html_to_bbc and bbc_to_html added back to preserve compatibility as much as possible

Signed-off-by: emanuele <[email protected]>

Sources/Subs-Editor.php

@@ -16,6 +16,900 @@
 
 if (!defined('SMF'))
 	die('Hacking attempt...');
+
+/**
+ * !!!Compatibility!!!
+ * Since we changed the editor we don't need it any more, but let's keep it if any mod wants to use it
+ * Convert only the BBC that can be edited in HTML mode for the editor.
+ * 
+ * @param string $text
+ * @param boolean $compat_mode if true will convert the text, otherwise not (default false)
+ * @return string
+ */
+function bbc_to_html($text, $compat_mode = false)
+{
+	global $modSettings, $smcFunc;
+
+	if (!$compat_mode)
+		return $text;
+
+	// Turn line breaks back into br's.
+	$text = strtr($text, array("\r" => '', "\n" => '<br />'));
+
+	// Prevent conversion of all bbcode inside these bbcodes.
+	// @todo Tie in with bbc permissions ?
+	foreach (array('code', 'php', 'nobbc') as $code)
+	{
+		if (strpos($text, '['. $code) !== false)
+		{
+			$parts = preg_split('~(\[/' . $code . '\]|\[' . $code . '(?:=[^\]]+)?\])~i', $text, -1, PREG_SPLIT_DELIM_CAPTURE);
+
+			// Only mess with stuff inside tags.
+			for ($i = 0, $n = count($parts); $i < $n; $i++)
+			{
+				// Value of 2 means we're inside the tag.
+				if ($i % 4 == 2)
+					$parts[$i] = strtr($parts[$i], array('[' => '&#91;', ']' => '&#93;', "'" => "'"));
+			}
+			// Put our humpty dumpty message back together again.
+			$text = implode('', $parts);
+		}
+	}
+
+	// What tags do we allow?
+	$allowed_tags = array('b', 'u', 'i', 's', 'hr', 'list', 'li', 'font', 'size', 'color', 'img', 'left', 'center', 'right', 'url', 'email', 'ftp', 'sub', 'sup');
+
+	$text = parse_bbc($text, true, '', $allowed_tags);
+
+	// Fix for having a line break then a thingy.
+	$text = strtr($text, array('<br /><div' => '<div', "\n" => '', "\r" => ''));
+
+	// Note that IE doesn't understand spans really - make them something "legacy"
+	$working_html = array(
+		'~<del>(.+?)</del>~i' => '<strike>$1</strike>',
+		'~<span\sclass="bbc_u">(.+?)</span>~i' => '<u>$1</u>',
+		'~<span\sstyle="color:\s*([#\d\w]+);" class="bbc_color">(.+?)</span>~i' => '<font color="$1">$2</font>',
+		'~<span\sstyle="font-family:\s*([#\d\w\s]+);" class="bbc_font">(.+?)</span>~i' => '<font face="$1">$2</font>',
+		'~<div\sstyle="text-align:\s*(left|right);">(.+?)</div>~i' => '<p align="$1">$2</p>',
+	);
+	$text = preg_replace(array_keys($working_html), array_values($working_html), $text);
+
+	// Parse unique ID's and disable javascript into the smileys - using the double space.
+	$i = 1;
+	$text = preg_replace('~(?:\s|&nbsp;)?<(img\ssrc="' . preg_quote($modSettings['smileys_url'], '~') . '/[^<>]+?/([^<>]+?)"\s*)[^<>]*?class="smiley" />~e', '\'<\' . ' . 'stripslashes(\'$1\') . \'alt="" title="" onresizestart="return false;" id="smiley_\' . ' . "\$" . 'i++ . \'_$2" style="padding: 0 3px 0 3px;" />\'', $text);
+
+	return $text;
+}
+
+/**
+ * !!!Compatibility!!!
+ * This is no more needed, but to avoid break mods let's keep it
+ * Run it it shouldn't even hurt either, so let's not bother remove it
+ * 
+ * The harder one - wysiwyg to BBC!
+ * 
+ * @param string $text
+ * @return string
+ */
+function html_to_bbc($text)
+{
+	global $modSettings, $smcFunc, $sourcedir, $scripturl, $context;
+
+	// Replace newlines with spaces, as that's how browsers usually interpret them.
+	$text = preg_replace("~\s*[\r\n]+\s*~", ' ', $text);
+
+	// Though some of us love paragraphs, the parser will do better with breaks.
+	$text = preg_replace('~</p>\s*?<p~i', '</p><br /><p', $text);
+	$text = preg_replace('~</p>\s*(?!<)~i', '</p><br />', $text);
+
+	// Safari/webkit wraps lines in Wysiwyg in <div>'s.
+	if (isBrowser('webkit'))
+		$text = preg_replace(array('~<div(?:\s(?:[^<>]*?))?' . '>~i', '</div>'), array('<br />', ''), $text);
+
+	// If there's a trailing break get rid of it - Firefox tends to add one.
+	$text = preg_replace('~<br\s?/?' . '>$~i', '', $text);
+
+	// Remove any formatting within code tags.
+	if (strpos($text, '[code') !== false)
+	{
+		$text = preg_replace('~<br\s?/?' . '>~i', '#smf_br_spec_grudge_cool!#', $text);
+		$parts = preg_split('~(\[/code\]|\[code(?:=[^\]]+)?\])~i', $text, -1, PREG_SPLIT_DELIM_CAPTURE);
+
+		// Only mess with stuff outside [code] tags.
+		for ($i = 0, $n = count($parts); $i < $n; $i++)
+		{
+			// Value of 2 means we're inside the tag.
+			if ($i % 4 == 2)
+				$parts[$i] = strip_tags($parts[$i]);
+		}
+
+		$text = strtr(implode('', $parts), array('#smf_br_spec_grudge_cool!#' => '<br />'));
+	}
+
+	// Remove scripts, style and comment blocks.
+	$text = preg_replace('~<script[^>]*[^/]?' . '>.*?</script>~i', '', $text);
+	$text = preg_replace('~<style[^>]*[^/]?' . '>.*?</style>~i', '', $text);
+	$text = preg_replace('~\\<\\!--.*?-->~i', '', $text);
+	$text = preg_replace('~\\<\\!\\[CDATA\\[.*?\\]\\]\\>~i', '', $text);
+
+	// Do the smileys ultra first!
+	preg_match_all('~<img\s+[^<>]*?id="*smiley_\d+_([^<>]+?)[\s"/>]\s*[^<>]*?/*>(?:\s)?~i', $text, $matches);
+	if (!empty($matches[0]))
+	{
+		// Easy if it's not custom.
+		if (empty($modSettings['smiley_enable']))
+		{
+			$smileysfrom = array('>:D', ':D', '::)', '>:(', ':)', ';)', ';D', ':(', ':o', '8)', ':P', '???', ':-[', ':-X', ':-*', ':\'(', ':-\\', '^-^', 'O0', 'C:-)', '0:)');
+			$smileysto = array('evil.gif', 'cheesy.gif', 'rolleyes.gif', 'angry.gif', 'smiley.gif', 'wink.gif', 'grin.gif', 'sad.gif', 'shocked.gif', 'cool.gif', 'tongue.gif', 'huh.gif', 'embarrassed.gif', 'lipsrsealed.gif', 'kiss.gif', 'cry.gif', 'undecided.gif', 'azn.gif', 'afro.gif', 'police.gif', 'angel.gif');
+
+			foreach ($matches[1] as $k => $file)
+			{
+				$found = array_search($file, $smileysto);
+				// Note the weirdness here is to stop double spaces between smileys.
+				if ($found)
+					$matches[1][$k] = '-[]-smf_smily_start#|#' . htmlspecialchars($smileysfrom[$found]) . '-[]-smf_smily_end#|#';
+				else
+					$matches[1][$k] = '';
+			}
+		}
+		else
+		{
+			// Load all the smileys.
+			$names = array();
+			foreach ($matches[1] as $file)
+				$names[] = $file;
+			$names = array_unique($names);
+
+			if (!empty($names))
+			{
+				$request = $smcFunc['db_query']('', '
+					SELECT code, filename
+					FROM {db_prefix}smileys
+					WHERE filename IN ({array_string:smiley_filenames})',
+					array(
+						'smiley_filenames' => $names,
+					)
+				);
+				$mappings = array();
+				while ($row = $smcFunc['db_fetch_assoc']($request))
+					$mappings[$row['filename']] = htmlspecialchars($row['code']);
+				$smcFunc['db_free_result']($request);
+
+				foreach ($matches[1] as $k => $file)
+					if (isset($mappings[$file]))
+						$matches[1][$k] = '-[]-smf_smily_start#|#' . $mappings[$file] . '-[]-smf_smily_end#|#';
+			}
+		}
+
+		// Replace the tags!
+		$text = str_replace($matches[0], $matches[1], $text);
+
+		// Now sort out spaces
+		$text = str_replace(array('-[]-smf_smily_end#|#-[]-smf_smily_start#|#', '-[]-smf_smily_end#|#', '-[]-smf_smily_start#|#'), ' ', $text);
+	}
+
+	// Only try to buy more time if the client didn't quit.
+	if (connection_aborted() && $context['server']['is_apache'])
+		@apache_reset_timeout();
+
+	$parts = preg_split('~(<[A-Za-z]+\s*[^<>]*?style="?[^<>"]+"?[^<>]*?(?:/?)>|</[A-Za-z]+>)~', $text, -1, PREG_SPLIT_DELIM_CAPTURE);
+	$replacement = '';
+	$stack = array();
+
+	foreach ($parts as $part)
+	{
+		if (preg_match('~(<([A-Za-z]+)\s*[^<>]*?)style="?([^<>"]+)"?([^<>]*?(/?)>)~', $part, $matches) === 1)
+		{
+			// If it's being closed instantly, we can't deal with it...yet.
+			if ($matches[5] === '/')
+				continue;
+			else
+			{
+				// Get an array of styles that apply to this element. (The strtr is there to combat HTML generated by Word.)
+				$styles = explode(';', strtr($matches[3], array('&quot;' => '')));
+				$curElement = $matches[2];
+				$precedingStyle = $matches[1];
+				$afterStyle = $matches[4];
+				$curCloseTags = '';
+				$extra_attr = '';
+
+				foreach ($styles as $type_value_pair)
+				{
+					// Remove spaces and convert uppercase letters.
+					$clean_type_value_pair = strtolower(strtr(trim($type_value_pair), '=', ':'));
+
+					// Something like 'font-weight: bold' is expected here.
+					if (strpos($clean_type_value_pair, ':') === false)
+						continue;
+
+					// Capture the elements of a single style item (e.g. 'font-weight' and 'bold').
+					list ($style_type, $style_value) = explode(':', $type_value_pair);
+
+					$style_value = trim($style_value);
+
+					switch (trim($style_type))
+					{
+						case 'font-weight':
+							if ($style_value === 'bold')
+							{
+								$curCloseTags .= '[/b]';
+								$replacement .= '[b]';
+							}
+						break;
+
+						case 'text-decoration':
+							if ($style_value == 'underline')
+							{
+								$curCloseTags .= '[/u]';
+								$replacement .= '[u]';
+							}
+							elseif ($style_value == 'line-through')
+							{
+								$curCloseTags .= '[/s]';
+								$replacement .= '[s]';
+							}
+						break;
+
+						case 'text-align':
+							if ($style_value == 'left')
+							{
+								$curCloseTags .= '[/left]';
+								$replacement .= '[left]';
+							}
+							elseif ($style_value == 'center')
+							{
+								$curCloseTags .= '[/center]';
+								$replacement .= '[center]';
+							}
+							elseif ($style_value == 'right')
+							{
+								$curCloseTags .= '[/right]';
+								$replacement .= '[right]';
+							}
+						break;
+
+						case 'font-style':
+							if ($style_value == 'italic')
+							{
+								$curCloseTags .= '[/i]';
+								$replacement .= '[i]';
+							}
+						break;
+
+						case 'color':
+							$curCloseTags .= '[/color]';
+							$replacement .= '[color=' . $style_value . ']';
+						break;
+
+						case 'font-size':
+							// Sometimes people put decimals where decimals should not be.
+							if (preg_match('~(\d)+\.\d+(p[xt])~i', $style_value, $dec_matches) === 1)
+								$style_value = $dec_matches[1] . $dec_matches[2];
+
+							$curCloseTags .= '[/size]';
+							$replacement .= '[size=' . $style_value . ']';
+						break;
+
+						case 'font-family':
+							// Only get the first freaking font if there's a list!
+							if (strpos($style_value, ',') !== false)
+								$style_value = substr($style_value, 0, strpos($style_value, ','));
+
+							$curCloseTags .= '[/font]';
+							$replacement .= '[font=' . strtr($style_value, array("'" => '')) . ']';
+						break;
+
+						// This is a hack for images with dimensions embedded.
+						case 'width':
+						case 'height':
+							if (preg_match('~[1-9]\d*~i', $style_value, $dimension) === 1)
+								$extra_attr .= ' ' . $style_type . '="' . $dimension[0] . '"';
+						break;
+
+						case 'list-style-type':
+							if (preg_match('~none|disc|circle|square|decimal|decimal-leading-zero|lower-roman|upper-roman|lower-alpha|upper-alpha|lower-greek|lower-latin|upper-latin|hebrew|armenian|georgian|cjk-ideographic|hiragana|katakana|hiragana-iroha|katakana-iroha~i', $style_value, $listType) === 1)
+								$extra_attr .= ' listtype="' . $listType[0] . '"';
+						break;
+					}
+				}
+
+				// Preserve some tags stripping the styling.
+				if (in_array($matches[2], array('a', 'font', 'td')))
+				{
+					$replacement .= $precedingStyle . $afterStyle;
+					$curCloseTags = '</' . $matches[2] . '>' . $curCloseTags;
+				}
+
+				// If there's something that still needs closing, push it to the stack.
+				if (!empty($curCloseTags))
+					array_push($stack, array(
+							'element' => strtolower($curElement),
+							'closeTags' => $curCloseTags
+						)
+					);
+				elseif (!empty($extra_attr))
+					$replacement .= $precedingStyle . $extra_attr . $afterStyle;
+			}
+		}
+
+		elseif (preg_match('~</([A-Za-z]+)>~', $part, $matches) === 1)
+		{
+			// Is this the element that we've been waiting for to be closed?
+			if (!empty($stack) && strtolower($matches[1]) === $stack[count($stack) - 1]['element'])
+			{
+				$byebyeTag = array_pop($stack);
+				$replacement .= $byebyeTag['closeTags'];
+			}
+
+			// Must've been something else.
+			else
+				$replacement .= $part;
+		}
+		// In all other cases, just add the part to the replacement.
+		else
+			$replacement .= $part;
+	}
+
+	// Now put back the replacement in the text.
+	$text = $replacement;
+
+	// We are not finished yet, request more time.
+	if (connection_aborted() && $context['server']['is_apache'])
+		@apache_reset_timeout();
+
+	// Let's pull out any legacy alignments.
+	while (preg_match('~<([A-Za-z]+)\s+[^<>]*?(align="*(left|center|right)"*)[^<>]*?(/?)>~i', $text, $matches) === 1)
+	{
+		// Find the position in the text of this tag over again.
+		$start_pos = strpos($text, $matches[0]);
+		if ($start_pos === false)
+			break;
+
+		// End tag?
+		if ($matches[4] != '/' && strpos($text, '</' . $matches[1] . '>', $start_pos) !== false)
+		{
+			$end_length = strlen('</' . $matches[1] . '>');
+			$end_pos = strpos($text, '</' . $matches[1] . '>', $start_pos);
+
+			// Remove the align from that tag so it's never checked again.
+			$tag = substr($text, $start_pos, strlen($matches[0]));
+			$content = substr($text, $start_pos + strlen($matches[0]), $end_pos - $start_pos - strlen($matches[0]));
+			$tag = str_replace($matches[2], '', $tag);
+
+			// Put the tags back into the body.
+			$text = substr($text, 0, $start_pos) . $tag . '[' . $matches[3] . ']' . $content . '[/' . $matches[3] . ']' . substr($text, $end_pos);
+		}
+		else
+		{
+			// Just get rid of this evil tag.
+			$text = substr($text, 0, $start_pos) . substr($text, $start_pos + strlen($matches[0]));
+		}
+	}
+
+	// Let's do some special stuff for fonts - cause we all love fonts.
+	while (preg_match('~<font\s+([^<>]*)>~i', $text, $matches) === 1)
+	{
+		// Find the position of this again.
+		$start_pos = strpos($text, $matches[0]);
+		$end_pos = false;
+		if ($start_pos === false)
+			break;
+
+		// This must have an end tag - and we must find the right one.
+		$lower_text = strtolower($text);
+
+		$start_pos_test = $start_pos + 4;
+		// How many starting tags must we find closing ones for first?
+		$start_font_tag_stack = 0;
+		while ($start_pos_test < strlen($text))
+		{
+			// Where is the next starting font?
+			$next_start_pos = strpos($lower_text, '<font', $start_pos_test);
+			$next_end_pos = strpos($lower_text, '</font>', $start_pos_test);
+
+			// Did we past another starting tag before an end one?
+			if ($next_start_pos !== false && $next_start_pos < $next_end_pos)
+			{
+				$start_font_tag_stack++;
+				$start_pos_test = $next_start_pos + 4;
+			}
+			// Otherwise we have an end tag but not the right one?
+			elseif ($start_font_tag_stack)
+			{
+				$start_font_tag_stack--;
+				$start_pos_test = $next_end_pos + 4;
+			}
+			// Otherwise we're there!
+			else
+			{
+				$end_pos = $next_end_pos;
+				break;
+			}
+		}
+		if ($end_pos === false)
+			break;
+
+		// Now work out what the attributes are.
+		$attribs = fetchTagAttributes($matches[1]);
+		$tags = array();
+		$sizes_equivalence = array(1 => '8pt', '10pt', '12pt', '14pt', '18pt', '24pt', '36pt');
+		foreach ($attribs as $s => $v)
+		{
+			if ($s == 'size')
+			{
+				// Cast before empty chech because casting a string results in a 0 and we don't have zeros in the array! ;)
+				$v = (int) trim($v);
+				$v = empty($v) ? 1 : $v;
+				$tags[] = array('[size=' . $sizes_equivalence[$v] . ']', '[/size]');
+			}
+			elseif ($s == 'face')
+				$tags[] = array('[font=' . trim(strtolower($v)) . ']', '[/font]');
+			elseif ($s == 'color')
+				$tags[] = array('[color=' . trim(strtolower($v)) . ']', '[/color]');
+		}
+
+		// As before add in our tags.
+		$before = $after = '';
+		foreach ($tags as $tag)
+		{
+			$before .= $tag[0];
+			if (isset($tag[1]))
+				$after = $tag[1] . $after;
+		}
+
+		// Remove the tag so it's never checked again.
+		$content = substr($text, $start_pos + strlen($matches[0]), $end_pos - $start_pos - strlen($matches[0]));
+
+		// Put the tags back into the body.
+		$text = substr($text, 0, $start_pos) . $before . $content . $after . substr($text, $end_pos + 7);
+	}
+
+	// Almost there, just a little more time.
+	if (connection_aborted() && $context['server']['is_apache'])
+		@apache_reset_timeout();
+
+	if (count($parts = preg_split('~<(/?)(li|ol|ul)([^>]*)>~i', $text, null, PREG_SPLIT_DELIM_CAPTURE)) > 1)
+	{
+		// A toggle that dermines whether we're directly under a <ol> or <ul>.
+		$inList = false;
+
+		// Keep track of the number of nested list levels.
+		$listDepth = 0;
+
+		// Map what we can expect from the HTML to what is supported by SMF.
+		$listTypeMapping = array(
+			'1' => 'decimal',
+			'A' => 'upper-alpha',
+			'a' => 'lower-alpha',
+			'I' => 'upper-roman',
+			'i' => 'lower-roman',
+			'disc' => 'disc',
+			'square' => 'square',
+			'circle' => 'circle',
+		);
+
+		// $i: text, $i + 1: '/', $i + 2: tag, $i + 3: tail.
+		for ($i = 0, $numParts = count($parts) - 1; $i < $numParts; $i += 4)
+		{
+			$tag = strtolower($parts[$i + 2]);
+			$isOpeningTag = $parts[$i + 1] === '';
+
+			if ($isOpeningTag)
+			{
+				switch ($tag)
+				{
+					case 'ol':
+					case 'ul':
+
+						// We have a problem, we're already in a list.
+						if ($inList)
+						{
+							// Inject a list opener, we'll deal with the ol/ul next loop.
+							array_splice($parts, $i, 0, array(
+								'',
+								'',
+								str_repeat("\t", $listDepth) . '[li]',
+								'',
+							));
+							$numParts = count($parts) - 1;
+
+							// The inlist status changes a bit.
+							$inList = false;
+						}
+
+						// Just starting a new list.
+						else
+						{
+							$inList = true;
+
+							if ($tag === 'ol')
+								$listType = 'decimal';
+							elseif (preg_match('~type="?(' . implode('|', array_keys($listTypeMapping)) . ')"?~', $parts[$i + 3], $match) === 1)
+								$listType = $listTypeMapping[$match[1]];
+							else
+								$listType = null;
+
+							$listDepth++;
+
+							$parts[$i + 2] = '[list' . ($listType === null ? '' : ' type=' . $listType) . ']' . "\n";
+							$parts[$i + 3] = '';
+						}
+					break;
+
+					case 'li':
+
+						// This is how it should be: a list item inside the list.
+						if ($inList)
+						{
+							$parts[$i + 2] = str_repeat("\t", $listDepth) . '[li]';
+							$parts[$i + 3] = '';
+
+							// Within a list item, it's almost as if you're outside.
+							$inList = false;
+						}
+
+						// The li is no direct child of a list.
+						else
+						{
+							// We are apparently in a list item.
+							if ($listDepth > 0)
+							{
+								$parts[$i + 2] = '[/li]' . "\n" . str_repeat("\t", $listDepth) . '[li]';
+								$parts[$i + 3] = '';
+							}
+
+							// We're not even near a list.
+							else
+							{
+								// Quickly create a list with an item.
+								$listDepth++;
+
+								$parts[$i + 2] = '[list]' . "\n\t" . '[li]';
+								$parts[$i + 3] = '';
+							}
+						}
+
+					break;
+				}
+			}
+
+			// Handle all the closing tags.
+			else
+			{
+				switch ($tag)
+				{
+					case 'ol':
+					case 'ul':
+
+						// As we expected it, closing the list while we're in it.
+						if ($inList)
+						{
+							$inList = false;
+
+							$listDepth--;
+
+							$parts[$i + 1] = '';
+							$parts[$i + 2] = str_repeat("\t", $listDepth) . '[/list]';
+							$parts[$i + 3] = '';
+						}
+
+						else
+						{
+							// We're in a list item.
+							if ($listDepth > 0)
+							{
+								// Inject closure for this list item first.
+								// The content of $parts[$i] is left as is!
+								array_splice($parts, $i + 1, 0, array(
+									'',				// $i + 1
+									'[/li]' . "\n",	// $i + 2
+									'',				// $i + 3
+									'',				// $i + 4
+								));
+								$numParts = count($parts) - 1;
+
+								// Now that we've closed the li, we're in list space.
+								$inList = true;
+							}
+
+							// We're not even in a list, ignore
+							else
+							{
+								$parts[$i + 1] = '';
+								$parts[$i + 2] = '';
+								$parts[$i + 3] = '';
+							}
+						}
+					break;
+
+					case 'li':
+
+						if ($inList)
+						{
+							// There's no use for a </li> after <ol> or <ul>, ignore.
+							$parts[$i + 1] = '';
+							$parts[$i + 2] = '';
+							$parts[$i + 3] = '';
+						}
+
+						else
+						{
+							// Remove the trailing breaks from the list item.
+							$parts[$i] = preg_replace('~\s*<br\s*' . '/?' . '>\s*$~', '', $parts[$i]);
+							$parts[$i + 1] = '';
+							$parts[$i + 2] = '[/li]' . "\n";
+							$parts[$i + 3] = '';
+
+							// And we're back in the [list] space.
+							$inList = true;
+						}
+
+					break;
+				}
+			}
+
+			// If we're in the [list] space, no content is allowed.
+			if ($inList && trim(preg_replace('~\s*<br\s*' . '/?' . '>\s*~', '', $parts[$i + 4])) !== '')
+			{
+				// Fix it by injecting an extra list item.
+				array_splice($parts, $i + 4, 0, array(
+					'', // No content.
+					'', // Opening tag.
+					'li', // It's a <li>.
+					'', // No tail.
+				));
+				$numParts = count($parts) - 1;
+			}
+		}
+
+		$text = implode('', $parts);
+
+		if ($inList)
+		{
+			$listDepth--;
+			$text .= str_repeat("\t", $listDepth) . '[/list]';
+		}
+
+		for ($i = $listDepth; $i > 0; $i--)
+			$text .= '[/li]' . "\n" . str_repeat("\t", $i - 1) . '[/list]';
+
+	}
+
+	// I love my own image...
+	while (preg_match('~<img\s+([^<>]*)/*>~i', $text, $matches) === 1)
+	{
+		// Find the position of the image.
+		$start_pos = strpos($text, $matches[0]);
+		if ($start_pos === false)
+			break;
+		$end_pos = $start_pos + strlen($matches[0]);
+
+		$params = '';
+		$had_params = array();
+		$src = '';
+
+		$attrs = fetchTagAttributes($matches[1]);
+		foreach ($attrs as $attrib => $value)
+		{
+			if (in_array($attrib, array('width', 'height')))
+				$params .= ' ' . $attrib . '=' . (int) $value;
+			elseif ($attrib == 'alt' && trim($value) != '')
+				$params .= ' alt=' . trim($value);
+			elseif ($attrib == 'src')
+				$src = trim($value);
+		}
+
+		$tag = '';
+		if (!empty($src))
+		{
+			// Attempt to fix the path in case it's not present.
+			if (preg_match('~^https?://~i', $src) === 0 && is_array($parsedURL = parse_url($scripturl)) && isset($parsedURL['host']))
+			{
+				$baseURL = (isset($parsedURL['scheme']) ? $parsedURL['scheme'] : 'http') . '://' . $parsedURL['host'] . (empty($parsedURL['port']) ? '' : ':' . $parsedURL['port']);
+
+				if (substr($src, 0, 1) === '/')
+					$src = $baseURL . $src;
+				else
+					$src = $baseURL . (empty($parsedURL['path']) ? '/' : preg_replace('~/(?:index\\.php)?$~', '', $parsedURL['path'])) . '/' . $src;
+			}
+
+			$tag = '[img' . $params . ']' . $src . '[/img]';
+		}
+
+		// Replace the tag
+		$text = substr($text, 0, $start_pos) . $tag . substr($text, $end_pos);
+	}
+
+	// The final bits are the easy ones - tags which map to tags which map to tags - etc etc.
+	$tags = array(
+		'~<b(\s(.)*?)*?' . '>~i' => '[b]',
+		'~</b>~i' => '[/b]',
+		'~<i(\s(.)*?)*?' . '>~i' => '[i]',
+		'~</i>~i' => '[/i]',
+		'~<u(\s(.)*?)*?' . '>~i' => '[u]',
+		'~</u>~i' => '[/u]',
+		'~<strong(\s(.)*?)*?' . '>~i' => '[b]',
+		'~</strong>~i' => '[/b]',
+		'~<em(\s(.)*?)*?' . '>~i' => '[i]',
+		'~</em>~i' => '[/i]',
+		'~<s(\s(.)*?)*?' . '>~i' => "[s]",
+		'~</s>~i' => "[/s]",
+		'~<strike(\s(.)*?)*?' . '>~i' => '[s]',
+		'~</strike>~i' => '[/s]',
+		'~<del(\s(.)*?)*?' . '>~i' => '[s]',
+		'~</del>~i' => '[/s]',
+		'~<center(\s(.)*?)*?' . '>~i' => '[center]',
+		'~</center>~i' => '[/center]',
+		'~<pre(\s(.)*?)*?' . '>~i' => '[pre]',
+		'~</pre>~i' => '[/pre]',
+		'~<sub(\s(.)*?)*?' . '>~i' => '[sub]',
+		'~</sub>~i' => '[/sub]',
+		'~<sup(\s(.)*?)*?' . '>~i' => '[sup]',
+		'~</sup>~i' => '[/sup]',
+		'~<tt(\s(.)*?)*?' . '>~i' => '[tt]',
+		'~</tt>~i' => '[/tt]',
+		'~<table(\s(.)*?)*?' . '>~i' => '[table]',
+		'~</table>~i' => '[/table]',
+		'~<tr(\s(.)*?)*?' . '>~i' => '[tr]',
+		'~</tr>~i' => '[/tr]',
+		'~<(td|th)\s[^<>]*?colspan="?(\d{1,2})"?.*?' . '>~ie' => 'str_repeat(\'[td][/td]\', $2 - 1) . \'[td]\'',
+		'~<(td|th)(\s(.)*?)*?' . '>~i' => '[td]',
+		'~</(td|th)>~i' => '[/td]',
+		'~<br(?:\s[^<>]*?)?' . '>~i' => "\n",
+		'~<hr[^<>]*>(\n)?~i' => "[hr]\n$1",
+		'~(\n)?\\[hr\\]~i' => "\n[hr]",
+		'~^\n\\[hr\\]~i' => "[hr]",
+		'~<blockquote(\s(.)*?)*?' . '>~i' => "&lt;blockquote&gt;",
+		'~</blockquote>~i' => "&lt;/blockquote&gt;",
+		'~<ins(\s(.)*?)*?' . '>~i' => "&lt;ins&gt;",
+		'~</ins>~i' => "&lt;/ins&gt;",
+	);
+	$text = preg_replace(array_keys($tags), array_values($tags), $text);
+
+	// Please give us just a little more time.
+	if (connection_aborted() && $context['server']['is_apache'])
+		@apache_reset_timeout();
+
+	// What about URL's - the pain in the ass of the tag world.
+	while (preg_match('~<a\s+([^<>]*)>([^<>]*)</a>~i', $text, $matches) === 1)
+	{
+		// Find the position of the URL.
+		$start_pos = strpos($text, $matches[0]);
+		if ($start_pos === false)
+			break;
+		$end_pos = $start_pos + strlen($matches[0]);
+
+		$tag_type = 'url';
+		$href = '';
+
+		$attrs = fetchTagAttributes($matches[1]);
+		foreach ($attrs as $attrib => $value)
+		{
+			if ($attrib == 'href')
+			{
+				$href = trim($value);
+
+				// Are we dealing with an FTP link?
+				if (preg_match('~^ftps?://~', $href) === 1)
+					$tag_type = 'ftp';
+
+				// Or is this a link to an email address?
+				elseif (substr($href, 0, 7) == 'mailto:')
+				{
+					$tag_type = 'email';
+					$href = substr($href, 7);
+				}
+
+				// No http(s), so attempt to fix this potential relative URL.
+				elseif (preg_match('~^https?://~i', $href) === 0 && is_array($parsedURL = parse_url($scripturl)) && isset($parsedURL['host']))
+				{
+					$baseURL = (isset($parsedURL['scheme']) ? $parsedURL['scheme'] : 'http') . '://' . $parsedURL['host'] . (empty($parsedURL['port']) ? '' : ':' . $parsedURL['port']);
+
+					if (substr($href, 0, 1) === '/')
+						$href = $baseURL . $href;
+					else
+						$href = $baseURL . (empty($parsedURL['path']) ? '/' : preg_replace('~/(?:index\\.php)?$~', '', $parsedURL['path'])) . '/' . $href;
+				}
+			}
+
+			// External URL?
+			if ($attrib == 'target' && $tag_type == 'url')
+			{
+				if (trim($value) == '_blank')
+					$tag_type == 'iurl';
+			}
+		}
+
+		$tag = '';
+		if ($href != '')
+		{
+			if ($matches[2] == $href)
+				$tag = '[' . $tag_type . ']' . $href . '[/' . $tag_type . ']';
+			else
+				$tag = '[' . $tag_type . '=' . $href . ']' . $matches[2] . '[/' . $tag_type . ']';
+		}
+
+		// Replace the tag
+		$text = substr($text, 0, $start_pos) . $tag . substr($text, $end_pos);
+	}
+
+	$text = strip_tags($text);
+
+	// Some tags often end up as just dummy tags - remove those.
+	$text = preg_replace('~\[[bisu]\]\s*\[/[bisu]\]~', '', $text);
+
+	// Fix up entities.
+	$text = preg_replace('~&#38;~i', '&#38;#38;', $text);
+
+	$text = legalise_bbc($text);
+
+	return $text;
+}
+
+/**
+ * !!!Compatibility!!!
+ * This is no more needed, but to avoid break mods let's keep it
+ * 
+ * Returns an array of attributes associated with a tag.
+ * 
+ * @param string $text
+ * @return string
+ */
+function fetchTagAttributes($text)
+{
+	$attribs = array();
+	$key = $value = '';
+	$strpos = 0;
+	$tag_state = 0; // 0 = key, 1 = attribute with no string, 2 = attribute with string
+	for ($i = 0; $i < strlen($text); $i++)
+	{
+		// We're either moving from the key to the attribute or we're in a string and this is fine.
+		if ($text[$i] == '=')
+		{
+			if ($tag_state == 0)
+				$tag_state = 1;
+			elseif ($tag_state == 2)
+				$value .= '=';
+		}
+		// A space is either moving from an attribute back to a potential key or in a string is fine.
+		elseif ($text[$i] == ' ')
+		{
+			if ($tag_state == 2)
+				$value .= ' ';
+			elseif ($tag_state == 1)
+			{
+				$attribs[$key] = $value;
+				$key = $value = '';
+				$tag_state = 0;
+			}
+		}
+		// A quote?
+		elseif ($text[$i] == '"')
+		{
+			// Must be either going into or out of a string.
+			if ($tag_state == 1)
+				$tag_state = 2;
+			else
+				$tag_state = 1;
+		}
+		// Otherwise it's fine.
+		else
+		{
+			if ($tag_state == 0)
+				$key .= $text[$i];
+			else
+				$value .= $text[$i];
+		}
+	}
+
+	// Anything left?
+	if ($key != '' && $value != '')
+		$attribs[$key] = $value;
+
+	return $attribs;
+}
+
+
 /**
  * Creates the javascript code for localization of the editor (SCEditor)
  */