src/Content/Text/BBCode.php

   1 <?php
   2 /**
   3  * @file src/Content/Text/BBCode.php
   4  */
   5
   6 namespace Friendica\Content\Text;
   7
   8 use DOMDocument;
   9 use DOMXPath;
  10 use Exception;
  11 use Friendica\BaseObject;
  12 use Friendica\Content\OEmbed;
  13 use Friendica\Content\Smilies;
  14 use Friendica\Core\Cache;
  15 use Friendica\Core\Config;
  16 use Friendica\Core\Hook;
  17 use Friendica\Core\L10n;
  18 use Friendica\Core\Logger;
  19 use Friendica\Core\Protocol;
  20 use Friendica\Core\Renderer;
  21 use Friendica\Core\System;
  22 use Friendica\Model\Contact;
  23 use Friendica\Model\Event;
  24 use Friendica\Network\Probe;
  25 use Friendica\Object\Image;
  26 use Friendica\Util\Map;
  27 use Friendica\Util\Network;
  28 use Friendica\Util\ParseUrl;
  29 use Friendica\Util\Proxy as ProxyUtils;
  30 use Friendica\Util\Strings;
  31 use Friendica\Util\XML;
  32
  33 class BBCode extends BaseObject
  34 {
  35         /**
  36          * @brief Fetches attachment data that were generated the old way
  37          *
  38          * @param string $body Message body
  39          * @return array
  40          *                     'type' -> Message type ("link", "video", "photo")
  41          *                     'text' -> Text before the shared message
  42          *                     'after' -> Text after the shared message
  43          *                     'image' -> Preview image of the message
  44          *                     'url' -> Url to the attached message
  45          *                     'title' -> Title of the attachment
  46          *                     'description' -> Description of the attachment
  47          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
  48          */
  49         private static function getOldAttachmentData($body)
  50         {
  51                 $post = [];
  52
  53                 // Simplify image codes
  54                 $body = preg_replace("/\[img\=([0-9]*)x([0-9]*)\](.*?)\[\/img\]/ism", '[img]$3[/img]', $body);
  55
  56                 if (preg_match_all("(\[class=(.*?)\](.*?)\[\/class\])ism", $body, $attached, PREG_SET_ORDER)) {
  57                         foreach ($attached as $data) {
  58                                 if (!in_array($data[1], ["type-link", "type-video", "type-photo"])) {
  59                                         continue;
  60                                 }
  61
  62                                 $post["type"] = substr($data[1], 5);
  63
  64                                 $pos = strpos($body, $data[0]);
  65                                 if ($pos > 0) {
  66                                         $post["text"] = trim(substr($body, 0, $pos));
  67                                         $post["after"] = trim(substr($body, $pos + strlen($data[0])));
  68                                 } else {
  69                                         $post["text"] = trim(str_replace($data[0], "", $body));
  70                                         $post["after"] = '';
  71                                 }
  72
  73                                 $attacheddata = $data[2];
  74
  75                                 if (preg_match("/\[img\](.*?)\[\/img\]/ism", $attacheddata, $matches)) {
  76
  77                                         $picturedata = Image::getInfoFromURL($matches[1]);
  78
  79                                         if ($picturedata) {
  80                                                 if (($picturedata[0] >= 500) && ($picturedata[0] >= $picturedata[1])) {
  81                                                         $post["image"] = $matches[1];
  82                                                 } else {
  83                                                         $post["preview"] = $matches[1];
  84                                                 }
  85                                         }
  86                                 }
  87
  88                                 if (preg_match("/\[bookmark\=(.*?)\](.*?)\[\/bookmark\]/ism", $attacheddata, $matches)) {
  89                                         $post["url"] = $matches[1];
  90                                         $post["title"] = $matches[2];
  91                                 }
  92                                 if (!empty($post["url"]) && (in_array($post["type"], ["link", "video"]))
  93                                         && preg_match("/\[url\=(.*?)\](.*?)\[\/url\]/ism", $attacheddata, $matches)) {
  94                                         $post["url"] = $matches[1];
  95                                 }
  96
  97                                 // Search for description
  98                                 if (preg_match("/\[quote\](.*?)\[\/quote\]/ism", $attacheddata, $matches)) {
  99                                         $post["description"] = $matches[1];
 100                                 }
 101                         }
 102                 }
 103                 return $post;
 104         }
 105
 106         /**
 107          * @brief Fetches attachment data that were generated with the "attachment" element
 108          *
 109          * @param string $body Message body
 110          * @return array
 111          *                     'type' -> Message type ("link", "video", "photo")
 112          *                     'text' -> Text before the shared message
 113          *                     'after' -> Text after the shared message
 114          *                     'image' -> Preview image of the message
 115          *                     'url' -> Url to the attached message
 116          *                     'title' -> Title of the attachment
 117          *                     'description' -> Description of the attachment
 118          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 119          */
 120         public static function getAttachmentData($body)
 121         {
 122                 $data = [];
 123
 124                 if (!preg_match("/(.*)\[attachment(.*?)\](.*?)\[\/attachment\](.*)/ism", $body, $match)) {
 125                         return self::getOldAttachmentData($body);
 126                 }
 127
 128                 $attributes = $match[2];
 129
 130                 $data["text"] = trim($match[1]);
 131
 132                 $type = "";
 133                 preg_match("/type='(.*?)'/ism", $attributes, $matches);
 134                 if (!empty($matches[1])) {
 135                         $type = strtolower($matches[1]);
 136                 }
 137
 138                 preg_match('/type="(.*?)"/ism', $attributes, $matches);
 139                 if (!empty($matches[1])) {
 140                         $type = strtolower($matches[1]);
 141                 }
 142
 143                 if ($type == "") {
 144                         return [];
 145                 }
 146
 147                 if (!in_array($type, ["link", "audio", "photo", "video"])) {
 148                         return [];
 149                 }
 150
 151                 if ($type != "") {
 152                         $data["type"] = $type;
 153                 }
 154
 155                 $url = "";
 156                 preg_match("/url='(.*?)'/ism", $attributes, $matches);
 157                 if (!empty($matches[1])) {
 158                         $url = $matches[1];
 159                 }
 160
 161                 preg_match('/url="(.*?)"/ism', $attributes, $matches);
 162                 if (!empty($matches[1])) {
 163                         $url = $matches[1];
 164                 }
 165
 166                 if ($url != "") {
 167                         $data["url"] = html_entity_decode($url, ENT_QUOTES, 'UTF-8');
 168                 }
 169
 170                 $title = "";
 171                 preg_match("/title='(.*?)'/ism", $attributes, $matches);
 172                 if (!empty($matches[1])) {
 173                         $title = $matches[1];
 174                 }
 175
 176                 preg_match('/title="(.*?)"/ism', $attributes, $matches);
 177                 if (!empty($matches[1])) {
 178                         $title = $matches[1];
 179                 }
 180
 181                 if ($title != "") {
 182                         $title = self::convert(html_entity_decode($title, ENT_QUOTES, 'UTF-8'), false, true);
 183                         $title = html_entity_decode($title, ENT_QUOTES, 'UTF-8');
 184                         $title = str_replace(["[", "]"], ["&#91;", "&#93;"], $title);
 185                         $data["title"] = $title;
 186                 }
 187
 188                 $image = "";
 189                 preg_match("/image='(.*?)'/ism", $attributes, $matches);
 190                 if (!empty($matches[1])) {
 191                         $image = $matches[1];
 192                 }
 193
 194                 preg_match('/image="(.*?)"/ism', $attributes, $matches);
 195                 if (!empty($matches[1])) {
 196                         $image = $matches[1];
 197                 }
 198
 199                 if ($image != "") {
 200                         $data["image"] = html_entity_decode($image, ENT_QUOTES, 'UTF-8');
 201                 }
 202
 203                 $preview = "";
 204                 preg_match("/preview='(.*?)'/ism", $attributes, $matches);
 205                 if (!empty($matches[1])) {
 206                         $preview = $matches[1];
 207                 }
 208
 209                 preg_match('/preview="(.*?)"/ism', $attributes, $matches);
 210                 if (!empty($matches[1])) {
 211                         $preview = $matches[1];
 212                 }
 213
 214                 if ($preview != "") {
 215                         $data["preview"] = html_entity_decode($preview, ENT_QUOTES, 'UTF-8');
 216                 }
 217
 218                 $data["description"] = trim($match[3]);
 219
 220                 $data["after"] = trim($match[4]);
 221
 222                 return $data;
 223         }
 224
 225         public static function getAttachedData($body, $item = [])
 226         {
 227                 /*
 228                 - text:
 229                 - type: link, video, photo
 230                 - title:
 231                 - url:
 232                 - image:
 233                 - description:
 234                 - (thumbnail)
 235                 */
 236
 237                 $has_title = !empty($item['title']);
 238                 $plink = defaults($item, 'plink', '');
 239                 $post = self::getAttachmentData($body);
 240
 241                 // if nothing is found, it maybe having an image.
 242                 if (!isset($post["type"])) {
 243                         // Simplify image codes
 244                         $body = preg_replace("/\[img\=([0-9]*)x([0-9]*)\](.*?)\[\/img\]/ism", '[img]$3[/img]', $body);
 245
 246                         $body = preg_replace("/\[img\=(.*?)\](.*?)\[\/img\]/ism", '[img]$1[/img]', $body);
 247
 248                         if (preg_match_all("(\[url=(.*?)\]\s*\[img\](.*?)\[\/img\]\s*\[\/url\])ism", $body, $pictures, PREG_SET_ORDER)) {
 249                                 if ((count($pictures) == 1) && !$has_title) {
 250                                         if (!empty($item['object-type']) && ($item['object-type'] == ACTIVITY_OBJ_IMAGE)) {
 251                                                 // Replace the preview picture with the real picture
 252                                                 $url = str_replace('-1.', '-0.', $pictures[0][2]);
 253                                                 $data = ['url' => $url, 'type' => 'photo'];
 254                                         } else {
 255                                                 // Checking, if the link goes to a picture
 256                                                 $data = ParseUrl::getSiteinfoCached($pictures[0][1], true);
 257                                         }
 258
 259                                         // Workaround:
 260                                         // Sometimes photo posts to the own album are not detected at the start.
 261                                         // So we seem to cannot use the cache for these cases. That's strange.
 262                                         if (($data["type"] != "photo") && strstr($pictures[0][1], "/photos/")) {
 263                                                 $data = ParseUrl::getSiteinfo($pictures[0][1], true);
 264                                         }
 265
 266                                         if ($data["type"] == "photo") {
 267                                                 $post["type"] = "photo";
 268                                                 if (isset($data["images"][0])) {
 269                                                         $post["image"] = $data["images"][0]["src"];
 270                                                         $post["url"] = $data["url"];
 271                                                 } else {
 272                                                         $post["image"] = $data["url"];
 273                                                 }
 274
 275                                                 $post["preview"] = $pictures[0][2];
 276                                                 $post["text"] = trim(str_replace($pictures[0][0], "", $body));
 277                                         } else {
 278                                                 $imgdata = Image::getInfoFromURL($pictures[0][1]);
 279                                                 if ($imgdata && substr($imgdata["mime"], 0, 6) == "image/") {
 280                                                         $post["type"] = "photo";
 281                                                         $post["image"] = $pictures[0][1];
 282                                                         $post["preview"] = $pictures[0][2];
 283                                                         $post["text"] = trim(str_replace($pictures[0][0], "", $body));
 284                                                 }
 285                                         }
 286                                 } elseif (count($pictures) > 0) {
 287                                         $post["type"] = "link";
 288                                         $post["url"] = $plink;
 289                                         $post["image"] = $pictures[0][2];
 290                                         $post["text"] = $body;
 291                                 }
 292                         } elseif (preg_match_all("(\[img\](.*?)\[\/img\])ism", $body, $pictures, PREG_SET_ORDER)) {
 293                                 if ((count($pictures) == 1) && !$has_title) {
 294                                         $post["type"] = "photo";
 295                                         $post["image"] = $pictures[0][1];
 296                                         $post["text"] = str_replace($pictures[0][0], "", $body);
 297                                 } elseif (count($pictures) > 0) {
 298                                         $post["type"] = "link";
 299                                         $post["url"] = $plink;
 300                                         $post["image"] = $pictures[0][1];
 301                                         $post["text"] = $body;
 302                                 }
 303                         }
 304
 305                         // Test for the external links
 306                         preg_match_all("(\[url\](.*?)\[\/url\])ism", $body, $links1, PREG_SET_ORDER);
 307                         preg_match_all("(\[url\=(.*?)\].*?\[\/url\])ism", $body, $links2, PREG_SET_ORDER);
 308
 309                         $links = array_merge($links1, $links2);
 310
 311                         // If there is only a single one, then use it.
 312                         // This should cover link posts via API.
 313                         if ((count($links) == 1) && !isset($post["preview"]) && !$has_title) {
 314                                 $post["type"] = "link";
 315                                 $post["text"] = trim($body);
 316                                 $post["url"] = $links[0][1];
 317                         }
 318
 319                         // Now count the number of external media links
 320                         preg_match_all("(\[vimeo\](.*?)\[\/vimeo\])ism", $body, $links1, PREG_SET_ORDER);
 321                         preg_match_all("(\[youtube\\](.*?)\[\/youtube\\])ism", $body, $links2, PREG_SET_ORDER);
 322                         preg_match_all("(\[video\\](.*?)\[\/video\\])ism", $body, $links3, PREG_SET_ORDER);
 323                         preg_match_all("(\[audio\\](.*?)\[\/audio\\])ism", $body, $links4, PREG_SET_ORDER);
 324
 325                         // Add them to the other external links
 326                         $links = array_merge($links, $links1, $links2, $links3, $links4);
 327
 328                         // Are there more than one?
 329                         if (count($links) > 1) {
 330                                 // The post will be the type "text", which means a blog post
 331                                 unset($post["type"]);
 332                                 $post["url"] = $plink;
 333                         }
 334
 335                         if (!isset($post["type"])) {
 336                                 $post["type"] = "text";
 337                                 $post["text"] = trim($body);
 338                         }
 339                 } elseif (isset($post["url"]) && ($post["type"] == "video")) {
 340                         $data = ParseUrl::getSiteinfoCached($post["url"], true);
 341
 342                         if (isset($data["images"][0])) {
 343                                 $post["image"] = $data["images"][0]["src"];
 344                         }
 345                 }
 346
 347                 return $post;
 348         }
 349
 350         /**
 351          * @brief Converts a BBCode text into plaintext
 352          *
 353          * @param      $text
 354          * @param bool $keep_urls Whether to keep URLs in the resulting plaintext
 355          *
 356          * @return string
 357          */
 358         public static function toPlaintext($text, $keep_urls = true)
 359         {
 360                 $naked_text = preg_replace('/\[.+?\]/','', $text);
 361                 if (!$keep_urls) {
 362                         $naked_text = preg_replace('#https?\://[^\s<]+[^\s\.\)]#i', '', $naked_text);
 363                 }
 364
 365                 return $naked_text;
 366         }
 367
 368         private static function proxyUrl($image, $simplehtml = false)
 369         {
 370                 // Only send proxied pictures to API and for internal display
 371                 if (in_array($simplehtml, [false, 2])) {
 372                         return ProxyUtils::proxifyUrl($image);
 373                 } else {
 374                         return $image;
 375                 }
 376         }
 377
 378         public static function scaleExternalImages($srctext, $include_link = true, $scale_replace = false)
 379         {
 380                 // Suppress "view full size"
 381                 if (intval(Config::get('system', 'no_view_full_size'))) {
 382                         $include_link = false;
 383                 }
 384
 385                 // Picture addresses can contain special characters
 386                 $s = htmlspecialchars_decode($srctext);
 387
 388                 $matches = null;
 389                 $c = preg_match_all('/\[img.*?\](.*?)\[\/img\]/ism', $s, $matches, PREG_SET_ORDER);
 390                 if ($c) {
 391                         foreach ($matches as $mtch) {
 392                                 Logger::log('scale_external_image: ' . $mtch[1]);
 393
 394                                 $hostname = str_replace('www.', '', substr(System::baseUrl(), strpos(System::baseUrl(), '://') + 3));
 395                                 if (stristr($mtch[1], $hostname)) {
 396                                         continue;
 397                                 }
 398
 399                                 // $scale_replace, if passed, is an array of two elements. The
 400                                 // first is the name of the full-size image. The second is the
 401                                 // name of a remote, scaled-down version of the full size image.
 402                                 // This allows Friendica to display the smaller remote image if
 403                                 // one exists, while still linking to the full-size image
 404                                 if ($scale_replace) {
 405                                         $scaled = str_replace($scale_replace[0], $scale_replace[1], $mtch[1]);
 406                                 } else {
 407                                         $scaled = $mtch[1];
 408                                 }
 409                                 $i = Network::fetchUrl($scaled);
 410                                 if (!$i) {
 411                                         return $srctext;
 412                                 }
 413
 414                                 // guess mimetype from headers or filename
 415                                 $type = Image::guessType($mtch[1], true);
 416
 417                                 if ($i) {
 418                                         $Image = new Image($i, $type);
 419                                         if ($Image->isValid()) {
 420                                                 $orig_width = $Image->getWidth();
 421                                                 $orig_height = $Image->getHeight();
 422
 423                                                 if ($orig_width > 640 || $orig_height > 640) {
 424                                                         $Image->scaleDown(640);
 425                                                         $new_width = $Image->getWidth();
 426                                                         $new_height = $Image->getHeight();
 427                                                         Logger::log('scale_external_images: ' . $orig_width . '->' . $new_width . 'w ' . $orig_height . '->' . $new_height . 'h' . ' match: ' . $mtch[0], Logger::DEBUG);
 428                                                         $s = str_replace(
 429                                                                 $mtch[0],
 430                                                                 '[img=' . $new_width . 'x' . $new_height. ']' . $scaled . '[/img]'
 431                                                                 . "\n" . (($include_link)
 432                                                                         ? '[url=' . $mtch[1] . ']' . L10n::t('view full size') . '[/url]' . "\n"
 433                                                                         : ''),
 434                                                                 $s
 435                                                         );
 436                                                         Logger::log('scale_external_images: new string: ' . $s, Logger::DEBUG);
 437                                                 }
 438                                         }
 439                                 }
 440                         }
 441                 }
 442
 443                 // replace the special char encoding
 444                 $s = htmlspecialchars($s, ENT_NOQUOTES, 'UTF-8');
 445                 return $s;
 446         }
 447
 448         /**
 449          * The purpose of this function is to apply system message length limits to
 450          * imported messages without including any embedded photos in the length
 451          *
 452          * @brief Truncates imported message body string length to max_import_size
 453          * @param string $body
 454          * @return string
 455          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 456          */
 457         public static function limitBodySize($body)
 458         {
 459                 $maxlen = Config::get('config', 'max_import_size', 0);
 460
 461                 // If the length of the body, including the embedded images, is smaller
 462                 // than the maximum, then don't waste time looking for the images
 463                 if ($maxlen && (strlen($body) > $maxlen)) {
 464
 465                         Logger::log('the total body length exceeds the limit', Logger::DEBUG);
 466
 467                         $orig_body = $body;
 468                         $new_body = '';
 469                         $textlen = 0;
 470
 471                         $img_start = strpos($orig_body, '[img');
 472                         $img_st_close = ($img_start !== false ? strpos(substr($orig_body, $img_start), ']') : false);
 473                         $img_end = ($img_start !== false ? strpos(substr($orig_body, $img_start), '[/img]') : false);
 474                         while (($img_st_close !== false) && ($img_end !== false)) {
 475
 476                                 $img_st_close++; // make it point to AFTER the closing bracket
 477                                 $img_end += $img_start;
 478                                 $img_end += strlen('[/img]');
 479
 480                                 if (!strcmp(substr($orig_body, $img_start + $img_st_close, 5), 'data:')) {
 481                                         // This is an embedded image
 482
 483                                         if (($textlen + $img_start) > $maxlen) {
 484                                                 if ($textlen < $maxlen) {
 485                                                         Logger::log('the limit happens before an embedded image', Logger::DEBUG);
 486                                                         $new_body = $new_body . substr($orig_body, 0, $maxlen - $textlen);
 487                                                         $textlen = $maxlen;
 488                                                 }
 489                                         } else {
 490                                                 $new_body = $new_body . substr($orig_body, 0, $img_start);
 491                                                 $textlen += $img_start;
 492                                         }
 493
 494                                         $new_body = $new_body . substr($orig_body, $img_start, $img_end - $img_start);
 495                                 } else {
 496
 497                                         if (($textlen + $img_end) > $maxlen) {
 498                                                 if ($textlen < $maxlen) {
 499                                                         Logger::log('the limit happens before the end of a non-embedded image', Logger::DEBUG);
 500                                                         $new_body = $new_body . substr($orig_body, 0, $maxlen - $textlen);
 501                                                         $textlen = $maxlen;
 502                                                 }
 503                                         } else {
 504                                                 $new_body = $new_body . substr($orig_body, 0, $img_end);
 505                                                 $textlen += $img_end;
 506                                         }
 507                                 }
 508                                 $orig_body = substr($orig_body, $img_end);
 509
 510                                 if ($orig_body === false) {
 511                                         // in case the body ends on a closing image tag
 512                                         $orig_body = '';
 513                                 }
 514
 515                                 $img_start = strpos($orig_body, '[img');
 516                                 $img_st_close = ($img_start !== false ? strpos(substr($orig_body, $img_start), ']') : false);
 517                                 $img_end = ($img_start !== false ? strpos(substr($orig_body, $img_start), '[/img]') : false);
 518                         }
 519
 520                         if (($textlen + strlen($orig_body)) > $maxlen) {
 521                                 if ($textlen < $maxlen) {
 522                                         Logger::log('the limit happens after the end of the last image', Logger::DEBUG);
 523                                         $new_body = $new_body . substr($orig_body, 0, $maxlen - $textlen);
 524                                 }
 525                         } else {
 526                                 Logger::log('the text size with embedded images extracted did not violate the limit', Logger::DEBUG);
 527                                 $new_body = $new_body . $orig_body;
 528                         }
 529
 530                         return $new_body;
 531                 } else {
 532                         return $body;
 533                 }
 534         }
 535
 536         /**
 537          * Processes [attachment] tags
 538          *
 539          * Note: Can produce a [bookmark] tag in the returned string
 540          *
 541          * @brief Processes [attachment] tags
 542          * @param string   $return
 543          * @param bool|int $simplehtml
 544          * @param bool     $tryoembed
 545          * @return string
 546          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 547          */
 548         private static function convertAttachment($return, $simplehtml = false, $tryoembed = true)
 549         {
 550                 $data = self::getAttachmentData($return);
 551                 if (empty($data) || empty($data["url"])) {
 552                         return $return;
 553                 }
 554
 555                 if (isset($data["title"])) {
 556                         $data["title"] = strip_tags($data["title"]);
 557                         $data["title"] = str_replace(["http://", "https://"], "", $data["title"]);
 558                 } else {
 559                         $data["title"] = null;
 560                 }
 561
 562                 if (((strpos($data["text"], "[img=") !== false) || (strpos($data["text"], "[img]") !== false) || Config::get('system', 'always_show_preview')) && !empty($data["image"])) {
 563                         $data["preview"] = $data["image"];
 564                         $data["image"] = "";
 565                 }
 566
 567                 $return = '';
 568                 if (in_array($simplehtml, [7, 9])) {
 569                         $return = self::convertUrlForOStatus($data["url"]);
 570                 } elseif (($simplehtml != 4) && ($simplehtml != 0)) {
 571                         $return = sprintf('<a href="%s" target="_blank">%s</a><br>', $data["url"], $data["title"]);
 572                 } else {
 573                         try {
 574                                 if ($tryoembed && OEmbed::isAllowedURL($data['url'])) {
 575                                         $return = OEmbed::getHTML($data['url'], $data['title']);
 576                                 } else {
 577                                         throw new Exception('OEmbed is disabled for this attachment.');
 578                                 }
 579                         } catch (Exception $e) {
 580                                 $data["title"] = defaults($data, 'title', $data['url']);
 581
 582                                 if ($simplehtml != 4) {
 583                                         $return = sprintf('<div class="type-%s">', $data["type"]);
 584                                 }
 585
 586                                 if (!empty($data['title']) && !empty($data['url'])) {
 587                                         if (!empty($data["image"]) && empty($data["text"]) && ($data["type"] == "photo")) {
 588                                                 $return .= sprintf('<a href="%s" target="_blank"><img src="%s" alt="" title="%s" class="attachment-image" /></a>', $data["url"], self::proxyUrl($data["image"], $simplehtml), $data["title"]);
 589                                         } else {
 590                                                 if (!empty($data["image"])) {
 591                                                         $return .= sprintf('<a href="%s" target="_blank"><img src="%s" alt="" title="%s" class="attachment-image" /></a><br />', $data["url"], self::proxyUrl($data["image"], $simplehtml), $data["title"]);
 592                                                 } elseif (!empty($data["preview"])) {
 593                                                         $return .= sprintf('<a href="%s" target="_blank"><img src="%s" alt="" title="%s" class="attachment-preview" /></a><br />', $data["url"], self::proxyUrl($data["preview"], $simplehtml), $data["title"]);
 594                                                 }
 595                                                 $return .= sprintf('<h4><a href="%s">%s</a></h4>', $data['url'], $data['title']);
 596                                         }
 597                                 }
 598
 599                                 if (!empty($data["description"]) && $data["description"] != $data["title"]) {
 600                                         // Sanitize the HTML by converting it to BBCode
 601                                         $bbcode = HTML::toBBCode($data["description"]);
 602                                         $return .= sprintf('<blockquote>%s</blockquote>', trim(self::convert($bbcode)));
 603                                 }
 604
 605                                 if (!empty($data['url'])) {
 606                                         $return .= sprintf('<sup><a href="%s">%s</a></sup>', $data['url'], parse_url($data['url'], PHP_URL_HOST));
 607                                 }
 608
 609                                 if ($simplehtml != 4) {
 610                                         $return .= '</div>';
 611                                 }
 612                         }
 613                 }
 614
 615                 return trim(defaults($data, 'text', '') . ' ' . $return . ' ' . defaults($data, 'after', ''));
 616         }
 617
 618         public static function removeShareInformation($Text, $plaintext = false, $nolink = false)
 619         {
 620                 $data = self::getAttachmentData($Text);
 621
 622                 if (!$data) {
 623                         return $Text;
 624                 } elseif ($nolink) {
 625                         return $data["text"] . defaults($data, 'after', '');
 626                 }
 627
 628                 $title = htmlentities(defaults($data, 'title', ''), ENT_QUOTES, 'UTF-8', false);
 629                 $text = htmlentities($data["text"], ENT_QUOTES, 'UTF-8', false);
 630                 if ($plaintext || (($title != "") && strstr($text, $title))) {
 631                         $data["title"] = $data["url"];
 632                 } elseif (($text != "") && strstr($title, $text)) {
 633                         $data["text"] = $data["title"];
 634                         $data["title"] = $data["url"];
 635                 }
 636
 637                 if (empty($data["text"]) && !empty($data["title"]) && empty($data["url"])) {
 638                         return $data["title"] . $data["after"];
 639                 }
 640
 641                 // If the link already is included in the post, don't add it again
 642                 if (!empty($data["url"]) && strpos($data["text"], $data["url"])) {
 643                         return $data["text"] . $data["after"];
 644                 }
 645
 646                 $text = $data["text"];
 647
 648                 if (!empty($data["url"]) && !empty($data["title"])) {
 649                         $text .= "\n[url=" . $data["url"] . "]" . $data["title"] . "[/url]";
 650                 } elseif (!empty($data["url"])) {
 651                         $text .= "\n[url]" . $data["url"] . "[/url]";
 652                 }
 653
 654                 return $text . "\n" . $data["after"];
 655         }
 656
 657         /**
 658          * Converts [url] BBCodes in a format that looks fine on Mastodon. (callback function)
 659          *
 660          * @brief Converts [url] BBCodes in a format that looks fine on Mastodon. (callback function)
 661          * @param array $match Array with the matching values
 662          * @return string reformatted link including HTML codes
 663          */
 664         private static function convertUrlForOStatusCallback($match)
 665         {
 666                 $url = $match[1];
 667
 668                 if (isset($match[2]) && ($match[1] != $match[2])) {
 669                         return $match[0];
 670                 }
 671
 672                 $parts = parse_url($url);
 673                 if (!isset($parts['scheme'])) {
 674                         return $match[0];
 675                 }
 676
 677                 return self::convertUrlForOStatus($url);
 678         }
 679
 680         /**
 681          * @brief Converts [url] BBCodes in a format that looks fine on OStatus systems.
 682          * @param string $url URL that is about to be reformatted
 683          * @return string reformatted link including HTML codes
 684          */
 685         private static function convertUrlForOStatus($url)
 686         {
 687                 $parts = parse_url($url);
 688                 $scheme = $parts['scheme'] . '://';
 689                 $styled_url = str_replace($scheme, '', $url);
 690
 691                 if (strlen($styled_url) > 30) {
 692                         $styled_url = substr($styled_url, 0, 30) . "…";
 693                 }
 694
 695                 $html = '<a href="%s" target="_blank">%s</a>';
 696
 697                 return sprintf($html, $url, $styled_url);
 698         }
 699
 700         /*
 701          * [noparse][i]italic[/i][/noparse] turns into
 702          * [noparse][ i ]italic[ /i ][/noparse],
 703          * to hide them from parser.
 704          */
 705         private static function escapeNoparseCallback($match)
 706         {
 707                 $whole_match = $match[0];
 708                 $captured = $match[1];
 709                 $spacefied = preg_replace("/\[(.*?)\]/", "[ $1 ]", $captured);
 710                 $new_str = str_replace($captured, $spacefied, $whole_match);
 711                 return $new_str;
 712         }
 713
 714         /*
 715          * The previously spacefied [noparse][ i ]italic[ /i ][/noparse],
 716          * now turns back and the [noparse] tags are trimed
 717          * returning [i]italic[/i]
 718          */
 719         private static function unescapeNoparseCallback($match)
 720         {
 721                 $captured = $match[1];
 722                 $unspacefied = preg_replace("/\[ (.*?)\ ]/", "[$1]", $captured);
 723                 return $unspacefied;
 724         }
 725
 726         /**
 727          * Returns the bracket character positions of a set of opening and closing BBCode tags, optionally skipping first
 728          * occurrences
 729          *
 730          * @param string $text        Text to search
 731          * @param string $name        Tag name
 732          * @param int    $occurrences Number of first occurrences to skip
 733          * @return boolean|array
 734          */
 735         public static function getTagPosition($text, $name, $occurrences = 0)
 736         {
 737                 if ($occurrences < 0) {
 738                         $occurrences = 0;
 739                 }
 740
 741                 $start_open = -1;
 742                 for ($i = 0; $i <= $occurrences; $i++) {
 743                         if ($start_open !== false) {
 744                                 $start_open = strpos($text, '[' . $name, $start_open + 1); // allow [name= type tags
 745                         }
 746                 }
 747
 748                 if ($start_open === false) {
 749                         return false;
 750                 }
 751
 752                 $start_equal = strpos($text, '=', $start_open);
 753                 $start_close = strpos($text, ']', $start_open);
 754
 755                 if ($start_close === false) {
 756                         return false;
 757                 }
 758
 759                 $start_close++;
 760
 761                 $end_open = strpos($text, '[/' . $name . ']', $start_close);
 762
 763                 if ($end_open === false) {
 764                         return false;
 765                 }
 766
 767                 $res = [
 768                         'start' => [
 769                                 'open' => $start_open,
 770                                 'close' => $start_close
 771                         ],
 772                         'end' => [
 773                                 'open' => $end_open,
 774                                 'close' => $end_open + strlen('[/' . $name . ']')
 775                         ],
 776                 ];
 777
 778                 if ($start_equal !== false) {
 779                         $res['start']['equal'] = $start_equal + 1;
 780                 }
 781
 782                 return $res;
 783         }
 784
 785         /**
 786          * Performs a preg_replace within the boundaries of all named BBCode tags in a text
 787          *
 788          * @param string $pattern Preg pattern string
 789          * @param string $replace Preg replace string
 790          * @param string $name    BBCode tag name
 791          * @param string $text    Text to search
 792          * @return string
 793          */
 794         public static function pregReplaceInTag($pattern, $replace, $name, $text)
 795         {
 796                 $occurrences = 0;
 797                 $pos = self::getTagPosition($text, $name, $occurrences);
 798                 while ($pos !== false && $occurrences++ < 1000) {
 799                         $start = substr($text, 0, $pos['start']['open']);
 800                         $subject = substr($text, $pos['start']['open'], $pos['end']['close'] - $pos['start']['open']);
 801                         $end = substr($text, $pos['end']['close']);
 802                         if ($end === false) {
 803                                 $end = '';
 804                         }
 805
 806                         $subject = preg_replace($pattern, $replace, $subject);
 807                         $text = $start . $subject . $end;
 808
 809                         $pos = self::getTagPosition($text, $name, $occurrences);
 810                 }
 811
 812                 return $text;
 813         }
 814
 815         private static function extractImagesFromItemBody($body)
 816         {
 817                 $saved_image = [];
 818                 $orig_body = $body;
 819                 $new_body = '';
 820
 821                 $cnt = 0;
 822                 $img_start = strpos($orig_body, '[img');
 823                 $img_st_close = ($img_start !== false ? strpos(substr($orig_body, $img_start), ']') : false);
 824                 $img_end = ($img_start !== false ? strpos(substr($orig_body, $img_start), '[/img]') : false);
 825                 while (($img_st_close !== false) && ($img_end !== false)) {
 826                         $img_st_close++; // make it point to AFTER the closing bracket
 827                         $img_end += $img_start;
 828
 829                         if (!strcmp(substr($orig_body, $img_start + $img_st_close, 5), 'data:')) {
 830                                 // This is an embedded image
 831                                 $saved_image[$cnt] = substr($orig_body, $img_start + $img_st_close, $img_end - ($img_start + $img_st_close));
 832                                 $new_body = $new_body . substr($orig_body, 0, $img_start) . '[$#saved_image' . $cnt . '#$]';
 833
 834                                 $cnt++;
 835                         } else {
 836                                 $new_body = $new_body . substr($orig_body, 0, $img_end + strlen('[/img]'));
 837                         }
 838
 839                         $orig_body = substr($orig_body, $img_end + strlen('[/img]'));
 840
 841                         if ($orig_body === false) {
 842                                 // in case the body ends on a closing image tag
 843                                 $orig_body = '';
 844                         }
 845
 846                         $img_start = strpos($orig_body, '[img');
 847                         $img_st_close = ($img_start !== false ? strpos(substr($orig_body, $img_start), ']') : false);
 848                         $img_end = ($img_start !== false ? strpos(substr($orig_body, $img_start), '[/img]') : false);
 849                 }
 850
 851                 $new_body = $new_body . $orig_body;
 852
 853                 return ['body' => $new_body, 'images' => $saved_image];
 854         }
 855
 856         private static function interpolateSavedImagesIntoItemBody($body, array $images)
 857         {
 858                 $newbody = $body;
 859
 860                 $cnt = 0;
 861                 foreach ($images as $image) {
 862                         // We're depending on the property of 'foreach' (specified on the PHP website) that
 863                         // it loops over the array starting from the first element and going sequentially
 864                         // to the last element
 865                         $newbody = str_replace('[$#saved_image' . $cnt . '#$]',
 866                                 '<img src="' . self::proxyUrl($image) . '" alt="' . L10n::t('Image/photo') . '" />', $newbody);
 867                         $cnt++;
 868                 }
 869
 870                 return $newbody;
 871         }
 872
 873         /**
 874          * This function converts a [share] block to text according to a provided callback function whose signature is:
 875          *
 876          * function(array $attributes, array $author_contact, string $content, boolean $is_quote_share): string
 877          *
 878          * Where:
 879          * - $attributes is an array of attributes of the [share] block itself. Missing keys will be completed by the contact
 880          * data lookup
 881          * - $author_contact is a contact record array
 882          * - $content is the inner content of the [share] block
 883          * - $is_quote_share indicates whether there's any content before the [share] block
 884          * - Return value is the string that should replace the [share] block in the provided text
 885          *
 886          * This function is intended to be used by addon connector to format a share block like the target network is expecting it.
 887          *
 888          * @param  string   $text     A BBCode string
 889          * @param  callable $callback
 890          * @return string The BBCode string with all [share] blocks replaced
 891          */
 892         public static function convertShare($text, callable $callback)
 893         {
 894                 $return = preg_replace_callback(
 895                         "/(.*?)\[share(.*?)\](.*?)\[\/share\]/ism",
 896                         function ($match) use ($callback) {
 897                                 $attribute_string = $match[2];
 898
 899                                 $attributes = [];
 900                                 foreach(['author', 'profile', 'avatar', 'link', 'posted'] as $field) {
 901                                         preg_match("/$field=(['\"])(.+?)\\1/ism", $attribute_string, $matches);
 902                                         $attributes[$field] = html_entity_decode(defaults($matches, 2, ''), ENT_QUOTES, 'UTF-8');
 903                                 }
 904
 905                                 // We only call this so that a previously unknown contact can be added.
 906                                 // This is important for the function "Model\Contact::getDetailsByURL()".
 907                                 // This function then can fetch an entry from the contact table.
 908                                 $default['url'] = $attributes['profile'];
 909
 910                                 if (!empty($attributes['author'])) {
 911                                         $default['name'] = $attributes['author'];
 912                                 }
 913
 914                                 if (!empty($attributes['avatar'])) {
 915                                         $default['photo'] = $attributes['avatar'];
 916                                 }
 917
 918                                 Contact::getIdForURL($attributes['profile'], 0, true, $default);
 919
 920                                 $author_contact = Contact::getDetailsByURL($attributes['profile']);
 921                                 $author_contact['addr'] = defaults($author_contact, 'addr' , Protocol::getAddrFromProfileUrl($attributes['profile']));
 922
 923                                 $attributes['author']   = defaults($author_contact, 'name' , $attributes['author']);
 924                                 $attributes['avatar']   = defaults($author_contact, 'micro', $attributes['avatar']);
 925                                 $attributes['profile']  = defaults($author_contact, 'url'  , $attributes['profile']);
 926
 927                                 if ($attributes['avatar']) {
 928                                         $attributes['avatar'] = ProxyUtils::proxifyUrl($attributes['avatar'], false, ProxyUtils::SIZE_THUMB);
 929                                 }
 930
 931                                 return $match[1] . $callback($attributes, $author_contact, $match[3], trim($match[1]) != '');
 932                         },
 933                         $text
 934                 );
 935
 936                 return $return;
 937         }
 938
 939         /**
 940          * Default [share] tag conversion callback
 941          *
 942          * Note: Can produce a [bookmark] tag in the output
 943          *
 944          * @see BBCode::convertShare()
 945          * @param array   $attributes     [share] block attribute values
 946          * @param array   $author_contact Contact row of the shared author
 947          * @param string  $content        Inner content of the [share] block
 948          * @param boolean $is_quote_share Whether there is content before the [share] block
 949          * @param integer $simplehtml     Mysterious integer value depending on the target network/formatting style
 950          * @return string
 951          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
 952          */
 953         private static function convertShareCallback(array $attributes, array $author_contact, $content, $is_quote_share, $simplehtml)
 954         {
 955                 $mention = Protocol::formatMention($attributes['profile'], $attributes['author']);
 956
 957                 switch ($simplehtml) {
 958                         case 1:
 959                                 $text = ($is_quote_share? '<br />' : '') . '<p>' . html_entity_decode('&#x2672; ', ENT_QUOTES, 'UTF-8') . ' <a href="' . $attributes['profile'] . '">' . $mention . '</a>: </p>' . "\n" . '«' . $content . '»';
 960                                 break;
 961                         case 2:
 962                                 $text = ($is_quote_share? '<br />' : '') . '<p>' . html_entity_decode('&#x2672; ', ENT_QUOTES, 'UTF-8') . ' ' . $author_contact['addr'] . ': </p>' . "\n" . $content;
 963                                 break;
 964                         case 3: // Diaspora
 965                                 if (stripos(Strings::normaliseLink($attributes['link']), 'http://twitter.com/') === 0) {
 966                                         $text = ($is_quote_share? '<hr />' : '') . '<p><a href="' . $attributes['link'] . '">' . $attributes['link'] . '</a></p>' . "\n";
 967                                 } else {
 968                                         $headline = '<p><b>♲ <a href="' . $attributes['profile'] . '">' . $attributes['author'] . '</a>:</b></p>' . "\n";
 969
 970                                         if (!empty($attributes['posted']) && !empty($attributes['link'])) {
 971                                                 $headline = '<p><b>♲ <a href="' . $attributes['profile'] . '">' . $attributes['author'] . '</a></b> - <a href="' . $attributes['link'] . '">' . $attributes['posted'] . ' GMT</a></p>' . "\n";
 972                                         }
 973
 974                                         $text = ($is_quote_share? '<hr />' : '') . $headline . '<blockquote>' . trim($content) . '</blockquote>' . "\n";
 975
 976                                         if (empty($attributes['posted']) && !empty($attributes['link'])) {
 977                                                 $text .= '<p><a href="' . $attributes['link'] . '">[Source]</a></p>' . "\n";
 978                                         }
 979                                 }
 980
 981                                 break;
 982                         case 4:
 983                                 $headline = '<p><b>' . html_entity_decode('&#x2672; ', ENT_QUOTES, 'UTF-8');
 984                                 $headline .= L10n::t('<a href="%1$s" target="_blank">%2$s</a> %3$s', $attributes['link'], $mention, $attributes['posted']);
 985                                 $headline .= ':</b></p>' . "\n";
 986
 987                                 $text = ($is_quote_share? '<hr />' : '') . $headline . '<blockquote class="shared_content">' . trim($content) . '</blockquote>' . "\n";
 988
 989                                 break;
 990                         case 5:
 991                                 $text = ($is_quote_share? '<br />' : '') . '<p>' . html_entity_decode('&#x2672; ', ENT_QUOTES, 'UTF-8') . ' ' . $author_contact['addr'] . ': </p>' . "\n" . $content;
 992                                 break;
 993                         case 7: // statusnet/GNU Social
 994                         case 9: // ActivityPub
 995                                 $text = ($is_quote_share? '<br />' : '') . '<p>' . html_entity_decode('&#x2672; ', ENT_QUOTES, 'UTF-8') . ' @' . $author_contact['addr'] . ': ' . $content . '</p>' . "\n";
 996                                 break;
 997                         default:
 998                                 // Transforms quoted tweets in rich attachments to avoid nested tweets
 999                                 if (stripos(Strings::normaliseLink($attributes['link']), 'http://twitter.com/') === 0 && OEmbed::isAllowedURL($attributes['link'])) {
1000                                         try {
1001                                                 $text = ($is_quote_share? '<br />' : '') . OEmbed::getHTML($attributes['link']);
1002                                         } catch (Exception $e) {
1003                                                 $text = ($is_quote_share? '<br />' : '') . sprintf('[bookmark=%s]%s[/bookmark]', $attributes['link'], $content);
1004                                         }
1005                                 } else {
1006                                         $text = ($is_quote_share? "\n" : '');
1007
1008                                         $tpl = Renderer::getMarkupTemplate('shared_content.tpl');
1009                                         $text .= Renderer::replaceMacros($tpl, [
1010                                                 '$profile' => $attributes['profile'],
1011                                                 '$avatar'  => $attributes['avatar'],
1012                                                 '$author'  => $attributes['author'],
1013                                                 '$link'    => $attributes['link'],
1014                                                 '$posted'  => $attributes['posted'],
1015                                                 '$content' => trim($content)
1016                                         ]);
1017                                 }
1018                                 break;
1019                 }
1020
1021                 return $text;
1022         }
1023
1024         private static function removePictureLinksCallback($match)
1025         {
1026                 $text = Cache::get($match[1]);
1027
1028                 if (is_null($text)) {
1029                         $a = self::getApp();
1030
1031                         $stamp1 = microtime(true);
1032
1033                         $ch = @curl_init($match[1]);
1034                         @curl_setopt($ch, CURLOPT_NOBODY, true);
1035                         @curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
1036                         @curl_setopt($ch, CURLOPT_USERAGENT, $a->getUserAgent());
1037                         @curl_exec($ch);
1038                         $curl_info = @curl_getinfo($ch);
1039
1040                         $a->getProfiler()->saveTimestamp($stamp1, "network", System::callstack());
1041
1042                         if (substr($curl_info["content_type"], 0, 6) == "image/") {
1043                                 $text = "[url=" . $match[1] . "]" . $match[1] . "[/url]";
1044                         } else {
1045                                 $text = "[url=" . $match[2] . "]" . $match[2] . "[/url]";
1046
1047                                 // if its not a picture then look if its a page that contains a picture link
1048                                 $body = Network::fetchUrl($match[1]);
1049
1050                                 $doc = new DOMDocument();
1051                                 @$doc->loadHTML($body);
1052                                 $xpath = new DOMXPath($doc);
1053                                 $list = $xpath->query("//meta[@name]");
1054                                 foreach ($list as $node) {
1055                                         $attr = [];
1056
1057                                         if ($node->attributes->length) {
1058                                                 foreach ($node->attributes as $attribute) {
1059                                                         $attr[$attribute->name] = $attribute->value;
1060                                                 }
1061                                         }
1062
1063                                         if (strtolower($attr["name"]) == "twitter:image") {
1064                                                 $text = "[url=" . $attr["content"] . "]" . $attr["content"] . "[/url]";
1065                                         }
1066                                 }
1067                         }
1068                         Cache::set($match[1], $text);
1069                 }
1070
1071                 return $text;
1072         }
1073
1074         private static function expandLinksCallback($match)
1075         {
1076                 if (($match[3] == "") || ($match[2] == $match[3]) || stristr($match[2], $match[3])) {
1077                         return ($match[1] . "[url]" . $match[2] . "[/url]");
1078                 } else {
1079                         return ($match[1] . $match[3] . " [url]" . $match[2] . "[/url]");
1080                 }
1081         }
1082
1083         private static function cleanPictureLinksCallback($match)
1084         {
1085                 $text = Cache::get($match[1]);
1086
1087                 if (is_null($text)) {
1088                         $a = self::getApp();
1089
1090                         $stamp1 = microtime(true);
1091
1092                         $ch = @curl_init($match[1]);
1093                         @curl_setopt($ch, CURLOPT_NOBODY, true);
1094                         @curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
1095                         @curl_setopt($ch, CURLOPT_USERAGENT, $a->getUserAgent());
1096                         @curl_exec($ch);
1097                         $curl_info = @curl_getinfo($ch);
1098
1099                         $a->getProfiler()->saveTimestamp($stamp1, "network", System::callstack());
1100
1101                         // if its a link to a picture then embed this picture
1102                         if (substr($curl_info["content_type"], 0, 6) == "image/") {
1103                                 $text = "[img]" . $match[1] . "[/img]";
1104                         } else {
1105                                 if (!empty($match[3])) {
1106                                         $text = "[img=" . $match[2] . "]" . $match[3] . "[/img]";
1107                                 } else {
1108                                         $text = "[img]" . $match[2] . "[/img]";
1109                                 }
1110
1111                                 // if its not a picture then look if its a page that contains a picture link
1112                                 $body = Network::fetchUrl($match[1]);
1113
1114                                 $doc = new DOMDocument();
1115                                 @$doc->loadHTML($body);
1116                                 $xpath = new DOMXPath($doc);
1117                                 $list = $xpath->query("//meta[@name]");
1118                                 foreach ($list as $node) {
1119                                         $attr = [];
1120                                         if ($node->attributes->length) {
1121                                                 foreach ($node->attributes as $attribute) {
1122                                                         $attr[$attribute->name] = $attribute->value;
1123                                                 }
1124                                         }
1125
1126                                         if (strtolower($attr["name"]) == "twitter:image") {
1127                                                 if (!empty($match[3])) {
1128                                                         $text = "[img=" . $attr["content"] . "]" . $match[3] . "[/img]";
1129                                                 } else {
1130                                                         $text = "[img]" . $attr["content"] . "[/img]";
1131                                                 }
1132                                         }
1133                                 }
1134                         }
1135                         Cache::set($match[1], $text);
1136                 }
1137
1138                 return $text;
1139         }
1140
1141         public static function cleanPictureLinks($text)
1142         {
1143                 $return = preg_replace_callback("&\[url=([^\[\]]*)\]\[img=(.*)\](.*)\[\/img\]\[\/url\]&Usi", 'self::cleanPictureLinksCallback', $text);
1144                 $return = preg_replace_callback("&\[url=([^\[\]]*)\]\[img\](.*)\[\/img\]\[\/url\]&Usi", 'self::cleanPictureLinksCallback', $return);
1145                 return $return;
1146         }
1147
1148         /**
1149          * @brief Converts a BBCode message to HTML message
1150          *
1151          * BBcode 2 HTML was written by WAY2WEB.net
1152          * extended to work with Mistpark/Friendica - Mike Macgirvin
1153          *
1154          * Simple HTML values meaning:
1155          * - 0: Friendica display
1156          * - 1: Unused
1157          * - 2: Used for Windows Phone push, Friendica API
1158          * - 3: Used before converting to Markdown in bb2diaspora.php
1159          * - 4: Used for WordPress, Libertree (before Markdown), pump.io and tumblr
1160          * - 5: Unused
1161          * - 6: Unused
1162          * - 7: Used for dfrn, OStatus
1163          * - 8: Used for WP backlink text setting
1164          * - 9: ActivityPub
1165          *
1166          * @param string $text
1167          * @param bool   $try_oembed
1168          * @param int    $simple_html
1169          * @param bool   $for_plaintext
1170          * @return string
1171          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
1172          */
1173         public static function convert($text, $try_oembed = true, $simple_html = 0, $for_plaintext = false)
1174         {
1175                 $a = self::getApp();
1176
1177                 /*
1178                  * preg_match_callback function to replace potential Oembed tags with Oembed content
1179                  *
1180                  * $match[0] = [tag]$url[/tag] or [tag=$url]$title[/tag]
1181                  * $match[1] = $url
1182                  * $match[2] = $title or absent
1183                  */
1184                 $try_oembed_callback = function ($match)
1185                 {
1186                         $url = $match[1];
1187                         $title = defaults($match, 2, null);
1188
1189                         try {
1190                                 $return = OEmbed::getHTML($url, $title);
1191                         } catch (Exception $ex) {
1192                                 $return = $match[0];
1193                         }
1194
1195                         return $return;
1196                 };
1197
1198                 // Extracting code blocks before the whitespace processing and the autolinker
1199                 $codeblocks = [];
1200
1201                 $text = preg_replace_callback("#\[code(?:=([^\]]*))?\](.*?)\[\/code\]#ism",
1202                         function ($matches) use (&$codeblocks) {
1203                                 $return = '#codeblock-' . count($codeblocks) . '#';
1204                                 if (strpos($matches[2], "\n") !== false) {
1205                                         $codeblocks[] = '<pre><code class="language-' . trim($matches[1]) . '">' . trim($matches[2], "\n\r") . '</code></pre>';
1206                                 } else {
1207                                         $codeblocks[] = '<code>' . $matches[2] . '</code>';
1208                                 }
1209
1210                                 return $return;
1211                         },
1212                         $text
1213                 );
1214
1215                 // Hide all [noparse] contained bbtags by spacefying them
1216                 // POSSIBLE BUG --> Will the 'preg' functions crash if there's an embedded image?
1217
1218                 $text = preg_replace_callback("/\[noparse\](.*?)\[\/noparse\]/ism", 'self::escapeNoparseCallback', $text);
1219                 $text = preg_replace_callback("/\[nobb\](.*?)\[\/nobb\]/ism", 'self::escapeNoparseCallback', $text);
1220                 $text = preg_replace_callback("/\[pre\](.*?)\[\/pre\]/ism", 'self::escapeNoparseCallback', $text);
1221
1222                 // Remove the abstract element. It is a non visible element.
1223                 $text = self::stripAbstract($text);
1224
1225                 // Move all spaces out of the tags
1226                 $text = preg_replace("/\[(\w*)\](\s*)/ism", '$2[$1]', $text);
1227                 $text = preg_replace("/(\s*)\[\/(\w*)\]/ism", '[/$2]$1', $text);
1228
1229                 // Extract the private images which use data urls since preg has issues with
1230                 // large data sizes. Stash them away while we do bbcode conversion, and then put them back
1231                 // in after we've done all the regex matching. We cannot use any preg functions to do this.
1232
1233                 $extracted = self::extractImagesFromItemBody($text);
1234                 $text = $extracted['body'];
1235                 $saved_image = $extracted['images'];
1236
1237                 // If we find any event code, turn it into an event.
1238                 // After we're finished processing the bbcode we'll
1239                 // replace all of the event code with a reformatted version.
1240
1241                 $ev = Event::fromBBCode($text);
1242
1243                 // Replace any html brackets with HTML Entities to prevent executing HTML or script
1244                 // Don't use strip_tags here because it breaks [url] search by replacing & with amp
1245
1246                 $text = str_replace("<", "&lt;", $text);
1247                 $text = str_replace(">", "&gt;", $text);
1248
1249                 // remove some newlines before the general conversion
1250                 $text = preg_replace("/\s?\[share(.*?)\]\s?(.*?)\s?\[\/share\]\s?/ism", "[share$1]$2[/share]", $text);
1251                 $text = preg_replace("/\s?\[quote(.*?)\]\s?(.*?)\s?\[\/quote\]\s?/ism", "[quote$1]$2[/quote]", $text);
1252
1253                 // when the content is meant exporting to other systems then remove the avatar picture since this doesn't really look good on these systems
1254                 if (!$try_oembed) {
1255                         $text = preg_replace("/\[share(.*?)avatar\s?=\s?'.*?'\s?(.*?)\]\s?(.*?)\s?\[\/share\]\s?/ism", "\n[share$1$2]$3[/share]", $text);
1256                 }
1257
1258                 // Convert new line chars to html <br /> tags
1259
1260                 // nlbr seems to be hopelessly messed up
1261                 //      $Text = nl2br($Text);
1262
1263                 // We'll emulate it.
1264
1265                 $text = trim($text);
1266                 $text = str_replace("\r\n", "\n", $text);
1267
1268                 // Remove linefeeds inside of the table elements. See issue #6799
1269                 $search = ["\n[th]", "[th]\n", " [th]", "\n[/th]", "[/th]\n", "[/th] ",
1270                         "\n[td]", "[td]\n", " [td]", "\n[/td]", "[/td]\n", "[/td] ",
1271                         "\n[tr]", "[tr]\n", " [tr]", "[tr] ", "\n[/tr]", "[/tr]\n", " [/tr]", "[/tr] ",
1272                         "[table]\n", "[table] ", " [table]", "\n[/table]", " [/table]", "[/table] "];
1273                 $replace = ["[th]", "[th]", "[th]", "[/th]", "[/th]", "[/th]",
1274                         "[td]", "[td]", "[td]", "[/td]", "[/td]", "[/td]",
1275                         "[tr]", "[tr]", "[tr]", "[tr]", "[/tr]", "[/tr]", "[/tr]", "[/tr]",
1276                         "[table]", "[table]", "[table]", "[/table]", "[/table]", "[/table]"];
1277                 do {
1278                         $oldtext = $text;
1279                         $text = str_replace($search, $replace, $text);
1280                 } while ($oldtext != $text);
1281
1282                 // Replace these here only once
1283                 $search = ["\n[table]", "[/table]\n"];
1284                 $replace = ["[table]", "[/table]"];
1285                 $text = str_replace($search, $replace, $text);
1286
1287                 // removing multiplicated newlines
1288                 if (Config::get("system", "remove_multiplicated_lines")) {
1289                         $search = ["\n\n\n", "\n ", " \n", "[/quote]\n\n", "\n[/quote]", "[/li]\n", "\n[li]", "\n[ul]", "[/ul]\n", "\n\n[share ", "[/attachment]\n",
1290                                         "\n[h1]", "[/h1]\n", "\n[h2]", "[/h2]\n", "\n[h3]", "[/h3]\n", "\n[h4]", "[/h4]\n", "\n[h5]", "[/h5]\n", "\n[h6]", "[/h6]\n"];
1291                         $replace = ["\n\n", "\n", "\n", "[/quote]\n", "[/quote]", "[/li]", "[li]", "[ul]", "[/ul]", "\n[share ", "[/attachment]",
1292                                         "[h1]", "[/h1]", "[h2]", "[/h2]", "[h3]", "[/h3]", "[h4]", "[/h4]", "[h5]", "[/h5]", "[h6]", "[/h6]"];
1293                         do {
1294                                 $oldtext = $text;
1295                                 $text = str_replace($search, $replace, $text);
1296                         } while ($oldtext != $text);
1297                 }
1298
1299                 // Handle attached links or videos
1300                 $text = self::convertAttachment($text, $simple_html, $try_oembed);
1301
1302                 // leave open the posibility of [map=something]
1303                 // this is replaced in Item::prepareBody() which has knowledge of the item location
1304                 if (strpos($text, '[/map]') !== false) {
1305                         $text = preg_replace_callback(
1306                                 "/\[map\](.*?)\[\/map\]/ism",
1307                                 function ($match) use ($simple_html) {
1308                                         return str_replace($match[0], '<p class="map">' . Map::byLocation($match[1], $simple_html) . '</p>', $match[0]);
1309                                 },
1310                                 $text
1311                         );
1312                 }
1313
1314                 if (strpos($text, '[map=') !== false) {
1315                         $text = preg_replace_callback(
1316                                 "/\[map=(.*?)\]/ism",
1317                                 function ($match) use ($simple_html) {
1318                                         return str_replace($match[0], '<p class="map">' . Map::byCoordinates(str_replace('/', ' ', $match[1]), $simple_html) . '</p>', $match[0]);
1319                                 },
1320                                 $text
1321                         );
1322                 }
1323
1324                 if (strpos($text, '[map]') !== false) {
1325                         $text = preg_replace("/\[map\]/", '<p class="map"></p>', $text);
1326                 }
1327
1328                 // Check for headers
1329                 $text = preg_replace("(\[h1\](.*?)\[\/h1\])ism", '<h1>$1</h1>', $text);
1330                 $text = preg_replace("(\[h2\](.*?)\[\/h2\])ism", '<h2>$1</h2>', $text);
1331                 $text = preg_replace("(\[h3\](.*?)\[\/h3\])ism", '<h3>$1</h3>', $text);
1332                 $text = preg_replace("(\[h4\](.*?)\[\/h4\])ism", '<h4>$1</h4>', $text);
1333                 $text = preg_replace("(\[h5\](.*?)\[\/h5\])ism", '<h5>$1</h5>', $text);
1334                 $text = preg_replace("(\[h6\](.*?)\[\/h6\])ism", '<h6>$1</h6>', $text);
1335
1336                 // Check for paragraph
1337                 $text = preg_replace("(\[p\](.*?)\[\/p\])ism", '<p>$1</p>', $text);
1338
1339                 // Check for bold text
1340                 $text = preg_replace("(\[b\](.*?)\[\/b\])ism", '<strong>$1</strong>', $text);
1341
1342                 // Check for Italics text
1343                 $text = preg_replace("(\[i\](.*?)\[\/i\])ism", '<em>$1</em>', $text);
1344
1345                 // Check for Underline text
1346                 $text = preg_replace("(\[u\](.*?)\[\/u\])ism", '<u>$1</u>', $text);
1347
1348                 // Check for strike-through text
1349                 $text = preg_replace("(\[s\](.*?)\[\/s\])ism", '<s>$1</s>', $text);
1350
1351                 // Check for over-line text
1352                 $text = preg_replace("(\[o\](.*?)\[\/o\])ism", '<span class="overline">$1</span>', $text);
1353
1354                 // Check for colored text
1355                 $text = preg_replace("(\[color=(.*?)\](.*?)\[\/color\])ism", "<span style=\"color: $1;\">$2</span>", $text);
1356
1357                 // Check for sized text
1358                 // [size=50] --> font-size: 50px (with the unit).
1359                 $text = preg_replace("(\[size=(\d*?)\](.*?)\[\/size\])ism", "<span style=\"font-size: $1px; line-height: initial;\">$2</span>", $text);
1360                 $text = preg_replace("(\[size=(.*?)\](.*?)\[\/size\])ism", "<span style=\"font-size: $1; line-height: initial;\">$2</span>", $text);
1361
1362                 // Check for centered text
1363                 $text = preg_replace("(\[center\](.*?)\[\/center\])ism", "<div style=\"text-align:center;\">$1</div>", $text);
1364
1365                 // Check for list text
1366                 $text = str_replace("[*]", "<li>", $text);
1367
1368                 // Check for style sheet commands
1369                 $text = preg_replace_callback(
1370                         "(\[style=(.*?)\](.*?)\[\/style\])ism",
1371                         function ($match) {
1372                                 return "<span style=\"" . HTML::sanitizeCSS($match[1]) . ";\">" . $match[2] . "</span>";
1373                         },
1374                         $text
1375                 );
1376
1377                 // Check for CSS classes
1378                 $text = preg_replace_callback(
1379                         "(\[class=(.*?)\](.*?)\[\/class\])ism",
1380                         function ($match) {
1381                                 return "<span class=\"" . HTML::sanitizeCSS($match[1]) . "\">" . $match[2] . "</span>";
1382                         },
1383                         $text
1384                 );
1385
1386                 // handle nested lists
1387                 $endlessloop = 0;
1388
1389                 while ((((strpos($text, "[/list]") !== false) && (strpos($text, "[list") !== false)) ||
1390                                 ((strpos($text, "[/ol]") !== false) && (strpos($text, "[ol]") !== false)) ||
1391                                 ((strpos($text, "[/ul]") !== false) && (strpos($text, "[ul]") !== false)) ||
1392                                 ((strpos($text, "[/li]") !== false) && (strpos($text, "[li]") !== false))) && (++$endlessloop < 20)) {
1393                         $text = preg_replace("/\[list\](.*?)\[\/list\]/ism", '<ul class="listbullet" style="list-style-type: circle;">$1</ul>', $text);
1394                         $text = preg_replace("/\[list=\](.*?)\[\/list\]/ism", '<ul class="listnone" style="list-style-type: none;">$1</ul>', $text);
1395                         $text = preg_replace("/\[list=1\](.*?)\[\/list\]/ism", '<ul class="listdecimal" style="list-style-type: decimal;">$1</ul>', $text);
1396                         $text = preg_replace("/\[list=((?-i)i)\](.*?)\[\/list\]/ism", '<ul class="listlowerroman" style="list-style-type: lower-roman;">$2</ul>', $text);
1397                         $text = preg_replace("/\[list=((?-i)I)\](.*?)\[\/list\]/ism", '<ul class="listupperroman" style="list-style-type: upper-roman;">$2</ul>', $text);
1398                         $text = preg_replace("/\[list=((?-i)a)\](.*?)\[\/list\]/ism", '<ul class="listloweralpha" style="list-style-type: lower-alpha;">$2</ul>', $text);
1399                         $text = preg_replace("/\[list=((?-i)A)\](.*?)\[\/list\]/ism", '<ul class="listupperalpha" style="list-style-type: upper-alpha;">$2</ul>', $text);
1400                         $text = preg_replace("/\[ul\](.*?)\[\/ul\]/ism", '<ul class="listbullet" style="list-style-type: circle;">$1</ul>', $text);
1401                         $text = preg_replace("/\[ol\](.*?)\[\/ol\]/ism", '<ul class="listdecimal" style="list-style-type: decimal;">$1</ul>', $text);
1402                         $text = preg_replace("/\[li\](.*?)\[\/li\]/ism", '<li>$1</li>', $text);
1403                 }
1404
1405                 $text = preg_replace("/\[th\](.*?)\[\/th\]/sm", '<th>$1</th>', $text);
1406                 $text = preg_replace("/\[td\](.*?)\[\/td\]/sm", '<td>$1</td>', $text);
1407                 $text = preg_replace("/\[tr\](.*?)\[\/tr\]/sm", '<tr>$1</tr>', $text);
1408                 $text = preg_replace("/\[table\](.*?)\[\/table\]/sm", '<table>$1</table>', $text);
1409
1410                 $text = preg_replace("/\[table border=1\](.*?)\[\/table\]/sm", '<table border="1" >$1</table>', $text);
1411                 $text = preg_replace("/\[table border=0\](.*?)\[\/table\]/sm", '<table border="0" >$1</table>', $text);
1412
1413                 $text = str_replace('[hr]', '<hr />', $text);
1414
1415                 // This is actually executed in Item::prepareBody()
1416
1417                 $nosmile = strpos($text, '[nosmile]') !== false;
1418                 $text = str_replace('[nosmile]', '', $text);
1419
1420                 // Check for font change text
1421                 $text = preg_replace("/\[font=(.*?)\](.*?)\[\/font\]/sm", "<span style=\"font-family: $1;\">$2</span>", $text);
1422
1423                 // Declare the format for [spoiler] layout
1424                 $SpoilerLayout = '<blockquote class="spoiler">$1</blockquote>';
1425
1426                 // Check for [spoiler] text
1427                 // handle nested quotes
1428                 $endlessloop = 0;
1429                 while ((strpos($text, "[/spoiler]") !== false) && (strpos($text, "[spoiler]") !== false) && (++$endlessloop < 20)) {
1430                         $text = preg_replace("/\[spoiler\](.*?)\[\/spoiler\]/ism", "$SpoilerLayout", $text);
1431                 }
1432
1433                 // Check for [spoiler=Author] text
1434
1435                 $t_wrote = L10n::t('$1 wrote:');
1436
1437                 // handle nested quotes
1438                 $endlessloop = 0;
1439                 while ((strpos($text, "[/spoiler]")!== false)  && (strpos($text, "[spoiler=") !== false) && (++$endlessloop < 20)) {
1440                         $text = preg_replace("/\[spoiler=[\"\']*(.*?)[\"\']*\](.*?)\[\/spoiler\]/ism",
1441                                 "<br /><strong class=".'"spoiler"'.">" . $t_wrote . "</strong><blockquote class=".'"spoiler"'.">$2</blockquote>",
1442                                 $text);
1443                 }
1444
1445                 // Declare the format for [quote] layout
1446                 $QuoteLayout = '<blockquote>$1</blockquote>';
1447
1448                 // Check for [quote] text
1449                 // handle nested quotes
1450                 $endlessloop = 0;
1451                 while ((strpos($text, "[/quote]") !== false) && (strpos($text, "[quote]") !== false) && (++$endlessloop < 20)) {
1452                         $text = preg_replace("/\[quote\](.*?)\[\/quote\]/ism", "$QuoteLayout", $text);
1453                 }
1454
1455                 // Check for [quote=Author] text
1456
1457                 $t_wrote = L10n::t('$1 wrote:');
1458
1459                 // handle nested quotes
1460                 $endlessloop = 0;
1461                 while ((strpos($text, "[/quote]")!== false)  && (strpos($text, "[quote=") !== false) && (++$endlessloop < 20)) {
1462                         $text = preg_replace("/\[quote=[\"\']*(.*?)[\"\']*\](.*?)\[\/quote\]/ism",
1463                                 "<p><strong class=".'"author"'.">" . $t_wrote . "</strong></p><blockquote>$2</blockquote>",
1464                                 $text);
1465                 }
1466
1467
1468                 // [img=widthxheight]image source[/img]
1469                 $text = preg_replace_callback(
1470                         "/\[img\=([0-9]*)x([0-9]*)\](.*?)\[\/img\]/ism",
1471                         function ($matches) use ($simple_html) {
1472                                 if (strpos($matches[3], "data:image/") === 0) {
1473                                         return $matches[0];
1474                                 }
1475
1476                                 $matches[3] = self::proxyUrl($matches[3], $simple_html);
1477                                 return "[img=" . $matches[1] . "x" . $matches[2] . "]" . $matches[3] . "[/img]";
1478                         },
1479                         $text
1480                 );
1481
1482                 $text = preg_replace("/\[img\=([0-9]*)x([0-9]*)\](.*?)\[\/img\]/ism", '<img src="$3" style="width: $1px;" >', $text);
1483                 $text = preg_replace("/\[zmg\=([0-9]*)x([0-9]*)\](.*?)\[\/zmg\]/ism", '<img class="zrl" src="$3" style="width: $1px;" >', $text);
1484
1485                 $text = preg_replace_callback("/\[img\=(.*?)\](.*?)\[\/img\]/ism",
1486                         function ($matches) use ($simple_html) {
1487                                 $matches[1] = self::proxyUrl($matches[1], $simple_html);
1488                                 $matches[2] = htmlspecialchars($matches[2], ENT_COMPAT);
1489                                 return '<img src="' . $matches[1] . '" alt="' . $matches[2] . '">';
1490                         },
1491                         $text);
1492
1493                 // Images
1494                 // [img]pathtoimage[/img]
1495                 $text = preg_replace_callback(
1496                         "/\[img\](.*?)\[\/img\]/ism",
1497                         function ($matches) use ($simple_html) {
1498                                 if (strpos($matches[1], "data:image/") === 0) {
1499                                         return $matches[0];
1500                                 }
1501
1502                                 $matches[1] = self::proxyUrl($matches[1], $simple_html);
1503                                 return "[img]" . $matches[1] . "[/img]";
1504                         },
1505                         $text
1506                 );
1507
1508                 $text = preg_replace("/\[img\](.*?)\[\/img\]/ism", '<img src="$1" alt="' . L10n::t('Image/photo') . '" />', $text);
1509                 $text = preg_replace("/\[zmg\](.*?)\[\/zmg\]/ism", '<img src="$1" alt="' . L10n::t('Image/photo') . '" />', $text);
1510
1511                 $text = preg_replace("/\[crypt\](.*?)\[\/crypt\]/ism", '<br/><img src="' .System::baseUrl() . '/images/lock_icon.gif" alt="' . L10n::t('Encrypted content') . '" title="' . L10n::t('Encrypted content') . '" /><br />', $text);
1512                 $text = preg_replace("/\[crypt(.*?)\](.*?)\[\/crypt\]/ism", '<br/><img src="' .System::baseUrl() . '/images/lock_icon.gif" alt="' . L10n::t('Encrypted content') . '" title="' . '$1' . ' ' . L10n::t('Encrypted content') . '" /><br />', $text);
1513                 //$Text = preg_replace("/\[crypt=(.*?)\](.*?)\[\/crypt\]/ism", '<br/><img src="' .System::baseUrl() . '/images/lock_icon.gif" alt="' . L10n::t('Encrypted content') . '" title="' . '$1' . ' ' . L10n::t('Encrypted content') . '" /><br />', $Text);
1514
1515                 // Try to Oembed
1516                 if ($try_oembed) {
1517                         $text = preg_replace("/\[video\](.*?\.(ogg|ogv|oga|ogm|webm|mp4).*?)\[\/video\]/ism", '<video src="$1" controls="controls" width="' . $a->videowidth . '" height="' . $a->videoheight . '" loop="true"><a href="$1">$1</a></video>', $text);
1518                         $text = preg_replace("/\[audio\](.*?\.(ogg|ogv|oga|ogm|webm|mp4|mp3).*?)\[\/audio\]/ism", '<audio src="$1" controls="controls"><a href="$1">$1</a></audio>', $text);
1519
1520                         $text = preg_replace_callback("/\[video\](.*?)\[\/video\]/ism", $try_oembed_callback, $text);
1521                         $text = preg_replace_callback("/\[audio\](.*?)\[\/audio\]/ism", $try_oembed_callback, $text);
1522                 } else {
1523                         $text = preg_replace("/\[video\](.*?)\[\/video\]/ism",
1524                                 '<a href="$1" target="_blank">$1</a>', $text);
1525                         $text = preg_replace("/\[audio\](.*?)\[\/audio\]/ism",
1526                                 '<a href="$1" target="_blank">$1</a>', $text);
1527                 }
1528
1529                 // html5 video and audio
1530
1531
1532                 if ($try_oembed) {
1533                         $text = preg_replace("/\[iframe\](.*?)\[\/iframe\]/ism", '<iframe src="$1" width="' . $a->videowidth . '" height="' . $a->videoheight . '"><a href="$1">$1</a></iframe>', $text);
1534                 } else {
1535                         $text = preg_replace("/\[iframe\](.*?)\[\/iframe\]/ism", '<a href="$1">$1</a>', $text);
1536                 }
1537
1538                 // Youtube extensions
1539                 if ($try_oembed) {
1540                         $text = preg_replace_callback("/\[youtube\](https?:\/\/www.youtube.com\/watch\?v\=.*?)\[\/youtube\]/ism", $try_oembed_callback, $text);
1541                         $text = preg_replace_callback("/\[youtube\](www.youtube.com\/watch\?v\=.*?)\[\/youtube\]/ism", $try_oembed_callback, $text);
1542                         $text = preg_replace_callback("/\[youtube\](https?:\/\/youtu.be\/.*?)\[\/youtube\]/ism", $try_oembed_callback, $text);
1543                 }
1544
1545                 $text = preg_replace("/\[youtube\]https?:\/\/www.youtube.com\/watch\?v\=(.*?)\[\/youtube\]/ism", '[youtube]$1[/youtube]', $text);
1546                 $text = preg_replace("/\[youtube\]https?:\/\/www.youtube.com\/embed\/(.*?)\[\/youtube\]/ism", '[youtube]$1[/youtube]', $text);
1547                 $text = preg_replace("/\[youtube\]https?:\/\/youtu.be\/(.*?)\[\/youtube\]/ism", '[youtube]$1[/youtube]', $text);
1548
1549                 if ($try_oembed) {
1550                         $text = preg_replace("/\[youtube\]([A-Za-z0-9\-_=]+)(.*?)\[\/youtube\]/ism", '<iframe width="' . $a->videowidth . '" height="' . $a->videoheight . '" src="https://www.youtube.com/embed/$1" frameborder="0" ></iframe>', $text);
1551                 } else {
1552                         $text = preg_replace("/\[youtube\]([A-Za-z0-9\-_=]+)(.*?)\[\/youtube\]/ism",
1553                                 '<a href="https://www.youtube.com/watch?v=$1" target="_blank">https://www.youtube.com/watch?v=$1</a>', $text);
1554                 }
1555
1556                 if ($try_oembed) {
1557                         $text = preg_replace_callback("/\[vimeo\](https?:\/\/player.vimeo.com\/video\/[0-9]+).*?\[\/vimeo\]/ism", $try_oembed_callback, $text);
1558                         $text = preg_replace_callback("/\[vimeo\](https?:\/\/vimeo.com\/[0-9]+).*?\[\/vimeo\]/ism", $try_oembed_callback, $text);
1559                 }
1560
1561                 $text = preg_replace("/\[vimeo\]https?:\/\/player.vimeo.com\/video\/([0-9]+)(.*?)\[\/vimeo\]/ism", '[vimeo]$1[/vimeo]', $text);
1562                 $text = preg_replace("/\[vimeo\]https?:\/\/vimeo.com\/([0-9]+)(.*?)\[\/vimeo\]/ism", '[vimeo]$1[/vimeo]', $text);
1563
1564                 if ($try_oembed) {
1565                         $text = preg_replace("/\[vimeo\]([0-9]+)(.*?)\[\/vimeo\]/ism", '<iframe width="' . $a->videowidth . '" height="' . $a->videoheight . '" src="https://player.vimeo.com/video/$1" frameborder="0" ></iframe>', $text);
1566                 } else {
1567                         $text = preg_replace("/\[vimeo\]([0-9]+)(.*?)\[\/vimeo\]/ism",
1568                                 '<a href="https://vimeo.com/$1" target="_blank">https://vimeo.com/$1</a>', $text);
1569                 }
1570
1571                 // oembed tag
1572                 $text = OEmbed::BBCode2HTML($text);
1573
1574                 // Avoid triple linefeeds through oembed
1575                 $text = str_replace("<br style='clear:left'></span><br /><br />", "<br style='clear:left'></span><br />", $text);
1576
1577                 // If we found an event earlier, strip out all the event code and replace with a reformatted version.
1578                 // Replace the event-start section with the entire formatted event. The other bbcode is stripped.
1579                 // Summary (e.g. title) is required, earlier revisions only required description (in addition to
1580                 // start which is always required). Allow desc with a missing summary for compatibility.
1581
1582                 if ((!empty($ev['desc']) || !empty($ev['summary'])) && !empty($ev['start'])) {
1583                         $sub = Event::getHTML($ev, $simple_html);
1584
1585                         $text = preg_replace("/\[event\-summary\](.*?)\[\/event\-summary\]/ism", '', $text);
1586                         $text = preg_replace("/\[event\-description\](.*?)\[\/event\-description\]/ism", '', $text);
1587                         $text = preg_replace("/\[event\-start\](.*?)\[\/event\-start\]/ism", $sub, $text);
1588                         $text = preg_replace("/\[event\-finish\](.*?)\[\/event\-finish\]/ism", '', $text);
1589                         $text = preg_replace("/\[event\-location\](.*?)\[\/event\-location\]/ism", '', $text);
1590                         $text = preg_replace("/\[event\-adjust\](.*?)\[\/event\-adjust\]/ism", '', $text);
1591                         $text = preg_replace("/\[event\-id\](.*?)\[\/event\-id\]/ism", '', $text);
1592                 }
1593
1594                 // Replace non graphical smilies for external posts
1595                 if (!$nosmile && !$for_plaintext) {
1596                         $text = Smilies::replace($text);
1597                 }
1598
1599                 // if the HTML is used to generate plain text, then don't do this search, but replace all URL of that kind to text
1600                 if (!$for_plaintext) {
1601                         $text = preg_replace(Strings::autoLinkRegEx(), '[url]$1[/url]', $text);
1602                         if (in_array($simple_html, [7, 9])) {
1603                                 $text = preg_replace_callback("/\[url\](.*?)\[\/url\]/ism", 'self::convertUrlForOStatusCallback', $text);
1604                                 $text = preg_replace_callback("/\[url\=(.*?)\](.*?)\[\/url\]/ism", 'self::convertUrlForOStatusCallback', $text);
1605                         }
1606                 } else {
1607                         $text = preg_replace("(\[url\](.*?)\[\/url\])ism", " $1 ", $text);
1608                         $text = preg_replace_callback("&\[url=([^\[\]]*)\]\[img\](.*)\[\/img\]\[\/url\]&Usi", 'self::removePictureLinksCallback', $text);
1609                 }
1610
1611                 $text = str_replace(["\r","\n"], ['<br />', '<br />'], $text);
1612
1613                 // Remove all hashtag addresses
1614                 if ((!$try_oembed || $simple_html) && !in_array($simple_html, [3, 7, 9])) {
1615                         $text = preg_replace("/([#@!])\[url\=(.*?)\](.*?)\[\/url\]/ism", '$1$3', $text);
1616                 } elseif ($simple_html == 3) {
1617                         // The ! is converted to @ since Diaspora only understands the @
1618                         $text = preg_replace("/([@!])\[url\=(.*?)\](.*?)\[\/url\]/ism",
1619                                 '@<a href="$2">$3</a>',
1620                                 $text);
1621                 } elseif (in_array($simple_html, [7, 9])) {
1622                         $text = preg_replace("/([@!])\[url\=(.*?)\](.*?)\[\/url\]/ism",
1623                                 '$1<span class="vcard"><a href="$2" class="url u-url mention" title="$3"><span class="fn nickname mention">$3</span></a></span>',
1624                                 $text);
1625                 } elseif (!$simple_html) {
1626                         $text = preg_replace("/([@!])\[url\=(.*?)\](.*?)\[\/url\]/ism",
1627                                 '$1<a href="$2" class="userinfo mention" title="$3">$3</a>',
1628                                 $text);
1629                 }
1630
1631                 // Bookmarks in red - will be converted to bookmarks in friendica
1632                 $text = preg_replace("/#\^\[url\](.*?)\[\/url\]/ism", '[bookmark=$1]$1[/bookmark]', $text);
1633                 $text = preg_replace("/#\^\[url\=(.*?)\](.*?)\[\/url\]/ism", '[bookmark=$1]$2[/bookmark]', $text);
1634                 $text = preg_replace("/#\[url\=.*?\]\^\[\/url\]\[url\=(.*?)\](.*?)\[\/url\]/i",
1635                                         "[bookmark=$1]$2[/bookmark]", $text);
1636
1637                 if (in_array($simple_html, [2, 6, 7, 8])) {
1638                         $text = preg_replace_callback("/([^#@!])\[url\=([^\]]*)\](.*?)\[\/url\]/ism", "self::expandLinksCallback", $text);
1639                         //$Text = preg_replace("/[^#@!]\[url\=([^\]]*)\](.*?)\[\/url\]/ism", ' $2 [url]$1[/url]', $Text);
1640                         $text = preg_replace("/\[bookmark\=([^\]]*)\](.*?)\[\/bookmark\]/ism", ' $2 [url]$1[/url]',$text);
1641                 }
1642
1643                 if ($simple_html == 5) {
1644                         $text = preg_replace("/[^#@!]\[url\=(.*?)\](.*?)\[\/url\]/ism", '[url]$1[/url]', $text);
1645                 }
1646
1647                 // Perform URL Search
1648                 if ($try_oembed) {
1649                         $text = preg_replace_callback("/\[bookmark\=([^\]]*)\](.*?)\[\/bookmark\]/ism", $try_oembed_callback, $text);
1650                 }
1651
1652                 if ($simple_html == 5) {
1653                         $text = preg_replace("/\[bookmark\=([^\]]*)\](.*?)\[\/bookmark\]/ism", '[url]$1[/url]', $text);
1654                 } else {
1655                         $text = preg_replace("/\[bookmark\=([^\]]*)\](.*?)\[\/bookmark\]/ism", '[url=$1]$2[/url]', $text);
1656                 }
1657
1658                 // Handle Diaspora posts
1659                 $text = preg_replace_callback(
1660                         "&\[url=/?posts/([^\[\]]*)\](.*)\[\/url\]&Usi",
1661                         function ($match) {
1662                                 return "[url=" . System::baseUrl() . "/display/" . $match[1] . "]" . $match[2] . "[/url]";
1663                         }, $text
1664                 );
1665
1666                 $text = preg_replace_callback(
1667                         "&\[url=/people\?q\=(.*)\](.*)\[\/url\]&Usi",
1668                         function ($match) {
1669                                 return "[url=" . System::baseUrl() . "/search?search=%40" . $match[1] . "]" . $match[2] . "[/url]";
1670                         }, $text
1671                 );
1672
1673                 // Server independent link to posts and comments
1674                 // See issue: https://github.com/diaspora/diaspora_federation/issues/75
1675                 $expression = "=diaspora://.*?/post/([0-9A-Za-z\-_@.:]{15,254}[0-9A-Za-z])=ism";
1676                 $text = preg_replace($expression, System::baseUrl()."/display/$1", $text);
1677
1678                 /* Tag conversion
1679                  * Supports:
1680                  * - #[url=<anything>]<term>[/url]
1681                  * - [url=<anything>]#<term>[/url]
1682                  */
1683                 $text = preg_replace_callback("/(?:#\[url\=.*?\]|\[url\=.*?\]#)(.*?)\[\/url\]/ism", function($matches) {
1684                         return '#<a href="'
1685                                 . System::baseUrl()     . '/search?tag=' . rawurlencode($matches[1])
1686                                 . '" class="tag" title="' . XML::escape($matches[1]) . '">'
1687                                 . XML::escape($matches[1])
1688                                 . '</a>';
1689                 }, $text);
1690
1691                 // We need no target="_blank" for local links
1692                 // convert links start with System::baseUrl() as local link without the target="_blank" attribute
1693                 $escapedBaseUrl = preg_quote(System::baseUrl(), '/');
1694                 $text = preg_replace("/\[url\](".$escapedBaseUrl.".*?)\[\/url\]/ism", '<a href="$1">$1</a>', $text);
1695                 $text = preg_replace("/\[url\=(".$escapedBaseUrl.".*?)\](.*?)\[\/url\]/ism", '<a href="$1">$2</a>', $text);
1696
1697                 $text = preg_replace("/\[url\](.*?)\[\/url\]/ism", '<a href="$1" target="_blank">$1</a>', $text);
1698                 $text = preg_replace("/\[url\=(.*?)\](.*?)\[\/url\]/ism", '<a href="$1" target="_blank">$2</a>', $text);
1699
1700                 // Red compatibility, though the link can't be authenticated on Friendica
1701                 $text = preg_replace("/\[zrl\=(.*?)\](.*?)\[\/zrl\]/ism", '<a href="$1" target="_blank">$2</a>', $text);
1702
1703
1704                 // we may need to restrict this further if it picks up too many strays
1705                 // link acct:user@host to a webfinger profile redirector
1706
1707                 $text = preg_replace('/acct:([^@]+)@((?!\-)(?:[a-zA-Z\d\-]{0,62}[a-zA-Z\d]\.){1,126}(?!\d+)[a-zA-Z\d]{1,63})/', '<a href="' . System::baseUrl() . '/acctlink?addr=$1@$2" target="extlink">acct:$1@$2</a>', $text);
1708
1709                 // Perform MAIL Search
1710                 $text = preg_replace("/\[mail\](.*?)\[\/mail\]/", '<a href="mailto:$1">$1</a>', $text);
1711                 $text = preg_replace("/\[mail\=(.*?)\](.*?)\[\/mail\]/", '<a href="mailto:$1">$2</a>', $text);
1712
1713                 // Unhide all [noparse] contained bbtags unspacefying them
1714                 // and triming the [noparse] tag.
1715
1716                 $text = preg_replace_callback("/\[noparse\](.*?)\[\/noparse\]/ism", 'self::unescapeNoparseCallback', $text);
1717                 $text = preg_replace_callback("/\[nobb\](.*?)\[\/nobb\]/ism", 'self::unescapeNoparseCallback', $text);
1718                 $text = preg_replace_callback("/\[pre\](.*?)\[\/pre\]/ism", 'self::unescapeNoparseCallback', $text);
1719
1720                 /// @todo What is the meaning of these lines?
1721                 $text = preg_replace('/\[\&amp\;([#a-z0-9]+)\;\]/', '&$1;', $text);
1722                 $text = preg_replace('/\&\#039\;/', '\'', $text);
1723
1724                 // Currently deactivated, it made problems with " inside of alt texts.
1725                 //$text = preg_replace('/\&quot\;/', '"', $text);
1726
1727                 // fix any escaped ampersands that may have been converted into links
1728                 $text = preg_replace('/\<([^>]*?)(src|href)=(.*?)\&amp\;(.*?)\>/ism', '<$1$2=$3&$4>', $text);
1729
1730                 // sanitizes src attributes (http and redir URLs for displaying in a web page, cid used for inline images in emails)
1731                 $allowed_src_protocols = ['http', 'redir', 'cid'];
1732                 $text = preg_replace('#<([^>]*?)(src)="(?!' . implode('|', $allowed_src_protocols) . ')(.*?)"(.*?)>#ism',
1733                                          '<$1$2=""$4 data-original-src="$3" class="invalid-src" title="' . L10n::t('Invalid source protocol') . '">', $text);
1734
1735                 // sanitize href attributes (only whitelisted protocols URLs)
1736                 // default value for backward compatibility
1737                 $allowed_link_protocols = Config::get('system', 'allowed_link_protocols', ['ftp', 'mailto', 'gopher', 'cid']);
1738
1739                 // Always allowed protocol even if config isn't set or not including it
1740                 $allowed_link_protocols[] = 'http';
1741                 $allowed_link_protocols[] = 'redir/';
1742
1743                 $regex = '#<([^>]*?)(href)="(?!' . implode('|', $allowed_link_protocols) . ')(.*?)"(.*?)>#ism';
1744                 $text = preg_replace($regex, '<$1$2="javascript:void(0)"$4 data-original-href="$3" class="invalid-href" title="' . L10n::t('Invalid link protocol') . '">', $text);
1745
1746                 // Shared content
1747                 $text = self::convertShare(
1748                         $text,
1749                         function (array $attributes, array $author_contact, $content, $is_quote_share) use ($simple_html) {
1750                                 return self::convertShareCallback($attributes, $author_contact, $content, $is_quote_share, $simple_html);
1751                         }
1752                 );
1753
1754                 if ($saved_image) {
1755                         $text = self::interpolateSavedImagesIntoItemBody($text, $saved_image);
1756                 }
1757
1758                 // Restore code blocks
1759                 $text = preg_replace_callback('/#codeblock-([0-9]+)#/iU',
1760                         function ($matches) use ($codeblocks) {
1761                                 $return = $matches[0];
1762                                 if (isset($codeblocks[intval($matches[1])])) {
1763                                         $return = $codeblocks[$matches[1]];
1764                                 }
1765                                 return $return;
1766                         },
1767                         $text
1768                 );
1769
1770                 // Clean up the HTML by loading and saving the HTML with the DOM.
1771                 // Bad structured html can break a whole page.
1772                 // For performance reasons do it only with activated item cache or at export.
1773                 if (!$try_oembed || (get_itemcachepath() != "")) {
1774                         $doc = new DOMDocument();
1775                         $doc->preserveWhiteSpace = false;
1776
1777                         $text = mb_convert_encoding($text, 'HTML-ENTITIES', "UTF-8");
1778
1779                         $doctype = '<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN" "http://www.w3.org/TR/REC-html40/loose.dtd">';
1780                         $encoding = '<?xml encoding="UTF-8">';
1781                         @$doc->loadHTML($encoding.$doctype."<html><body>".$text."</body></html>");
1782                         $doc->encoding = 'UTF-8';
1783                         $text = $doc->saveHTML();
1784                         $text = str_replace(["<html><body>", "</body></html>", $doctype, $encoding], ["", "", "", ""], $text);
1785
1786                         $text = str_replace('<br></li>', '</li>', $text);
1787
1788                         //$Text = mb_convert_encoding($Text, "UTF-8", 'HTML-ENTITIES');
1789                 }
1790
1791                 // Clean up some useless linebreaks in lists
1792                 //$Text = str_replace('<br /><ul', '<ul ', $Text);
1793                 //$Text = str_replace('</ul><br />', '</ul>', $Text);
1794                 //$Text = str_replace('</li><br />', '</li>', $Text);
1795                 //$Text = str_replace('<br /><li>', '<li>', $Text);
1796                 //$Text = str_replace('<br /><ul', '<ul ', $Text);
1797
1798                 Hook::callAll('bbcode', $text);
1799
1800                 return trim($text);
1801         }
1802
1803         /**
1804          * @brief Strips the "abstract" tag from the provided text
1805          *
1806          * @param string $text The text with BBCode
1807          * @return string The same text - but without "abstract" element
1808          */
1809         public static function stripAbstract($text)
1810         {
1811                 $text = preg_replace("/[\s|\n]*\[abstract\].*?\[\/abstract\][\s|\n]*/ism", '', $text);
1812                 $text = preg_replace("/[\s|\n]*\[abstract=.*?\].*?\[\/abstract][\s|\n]*/ism", '', $text);
1813
1814                 return $text;
1815         }
1816
1817         /**
1818          * @brief Returns the value of the "abstract" element
1819          *
1820          * @param string $text The text that maybe contains the element
1821          * @param string $addon The addon for which the abstract is meant for
1822          * @return string The abstract
1823          */
1824         public static function getAbstract($text, $addon = "")
1825         {
1826                 $abstract = "";
1827                 $abstracts = [];
1828                 $addon = strtolower($addon);
1829
1830                 if (preg_match_all("/\[abstract=(.*?)\](.*?)\[\/abstract\]/ism", $text, $results, PREG_SET_ORDER)) {
1831                         foreach ($results AS $result) {
1832                                 $abstracts[strtolower($result[1])] = $result[2];
1833                         }
1834                 }
1835
1836                 if (isset($abstracts[$addon])) {
1837                         $abstract = $abstracts[$addon];
1838                 }
1839
1840                 if ($abstract == "" && preg_match("/\[abstract\](.*?)\[\/abstract\]/ism", $text, $result)) {
1841                         $abstract = $result[1];
1842                 }
1843
1844                 return $abstract;
1845         }
1846
1847         /**
1848          * @brief Callback function to replace a Friendica style mention in a mention for Diaspora
1849          *
1850          * @param array $match Matching values for the callback
1851          *                     [1] = Mention type (! or @)
1852          *                     [2] = Name
1853          *                     [3] = Address
1854          * @return string Replaced mention
1855          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
1856          * @throws \ImagickException
1857          */
1858         private static function bbCodeMention2DiasporaCallback($match)
1859         {
1860                 $contact = Contact::getDetailsByURL($match[3]);
1861
1862                 if (empty($contact['addr'])) {
1863                         $contact = Probe::uri($match[3]);
1864                 }
1865
1866                 if (empty($contact['addr'])) {
1867                         return $match[0];
1868                 }
1869
1870                 $mention = $match[1] . '{' . $match[2] . '; ' . $contact['addr'] . '}';
1871                 return $mention;
1872         }
1873
1874         /**
1875          * @brief Converts a BBCode text into Markdown
1876          *
1877          * This function converts a BBCode item body to be sent to Markdown-enabled
1878          * systems like Diaspora and Libertree
1879          *
1880          * @param string $text
1881          * @param bool   $for_diaspora Diaspora requires more changes than Libertree
1882          * @return string
1883          * @throws \Friendica\Network\HTTPException\InternalServerErrorException
1884          */
1885         public static function toMarkdown($text, $for_diaspora = true)
1886         {
1887                 $a = self::getApp();
1888
1889                 $original_text = $text;
1890
1891                 // Since Diaspora is creating a summary for links, this function removes them before posting
1892                 if ($for_diaspora) {
1893                         $text = self::removeShareInformation($text);
1894                 }
1895
1896                 /**
1897                  * Transform #tags, strip off the [url] and replace spaces with underscore
1898                  */
1899                 $url_search_string = "^\[\]";
1900                 $text = preg_replace_callback("/#\[url\=([$url_search_string]*)\](.*?)\[\/url\]/i",
1901                         function ($matches) {
1902                                 return '#' . str_replace(' ', '_', $matches[2]);
1903                         },
1904                         $text
1905                 );
1906
1907                 // Converting images with size parameters to simple images. Markdown doesn't know it.
1908                 $text = preg_replace("/\[img\=([0-9]*)x([0-9]*)\](.*?)\[\/img\]/ism", '[img]$3[/img]', $text);
1909
1910                 // Convert it to HTML - don't try oembed
1911                 if ($for_diaspora) {
1912                         $text = self::convert($text, false, 3);
1913
1914                         // Add all tags that maybe were removed
1915                         if (preg_match_all("/#\[url\=([$url_search_string]*)\](.*?)\[\/url\]/ism", $original_text, $tags)) {
1916                                 $tagline = "";
1917                                 foreach ($tags[2] as $tag) {
1918                                         $tag = html_entity_decode($tag, ENT_QUOTES, 'UTF-8');
1919                                         if (!strpos(html_entity_decode($text, ENT_QUOTES, 'UTF-8'), '#' . $tag)) {
1920                                                 $tagline .= '#' . $tag . ' ';
1921                                         }
1922                                 }
1923                                 $text = $text . " " . $tagline;
1924                         }
1925                 } else {
1926                         $text = self::convert($text, false, 4);
1927                 }
1928
1929                 // mask some special HTML chars from conversation to markdown
1930                 $text = str_replace(['&lt;', '&gt;', '&amp;'], ['&_lt_;', '&_gt_;', '&_amp_;'], $text);
1931
1932                 // If a link is followed by a quote then there should be a newline before it
1933                 // Maybe we should make this newline at every time before a quote.
1934                 $text = str_replace(["</a><blockquote>"], ["</a><br><blockquote>"], $text);
1935
1936                 $stamp1 = microtime(true);
1937
1938                 // Now convert HTML to Markdown
1939                 $text = HTML::toMarkdown($text);
1940
1941                 // unmask the special chars back to HTML
1942                 $text = str_replace(['&\_lt\_;', '&\_gt\_;', '&\_amp\_;'], ['&lt;', '&gt;', '&amp;'], $text);
1943
1944                 $a->getProfiler()->saveTimestamp($stamp1, "parser", System::callstack());
1945
1946                 // Libertree has a problem with escaped hashtags.
1947                 $text = str_replace(['\#'], ['#'], $text);
1948
1949                 // Remove any leading or trailing whitespace, as this will mess up
1950                 // the Diaspora signature verification and cause the item to disappear
1951                 $text = trim($text);
1952
1953                 if ($for_diaspora) {
1954                         $url_search_string = "^\[\]";
1955                         $text = preg_replace_callback(
1956                                 "/([@!])\[(.*?)\]\(([$url_search_string]*?)\)/ism",
1957                                 ['self', 'bbCodeMention2DiasporaCallback'],
1958                                 $text
1959                         );
1960                 }
1961
1962                 Hook::callAll('bb2diaspora', $text);
1963
1964                 return $text;
1965         }
1966
1967         /**
1968      * @brief Pull out all #hashtags and @person tags from $string.
1969      *
1970      * We also get @person@domain.com - which would make
1971      * the regex quite complicated as tags can also
1972      * end a sentence. So we'll run through our results
1973      * and strip the period from any tags which end with one.
1974      * Returns array of tags found, or empty array.
1975      *
1976      * @param string $string Post content
1977      *
1978      * @return array List of tag and person names
1979      */
1980     public static function getTags($string)
1981     {
1982         $ret = [];
1983
1984         // Convert hashtag links to hashtags
1985         $string = preg_replace('/#\[url\=([^\[\]]*)\](.*?)\[\/url\]/ism', '#$2', $string);
1986
1987         // ignore anything in a code block
1988         $string = preg_replace('/\[code.*?\].*?\[\/code\]/sm', '', $string);
1989
1990         // Force line feeds at bbtags
1991         $string = str_replace(['[', ']'], ["\n[", "]\n"], $string);
1992
1993         // ignore anything in a bbtag
1994         $string = preg_replace('/\[(.*?)\]/sm', '', $string);
1995
1996         // Match full names against @tags including the space between first and last
1997         // We will look these up afterward to see if they are full names or not recognisable.
1998
1999         if (preg_match_all('/(@[^ \x0D\x0A,:?]+ [^ \x0D\x0A@,:?]+)([ \x0D\x0A@,:?]|$)/', $string, $matches)) {
2000             foreach ($matches[1] as $match) {
2001                 if (strstr($match, ']')) {
2002                     // we might be inside a bbcode color tag - leave it alone
2003                     continue;
2004                 }
2005
2006                 if (substr($match, -1, 1) === '.') {
2007                     $ret[] = substr($match, 0, -1);
2008                 } else {
2009                     $ret[] = $match;
2010                 }
2011             }
2012         }
2013
2014         // Otherwise pull out single word tags. These can be @nickname, @first_last
2015         // and #hash tags.
2016
2017         if (preg_match_all('/([!#@][^\^ \x0D\x0A,;:?]+)([ \x0D\x0A,;:?]|$)/', $string, $matches)) {
2018             foreach ($matches[1] as $match) {
2019                 if (strstr($match, ']')) {
2020                     // we might be inside a bbcode color tag - leave it alone
2021                     continue;
2022                 }
2023                 if (substr($match, -1, 1) === '.') {
2024                     $match = substr($match,0,-1);
2025                 }
2026                 // ignore strictly numeric tags like #1
2027                 if ((strpos($match, '#') === 0) && ctype_digit(substr($match, 1))) {
2028                     continue;
2029                 }
2030                 // try not to catch url fragments
2031                 if (strpos($string, $match) && preg_match('/[a-zA-z0-9\/]/', substr($string, strpos($string, $match) - 1, 1))) {
2032                     continue;
2033                 }
2034                 $ret[] = $match;
2035             }
2036         }
2037
2038         return $ret;
2039     }
2040 }