Experimental Physics and
Industrial Control System

Andrew Johnson via Core-talk <[email protected]> · Mon, 17 Jun 2019 22:16:13 -0000

Disabling individual tests on the specific architectures where they fail isn't something we have done in the past. I'm not saying it can't be done, but that merely removes the test lockups from ever being seen. When tests fail to exit that mostly only affects me because I have to see that they've been running for hours and kill that job on the APS Jenkins server.

I would like to make it easier for test authors to build safeguards like timeouts into their code. If a test never fails for an author they aren't as likely to want to work on fixing the actual problem, and adding a timeout is an easy workaround that would save work for me. A test program will be reported as failing if it times out, whereas if I have to kill the job the test results don't get analyzed at all so the author doesn't see the problem (Jenkins doesn't send email when I kill a job).

In a perfect world I might agree that epicsThreadJoin() doesn't need a timeout, but it could make life a little easier in this world IMHO.

If it's too hard to implement a timeout on Posix I would accept that as a reason for not adding it, but I still think that epicsThreadJoin() should return a status, and epicsThreadMustJoin() should be added with the call to cantProceed() — this has been our practice in other APIs e.g. callocMustSucceed(), epicsMutexMustCreate(), epicsMutexMustLock(), epicsEventMustWait(), epicsEventMustTrigger(), epicsSpinMustCreate().
-- 
https://code.launchpad.net/~epics-core/epics-base/+git/Com/+merge/361379
Your team EPICS Core Developers is subscribed to branch epics-base:7.0.

Experimental Physics and Industrial Control System

Experimental Physics and
Industrial Control System