includes/libs/MultiHttpClient.php

   1 <?php
   2 /**
   3  * HTTP service client
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 /**
  24  * Class to handle concurrent HTTP requests
  25  *
  26  * HTTP request maps are arrays that use the following format:
  27  *   - method   : GET/HEAD/PUT/POST/DELETE
  28  *   - url      : HTTP/HTTPS URL
  29  *   - query    : <query parameter field/value associative array> (uses RFC 3986)
  30  *   - headers  : <header name/value associative array>
  31  *   - body     : source to get the HTTP request body from;
  32  *                this can simply be a string (always), a resource for
  33  *                PUT requests, and a field/value array for POST request;
  34  *                array bodies are encoded as multipart/form-data and strings
  35  *                use application/x-www-form-urlencoded (headers sent automatically)
  36  *   - stream   : resource to stream the HTTP response body to
  37  * Request maps can use integer index 0 instead of 'method' and 1 instead of 'url'.
  38  *
  39  * @author Aaron Schulz
  40  * @since 1.23
  41  */
  42 class MultiHttpClient {
  43         /** @var resource */
  44         protected $multiHandle = null; // curl_multi handle
  45         /** @var string|null SSL certificates path  */
  46         protected $caBundlePath;
  47         /** @var integer */
  48         protected $connTimeout = 10;
  49         /** @var integer */
  50         protected $reqTimeout = 300;
  51         /** @var bool */
  52         protected $usePipelining = false;
  53         /** @var integer */
  54         protected $maxConnsPerHost = 50;
  55
  56         /**
  57          * @param array $options
  58          *   - connTimeout     : default connection timeout
  59          *   - reqTimeout      : default request timeout
  60          *   - usePipelining   : whether to use HTTP pipelining if possible (for all hosts)
  61          *   - maxConnsPerHost : maximum number of concurrent connections (per host)
  62          * @throws Exception
  63          */
  64         public function __construct( array $options ) {
  65                 if ( isset( $options['caBundlePath'] ) ) {
  66                         $this->caBundlePath = $options['caBundlePath'];
  67                         if ( !file_exists( $this->caBundlePath ) ) {
  68                                 throw new Exception( "Cannot find CA bundle: " . $this->caBundlePath );
  69                         }
  70                 }
  71                 static $opts = array( 'connTimeout', 'reqTimeout', 'usePipelining', 'maxConnsPerHost' );
  72                 foreach ( $opts as $key ) {
  73                         if ( isset( $options[$key] ) ) {
  74                                 $this->$key = $options[$key];
  75                         }
  76                 }
  77         }
  78
  79         /**
  80          * Execute an HTTP(S) request
  81          *
  82          * This method returns a response map of:
  83          *   - code    : HTTP response code or 0 if there was a serious cURL error
  84          *   - reason  : HTTP response reason (empty if there was a serious cURL error)
  85          *   - headers : <header name/value associative array>
  86          *   - body    : HTTP response body or resource (if "stream" was set)
  87          *   - error     : Any cURL error string
  88          * The map also stores integer-indexed copies of these values. This lets callers do:
  89          *      <code>
  90          *              list( $rcode, $rdesc, $rhdrs, $rbody, $rerr ) = $http->run( $req );
  91          *  </code>
  92          * @param array $req HTTP request array
  93          * @param array $opts
  94          *   - connTimeout    : connection timeout per request
  95          *   - reqTimeout     : post-connection timeout per request
  96          * @return array Response array for request
  97          */
  98         final public function run( array $req, array $opts = array() ) {
  99                 $req = $this->runMulti( array( $req ), $opts );
 100                 return $req[0]['response'];
 101         }
 102
 103         /**
 104          * Execute a set of HTTP(S) requests concurrently
 105          *
 106          * The maps are returned by this method with the 'response' field set to a map of:
 107          *   - code    : HTTP response code or 0 if there was a serious cURL error
 108          *   - reason  : HTTP response reason (empty if there was a serious cURL error)
 109          *   - headers : <header name/value associative array>
 110          *   - body    : HTTP response body or resource (if "stream" was set)
 111          *   - error   : Any cURL error string
 112          * The map also stores integer-indexed copies of these values. This lets callers do:
 113          *    <code>
 114          *        list( $rcode, $rdesc, $rhdrs, $rbody, $rerr ) = $req['response'];
 115          *  </code>
 116          * All headers in the 'headers' field are normalized to use lower case names.
 117          * This is true for the request headers and the response headers. Integer-indexed
 118          * method/URL entries will also be changed to use the corresponding string keys.
 119          *
 120          * @param array $reqs Map of HTTP request arrays
 121          * @param array $opts
 122          *   - connTimeout     : connection timeout per request
 123          *   - reqTimeout      : post-connection timeout per request
 124          *   - usePipelining   : whether to use HTTP pipelining if possible
 125          *   - maxConnsPerHost : maximum number of concurrent connections (per host)
 126          * @return array $reqs With response array populated for each
 127          * @throws Exception
 128          */
 129         public function runMulti( array $reqs, array $opts = array() ) {
 130                 $chm = $this->getCurlMulti();
 131
 132                 // Normalize $reqs and add all of the required cURL handles...
 133                 $handles = array();
 134                 foreach ( $reqs as $index => &$req ) {
 135                         $req['response'] = array(
 136                                 'code'     => 0,
 137                                 'reason'   => '',
 138                                 'headers'  => array(),
 139                                 'body'     => '',
 140                                 'error'    => ''
 141                         );
 142                         if ( isset( $req[0] ) ) {
 143                                 $req['method'] = $req[0]; // short-form
 144                                 unset( $req[0] );
 145                         }
 146                         if ( isset( $req[1] ) ) {
 147                                 $req['url'] = $req[1]; // short-form
 148                                 unset( $req[1] );
 149                         }
 150                         if ( !isset( $req['method'] ) ) {
 151                                 throw new Exception( "Request has no 'method' field set." );
 152                         } elseif ( !isset( $req['url'] ) ) {
 153                                 throw new Exception( "Request has no 'url' field set." );
 154                         }
 155                         $req['query'] = isset( $req['query'] ) ? $req['query'] : array();
 156                         $headers = array(); // normalized headers
 157                         if ( isset( $req['headers'] ) ) {
 158                                 foreach ( $req['headers'] as $name => $value ) {
 159                                         $headers[strtolower( $name )] = $value;
 160                                 }
 161                         }
 162                         $req['headers'] = $headers;
 163                         if ( !isset( $req['body'] ) ) {
 164                                 $req['body'] = '';
 165                                 $req['headers']['content-length'] = 0;
 166                         }
 167                         $handles[$index] = $this->getCurlHandle( $req, $opts );
 168                         if ( count( $reqs ) > 1 ) {
 169                                 // https://github.com/guzzle/guzzle/issues/349
 170                                 curl_setopt( $handles[$index], CURLOPT_FORBID_REUSE, true );
 171                         }
 172                 }
 173                 unset( $req ); // don't assign over this by accident
 174
 175                 $indexes = array_keys( $reqs );
 176                 if ( function_exists( 'curl_multi_setopt' ) ) { // PHP 5.5
 177                         if ( isset( $opts['usePipelining'] ) ) {
 178                                 curl_multi_setopt( $chm, CURLMOPT_PIPELINING, (int)$opts['usePipelining'] );
 179                         }
 180                         if ( isset( $opts['maxConnsPerHost'] ) ) {
 181                                 // Keep these sockets around as they may be needed later in the request
 182                                 curl_multi_setopt( $chm, CURLMOPT_MAXCONNECTS, (int)$opts['maxConnsPerHost'] );
 183                         }
 184                 }
 185
 186                 // @TODO: use a per-host rolling handle window (e.g. CURLMOPT_MAX_HOST_CONNECTIONS)
 187                 $batches = array_chunk( $indexes, $this->maxConnsPerHost );
 188
 189                 foreach ( $batches as $batch ) {
 190                         // Attach all cURL handles for this batch
 191                         foreach ( $batch as $index ) {
 192                                 curl_multi_add_handle( $chm, $handles[$index] );
 193                         }
 194                         // Execute the cURL handles concurrently...
 195                         $active = null; // handles still being processed
 196                         do {
 197                                 // Do any available work...
 198                                 do {
 199                                         $mrc = curl_multi_exec( $chm, $active );
 200                                 } while ( $mrc == CURLM_CALL_MULTI_PERFORM );
 201                                 // Wait (if possible) for available work...
 202                                 if ( $active > 0 && $mrc == CURLM_OK ) {
 203                                         if ( curl_multi_select( $chm, 10 ) == -1 ) {
 204                                                 // PHP bug 63411; http://curl.haxx.se/libcurl/c/curl_multi_fdset.html
 205                                                 usleep( 5000 ); // 5ms
 206                                         }
 207                                 }
 208                         } while ( $active > 0 && $mrc == CURLM_OK );
 209                 }
 210
 211                 // Remove all of the added cURL handles and check for errors...
 212                 foreach ( $reqs as $index => &$req ) {
 213                         $ch = $handles[$index];
 214                         curl_multi_remove_handle( $chm, $ch );
 215                         if ( curl_errno( $ch ) !== 0 ) {
 216                                 $req['response']['error'] = "(curl error: " .
 217                                         curl_errno( $ch ) . ") " . curl_error( $ch );
 218                         }
 219                         // For convenience with the list() operator
 220                         $req['response'][0] = $req['response']['code'];
 221                         $req['response'][1] = $req['response']['reason'];
 222                         $req['response'][2] = $req['response']['headers'];
 223                         $req['response'][3] = $req['response']['body'];
 224                         $req['response'][4] = $req['response']['error'];
 225                         curl_close( $ch );
 226                         // Close any string wrapper file handles
 227                         if ( isset( $req['_closeHandle'] ) ) {
 228                                 fclose( $req['_closeHandle'] );
 229                                 unset( $req['_closeHandle'] );
 230                         }
 231                 }
 232                 unset( $req ); // don't assign over this by accident
 233
 234                 // Restore the default settings
 235                 if ( function_exists( 'curl_multi_setopt' ) ) { // PHP 5.5
 236                         curl_multi_setopt( $chm, CURLMOPT_PIPELINING, (int)$this->usePipelining );
 237                         curl_multi_setopt( $chm, CURLMOPT_MAXCONNECTS, (int)$this->maxConnsPerHost );
 238                 }
 239
 240                 return $reqs;
 241         }
 242
 243         /**
 244          * @param array $req HTTP request map
 245          * @param array $opts
 246          *   - connTimeout    : default connection timeout
 247          *   - reqTimeout     : default request timeout
 248          * @return resource
 249          * @throws Exception
 250          */
 251         protected function getCurlHandle( array &$req, array $opts = array() ) {
 252                 $ch = curl_init();
 253
 254                 curl_setopt( $ch, CURLOPT_CONNECTTIMEOUT,
 255                         isset( $opts['connTimeout'] ) ? $opts['connTimeout'] : $this->connTimeout );
 256                 curl_setopt( $ch, CURLOPT_TIMEOUT,
 257                         isset( $opts['reqTimeout'] ) ? $opts['reqTimeout'] : $this->reqTimeout );
 258                 curl_setopt( $ch, CURLOPT_FOLLOWLOCATION, 1 );
 259                 curl_setopt( $ch, CURLOPT_MAXREDIRS, 4 );
 260                 curl_setopt( $ch, CURLOPT_HEADER, 0 );
 261                 if ( !is_null( $this->caBundlePath ) ) {
 262                         curl_setopt( $ch, CURLOPT_SSL_VERIFYPEER, true );
 263                         curl_setopt( $ch, CURLOPT_CAINFO, $this->caBundlePath );
 264                 }
 265                 curl_setopt( $ch, CURLOPT_RETURNTRANSFER, 1 );
 266
 267                 $url = $req['url'];
 268                 // PHP_QUERY_RFC3986 is PHP 5.4+ only
 269                 $query = str_replace(
 270                         array( '+', '%7E' ),
 271                         array( '%20', '~' ),
 272                         http_build_query( $req['query'], '', '&' )
 273                 );
 274                 if ( $query != '' ) {
 275                         $url .= strpos( $req['url'], '?' ) === false ? "?$query" : "&$query";
 276                 }
 277                 curl_setopt( $ch, CURLOPT_URL, $url );
 278
 279                 curl_setopt( $ch, CURLOPT_CUSTOMREQUEST, $req['method'] );
 280                 if ( $req['method'] === 'HEAD' ) {
 281                         curl_setopt( $ch, CURLOPT_NOBODY, 1 );
 282                 }
 283
 284                 if ( $req['method'] === 'PUT' ) {
 285                         curl_setopt( $ch, CURLOPT_PUT, 1 );
 286                         if ( is_resource( $req['body'] ) ) {
 287                                 curl_setopt( $ch, CURLOPT_INFILE, $req['body'] );
 288                                 if ( isset( $req['headers']['content-length'] ) ) {
 289                                         curl_setopt( $ch, CURLOPT_INFILESIZE, $req['headers']['content-length'] );
 290                                 } elseif ( isset( $req['headers']['transfer-encoding'] ) &&
 291                                         $req['headers']['transfer-encoding'] === 'chunks'
 292                                 ) {
 293                                         curl_setopt( $ch, CURLOPT_UPLOAD, true );
 294                                 } else {
 295                                         throw new Exception( "Missing 'Content-Length' or 'Transfer-Encoding' header." );
 296                                 }
 297                         } elseif ( $req['body'] !== '' ) {
 298                                 $fp = fopen( "php://temp", "wb+" );
 299                                 fwrite( $fp, $req['body'], strlen( $req['body'] ) );
 300                                 rewind( $fp );
 301                                 curl_setopt( $ch, CURLOPT_INFILE, $fp );
 302                                 curl_setopt( $ch, CURLOPT_INFILESIZE, strlen( $req['body'] ) );
 303                                 $req['_closeHandle'] = $fp; // remember to close this later
 304                         } else {
 305                                 curl_setopt( $ch, CURLOPT_INFILESIZE, 0 );
 306                         }
 307                         curl_setopt( $ch, CURLOPT_READFUNCTION,
 308                                 function ( $ch, $fd, $length ) {
 309                                         $data = fread( $fd, $length );
 310                                         $len = strlen( $data );
 311                                         return $data;
 312                                 }
 313                         );
 314                 } elseif ( $req['method'] === 'POST' ) {
 315                         curl_setopt( $ch, CURLOPT_POST, 1 );
 316                         curl_setopt( $ch, CURLOPT_POSTFIELDS, $req['body'] );
 317                 } else {
 318                         if ( is_resource( $req['body'] ) || $req['body'] !== '' ) {
 319                                 throw new Exception( "HTTP body specified for a non PUT/POST request." );
 320                         }
 321                         $req['headers']['content-length'] = 0;
 322                 }
 323
 324                 $headers = array();
 325                 foreach ( $req['headers'] as $name => $value ) {
 326                         if ( strpos( $name, ': ' ) ) {
 327                                 throw new Exception( "Headers cannot have ':' in the name." );
 328                         }
 329                         $headers[] = $name . ': ' . trim( $value );
 330                 }
 331                 curl_setopt( $ch, CURLOPT_HTTPHEADER, $headers );
 332
 333                 curl_setopt( $ch, CURLOPT_HEADERFUNCTION,
 334                         function ( $ch, $header ) use ( &$req ) {
 335                                 $length = strlen( $header );
 336                                 $matches = array();
 337                                 if ( preg_match( "/^(HTTP\/1\.[01]) (\d{3}) (.*)/", $header, $matches ) ) {
 338                                         $req['response']['code'] = (int)$matches[2];
 339                                         $req['response']['reason'] = trim( $matches[3] );
 340                                         return $length;
 341                                 }
 342                                 if ( strpos( $header, ":" ) === false ) {
 343                                         return $length;
 344                                 }
 345                                 list( $name, $value ) = explode( ":", $header, 2 );
 346                                 $req['response']['headers'][strtolower( $name )] = trim( $value );
 347                                 return $length;
 348                         }
 349                 );
 350
 351                 if ( isset( $req['stream'] ) ) {
 352                         // Don't just use CURLOPT_FILE as that might give:
 353                         // curl_setopt(): cannot represent a stream of type Output as a STDIO FILE*
 354                         // The callback here handles both normal files and php://temp handles.
 355                         curl_setopt( $ch, CURLOPT_WRITEFUNCTION,
 356                                 function ( $ch, $data ) use ( &$req ) {
 357                                         return fwrite( $req['stream'], $data );
 358                                 }
 359                         );
 360                 } else {
 361                         curl_setopt( $ch, CURLOPT_WRITEFUNCTION,
 362                                 function ( $ch, $data ) use ( &$req ) {
 363                                         $req['response']['body'] .= $data;
 364                                         return strlen( $data );
 365                                 }
 366                         );
 367                 }
 368
 369                 return $ch;
 370         }
 371
 372         /**
 373          * @return resource
 374          */
 375         protected function getCurlMulti() {
 376                 if ( !$this->multiHandle ) {
 377                         $cmh = curl_multi_init();
 378                         if ( function_exists( 'curl_multi_setopt' ) ) { // PHP 5.5
 379                                 curl_multi_setopt( $cmh, CURLMOPT_PIPELINING, (int)$this->usePipelining );
 380                                 curl_multi_setopt( $cmh, CURLMOPT_MAXCONNECTS, (int)$this->maxConnsPerHost );
 381                         }
 382                         $this->multiHandle = $cmh;
 383                 }
 384                 return $this->multiHandle;
 385         }
 386
 387         function __destruct() {
 388                 if ( $this->multiHandle ) {
 389                         curl_multi_close( $this->multiHandle );
 390                 }
 391         }
 392 }