The below list is a semi-complete list of all the “pages” resources on pages.uwwebpub.com. I try to keep the list updated (and am not interested in making it automated), but I’m not quite perfect.

All of the below works are original copyright (C) by Ryan Timmons of the Catalyst Instructor Team. You are free to reproduce or distribute them provided that they are credited with the source’s URL and aren’t distributed for profit.

List of Pages Resources:

Shop Talk

I searched high and low for a good lightweight platform to publish these content pages. The “lightest weight” thing I could find was Drupal, which is far too heavy for my liking.

The Requirements

The Solution

I should probably surround “platform” with quotes whenever speaking of this setup, though. It’s really just a bunch of other people’s work glued together in a crazy frankenstein script-slash-configuration.

In particular, requests directed to http://pages.uwwebpub.com are handled by an instance of Apache with mod_php and mod_rewrite. The server’s source for web documents, called DocumentRoot in Apache lingo, are at, let’s say, /users/Pages/lib.

RewriteEngine On
RewriteBase /
RewriteCond %{REQUEST_FILENAME} !-d
RewriteCond %{REQUEST_FILENAME} !-f
RewriteRule ^(.*)$ ./index.php?url=$1 [QSA,L]

So basically, a page request is received by the server and the dispatched to a parser script called index.php (how original, I know). So going to http://pages.uwwebpub.com/page.html gets, in effect, routed to http://pages.uwwebpub.com/index.php?url=page.html.

The parser script takes the requested URL and finds the respective Pages document and grabs its meta data (such as title, last updated date, stylesheets/scripts, etc.). Then depending on the format of the request the parser script then performs various text transformations.

I consider the “format” of the request to be its extension. So if the user requests php_join.html, the parser runs the php_join file through some text-html filters. If the user requests php_join.txt, the parser runs it through some text-text filters.

For the text-html filters, I’m using PHPMarkdownExtra (a PHP implementation of Markdown, which was originally written in Perl) and then GeSHi for syntax highlighting with a couple modified syntax files (I hated the default PHP syntax definition). I had to modify PHPMarkdownExtra slightly to integrate it with GeSHi (it recognizes the {{lang:xxx}} lines, shown below, and then hands the block off to GeSHi). Unfortunately the server still re-parses the files at each request, and GeSHi is slow as heck (although it does a pretty good job). Ideally the server would parse the base file once and cache the results, only re-parsing when the source file’s mod-time has changed, but that’s something for down the road.

The text-html filters are pretty simple. Here is how a local content file looks:

Title:   Some Sort of Example Document
Author:  Sir Arthur McFinnigan
<css defaults="true">
a:link { color: green; }
</css>
===
# Example Page
 
This is an example page. This page as its own internal stylesheet making
[links](http://example.org) green.
 
I can add source code easily in lots of languages such as PHP
 
    {{lang:php}}
    <?php
    function counter()
    {
        static $i = 0;
        return $i++;
    }
 
or maybe HTML
 
    {{lang:html}}
    <table> <tr><td>Foo</td>  <td>Bar</td></tr>
            <tr><td>Fud</td>  <td>Baz</td></tr> </table>
 
and a "hidden" block:
 
{hidden}
 
This is a hidden block. It will be displayed as usual in the plaintext parsed
version, but the HTML version will replace the `{hidden}` lines with `<div
class="showhide">` which are then turned into what you see normally on the
page using jQuery.
 
{/hidden}
 
Everything above the tripe-equals sign above is optional. The parser assumes
default css and javascript and will assume no title if one isn't given and
will assume that I (Ryan Timmons) am the author if one is not specified.

And gets parsed as both HTML output (HTML results) and plaintext (in which mode the parser moves meta-info to the bottom and removes the {{lang:xxx}} lines) (plaintext results).

The HTML results are displayed in standard HTML/CSS/JavaScript. The CSS styles for this site were influenced heavily by the excellent “Dawn” theme from TextMate. The default JavaScript script includes the excellent jQuery library for taking care of showing/hiding solution blocks.

And then, finally, to get the data between my local computer and the server, I’m using the excellent file-synchronization utility Unison. I have a very simple script setup that I run whenever I want to synchronize my local and remote copies.

And that’s pretty much it. I’m considering posting more of the code to make this all happen, but it’s all just a product of a caffeine-induced coding marathon one Saturday afternoon, so it’s all a bit rough around the edges. If you’re at all interested in the code, let me know at {rtimmons}atsign{u.washington.edu}, and I’d be happy to send you the code.

Author:  Ryan Timmons
Last Modified:  06 August 2008 15:23:13 PDT
URL:  http://uwwebpub.com/

UWWebPub.com Pages

List of Pages Resources:

Shop Talk

The Requirements

The Solution