src/Network/HTTPRequest.php

   1 <?php
   2 /**
   3  * @copyright Copyright (C) 2010-2021, the Friendica project
   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 DOMDocument;
  25 use DomXPath;
  26 use Friendica\App;
  27 use Friendica\Core\Config\IConfig;
  28 use Friendica\Core\System;
  29 use Friendica\Util\Network;
  30 use Friendica\Util\Profiler;
  31 use Psr\Log\LoggerInterface;
  32
  33 /**
  34  * Performs HTTP requests to a given URL
  35  */
  36 class HTTPRequest implements IHTTPRequest
  37 {
  38         /** @var LoggerInterface */
  39         private $logger;
  40         /** @var Profiler */
  41         private $profiler;
  42         /** @var IConfig */
  43         private $config;
  44         /** @var string */
  45         private $baseUrl;
  46
  47         public function __construct(LoggerInterface $logger, Profiler $profiler, IConfig $config, App\BaseURL $baseUrl)
  48         {
  49                 $this->logger   = $logger;
  50                 $this->profiler = $profiler;
  51                 $this->config   = $config;
  52                 $this->baseUrl  = $baseUrl->get();
  53         }
  54
  55         /** {@inheritDoc}
  56          *
  57          * @throws HTTPException\InternalServerErrorException
  58          */
  59         public function head(string $url, array $opts = [])
  60         {
  61                 $opts['nobody'] = true;
  62
  63                 return $this->get($url, $opts);
  64         }
  65
  66         /**
  67          * {@inheritDoc}
  68          *
  69          * @param int $redirects The recursion counter for internal use - default 0
  70          *
  71          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
  72          */
  73         public function get(string $url, 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                         $this->profiler->saveTimestamp($stamp1, 'network');
  80                         return CurlResult::createErrorCurl(substr($url, 0, 200));
  81                 }
  82
  83                 $parts2     = [];
  84                 $parts      = parse_url($url);
  85                 $path_parts = explode('/', $parts['path'] ?? '');
  86                 foreach ($path_parts as $part) {
  87                         if (strlen($part) <> mb_strlen($part)) {
  88                                 $parts2[] = rawurlencode($part);
  89                         } else {
  90                                 $parts2[] = $part;
  91                         }
  92                 }
  93                 $parts['path'] = implode('/', $parts2);
  94                 $url           = Network::unparseURL($parts);
  95
  96                 if (Network::isUrlBlocked($url)) {
  97                         $this->logger->info('Domain is blocked.', ['url' => $url]);
  98                         $this->profiler->saveTimestamp($stamp1, 'network');
  99                         return CurlResult::createErrorCurl($url);
 100                 }
 101
 102                 $ch = @curl_init($url);
 103
 104                 if (($redirects > 8) || (!$ch)) {
 105                         $this->profiler->saveTimestamp($stamp1, 'network');
 106                         return CurlResult::createErrorCurl($url);
 107                 }
 108
 109                 @curl_setopt($ch, CURLOPT_HEADER, true);
 110
 111                 if (!empty($opts['cookiejar'])) {
 112                         curl_setopt($ch, CURLOPT_COOKIEJAR, $opts["cookiejar"]);
 113                         curl_setopt($ch, CURLOPT_COOKIEFILE, $opts["cookiejar"]);
 114                 }
 115
 116                 // These settings aren't needed. We're following the location already.
 117                 //      @curl_setopt($ch, CURLOPT_FOLLOWLOCATION, true);
 118                 //      @curl_setopt($ch, CURLOPT_MAXREDIRS, 5);
 119
 120                 if (!empty($opts['accept_content'])) {
 121                         curl_setopt(
 122                                 $ch,
 123                                 CURLOPT_HTTPHEADER,
 124                                 ['Accept: ' . $opts['accept_content']]
 125                         );
 126                 }
 127
 128                 if (!empty($opts['header'])) {
 129                         curl_setopt($ch, CURLOPT_HTTPHEADER, $opts['header']);
 130                 }
 131
 132                 @curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 133                 @curl_setopt($ch, CURLOPT_USERAGENT, $this->getUserAgent());
 134
 135                 $range = intval($this->config->get('system', 'curl_range_bytes', 0));
 136
 137                 if ($range > 0) {
 138                         @curl_setopt($ch, CURLOPT_RANGE, '0-' . $range);
 139                 }
 140
 141                 // Without this setting it seems as if some webservers send compressed content
 142                 // This seems to confuse curl so that it shows this uncompressed.
 143                 /// @todo  We could possibly set this value to "gzip" or something similar
 144                 curl_setopt($ch, CURLOPT_ENCODING, '');
 145
 146                 if (!empty($opts['headers'])) {
 147                         $this->logger->notice('Wrong option \'headers\' used.');
 148                         @curl_setopt($ch, CURLOPT_HTTPHEADER, $opts['headers']);
 149                 }
 150
 151                 if (!empty($opts['nobody'])) {
 152                         @curl_setopt($ch, CURLOPT_NOBODY, $opts['nobody']);
 153                 }
 154
 155                 @curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 10);
 156
 157                 if (!empty($opts['timeout'])) {
 158                         @curl_setopt($ch, CURLOPT_TIMEOUT, $opts['timeout']);
 159                 } else {
 160                         $curl_time = $this->config->get('system', 'curl_timeout', 60);
 161                         @curl_setopt($ch, CURLOPT_TIMEOUT, intval($curl_time));
 162                 }
 163
 164                 // by default we will allow self-signed certs
 165                 // but you can override this
 166
 167                 $check_cert = $this->config->get('system', 'verifyssl');
 168                 @curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
 169
 170                 if ($check_cert) {
 171                         @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2);
 172                 }
 173
 174                 $proxy = $this->config->get('system', 'proxy');
 175
 176                 if (!empty($proxy)) {
 177                         @curl_setopt($ch, CURLOPT_HTTPPROXYTUNNEL, 1);
 178                         @curl_setopt($ch, CURLOPT_PROXY, $proxy);
 179                         $proxyuser = $this->config->get('system', 'proxyuser');
 180
 181                         if (!empty($proxyuser)) {
 182                                 @curl_setopt($ch, CURLOPT_PROXYUSERPWD, $proxyuser);
 183                         }
 184                 }
 185
 186                 if ($this->config->get('system', 'ipv4_resolve', false)) {
 187                         curl_setopt($ch, CURLOPT_IPRESOLVE, CURL_IPRESOLVE_V4);
 188                 }
 189
 190                 $s         = @curl_exec($ch);
 191                 $curl_info = @curl_getinfo($ch);
 192
 193                 // Special treatment for HTTP Code 416
 194                 // See https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/416
 195                 if (($curl_info['http_code'] == 416) && ($range > 0)) {
 196                         @curl_setopt($ch, CURLOPT_RANGE, '');
 197                         $s         = @curl_exec($ch);
 198                         $curl_info = @curl_getinfo($ch);
 199                 }
 200
 201                 $curlResponse = new CurlResult($url, $s, $curl_info, curl_errno($ch), curl_error($ch));
 202
 203                 if (!Network::isRedirectBlocked($url) && $curlResponse->isRedirectUrl()) {
 204                         $redirects++;
 205                         $this->logger->notice('Curl redirect.', ['url' => $url, 'to' => $curlResponse->getRedirectUrl()]);
 206                         @curl_close($ch);
 207                         $this->profiler->saveTimestamp($stamp1, 'network');
 208                         return $this->get($curlResponse->getRedirectUrl(), $opts, $redirects);
 209                 }
 210
 211                 @curl_close($ch);
 212
 213                 $this->profiler->saveTimestamp($stamp1, 'network');
 214
 215                 return $curlResponse;
 216         }
 217
 218         /**
 219          * {@inheritDoc}
 220          *
 221          * @param int $redirects The recursion counter for internal use - default 0
 222          *
 223          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 224          */
 225         public function post(string $url, $params, array $headers = [], int $timeout = 0, int &$redirects = 0)
 226         {
 227                 $stamp1 = microtime(true);
 228
 229                 if (Network::isUrlBlocked($url)) {
 230                         $this->logger->info('Domain is blocked.' . ['url' => $url]);
 231                         $this->profiler->saveTimestamp($stamp1, 'network');
 232                         return CurlResult::createErrorCurl($url);
 233                 }
 234
 235                 $ch = curl_init($url);
 236
 237                 if (($redirects > 8) || (!$ch)) {
 238                         $this->profiler->saveTimestamp($stamp1, 'network');
 239                         return CurlResult::createErrorCurl($url);
 240                 }
 241
 242                 $this->logger->debug('Post_url: start.', ['url' => $url]);
 243
 244                 curl_setopt($ch, CURLOPT_HEADER, true);
 245                 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 246                 curl_setopt($ch, CURLOPT_POST, 1);
 247                 curl_setopt($ch, CURLOPT_POSTFIELDS, $params);
 248                 curl_setopt($ch, CURLOPT_USERAGENT, $this->getUserAgent());
 249
 250                 if ($this->config->get('system', 'ipv4_resolve', false)) {
 251                         curl_setopt($ch, CURLOPT_IPRESOLVE, CURL_IPRESOLVE_V4);
 252                 }
 253
 254                 @curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 10);
 255
 256                 if (intval($timeout)) {
 257                         curl_setopt($ch, CURLOPT_TIMEOUT, $timeout);
 258                 } else {
 259                         $curl_time = $this->config->get('system', 'curl_timeout', 60);
 260                         curl_setopt($ch, CURLOPT_TIMEOUT, intval($curl_time));
 261                 }
 262
 263                 if (!empty($headers)) {
 264                         curl_setopt($ch, CURLOPT_HTTPHEADER, $headers);
 265                 }
 266
 267                 $check_cert = $this->config->get('system', 'verifyssl');
 268                 curl_setopt($ch, CURLOPT_SSL_VERIFYPEER, (($check_cert) ? true : false));
 269
 270                 if ($check_cert) {
 271                         @curl_setopt($ch, CURLOPT_SSL_VERIFYHOST, 2);
 272                 }
 273
 274                 $proxy = $this->config->get('system', 'proxy');
 275
 276                 if (!empty($proxy)) {
 277                         curl_setopt($ch, CURLOPT_HTTPPROXYTUNNEL, 1);
 278                         curl_setopt($ch, CURLOPT_PROXY, $proxy);
 279                         $proxyuser = $this->config->get('system', 'proxyuser');
 280                         if (!empty($proxyuser)) {
 281                                 curl_setopt($ch, CURLOPT_PROXYUSERPWD, $proxyuser);
 282                         }
 283                 }
 284
 285                 // don't let curl abort the entire application
 286                 // if it throws any errors.
 287
 288                 $s = @curl_exec($ch);
 289
 290                 $curl_info = curl_getinfo($ch);
 291
 292                 $curlResponse = new CurlResult($url, $s, $curl_info, curl_errno($ch), curl_error($ch));
 293
 294                 if (!Network::isRedirectBlocked($url) && $curlResponse->isRedirectUrl()) {
 295                         $redirects++;
 296                         $this->logger->info('Post redirect.', ['url' => $url, 'to' => $curlResponse->getRedirectUrl()]);
 297                         curl_close($ch);
 298                         $this->profiler->saveTimestamp($stamp1, 'network');
 299                         return $this->post($curlResponse->getRedirectUrl(), $params, $headers, $redirects, $timeout);
 300                 }
 301
 302                 curl_close($ch);
 303
 304                 $this->profiler->saveTimestamp($stamp1, 'network');
 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                         $this->logger->info('Server responds with 417, applying workaround', ['url' => $url]);
 318                         return $this->post($url, $params, $headers, $redirects, $timeout);
 319                 }
 320
 321                 $this->logger->debug('Post_url: End.', ['url' => $url]);
 322
 323                 return $curlResponse;
 324         }
 325
 326         /**
 327          * {@inheritDoc}
 328          */
 329         public function finalUrl(string $url, int $depth = 1, bool $fetchbody = false)
 330         {
 331                 if (Network::isUrlBlocked($url)) {
 332                         $this->logger->info('Domain is blocked.', ['url' => $url]);
 333                         return $url;
 334                 }
 335
 336                 if (Network::isRedirectBlocked($url)) {
 337                         $this->logger->info('Domain should not be redirected.', ['url' => $url]);
 338                         return $url;
 339                 }
 340
 341                 $url = Network::stripTrackingQueryParams($url);
 342
 343                 if ($depth > 10) {
 344                         return $url;
 345                 }
 346
 347                 $url = trim($url, "'");
 348
 349                 $stamp1 = microtime(true);
 350
 351                 $ch = curl_init();
 352                 curl_setopt($ch, CURLOPT_URL, $url);
 353                 curl_setopt($ch, CURLOPT_HEADER, 1);
 354                 curl_setopt($ch, CURLOPT_NOBODY, 1);
 355                 curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 10);
 356                 curl_setopt($ch, CURLOPT_TIMEOUT, 10);
 357                 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 358                 curl_setopt($ch, CURLOPT_USERAGENT, $this->getUserAgent());
 359
 360                 curl_exec($ch);
 361                 $curl_info = @curl_getinfo($ch);
 362                 $http_code = $curl_info['http_code'];
 363                 curl_close($ch);
 364
 365                 $this->profiler->saveTimestamp($stamp1, "network");
 366
 367                 if ($http_code == 0) {
 368                         return $url;
 369                 }
 370
 371                 if (in_array($http_code, ['301', '302'])) {
 372                         if (!empty($curl_info['redirect_url'])) {
 373                                 return $this->finalUrl($curl_info['redirect_url'], ++$depth, $fetchbody);
 374                         } elseif (!empty($curl_info['location'])) {
 375                                 return $this->finalUrl($curl_info['location'], ++$depth, $fetchbody);
 376                         }
 377                 }
 378
 379                 // Check for redirects in the meta elements of the body if there are no redirects in the header.
 380                 if (!$fetchbody) {
 381                         return $this->finalUrl($url, ++$depth, true);
 382                 }
 383
 384                 // if the file is too large then exit
 385                 if ($curl_info["download_content_length"] > 1000000) {
 386                         return $url;
 387                 }
 388
 389                 // if it isn't a HTML file then exit
 390                 if (!empty($curl_info["content_type"]) && !strstr(strtolower($curl_info["content_type"]), "html")) {
 391                         return $url;
 392                 }
 393
 394                 $stamp1 = microtime(true);
 395
 396                 $ch = curl_init();
 397                 curl_setopt($ch, CURLOPT_URL, $url);
 398                 curl_setopt($ch, CURLOPT_HEADER, 0);
 399                 curl_setopt($ch, CURLOPT_NOBODY, 0);
 400                 curl_setopt($ch, CURLOPT_CONNECTTIMEOUT, 10);
 401                 curl_setopt($ch, CURLOPT_TIMEOUT, 10);
 402                 curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
 403                 curl_setopt($ch, CURLOPT_USERAGENT, $this->getUserAgent());
 404
 405                 $body = curl_exec($ch);
 406                 curl_close($ch);
 407
 408                 $this->profiler->saveTimestamp($stamp1, "network");
 409
 410                 if (trim($body) == "") {
 411                         return $url;
 412                 }
 413
 414                 // Check for redirect in meta elements
 415                 $doc = new DOMDocument();
 416                 @$doc->loadHTML($body);
 417
 418                 $xpath = new DomXPath($doc);
 419
 420                 $list = $xpath->query("//meta[@content]");
 421                 foreach ($list as $node) {
 422                         $attr = [];
 423                         if ($node->attributes->length) {
 424                                 foreach ($node->attributes as $attribute) {
 425                                         $attr[$attribute->name] = $attribute->value;
 426                                 }
 427                         }
 428
 429                         if (@$attr["http-equiv"] == 'refresh') {
 430                                 $path = $attr["content"];
 431                                 $pathinfo = explode(";", $path);
 432                                 foreach ($pathinfo as $value) {
 433                                         if (substr(strtolower($value), 0, 4) == "url=") {
 434                                                 return $this->finalUrl(substr($value, 4), ++$depth);
 435                                         }
 436                                 }
 437                         }
 438                 }
 439
 440                 return $url;
 441         }
 442
 443         /**
 444          * {@inheritDoc}
 445          *
 446          * @param int $redirects The recursion counter for internal use - default 0
 447          *
 448          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 449          */
 450         public function fetch(string $url, int $timeout = 0, string $accept_content = '', string $cookiejar = '', int &$redirects = 0)
 451         {
 452                 $ret = $this->fetchFull($url, $timeout, $accept_content, $cookiejar, $redirects);
 453
 454                 return $ret->getBody();
 455         }
 456
 457         /**
 458          * {@inheritDoc}
 459          *
 460          * @param int $redirects The recursion counter for internal use - default 0
 461          *
 462          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 463          */
 464         public function fetchFull(string $url, int $timeout = 0, string $accept_content = '', string $cookiejar = '', int &$redirects = 0)
 465         {
 466                 return $this->get(
 467                         $url,
 468                         [
 469                                 'timeout'        => $timeout,
 470                                 'accept_content' => $accept_content,
 471                                 'cookiejar'      => $cookiejar
 472                         ],
 473                         $redirects
 474                 );
 475         }
 476
 477         /**
 478          * {@inheritDoc}
 479          */
 480         public function getUserAgent()
 481         {
 482                 return
 483                         FRIENDICA_PLATFORM . " '" .
 484                         FRIENDICA_CODENAME . "' " .
 485                         FRIENDICA_VERSION . '-' .
 486                         DB_UPDATE_VERSION . '; ' .
 487                         $this->baseUrl;
 488         }
 489 }