Patch metadata (Was: Plan for tomorrows (20080424) FESCO meeting)

[Jason L Tibbitts III]

>>>>> "CW" == Colin Walters <walters at verbum.org> writes:

CW> What's the ultimate problem you want to solve with that metadata?

The recurring issue: users and upstream want to know how we're
patching the software we distribute, and it's rather difficult to
figure that out.  PackageDB gets them to CVS, but reading specfiles or
poring over a CVS listing of patches is a bit difficult for many.
Even without such metadata it's possible to present a list of patches
which are applied, and I intend to pursue that in any case.

CW> I'm worried that if we require extensive/structured metadata,
CW> people won't do it.

I hadn't intended to require it, although it's certainly a topic for
discussion.  I had only intended to make it possible to present such
data in an accessible manner.  Even in the simple case, it's possible
to extract comment lines from the specfile above the Patch: lines and
treat them as comments.  But structured metadata enables us to do
things like present links for the patches to the upstream trackers, to
the CVE pages for security issues fixed by the patches, etc.

 - J<