Bug 3864

Summary: Tests timeout is too short
Product: frysk Reporter: Jan Kratochvil <jan>
Component: generalAssignee: Unassigned <frysk-bugzilla>
Severity: normal    
Priority: P2    
Version: unspecified   
Target Milestone: ---   
Host: i686-pc-linux-gnu Target:
Build: Last reconfirmed:
Bug Depends on:    
Bug Blocks: 2234    
Attachments: Temporary timeout increase to at least 60 seconds.

Description Jan Kratochvil 2007-01-12 09:12:40 UTC
Debugging frysk under gdb adds too much overhead and with the all-time-swapping
frysk it will fail on:

(gdb) r
Starting program: /home/jkratoch/redhat/frysk-build/frysk-core/TestRunner -r
1000 testStressMultiThreadedDetach\(frysk.util.StressTestFStack\)
[Thread debugging using libthread_db enabled]
[New Thread -1208522512 (LWP 15307)]
[New Thread -1210623088 (LWP 15702)]
Running testStressMultiThreadedDetach(frysk.util.StressTestFStack) ...[New
Thread -1221112944 (LWP 15770)]

Time: 36.884

OK (1 test)
Running testStressMultiThreadedDetach(frysk.util.StressTestFStack) ...FAIL
  junit.framework.AssertionFailedError: event loop run explictly stopped
(perform backtrace)

Time: 9.603
There was 1 failure:
event loop run explictly stopped (perform backtrace)
   at frysk.proc.TestLib.assertRunUntilStop(TestRunner)
   at frysk.proc.TestLib.assertRunUntilStop(TestRunner)
   at frysk.util.TestFStack.multiThreaded(TestRunner)
   at frysk.util.StressTestFStack.testStressMultiThreadedDetach(TestRunner)
   at frysk.junit.Runner.runCases(TestRunner)
   at frysk.junit.Runner.runArchCases(TestRunner)
   at frysk.junit.Runner.runTestCases(TestRunner)
   at TestRunner.main(TestRunner)

Tests run: 1,  Failures: 1,  Errors: 0

It looks as frysk testsuite goes the same way as gdb testsuite where many
testcases have instable results depending on various races.  Any timeout should
be >=300 seconds and it should be exhausted only in the FAIL case.
Comment 1 Jan Kratochvil 2007-01-12 09:14:06 UTC
Created attachment 1490 [details]
Temporary timeout increase to at least 60 seconds.

Test started to behave the same way in gdb as without it.
Comment 2 Andrew Cagney 2007-01-19 22:39:48 UTC
A 300 second timeout would mean waiting 5 minutes for each test to fail, outch!

Since each frysk test (ignoring the stress tests which are not run by default)
should complete in a second or so, a 5 second timeout is sufficient as a default.

For the cases where something such as gdb, or the logger, is slowing things
down, a [recently added] -timeout N option is available.