Machina improba

drupal

Now brought to you by Drupal 6

After much work, I've finally been successful in upgrading this blog to Drupal 6. This is the third attempt, and the first one that wouldn't have had me editing the database to fix the bork-ups that 'upgrading' modules between versions created. All of this caused me to remember why I was hesitant to upgrade before now.

Drupal Security Updates - 5.8 / 6.3 and OpenID module 5.x-1.2

This morning I received notice of two security update releases in my inbox: one about Drupal core being upgraded, another about the OpenID module containing a security vulnerability before 1.2.

Since upgrading Drupal core is such a pain, I've generated a patch (it's attached to this post: click on '1 attachment' to retrieve it) for upgrading your existing sites from 5.7. If you're not running a 5.7 site, please don't try to use it.

Using Disposable Email accounts for Drupal testing.

If you're working on some feature that requires you to create many user accounts, here's a tip I find helpful. Use a service like Spaml that creates disposable email accounts just by going to the site.

Drupal allows valid email addresses as usernames. If you can reduce your user creation process to just those fields which accept as valid an email address, you can just go down the form, tab, ctrl+v, tab... and then hit 'Submit'.

Categorizing Site Slowdowns

In my experience, there are three different categories of site slowdown and delay:

  • Imperceptible
  • Perceptible
  • Unusable

Other distinctions really don't matter so much. The user doesn't care whether the page load latency is 10 seconds or 15 seconds. They're just going to leave.

The progression of a delay is:

Imperceptible -> Perceptible -> Unusable

The thing about predictions is.... ("Drupal is Dying!")

... that they seldom come true. And in the age of the forever web making bold but inevitably false predictions is eventually going to get fingers poked at you.

Almost a year ago now, podcasting jockey Chris Pirillo predicted that Drupal was dying because of a shortage of 'intelligent' developers.

I don't know if he's forgotten his economics lessons, but low supply means one of two things:

Building on Drupal to Reduce Costs

Building upon Drupal because it's 'cheaper' is a very poor reason to build on Drupal. It may be cheaper in the end than hacking up some closed-source commercial CMS to include the functionality that Drupal provides in its core distribution. But there's at least one thing to keep in mind:

Three Basic Steps to Optimizing Drupal

There are a few simple steps you can take to optimize Drupal. Drupal is a large, heavy system: every advantage counts.

This initial benchmark is taken without any of the variables which I'm going to test: css aggregation, mod_deflate, and apc.


ab -n100 http://drupal.shadowlife.ca/drupal-5.7/


Connection Times (ms)
min mean[+/-sd] median max
Connect: 0 0 0.0 0 0
Processing: 24 40 20.4 31 123
Waiting: 21 34 17.2 28 119
Total: 24 40 20.4 31 123