You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Just a guess, but maybe CI missed it because there was a four day gap between the CI run and when the commit was landed. So maybe some other change landed during that time. And that change by itself is fine. And this change by itself is fine. But the two changes together mean failing test.
That's the first explanation that leaps to mind, at least...
Probably I just don’t know enough about the magic behind the CI, but does cctest even run on CI? I can’t find anything that looks like it in the console output.
Ah. that explains why the CI missed it. I can still reproduce the issue with the latest master. I believe @eugeneo is planning to investigate the failure itself.
@bnoordhuis: Correct. We could look at parsing xunit output through a plugin? The difference between xunit and junit (what Jenkins natively prefers) seems pretty small. Stackoverflow suggests awk:ing it which we also could give a go.
@bnoordhuishttps://github.com/bnoordhuis: Correct. We could look at
parsing xunit output through a plugin? The difference between xunit and
junit (what Jenkins natively prefers) seems pretty small. Stackoverflow
suggests awk:ing it which we also could give a go.
Activity
Trott commentedon Aug 8, 2016
Just a guess, but maybe CI missed it because there was a four day gap between the CI run and when the commit was landed. So maybe some other change landed during that time. And that change by itself is fine. And this change by itself is fine. But the two changes together mean failing test.
That's the first explanation that leaps to mind, at least...
Trott commentedon Aug 8, 2016
Oh, wait, if CI is still missing this than my speculation isn't the issue. Sorry for the noise.
bnoordhuis commentedon Aug 8, 2016
FWIW, I could initially reproduce with a build from last week but after updating to today's HEAD of master and rebuilding, cctest passes again.
jbergstroem commentedon Aug 8, 2016
OS X version discrepancy in CI vs local perhaps?
addaleax commentedon Aug 8, 2016
Probably I just don’t know enough about the magic behind the CI, but does cctest even run on CI? I can’t find anything that looks like it in the console output.
bnoordhuis commentedon Aug 8, 2016
It doesn't seem to be part of the test-ci target. For that matter, gtest doesn't support tap output, AFAIK.
ofrobots commentedon Aug 8, 2016
Ah. that explains why the CI missed it. I can still reproduce the issue with the latest
master. I believe @eugeneo is planning to investigate the failure itself.jbergstroem commentedon Aug 8, 2016
@bnoordhuis: Correct. We could look at parsing xunit output through a plugin? The difference between xunit and junit (what Jenkins natively prefers) seems pretty small. Stackoverflow suggests awk:ing it which we also could give a go.
eugeneo commentedon Aug 8, 2016
I'm looking into the test failure.
On Mon, Aug 8, 2016 at 8:49 AM Johan Bergström notifications@github.com
wrote:
test: fix failing inspector cctest
test: fix failing inspector cctest