The class EventManager defines a singleton, which you access with the EventManager::getInstance() expression.

Here is the logic:

1. bind a handler to an event: EventManager::getInstance()->bind( ‘event’, ‘handler’ )
   An event can be a string or a PHP regular expression. A handler must be a PHP callable expression.
2. trigger an event: EventManager::getInstance()->trigger( ‘event’ )
   A triggered event should be  a string. That string is going to be matched against all the bound events. Whenever a match is found, all the handlers bound to the matching events will be called. All the arguments passed to trigger will be passed along to the handlers. If a handler returns a false value, all other handlers won’t be called.

class EventManager {
    
    /**
     * Singleton setup
     */
	static protected $_instance = null;
	protected function __construct() {}
	protected function __clone() {}
	static public function getInstance() {
		if (is_null( self::$_instance )) {
		    $class = __CLASS__;
			self::$_instance = new $class();
		}
		return self::$_instance;
	}
	
	
	/**
	 * Stores bindings between events and their handlers
	 *
	 * @var array
	 */
	protected $_bindings = array();
	
	
	/**
	 * Returns the id of the given $handler
	 *
	 * @param $handler a callable expression
	 * @return string
	 */
	protected function getId( $handler ) {
	    if (! is_callable( $handler )) {
	        throw new Exception('Expected a callable expression');
	    }
	    if (is_string( $handler )) {
	        $id = "function: $handler";
	    } else {
	        $container = $hadler[0];
	        $method = $handler[1];
	        if (is_string( $container )) {
	            $class = $container;
	            $id = "class: $class, method: $method";
	        } else {
	            $class = get_class( $container );
	            $object = spl_object_hash( $container );
	            $id = "class: $class, object: $object, method: $method";
	        }
	    }
	    return $id;
	}
	
	
	/**
	 * Returns TRUE if the given $expression is a regular expression in PHP
	 * TODO improve EventManager::isRegExp by allowing matching delimiters (), [], {}, <>
	 *
	 * @param $regex
	 * @return boolean
	 */
	protected function isRegExp( $expression ) {
	    if (is_string( $expression ) && preg_match( '/([^\w \\])[^\1\\\r\n]*(?:\\.[^\1\\\r\n]*)*\1/', $expression )) {
	        return true;
	    }
	    return false;
	}
	
	
	/**
	 * Binds a $handler to an $event
	 *
	 * @param string $event
	 * @param callable $handler
	 * @return EventManager
	 */
	public function bind( $event, $handler ) {
	    $id = $this->getId( $handler );
	    $this->_bindings[ $event ][ $id ] = $handler;
		return $this;
	}
	
	
	/**
	 * Unbinds a $handler from an $event
	 *
	 * @param string $event
	 * @param callable $handler
	 * @return EventManager
	 */
	public function unbind( $event, $handler = null ) {
		if (is_null( $handler )) {
			unset( $this->_bindings[ $event ] );
		} else {
		    $id = $this->getId( $handler );
    	    unset( $this->_bindings[ $event ][ $id ] );
		}
		return $this;
	}
	
	
	/**
	 * Dispatches a $triggered event to all its matching handlers
	 *
	 * @param string $triggered
	 * @return EventManager
	 */
	public function trigger( $triggered ) {
	    $result = true;
	    $args = func_get_args();
	    foreach ($this->_bindings as $event => $handlers) {
	        if ($this->isRegExp( $event )) {
    	        if (! preg_match( $event, $triggered )) {
    	            continue;
    	        }
	        } else {
	            if ($event != $triggered) {
	                continue;
	            }
	        }
    	    if (is_array( $handlers )) {
    			foreach ($handlers as $handler) {
    			    if (is_callable( $handler )) {
    			        $result = call_user_func_array( $handler, $args );
        				if (false === $result) {
        					break;
        				}
    			    }
    			}
    		}
    		if (false === $result) {
				break;
			}
	    }
		return $this;
	}
	
}

Shouldn’t it be null is null and is not an object ?

• UPDATED 2009-10-11

Solution – If you can’t beat them, join them. You can show a progress indicator to let your web users know that your server is thinking hard the answer to their request, but how do you do that?

Requirement 1 – You cannot put the progress indicator on the leaving page, because you don’t want to modify every single reference to a slow page. It doesn’t make sense either: a slow page is slow by itself! You really don’t need that coupling.

Requirement 2 - You cannot put the progress indicator on the landing page, the slow one, because the browser needs it before the slow page has arrived. If they arrive together, the progress indicator is useless!

There is a good article about this issue but it fails at implementing the first requirement. Anyway that was the idea I worked on when starting this project.

Here is the waiting page that get’s loaded while loading the slow page. For browsers other than IE, the trick is to change the document.location.href property after loading the progress indicator image. In IE we change the document.location.href and then add the image.

I used jQuery to make it simple.

waiting.php

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "DTD/xhtml1-transitional.dtd">

<html xml:lang="en" lang="en" xmlns="http://www.w3.org/1999/xhtml"><head>

<title>Loading Page...</title>

<script type="text/javascript" src="http://ajax.googleapis.com/ajax/libs/jquery/1.3/jquery.min.js"></script>
<script type="text/javascript">
jQuery( function( $ ) {
	function redirect() {
		$('form').submit();
	}
	
	if($.browser.msie) {
		redirect();
		$('img').appendTo('body').attr('src', 'loading-page.gif');
	} else {
		$('#loading').bind('load', function() {
	        redirect();
	    }).attr('src', 'loading-page.gif');
	}
} );
</script>

</head><body>

<h1>Loading Page...</h1>
<img id="loading" />
<form method="<?php echo $method ?>" action="<?php echo $url ?>">
 <input type="hidden" name="waiting" value="<?php echo $waiting ?>" />
</form>

</body></html>

And here is the slow page.

slow-page.php

<?php

if ( ! isset($_REQUEST['waiting']) ) {
    $url = htmlspecialchars( $_SERVER['REQUEST_URI'] );
	$method = $_SERVER['REQUEST_METHOD'];
	$waiting = time();
	include "waiting.php";
    return;
}

sleep(10); //simulate a slow task

?>

<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "DTD/xhtml1-transitional.dtd">

<html xml:lang="en" lang="en" xmlns="http://www.w3.org/1999/xhtml"><head>
 <title>Slow Page</title>
</head><body>
 <h1>Hello World!</h1>
</body></html>

You can test it here.

I was experimenting some issues with Toad for MySQL, totally unrelated by the way and still to be solved, so I was sensitive to malfunctions, and decided to go through it.

The problem is that WampServer 2.0f installs MySQL 5.1.30 and PHP 5.2.8, which only supports MySQL 5.0.51a. So the only available solution, if you  want to stay with the last stable PHP version, is to downgrade MySQL to version 5.0.51a.

You can download and install a MySQL add on, by using the menu option WampServer/MySQL/Version/Get more… Unfortunately, the installation does not end when the installer finishes. In fact you’ll need to manually configure your new database server.

Switch to it by using the menu option WampServer/MySQL/Version/5.0.51a, wait for the WampServer status icon to get back to white, and reload phpMyAdmin. Works? Mine didn’t. The problem is that the new MySQL server is still not configured and very empty.

• Switch back to the first server (5.1.30) and  export all your databases (don’t export information_schema nor mysql). Then switch to the second server (5.0.51a) for the last time.
• Edit the file wamp/apps/phpmyadmin3.1.1/config.inc.php and remove the password.
• Delete cookies and cache of the browser. This should be enough for allowing phpMyAdmin to load again.
• Go to the Priviliges tab, and remove the users rows with production.mysql.com in them: they are garbage of this MySQL version.
• Edit the user root/localhost and set a password.
• Reload phpMyAdmin and check that it fails.
• Edit again the file wamp/apps/phpmyadmin3.1.1/config.inc.php and insert the password.
• Reload phpMyAdmin and check that it doesn’t fail.
• Add new users/databases as needed.
• Select each database and import its schema and data, from your previous exports.

Done.

$('body').width( $('body').width() - 1 ).width( $('body').width() + 1 );

If you want to match a large text, these expressions are not the best. I’ve prepared a simple test page where the GeSHi’s engine file, which is almost 120KB, is going to be matched by the regular expression you input.

In Firefox 2 the performance is quite good, about 100 ms on my PC, but in Internet Explorer 7 it takes more than 7 minutes !!!

The best catch-all regular expression is [\w\W]*, which employs about 50 ms in FF2, and 0 ms in IE7 !!! (yes, zero milliseconds)

This bug affects content and custom fields, which means that it’s more of a conceptual bug than a coding bug. I wrote an article with a workaround for the content. Now, let’s see a simple fix for custom fields.

I’ll describe both problems I detected and give a solution based on changing some WordPress code. I know it’s not a perfect solution, but it works and it’s also relatively easy. In the snippets you’ll see the code as it’s supposed to get changed to. (with 3 lines of context before and after changed lines, each preceded by a //noteslog.com comment)

WordPress Version: 2.3.1

Problem: HTML Entities Conversion

Custom fields that have HTML entities in their key or value are treated by WordPress weirdly. Frustration will appear soon after realizing that WordPress won’t return what you put in before.

If I write ¶ inside the visual editor of a post, WordPress should assume that I want to visualize ¶. In fact it will pretend to do the right thing the first time I save the post. All subsequent times it will show a ¶ where I put a ¶.

If I write ¶ inside the code editor of a post, WordPress should assume that I want to visualize ¶, and it does the right thing (sort of, because I now have a ¶ inside the code too).

If I write ¶ inside a custom field’s key or value, WordPress should assume that I want to visualize ¶ for my readers, and ¶ for myself when I’m authoring the custom field. In fact WordPress does the wrong thing here again, because it’ll immediately convert ¶ to ¶.

Solution

This fix requires changing four lines in two files. We’ll use the PHP function htmlspecialchars in place of the WordPress function attribute_escape. After the fix, filters for “attribute_escape” won’t get applied to custom fields keys and values. (this shouldn’t be a problem)

Open: wp-admin/admin-ajax.php
Search: function wp_ajax_meta_row

add_action( 'shutdown', 'get_out_now', -1 );

function wp_ajax_meta_row( $pid, $mid, $key, $value ) {
//noteslog.com
	$value = htmlspecialchars( $value ); //attribute_escape($value);
	$key_js = addslashes(wp_specialchars($key, 'double'));
//noteslog.com
	$key = htmlspecialchars( $key ); //attribute_escape($key);
	$r .= "<tr id='meta-$mid'><td valign='top'>";
	$r .= "<input name='meta[$mid][key]' tabindex='6' onkeypress='return killSubmit(\"theList.ajaxUpdater(&#039;meta&#039;,&#039;meta-$mid&#039;);\",event);' type='text' size='20' value='$key' />";
	$r .= "</td><td><textarea name='meta[$mid][value]' tabindex='6' rows='2' cols='30'>$value</textarea></td><td align='center'>";

Open: wp-admin/includes/template.php
Search: function list_meta

}

		$key_js = js_escape( $entry['meta_key'] );
//noteslog.com
		$entry['meta_key']   = htmlspecialchars($entry['meta_key']);//attribute_escape($entry['meta_key']);
//noteslog.com
		$entry['meta_value'] = htmlspecialchars($entry['meta_value']);//attribute_escape($entry['meta_value']);
		$entry['meta_id'] = (int) $entry['meta_id'];
		$r .= "\n\t<tr id='meta-{$entry['meta_id']}' class='$style'>";
		$r .= "\n\t\t<td valign='top'><input name='meta[{$entry['meta_id']}][key]' tabindex='6' type='text' size='20' value='{$entry['meta_key']}' /></td>";

Problem: White Space Trimming

Custom fields that begin or end with white space in their key or value are trimmed. I can undestand that a key is more easily dealt with if it is trimmed before saving it into the database. But values should definitely retain all their white space AS IS.

Solution

This fix requires changing four lines in three files. We’ll add a $trim parameter to the wordpress function maybe_serialize, by default set to true. Wherever we do not want a $data value trimmed, we’ll call maybe_serialize with a false second argument.

Open: wp-includes/functions.php
Search: function maybe_serialize

return true;
}

//noteslog.com
function maybe_serialize($data, $trim=true) { //($data)
//noteslog.com
	if ( $trim && is_string($data) ) //( is_string($data) )
		$data = trim($data);
	elseif ( is_array($data) || is_object($data) )
		return serialize($data);

Open: wp-admin/includes/post.php
Search: function add_meta

$metakeyselect = $wpdb->escape( stripslashes( trim( $_POST['metakeyselect'] ) ) );
	$metakeyinput = $wpdb->escape( stripslashes( trim( $_POST['metakeyinput'] ) ) );
//noteslog.com
	$metavalue = maybe_serialize( stripslashes( ( $_POST['metavalue'] ) ), false ); //( stripslashes( trim( $_POST['metavalue'] ) ) );
	$metavalue = $wpdb->escape( $metavalue );

	if ( ('0' === $metavalue || !empty ( $metavalue ) ) && ((('#NONE#' != $metakeyselect) && !empty ( $metakeyselect) ) || !empty ( $metakeyinput) ) ) {
		// We have a key/value pair. If both the select and the

Open: wp-admin/includes/post.php
Search: function update_meta

if ( in_array($mkey, $protected) )
		return false;

//noteslog.com
	$mvalue = maybe_serialize( stripslashes( $mvalue ), false ); //( stripslashes( $mvalue ) );
	$mvalue = $wpdb->escape( $mvalue );
	$mid = (int) $mid;
	return $wpdb->query( "UPDATE $wpdb->postmeta SET meta_key = '$mkey', meta_value = '$mvalue' WHERE meta_id = '$mid'" );

After quite a long session of R&D, I’ve got to a pretty good solution, in the form of a jQuery plugin: jquery.wresize.js

/* =============================================================================== WResize is the jQuery plugin for fixing the IE window resize bug ............................................................................... Copyright 2007 / Andrea Ercolino ------------------------------------------------------------------------------- LICENSE: http://www.opensource.org/licenses/mit-license.php WEBSITE: http://noteslog.com/ =============================================================================== */ ( function( $ ) { $.fn.wresize = function( f ) { version = '1.1'; wresize = {fired: false, width: 0}; function resizeOnce() { if ( $.browser.msie ) { if ( ! wresize.fired ) { wresize.fired = true; } else { var version = parseInt( $.browser.version, 10 ); wresize.fired = false; if ( version < 7 ) { return false; } else if ( version == 7 ) { //a vertical resize is fired once, an horizontal resize twice var width = $( window ).width(); if ( width != wresize.width ) { wresize.width = width; return false; } } } } return true; } function handleWResize( e ) { if ( resizeOnce() ) { return f.apply(this, [e]); } } this.each( function() { if ( this == window ) { $( this ).resize( handleWResize ); } else { $( this ).resize( f ); } } ); return this; }; } ) ( jQuery );

If you want to try it, here is a test page, where a div is automatically resized when the window is resized.

References:

http://dev.jquery.com
http://snook.ca
http://ecmascript.stchur.com

Internet Explorer and Firefox accept such a giant, but Opera does not. No way. It’s a problem at the core level. So the only solution, very simple though, is not to use a literal expression: use new RegExp( "..." ); instead of /.../.

Remember to escape the expression for strings: for my recipe it was very simple because the step only contains a long list of words, between two word boundaries, which need an extra backslash.

If you want to write HTML entity codes in a post, and you need to represent that of an ampersand for example (it’s &amp;), then there is no way to get it right. In fact, WordPress will always resolve an entity code.

Thus, if you write &amp;, you’ll get & after the first save roundtrip, and if you try to escape the & into &amp; with &amp;, obtaining &amp;amp; (this looks weird but it’s the way to do it in plain HTML), then the first time you save the post you get &amp; back, and the second time you save the post you get & back.

In general if you write & followed by any number of amp; (like &amp;amp;amp;amp;amp;amp;amp;) then WordPress will make the & eat up an amp; at each saving roundtrip, hence the pacman effect.

But the title of my previous post about this issue was “the pacman effect strikes back.” In fact it’s not only a problem of the post content but also of the custom fields, thus corrupting the last resort we WordPress bloggers have for content AS IS. And this is where I get really upset.

Last tuesday I found the culprit and asked the wp-hackers list wether they considered it a bug or not. I’m still waiting for an answer, so I hope this post will help me to broaden the question and understand if I need to submit it to WordPress Trac or go in for a hack myself.