<lavamind> hello
<lavamind> is there a known issue with regards to OpenSSL "noverify" connections?
<lavamind> I've been seeing flaky tests related to it, like TestOpenURISSL#test_validation_noverify
<enebo> Not that I know of
<enebo> you just opened the issue and I see you are looping by calling successive jruby runs and not looping in an individual runtime
<enebo> So it seems like it indicates some level of transience
<enebo> lavamind: 9.4.6.0 also shipped with a newer version of jruby-openssl but not a ton changed. Still worth repeating just to be sure
<lavamind> enebo: the failure happens on "regular" test runs also
<lavamind> which I've been chasing fixes by hitting the Retry button
<lavamind> I can try and pull the latest jruby-openssl to see if that helps
<enebo> I am just checking our release notes for 9.4.6.0 because we did fix something with IO that might affect this
<enebo> I wonder if this is involved
<enebo> I am up to 93 runs with no fail but I am using Java 11 on Fedora using effective a few commits past 9.4.6.0 (which would not affect IO or networking at all)
<enebo> I will let this run a while longer then try much newer java
<enebo> I stopped at 140 good runs
<enebo> lavamind: I hate to say WFM but I am not use 9.4.5 or on your system. If you are making a package for 9.4.5.0 you may just want to skip to 9.4.6.0 (make sure it is actually working first) since it is out now
<lavamind> yeah I understand
<enebo> lavamind: also kares is openssl guy lately and headius made the most recent changes to IO and they both are on our matrix (I use app.element.io as web client.
<lavamind> I'm going to try with that patch
<lavamind> ah ok gotcha
<enebo> ok yeah I am not positive but that logic was backwards so it gives me some hope it could be the problem
<enebo> but it is amazing if we had blocking as non-blocking and things are so quick we don't see a problem most of the time
<lavamind> so 9.4.5.0 with the patch and jruby-openssl 0.14.3 failed the same way on try #13