($parsed_url['path']);
|
370
|
|
371
|
// Adding to the existing path.
|
372
|
if ($url{0} == '/') {
|
373
|
$cparts = array_filter(explode("/", $url));
|
374
|
}
|
375
|
else {
|
376
|
// Backtracking from the existing path.
|
377
|
$cparts = array_merge(array_filter(explode("/", $path)), array_filter(explode("/", $url)));
|
378
|
foreach ($cparts as $i => $part) {
|
379
|
if ($part == '.') {
|
380
|
$cparts[$i] = NULL;
|
381
|
}
|
382
|
if ($part == '..') {
|
383
|
$cparts[$i - 1] = NULL;
|
384
|
$cparts[$i] = NULL;
|
385
|
}
|
386
|
}
|
387
|
$cparts = array_filter($cparts);
|
388
|
}
|
389
|
$path = implode("/", $cparts);
|
390
|
|
391
|
// Build the prefix to the path.
|
392
|
$absolute_url = '';
|
393
|
if (isset($parsed_url['scheme'])) {
|
394
|
$absolute_url = $parsed_url['scheme'] . '://';
|
395
|
}
|
396
|
|
397
|
if (isset($parsed_url['user'])) {
|
398
|
$absolute_url .= $parsed_url['user'];
|
399
|
if (isset($pass)) {
|
400
|
$absolute_url .= ':' . $parsed_url['pass'];
|
401
|
}
|
402
|
$absolute_url .= '@';
|
403
|
}
|
404
|
if (isset($parsed_url['host'])) {
|
405
|
$absolute_url .= $parsed_url['host'] . '/';
|
406
|
}
|
407
|
|
408
|
$absolute_url .= $path;
|
409
|
|
410
|
if (valid_url($absolute_url, TRUE)) {
|
411
|
return $absolute_url;
|
412
|
}
|
413
|
}
|
414
|
return FALSE;
|
415
|
}
|