On 2011-10-28 08:09 +0200, Raphael Hertzog wrote:
> On Thu, 27 Oct 2011, Sven Joachim wrote:
>> Whether it's multi-arch enabled depends on the debhelper and dpkg-dev
>> versions on the build system. With debhelper from squeeze-backports and
>> dpkg-dev from squeeze, ncurses builds fine (without multiarch support, of
>> course).
>
> Just because dpkg-architecture -qDEB_HOST_MULTIARCH fails... and thus
> $(DEB_HOST_MULTIARCH) happens to be empty.
>
> Is this really something that was made on purpose?
Not on my side, but dh(1) also does its multiarch stuff in v9 mode only
if dpkg-architecture -qDEB_HOST_MULTIARCH does not fail.
> If you want to support both a traditional build (for backport purpose) and
> a multiarch build, I think you'd better be explicit about it in
> debian/rules (use a dedicated variable that you can easily switch) instead
> of relying on dpkg-architecture -qDEB_HOST_MULTIARCH to fail and thus
> return an empty string.
That may be cleaner, yes.
> Adding the multiarch-support dependency on dpkg-dev means that your
> package would build but would get its multiarch-support pre-dependency
> too, making it uninstallable in squeeze without upgrading the libc6.
That's true. Actually the FTBFS is not my biggest concern, packages
with broken dependencies are a worse problem and are just as likely to
result.
Cheers,
Sven
--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
