Anthony Williams schrieb:
Thanks for fixing this that quickly. Any idea when this fix is going to be included in an official release? (1.35.1 or 1.36?)
I'll make sure it goes in the next one, but I don't know when that will be.
Hmm, I'm still having problems, but mostly only under heavy system load. Seems a notifyOne gets lost somewhere (at least, the result is that a thread is waiting at a point where it shouldn't be). I've tried the latest SVN trunk (as the commit log said that another similar bug to the one I reported was fixed) -- no luck. Funnily enough, it's exactly at the same place again where I had my original problem. Seems that the switch during condition::wait is not fully atomic or something similar, as everything should be protected by a mutex and yet the state I'm getting into is inconsistent. I'll try to investigate a bit further. Any ideas? Cheers, Anteru