Seems that the last release of flake8-quotes broke it (?).
On one hand we are testing on test_flake8.py code that is not within p.r.codeanalysis but instead on flake8 plugins (flake8-quotes just being one of them).
On the other hand we are not pinning a version thus being exposed to these kind of failures.
I would say that we have to do both:
pin versions and bump them on a pull request to ensure master is always green
(re)move the tests on p.r.codeanalysis and put them on the upstream flake8 plugins, and fix the issues there
Travis is failing on master: https://travis-ci.org/plone/plone.recipe.codeanalysis/jobs/90321940
Seems that the last release of flake8-quotes broke it (?).
On one hand we are testing on test_flake8.py code that is not within p.r.codeanalysis but instead on flake8 plugins (flake8-quotes just being one of them).
On the other hand we are not pinning a version thus being exposed to these kind of failures.
I would say that we have to do both:
Opinions?