Bug#655975: [PATCH] Support :any architecture qualifiers for multiarch

[Niels Thykier <niels@thykier.net>]

On 2013-09-17 22:34, Adam D. Barratt wrote:
> [Dropped -python CC; replaced -release with an existing bug relating to
> britney and multi-arch]
> 
> On Mon, 2013-09-16 at 15:02 +0100, Colin Watson wrote:
>>   i18nspector/i386 unsatisfiable Depends: python3:any (>= 3.2.3-3~)
>>
>> I've fixed this for Ubuntu's britney instance with the following patch.
> 
> Thanks very much!
> 

Thanks indeed,

>> I'm afraid I haven't had a chance to rebase it on top of Debian or to
>> test it there, but that should hopefully not be very much work, and
>> certainly easier than recreating all this from scratch ...
> 
> Indeed. I've done the rebasing (which was mostly handling the fact that
> Debian has a "consts" module) and attached the result.
> 
> The resulting britney successfully passes our test-suite, including the
> four "live-data" datasets. Admittedly none of those tests involve
> multi-arch dependencies, but it does at least demonstrate a lack of
> regressions.
> 
> Regards,
> 
> Adam
> 

I think we should take this opportunity to get rid of the "empty"
PREDEPENDS slot (i.e. have the MULTIARCH field replace PREDEPENDS).
There is no point in having the empty slot for PREDEPENDS; yours truly
was just too lazy to figure out how to remove it from the C code when I
merged PREDEPENDS into DEPENDS.

~Niels