Re: heads up: poppler soname bump

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

 



On Sat, Aug 5, 2017 at 10:49 AM, Zbigniew Jędrzejewski-Szmek
<zbyszek@xxxxxxxxx> wrote:
> On Sat, Aug 05, 2017 at 04:15:24PM +0900, Mamoru TASAKA wrote:
>> Adam Williamson wrote on 08/05/2017 05:00 AM:
>> >On Wed, 2017-08-02 at 13:19 +0200, David Tardon wrote:
>> >>Hi,
>> >>
>> >>I will build a new release of poppler this week, which includes soname
>> >>bump. I will take care of rebuilding the affected packages:
>> >>
>> >>boomaga
>> >>calligra
>> >>cups-filters
>> >>evas-generic-loaders
>> >>gambas3
>> >>gdal
>> >>gdcm
>> >>inkscape
>> >>kf5-kfilemetadata
>> >>libreoffice
>> >>okular
>> >>pdf2djvu
>> >>poppler-sharp
>> >>texlive
>> >>texworks
>> >
>> >Welp, the rebuild of texlive failed, which means gdal can't be built,
>> >and I can't build openqa. And none of this seems to have been touched
>> >since yesterday,
>>
>> Umm? Did you really check?
>> https://koji.fedoraproject.org/koji/builds?order=-completion_time&userID=803
>>
>> >so now it's probably going to be stuck over the
>> >weekend, right?
>> >
>> >Perhaps next time, you could test rebuilds of at least major things
>> >like texlive against the new poppler *before* you land the new poppler
>> >into rawhide? Thanks.
>>
>> And then, if test rebuild of such a big package like texlive against
>> new poppler fails, poppler maintainer has to wait until texlive is
>> to be ported into new poppler?
>
> Yes! That's how the new "no alphas" rawhide is supposed to
> look. Maintainers are supposed to avoid a state where rawhide is
> broken.
>
> poppler shouldn't be *blocked*, but it should be delayed a bit, so
> that maintainers of the dependent packages are given a few days to
> find a resolution.
>

This is garbage. How the heck are we supposed to do crap like this
when Koji itself isn't built to support this kind of thing freely? If
anything, Koji (or something) should be doing the bloody rebuilds for
us, automatically doing a scratch build and if that passes, do a "real
build" with the bumped Release and changelog entry.

We don't have a concept of projects in Koji where people can do these
things without affecting the Rawhide package set, and COPR has no
dependency tracking to indicate what depends on it in other
repositories/linked projects, much less auto-rebuilding.

For example, SUSE's OBS gives you this information on the package
detail page. For example:
https://build.opensuse.org/package/binary/openSUSE:Factory/rpm-python?arch=x86_64&filename=rpm-python-4.13.0.1-5.1.x86_64.rpm&repository=standard

(And of course, OBS automatically bumps packages and rebuilds for you,
so that you *don't* get stuck with stupid grunt work like this!)

We keep pushing for this stuff, but we're not even bothering to make
it easier for packagers to keep up. What about the poor souls who
aren't proven packagers? They're out of luck and they get blamed for
"breaking Rawhide".

I implore someone, please fix this before we keep going down this
path! I'm not against the "no more alphas" thing, but it seems like
everyone who is "for" this has forgotten that the overwhelming
majority of packagers are not actually capable of dealing with the
fallout of that kind of change.


-- 
真実はいつも一つ!/ Always, there's only one truth!
_______________________________________________
devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to devel-leave@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