Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
Total | |
78.00% |
78 / 100 |
|
85.71% |
6 / 7 |
CRAP | |
0.00% |
0 / 1 |
GuzzleHttpRequest | |
78.00% |
78 / 100 |
|
85.71% |
6 / 7 |
51.58 | |
0.00% |
0 / 1 |
__construct | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
3 | |||
setCallback | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
1 | |||
doSetCallback | |
100.00% |
3 / 3 |
|
100.00% |
1 / 1 |
2 | |||
execute | |
72.84% |
59 / 81 |
|
0.00% |
0 / 1 |
35.54 | |||
prepare | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
1 | |||
usingCurl | |
100.00% |
2 / 2 |
|
100.00% |
1 / 1 |
3 | |||
parseHeader | |
100.00% |
5 / 5 |
|
100.00% |
1 / 1 |
3 |
1 | <?php |
2 | /** |
3 | * This program is free software; you can redistribute it and/or modify |
4 | * it under the terms of the GNU General Public License as published by |
5 | * the Free Software Foundation; either version 2 of the License, or |
6 | * (at your option) any later version. |
7 | * |
8 | * This program is distributed in the hope that it will be useful, |
9 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
10 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
11 | * GNU General Public License for more details. |
12 | * |
13 | * You should have received a copy of the GNU General Public License along |
14 | * with this program; if not, write to the Free Software Foundation, Inc., |
15 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
16 | * http://www.gnu.org/copyleft/gpl.html |
17 | * |
18 | * @file |
19 | */ |
20 | |
21 | use GuzzleHttp\Client; |
22 | use GuzzleHttp\Handler\CurlHandler; |
23 | use GuzzleHttp\HandlerStack; |
24 | use GuzzleHttp\MessageFormatter; |
25 | use GuzzleHttp\Middleware; |
26 | use GuzzleHttp\Psr7\Request; |
27 | use MediaWiki\Status\Status; |
28 | use Psr\Http\Message\RequestInterface; |
29 | use Psr\Log\NullLogger; |
30 | |
31 | /** |
32 | * MWHttpRequest implemented using the Guzzle library |
33 | * |
34 | * @note a new 'sink' option is available as an alternative to callbacks. |
35 | * See: http://docs.guzzlephp.org/en/stable/request-options.html#sink) |
36 | * The 'callback' option remains available as well. If both 'sink' and 'callback' are |
37 | * specified, 'sink' is used. |
38 | * @note Callers may set a custom handler via the 'handler' option. |
39 | * If this is not set, Guzzle will use curl (if available) or PHP streams (otherwise) |
40 | * @note Setting either sslVerifyHost or sslVerifyCert will enable both. |
41 | * Guzzle does not allow them to be set separately. |
42 | * |
43 | * @since 1.33 |
44 | */ |
45 | class GuzzleHttpRequest extends MWHttpRequest { |
46 | public const SUPPORTS_FILE_POSTS = true; |
47 | |
48 | protected $handler = null; |
49 | protected $sink = null; |
50 | /** @var array */ |
51 | protected $guzzleOptions = [ 'http_errors' => false ]; |
52 | |
53 | /** |
54 | * @internal Use HttpRequestFactory |
55 | * |
56 | * @param string $url Url to use. If protocol-relative, will be expanded to an http:// URL |
57 | * @param array $options (optional) extra params to pass (see HttpRequestFactory::create()) |
58 | * @param string $caller The method making this request, for profiling |
59 | * @param Profiler|null $profiler An instance of the profiler for profiling, or null |
60 | * @throws Exception |
61 | */ |
62 | public function __construct( |
63 | $url, array $options = [], $caller = __METHOD__, Profiler $profiler = null |
64 | ) { |
65 | parent::__construct( $url, $options, $caller, $profiler ); |
66 | |
67 | if ( isset( $options['handler'] ) ) { |
68 | $this->handler = $options['handler']; |
69 | } |
70 | if ( isset( $options['sink'] ) ) { |
71 | $this->sink = $options['sink']; |
72 | } |
73 | } |
74 | |
75 | /** |
76 | * Set a read callback to accept data read from the HTTP request. |
77 | * By default, data is appended to an internal buffer which can be |
78 | * retrieved through $req->getContent(). |
79 | * |
80 | * To handle data as it comes in -- especially for large files that |
81 | * would not fit in memory -- you can instead set your own callback, |
82 | * in the form function($resource, $buffer) where the first parameter |
83 | * is the low-level resource being read (implementation specific), |
84 | * and the second parameter is the data buffer. |
85 | * |
86 | * You MUST return the number of bytes handled in the buffer; if fewer |
87 | * bytes are reported handled than were passed to you, the HTTP fetch |
88 | * will be aborted. |
89 | * |
90 | * This function overrides any 'sink' or 'callback' constructor option. |
91 | * |
92 | * @param callable|null $callback |
93 | * @throws InvalidArgumentException |
94 | */ |
95 | public function setCallback( $callback ) { |
96 | $this->sink = null; |
97 | $this->doSetCallback( $callback ); |
98 | } |
99 | |
100 | /** |
101 | * Worker function for setting callbacks. Calls can originate both internally and externally |
102 | * via setCallback). Defaults to the internal read callback if $callback is null. |
103 | * |
104 | * If a sink is already specified, this does nothing. This causes the 'sink' constructor |
105 | * option to override the 'callback' constructor option. |
106 | * |
107 | * @param callable|null $callback |
108 | * @throws InvalidArgumentException |
109 | */ |
110 | protected function doSetCallback( $callback ) { |
111 | if ( !$this->sink ) { |
112 | parent::doSetCallback( $callback ); |
113 | $this->sink = new MWCallbackStream( $this->callback ); |
114 | } |
115 | } |
116 | |
117 | /** |
118 | * @see MWHttpRequest::execute |
119 | * |
120 | * @return Status |
121 | */ |
122 | public function execute() { |
123 | $this->prepare(); |
124 | |
125 | if ( !$this->status->isOK() ) { |
126 | return Status::wrap( $this->status ); // TODO B/C; move this to callers |
127 | } |
128 | |
129 | if ( $this->proxy ) { |
130 | $this->guzzleOptions['proxy'] = $this->proxy; |
131 | } |
132 | |
133 | $this->guzzleOptions['timeout'] = $this->timeout; |
134 | $this->guzzleOptions['connect_timeout'] = $this->connectTimeout; |
135 | $this->guzzleOptions['version'] = '1.1'; |
136 | |
137 | if ( !$this->followRedirects ) { |
138 | $this->guzzleOptions['allow_redirects'] = false; |
139 | } else { |
140 | $this->guzzleOptions['allow_redirects'] = [ |
141 | 'max' => $this->maxRedirects |
142 | ]; |
143 | } |
144 | |
145 | if ( $this->method == 'POST' ) { |
146 | $postData = $this->postData; |
147 | if ( is_array( $postData ) ) { |
148 | $this->guzzleOptions['form_params'] = $postData; |
149 | } else { |
150 | $this->guzzleOptions['body'] = $postData; |
151 | // mimic CURLOPT_POST option |
152 | if ( !isset( $this->reqHeaders['Content-Type'] ) ) { |
153 | $this->reqHeaders['Content-Type'] = 'application/x-www-form-urlencoded'; |
154 | } |
155 | } |
156 | |
157 | // Suppress 'Expect: 100-continue' header, as some servers |
158 | // will reject it with a 417 and Curl won't auto retry |
159 | // with HTTP 1.0 fallback |
160 | $this->guzzleOptions['expect'] = false; |
161 | } |
162 | |
163 | $stack = HandlerStack::create( $this->handler ); |
164 | |
165 | // Create Middleware to use cookies from $this->getCookieJar(), |
166 | // which is in MediaWiki CookieJar format, not in Guzzle-specific CookieJar format. |
167 | // Note: received cookies (from HTTP response) don't need to be handled here, |
168 | // they will be added back into the CookieJar by MWHttpRequest::parseCookies(). |
169 | // @phan-suppress-next-line PhanUndeclaredFunctionInCallable |
170 | $stack->remove( 'cookies' ); |
171 | $mwCookieJar = $this->getCookieJar(); |
172 | $stack->push( Middleware::mapRequest( |
173 | static function ( RequestInterface $request ) use ( $mwCookieJar ) { |
174 | $uri = $request->getUri(); |
175 | $cookieHeader = $mwCookieJar->serializeToHttpRequest( |
176 | $uri->getPath() ?: '/', |
177 | $uri->getHost() |
178 | ); |
179 | if ( !$cookieHeader ) { |
180 | return $request; |
181 | } |
182 | |
183 | return $request->withHeader( 'Cookie', $cookieHeader ); |
184 | } |
185 | ), 'cookies' ); |
186 | |
187 | if ( !$this->logger instanceof NullLogger ) { |
188 | $stack->push( Middleware::log( $this->logger, new MessageFormatter( |
189 | // TODO {error} will be 'NULL' on success which is unfortunate, but |
190 | // doesn't seem fixable without a custom formatter. Same for using |
191 | // PSR-3 variable replacement instead of raw strings. |
192 | '[{ts}] {method} {uri} HTTP/{version} - {code} {error}' |
193 | ) ), 'logger' ); |
194 | } |
195 | |
196 | $this->guzzleOptions['handler'] = $stack; |
197 | |
198 | if ( $this->sink ) { |
199 | $this->guzzleOptions['sink'] = $this->sink; |
200 | } |
201 | |
202 | if ( $this->caInfo ) { |
203 | $this->guzzleOptions['verify'] = $this->caInfo; |
204 | } elseif ( !$this->sslVerifyHost && !$this->sslVerifyCert ) { |
205 | $this->guzzleOptions['verify'] = false; |
206 | } |
207 | |
208 | $client = new Client( $this->guzzleOptions ); |
209 | $request = new Request( $this->method, $this->url ); |
210 | foreach ( $this->reqHeaders as $name => $value ) { |
211 | $request = $request->withHeader( $name, $value ); |
212 | } |
213 | |
214 | try { |
215 | $response = $client->send( $request ); |
216 | $this->headerList = $response->getHeaders(); |
217 | |
218 | $this->respVersion = $response->getProtocolVersion(); |
219 | $this->respStatus = $response->getStatusCode() . ' ' . $response->getReasonPhrase(); |
220 | } catch ( GuzzleHttp\Exception\ConnectException $e ) { |
221 | // ConnectException is thrown for several reasons besides generic "timeout": |
222 | // Connection refused |
223 | // couldn't connect to host |
224 | // connection attempt failed |
225 | // Could not resolve IPv4 address for host |
226 | // Could not resolve IPv6 address for host |
227 | if ( $this->usingCurl() ) { |
228 | $handlerContext = $e->getHandlerContext(); |
229 | if ( $handlerContext['errno'] == CURLE_OPERATION_TIMEOUTED ) { |
230 | $this->status->fatal( 'http-timed-out', $this->url ); |
231 | } else { |
232 | $this->status->fatal( 'http-curl-error', $handlerContext['error'] ); |
233 | } |
234 | } else { |
235 | $this->status->fatal( 'http-request-error' ); |
236 | } |
237 | } catch ( GuzzleHttp\Exception\RequestException $e ) { |
238 | if ( $this->usingCurl() ) { |
239 | $handlerContext = $e->getHandlerContext(); |
240 | $this->status->fatal( 'http-curl-error', $handlerContext['error'] ); |
241 | } else { |
242 | // Non-ideal, but the only way to identify connection timeout vs other conditions |
243 | $needle = 'Connection timed out'; |
244 | if ( strpos( $e->getMessage(), $needle ) !== false ) { |
245 | $this->status->fatal( 'http-timed-out', $this->url ); |
246 | } else { |
247 | $this->status->fatal( 'http-request-error' ); |
248 | } |
249 | } |
250 | } catch ( GuzzleHttp\Exception\GuzzleException $e ) { |
251 | $this->status->fatal( 'http-internal-error' ); |
252 | } |
253 | |
254 | if ( $this->profiler ) { |
255 | $profileSection = $this->profiler->scopedProfileIn( |
256 | __METHOD__ . '-' . $this->profileName |
257 | ); |
258 | } |
259 | |
260 | if ( $this->profiler ) { |
261 | $this->profiler->scopedProfileOut( $profileSection ); |
262 | } |
263 | |
264 | $this->parseHeader(); |
265 | $this->setStatus(); |
266 | |
267 | return Status::wrap( $this->status ); // TODO B/C; move this to callers |
268 | } |
269 | |
270 | protected function prepare() { |
271 | $this->doSetCallback( $this->callback ); |
272 | parent::prepare(); |
273 | } |
274 | |
275 | protected function usingCurl(): bool { |
276 | return $this->handler instanceof CurlHandler || |
277 | ( !$this->handler && extension_loaded( 'curl' ) ); |
278 | } |
279 | |
280 | /** |
281 | * Guzzle provides headers as an array. Reprocess to match our expectations. Guzzle will |
282 | * have already parsed and removed the status line (in EasyHandle::createResponse). |
283 | */ |
284 | protected function parseHeader() { |
285 | // Failure without (valid) headers gets a response status of zero |
286 | if ( !$this->status->isOK() ) { |
287 | $this->respStatus = '0 Error'; |
288 | } |
289 | |
290 | foreach ( $this->headerList as $name => $values ) { |
291 | $this->respHeaders[strtolower( $name )] = $values; |
292 | } |
293 | |
294 | $this->parseCookies(); |
295 | } |
296 | } |