< Back to James' Blog

Hello everyone,

You're currently using the latest version of the site. It's got a hell of a lot of improvements, including the blogs (which you could read before on beta.matchstd.com), a new search system, upgrades to the forums, a new registration form, and a whole load of bug fixes.

I hate to say it, but there were a few members out there who were hiding away in the system, unable to be reached by search. That's changed. These members were few and far between, but they were still there. This was due to a little glitch in the SQL query that I used, which has been repaired. For those of you who write queries on a day to day basis, you understand how difficult it can be at times to get the syntax right when you're including them as strings within your application. To make matters worse, I use SQLObject for about half of the system, and psycopg2 for the rest. Combining the two can be a little bit of an issue when you're trying to keep track of what you're doing with the database. The search itself is done with a raw query through psycopg2, and some items are weeded out through SQLObject models.

So why do I program it like that? Because with psycopg2 and the raw queries you can get a lot more control over your database, but SQLObject keeps accessing properties of rarely grabbed items as easy as, heck, I don't know. Anyway, it was a damned little parantheses issue in the initial query. I need to get syntax highlighting to work it's magic on sql queries in vim. But first things first. Like system stability.

I've got a memory leak in the code base somewhere. But despite all my tests, I can't replicate it on my testing machine. It only happens on the server. I recently installed python 2.5 alpha 1 on my development computer, and tried the application under it. There are minor improvements in stability, and I'm hoping that it might fix the memory leak, as well. I do, however, want to wait a while for that to be more stable, and also for all the accompanying software I'm using to support it. CherryPy looks like it works properly, but the PIL throws up some errors. I had to fix the locations of a few import statements in kid so that those modules would load, but other than a few warning statements thrown up every now and then, everything is functional. This is a very good sign...

Heh. I ramble on... In a few hours I get to take a much needed break from it all with a trip to the mesa, something I'm looking forward to quite a bit. I haven't really left the computer for more than an hour or two since we hit CNN, but now everything is looking really good, so I'm going to give it a shot.

Let's see... That's about it. Everybody take it easy, and have a good time on the site,

James

Comments