Re: Set of patches
Raphael Geissert <firstname.lastname@example.org> writes:
> Russ Allbery wrote:
>> I'd be good with making this a real global, but it's hard to do that
>> in frontend/lintian. I've been thinking for a while that the right
>> solution is to move all the changes file checking into some other
> I couldn't agree more. I've actually added it to my TODO list, but
> hadn't yet though very much how to do it. One possibility was to add a
> collection script that would create a 'changes' file that could later
> be checked if not empty.
There isn't a reasonable place to put it in the lab, though. Consider
the case of a *.changes file for a binary-only build where there's no
source package and multiple binary packages.
It's really a third type of "thing," separate from either source or
>> Ideally, I'd like to see changes be a fourth type of "package" that can
> Time for me to say "ew" :)
Why? I mean, it's a little odd in sound, but it fits into the Lintian
architecture really well, I think. All of the changes data could can be
passed in via the $info argument as a Lintian::Collect::Changes object,
so there's no need to do strange things with the lab.
Russ Allbery (email@example.com) <http://www.eyrie.org/~eagle/>