This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
Re: Success report on Linux/PPC, small Ada problem
- From: mike stump <mrs at windriver dot com>
- To: bosch at gnat dot com
- Cc: dewar at gnat dot com, fw at deneb dot enyo dot de, gcc at gcc dot gnu dot org, jsm28 at cam dot ac dot uk, minyard at acm dot org, rth at redhat dot com
- Date: Wed, 6 Feb 2002 13:47:23 -0800 (PST)
- Subject: Re: Success report on Linux/PPC, small Ada problem
> Date: Wed, 6 Feb 2002 15:23:40 -0500
> To: mike stump <mrs@windriver.com>
> From: Geert Bosch <bosch@gnat.com>
> On Wednesday, February 6, 2002, at 03:02 , mike stump wrote:
> > Now, if what you would want to merge, isn't `better' than what it
> > replaces, then certainly I would agree to hold off merging. I cannot
> > tell from your message which is the case.
> If we can get an agreement to update the FSF Ada tree to include all
> changes that have been made by ACT since the last merge, in a single
> shot, that would be great.
> We know that current GCC sources and current FSF sources plus all
> changes results in a compiler that is in very good shape for
> x86/Linux and just slightly behind on a number of other
> platforms. Unraveling all development work into separate
> patches/changelogs and test these individually is an enormous job
> though, which I don't know how to do efficiently.
My take is this is for the maintainer of the code to call. If you are
that person, I think we should let _you_ do what you think is best.
Sure, some people would have preferred if they all went in one by one,
but the more we delay, the worse the problem gets. The right solution
is to merge as early and as often as you can tolerate. The only
limitation I would place on you is, you have to think that the result
of doing that is `better', than what we had before. If you aren't the
maintainer, then you'll need to do whatever they want, well, cept
washing his car.
As the Ada tree stabilizes, we'd expect that more and more, it will
follow all the usual conventions, for example, each patch in solation
to {gcc,ada}-patches. I think we should be somewhat flexible for new
ports, new frontends, new optimization passes.
Does anyone have objections to the Ada folks bulk merging bits they
own to sync the tree up? (Even I would object to bits you don't own.)