Santhosh Raju <fox%netbsd.org@localhost> writes:
> On Sun, Mar 21, 2021 at 10:21 PM Greg Troxel <gdt%lexort.com@localhost> wrote:
>>
>>
>> Santhosh Raju <fox%NetBSD.org@localhost> writes:
>>
>> > Module Name: pkgsrc-wip
>> > Committed By: Santhosh Raju <fox%NetBSD.org@localhost>
>> > Pushed By: fox
>> > Date: Sun Mar 21 22:15:03 2021 +0530
>> > Changeset: e9daf7b1eb4e9570eeaa02ff2688c4331f54a125
>> >
>> > Added Files:
>> > pypot/DESCR
>> > pypot/Makefile
>> > pypot/PLIST
>> > pypot/distinfo
>> > Removed Files:
>> > py-pypot/DESCR
>> > py-pypot/Makefile
>> > py-pypot/PLIST
>> > py-pypot/distinfo
>> >
>> > Log Message:
>> > py-pypot: Rename to pypot, remove the redundant py prefix.
>>
>> This is not really ok, as now the directory doesn't follow our norms.
>
> I probably misread your message on the removal of the redundant py, I
> thought you were talking about the "py-" part and not the "py" in the
> "pypot" part.
Sorry for being unclear. Yes, it is a very firm notion that python
programs have directory names py-foo for some value of foo. The
PKGNAME is then
${PYPKGPREFIX}-foo
which ends up being py27-foo or py38-foo which allows building both at
once, assuming the PLISTs don't collide. If you have something that
doesn't start py- to be changed to pyNN-, then that doesn't work.
>> It might be good if you reverted back to py-OT and py-pot and left it
>> for Kamal.
>
> But naming the package wrong is alright with the norms?
No, not really, but it's a smaller error. The right thing to do is to
file bugs with both upstreams that they have colliding package names and
wait for them to fix it! Before importing either...
But seriously, in general, there might be two packages with the same
name, and we have to rename one of them to have things fit. Imagine
there are two packages py-pot which are different. So one has to be
renamed, and there's no way around this. Here, we have py-POT and
py-pot as the upstream values, and we consider those to be the same
name.
> I shall revert both the changes.
I had thought py-pypot was irregular but as I read the guide it doeesn't
say that, and we have a lot of those already. So if you go back to
py-pypot that seems ok.
Attachment:
signature.asc
Description: PGP signature