On January 7, 2018 5:25:21 PM UTC, Mattia Rizzolo <mat...@debian.org> wrote:
>On Sun, Jan 07, 2018 at 11:40:41AM -0500, Scott Kitterman wrote:
>> As discussed, stepic is another case of this that I just ran into.
>The test
>> was a useful reminder that it hadn't been ported to python3 in 0.3-5,
>but once
>> I did the port in 0.3-6 it was still there due to the way the package
>renaming
>> test was done.
>>
>> Maybe if you look for python:Depends, but no python3:Depends that
>would reduce
>> the false positives without being excessively magical.
>
>Personally, to catch this (rather special, I dare say) case, I'd ignore
>transitional packages (I don't know if lintian already has a definition
>of those, but I'd define them as "priority extra/optional, section
>oldlibs, any('dummy transitional', 'transitional package') in the long
>description").
That would work for this new case, but not the one I originally filed the bug
for. My suggestion solves both.
Scott K
