May 29, 2011

f Comment

How Does Apache Server Work With PHP Module?

Amazon How the HECK does Apache server work with PHP module? What is the flow in which an HTTP request makes through a PHP enabled Apache web server? How does Apache know whether the request is supposed to go through PHP engine or to go through Apache processing directly?

I have WAMP (Version 2.0) installed on my Windows 7 operating system. For those who don't know, WAMP is an acronym that stands for Windows Apache MySQL PHP. Just so you know this versino of WAMP comes with the following versions of Apache, MySQL, PHP:

Apache version 2.2.11
PHP version 5.3.0
MySQL version 5.1.36

I need to enable the rewrite module of Apache to process URL rewrite rules, so I have the following line in httpd.conf:

LoadModule rewrite_module modules/
I enable rewrite logging at "C:\wamp\logs\rewrite.log" so I can see how rewriting is done via adding the following line in httpd.conf right after Loadmodule:

RewriteLog "C:\wamp\logs\rewrite.log"
RewriteLogLevel 9

I also have the following line in httpd.conf to enable PHP processing:

LoadModule php5_module "c:/wamp/bin/php/php5.3.0/php5apache2_2.dll"
Now an HTTP request goes through Apache, which hands it to the rewrite module to process URL rewrite handling, and has the resulting URL go through PHP engine. But how does Apache rewrite module supposed to know when it should stop rewriting the URL and hand it to PHP module? Is there any way we can control when the processing goes to PHP engine through rewrite rules? Let's answer these questions below.

By the way if you need to know how Apache server variables including %{DOCUMENT_ROOT}, %{REQUEST_URI}, %{REQUEST_FILENAME} work refer to Test Whether a Server Variable is Empty in Popular Web Servers.
My Answers
Simply by having the line 'LoadModule php5_module "c:/wamp/bin/php/php5.3.0/php5apache2_2.dll"' in httpd.conf, PHP is enabled to handle pages with extension 'php' (e.g. black-jacket.php) after Apache passes the page through to PHP module. By the way if you'd like PHP to also be able to handle pages with other extensions add the following line after LoadModule in httpd.conf (so PHP handles .html too):

AddType application/x-httpd-php .html
Now here's the interesting part. How does Apache rewrite module know when to pass through the request to PHP module? The answer is ONLY and ONLY if when the request is NOT matched by ANY rewrite rule, the request is passed through to PHP handler. Even funnier is that there is absolutely NO control you have in the rewrite rules to tell Apache to pass through to PHP handler right away! If you don't agree let me know.

I've tried [L] and [PT] on the RewriteRule directive but when the request matches that particular directive, the request is "internally redirected" to rewrite module again and it's processed by the rewrite module again! In rewrite log it looks like this: - - [29/May/2011:22:26:46 +0800] [localhost/sid#d33140][rid#2254c80/initial/redir#1] (4) [perdir C:/repository/trunk-php/] RewriteCond: input='C:/repository/trunk-php/cache/index.php' pattern='-f' => matched - - [29/May/2011:22:26:46 +0800] [localhost/sid#d33140][rid#2254c80/initial/redir#1] (2) [perdir C:/repository/trunk-php/] rewrite 'index.php' -> '/cache/index.php' # this rule has [L] but it's still internally redirected to rewrite module as the following log statement suggests - - [29/May/2011:22:26:46 +0800] [localhost/sid#d33140][rid#2254c80/initial/redir#1] (1) [perdir C:/repository/trunk-php/] internal redirect with /cache/index.php [INTERNAL REDIRECT] - - [29/May/2011:22:26:46 +0800] [localhost/sid#d33140][rid#2351398/initial/redir#2] (3) [perdir C:/repository/trunk-php/] strip per-dir prefix: C:/repository/trunk-php/cache/index.php -> cache/index.php - - [29/May/2011:22:26:46 +0800] [localhost/sid#d33140][rid#2351398/initial/redir#2] (3) [perdir C:/repository/trunk-php/] applying pattern '(.+)$' to uri 'cache/index.php'

The log statement in RED tells you the request is internally redirected to Apache rewrite module for handling again. Only when no rewrite rules have been matched by the current request will you see the following line in the rewrite log: - - [29/May/2011:22:27:52 +0800] [localhost/sid#d33140][rid#228b510/initial/redir#2] (1) [perdir C:/repository/trunk-php/] pass through C:/repository/trunk-php/cache/index.php
Meaning that the Apache rewrite module is finally done with handling the request, and it's the next handler's job to handle it. In this case PHP engine will simply render C:/repository/trunk-php/cache/index.php.

So how do you add a handler and control the order of the handlers? To my disappointment I have NOT found any document online that answers this question. There is "AddHandler" directive in httpd.conf that's supposed to do that but loading the PHP module does that implicitly already.

So the conclusion is that ONLY and ONLY if when the request is NOT matched by ANY rewrite rule, the request is passed through to PHP handler. There is absolutely NO control you have in the rewrite rules to tell Apache to pass through to PHP handler right away! Not by using 'last' option [L] or 'pass through' option [PT] at the end of RewriteRule directive. If you don't agree let me know.

Knowing this fact you may find it impossible to handle the rewrite logic you have in mind. Think deeper and the solution will come. For example you may have a rewrite rule that you'd like the rewritten URL to go straight to PHP engine right away, and you add [L] to that rule which immediately internally redirects the rewritten URL to rewrite engine again (and therefore all the server variables such as %{QUERY_STRING}, %{REQUEST_FILENAME} and %{REQUEST_URI} are updated accordingly) and make the request go through each rule again. In that case you'll have to make sure the request is NOT matched by any RewriteRule and therefore passed through to PHP handler.

If you have any questions please let me know and I will do my best to help you!
Please leave a comment here!
One Minute Information - by Michael Wen
ADVERTISING WITH US - Direct your advertising requests to Michael