This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: add #define to revert C99 inlining behaviour

Ian Lance Taylor <> writes:

> > 2006-11-03  Geoffrey Keating  <>
> > 
> > 	* c-decl.c (WANT_C99_INLINE_SEMANTICS): New, set to 1.
> > 	(merge_decls): Implement WANT_C99_INLINE_SEMANTICS.
> > 	(grokdeclarator): Likewise.
> This is completely pointless.  Please do not check this in.  Thanks.

Great, you already checked it in.

Geoff, please do not use the fact that you have global write
privileges to ignore the need to build consensus among maintainers.

You observed earlier that you presented this patch in May and that the
time to object was back then.  In fact, people did object.  Here is
the thread:
You did nothing to actually address the objections, except to
encourage other people to work on them, and to observe that Darwin
does not have a problem.

Then you went ahead and checked the patch in a couple of days ago, and
now we can not build glibc with gcc.

I'm not asking you to fix any of this.  I proposed a plan, based on
our offline discussion, for how we can implement C99 semantics in gcc
in what I consider to be a sane manner which will work for our users:

If you will back out your patches, I will volunteer to make sure that
this plan gets implemented, as modified by other people's suggestions.
I won't do it today or next week.  But I will do it before the 4.3

Please do not get so attached to your issues that you ignore other
people's concerns.  I understand and respect your frustration with
forward progress in gcc.  Making progress in areas with complex
interdependencies is hard work.  It is much harder than hacking code

Thanks for listening.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]