Ricardo Wurmus writes: > Ricardo Wurmus writes: > >> Leo Famulari writes: >> >>> On Tue, Feb 21, 2017 at 06:30:12PM +0100, Ricardo Wurmus wrote: >>>> From 1ac5166df11766b47cd1ac723a464063a89afc96 Mon Sep 17 00:00:00 2001 >>>> From: Ricardo Wurmus >>>> Date: Tue, 21 Feb 2017 18:28:21 +0100 >>>> Subject: [PATCH] gnu: python-dendropy: Disable failing tests. >>>> >>>> * gnu/packages/bioinformatics.scm (python-dendropy): Disable two failing >>>> tests. >>> >>> LGTM >> >> The problem was fixed in the upstream commit 93f984b in response to my >> bug report: >> >> https://github.com/jeetsukumaran/DendroPy/commit/93f984bba7a6c588a28ca87f4e557ce283809453 >> >> I believe we can just backport this fix instead of deleting the test. > > Here’s a patch that fixes the tests by applying the changes of the > upstream commit. If this is fine I’ll push it to the python-tests > branch. I tried this patch, and guess the reason it was not pushed is because dendropy now has a different failure. Can you look into it? Currently, I can build all of 'guix package -A python' on amd64, and have fixed the few i686 problems on Hydra. Except for these three: python-dendropy python-openid python-axolotl Given that the two latter packages have security implications I would rather leave them broken, than ignoring the test failures. They don't have any in-tree users. I don't always break packages wittingly, but when I do, it's on 'master'. Should we do a final Hydra round, or just merge it to the master branch first? The 'python-tests' branch haven't been synced in a while, so many substitutes will be out of date.