Re: Summary of the 2008-04-08 Packaging Committee meeting

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

 



Jason Corley wrote:
Toshio Kuratomi wrote:
Which I think illustrates the point that JPackage isn't the upstream for the tomcat5 package, yes?

Tom "spot" Callaway wrote:
It seems far more appropriate to say that the Fedora package is derived from JPackage.

Perhaps you two can explain to me the difference between what you
consider package derivation vs. upstream in the context of
packaging... In my mind if you base your package off of the packaging
work of others, those others are (or in this instance should be)
upstream for the package.  Again the Fedora/RHEL analogy seems to fit.
 It feels like splitting semantic hairs to me but you may have a real
distinction which I'd like to hear.  If there is debate there that is
likely the part I'm not understanding, but then I'm just a pig
wallowing in mud so what do I know.

If we were talking about source code it would be the difference between having an upstream and having a fork. Even though xorg and XFree86 started with the same code base, one is not the upstream for the other.

So in this context, derivation would be if I took a package from JPackage, Mandriva, etc, had it reviewed for Fedora and then proceeded to maintain the package in Fedora, syncing against new upstream releases of the source, fixing bugs reported in the Fedora bugzilla, and generally, considering the package I'm maintaining in Fedora to be independent of the package the original work was based on.

Upstream in this context would be when I take a package from an upstream repository and submit it for Fedora review. Changes implemented in the review would be pushed back into the upstream package unless they were truly Fedora-only changes. Once imported and built the normal flow of events for the package would be to wait for changes from the new upstream package release (or aid in making new upstream package releases) and then syncing those to the Fedora package. As packaging bugs were filed with Fedora, the changes could be made locally and pushed to the upstream repository's packages or made in the upstream's packages and then backported to the Fedora package until the next upstream release.

So using the term upstream says there's an ongoing relationship between the packages where the Fedora package tracks the changes made in the upstream package.

The grey area is when a Fedora package tracks changes in another repository but the packager still thinks of it as an independent work. This happens more often when a bug occurs in source code and the packager looks for patches in other distros/repos that can fix the problem. It could happen in the context we care about here but I don't think it will happen as often. A package maintainer is used to looking for help with source code upstream but doing all the work of packaging themselves -- so a packager who is deriving from JPackage would be more likely to rely on their own resources or by asking fedora-devel list for advice while a packager that is using JPackage as an upstream would look to JPackage's cvs for ideas.

-Toshio

--
fedora-devel-list mailing list
fedora-devel-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/fedora-devel-list

[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