On Wed, Oct 09, 2019 at 08:50:44PM -0500, Richard Shaw wrote:
On Tue, Oct 8, 2019 at 3:54 AM Zbigniew Jędrzejewski-Szmek <
zbyszek(a)in.waw.pl> wrote:
> On Tue, Oct 08, 2019 at 08:32:47AM +0200, Ralf Corsepius wrote:
> > >
> > >Those are fairly substantial changes, but time is of essence here.
> > I could not disagree more. Quality and stability is of more essence,
> here.
>
> Richard is working on updating Coin to the latest version along with the
> dependent packages. The PRs were for rawhide. I don't think there's much
> choice: we need to update to latest versions of packages and rawhide is
> the appropriate place to do it, and we are early in the release cycle.
>
So the PRs were for Rawhide, but the bug I'm trying to fix exists on all
supported Fedora releases. I wasn't planning on updating F29 at this point
but F30 does have a lot of life left.
I don't like the idea of major upgrades within a release but the list of
dependencies (as noted by the list of PRs) is fairly small and through my
COPR I have found no *build* issues with the update.
I'm open to suggestion here but I don't like leaving broken software in
Fedora and basically having to tell the user, "It's fixed in Rawhide so
you'll get it eventually..."
Thoughts?
Let's get everything built and tested in rawhide first. Based
on how this turns out, an update in F30 and F31 might be appropriate
(with a suitably long testing period, etc). I agree that if the version
in stable Fedora is sufficiently broken, it's better to release an
working update with major version changes than to do nothing.
Zbyszek