Re: Reviving Debian QA
Adam Di Carlo wrote:
> Why not just extend lintian, or parse lintian report output, to be
> able to submit bugs. We'd need to do a bit a work: (a) getting
> approval from this list to submit bugs based on this or that lintian
> tags (b) the actual work
This would have to be well organized, to prevent lots of duplicate
I don't see much value in bug reports generated from Lintian output
unless they have been verified by a human. In other words, Lintian
is used as a tool to discover bugs, which are then reported on their
own merits. It should be possible to phrase the bug report without
referring to Lintian.
For some kinds of bugs, I think the Lintian pages are much better suited
for tracking them than the BTS is. This is because Lintian automatically
tracks the current state of the archive, thus cutting out a lot of
administrative overhead. For those bugs, sending a simple reminder
to the maintainer may be more useful than submitting a bugreport.
> John> Yes lintian should be used, but it is currently not effective
> John> in getting the bugs fixed. I suppose if an error were reported
> John> for each package which does not use doc-base, it would be
> John> helpful.
> Well, technically, doc-base isn't required by Policy. So that work
> would have to be done first.
Indeed. The procedure for registering documentation seems quite
confused right now. I don't remember a consensus the last time
this came up.
> There are some serious design flaws in doc-base (and dwww) which I
> hope to address soon.
This may be one of the reasons :-)