You are here

Where Have I Been?

Where indeed. To start with, I had decided to take a bit of a break to catch up on some things I had been neglecting before. Unfortunately the break became indefinite when I made the mistake of "upgrading" to KDE 4. Sadly, two years after the first allegedly stable KDE release, KDevelop 4 is still not only considered beta, but it's not a Google-style beta that is perfectly usable. From my experience, it's really only useful to developers who are actually working on KDevelop itself. It crashes on a regular basis doing simple operations, most of the functionality doesn't actually seem to work, and there's plenty of stuff just plain missing.

To be fair to the KDevelop devs, they don't recommend using it either. Unfortunately for those of us living in the real world, KDE 3 support has been dropped from all of the distros that I actively use so I had very little choice, unless I wanted to roll back KDevelop after doing the upgrade, which would have pulled back in all of the KDE 3 deps. While it wasn't a big deal to clean out KDE 3 for the upgrade (in Gentoo, where that was the recommendation), it would be a pain now that KDE 4 is installed. So I thought maybe KDevelop 4 would be at least minimally usable for my very basic project. Sadly that has not proven true thus far.

Lest this turn into a (bigger) rant on KDE 4, I'll leave this topic with one final thought: In order to seriously continue Coldest development it's looking like I'll need to switch to a different IDE. In case anyone cares.:-)

One final issue that's holding me up right now is continuing collision detection problems. I know, I said that I had resolved them. If it makes you feel any better I was lying to myself too, and my changes had merely reduced the problem somewhat. Ultimately the problem comes down to the algorithm. Checking and resolving collisions against a curved shape using only straight lines causes problems when you start out very close to the curved object and try to slide around it. At least that's my working theory.

I've investigated a couple of solutions, but so far they've all had problems of their own. If I can't come up with a proper solution, I may have to use an improper one, but even if it does come to that I'm hopeful that it won't result in a repeat of the old problems that prompted me to implement this stricter collision detection.

Anyway, this post has gotten rather long now, but I suppose that's not surprising given the length of time since my last update. In the interest of not having any boredom deaths on my conscience I'll call it a night and go watch the Olympics.