githubFork/roundcubemail.git

Aleksander Machniak

2013-10-13 ff6de99ae4f353f793c95208ad6407ac59a63376

commit \| author \| age
4e17e6	1	<?php
T	2
66afd7	3	/**
AM	4	+-----------------------------------------------------------------------+
	5	\| This file is part of the Roundcube Webmail client \|
	6	\| Copyright (C) 2008-2012, The Roundcube Dev Team \|
	7	\| Copyright (c) 2005-2007, Jon Abernathy <jon@chuggnutt.com> \|
	8	\| \|
	9	\| Licensed under the GNU General Public License version 3 or \|
	10	\| any later version with exceptions for skins & plugins. \|
	11	\| See the README file for a full license statement. \|
	12	\| \|
	13	\| PURPOSE: \|
	14	\| Converts HTML to formatted plain text (based on html2text class) \|
	15	+-----------------------------------------------------------------------+
	16	\| Author: Thomas Bruederli <roundcube@gmail.com> \|
	17	\| Author: Aleksander Machniak <alec@alec.pl> \|
	18	\| Author: Jon Abernathy <jon@chuggnutt.com> \|
	19	+-----------------------------------------------------------------------+
	20	*/
4e17e6	21
T	22	/**
8ac6fd	23	* Takes HTML and converts it to formatted, plain text.
A	24	*
	25	* Thanks to Alexander Krug (http://www.krugar.de/) to pointing out and
	26	* correcting an error in the regexp search array. Fixed 7/30/03.
	27	*
	28	* Updated set_html() function's file reading mechanism, 9/25/03.
	29	*
	30	* Thanks to Joss Sanglier (http://www.dancingbear.co.uk/) for adding
	31	* several more HTML entity codes to the $search and $replace arrays.
	32	* Updated 11/7/03.
	33	*
	34	* Thanks to Darius Kasperavicius (http://www.dar.dar.lt/) for
	35	* suggesting the addition of $allowed_tags and its supporting function
	36	* (which I slightly modified). Updated 3/12/04.
	37	*
	38	* Thanks to Justin Dearing for pointing out that a replacement for the
	39	* <TH> tag was missing, and suggesting an appropriate fix.
	40	* Updated 8/25/04.
	41	*
	42	* Thanks to Mathieu Collas (http://www.myefarm.com/) for finding a
	43	* display/formatting bug in the _build_link_list() function: email
	44	* readers would show the left bracket and number ("[1") as part of the
	45	* rendered email address.
	46	* Updated 12/16/04.
	47	*
	48	* Thanks to Wojciech Bajon (http://histeria.pl/) for submitting code
	49	* to handle relative links, which I hadn't considered. I modified his
	50	* code a bit to handle normal HTTP links and MAILTO links. Also for
	51	* suggesting three additional HTML entity codes to search for.
	52	* Updated 03/02/05.
	53	*
	54	* Thanks to Jacob Chandler for pointing out another link condition
	55	* for the _build_link_list() function: "https".
	56	* Updated 04/06/05.
	57	*
	58	* Thanks to Marc Bertrand (http://www.dresdensky.com/) for
	59	* suggesting a revision to the word wrapping functionality; if you
	60	* specify a $width of 0 or less, word wrapping will be ignored.
	61	* Updated 11/02/06.
	62	*
	63	* *** Big housecleaning updates below:
	64	*
	65	* Thanks to Colin Brown (http://www.sparkdriver.co.uk/) for
	66	* suggesting the fix to handle </li> and blank lines (whitespace).
	67	* Christian Basedau (http://www.movetheweb.de/) also suggested the
	68	* blank lines fix.
	69	*
	70	* Special thanks to Marcus Bointon (http://www.synchromedia.co.uk/),
	71	* Christian Basedau, Norbert Laposa (http://ln5.co.uk/),
	72	* Bas van de Weijer, and Marijn van Butselaar
	73	* for pointing out my glaring error in the <th> handling. Marcus also
	74	* supplied a host of fixes.
	75	*
	76	* Thanks to Jeffrey Silverman (http://www.newtnotes.com/) for pointing
	77	* out that extra spaces should be compressed--a problem addressed with
	78	* Marcus Bointon's fixes but that I had not yet incorporated.
	79	*
21d463	80	* Thanks to Daniel Schledermann (http://www.typoconsult.dk/) for
8ac6fd	81	* suggesting a valuable fix with <a> tag handling.
A	82	*
	83	* Thanks to Wojciech Bajon (again!) for suggesting fixes and additions,
	84	* including the <a> tag handling that Daniel Schledermann pointed
	85	* out but that I had not yet incorporated. I haven't (yet)
	86	* incorporated all of Wojciech's changes, though I may at some
	87	* future time.
	88	*
	89	* *** End of the housecleaning updates. Updated 08/08/07.
	90	*/
66afd7	91
AM	92	/**
	93	* Converts HTML to formatted plain text
	94	*
	95	* @package Framework
	96	* @subpackage Utils
	97	*/
	98	class rcube_html2text
4e17e6	99	{
66afd7	100	/**
AM	101	* Contains the HTML content to convert.
	102	*
	103	* @var string $html
	104	*/
	105	protected $html;
4e17e6	106
T	107	/**
66afd7	108	* Contains the converted, formatted text.
4e17e6	109	*
66afd7	110	* @var string $text
4e17e6	111	*/
66afd7	112	protected $text;
4e17e6	113
T	114	/**
66afd7	115	* Maximum width of the formatted text, in columns.
4e17e6	116	*
66afd7	117	* Set this value to 0 (or less) to ignore word wrapping
AM	118	* and not constrain text to a fixed-width column.
	119	*
	120	* @var integer $width
4e17e6	121	*/
66afd7	122	protected $width = 70;
4e17e6	123
T	124	/**
66afd7	125	* Target character encoding for output text
4e17e6	126	*
66afd7	127	* @var string $charset
4e17e6	128	*/
66afd7	129	protected $charset = 'UTF-8';
4e17e6	130
T	131	/**
66afd7	132	* List of preg* regular expression patterns to search for,
AM	133	* used in conjunction with $replace.
c72a96	134	*
66afd7	135	* @var array $search
AM	136	* @see $replace
c72a96	137	*/
66afd7	138	protected $search = array(
4e17e6	139	"/\r/", // Non-legal carriage return
T	140	"/[\n\t]+/", // Newlines and tabs
dc8f29	141	'/<head[^>]>.?<\/head>/i', // <head>
4e17e6	142	'/<script[^>]>.?<\/script>/i', // <script>s -- which strip_tags supposedly has problems with
8ac6fd	143	'/<style[^>]>.?<\/style>/i', // <style>s -- which strip_tags supposedly has problems with
4e17e6	144	'/<p[^>]*>/i', // <P>
T	145	'/<br[^>]*>/i', // <br>
8ac6fd	146	'/<i[^>]>(.?)<\/i>/i', // <i>
A	147	'/<em[^>]>(.?)<\/em>/i', // <em>
4e17e6	148	'/(<ul[^>]*>\|<\/ul>)/i', // <ul> and </ul>
T	149	'/(<ol[^>]*>\|<\/ol>)/i', // <ol> and </ol>
8ac6fd	150	'/<li[^>]>(.?)<\/li>/i', // <li> and </li>
4e17e6	151	'/<li[^>]*>/i', // <li>
T	152	'/<hr[^>]*>/i', // <hr>
f6b282	153	'/<div[^>]*>/i', // <div>
4e17e6	154	'/(<table[^>]*>\|<\/table>)/i', // <table> and </table>
T	155	'/(<tr[^>]*>\|<\/tr>)/i', // <tr> and </tr>
8ac6fd	156	'/<td[^>]>(.?)<\/td>/i', // <td> and </td>
4e17e6	157	);
T	158
	159	/**
66afd7	160	* List of pattern replacements corresponding to patterns searched.
4e17e6	161	*
66afd7	162	* @var array $replace
AM	163	* @see $search
4e17e6	164	*/
66afd7	165	protected $replace = array(
4e17e6	166	'', // Non-legal carriage return
T	167	' ', // Newlines and tabs
dc8f29	168	'', // <head>
4e17e6	169	'', // <script>s -- which strip_tags supposedly has problems with
8ac6fd	170	'', // <style>s -- which strip_tags supposedly has problems with
f6b282	171	"\n\n", // <P>
4e17e6	172	"\n", // <br>
T	173	'_\\1_', // <i>
8ac6fd	174	'_\\1_', // <em>
4e17e6	175	"\n\n", // <ul> and </ul>
T	176	"\n\n", // <ol> and </ol>
8ac6fd	177	"\t* \\1\n", // <li> and </li>
A	178	"\n\t* ", // <li>
6972cc	179	"\n-------------------------\n", // <hr>
11bcac	180	"<div>\n", // <div>
6972cc	181	"\n\n", // <table> and </table>
4e17e6	182	"\n", // <tr> and </tr>
T	183	"\t\t\\1\n", // <td> and </td>
ca0cd0	184	);
A	185
	186	/**
66afd7	187	* List of preg* regular expression patterns to search for,
AM	188	* used in conjunction with $ent_replace.
ca0cd0	189	*
66afd7	190	* @var array $ent_search
AM	191	* @see $ent_replace
ca0cd0	192	*/
66afd7	193	protected $ent_search = array(
ca0cd0	194	'/&(nbsp\|#160);/i', // Non-breaking space
A	195	'/&(quot\|rdquo\|ldquo\|#8220\|#8221\|#147\|#148);/i',
21d463	196	// Double quotes
ca0cd0	197	'/&(apos\|rsquo\|lsquo\|#8216\|#8217);/i', // Single quotes
A	198	'/>/i', // Greater-than
	199	'/</i', // Less-than
	200	'/&(copy\|#169);/i', // Copyright
	201	'/&(trade\|#8482\|#153);/i', // Trademark
	202	'/&(reg\|#174);/i', // Registered
	203	'/&(mdash\|#151\|#8212);/i', // mdash
	204	'/&(ndash\|minus\|#8211\|#8722);/i', // ndash
	205	'/&(bull\|#149\|#8226);/i', // Bullet
	206	'/&(pound\|#163);/i', // Pound sign
	207	'/&(euro\|#8364);/i', // Euro sign
	208	'/&(amp\|#38);/i', // Ampersand: see _converter()
	209	'/[ ]{2,}/', // Runs of spaces, post-handling
	210	);
	211
	212	/**
66afd7	213	* List of pattern replacements corresponding to patterns searched.
ca0cd0	214	*
66afd7	215	* @var array $ent_replace
AM	216	* @see $ent_search
ca0cd0	217	*/
66afd7	218	protected $ent_replace = array(
8ac6fd	219	' ', // Non-breaking space
A	220	'"', // Double quotes
	221	"'", // Single quotes
4e17e6	222	'>',
T	223	'<',
	224	'(c)',
	225	'(tm)',
	226	'(R)',
8ac6fd	227	'--',
A	228	'-',
4e17e6	229	'*',
300fc6	230	'£',
8ac6fd	231	'EUR', // Euro sign. ?
6084d7	232	'\|+\|amp\|+\|', // Ampersand: see _converter()
ca0cd0	233	' ', // Runs of spaces, post-handling
f50cc7	234	);
A	235
	236	/**
66afd7	237	* List of preg* regular expression patterns to search for
AM	238	* and replace using callback function.
f50cc7	239	*
66afd7	240	* @var array $callback_search
f50cc7	241	*/
66afd7	242	protected $callback_search = array(
8c1880	243	'/<(a) [^>]href=("\|\')([^"\']+)\2[^>]>(.*?)<\/a>/i', // <a href="">
AM	244	'/<(h)[123456]( [^>])?>(.?)<\/h[123456]>/i', // h1 - h6
	245	'/<(b)( [^>])?>(.?)<\/b>/i', // <b>
	246	'/<(strong)( [^>])?>(.?)<\/strong>/i', // <strong>
	247	'/<(th)( [^>])?>(.?)<\/th>/i', // <th> and </th>
7353fa	248	);
A	249
8ac6fd	250	/**
66afd7	251	* List of preg* regular expression patterns to search for in PRE body,
AM	252	* used in conjunction with $pre_replace.
8ac6fd	253	*
66afd7	254	* @var array $pre_search
AM	255	* @see $pre_replace
8ac6fd	256	*/
66afd7	257	protected $pre_search = array(
6972cc	258	"/\n/",
T	259	"/\t/",
	260	'/ /',
	261	'/<pre[^>]*>/',
	262	'/<\/pre>/'
7353fa	263	);
A	264
	265	/**
66afd7	266	* List of pattern replacements corresponding to patterns searched for PRE body.
7353fa	267	*
66afd7	268	* @var array $pre_replace
AM	269	* @see $pre_search
7353fa	270	*/
66afd7	271	protected $pre_replace = array(
6972cc	272	'<br>',
T	273	'    ',
	274	' ',
	275	'',
	276	''
4e17e6	277	);
T	278
	279	/**
66afd7	280	* Contains a list of HTML tags to allow in the resulting text.
4e17e6	281	*
66afd7	282	* @var string $allowed_tags
AM	283	* @see set_allowed_tags()
4e17e6	284	*/
66afd7	285	protected $allowed_tags = '';
4e17e6	286
T	287	/**
66afd7	288	* Contains the base URL that relative links should resolve to.
4e17e6	289	*
66afd7	290	* @var string $url
4e17e6	291	*/
66afd7	292	protected $url;
4e17e6	293
T	294	/**
66afd7	295	* Indicates whether content in the $html variable has been converted yet.
4e17e6	296	*
66afd7	297	* @var boolean $_converted
AM	298	* @see $html, $text
4e17e6	299	*/
66afd7	300	protected $_converted = false;
4e17e6	301
T	302	/**
66afd7	303	* Contains URL addresses from links to be rendered in plain text.
4e17e6	304	*
66afd7	305	* @var array $_link_list
AM	306	* @see _build_link_list()
4e17e6	307	*/
66afd7	308	protected $_link_list = array();
4e17e6	309
ca0cd0	310	/**
A	311	* Boolean flag, true if a table of link URLs should be listed after the text.
	312	*
	313	* @var boolean $_do_links
66afd7	314	* @see __construct()
e7f85b	315	*/
66afd7	316	protected $_do_links = true;
ca0cd0	317
4e17e6	318	/**
66afd7	319	* Constructor.
4e17e6	320	*
66afd7	321	* If the HTML source string (or file) is supplied, the class
AM	322	* will instantiate with that source propagated, all that has
	323	* to be done it to call get_text().
4e17e6	324	*
66afd7	325	* @param string $source HTML content
AM	326	* @param boolean $from_file Indicates $source is a file to pull content from
	327	* @param boolean $do_links Indicate whether a table of link URLs is desired
	328	* @param integer $width Maximum width of the formatted text, 0 for no limit
4e17e6	329	*/
66afd7	330	function __construct($source = '', $from_file = false, $do_links = true, $width = 75, $charset = 'UTF-8')
4e17e6	331	{
66afd7	332	if (!empty($source)) {
4e17e6	333	$this->set_html($source, $from_file);
T	334	}
6972cc	335
4e17e6	336	$this->set_base_url();
66afd7	337
6972cc	338	$this->_do_links = $do_links;
66afd7	339	$this->width = $width;
AM	340	$this->charset = $charset;
4e17e6	341	}
T	342
	343	/**
66afd7	344	* Loads source HTML into memory, either from $source string or a file.
4e17e6	345	*
66afd7	346	* @param string $source HTML content
AM	347	* @param boolean $from_file Indicates $source is a file to pull content from
4e17e6	348	*/
66afd7	349	function set_html($source, $from_file = false)
4e17e6	350	{
66afd7	351	if ($from_file && file_exists($source)) {
8c1880	352	$this->html = file_get_contents($source);
4e17e6	353	}
66afd7	354	else {
6972cc	355	$this->html = $source;
66afd7	356	}
4e17e6	357
T	358	$this->_converted = false;
	359	}
	360
	361	/**
66afd7	362	* Returns the text, converted from HTML.
4e17e6	363	*
66afd7	364	* @return string Plain text
4e17e6	365	*/
T	366	function get_text()
	367	{
66afd7	368	if (!$this->_converted) {
4e17e6	369	$this->_convert();
T	370	}
	371
	372	return $this->text;
	373	}
	374
	375	/**
66afd7	376	* Prints the text, converted from HTML.
4e17e6	377	*/
T	378	function print_text()
	379	{
	380	print $this->get_text();
	381	}
	382
	383	/**
66afd7	384	* Sets the allowed HTML tags to pass through to the resulting text.
4e17e6	385	*
66afd7	386	* Tags should be in the form "<p>", with no corresponding closing tag.
4e17e6	387	*/
66afd7	388	function set_allowed_tags($allowed_tags = '')
4e17e6	389	{
66afd7	390	if (!empty($allowed_tags)) {
4e17e6	391	$this->allowed_tags = $allowed_tags;
T	392	}
	393	}
	394
	395	/**
66afd7	396	* Sets a base URL to handle relative links.
4e17e6	397	*/
66afd7	398	function set_base_url($url = '')
4e17e6	399	{
66afd7	400	if (empty($url)) {
AM	401	if (!empty($_SERVER['HTTP_HOST'])) {
21d463	402	$this->url = 'http://' . $_SERVER['HTTP_HOST'];
66afd7	403	}
AM	404	else {
21d463	405	$this->url = '';
AM	406	}
66afd7	407	}
AM	408	else {
4e17e6	409	// Strip any trailing slashes for consistency (relative
T	410	// URLs may already start with a slash like "/file.html")
66afd7	411	if (substr($url, -1) == '/') {
4e17e6	412	$url = substr($url, 0, -1);
T	413	}
	414	$this->url = $url;
	415	}
	416	}
	417
	418	/**
66afd7	419	* Workhorse function that does actual conversion (calls _converter() method).
4e17e6	420	*/
66afd7	421	protected function _convert()
4e17e6	422	{
T	423	// Variables used for building the link list
43c40f	424	$this->_link_list = array();
4e17e6	425
T	426	$text = trim(stripslashes($this->html));
11bcac	427
A	428	// Convert HTML to TXT
	429	$this->_converter($text);
	430
	431	// Add link list
43c40f	432	if (!empty($this->_link_list)) {
A	433	$text .= "\n\nLinks:\n------\n";
	434	foreach ($this->_link_list as $idx => $url) {
	435	$text .= '[' . ($idx+1) . '] ' . $url . "\n";
	436	}
11bcac	437	}
A	438
66afd7	439	$this->text = $text;
11bcac	440	$this->_converted = true;
A	441	}
	442
	443	/**
66afd7	444	* Workhorse function that does actual conversion.
11bcac	445	*
66afd7	446	* First performs custom tag replacement specified by $search and
AM	447	* $replace arrays. Then strips any remaining HTML tags, reduces whitespace
	448	* and newlines to a readable format, and word wraps the text to
	449	* $width characters.
11bcac	450	*
66afd7	451	* @param string Reference to HTML content string
11bcac	452	*/
66afd7	453	protected function _converter(&$text)
11bcac	454	{
A	455	// Convert <BLOCKQUOTE> (before PRE!)
	456	$this->_convert_blockquotes($text);
4e17e6	457
6972cc	458	// Convert <PRE>
8ac6fd	459	$this->_convert_pre($text);
300fc6	460
ca0cd0	461	// Run our defined tags search-and-replace
4e17e6	462	$text = preg_replace($this->search, $this->replace, $text);
ca0cd0	463
A	464	// Run our defined tags search-and-replace with callback
66afd7	465	$text = preg_replace_callback($this->callback_search, array($this, 'tags_preg_callback'), $text);
ca0cd0	466
A	467	// Strip any other HTML tags
	468	$text = strip_tags($text, $this->allowed_tags);
	469
	470	// Run our defined entities/characters search-and-replace
	471	$text = preg_replace($this->ent_search, $this->ent_replace, $text);
4e17e6	472
6972cc	473	// Replace known html entities
c72a96	474	$text = html_entity_decode($text, ENT_QUOTES, $this->charset);
4e0419	475
755900	476	// Remove unknown/unhandled entities (this cannot be done in search-and-replace block)
6084d7	477	$text = preg_replace('/&([a-zA-Z0-9]{2,6}\|#[0-9]{2,4});/', '', $text);
A	478
	479	// Convert "\|+\|amp\|+\|" into "&", need to be done after handling of unknown entities
	480	// This properly handles situation of "&quot;" in input string
	481	$text = str_replace('\|+\|amp\|+\|', '&', $text);
755900	482
4e17e6	483	// Bring down number of empty lines to 2 max
8ac6fd	484	$text = preg_replace("/\n\s+\n/", "\n\n", $text);
4e17e6	485	$text = preg_replace("/[\n]{3,}/", "\n\n", $text);
T	486
4d7fbd	487	// remove leading empty lines (can be produced by eg. P tag on the beginning)
ca0cd0	488	$text = ltrim($text, "\n");
4d7fbd	489
4e17e6	490	// Wrap the text to a readable format
T	491	// for PHP versions >= 4.0.2. Default width is 75
8ac6fd	492	// If width is 0 or less, don't wrap the text.
A	493	if ( $this->width > 0 ) {
21d463	494	$text = wordwrap($text, $this->width);
8ac6fd	495	}
4e17e6	496	}
T	497
	498	/**
66afd7	499	* Helper function called by preg_replace() on link replacement.
4e17e6	500	*
66afd7	501	* Maintains an internal list of links to be displayed at the end of the
AM	502	* text, with numeric indices to the original point in the text they
	503	* appeared. Also makes an effort at identifying and handling absolute
	504	* and relative links.
4e17e6	505	*
66afd7	506	* @param string $link URL of the link
AM	507	* @param string $display Part of the text to associate number with
8ac6fd	508	*/
66afd7	509	protected function _build_link_list( $link, $display )
8ac6fd	510	{
21d463	511	if (!$this->_do_links \|\| empty($link)) {
AM	512	return $display;
4e17e6	513	}
T	514
21d463	515	// Ignored link types
AM	516	if (preg_match('!^(javascript:\|mailto:\|#)!i', $link)) {
	517	return $display;
	518	}
	519
	520	if (preg_match('!^([a-z][a-z0-9.+-]+:)!i', $link)) {
43c40f	521	$url = $link;
A	522	}
	523	else {
	524	$url = $this->url;
	525	if (substr($link, 0, 1) != '/') {
	526	$url .= '/';
	527	}
	528	$url .= "$link";
	529	}
	530
	531	if (($index = array_search($url, $this->_link_list)) === false) {
	532	$index = count($this->_link_list);
8c1880	533	$this->_link_list[] = $url;
43c40f	534	}
A	535
	536	return $display . ' [' . ($index+1) . ']';
8ac6fd	537	}
11bcac	538
7353fa	539	/**
66afd7	540	* Helper function for PRE body conversion.
7353fa	541	*
66afd7	542	* @param string HTML content
8ac6fd	543	*/
66afd7	544	protected function _convert_pre(&$text)
8ac6fd	545	{
d483cd	546	// get the content of PRE element
11bcac	547	while (preg_match('/<pre[^>]>(.)<\/pre>/ismU', $text, $matches)) {
8c1880	548	$this->pre_content = $matches[1];
AM	549
	550	// Run our defined tags search-and-replace with callback
	551	$this->pre_content = preg_replace_callback($this->callback_search,
66afd7	552	array($this, 'tags_preg_callback'), $this->pre_content);
8c1880	553
d483cd	554	// convert the content
A	555	$this->pre_content = sprintf('<div><br>%s<br></div>',
8c1880	556	preg_replace($this->pre_search, $this->pre_replace, $this->pre_content));
AM	557
d483cd	558	// replace the content (use callback because content can contain $0 variable)
8c1880	559	$text = preg_replace_callback('/<pre[^>]>.<\/pre>/ismU',
66afd7	560	array($this, 'pre_preg_callback'), $text, 1);
8c1880	561
d483cd	562	// free memory
A	563	$this->pre_content = '';
11bcac	564	}
A	565	}
	566
	567	/**
66afd7	568	* Helper function for BLOCKQUOTE body conversion.
11bcac	569	*
66afd7	570	* @param string HTML content
11bcac	571	*/
66afd7	572	protected function _convert_blockquotes(&$text)
11bcac	573	{
bb6f4b	574	$level = 0;
TB	575	$offset = 0;
	576	while (($start = strpos($text, '<blockquote', $offset)) !== false) {
	577	$offset = $start + 12;
	578	do {
	579	$end = strpos($text, '</blockquote>', $offset);
	580	$next = strpos($text, '<blockquote', $offset);
	581
	582	// nested <blockquote>, skip
	583	if ($next !== false && $next < $end) {
	584	$offset = $next + 12;
	585	$level++;
	586	}
	587	// nested </blockquote> tag
	588	if ($end !== false && $level > 0) {
	589	$offset = $end + 12;
11bcac	590	$level--;
A	591	}
bb6f4b	592	// found matching end tag
TB	593	else if ($end !== false && $level == 0) {
	594	$taglen = strpos($text, '>', $start) - $start;
	595	$startpos = $start + $taglen + 1;
	596
	597	// get blockquote content
	598	$body = trim(substr($text, $startpos, $end - $startpos));
	599
737b62	600	// adjust text wrapping width
TB	601	$p_width = $this->width;
	602	if ($this->width > 0) $this->width -= 2;
	603
bb6f4b	604	// replace content with inner blockquotes
TB	605	$this->_converter($body);
	606
737b62	607	// resore text width
TB	608	$this->width = $p_width;
	609
bb6f4b	610	// Add citation markers and create <pre> block
TB	611	$body = preg_replace_callback('/((?:^\|\n)>)([^\n])/', array($this, 'blockquote_citation_ballback'), trim($body));
	612	$body = '<pre>' . htmlspecialchars($body) . '</pre>';
	613
ff6de9	614	$text = substr_replace($text, $body . "\n", $start, $end + 13 - $start);
bb6f4b	615	$offset = 0;
ff6de9	616
bb6f4b	617	break;
11bcac	618	}
ff6de9	619	}
AM	620	while ($end \|\| $next);
6972cc	621	}
8ac6fd	622	}
f50cc7	623
A	624	/**
bb6f4b	625	* Callback function to correctly add citation markers for blockquote contents
TB	626	*/
	627	public function blockquote_citation_ballback($m)
	628	{
ff6de9	629	$line = ltrim($m[2]);
bb6f4b	630	$space = $line[0] == '>' ? '' : ' ';
ff6de9	631
bb6f4b	632	return $m[1] . '>' . $space . $line;
TB	633	}
	634
	635	/**
66afd7	636	* Callback function for preg_replace_callback use.
f50cc7	637	*
66afd7	638	* @param array PREG matches
AM	639	* @return string
f50cc7	640	*/
66afd7	641	public function tags_preg_callback($matches)
f50cc7	642	{
2d7b4f	643	switch (strtolower($matches[1])) {
6972cc	644	case 'b':
T	645	case 'strong':
8c1880	646	return $this->_toupper($matches[3]);
da1722	647	case 'th':
8c1880	648	return $this->_toupper("\t\t". $matches[3] ."\n");
6972cc	649	case 'h':
8c1880	650	return $this->_toupper("\n\n". $matches[3] ."\n\n");
6972cc	651	case 'a':
29c542	652	// Remove spaces in URL (#1487805)
A	653	$url = str_replace(' ', '', $matches[3]);
	654	return $this->_build_link_list($url, $matches[4]);
6972cc	655	}
f50cc7	656	}
29c542	657
f50cc7	658	/**
66afd7	659	* Callback function for preg_replace_callback use in PRE content handler.
d483cd	660	*
66afd7	661	* @param array PREG matches
AM	662	* @return string
d483cd	663	*/
66afd7	664	public function pre_preg_callback($matches)
d483cd	665	{
A	666	return $this->pre_content;
	667	}
	668
	669	/**
f35995	670	* Strtoupper function with HTML tags and entities handling.
f50cc7	671	*
f35995	672	* @param string $str Text to convert
A	673	* @return string Converted text
	674	*/
	675	private function _toupper($str)
	676	{
	677	// string can containg HTML tags
	678	$chunks = preg_split('/(<[^>]*>)/', $str, null, PREG_SPLIT_NO_EMPTY \| PREG_SPLIT_DELIM_CAPTURE);
	679
	680	// convert toupper only the text between HTML tags
	681	foreach ($chunks as $idx => $chunk) {
	682	if ($chunk[0] != '<') {
	683	$chunks[$idx] = $this->_strtoupper($chunk);
	684	}
	685	}
	686
	687	return implode($chunks);
	688	}
	689
	690	/**
	691	* Strtoupper multibyte wrapper function with HTML entities handling.
	692	*
	693	* @param string $str Text to convert
	694	* @return string Converted text
f50cc7	695	*/
d483cd	696	private function _strtoupper($str)
f50cc7	697	{
c72a96	698	$str = html_entity_decode($str, ENT_COMPAT, $this->charset);
66afd7	699	$str = mb_strtoupper($str);
c72a96	700	$str = htmlspecialchars($str, ENT_COMPAT, $this->charset);
67e592	701
A	702	return $str;
f50cc7	703	}
4e17e6	704	}