Re: python F23 update testing needed

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

 



On 04/07/2016 07:28 AM, Jan Pokorný wrote:
On 02/04/16 13:58 +0200, Jan Pokorný wrote:
On 02/04/16 12:26 +0200, Jan Pokorný wrote:
On 31/03/16 21:28 -0600, Orion Poplawski wrote:
I've just submitted
https://bodhi.fedoraproject.org/updates/FEDORA-2016-dcddcc1f06

This update does two things:

- Updates python to version 2.7.11
- Splits out the python macros into separate packages

It would be helpful if packagers would try building python packages with
these packages installed.

This might be the cause of a recent failure to build Pacemaker while
other Fedoras (24 + 25) were OK:
http://koji.fedoraproject.org/koji/taskinfo?taskID=13522053

See
https://kojipkgs.fedoraproject.org//work/tasks/2053/13522053/mock_output.log

Package python-macros is obsoleted by python-rpm-macros, but
obsoleting package does not provide for requirements

Subsequent build succeeded, though:
http://koji.fedoraproject.org/koji/buildinfo?buildID=750257

So it must have been an intermittent issue.
Can this be prevented next time a basic build dependency
is restructured?  Or would it require additional support
in the build infra?

I tried to investigate this a little bit.

6 days ago, at around the point the build failed, several overrides
for the build root have been requested:

python-rpm-macros-3-7.fc23:
https://bodhi.fedoraproject.org/overrides/python-rpm-macros-3-7.fc23
redhat-rpm-config-36-1.fc23.1:
https://bodhi.fedoraproject.org/overrides/redhat-rpm-config-36-1.fc23.1
python-2.7.11-2.fc23:
https://bodhi.fedoraproject.org/overrides/python-2.7.11-2.fc23

All but the last, expired on 2016-04-01, appear active at the moment.
The last one was superseded with python-2.7.11-3.fc23:
https://bodhi.fedoraproject.org/overrides/python-2.7.11-3.fc23,
which is also active right now.

But clearly, at the moment of the linked build that failed,
python-rpm-macros and redhat-rpm-config were overridden whereas
python-macros-2.7.10-8.fc23+ wasn't.  It looks to me I hit a racy
window in which python-2.7.11-2.fc23 override hasn't yet been
applied.

And later, at the moment of the linked build that succeeded,
no such override was active.

I would conclude, please be more careful about management of
overrides to prevent this kind of mess-ups.

Does bodhi allow to specify several builds as a means of an atomic
(non-racy) override?   Should I file a RFE for that if not?

You have to submit each package override separately. There also was a moment when I had to submit an initial version of a couple things, and then rebuild some dependent packages.

So, there may have been a nearly unavoidable period of breakage, but it also may be useful to be able to submit a group of packages as overrides in a single submission.


--
Orion Poplawski
Technical Manager                     303-415-9701 x222
NWRA/CoRA Division                    FAX: 303-415-9702
3380 Mitchell Lane                  orion@xxxxxxxxxxxxx
Boulder, CO 80301              http://www.cora.nwra.com
--
devel mailing list
devel@xxxxxxxxxxxxxxxxxxxxxxx
http://lists.fedoraproject.org/admin/lists/devel@xxxxxxxxxxxxxxxxxxxxxxx




[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Index of Archives]     [Fedora Announce]     [Fedora Kernel]     [Fedora Testing]     [Fedora Formulas]     [Fedora PHP Devel]     [Kernel Development]     [Fedora Legacy]     [Fedora Maintainers]     [Fedora Desktop]     [PAM]     [Red Hat Development]     [Gimp]     [Yosemite News]
  Powered by Linux