On 12. 08. 19 16:19, Zbigniew Jędrzejewski-Szmek wrote:
On Tue, Aug 06, 2019 at 09:35:50PM +0200, Petr Stodulka wrote:
> Hi guys,
> as discussion was started week ago, Python2 is dying. As that, some
> dependencies of mercurial will be orphaned soon (or they are already)
> and mercurial as it is has to move in weeks on Python3. As I wrote
> in [0], I already started some testing and investigation.
>
> Currently it seems that with ugly hacky fix, we are able to run
> somi-working mercurial with Python3. I did just simple testing
> that worked for me and in the latest copr-build (below), it seems
> that hgk extension is workin as well. But many of you extensions
> will be probably broken. So, I guess the most probably mercurial
> will be broken for the others who use it.
>
> So it's question, should I rebase it in rawhide and setup for Python3
> already even when it is so broken, or should I wait several weeks yet
> for additional fixes?
I think it's reasonable to ship in rawhide, if the basic clone
operations work. This will satisfy 99% of use cases, i.e. cloning
of repos from the web without any further interaction with the VCS.
I don't have any data to back this up, but since Mercurial has lost
a lot of it's erstwhile popularity, based on how things go with
other python2-only projects, chances are that most extensions will not
be updated in time for F31. And it's better to have somewhat functional
clone operation than nothing.
I suggest to switch after the branching in F32 only and than (re)consider doing
a backport before the F31 beta freeze.
--
Miro Hrončok
--
Phone: +420777974800
IRC: mhroncok