That’s fine. My general approach is to merge things as they’re ready. That’s the point of continuous integration. =) Would you like to have a stab at the rebase? If you ping me here I’ll review ASAP.CC list: sorry, I forgot to reply-all earlier. Full (tiny) thread below.On 25 Mar 2017, 12:09 PM -0400, Nelle Varoquaux <nelle.varoquaux@gmail.com>, wrote:His point was that backporting would be easier if it was merged before.On 25 March 2017 at 09:03, Juan Nunez-Iglesias <jni.soma@gmail.com> wrote:Oh! I thought the consensus was to have it *after* the release! #releasemanagerfail =P But it’s still on the 0.14 milestone. And looking at the comments it’s not clear that he wanted that? Anyway, I’m personally happy to merge if a rebase fixes the failing travis build.On 25 Mar 2017, 11:57 AM -0400, Nelle Varoquaux <nelle.varoquaux@gmail.com>, wrote:On 25 March 2017 at 08:37, Juan Nunez-Iglesias <jni.soma@gmail.com> wrote:Hi everyone,We’ve had these two 32-bit blockers holding up 0.13 for a couple of months now. Importantly, both of these bugs:- existed in 0.12- are only testing bugs, not actual bugs (as far as I can tell)Therefore, I’ve proposed to ship 0.13.0 before fixing them. When we do fix them, we can back-port to 0.13.1/2/3. Stéfan was on board with this plan. If there are no objections, I’ll get the ball rolling shortly on the release. But, I wanted to give people a chance to comment on the decision before starting. =)Stéfan wanted my pytest PR in before the release. Is that still the case?Cheers,NJuan._______________________________________________ scikit-image mailing list_______________________________________________scikit-image mailing list