Re: [gentoo-dev] [RFC] Initial python-r1.eclass & distutils-r1.eclass

[Michał Górny]

On Sat, 29 Sep 2012 15:49:32 +0200
hasufell <hasufell@g.o> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> On 09/29/2012 12:49 PM, Michał Górny wrote:
> > On Sat, 29 Sep 2012 11:20:31 +0100 Markos Chandras
> > <hwoarang@g.o> wrote:
> >
> >> -----BEGIN PGP SIGNED MESSAGE----- Hash: SHA512
> >>
> >> On 09/29/2012 09:53 AM, Micha? Górny wrote:
> >>> Hello,
> >>>
> >>> Instead of the floating patches and p-d-ng modifications I sent
> >>>  earlier, here are the two complete (so far, well, initial :P)
> >>> eclasses for review.
> >>>
> >>> They are designed as 'mostly' drop-in python-distutils-ng
> >>> replacement.
> >>>
> >> Hi,
> >>
> >> Are you saying that you are going to remove the
> >> python-distutils-ng eclass in favour of the new eclasses? I don't
> >> quite understand the reasons to be honest.
> >
> > The reason is simple -- I can't fix it without changing the API.
> > Changing the API on a live eclass is confusing, and considering
> > that it is not used by many packages, it's easier to lastrite it.
> >
> > Also, this fixes the name not to have any '-ng' nor '-ds9'.
> >
>
> What are the reasons to change the API in the first place? There has
> to be a good reason, cause this will involve yet another migration of
> many ebuilds. I don't see any bugreports.

I have pointed out what changes to the API are _necessary_ to introduce
a good eclass on gentoo-python@.

Otherwise, the eclass would have to have at least two almost identical
functions doing the same thing, one universal and one for specific case
where specific parameters are passed (and not used in a single ebuild).

> I fear this will cause more confusion, i.e. some ebuilds using the old
> distutils, some using python-distutils-ng and some using distutils-r1
> resulting in weird tree behavior.

[example needed]

-- 
Best regards,
Michał Górny