Re: lintian support in dinstall
* "Santiago" == Santiago Vila <sanvila@unex.es> wrote:
Santiago> A lintian error or a lintian warning is a just a bug, and a
Santiago> bug reported by lintian is no more important than any other
Santiago> ordinary bug.
This is not completely true. Bugs reported in the BTS are all kind of
things. Lintian reports only policy violation. This is of a different
quality.
And to me, it is reasonable to not let packages pass that violate
policy (and we only speak of Lintain errors, not warnings). [ It was
said that packages should pass, that decrese the number of bugs, so
that one can remove them step by step. ]
Maintainer should always use lintian to check their packages, and
debuild calls it automatically. Maintainers should not upload packages
that violate policy by themself.
Even if a lintian error is just an ordinary bug, we are preventing the
inclusion of a new bug. This is an QA activity.
Just saying "well, it's a bug, it will be noted in the BTS, so what"
is the current behaviour, but it should be changed to a more active
one of bug prevention. The BTS should be a list of real bugs, not ones
we introduce by ignoring the red lights of lintian.
But I don't opt for the mandatory check, until we can assure that
packages that have a lintian error, but this error is not justified,
will pass nevertheless in a _timely_ matter, or we will end up like
with the Incoming mess.
Maybe the maintainer himself could override this, and his
justification would made public (so noone will override just because,
as this will make him a fool in the public view)
Ciao,
Martin
Reply to: