I'm a big fan of partitioning my Rails applications such that pages are either public or private. Private pages are all name spaced through /admin/* which enforces authentication. Public pages are directly accessible and cached. Recently, however, I ran into a problem when trying to implement action caching with this name spaced design. Whenever the sweeper would try to expire a cached action it would include the /admin portion of the url and thus fail to sweep the public page. For example, when I create a new blog post I need to expire the /posts index: « read the full post »

After posting blog entries I like to log into my Google Webmaster Tools and force a re-download of my site map to cut down on the time it takes to be indexed. Normally this is a manual process, but it is pretty simple to automate. I develop with Ruby on Rails and host on a Linux server but the same strategy could easily be applied to other environments. « read the full post »

After installing Windows XP Service Pack 3 I kept getting this message when attempting new updates: 'Error: A problem on your computer is preventing updates from being downloaded.' Re-registering the dlls seemed to fix the issue: « read the full post »

After a couple of weeks working with Blogger I decided to throw in the towel and code up my own blog software. The motivating factor for me was trying to deal with their hideous template engine. Something as simple as trying to change the title of a blog post so that it is more SEO friendly required hideous hacks that created broken HTML. « read the full post »

I help out at a local coffee shop (see blog title) from time to time and recently was asked to setup a donated PC for public internet access. There was no budget for commercial software so I decided to see what the open source world had to offer. After a quick 10 minutes of Googling I found a few interesting posts that looked promising. I decided to go with Ubuntu 7.10 desktop and use a utility on KDE called the kiosktool to lock things down. « read the full post »

In my last post I showed you how to automatically update an Ubuntu machine without human intervention. The problem with the fully automated solution, though, is that it could wreak havoc if one of those updates is flaky. Sometimes you just want to be notified when there are updates available but manually chose to install them. Enter cron-apt: « read the full post »

This weekend I finally got around to figuring out why my Ubuntu 7.10 systems were not auto-updating like I wanted. The problem was that my crontab entry that was supposed to do the work was missing some paths to binaries that it was trying to execute. « read the full post »

After a fresh install of windows you inevitably have a ton of updates to perform. For those that directly update through the Windows Update site this is just a matter of updating, rebooting, rinse & repeat. For those of us that utilize WSUS, however, it can naturally take a bit longer if you let the automatic update client do its thing. For security reasons it is always preferable to fully patch a system as soon as possible so you don't end up with a compromised machine on your hands. « read the full post »