> Hi Patrice,
>
> I'm afraid Ralf is right here. The details (especially small details)
> of upstream specs are basically irrelevant.
As I said above I agree that it is more important to follow guidelines,
but being in sync with upstream spec is an added bonus.
> If you go back through the email list archives there was a long
> discussion about not relying on the fact that:
>
> fc3 < fc4 < fc5
But... I always rely on that. Especially when packages are kept in sync
for all branches. And even more when there is a need to modify a previous
branch. For example, I have
fc3:
2%{?dist}
fc4:
2%{?dist}
And then I modify the fc3 package. What I did previously was to add .1 to
the fc3 release such that I get
fc4:
2.fc4
fc3:
2.fc3.1
such that there is 2.fc3.1 < 2.fc4 for upgrades.
And if I remember well Ville gave me the trick, and if I'm not wrong Spot
uses the %{?dist} tag to keep spec files in sync for all the branches...
One solution could be to ensure that we never get from a situation where
in fc3 and fc4 there is the same release number (ie bump release twice
on devel and once on FC-4 just after the import of a package in cvs).
This should really be stated somewhere.
This would also render the dist tag in release only informational.
If so, it should be stated explicitely on the page
http://fedoraproject.org/wiki/DistTag
> when rpm and yum (and other tools) do comparisons of the EVR. So, you
> are in fact violating our long-debated (way too long-debated, IMHO)
> policy.
I haven't found the thread, looking in gmane, do you have an idea of
the date of the thread or words I could use to find it?
--
Pat
--
fedora-extras-list mailing list
fedora-extras-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-extras-list
