Bug #8997

TestSignal#test_hup_me hangs up the test suite

Added by Vit Ondruch 6 months ago. Updated 4 months ago.

[ruby-core:57726]
Status:Third Party's Issue
Priority:Normal
Assignee:Motohiro KOSAKI
Category:-
Target version:-
ruby -v:ruby 2.1.0dev (2013-09-22 trunk 43011) [i386-linux] Backport:1.9.3: UNKNOWN, 2.0.0: UNKNOWN

Description

The TestSignal#testhupme was introduced in r39819 and there were attempt to fix some deadlock issue in r39862. However, I can still observe deadlock on Fedora Rawhide's builders. Not sure what is the cause, since I cannot reproduce it locally on my computer. Any idea what might be the cause? There are only two differences I can think of between my environment and the builders (1) the builders are virtualized guest systems (2) there is no network connectivity available. Both of these looks to be improbable cause of the deadlock.

Going to disable the test case for now:
sed -i '/def testhupme/,/end if Process.respondto/ s//#/' test/ruby/testsignal.rb

History

#1 Updated by Motohiro KOSAKI 6 months ago

  • Status changed from Open to Third Party's Issue

Please contact us after you've investigated. Fedora Rawhide is not supported platform of us.

#2 Updated by Geoffrey Giesemann 4 months ago

Also seeing a deadlock on our (CentOS) koji farm which are all on vmware.

Jumping on to the one of the build nodes, I don't seem to be able to reproduce the hang manually running the specific test, or running the entire test suite.

Something to do with mock? I don't have any ideas :S

Also available in: Atom PDF