Roterhead wrote:This probably has no bearing but am offering just in case. We also had application lockups that did not show until quad core machines started being used. We discovered all DotNet GUI updates are actually performed on a single thread. Consequently we had to ensure all GUI activity was marshaled to a single thread.
Thanks for that -- yeah, it's a similar sort of thing. for the preview thread, we're actually explicitly starting a separate runqueue and doing all the UI on an independent thread, since it's not really interactive and is more of a read-only display so we can get away with that (and then it means it doesn't impact the performance of the sequencer when under heavy load).
So it's a similar issue -- the most annoying thing is that it works fine for most people (including all the developers, even outside the Visual Studio environment!), which makes it hard to track down what's going on! 🙂
Thanks for the comments.
Cheers,
Michael
OK, we took this off-thread, and Dave helped me test out a few versions to track down the problem -- much thanks to dlovely.
If anyone else is having the same problem, try testing one of the recent dev builds if you're comfortable testing bleeding-edge stuff:
http://www.vixenlights.com/builds/Vixen-3.0.6.22-Setup.exe
http://www.vixenlights.com/builds/Vixen-3.0.6.23-Setup.exe
Keep in mind they're development builds only, and things may change, and there's no guarantees about them. However, they're useful to test something like this preview quickly, etc. If you're having problems, please let me know (and post your log files, as there may be clues in there!)
Thanks!
Michael
Hallelujah! Build 25 appears to work for me!
Thanks for whatever you fixed. And thanks for others that provided more information that allowed this to be root caused. (I have to say I was selfishly glad when I started to see more folks having the same problem. I was starting to think it was something unique to my computer that I screwed up.)
Back to putting pixel strings together...
Mark
Thanks Mark! 🙂 Yeah, it was a weird one -- it seems to affect only some particular machines or installations, but they reliably fail every time. The only possible thing I could come up with is that it might be a slightly different version of the .NET framework, or some patches applied/not applied, which caused the multi-threading behaviour to operate slightly differently on some machines (triggering the bug) and not other.
Still confuses me, but it's worked around for now, which is the key thing. 🙂
