src/Network/HTTPRequest.php

   1 <?php
   2 /**
   3  * @copyright Copyright (C) 2020, Friendica
   4  *
   5  * @license GNU APGL version 3 or any later version
   6  *
   7  * This program is free software: you can redistribute it and/or modify
   8  * it under the terms of the GNU Affero General Public License as
   9  * published by the Free Software Foundation, either version 3 of the
  10  * License, or (at your option) any later version.
  11  *
  12  * This program is distributed in the hope that it will be useful,
  13  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  14  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  15  * GNU Affero General Public License for more details.
  16  *
  17  * You should have received a copy of the GNU Affero General Public License
  18  * along with this program.  If not, see <https://www.gnu.org/licenses/>.
  19  *
  20  */
  21
  22 namespace Friendica\Network;
  23
  24 use Friendica\App;
  25 use Friendica\Core\Config\IConfig;
  26 use Friendica\Core\Logger;
  27 use Friendica\Core\System;
  28 use Friendica\Util\Network;
  29 use Friendica\Util\Profiler;
  30 use Psr\Log\LoggerInterface;
  31
  32 /**
  33  * Performs HTTP requests to a given URL
  34  */
  35 class HTTPRequest
  36 {
  37         /** @var LoggerInterface */
  38         private $logger;
  39         /** @var Profiler */
  40         private $profiler;
  41         /** @var IConfig */
  42         private $config;
  43         /** @var string */
  44         private $userAgent;
  45
  46         public function __construct(LoggerInterface $logger, Profiler $profiler, IConfig $config, App $a)
  47         {
  48                 $this->logger    = $logger;
  49                 $this->profiler  = $profiler;
  50                 $this->config    = $config;
  51                 $this->userAgent = $a->getUserAgent();
  52         }
  53
  54         /**
  55          * fetches an URL.
  56          *
  57          * @param string $url        URL to fetch
  58          * @param bool   $binary     default false
  59          *                           TRUE if asked to return binary results (file download)
  60          * @param array  $opts       (optional parameters) assoziative array with:
  61          *                           'accept_content' => supply Accept: header with 'accept_content' as the value
  62          *                           'timeout' => int Timeout in seconds, default system config value or 60 seconds
  63          *                           'http_auth' => username:password
  64          *                           'novalidate' => do not validate SSL certs, default is to validate using our CA list
  65          *                           'nobody' => only return the header
  66          *                           'cookiejar' => path to cookie jar file
  67          *                           'header' => header array
  68          * @param int    $redirects  The recursion counter for internal use - default 0
  69          *
  70          * @return CurlResult
  71          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
  72          */
  73         public function curl(string $url, bool $binary = false, array $opts = [], int &$redirects = 0)
  74         {
  75                 $stamp1 = microtime(true);
  76
  77                 if (strlen($url) > 1000) {
  78                         $this->logger->debug('URL is longer than 1000 characters.', ['url' => $url, 'callstack' => System::callstack(20)]);
  79                         return CurlResult::createErrorCurl(substr($url, 0, 200));
  80                 }
  81
  82                 $parts2     = [];
  83                 $parts      = parse_url($url);
  84                 $path_parts = explode('/', $parts['path'] ?? '');
  85                 foreach ($path_parts as $part) {
  86                         if (strlen($part) <> mb_strlen($part)) {
  87                                 $parts2[] = rawurlencode($part);
  88                         } else {
  89                                 $parts2[] = $part;
  90                         }
  91                 }
  92                 $parts['path'] = implode('/', $parts2);
  93                 $url           = Network::unparseURL($parts);
  94
  95                 if (Network::isUrlBlocked($url)) {
  96                         $this->logger->info('Domain is blocked.', ['url' => $url]);
  97                         return CurlResult::createErrorCurl($url);
  98                 }
  99
 100                 $ch = @curl_init($url);
 101
 102                 if (($redirects > 8) || (!$ch)) {
 103                         return CurlResult::createErrorCurl($url);
 104                 }
 105
 106                 @curl_setopt($ch, CURLOPT_HEADER, true);
 107
 108                 if (!empty($opts['cookiejar'])) {
 109                         curl_setopt($ch, CURLOPT_COOKIEJAR, $opts["cookiejar"]);
 110                         curl_setopt($ch, CURLOPT_COOKIEFILE, $opts["cookiejar"]);
 111                 }
 112
 113                 // These settings aren't needed. We're following the location already.
 114                 //      @curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
 115                 //      @curl_setopt($ch, CURLOPT_MAXREDIRS, 5);
 116
 117                 if (!empty($opts['accept_content'])) {
 118                         curl_setopt(
 119                                 $ch,
 120                                 CURLOPT_HTTPHEADER,
 121                                 ['Accept: ' . $opts['accept_content']]
 122                         );
 123                 }
 124
 125                 if (!empty($opts['header'])) {
 126                         curl_setopt($ch, CURLOPT_HTTPHEADER, $opts['header']);
 127                 }
 128
 129                 @curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 130                 @curl_setopt($ch, CURLOPT_USERAGENT, $this->userAgent);
 131
 132                 $range = intval($this->config->get('system', 'curl_range_bytes', 0));
 133
 134                 if ($range > 0) {
 135                         @curl_setopt($ch, CURLOPT_RANGE, '0-' . $range);
 136                 }
 137
 138                 // Without this setting it seems as if some webservers send compressed content
 139                 // This seems to confuse curl so that it shows this uncompressed.
 140                 /// @todo  We could possibly set this value to "gzip" or something similar
 141                 curl_setopt($ch, CURLOPT_ENCODING, '');
 142
 143                 if (!empty($opts['headers'])) {
 144                         @curl_setopt($ch, CURLOPT_HTTPHEADER, $opts['headers']);
 145                 }
 146
 147                 if (!empty($opts['nobody'])) {
 148                         @curl_setopt($ch, CURLOPT_NOBODY, $opts['nobody']);
 149                 }
 150
 151                 if (!empty($opts['timeout'])) {
 152                         @curl_setopt($ch, CURLOPT_TIMEOUT, $opts['timeout']);
 153                 } else {
 154                         $curl_time = $this->config->get('system', 'curl_timeout', 60);
 155                         @curl_setopt($ch, CURLOPT_TIMEOUT, intval($curl_time));
 156                 }
 157
 158                 // by default we will allow self-signed certs
 159                 // but you can override this
 160
 161                 $check_cert = $this->config->get('system', 'verifyssl');
 162                 @curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
 163
 164                 if ($check_cert) {
 165                         @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2);
 166                 }
 167
 168                 $proxy = $this->config->get('system', 'proxy');
 169
 170                 if (!empty($proxy)) {
 171                         @curl_setopt($ch, CURLOPT_HTTPPROXYTUNNEL, 1);
 172                         @curl_setopt($ch, CURLOPT_PROXY, $proxy);
 173                         $proxyuser = $this->config->get('system', 'proxyuser');
 174
 175                         if (!empty($proxyuser)) {
 176                                 @curl_setopt($ch, CURLOPT_PROXYUSERPWD, $proxyuser);
 177                         }
 178                 }
 179
 180                 if ($this->config->get('system', 'ipv4_resolve', false)) {
 181                         curl_setopt($ch, CURLOPT_IPRESOLVE, CURL_IPRESOLVE_V4);
 182                 }
 183
 184                 if ($binary) {
 185                         @curl_setopt($ch, CURLOPT_BINARYTRANSFER, 1);
 186                 }
 187
 188                 // don't let curl abort the entire application
 189                 // if it throws any errors.
 190
 191                 $s         = @curl_exec($ch);
 192                 $curl_info = @curl_getinfo($ch);
 193
 194                 // Special treatment for HTTP Code 416
 195                 // See https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/416
 196                 if (($curl_info['http_code'] == 416) && ($range > 0)) {
 197                         @curl_setopt($ch, CURLOPT_RANGE, '');
 198                         $s         = @curl_exec($ch);
 199                         $curl_info = @curl_getinfo($ch);
 200                 }
 201
 202                 $curlResponse = new CurlResult($url, $s, $curl_info, curl_errno($ch), curl_error($ch));
 203
 204                 if ($curlResponse->isRedirectUrl()) {
 205                         $redirects++;
 206                         $this->logger->notice('Curl redirect.', ['url' => $url, 'to' => $curlResponse->getRedirectUrl()]);
 207                         @curl_close($ch);
 208                         return self::curl($curlResponse->getRedirectUrl(), $binary, $opts, $redirects);
 209                 }
 210
 211                 @curl_close($ch);
 212
 213                 $this->profiler->saveTimestamp($stamp1, 'network', System::callstack());
 214
 215                 return $curlResponse;
 216         }
 217
 218         /**
 219          * Send POST request to $url
 220          *
 221          * @param string $url       URL to post
 222          * @param mixed  $params    array of POST variables
 223          * @param array  $headers   HTTP headers
 224          * @param int    $redirects Recursion counter for internal use - default = 0
 225          * @param int    $timeout   The timeout in seconds, default system config value or 60 seconds
 226          *
 227          * @return CurlResult The content
 228          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 229          */
 230         public function post(string $url, $params, array $headers = [], int $timeout = 0, int &$redirects = 0)
 231         {
 232                 $stamp1 = microtime(true);
 233
 234                 if (Network::isUrlBlocked($url)) {
 235                         $this->logger->info('Domain is blocked.'. ['url' => $url]);
 236                         return CurlResult::createErrorCurl($url);
 237                 }
 238
 239                 $ch = curl_init($url);
 240
 241                 if (($redirects > 8) || (!$ch)) {
 242                         return CurlResult::createErrorCurl($url);
 243                 }
 244
 245                 $this->logger->debug('Post_url: start.', ['url' => $url]);
 246
 247                 curl_setopt($ch, CURLOPT_HEADER, true);
 248                 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 249                 curl_setopt($ch, CURLOPT_POST, 1);
 250                 curl_setopt($ch, CURLOPT_POSTFIELDS, $params);
 251                 curl_setopt($ch, CURLOPT_USERAGENT, $a->getUserAgent());
 252
 253                 if ($this->config->get('system', 'ipv4_resolve', false)) {
 254                         curl_setopt($ch, CURLOPT_IPRESOLVE, CURL_IPRESOLVE_V4);
 255                 }
 256
 257                 if (intval($timeout)) {
 258                         curl_setopt($ch, CURLOPT_TIMEOUT, $timeout);
 259                 } else {
 260                         $curl_time = $this->config->get('system', 'curl_timeout', 60);
 261                         curl_setopt($ch, CURLOPT_TIMEOUT, intval($curl_time));
 262                 }
 263
 264                 if (!empty($headers)) {
 265                         curl_setopt($ch, CURLOPT_HTTPHEADER, $headers);
 266                 }
 267
 268                 $check_cert = $this->config->get('system', 'verifyssl');
 269                 curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
 270
 271                 if ($check_cert) {
 272                         @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2);
 273                 }
 274
 275                 $proxy = $this->config->get('system', 'proxy');
 276
 277                 if (!empty($proxy)) {
 278                         curl_setopt($ch, CURLOPT_HTTPPROXYTUNNEL, 1);
 279                         curl_setopt($ch, CURLOPT_PROXY, $proxy);
 280                         $proxyuser = $this->config->get('system', 'proxyuser');
 281                         if (!empty($proxyuser)) {
 282                                 curl_setopt($ch, CURLOPT_PROXYUSERPWD, $proxyuser);
 283                         }
 284                 }
 285
 286                 // don't let curl abort the entire application
 287                 // if it throws any errors.
 288
 289                 $s = @curl_exec($ch);
 290
 291                 $curl_info = curl_getinfo($ch);
 292
 293                 $curlResponse = new CurlResult($url, $s, $curl_info, curl_errno($ch), curl_error($ch));
 294
 295                 if ($curlResponse->isRedirectUrl()) {
 296                         $redirects++;
 297                         $this->logger->info('Post redirect.', ['url' => $url, 'to' => $curlResponse->getRedirectUrl()]);
 298                         curl_close($ch);
 299                         return self::post($curlResponse->getRedirectUrl(), $params, $headers, $redirects, $timeout);
 300                 }
 301
 302                 curl_close($ch);
 303
 304                 $this->profiler->saveTimestamp($stamp1, 'network', System::callstack());
 305
 306                 // Very old versions of Lighttpd don't like the "Expect" header, so we remove it when needed
 307                 if ($curlResponse->getReturnCode() == 417) {
 308                         $redirects++;
 309
 310                         if (empty($headers)) {
 311                                 $headers = ['Expect:'];
 312                         } else {
 313                                 if (!in_array('Expect:', $headers)) {
 314                                         array_push($headers, 'Expect:');
 315                                 }
 316                         }
 317                         Logger::info('Server responds with 417, applying workaround', ['url' => $url]);
 318                         return self::post($url, $params, $headers, $redirects, $timeout);
 319                 }
 320
 321                 Logger::log('post_url: end ' . $url, Logger::DATA);
 322
 323                 return $curlResponse;
 324         }
 325
 326         /**
 327          * Curl wrapper
 328          *
 329          * If binary flag is true, return binary results.
 330          * Set the cookiejar argument to a string (e.g. "/tmp/friendica-cookies.txt")
 331          * to preserve cookies from one request to the next.
 332          *
 333          * @param string $url             URL to fetch
 334          * @param bool   $binary          default false
 335          *                                TRUE if asked to return binary results (file download)
 336          * @param int    $timeout         Timeout in seconds, default system config value or 60 seconds
 337          * @param string $accept_content  supply Accept: header with 'accept_content' as the value
 338          * @param string $cookiejar       Path to cookie jar file
 339          * @param int    $redirects       The recursion counter for internal use - default 0
 340          *
 341          * @return string The fetched content
 342          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 343          */
 344         public static function fetchUrl(string $url, bool $binary = false, int $timeout = 0, string $accept_content = '', string $cookiejar = '', int &$redirects = 0)
 345         {
 346                 $ret = self::fetchUrlFull($url, $binary, $timeout, $accept_content, $cookiejar, $redirects);
 347
 348                 return $ret->getBody();
 349         }
 350
 351         /**
 352          * Curl wrapper with array of return values.
 353          *
 354          * Inner workings and parameters are the same as @ref fetchUrl but returns an array with
 355          * all the information collected during the fetch.
 356          *
 357          * @param string $url             URL to fetch
 358          * @param bool   $binary          default false
 359          *                                TRUE if asked to return binary results (file download)
 360          * @param int    $timeout         Timeout in seconds, default system config value or 60 seconds
 361          * @param string $accept_content  supply Accept: header with 'accept_content' as the value
 362          * @param string $cookiejar       Path to cookie jar file
 363          * @param int    $redirects       The recursion counter for internal use - default 0
 364          *
 365          * @return CurlResult With all relevant information, 'body' contains the actual fetched content.
 366          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 367          */
 368         public static function fetchUrlFull(string $url, bool $binary = false, int $timeout = 0, string $accept_content = '', string $cookiejar = '', int &$redirects = 0)
 369         {
 370                 return self::curl(
 371                         $url,
 372                         $binary,
 373                         [
 374                                 'timeout'        => $timeout,
 375                                 'accept_content' => $accept_content,
 376                                 'cookiejar'      => $cookiejar
 377                         ],
 378                         $redirects
 379                 );
 380         }
 381 }