Closed fgregg closed 7 years ago
Yup, that would explain why the PR build worked but the branch build failed. At this moment, the 'manylinux27' branch is 6 commits behind master.
Curiously it doesn't explain why the appveyor pr build failed. It looks like both Travis and Appveyor do the same git commands to merge master and the branch for the pr build.
Ugh, I hate merges that have merge commits. Any particular reason you didn't git rebase like I asked?
sorry, let me try this again
I'm not sure what's going on with the windows 3.5 builds, but those failures started here e348415f3ea22b3ea396ab48c22708b1cdafc3d5
I don't think the Appveyor failure should block this merge.
Wait, I thought for some reason that this PR didn't touch appveyor.yml
. Let me take a look.
Oh, I see you've fixed Appveyor in #17.
It looks like you'll need another rebase, since this PR will conflict with #17. On the plus side, you won't need to touch appveyor.yml
at all in this PR once #17 is merged.
Could we have a new zope.index release since it seems like there has been a number of changes beyond the manylinux bits.
Will do!
Strange: the previous master build succeeded (and was running
zope-testrunner --test-path=src
), while this build fails while runningpython setup.py test -q
. The PR is not modifying.travis.yml
, so how come the test command is different?Perhaps you accidentally based your PR on an older commit? Could you rebase your branch on latest master and force-push?