On 20/10/2019 00:15, Robie Basak wrote:
On Sat, Oct 19, 2019 at 11:59:45PM +0200, Nicolas Dandrimont wrote:
2. while python3-trio has passed new, it's maintainer (Robie Basak) has filed a
RC bug saying he does not belive it is currently suitable for a stable release.
In part I just followed on from the opinion of the person who filed the
ITP originally (I took it over).
Thanks for the info.
I've only glanced at the upstream issue for a moment, but I have the feeling
that's a little bit too conservative: the only difference between this and most
other Python modules packaged in Debian is that upstream actually seems to care
about their API being unstable. We've shipped way worse stuff, and with
autopkgtests it's fairly easy to notice that you're breaking a reverse
dependency. Of course I'll defer to the maintainer's judgement.
Thoughts on the best way forward? should the upload of 0.6.1 be reverted?
ISTR the trio stuff is not needed for most of msrest, only some of the async
stuff. If Robie confirms that he doesn't want trio in testing, I'll look into
patching it out of msrest.
I don't intend to make a decision on this alone unless upstream declare
the API stable. I welcome discussion on the bug itself;
I have added 913...@bugs.debian.org to cc.
I don't think
it's appropriate for this discussion to remain private.
It's not private, these mails were set to "python-msr...@packages.debian.org" which
forwards to "python-modules-t...@lists.alioth.debian.org". This mail thread can be found
at https://alioth-lists.debian.net/pipermail/python-modules-team/2019-October/059368.html.
I'm open to
closing the bug if that's the consensus _and_ everyone involved,
including upstream, has had an opportunity to comment.
I have posted a summary at
https://github.com/python-trio/trio/issues/1#issuecomment-544304785 and asked
for an upstream opinion.
I notice that python-msrest has already been included in ubuntu eoan, so
version 0.11.0 is going to be around for a while in at least one major Linux
distro whatever we do now in Debian.
