Re: Lintian tag classification
Jordà Polo <jorda@ettin.org> writes:
> My view isn't that different. A fourth step I also thought about is
> enhancing lintian.debian.org to support the new classification, but it
> probably isn't a "basic" step.
Yes, that's a good point. Currently, there are several places where the
type is used in that code:
* We gather statistics for QA purposes that includes a count of errors and
warnings (separately) for each package.
* There are two views of the maintainer page, one of which showing all of
the tags and the other only showing errors and warnings.
When you change the representation of tag severities, there are a lot of
little things like that which will need tweaking. Another is the
colorization of tags, which currently is done only based on the E/W/I
status.
It would be particularly cool to support using JavaScript or something
similar to selectively hide and show tags from different sources,
certanties, or severities, but that verges on a whole additional project.
:)
> Thanks for your comments, they are really appreciated. You helped me
> solve a few misunderstandings I had, and made me realize about some
> flaws in my proposal.
Absolutely. I'm very happy to answer any questions. Please feel free to
ping me if mail goes without an answer for a few days; sometimes mail gets
buried.
--
Russ Allbery (rra@debian.org) <http://www.eyrie.org/~eagle/>
Reply to: