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