Re: [PATCH, RFC] Enable IBM long double for PPC32 Linux

[Mark Mitchell <mark at codesourcery dot com>]

Giovanni Bajo wrote:

> This is a little unfair, though. So now the burden on enforcing the policy is
> not on the maintainers that prepare the patches? 

No, that burden falls on the Release Manager.  However, the SC has also
given me considerable latitude to exercise my judgement, which I did.  I
made the decision that, if safety could be prserved, this patch was
worthwhile.  My point was that people who did not like the direction in
which I was moving could have gone over my head to the SC right away.

> it misses the point that many important resources in GCC are being used in
> fixing and testing this new feature, instead of putting GCC in shape for the
> release. So the release has been already delayed because of this, and will be
> even more. That's something which already happened.

I agree.

I am not happy about being put in this position, and I've made that
clear.  I have also indicated that this represents a communication
problem.  As expressed before, part of that is GLIBC/GCC communication
-- but, in addition, communication on the part of the PowerPC GNU/Linux
developers, who participate in both GLIBC and GCC development.

Your objections, and those of others, have been noted, and I will bear
your concern in mind in future.  However, my decision for 4.1, right or
wrong, has been made; the compiler portions of these changes, if I deem
them safe, and if off by default, will be part of GCC 4.1.  At this
point, we can live with that decision, criticize it, or ask the SC to
intervene.

-- 
Mark Mitchell
CodeSourcery
mark@codesourcery.com
(650) 331-3385 x713