Hi,
On Mon, Jun 20, 2022 at 01:45:07PM +0200, Tomas Hrnciar wrote:
[...]
If your package fails because there is a non-dependency problem, you
might have already received a bugzilla from us in the past. If the build
failure is related to changes in Python 3.11, it should contain some
hints about the problem.
[...]
## My package fails to build because it has test failures in %check
Please, try to resolve the failures. If you are confident that the package
works fine, but the tests are wrong, skip some failing tests, ideally with a
link to an upstream issue. Do not disable (e.g. comment out) all tests just to
unblock the rebuild of your package, it usually only hides the problem.
[...]
## Where to get help
Reply to this thread or find us (thrnciar, mhroncok) on #fedora-python IRC
channel (libera.chat)or Matrix (#python:fedoraproject.org).
---
Let us know if you have other questions.
Maintainers by package:
[...]
python-migen somlo
[...]
There's currently
https://bugzilla.redhat.com/show_bug.cgi?id=2069537
(about the failing test), and
https://github.com/m-labs/migen/issues/259
for the upstream issue I filed.
Some preliminary analysis (thanks to @music) indicates the problem
is due to the restructuring of CPython bytecode w.r.t. CALL_*
instructions.
If anyone reading this is deeply familiar with CPython and willing to
take a quick look, that'd be much appreciated!
Otherwise, what's the avilable timeframe I have to try and figure
it out in time for F37 (assuming upstream doesn't start caring about
Python 3.11 *before* then)?
Thanks much,
--Gabriel