On Fr November 2 2007, Michael Schwendt wrote:
On Fri, 02 Nov 2007 10:36:01 +0100, Till Maas wrote:
> Imho it should be valid to have a package in F7 updates-testing with a
> higher nevr than in F8.
Why isn't the same update built also for F8?
I did not say, that the update is not built for F8, but it will be in F8
updates-testing and not in F8.
Why do you want to limit the testing to one dist?
I want to have testing simultaneously for two dists, with a package in F-7
updates-testing and F-8 updates-testing. But then there will be no upgrade
path from F-7 updates-testing to F-8. There will be one to F-8
updatest-testing.
> Afaik targets updates-testing at experienced users
And rawhide also targets experienced users, doesn't it?
Yes and not every rpm/update in Rawhide makes it into the next stable release
and there are also rpms that are removed from rawhide without adding a newer
version afaik.
> who are
> willing to test packages and should be able to handle situations, where
> there is no upgrade path to the next stable release.
The fun will start when we get multiple updates-testing repositories.
So far, F7 updates-testing has been the only one.
For some time we've excluded F7 updates-testing from the upgradepath
check, only to notice that F8 test1 got nearer and nearer with
packagers not preparing their updates (often even version upgrades)
for 'devel' at all. Unless it's a version upgrade, it's always
possible to update an old dist without breaking the upgrade path
(just increase %{release} appropriately).
Imho the only solution here is to warn the maintainer when a push to stable is
requested and this would break the upgrade path. But when there is a valid
upgrade path from F-7 updates-testing to F-8 updates-testing (but not from
F-7 to F-8), then this should not be a problem.
Regards,
Till