On Wed, 2019-06-19 at 19:21 +0200, Ján Tomko wrote:
On Tue, May 28, 2019 at 01:51:06PM +0200, Andrea Bolognani wrote:
> On Tue, 2019-05-28 at 13:34 +0200, Daniel Veillard wrote:
> > On Tue, May 28, 2019 at 01:14:19PM +0200, Ján Tomko wrote:
> > > This reverts commit 8a1179831b5edc0a3590489eda693914fc0ff94f.
> > >
> > > It bumped the version prematurely, before the EOL of Fedora 28.
>
> Yeah, that was my bad :( I'll be more careful about releases having
> actually hit EOL before dropping support for them in the future.
I'd say for Fedora's case, we could happily wait even for half a year
more or so - by dropping it here we did not really gain anything - all
the libraries needed for libvirt to work on F28 are still way newer than
stuff from the long-term supported distros.
I see where you're coming from, but at the same time we don't have
the hardware capacity to keep more than two (three, if you count
Rawhide) versions of Fedora at the same time on
ci.centos.org, so it
would be a bit disingenuous to keep allowing RPM builds to succeed
months after we've considered the target out of support based on our
target platform policy and even stopped running builds on it...
I think a reasonable compromise would be to remove support in the
release *after* the one during whose development cycle the Fedora
project itself has dropped support for the OS.
--
Andrea Bolognani / Red Hat / Virtualization