Re: pmake patch
On Wed, Sep 06, 2000 at 12:53:11AM +0200, Marcus Brinkmann wrote:
> On Tue, Sep 05, 2000 at 07:28:13AM -0400, Igor Khavkine wrote:
[snip]
> > > Note the second #if, which takes effect when you compile on an older linux
> > > version with libc5 or libc4.
> > >
> > > By changing from linux to __GLIBC__, you actually prevent the getmode and
> > > setmode declarations on those platforms. Now you can either say, let's drop
> > > support for those, and then you can remove the second, inner, #if block, or
> > > remain compatible and change to #if defined(__GLIBC__) || defined(linux)
> > > (or something similar).
> > >
> >
> > I think we should keep support for old libraries since packages for
> > all platforms are compiled from the same source tree. However I still
> > object to using defined(linux) since the features in the #if block are
> > not linux specific but specific to some version of glibc. A better idea
> > would be to investigate if all the defined(__GLIBC__) that I added
> > and see if that alone would suffice or glibc version information has to
> > be added.
>
> Are we now talking about the one change above, or in general? In general, I
> mean in all cases except the above, the change from linux -> __GLIBC__ is
> the right thing to do. Even in the second paragraph about ("By changing...")
> I was only talking about the one change in miscbltin.
>
> Does anybody know which purpose MACHINE and MACHINEARCH has? The
> debian/rules file probably needs to be changed slightly to be correct on the
> Hurd.
I was talking about it in general. The case you singled out is just a
special case, actually feel free to modify that portion of the patch
as you see fit. The general discussion comes into play if this
patch is submitted upstream, whether it will be accepted there is yet
another question. But if it is, it would probably be a good idea to
let the upstream maintainers to investigate if simply the presence of
__GLIBC__ is enough or some version checking should be done as well.
I say if the patch currently works and no-one is objecting to any of the
changes it should be at least submitted to the debian ash maintainer.
Igor
--
To UNSUBSCRIBE, email to debian-hurd-request@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org
Reply to: