On Sun, Aug 30, 2009 at 4:41 PM, Stefan Sperling<stsp_at_elego.de> wrote:
> On Sun, Aug 30, 2009 at 10:20:11PM +0200, Arfrever Frehtes Taifersar Arahesis wrote:
>> 2009-08-30 22:11:15 Stefan Sperling napisa??(a):
>> > On Sat, Aug 29, 2009 at 01:00:53AM +0200, Arfrever Frehtes Taifersar Arahesis wrote:
>> > > > @@ -759,9 +496,7 @@ def patch_unidiff_offset(sbox):
>> > > >
>> > > > # list all tests here, starting with None:
>> > > > test_list = [ None,
>> > > > - patch_basic,
>> > > > patch_unidiff,
>> > > > - patch_copy_and_move,
>> > > > patch_unidiff_absolute_paths,
>> > > > patch_unidiff_offset,
>> > > > ]
>> > >
>> > > Removing references to these tests from test_list would be sufficient.
>> > > Removing definitions of these tests would cause needless problems for
>> > > maintainers of branches on which these tests have been modified.
>> > I understand. But if we'll switch this to the git patch format,
>> > the tests in their current state won't be needed at all, right?
>> At least a part (which might include code modified on other branches) of these
>> tests won't need to be rewritten.
> Right. Then we'll just disable the tests by removing (or commenting)
> them from the tests list, and put comments above them saying that they
> will need to be rewritten to use git diff extensions instead of the
> svnpatch format before they can be enabled again. OK?
It seems unlikely we will need these tests unless we decide to
maintain the same patch format as was originally taken. If this is
done, the tests can be revived on Arfrever's Python 3 branch just as
easily as it will be for the person that revives them on trunk.
I say we should remove the tests if we remove the code that goes with them.
Received on 2009-08-30 23:28:57 CEST