Re: GCC vs GLIBC: why this stance, Drepper ?!?

["Zack Weinberg" <zackw at Stanford dot EDU>]

Please honor the Followup-To: this time.  I do not want to be cc:ed.

On Sat, Jun 30, 2001 at 03:59:51PM -0700, H . J . Lu wrote:
> On Sat, Jun 30, 2001 at 11:41:06PM +0200, Paolo Carlini wrote:
> > 
> > No. It's time for you, the developers, to discuss technical solutions to the
> > problem.
> > 
> 
> Well, like I said, the gcc developers have to ACKNOWLEDGE there is a
> problem first.

I will repeat, once, what I said the last time you brought this up.
The following is a verbatim quote from
http://gcc.gnu.org/ml/gcc/2001-02/msg00972.html.

# On Wed, Feb 21, 2001 at 11:13:09AM -0800, H . J . Lu wrote:
# > That is one reason I call it a mess. The current scheme is a huge
# > problem for Linux and the gcc people don't even want to admit it.
# > I will be more than happy to be proved wrong.
# 
# This is the exact same thing as when you post obscure Makefile patches
# with no explanation.  We have no way of evaluating your assertion that
# "the current scheme is a huge problem for Linux."  All we know is that
# you say it's a problem.  Lots of the GCC developers don't concern
# themselves with EH runtime issues at all.  Those of us who know about
# the EH runtime, are not also system integrators.  So we don't have any
# context.
# 
# You need to describe in excruciating detail exactly what will break,
# and why.  It is not at all obvious to us why it is a huge problem.
# Convince us.

You never provided an adequate response to that question.  Until you
do, I will not know why you think "the current scheme is a huge
problem for Linux."  To the best of my knowledge, the current problems
with gcc 3.0 and glibc 2.2.x are minor, fixable bugs.  I believe that
most of the gcc developers are of the same opinion.

You know perfectly well how to answer this sort of question in
adequate detail; I've seen you do it on the binutils list.  I can only
assume that you don't care enough to do it for us.  Why, then, should
we care about your problem reports?

> BTW, I, and other glibc developers, can provide patches to gcc and
> glibc. But they have to be acceptable to both gcc and glibc developers.
> If the gcc developers won't even admit there is a serious problem, do
> you think those gcc patches will be considered?

If you post a patch for GCC to gcc-patches, I hereby promise that I
will not ignore it.  (Be aware that I may have only sporadic net
access for the next month or so, however.)

However, in exchange, I expect you to provide a complete explanation
and justification of the patch.  If you don't, I will not recommend
the patch for inclusion, even if it is correct.

zw