qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [PATCH] configure: Require Python >= 3.5


From: Kevin Wolf
Subject: Re: [PATCH] configure: Require Python >= 3.5
Date: Fri, 18 Oct 2019 10:15:36 +0200
User-agent: Mutt/1.12.1 (2019-06-15)

Am 17.10.2019 um 21:39 hat John Snow geschrieben:
> On 10/17/19 7:21 AM, Kevin Wolf wrote:
> > Am 17.10.2019 um 00:48 hat John Snow geschrieben:
> >> On 10/16/19 6:42 PM, Eduardo Habkost wrote:
> >>> Python 3.5 is the oldest Python version available on our
> >>> supported build platforms, and Python 2 end of life will be 3
> >>> weeks after the planned release date of QEMU 4.2.0.  Drop Python
> >>> 2 support from configure completely, and require Python 3.5 or
> >>> newer.
> >>>
> >>> Signed-off-by: Eduardo Habkost <address@hidden>
> >>
> >> Seems like a good time and place to mention this. Kevin, you require
> >> 3.6+ for iotests, which are -- at present -- invoked as part of "make
> >> check".
> >>
> >> Do we care? Basically, this just means that iotests won't run for
> >> systems that don't have 3.6+, which would be platforms like Debian 9 --
> >> which is why ehabkost is choosing 3.5 here.
> > 
> > I think we were aware of this when we made the change to iotests. That
> > all tests of the current upstream QEMU version are run on Debian
> > oldstable (with the distro Python version) is, to say the least, not a
> > priority for me. They must not fail, but I'd say skipping is fine.
> > 
> > And actually, we should still have a reasonable coverage there with the
> > shell-based test cases.
> 
> This seems like a weirdly arbitrary decision for a benefit that's not
> clear to me. Is it because you want variable annotations?

Yes, the discussion about type annotations is what made me check whether
we could do 3.6, because if we want to make use of type checking, we'll
need it for both functions and variables to get reasonable results.

And actually, we currently don't have any Python tests in the auto
group, so the only effect is for people manually running ./check on
Debian oldstable. I'm not sure, but I suspect this might be the empty
set.

Kevin



reply via email to

[Prev in Thread] Current Thread [Next in Thread]