TestEvents test case causes python runner to dump core

Hi all,

In a build on 64-bit linux from TOT, I'm experiencing the test case
"TestEvents" causing python to segfault.

I believe this to be a fairly recent regression. Anyone else seeing
this/or know what's responsible?

I'm starting to analyse this...

Matt

Member of the CSR plc group of companies. CSR plc registered in England and Wales, registered number 4187346, registered office Churchill House, Cambridge Business Park, Cowley Road, Cambridge, CB4 0WZ, United Kingdom
More information can be found at www.csr.com. Keep up to date with CSR on our technical blog, www.csr.com/blog, CSR people blog, www.csr.com/people, YouTube, www.youtube.com/user/CSRplc, Facebook, www.facebook.com/pages/CSR/191038434253534, or follow us on Twitter at www.twitter.com/CSR_plc.
New for 2014, you can now access the wide range of products powered by aptX at www.aptx.com.

Sorry, I forgot to mention, the core dump emanates from the failed
assertion below:

UNSUPPORTED: LLDB (gcc-x86_64) ::
test_add_listener_to_broadcaster_with_dsym (TestEvents.EventAPITestCase)
(requires Darwin)
python: /home/mg11/src/vendor/llvm/tools/lldb/source/Plugins/Process/POSIX/ProcessPOSIX.cpp:270: virtual lldb_private::Error ProcessPOSIX::DoResume(): Assertion `state == eStateStopped' failed.
Aborted (core dumped)

Matt

If I revert Greg Clayton's change r220113, then the core dump problem
vanishes.

I also found that running with r220113, but with the invocations of
SetAsync(True) removed, see below:

I don't know what is going on on the Linux side, but r220113 is a patch that only changed the test case. It really shouldn't be possible to cause a core dump by driving the SB API's. Plus, what the test is doing is something the SB API's should support. So I think this change is not in itself the bug, but rather it is showing up a real bug somewhere in the Linux process support. IMO the correct solution should not be to revert the test change, but to figure out what is going on in the underlying Linux support.

Not sure Greg is the proper person to do that, but anyway it seems to me the correct thing to do for the nonce is skip the test on Linux and file a PR to get to the underlying cause.

Jim

I filed llvm.org/pr21325 for TestEvents.EventAPITestCase dumping core
on FreeBSD. I'd suggest just adding a comment that this happens on
Linux as well.

I apologize for the breakage. If anyone can help out and try to figure out what is happening on Linux or FreeBSD, please do help out. I don't have a linux box and this test should work. The "sync" mode was completely broken prior to this fix, so it would really help out LLDB in general if we can get a rock solid "sync" mode.

Greg

Thanks Greg - I don't mean to imply that submitting the PR was the
extent of my involvement here. For now I just wanted to get the
testrun to stop dumping core and get the buildbot back to green to
keep on top of new failures, but hope that I (or one of the Linux
people) will have some time to look at it a bit later on.

-Ed