@relentless said in Patch 3721 / hotfix 3722 / hotfix 3723:
Hey Jip
I've been tracking down a sim freeze today trying to figure out if it was my AI causing it or not. But it looks like its not. This freeze didn't previously happen before the patch and I only see it happen on the fields of isis map (I assume its related to whatever conditions are met by that style of map) I ran about 30 odd test today and this particular map had it happen probably 4 out of 15 games with other maps not having it happen at all.
Unfortunately the log show absolutely nothing out of ordinary, looks like a normal game up until it isn't, the sim doesn't log anything afterwards, you still get info messages e.g minimized game etc.As far as symptoms go the sim will freeze but the UI is still running happily, you can still hover over buildings etc. No errors, no nothing.
If you want to reproduce I think the best method is to use uveso's rush AI and setup a 2v2 on it then run it at 10 speed. His AI seems to make it happen more often, which could be related to the unit production numbers it does, but it also happens with others where it didn't previously. In the times I saw it, the freeze would happen between 20-50 mins of game time.
What you describe sounds like a dead loop - this can happen if a forked thread doesn't yield properly while being in an infinite statement and therefore takes up all the time of the sim indefinitely. Only question that remains is what causes it.
I'll see if I can reproduce it - sadly, as you mentioned without any form of errors or logs it becomes guess work as to where it is. Do you use the latest version of Uveso from the vault or his version on Github?