-
Notifications
You must be signed in to change notification settings - Fork 26
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Bus error / Segfault when calling finalizer of an actor #274
Comments
I can confirm this bug on my Linux box. Only encounterred segfault, though. Spent some time on this, but no clue why it occurs. |
I'll send an email to Sylvan today. I know where it happens and why. There's just a detail that I don't understand and therefore I cannot fix (for now) |
The proposed solution is to move to the new |
Spent some time on debugging this problem, my understanding is that it's a bug from upstream. The following is the comparatively shorter program to reproduce the bug.
I don't fully understand
It would be good to get some comments from someone who's more knowledgeable on this. |
@albertnetymk I can try to swap those two lines and test 100.000 times but I doubt that there is anyone with more knowledge on the subject than you, me and Sylvan. BTW, when you say Could you check if the latest version of the PonyRT has changed those lines to something else? We need to update the EncoreRT and it would be good to know if that could solve everything or if the PonyRT has changed that part and we may not rely on that change. |
I meant PonyRT. The relevant code is the same in master of |
I just ran a benchmark with an old version of pony and I also got a seg fault (probably because of a different reason). |
@jupvfranco it might... the problem is that the new PonyRT modified the header files, scheduler and some other parts that we were relying on. This is short list of things that I have detected to cause troubles in Encore (issues with new PonyRT). |
@albertnetymk could you test this on Linux to confirm that the issue has been solved? |
The larger example doesn't compile; the smaller one seem to work on my box using |
tested on my Mac in |
It seems that there's a bug in
upstream
/master
.I am using plain active / passive objects in the following test and get
segfaults
andbus error
s.It would be good if someone else could check that the test doesn't do anything that should fail (cyclic dependency, e.g.).
(This has been taken from the
extract.enc
in the parallel combinators and I have removed everything that has to do with them)This code crashes 145 / 10.000 times:
Bus error
Segfault
If we use the single threaded approach, it does not crash.
The text was updated successfully, but these errors were encountered: