In this case I found interesting technique how to overcome this.

The idea is simple and is borrowed from official documentation comments. In PHP it is possible to sent browser output and continue to execute other PHP code.

This means that on each page request, PHP generates full HTML, sends to the user and closes the connection. User is happy to view the page, however on the server PHP script may execute useful stuff.

Code looks like this:

ob_start();
 
// Generage HTML page here
generate_full_html_page();
 
// All magic goes here
$output = ob_get_clean();
ignore_user_abort(true);
set_time_limit(0);
header("Connection: close");
header("Content-Length: ".strlen($output));
header("Content-Encoding: none");
echo $output;
flush();
 
// Now page is sent and it safe to do all needed stuff here
cron_task1();
cron_task2();

If you need to run CRON jobs continuously and not to depend on the user visits, you may initiate a SELF request to the same page using CURL after all tasks are done.

Based on this technique very easy to make a full CRON class which will handle all the dirty job, may be in future )

Tweet

I did research several times about it and found very few information.

RSS merge for WordPress

Very few plugins exists for WordPress and they do not work with recent 3.X versions. Unfortunately.

Yahoo pipes

Yahoo pipes is a very flexible and feature rich tool for managing RSS feeds. It is possible fetch, merge, split, filter for any RSS source. It has nice graphical interface for all this.

But I think it is a little but overkill for my two feeds.

PHP Solution for mixing RSS feeds

I did not want to bother with PHP coding for such a simple task. But if to compare actual time for coding it is far less then trying different plugins and research.

So I found somewhere on the Internet these two nice libraries:

• magpierss - for fetching feeds
• feedcreator.class.php - for publishing feeds

With a small PHP glue between them I created simple script for merging several RSS feeds into one. Below is simple script that does the job.

You can download full package with libraries included below in the end of post.

<?php
/**
 * Setup part
 * 
 */
$DOMAIN_NAME = 'http://a32.me/';
$FEED_URL = $DOMAIN_NAME . 'rss/full.php';
$SITE_TITLE = 'a32.blog by Constantin Bosneaga';
$SITE_DESRIPTION = 'Its all about. IT and "stuff" :)';
$SITE_AUTHOR = 'Constantin Bosneaga';
 
$RSS_CACHE = "/tmp/rsscache";
$RSS_CACHE_EXP = 7200;
 
$FEED_LIST = array(
	'http://a32.me/feed/',
	'http://a32.me/feed/?lang=ru'
);
 
 
/**
 * Do not modify below this point
 * 
 */
define('MAGPIE_CACHE_ON', true);
define('MAGPIE_CACHE_DIR', $RSS_CACHE);
define('MAGPIE_CACHE_AGE', $RSS_CACHE_EXP);
define('MAGPIE_OUTPUT_ENCODING', 'utf-8');
 
 
// include required files 
require_once ('magpierss-0.72/rss_fetch.inc');
include ('feedcreator.class.php');
 
/* Set RSS properties */
$rss = new UniversalFeedCreator();
$rss->useCached();
$rss->title = $SITE_TITLE;
$rss->description = $SITE_DESRIPTION;
$rss->link = $DOMAIN_NAME;
$rss->syndicationURL = $FEED_URL;
$rss->encoding = 'utf8';
 
/* Set Image properties 
$image = new FeedImage();
$image->title = $SITE_TITLE . " Logo";
$image->url = $SITE_LOG_URL;
$image->link = $DOMAIN_NAME;
$image->description = "Feed provided by " . $SITE_TITLE . ". Click to visit.";
$rss->image = $image;
*/
 
function showSummary($url, $num = 10, $showfullfeed = false) {
	global $rss, $DOMAIN_NAME, $SITE_AUTHOR, $SITE_TITLE;
	$num_items = $num;
	@ $rss1 = fetch_rss($url);
	if ($rss1) {
		$items = array_slice($rss1->items, 0, $num_items);
		foreach ($items as $item) {
			$href = $item['link'];
			$title = $item['title'];
			if (!$showfullfeed) {
				$desc = $item['description'];
			} else {
				$desc = $item['content']['encoded'];
			}
			//                $desc .=  '
			//Copyright &copy; <a href="'.$DOMAIN_NAME.'">'.$SITE_TITLE.'</a>.  All Rights Reserved.
			//';
			$pdate = $item['pubdate'];
			$rss_item = new FeedItem();
			$rss_item->title = $item['title'];
			$rss_item->link = $item['link'];
			$rss_item->description = $item['content']['encoded'];
			$rss_item->date = $item['pubdate'];
			$rss_item->source = $DOMAIN_NAME;
			$rss_item->author = $SITE_AUTHOR;
			$rss->addItem($rss_item);
		}
 
	} else {
		echo "Error: Cannot fetch feed url - " . $url;
	}
}
 
// Fetch all feeds
foreach($FEED_LIST as $v) showSummary($v);
 
// Sort items by date
function __usort($ad, $bd) {return strtotime($bd->date) - strtotime($ad->date);}
usort($rss->items, '__usort');
 
// Display items
$rss->saveFeed("RSS1.0", $RSS_CACHE . "/feed.xml");
 

<?php
/**
 * Setup part
 * 
 */
$DOMAIN_NAME = 'http://a32.me/';
$FEED_URL = $DOMAIN_NAME . 'rss/full.php';
$SITE_TITLE = 'a32.blog by Constantin Bosneaga';
$SITE_DESRIPTION = 'Its all about. IT and "stuff" :)';
$SITE_AUTHOR = 'Constantin Bosneaga';
 
$RSS_CACHE = "/tmp/rsscache";
$RSS_CACHE_EXP = 7200;
 
$FEED_LIST = array(
	'http://a32.me/feed/',
	'http://a32.me/feed/?lang=ru'
);
 
 
/**
 * Do not modify below this point
 * 
 */
define('MAGPIE_CACHE_ON', true);
define('MAGPIE_CACHE_DIR', $RSS_CACHE);
define('MAGPIE_CACHE_AGE', $RSS_CACHE_EXP);
define('MAGPIE_OUTPUT_ENCODING', 'utf-8');
 
 
// include required files 
require_once ('magpierss-0.72/rss_fetch.inc');
include ('feedcreator.class.php');
 
/* Set RSS properties */
$rss = new UniversalFeedCreator();
$rss->useCached();
$rss->title = $SITE_TITLE;
$rss->description = $SITE_DESRIPTION;
$rss->link = $DOMAIN_NAME;
$rss->syndicationURL = $FEED_URL;
$rss->encoding = 'utf8';
 
/* Set Image properties 
$image = new FeedImage();
$image->title = $SITE_TITLE . " Logo";
$image->url = $SITE_LOG_URL;
$image->link = $DOMAIN_NAME;
$image->description = "Feed provided by " . $SITE_TITLE . ". Click to visit.";
$rss->image = $image;
*/
 
function showSummary($url, $num = 10, $showfullfeed = false) {
	global $rss, $DOMAIN_NAME, $SITE_AUTHOR, $SITE_TITLE;
	$num_items = $num;
	@ $rss1 = fetch_rss($url);
	if ($rss1) {
		$items = array_slice($rss1->items, 0, $num_items);
		foreach ($items as $item) {
			$href = $item['link'];
			$title = $item['title'];
			if (!$showfullfeed) {
				$desc = $item['description'];
			} else {
				$desc = $item['content']['encoded'];
			}
			//                $desc .=  '
			//Copyright &copy; <a href="'.$DOMAIN_NAME.'">'.$SITE_TITLE.'</a>.  All Rights Reserved.
			//';
			$pdate = $item['pubdate'];
			$rss_item = new FeedItem();
			$rss_item->title = $item['title'];
			$rss_item->link = $item['link'];
			$rss_item->description = $item['content']['encoded'];
			$rss_item->date = $item['pubdate'];
			$rss_item->source = $DOMAIN_NAME;
			$rss_item->author = $SITE_AUTHOR;
			$rss->addItem($rss_item);
		}
 
	} else {
		echo "Error: Cannot fetch feed url - " . $url;
	}
}
 
// Fetch all feeds
foreach($FEED_LIST as $v) showSummary($v);
 
// Sort items by date
function __usort($ad, $bd) {return strtotime($bd->date) - strtotime($ad->date);}
usort($rss->items, '__usort');
 
// Display items
$rss->saveFeed("RSS1.0", $RSS_CACHE . "/feed.xml");
 

Download RSS feed mixer/combiner

Download RSS feed mixer/combiner. Unpack and change the Setup part of the script.

Vote for wordpress plugin

If you like to see these functionality done as WordPress plugiun, please leave your vote in comments!

Tweet

History API basics

History APS bases on the single DOM interface - History object. Each browser tab has its own unique History object, which is accessible  in window.history.

History has a number of methods, events and properties that we can control from JavaScript. Each page of the browser tab (Document object) is a collection of History objects.

Each element of the history consist of URL, state of the object (state object), can have a title, document object, form data, scroll position and other information related to the page.

The main methods of the History object are:

1. window.history.length: Number of entries in the current session history
2. window.history.state: Returns current history object
3. window.history.go(n): Method allows to walk through history. The argument passed is the offset the current position. If you pass 0, the current page will be updated. If the index goes beyond the history, nothing happens.
4. window.history.back(): The method is identical to calling go(-1)
5. window.history.forward(): The method is identical to calling go(1)
6. window.history.pushState(data, title [, url]): Adds an element to the history
7. window.history.replaceState(data, title [, url]): Updates the current element of history

To go two steps back in history you can use:

history.go(-2)

To add elements of history, you can use history.pushState:

history.pushState({foo: 'bar'}, 'Title', '/baz.html')

To change the history records we can use history.replaceState:

history.replaceState({foo: 'bat'}, 'New Title')

Live example

Now we know the basics, let's look at a live example. We will do a web file manager, which allows you to find the URI of selected image (look what you get in the end). The File Manager uses a simple file structure that is written in JavaScript. When you select a picture file or folder is updated dynamically.

We use data-* attributes to store the header of each image and use the property dataset for this property:

<li class="photo"> <a href="crab2.png" data-note="Grey crab!">crab2.png</a> </li>

To get everything working quickly we load all the pictures and update the src attribute dynamically. But, this acceleration creates a problem - it breaks the back button, so you can not go by pictures forward or backward.

HTML5 history comes to the rescue!

Every time we choose a file new history record is created and document location is updated(it contains a unique URL of the image). This means that we can use the back and forward button navigate through then our images, while in the address bar, we have a direct link to the picture that we can save a bookmark, or send someone else.

Demo code

We have 2 divs. One contains a folder structure, the other contains the current picture. Here I will show only the most important moments. Source code of example is very short (80 lines) look at it after reading the entire article.

The method bindEvents add handlers for the event popstate, which is called when the user navigates through the history and allows the application to updateits state.

window.addEventListener('popstate', function(e){
  self.loadImage(e.state.path, e.state.note);
}, false);

The object event, which is passed to the event handler has a property state - the data that we have passed as first argument to pushState or replaceState.

We bind event handler click event on the DIV element, which is our filestructure. Using the delegation event, we open or close the folder or download a picture (with the addition of history record). We look at the className of the parent element in order to understand which of the items we clicked:

• If this folder we open or close it
• If this is a picture, then we show it, and add an element of history

   dir.addEventListener('click', function(e){
        e.preventDefault();
        var f = e.target;
 
        if (f.parentNode.classList.contains('folder')) {
          self.toggleFolders(f);
        }
 
        else if (f.parentNode.classList.contains('photo')){
          //be nice to FF 4 & 5, as they don't have .dataset yet
          //https://bugzilla.mozilla.org/show_bug.cgi?id=560112
          note = f.dataset ? f.dataset.note : f.getAttribute('data-note');
 
          self.loadImage(f.textContent, note);
          history.pushState({note: note, path:f.textContent}, '', f.href);
        }
   }, false);

A method that modifies the contents of pictures and updates its signature is very simple:

loadImage: function(path, note){
  img.src = path;
  h2.textContent = note;
}

I have a simple html5 history demo application that demonstrates the possibility of the updated interface of History object. We use pushState to add an element of history and event popstate to update page content.

In addition, when you click on the picture we get a valid URL in the address bar, so you can save or send it to someone.

When can I use it?

Firefox 4+
Safari 5+
Chrome 10+
Opera 11.5+
iOS Safari 4.2+
Android 2.2+
IE ???
See browser list with History API supported

Additional reading

See my post about HTML5 FileUpload API

1. Manipulating History for Fun & Profit
2. WHATWG HTML5 history API
3. W3C history API Spec
4. MDC Manipulating the browser history
5. History.js — emulates HTML5 history API(location.hash magic) in browser, which does not support it

rss_only

Tweet

I use simple lftp command line tool to automate project deployment via FTP and eliminate any possible deployment errors.

The problem

When the project becomes complex, deployment is going to be complex too. It includes many steps and forgeting just one can lead to hard tracking situations. Likely you are going to catch bugs like: it works locally but does not on remote server.

Once it happens over and over again and there is room for an error, this needs to be automated, no more manual actions are needed. I want it just run-and-forget.

Here is the list of my basic requirements

• Mirror application code files, this involves only updating files that have been changed, project may contains thousands of files and this will take time
• Remove stale application files on server side which are no longer in use
• Skip uploading configuration files, This is a common case, overwriting database configuration on the server with local configuration.
• Skip uploading data folders, while some folders on the server may have user data in it, do not touch these folders
• Delete remote cache files, some frameworks and template engines create remote cache files, which needs to be removed
• Change remote permissions, to be sure folders can be used for uploaded files
• Fully automatic, setup once and run when needed

Lftp - seems like a solution

lftp is a program that allows sophisticated ftp and http connections to other hosts. lftp can handle six file access methods - ftp, ftps, http, https, hftp.   Every operation in lftp is reliable, that is any not fatal error is ignored and the operation is repeated. So if downloading breaks, it will be restarted from the point automatically.

lftp has shell-like command syntax allowing, it accepts on the input script like program. Each command tells lftp what to do. This gives perfect control of what is going to be uploaded or deleted.

Installing under Ubuntu is pretty straightforward: apt-get install lftp. There is also windows version available.

Usage example

In this example, lftp script is used for uploading CakePHP project, which is not a simple upload task.

In the project folder I created file sync.sh which handles all the dirty job, below is the content of the file.

www is local folder. public_html is remote folder.

#!/bin/bash
lftp << EOF
set ftp:ssl-allow no
set ftp:passive-mode true
set ftp:list-options -a
open -u [USERNAME],[PASSWORD] ftp.host.com
# Upload site configuration
put -O public_html/dev/app/config/ www/app/config/bootstrap.php
put -O public_html/dev/app/config/ www/app/config/routes.php
# Mirror program code
mirror -R --delete --exclude config --exclude tmp --exclude webroot www/app public_html/dev/app
# Mirror webroot (not delete remote user files)
mirror -R --exclude user_data www/app/webroot public_html/dev/app/webroot
cd public_html/dev/app/tmp/cache/models
mrm *
cd public_html/dev/app/tmp/cache/persistent
mrm *
EOF

FTP username or password can be embedded into script itself, which is less secure. More secure way is to add additional lines of bash scripting to ask for password.

put - uploads one sinlge file. First argument is REMOTE file name, second argument is LOCAL file name.

mirror - synchronizes local and remote folder. Used swithes are:

• -R - sync from local to remote (uploading, NOT downloading)
• --delete - delete files on remote side which do not exist on local side
• --exclude - exclude directory by name
• cd - remote change directory command
• mrm - remove multiple files

Debugging

When lftp is run with "-d" switch it shows all FTP commands, which is good for debugging. One more good switch of mirror command is "--dry-run", it just shows what is going to be uploaded without any actual action performed.

Beware! Do not do any experiments on live project with uncommitted changes!!! You may loose your changes. Do a project copy or commit all changes before.

Conclusion

Now I am sure that after development round all changes will be uploaded to the server and nothing left behind. Upload script is easy customizable for future projects.

Resources

• LFTP man page

Tweet

File uploads have traditionally had very bad usability on the web. The standard solution was uploading files as part of a form, leaving the user to just wait until the process was done. We could offer barely any feedback of what was going on. Several options appeared in HTML5 to make the process more bearable for the user.

In HTML5 input element have got new attribute multiple. It allows to upload multiple files at once without using Flash based uploaders. This attribute receives one value "multiple" and looks like this:

<input type="file" multiple="multiple" name="files[]" />

Pay attention to name, I specified it as an array to receive array of files in PHP script.

Once there is the question, what kind of browser does not understand this html5 innovation, the answer is easily predicted, a family of Internet Explorer. Starting with version 9 and below, they do not support this functionality and simply ignore the attribute and sincerely hope that in the final 10 version, they are corrected.

Besides that, we gave the users to upload many files at once, we must take care of them and give them the opportunity to download just the files we need. And then comes to help another new attribute accept. Which takes the "MIME Media Types".

<input type="file" multiple="multiple" name="files[]" accept="image" />

<input type="file" multiple="multiple" name="files[]" accept="image" required="required" title="Please choose at least one file" />

How will it work when you try to send the form with an empty input, a warning about an error with the following text:

Firefox html5 file uploads view

Chrome html5 file uploads view

Opera html5 file uploads view

To my surprise, Safari did not show a warning and sent a blank form. The pictures above show how it looks in the three browsers, Chrome does not show a very pleasing look, but once he showed an addition to the errors that took the title of the item.

Html5 file upload and javascript

There are still several problems  to be solved for the convenience of the user.

• Check the file type
• File sizes
• Number of Files

And we have a very simple way to check things out with JavaScript. In the file input element there is an invisible attribute "files", when accessing it returns FilleList object

var inputFile = document.getElementById('input').files;

FileList : {
	0,
	1,
	2,
	length: 3,
	item
}

Where 0,1,2 are keys of uploading files.
Length, number of files
Method item() returns file item object.

Refer to the file in two ways:

// First way
var file = inputFile[0];
// Second way
var file = inputFile.item(0);

The variable file will contain a file object:

File : {
	constructor: File {...},
	fileName: 'image.png',
	fileSize: 879394,
	name: 'image.png',
	size: 879394,
	type: 'image/png',
	getAsBinary,
	getAsDataURL,
	getAsText,
	lastModifiedDate: 'Thu May 16 2011 00:01:34 GMT+0300 (Eastern Europe Daylight Time)'
}

Opera has no properties fileSize, FileName but there is a name and size. In Chrome, Firefox, Safari has all four properties. Only in Chrome is the property of lastModifiedDate.

Tweet

This book is about  applying your PHP skills in a clouds.

Most developers gloss over when they hear “the cloud” as the term has been appropriated by marketing types and has very little true meaning anymore. In this book, Vito and Ivo strip bare the buzzwords and help PHP developers figure out what the cloud is actually about and how they can take advantage of what cloud computing has to offer. After reading this book, you will not only have a clear picture of what cloud computing is, but you’ll also be able to utilize the various cloud services that are out there.

Topics covered include:

• Cloud Computing primer (characteristics and models)
• Cloud Basics (prerequisites and programming)
• Cloud Architecture (horizontal scalability, preventing bottlenecks, abstraction, multi-tenancy)
• Working with Popular Cloud Infrastructures (Amazon, Rackspace, Azure)
• Working with Popular Cloud Platforms (Google, Rackspace, and others)
• Working with Popular Cloud Software and Services (OpenID, OAuth, Search, Payments, CRMs, Maps, Storage)

Buy on Amazon

ISBN 9780981034522
Pages 172
Authors Ivo Jansch Vito Chin
Print $39.99
PDF-only $36.99

Tweet

OpenID is an open standard that describes how users can be authenticated in a decentralized manner, obviating the need for services to provide their own ad hoc systems and allowing users to consolidate their digital identities

The OpenID protocol does not rely on a central authority to authenticate a user's identity. Moreover, neither services nor the OpenID standard may mandate a specific means by which to authenticate users, allowing for approaches ranging from the common (such as passwords) to the novel (such as smart cards or biometrics).

See original article here for DOWNLOADS and DEMO

Tweet

Nowadays many companies have gone cloud, which covers most needs, but leaves space for homegrown small applications, ideas, prototypes, etc. It is convenient to organize user login into these applications using already existing Google Apps login infrastructure.

This is about using it with PHP.

Since previous post, I've updated PHP library used to the latest versions and tested them again. Also I rewrite bundled example to work in more transparent way and made it more clear.

Libraries version

php-openid-apps-discover-1.0.2.zip
openid-php-openid-2.2.2-0-ga287b2d.zip

Online demo

http://a32.me/other/ga-open-id/

Downloads

All-in-one-pack includes libraries and example

Installation

Unpack, make sure make TMP folder with sub-folders are writable.

Example code

<?php
/**
 * @date Created on Feb 22, 2011
 * @author		Constantin Bosneaga <ameoba32@gmail.com>
 * 
 */
 
// Include files
require_once "Auth/OpenID/Consumer.php";
require_once "Auth/OpenID/AX.php";
require_once "Auth/OpenID/google_discovery.php";
require_once "Auth/OpenID/FileStore.php";
require_once "Auth/OpenID/SReg.php";
require_once "Auth/OpenID/PAPE.php";
 
// Init login
$tmp = dirname(realpath(__FILE__)).DIRECTORY_SEPARATOR."tmp";
if (!file_exists($tmp)) die('Temp path '.$tmp.' does not exists'); 
if (!is_writable($tmp)) die('Temp path '.$tmp.' is not writable');
$config['tmp_path'] = $tmp;
 
// Return URL
$config['return_server'] = ($_SERVER["HTTPS"]?'https://':'http://').$_SERVER['SERVER_NAME'].":".$_SERVER['SERVER_PORT'];
$config['return_url'] = $config['return_server'].$_SERVER['REQUEST_URI']."?module=return";
 
// Cache for google discovery (much faster)
$config['cache'] = new FileCache($config['tmp_path']);
 
// Open id lib has many warnig and notices
error_reporting(E_ALL ^ E_NOTICE ^ E_WARNING ^ E_USER_NOTICE);
 
$module = $_GET['module']?$_GET['module']:$_POST['module'];
switch ($module) {
 
	/**
	 * Login form
	 * 
	 * 
	 */	
	case '':
	?>
	<h1>Login with Google Apps</h1>
	<form method="POST">
	<input type="hidden" name="module" value="login">
	Your Google Apps domain:
	<input type="text" name="domain" value="">
	<input type="submit" name="submit" value="Login">
	</form>	
 
	<h1>Login with Google</h1>
	<form method="POST">
	<input type="hidden" name="module" value="login">
	<input type="hidden" name="domain" value="https://www.google.com/accounts/o8/id">
	<input type="submit" name="submit" value="Login">
	</form>
 
	<h1>Login with Yahoo</h1>
	<form method="POST">
	<input type="hidden" name="module" value="login">
	<input type="hidden" name="domain" value="http://me.yahoo.com">
	<input type="submit" name="submit" value="Login">
	</form>
		<?	
	break;	
 
 
	/**
	 * Process login 
	 */
	case 'login':
		$store = new Auth_OpenID_FileStore($config['tmp_path']);
		$consumer = new Auth_OpenID_Consumer($store);
		new GApps_OpenID_Discovery($consumer, null, $config['cache']);
 
		try {
			$auth_request = $consumer->begin($_POST['domain']);
			if (!is_object($auth_request)) die('Auth request object error. Try again');
		} catch (Exception $error) {
			die($error->getMessage());
		}
 
		/// Request additional parameters
		$ax = new Auth_OpenID_AX_FetchRequest;
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/contact/email',2,1,'email') );
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/namePerson/first',1,1, 'firstname') );
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/namePerson/last',1,1, 'lastname') );
 
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/namePerson/friendly',1,1,'friendly') );
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/namePerson',1,1,'fullname') );
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/birthDate',1,1,'dob') );
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/person/gender',1,1,'gender') );
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/contact/postalCode/home',1,1,'postcode') );
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/contact/country/home',1,1,'country') );
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/pref/language',1,1,'language') );
		$ax->add( Auth_OpenID_AX_AttrInfo::make('http://axschema.org/pref/timezone',1,1,'timezone') );
 
		$auth_request->addExtension($ax);
 
		// Request URL for auth dialog url 
		$redirect_url = $auth_request->redirectURL(
			$config['return_server'],
			$config['return_url']
		);
 
		if (Auth_OpenID::isFailure($redirect_url)) {
			die('Could not redirect to server: ' . $redirect_url->message);
		} else {
			header('Location: '.$redirect_url);
		}	
	break;
 
 
	/**
	 * Return URL, google redirects here after login
	 */
	case 'return':
		$store = new Auth_OpenID_FileStore($config['tmp_path']);
		$consumer = new Auth_OpenID_Consumer($store);
		new GApps_OpenID_Discovery($consumer, null, $config['cache']);
 
		$response = $consumer->complete($config['return_url']);
 
		// Check the response status.
		if ($response->status == Auth_OpenID_CANCEL) die('Verification cancelled.');
		if ($response->status == Auth_OpenID_FAILURE) die("OpenID authentication failed: " . $response->message);
		if ($response->status != Auth_OpenID_SUCCESS) die('Other error');
 
		// Successful login
 
		// Extract returned information
		$openid = $response->getDisplayIdentifier();
		$ax = new Auth_OpenID_AX_FetchResponse();
		if ($ax) $ax = $ax->fromSuccessResponse($response);
 
		$sreg = Auth_OpenID_SRegResponse::fromSuccessResponse($response);
		if ($sreg ) $sreg = $sreg->contents();
 
		# print response
		?>
		<h1>OK</h1>
		You have successfully verified <a href="<?=$openid?>"><?=$openid?></a> as your identity.<br/><br/>
		<p>The following AX response received:</p>
		<pre><?=nl2br(print_r($ax->data,true))?></pre>
 
		<p>The following sreg response received:</p>
		<pre><?=nl2br(print_r($sreg,true))?></pre>
		<?
	break;	
 
}
 
 
class FileCache {
	var $cache_file;
 
	function __construct($tmp_path) {
		$this->cache_file = $tmp_path.DIRECTORY_SEPARATOR."google.tmp";
	}
 
	function get($name) {
		$cache = unserialize(file_get_contents($this->cache_file));
		return $cache[$name];
	}
 
	function set($name, $value) {
		$cache = unserialize(file_get_contents($this->cache_file));
		$cache[$name] = $value;
		file_put_contents($this->cache_file, serialize($cache));
	}
 
}
 
 
?>
 

Tweet

In projects with 200+ code files, one simple return at the end of file can cause bugs hard to find.

UPD: My discovery is clearly described in documentation http://php.net/manual/en/language.basic-syntax.instruction-separation.php

Tweet

Use nice JavaScript code formatting library. Just enter program text in <pre><code></code></pre> tags in any part of post, include CSS and library and voila:

require_once 'Zend.php';
require_once 'Zend/Uri/Exception.php';
require_once 'Zend/Uri/Http.php';
require_once 'Zend/Uri/Mailto.php';

abstract class Zend_Uri
{

  /**
   * Return a string representation of this URI.
   *
   * @see     getUri()
   * @return  string
   */
  public function __toString()
  {
      return $this->getUri();
  }

  static public function factory($uri = 'http')
  {
      $uri = explode(':', $uri, 2);
      $scheme = strtolower($uri[0]);
      $schemeSpecific = isset($uri[1]) ? $uri[1] : '';

      // Security check: $scheme is used to load a class file,
      // so only alphanumerics are allowed.
      if (!ctype_alnum($scheme)) {
          throw new Zend_Uri_Exception('Illegal scheme');
      }
  }
}

This code should be appended into HTML:

<script src="http://yandex.st/highlightjs/5.12/highlight.min.js"></script>

<link rel="stylesheet" href="http://yandex.st/highlightjs/5.12/styles/default.min.css">

<link rel="stylesheet" href="http://yandex.st/highlightjs/5.12/styles/vs.min.css">

<script>

  hljs.initHighlightingOnLoad();

</script>

Library itself is available through Yandex content distribution network(CDN), so no installation is needed.

Links

• Test page with more themes
• Official library page

Thanks to Ivan Sagalaev.

Tweet