Re: invalid and duplicate architectures
On 2014-05-24 04:16, Russ Allbery wrote:
> Paul Wise <pabs@debian.org> writes:
>> On Fri, May 23, 2014 at 10:00 PM, Johannes Schauer wrote:
>
>>> In my code I counted all debian architectures as "valid" which are
>>> listed on packages.debian.net. Is there a better way to retrieve
>>> "valid" architectures in this context?
>
>> Dpkg::Arch::get_valid_arches, which looks at /usr/share/dpkg/*table
>
> I'm pretty sure that Lintian already uses that data to check whether
> architectures are valid (I vaguely remember working on that code), but it
> may well be that the existing check doesn't get the wildcards right.
>
Hi,
Honestly, I am pretty sure the validator gets wildcards right. I
remember re-implementing that part of the Lintian code and ran into the
exact problem[1].
I am more inclined to believe that the problem is that there is /no/
validation of the "Architecture" field in d/control of the source
package. At least a quick scan of checks/control-file.pm suggests this
to be the case.
~Niels
[1]
http://anonscm.debian.org/gitweb/?p=lintian/lintian.git;a=blob;f=lib/Lintian/Architecture.pm;h=f03701c812dea4f9ac9a803c8697313211eb642b;hb=e3015d2327847d729183a196e0b8db9743134ea5#l103
"""
# NB: "$os-$cpu" ne $archstr in some cases
"""
Which happens in cases like "armhf", which is a part of "any-arm /
linux-any".
Reply to: