Branko Čibej wrote on Sat, 16 May 2020 13:46 +0200: > On 16.05.2020 12:00, Stefan Sperling wrote: > > On Fri, May 15, 2020 at 09:13:38PM +0000, Daniel Shahaf wrote: > >> IIRC the soak was to end on Wednesday and any change requires at least > >> a week's extension. > >> > >> Here's the patch. I smoke tested it and it passed, but I'm seeing tests > >> randomly fail even in unpatched trunk so I can't an unqualified +1. > > I would prefer if we just nominated this for 1.14.1 as usual. > > > > This is something that is very, very easy to work around for packagers. > > And I don't think we are seeing a lot of people making their own builds > > from source these days, anyway. And even then, in the worst case they ask > > for help and it's easy to solve. > > > > So I see no reason to fix this for .0, not even with the hypothetical case > > in mind where we would drop python2 support to a stronger degree than > > we already have. > > > > If it is deemed very important, we could just document this in release > > notes under "Known issues" until 1.14.1 comes out. > > Sure, that makes sense, too. After all, we have a bunch of > python-releated fixes already on trunk that are easily candidates for .1.
Done in r1877835. Revisions welcome. ☺
