On the Symfony Finland blog there's a new post showing you how to serve PHP over HTTP/2 with HHVM and H2O. H2O describes itself as a "new generation HTTP server providing quicker response to users when compared to older generation of web servers".

This article is not about improvements made in HTTP/2 - as there are plenty of locations for you to read up on the internals. It's a hands on article to get started using HTTP/2 today with popular tools such as Symfony, WordPress and Drupal with the HHVM PHP runtime from Facebook. You can just as well use PHP-FPM.

They start with a bit of a look at the current state of PHP and HTTP/2 on the various major web server types. H2O, while younger, natively supports HTTP/2, he does offer the caveat that "waiting won't kill you". Despite this, they go on to show you how to set up the PHP+H2O+HHVM combination complete with configuration examples and what to look for in the logs to ensure HTTP/2 functionality.

On the SmartyCode.com site, there was a new article posted recently looking at making the output of your Zend Framework application XHTML compliant.

Serving XHTML is often misunderstood by php developers. Frontend engineers simply include the XHTML doctype to their documents, without actually serving document as XHTML. This triggers majority of the browsers to treat such pages as 'tag-soup'. [...] This front controller plugin's code mostly takes concepts from the excellent article by Keystone Websites, but implements in Zend Framework environment in an object-oriented way.

The code works as a plugin to the controller and runs a few checks on the contents of the data being pushed out (dispatchLoopShutdown) and returns the correct header information (DOCTYPE, language attribute) in the correct XHTML format for the data. When the plugin is registered, all it takes is a call to the "doctype()" method to output the correct information.

Kae Verens has posted a quick tutorial about serving up files by routing them through a "fetch" script, pulling their contents in one side and back out the other.

One thing I need to do while building the multi-user version of webme is to convert it so file references such as /f/photos/an_image.jpg get transparently converted so they serve correctly, even though the actual file may be located somewhere entirely else.

There's two steps involved - rewriting the URL request for the types of files you'd like to pull through the script (using some mod_rewrite magic in Apache) and make the script to do the actual work. Source for that is included too. Not only can something like this help you keep things organized but it also allows for extra security if you need to store the files outside of the webserver's document root.