On Fri, 24 Jun 2016, Andrew Cagney wrote:
It is important not to shutdown when you want to manually debug single test.
The current setup is easier to leave the host running and manually login debug
further.
I have a bias to keep doing this. Also currently when runng with TESTLIST it is
alway shutdown.
What is missing in this model?
This lacks consistency, the test results should be the same whether
the test is run individually or as part of a batch.
Why is it capturing the logs of the shutdown? If the test does not do a
shutdown, and final.sh has finished, there should be no further
relevance to the output.
As my original post tried to point out, if pluto crashes (not aborts)
while the machine is shutting down we'll miss that - I just happened
to notice the problem because I had an inconsistency between some
counters.
As long as it shows up in the tests that _do_ test pluto shutdown,
that's fine.
Perhaps we do need a special post-processing for the leak-detective
logs, which are generated on pluto shutdown.
If we really want to capture any cores that happen on shutdowns
not specifically tested, perhaps we should move dumping core from
/var/run/pluto to /testing/pluto/testname/OUTPUT/ although we would have
to make sure selinux allows core dumping there.
Paul
_______________________________________________
Swan-dev mailing list
[email protected]
https://lists.libreswan.org/mailman/listinfo/swan-dev
