Re: RFC round 3: DEP-3: Patch Tagging Guidelines
On Wed, Jul 15, 2009 at 06:06:38PM -0400, Lucas Nussbaum wrote:
> I think that this information should be stored outside of the patch (in the
> history of a VCS, for example).
Can you explain how you envisage such a thing working, outside of the patch?
Not all packages are maintained inside a VCS.
The system I envisaged in my proposal was (back of the envelope) summary
information generated like this
* there are <X> debian packages, <Y>/<X> using a recognisable patching
system, constitution <Z> patches. <W>/<Z> patches use the patch
* <V>/<W> of these patches have been reviewed.
* <U>/<V> of these patches have been reviewed less than <T> days ago (where
<U> is some number defined to try to describe acceptable patch "staleness")
and a list of patches in need of review <W> - <V>
and a list of patches last reviewed too long ago <V> - <U>
And a system for people to search out such patches by package, maintainer, or
other criteria, to facilitate reviewing.
Such figures could be determined purely from a source archive if the
information was encoded in the headers. It would not be feasible I don't think
to do something like this if the system needed to query lots of different
types of VCSes or other systems too.
I suppose a system could be put together that stored all of the patch review
data inside of itself.