Steve Langasek wrote: > - Is bumping the dpkg-dev dependency ok, or should we instead > opportunistically detect the DEB_HOST_MULTIARCH field and fall back to > /usr/lib otherwise?
I've chosen to do both. This way, backporters can simply make a version
of debhelper that drops the dependency, and it will fall back to
non-multiarch behavior.
> - Do you think adding a new {misc:Pre-Depends} substvar is sensible here?
> Since no one will have this substvar anywhere yet, all affected packages
> will need to update debian/control, so arguably they could be updated to
> declare Pre-Depends: multiarch-support directly; but this would be less
> backportable, and it may be easier for tools to warn about an unused
> substvar than a missing pre-depends?
It may be easier if those pre-depends later need to be removed (or
changed). I've kept it for now.
> - What accompanying documentation would you like for this change?
Taken care of.
> - How soon could we have this in unstable? :D
After dpkg-dev gets in, and someone checks what's in git and makes sure
it works, ideally.
--
see shy jo
signature.asc
Description: Digital signature
