On Sat, Jan 13, 2018 at 09:17:25AM +0100, Mattia Rizzolo wrote:
> On Sat, Jan 13, 2018 at 01:11:11PM +0530, Chris Lamb wrote:
> > As an aside, I was wondering whether a Lintian check for this would be sane,
> > like an inverse of "arch-independent-package-contains-binary-or-object"?
>
> I think not.
> There are several packages in the archive that totally ought to be
> arch:all, but are otherwise arch:any due to how architectures
> costrainsts propagates in a Multi-Arch situation (go ask helmut for more
> details). It's a workaround, but it's the best way to deal with that
> particular problem, I'd rather not have lintian warn about it.
There are also plenty packages that are arch-dependent even though they
don't contain any binary or object files.
linux-libc-dev would be an obvious example,
and there are plenty of less obvious ones.
> Having a package arch:any instead of arch:all doesn't really inpact
> anything except some extra buildd time and mirror space (and looks odd
> to those who notice it).
Agreed.
cu
Adrian
--
"Is there not promise of rain?" Ling Tan asked suddenly out
of the darkness. There had been need of rain for many days.
"Only a promise," Lao Er said.
Pearl S. Buck - Dragon Seed
