On 22/08/15 16:48, Andrey Rahmatullin wrote:
> On Sat, Aug 22, 2015 at 04:35:15PM +0200, Adam Borowski wrote:
>> -- are arch:all packages not building on other architectures worth filing
>> FTBFS bugs for?
>
> I think this should be reported (it's probably common sense that you
> should be able to rebuild a package on a given supported arch?)
I would be inclined to say that an arch-all package failing to build
from source is "important" if it's on some reasonably fast/capable
architecture like arm64, but not really worth reporting (and if
reported, "minor" at most) if it's on a machine that no sensible person
would choose for their non-architecture-specific builds.
One example is that I recently proposed an imagemagick patch to do the
image-resizing for its icons in build-indep, not in build-arch, so that
in practice weaker architectures wouldn't do it; it takes my laptop a
matter of seconds, and takes a mips buildd with no FPU about 15 hours
(which gets the build killed for inactivity).
S
