Verbose Logging

software development with some really amazing hair

T + S G I F R

Get Your Wordpress Out Of My Blogging Software

· · Posted in Programming
Tagged with

Hopefully you didn't notice much, except the speed increase, but my blog is no longer on the behemoth that is Wordpress.

While Wordpress served me well for quite some time on my other blog it always kind of bugged me.

Well no more! I took a couple days (basically 2 full days and 2 evenings) and wrote my own little blogging engine from scratch. You can find it on github and get forking if you so choose.

What I used:

ruby and sinatra

They are pretty much my tools of choice as of late, so of course.

haml and builder

Haml for HTML, builder for sitemap and the RSS feed.

will_paginate

For paging things. I had to grab the branch for non-rails usage, since it doesn't play nice otherwise.

ActiveRecord

The ORM. I love AR, and it plays with…

acts_as_taggable_on_steroids

For handling tags on posts.

RedCloth

For markup. Textile is the bomb!

messagepub

I send notifications of problems to my hosted FogBugz account using messagepub.

DIsqus

Disqus handles all the comments.

Extra Bits

Some extra things I use are:

  • crack
  • restclient
  • www-delicious
  • feedzirra
  • twitter

Those are mainly to handle the top panel.

I had the theme HTML and CSS already done from my Wordpress setup, so that was easy to port over. I didn't even touch the stylesheet. Using haml and partials made life easy, especially when it came to porting my jQuery post inline script (in action), since I could just render the partial when the request was made from jQuery. In Wordpress I rendered the entire page and parsed out the part I wanted. Lame.

Heroku hosts everything, and I've had a few problems here and there, but most have been little stupid bugs on my part.

I use etags (rack-etag middleware) and max-age so that the Heroku system and other systems can cache everything they need. Static assets are set to 1 year with the rack-static_cache middleware.

The top panel holds my shared items from Google Reader, my latest bookmarks from Delicious, my latest posts to Twitter, and some of my random Github repositories. On top of the entire request being cached, these items are retrieved and then cached in the database using ActiveRecord. The cache checks the age of items, and returns either the database contents if the item is still valid, or if the item is old, yields a block, stores that result in the database and then returns it. Basically my little poor-man's memcached. The entire cache code is:

I also implemented a simple redirection system. When I update a post, if will check to see if the permalink is going to change, and if it does, it creates a Redirection. When you hit the old URL, it first checks to see if any post matches the permalink, and then if none is found, checks redirections. If it find a redirection, it will 301 redirect you to the new URL. Win.

ActiveRecord stores things in UTC, and display times are all America/Edmonton thanks to tzinfo.

What did I gain?

  • My blog is more stable. This meaning that occasionally Shaw likes to drop my internet connection, sometimes for a few hours. Not that thousands of people are reading this, but having 100% uptime (pretty much) is nice.
  • My blog is faster. Much faster. Doing everything myself, instead of working with Wordpress, I could make it all work the way I wanted. Things are cached appropriately, and as a result the blog is much snappier. Wordpress was a dog, even with all the caching tweaks I added to it.
  • My blog has the bandwidth. Running my Wordpress based blog from home wasn't a problem when it came to CPU power. My server is a dual core AMD running 64-bit Ubuntu 8.04.2 with 4 GB of RAM, so my upload bandwidth was the bottleneck. With Heroku, at least I have some bandwidth.

Overall I feel much better about my blog, and am very happy with it. What do you think?


Sharing is Caring