', $data['url'], $data['title']);
}
}
if (!empty($data["description"]) && $data["description"] != $data["title"]) {
// Sanitize the HTML by converting it to BBCode
$bbcode = HTML::toBBCode($data["description"]);
$return .= sprintf('
';
}
break;
default:
// Transforms quoted tweets in rich attachments to avoid nested tweets
if (stripos(Strings::normaliseLink($attributes['link']), 'http://twitter.com/') === 0 && OEmbed::isAllowedURL($attributes['link'])) {
try {
$text = ($is_quote_share? ' ' : '') . OEmbed::getHTML($attributes['link']);
} catch (Exception $e) {
$text = ($is_quote_share? ' ' : '') . sprintf('[bookmark=%s]%s[/bookmark]', $attributes['link'], $content);
}
} else {
$text = ($is_quote_share? "\n" : '');
$tpl = Renderer::getMarkupTemplate('shared_content.tpl');
$text .= Renderer::replaceMacros($tpl, [
'$profile' => $attributes['profile'],
'$avatar' => $attributes['avatar'],
'$author' => $attributes['author'],
'$link' => $attributes['link'],
'$posted' => $attributes['posted'],
'$content' => trim($content)
]);
}
break;
}
return $text;
}
private static function removePictureLinksCallback($match)
{
$text = Cache::get($match[1]);
if (is_null($text)) {
$a = self::getApp();
$stamp1 = microtime(true);
$ch = @curl_init($match[1]);
@curl_setopt($ch, CURLOPT_NOBODY, true);
@curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
@curl_setopt($ch, CURLOPT_USERAGENT, $a->getUserAgent());
@curl_exec($ch);
$curl_info = @curl_getinfo($ch);
$a->saveTimestamp($stamp1, "network");
if (substr($curl_info["content_type"], 0, 6) == "image/") {
$text = "[url=" . $match[1] . "]" . $match[1] . "[/url]";
} else {
$text = "[url=" . $match[2] . "]" . $match[2] . "[/url]";
// if its not a picture then look if its a page that contains a picture link
$body = Network::fetchUrl($match[1]);
$doc = new DOMDocument();
@$doc->loadHTML($body);
$xpath = new DOMXPath($doc);
$list = $xpath->query("//meta[@name]");
foreach ($list as $node) {
$attr = [];
if ($node->attributes->length) {
foreach ($node->attributes as $attribute) {
$attr[$attribute->name] = $attribute->value;
}
}
if (strtolower($attr["name"]) == "twitter:image") {
$text = "[url=" . $attr["content"] . "]" . $attr["content"] . "[/url]";
}
}
}
Cache::set($match[1], $text);
}
return $text;
}
private static function expandLinksCallback($match)
{
if (($match[3] == "") || ($match[2] == $match[3]) || stristr($match[2], $match[3])) {
return ($match[1] . "[url]" . $match[2] . "[/url]");
} else {
return ($match[1] . $match[3] . " [url]" . $match[2] . "[/url]");
}
}
private static function cleanPictureLinksCallback($match)
{
$text = Cache::get($match[1]);
if (is_null($text)) {
$a = self::getApp();
$stamp1 = microtime(true);
$ch = @curl_init($match[1]);
@curl_setopt($ch, CURLOPT_NOBODY, true);
@curl_setopt($ch, CURLOPT_RETURNTRANSFER, true);
@curl_setopt($ch, CURLOPT_USERAGENT, $a->getUserAgent());
@curl_exec($ch);
$curl_info = @curl_getinfo($ch);
$a->saveTimestamp($stamp1, "network");
// if its a link to a picture then embed this picture
if (substr($curl_info["content_type"], 0, 6) == "image/") {
$text = "[img]" . $match[1] . "[/img]";
} else {
$text = "[img]" . $match[2] . "[/img]";
// if its not a picture then look if its a page that contains a picture link
$body = Network::fetchUrl($match[1]);
$doc = new DOMDocument();
@$doc->loadHTML($body);
$xpath = new DOMXPath($doc);
$list = $xpath->query("//meta[@name]");
foreach ($list as $node) {
$attr = [];
if ($node->attributes->length) {
foreach ($node->attributes as $attribute) {
$attr[$attribute->name] = $attribute->value;
}
}
if (strtolower($attr["name"]) == "twitter:image") {
$text = "[img]" . $attr["content"] . "[/img]";
}
}
}
Cache::set($match[1], $text);
}
return $text;
}
public static function cleanPictureLinks($text)
{
$return = preg_replace_callback("&\[url=([^\[\]]*)\]\[img\](.*)\[\/img\]\[\/url\]&Usi", 'self::cleanPictureLinksCallback', $text);
return $return;
}
/**
* @brief Converts a BBCode message to HTML message
*
* BBcode 2 HTML was written by WAY2WEB.net
* extended to work with Mistpark/Friendica - Mike Macgirvin
*
* Simple HTML values meaning:
* - 0: Friendica display
* - 1: Unused
* - 2: Used for Google+, Windows Phone push, Friendica API
* - 3: Used before converting to Markdown in bb2diaspora.php
* - 4: Used for WordPress, Libertree (before Markdown), pump.io and tumblr
* - 5: Unused
* - 6: Unused
* - 7: Used for dfrn, OStatus
* - 8: Used for WP backlink text setting
*
* @param string $text
* @param bool $try_oembed
* @param int $simple_html
* @param bool $for_plaintext
* @return string
*/
public static function convert($text, $try_oembed = true, $simple_html = false, $for_plaintext = false)
{
$a = self::getApp();
/*
* preg_match_callback function to replace potential Oembed tags with Oembed content
*
* $match[0] = [tag]$url[/tag] or [tag=$url]$title[/tag]
* $match[1] = $url
* $match[2] = $title or absent
*/
$try_oembed_callback = function ($match)
{
$url = $match[1];
$title = defaults($match, 2, null);
try {
$return = OEmbed::getHTML($url, $title);
} catch (Exception $ex) {
$return = $match[0];
}
return $return;
};
// Extracting multi-line code blocks before the whitespace processing
$codeblocks = [];
$text = preg_replace_callback("#\[code(?:=([^\]]*))?\](.*?)\[\/code\]#is",
function ($matches) use (&$codeblocks) {
$return = $matches[0];
if (strpos($matches[2], "\n") !== false) {
$return = '#codeblock-' . count($codeblocks) . '#';
$codeblocks[] = '
' . trim($matches[2], "\n\r") . '
';
}
return $return;
},
$text
);
// Hide all [noparse] contained bbtags by spacefying them
// POSSIBLE BUG --> Will the 'preg' functions crash if there's an embedded image?
$text = preg_replace_callback("/\[noparse\](.*?)\[\/noparse\]/ism", 'self::escapeNoparseCallback', $text);
$text = preg_replace_callback("/\[nobb\](.*?)\[\/nobb\]/ism", 'self::escapeNoparseCallback', $text);
$text = preg_replace_callback("/\[pre\](.*?)\[\/pre\]/ism", 'self::escapeNoparseCallback', $text);
// Remove the abstract element. It is a non visible element.
$text = self::stripAbstract($text);
// Move all spaces out of the tags
$text = preg_replace("/\[(\w*)\](\s*)/ism", '$2[$1]', $text);
$text = preg_replace("/(\s*)\[\/(\w*)\]/ism", '[/$2]$1', $text);
// Extract the private images which use data urls since preg has issues with
// large data sizes. Stash them away while we do bbcode conversion, and then put them back
// in after we've done all the regex matching. We cannot use any preg functions to do this.
$extracted = self::extractImagesFromItemBody($text);
$text = $extracted['body'];
$saved_image = $extracted['images'];
// If we find any event code, turn it into an event.
// After we're finished processing the bbcode we'll
// replace all of the event code with a reformatted version.
$ev = Event::fromBBCode($text);
// Replace any html brackets with HTML Entities to prevent executing HTML or script
// Don't use strip_tags here because it breaks [url] search by replacing & with amp
$text = str_replace("<", "<", $text);
$text = str_replace(">", ">", $text);
// remove some newlines before the general conversion
$text = preg_replace("/\s?\[share(.*?)\]\s?(.*?)\s?\[\/share\]\s?/ism", "[share$1]$2[/share]", $text);
$text = preg_replace("/\s?\[quote(.*?)\]\s?(.*?)\s?\[\/quote\]\s?/ism", "[quote$1]$2[/quote]", $text);
// when the content is meant exporting to other systems then remove the avatar picture since this doesn't really look good on these systems
if (!$try_oembed) {
$text = preg_replace("/\[share(.*?)avatar\s?=\s?'.*?'\s?(.*?)\]\s?(.*?)\s?\[\/share\]\s?/ism", "\n[share$1$2]$3[/share]", $text);
}
// Convert new line chars to html tags
// nlbr seems to be hopelessly messed up
// $Text = nl2br($Text);
// We'll emulate it.
$text = trim($text);
$text = str_replace("\r\n", "\n", $text);
// removing multiplicated newlines
if (Config::get("system", "remove_multiplicated_lines")) {
$search = ["\n\n\n", "\n ", " \n", "[/quote]\n\n", "\n[/quote]", "[/li]\n", "\n[li]", "\n[ul]", "[/ul]\n", "\n\n[share ", "[/attachment]\n",
"\n[h1]", "[/h1]\n", "\n[h2]", "[/h2]\n", "\n[h3]", "[/h3]\n", "\n[h4]", "[/h4]\n", "\n[h5]", "[/h5]\n", "\n[h6]", "[/h6]\n"];
$replace = ["\n\n", "\n", "\n", "[/quote]\n", "[/quote]", "[/li]", "[li]", "[ul]", "[/ul]", "\n[share ", "[/attachment]",
"[h1]", "[/h1]", "[h2]", "[/h2]", "[h3]", "[/h3]", "[h4]", "[/h4]", "[h5]", "[/h5]", "[h6]", "[/h6]"];
do {
$oldtext = $text;
$text = str_replace($search, $replace, $text);
} while ($oldtext != $text);
}
// Set up the parameters for a URL search string
$URLSearchString = "^\[\]";
// Set up the parameters for a MAIL search string
$MAILSearchString = $URLSearchString;
// if the HTML is used to generate plain text, then don't do this search, but replace all URL of that kind to text
if (!$for_plaintext) {
// Autolink feature (thanks to http://code.seebz.net/p/autolink-php/)
// Currently disabled, since the function is too greedy
// $autolink_regex = "`([^\]\=\"']|^)(https?\://[^\s<]+[^\s<\.\)])`ism";
$autolink_regex = "/([^\]\='".'"'."]|^)(https?\:\/\/[a-zA-Z0-9\:\/\-\?\&\;\.\=\_\~\#\%\$\!\+\,]+)/ism";
$text = preg_replace($autolink_regex, '$1[url]$2[/url]', $text);
if ($simple_html == 7) {
$text = preg_replace_callback("/\[url\]([$URLSearchString]*)\[\/url\]/ism", 'self::convertUrlForOStatusCallback', $text);
$text = preg_replace_callback("/\[url\=([$URLSearchString]*)\]([$URLSearchString]*)\[\/url\]/ism", 'self::convertUrlForOStatusCallback', $text);
}
} else {
$text = preg_replace("(\[url\]([$URLSearchString]*)\[\/url\])ism", " $1 ", $text);
$text = preg_replace_callback("&\[url=([^\[\]]*)\]\[img\](.*)\[\/img\]\[\/url\]&Usi", 'self::removePictureLinksCallback', $text);
}
// Handle attached links or videos
$text = self::convertAttachment($text, $simple_html, $try_oembed);
$text = str_replace(["\r","\n"], [' ', ' '], $text);
// Remove all hashtag addresses
if ((!$try_oembed || $simple_html) && !in_array($simple_html, [3, 7])) {
$text = preg_replace("/([#@!])\[url\=([$URLSearchString]*)\](.*?)\[\/url\]/ism", '$1$3', $text);
} elseif ($simple_html == 3) {
// The ! is converted to @ since Diaspora only understands the @
$text = preg_replace("/([@!])\[url\=([$URLSearchString]*)\](.*?)\[\/url\]/ism",
'@$3',
$text);
} elseif ($simple_html == 7) {
$text = preg_replace("/([@!])\[url\=([$URLSearchString]*)\](.*?)\[\/url\]/ism",
'$1$3',
$text);
} elseif (!$simple_html) {
$text = preg_replace("/([@!])\[url\=([$URLSearchString]*)\](.*?)\[\/url\]/ism",
'$1$3',
$text);
}
// Bookmarks in red - will be converted to bookmarks in friendica
$text = preg_replace("/#\^\[url\]([$URLSearchString]*)\[\/url\]/ism", '[bookmark=$1]$1[/bookmark]', $text);
$text = preg_replace("/#\^\[url\=([$URLSearchString]*)\](.*?)\[\/url\]/ism", '[bookmark=$1]$2[/bookmark]', $text);
$text = preg_replace("/#\[url\=[$URLSearchString]*\]\^\[\/url\]\[url\=([$URLSearchString]*)\](.*?)\[\/url\]/i",
"[bookmark=$1]$2[/bookmark]", $text);
if (in_array($simple_html, [2, 6, 7, 8, 9])) {
$text = preg_replace_callback("/([^#@!])\[url\=([^\]]*)\](.*?)\[\/url\]/ism", "self::expandLinksCallback", $text);
//$Text = preg_replace("/[^#@!]\[url\=([^\]]*)\](.*?)\[\/url\]/ism", ' $2 [url]$1[/url]', $Text);
$text = preg_replace("/\[bookmark\=([^\]]*)\](.*?)\[\/bookmark\]/ism", ' $2 [url]$1[/url]',$text);
}
if ($simple_html == 5) {
$text = preg_replace("/[^#@!]\[url\=([$URLSearchString]*)\](.*?)\[\/url\]/ism", '[url]$1[/url]', $text);
}
// Perform URL Search
if ($try_oembed) {
$text = preg_replace_callback("/\[bookmark\=([^\]]*)\](.*?)\[\/bookmark\]/ism", $try_oembed_callback, $text);
}
if ($simple_html == 5) {
$text = preg_replace("/\[bookmark\=([^\]]*)\](.*?)\[\/bookmark\]/ism", '[url]$1[/url]', $text);
} else {
$text = preg_replace("/\[bookmark\=([^\]]*)\](.*?)\[\/bookmark\]/ism", '[url=$1]$2[/url]', $text);
}
// Handle Diaspora posts
$text = preg_replace_callback(
"&\[url=/?posts/([^\[\]]*)\](.*)\[\/url\]&Usi",
function ($match) {
return "[url=" . System::baseUrl() . "/display/" . $match[1] . "]" . $match[2] . "[/url]";
}, $text
);
$text = preg_replace_callback(
"&\[url=/people\?q\=(.*)\](.*)\[\/url\]&Usi",
function ($match) {
return "[url=" . System::baseUrl() . "/search?search=%40" . $match[1] . "]" . $match[2] . "[/url]";
}, $text
);
// Server independent link to posts and comments
// See issue: https://github.com/diaspora/diaspora_federation/issues/75
$expression = "=diaspora://.*?/post/([0-9A-Za-z\-_@.:]{15,254}[0-9A-Za-z])=ism";
$text = preg_replace($expression, System::baseUrl()."/display/$1", $text);
/* Tag conversion
* Supports:
* - #[url=][/url]
* - [url=]#[/url]
*/
$text = preg_replace_callback("/(?:#\[url\=[$URLSearchString]*\]|\[url\=[$URLSearchString]*\]#)(.*?)\[\/url\]/ism", function($matches) {
return '#'
. XML::escape($matches[1])
. '';
}, $text);
$text = preg_replace("/\[url\]([$URLSearchString]*)\[\/url\]/ism", '$1', $text);
$text = preg_replace("/\[url\=([$URLSearchString]*)\](.*?)\[\/url\]/ism", '$2', $text);
// Red compatibility, though the link can't be authenticated on Friendica
$text = preg_replace("/\[zrl\=([$URLSearchString]*)\](.*?)\[\/zrl\]/ism", '$2', $text);
// we may need to restrict this further if it picks up too many strays
// link acct:user@host to a webfinger profile redirector
$text = preg_replace('/acct:([^@]+)@((?!\-)(?:[a-zA-Z\d\-]{0,62}[a-zA-Z\d]\.){1,126}(?!\d+)[a-zA-Z\d]{1,63})/', 'acct:$1@$2', $text);
// Perform MAIL Search
$text = preg_replace("/\[mail\]([$MAILSearchString]*)\[\/mail\]/", '$1', $text);
$text = preg_replace("/\[mail\=([$MAILSearchString]*)\](.*?)\[\/mail\]/", '$2', $text);
// leave open the posibility of [map=something]
// this is replaced in Item::prepareBody() which has knowledge of the item location
if (strpos($text, '[/map]') !== false) {
$text = preg_replace_callback(
"/\[map\](.*?)\[\/map\]/ism",
function ($match) use ($simple_html) {
return str_replace($match[0], '
' . Map::byLocation($match[1], $simple_html) . '
', $match[0]);
},
$text
);
}
if (strpos($text, '[map=') !== false) {
$text = preg_replace_callback(
"/\[map=(.*?)\]/ism",
function ($match) use ($simple_html) {
return str_replace($match[0], '
', $text);
$text = str_replace('[hr]', '', $text);
// This is actually executed in Item::prepareBody()
$text = str_replace('[nosmile]', '', $text);
// Check for font change text
$text = preg_replace("/\[font=(.*?)\](.*?)\[\/font\]/sm", "$2", $text);
// Declare the format for [code] layout
$CodeLayout = '$1';
// Check for [code] text
$text = preg_replace("/\[code\](.*?)\[\/code\]/ism", "$CodeLayout", $text);
// Declare the format for [spoiler] layout
$SpoilerLayout = '
", $text);
// If we found an event earlier, strip out all the event code and replace with a reformatted version.
// Replace the event-start section with the entire formatted event. The other bbcode is stripped.
// Summary (e.g. title) is required, earlier revisions only required description (in addition to
// start which is always required). Allow desc with a missing summary for compatibility.
if ((!empty($ev['desc']) || !empty($ev['summary'])) && !empty($ev['start'])) {
$sub = Event::getHTML($ev, $simple_html);
$text = preg_replace("/\[event\-summary\](.*?)\[\/event\-summary\]/ism", '', $text);
$text = preg_replace("/\[event\-description\](.*?)\[\/event\-description\]/ism", '', $text);
$text = preg_replace("/\[event\-start\](.*?)\[\/event\-start\]/ism", $sub, $text);
$text = preg_replace("/\[event\-finish\](.*?)\[\/event\-finish\]/ism", '', $text);
$text = preg_replace("/\[event\-location\](.*?)\[\/event\-location\]/ism", '', $text);
$text = preg_replace("/\[event\-adjust\](.*?)\[\/event\-adjust\]/ism", '', $text);
$text = preg_replace("/\[event\-id\](.*?)\[\/event\-id\]/ism", '', $text);
}
// Replace non graphical smilies for external posts
if ($simple_html) {
$text = Smilies::replace($text, false, true);
}
// Unhide all [noparse] contained bbtags unspacefying them
// and triming the [noparse] tag.
$text = preg_replace_callback("/\[noparse\](.*?)\[\/noparse\]/ism", 'self::unescapeNoparseCallback', $text);
$text = preg_replace_callback("/\[nobb\](.*?)\[\/nobb\]/ism", 'self::unescapeNoparseCallback', $text);
$text = preg_replace_callback("/\[pre\](.*?)\[\/pre\]/ism", 'self::unescapeNoparseCallback', $text);
/// @todo What is the meaning of these lines?
$text = preg_replace('/\[\&\;([#a-z0-9]+)\;\]/', '&$1;', $text);
$text = preg_replace('/\&\#039\;/', '\'', $text);
// Currently deactivated, it made problems with " inside of alt texts.
//$text = preg_replace('/\"\;/', '"', $text);
// fix any escaped ampersands that may have been converted into links
$text = preg_replace('/\<([^>]*?)(src|href)=(.*?)\&\;(.*?)\>/ism', '<$1$2=$3&$4>', $text);
// sanitizes src attributes (http and redir URLs for displaying in a web page, cid used for inline images in emails)
$allowed_src_protocols = ['http', 'redir', 'cid'];
$text = preg_replace('#<([^>]*?)(src)="(?!' . implode('|', $allowed_src_protocols) . ')(.*?)"(.*?)>#ism',
'<$1$2=""$4 data-original-src="$3" class="invalid-src" title="' . L10n::t('Invalid source protocol') . '">', $text);
// sanitize href attributes (only whitelisted protocols URLs)
// default value for backward compatibility
$allowed_link_protocols = Config::get('system', 'allowed_link_protocols', ['ftp', 'mailto', 'gopher', 'cid']);
// Always allowed protocol even if config isn't set or not including it
$allowed_link_protocols[] = 'http';
$allowed_link_protocols[] = 'redir/';
$regex = '#<([^>]*?)(href)="(?!' . implode('|', $allowed_link_protocols) . ')(.*?)"(.*?)>#ism';
$text = preg_replace($regex, '<$1$2="javascript:void(0)"$4 data-original-href="$3" class="invalid-href" title="' . L10n::t('Invalid link protocol') . '">', $text);
if ($saved_image) {
$text = self::interpolateSavedImagesIntoItemBody($text, $saved_image);
}
// Restore code blocks
$text = preg_replace_callback('/#codeblock-([0-9]+)#/iU',
function ($matches) use ($codeblocks) {
$return = $matches[0];
if (isset($codeblocks[intval($matches[1])])) {
$return = $codeblocks[$matches[1]];
}
return $return;
},
$text
);
// Clean up the HTML by loading and saving the HTML with the DOM.
// Bad structured html can break a whole page.
// For performance reasons do it only with ativated item cache or at export.
if (!$try_oembed || (get_itemcachepath() != "")) {
$doc = new DOMDocument();
$doc->preserveWhiteSpace = false;
$text = mb_convert_encoding($text, 'HTML-ENTITIES', "UTF-8");
$doctype = '';
$encoding = '';
@$doc->loadHTML($encoding.$doctype."".$text."");
$doc->encoding = 'UTF-8';
$text = $doc->saveHTML();
$text = str_replace(["", "", $doctype, $encoding], ["", "", "", ""], $text);
$text = str_replace(' ', '', $text);
//$Text = mb_convert_encoding($Text, "UTF-8", 'HTML-ENTITIES');
}
// Clean up some useless linebreaks in lists
//$Text = str_replace('
"], $text);
$stamp1 = microtime(true);
// Now convert HTML to Markdown
$text = HTML::toMarkdown($text);
// unmask the special chars back to HTML
$text = str_replace(['&\_lt\_;', '&\_gt\_;', '&\_amp\_;'], ['<', '>', '&'], $text);
$a->saveTimestamp($stamp1, "parser");
// Libertree has a problem with escaped hashtags.
$text = str_replace(['\#'], ['#'], $text);
// Remove any leading or trailing whitespace, as this will mess up
// the Diaspora signature verification and cause the item to disappear
$text = trim($text);
if ($for_diaspora) {
$url_search_string = "^\[\]";
$text = preg_replace_callback(
"/([@]\[(.*?)\])\(([$url_search_string]*?)\)/ism",
['self', 'bbCodeMention2DiasporaCallback'],
$text
);
}
Addon::callHooks('bb2diaspora', $text);
return $text;
}
/**
* @brief Pull out all #hashtags and @person tags from $string.
*
* We also get @person@domain.com - which would make
* the regex quite complicated as tags can also
* end a sentence. So we'll run through our results
* and strip the period from any tags which end with one.
* Returns array of tags found, or empty array.
*
* @param string $string Post content
*
* @return array List of tag and person names
*/
public static function getTags($string)
{
$ret = [];
// Convert hashtag links to hashtags
$string = preg_replace('/#\[url\=([^\[\]]*)\](.*?)\[\/url\]/ism', '#$2', $string);
// ignore anything in a code block
$string = preg_replace('/\[code\](.*?)\[\/code\]/sm', '', $string);
// Force line feeds at bbtags
$string = str_replace(['[', ']'], ["\n[", "]\n"], $string);
// ignore anything in a bbtag
$string = preg_replace('/\[(.*?)\]/sm', '', $string);
// Match full names against @tags including the space between first and last
// We will look these up afterward to see if they are full names or not recognisable.
if (preg_match_all('/(@[^ \x0D\x0A,:?]+ [^ \x0D\x0A@,:?]+)([ \x0D\x0A@,:?]|$)/', $string, $matches)) {
foreach ($matches[1] as $match) {
if (strstr($match, ']')) {
// we might be inside a bbcode color tag - leave it alone
continue;
}
if (substr($match, -1, 1) === '.') {
$ret[] = substr($match, 0, -1);
} else {
$ret[] = $match;
}
}
}
// Otherwise pull out single word tags. These can be @nickname, @first_last
// and #hash tags.
if (preg_match_all('/([!#@][^\^ \x0D\x0A,;:?]+)([ \x0D\x0A,;:?]|$)/', $string, $matches)) {
foreach ($matches[1] as $match) {
if (strstr($match, ']')) {
// we might be inside a bbcode color tag - leave it alone
continue;
}
if (substr($match, -1, 1) === '.') {
$match = substr($match,0,-1);
}
// ignore strictly numeric tags like #1
if ((strpos($match, '#') === 0) && ctype_digit(substr($match, 1))) {
continue;
}
// try not to catch url fragments
if (strpos($string, $match) && preg_match('/[a-zA-z0-9\/]/', substr($string, strpos($string, $match) - 1, 1))) {
continue;
}
$ret[] = $match;
}
}
return $ret;
}
}