gproc_dist_tests: dist_test_...*timed out* about gproc HOT 10 CLOSED

Sorry about the delay. I have made some changes - most importantly, changed the 'inparallel' macro to 'inorder'. I think the nondeterminism hides somewhere in the test code, but haven't been able to pin it down. After making minor changes here and there, I could no longer reproduce the error, but the parallel runs also make it harder to debug failing tests.

from gproc.

Hi,

Sorry to say, but I've found yet another(?) stochastic issue after your fix:
Could you check whether It is the same?
http://susepaste.org/51044195

from gproc.

Yes, there seems to be some timing issue with gproc_monitor.

Hopefully, as the uw-standby-monitors branch is starting to look stable (issue #53), I will be able to deprecate gproc_monitor in favor of gproc:monitor(Key, follow).

from gproc.

I did tweak gproc_monitor a bit. At least on my laptop, this seems to have made it harder to reproduce the problem. Pushed to master.

from gproc.

Have you seen any other intermittent failures?

from gproc.

With the latest master I still sometimes see: http://susepaste.org/25878998

from gproc.

Ok, I'll keep the issue open then.

from gproc.

I've just merged some changes that should at least take care of the timeout in t_follow_monitor/1.

from gproc.

if @matwey isn't complaining it's meaning it's fixed. @uwiger I think you can close the ticket.

from gproc.

Lets close this. I'll reopen if required

from gproc.