Pinned task problem about enkits HOT 12 CLOSED

I'll have a check through the logic again for this over the weekend, but if you're able to reproduce the issue in a small sample then that would be great, but it's not a requirement for me to check this.

from enkits.

Thanks, I'll try to reproduce it in a sample later.

In the meanwhile, here's a picture of missed signals:

from enkits.

I've created a branch dev_waketasks with a potential fix for this - it looks like I broke the logic when moving to semaphores. I'm still doing some tests and running through the logic to check, but would appreciate knowing if this resolves your issue.

from enkits.

I've been able to find a reproduction for this, which will be added to my tests. The above fix doesn't resolve it but now I have a reproduction I should be able to fix this fairly rapidly - however I'm not sure how much time I'll have over the next few days.

from enkits.

Yeah, unfortunately the change hasn't helped, hopefully it will be fixed soon enough.
And I would like to say thank you, Doug, and wish you a Merry Christmas and a Happy New Year! 🎄

from enkits.

I have a good idea what's going on now, and a potential fix in progress so this shouldn't take too long.

Merry Christmas and a Happy New Year to you too!

from enkits.

I've pushed a fix for the issue I was able to reproduce in the revised TestWaitforTask.cpp test, though there's always a chance you have another related issue.

The current solution is slightly heavy handed, but the alternative solutions will take longer to program.

from enkits.

Update I'm still tracking a bug with this which happens extremely rarely, so the current dev_waketasks branch changes improve but do not resolve the issue.

from enkits.

The latest commit on dev_waketasks resolves the above issue, so may be worth testing on your code when you get time.

Meanwhile I'll continue checking logic and testing.

from enkits.

I've now merged this into master branch.

I'm going to pre-emptively close this assuming the fix works for your situation as well as my test, but if not please re-open.

from enkits.

I'm reopening this issue as I can still replicate the problem on OSX now I've fixed #49

from enkits.

I found the logic error responsible for the recent issue - somehow I'd missed that HaveTasks() was not being called by WaitForTaskCompletion() after wait variables were set. I found a few other minor perf issues at the same time.

I'm going to close this again, hopefully it will stay closed for a while :) Please do update this thread if your able to test with whether your issue is resolved or not and if not reopen.

Many thanks for finding this!

from enkits.