Taking off from Toronto City Centre Airport*

Last friday I flew up to Toronto and returned on sunday. In theory, it was to build up my cross-country time so I could take my instrument practical test soon, but it was also nice to see friends and the city that I miss.

It rained most of the day on Saturday while I was there and that weather was still hanging around by the time I was ready to return on Sunday. For the leg from Toronto to Watertown, NY there were stable ceilings well with my comfort zone and by the time I got there it was almost entirely clear. For the leg from Watertown back to Bedford, I planned a route that took me north over to Portland, ME and then down to the Boston area because most of Massachusetts was still experiencing bad weather. However, by the time I was about an hour outside of Watertown, flight service said that most of the state had cleared out so I diverted directly to Bedford from where I was. Turned out to be a beautful flight and only took about 4 hours thanks to the favorable winds aloft.

By the way, I'm pretty sure that I saw shaver dry-humping one of the monitors I brought up with me when he thought I wasn't looking.

* Frank took these pictures at the airport. Nice of him to placate my ego like that.

Thunderbird 0.6 is out. Lots of good stuff in this release, especially the improved spam filters. I've been having pretty good luck with them. They are able to identify a lot of spam that spamassassin misses and has a pretty low incidence of marking legitimate mail as spam. I'm still handholding it, though. People on the mac will like the new theme as well.

I've also added a little button to the bottom of my web log for Firefox. I haven't done this in years and it fills me with nostalgia[*].

I'm slowly unburying myself from the email pile that was created while I was out of town for a week, away from Internet access. This includes bugzilla mail, of course, so I'm a bit behind on that, too.

I spent a lot of yesterday grovelling through pango code to figure out how some of the internals work. I have to work on a patch to try and get custom encodings working for some fonts so we can have a proper pango backend for Mozilla. The MathML support in Mozilla uses fonts that have custom encodings, which means that Mozilla knows how to translate internal representations of the characters used in MathML into the encodings for those fonts based on the family name. Since pango groks unicode and only unicode, we're going to need to extend the pango interface a bit to support this kind of functionality.

Of course, the reality is that me writing a patch for this will go something like this:

<blizzard> OK, I made a patch to add this functionality
* owen looks at the patch
<owen> no, no this is all wrong.  let me rewrite this.

But at least I have at least some knowledge of how pango works internally. It's surprisingly similar to the way that our current Xft code in Mozilla works, except it's, you know, smarter.

*: Here's a good line from the way back machine: "When glibc ( which does have getpwent_r() ) is in wider use this won't be an issue anymore." I can't even remember a world without glibc.