Raiders of the Lost Device

· by Steve · Read in about 2 min · (384 Words)

Over the last few days I’ve been tackling the straggling few lost device issues that the D3D9 rendersystem had remaining, namely the gui demo, that manual textures would bomb the device restore, and an odd colouring issue after a device restore on some (but not all) demos. All are now fixed, thanks to some help from the community providing test cases and in some cases just pointing at what I’d missed (“It’s THERE you moron!” “Doh”). There’s the beauty and the humbling nature of open source in one - other people can help find bugs, but you also get to feel like a tit sometimes. In the cosmic scheme of things, it’s still a win though.

I applied one community patch a little too fast and later found it made some things worse (oops), but a swift rewrite of the way textures have their internal resources tracked (versus being actually loaded) sorted it all out, and it feels a lot more natural now. I believe, finally, that every lost device issue is resolved in Azathoth. Dagon is fine too I think, although there have been a lot more changes there, particurly to do with textures (the RTT code has been mostly rewritten to support rendering to any surface, including mips and volume slices, MRTs and to use FBOs in GL) so it will need a bit more testing, but I’m pretty confident.

There has been lots of activity in the community lately, alphageek has released the first commercial Axiomatic version, FragFist released a very nice looking demo, and unbeknowst to me people have been building OGRE environments into Eclipse, in one case as a project for a leading car manufacturer (more details when they’re officially announced). I’ve used Eclipse for years at work - or rather, the IBM version WebSphere Studio, since we use the WebSphere appserver test environment - and have witnessed it’s flexibility but I really didn’t see that coming.

So generally it’s been 100% work since I got back! Marie still has arm issues, the xrays were inconclusive but the suspicion is that there’s a fracture around the elbow joint which is difficult to spot. So she gets to remain in the sling and be irradiated still further next week to try to establish one way or the other for sure.