Join GitHub today
FIX: No more workaround #387
@@ Coverage Diff @@ ## master #387 +/- ## ========================================== + Coverage 93.32% 93.33% +<.01% ========================================== Files 27 27 Lines 1949 1951 +2 ========================================== + Hits 1819 1821 +2 Misses 130 130
Jun 5, 2018
5 checks passed
I should have tested this on sphinx master, it looks like the problem is still here: https://travis-ci.org/sphinx-gallery/sphinx-gallery/jobs/388594944 (sphinx master is only tested during the daily Travis cron job).
The reason I put it in a daily Travis cron job is that in 99% of the cases, it's not the fault of the PR if the build with sphinx master fails at all. Also it's not really urgent generally so we can leave it fail for a few days, whereas I find having all PRs red for a few days without a good reason is annoying.
To be honest: I just followed what we do in scikit-learn for numpy, scipy, cython master and that has worked quite well for us so far.