This is the mail archive of the gcc-patches@gcc.gnu.org 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: [rfc] atomic memory operation builtins


On Thu, 14 Apr 2005, Gabriel Dos Reis wrote:

> I did not understand the first part of your message, maybe because I
> was not supposed to understand it.  As of WG14, it has been repeated

X = http://gcc.gnu.org/ml/gcc-patches/2005-04/msg00627.html
Z = any C/C++ front end merger (vaporware of uncertain merit)

You appeared <http://gcc.gnu.org/ml/gcc-patches/2005-04/msg00636.html> to 
suggest that X was a move away from Z and was bad on those grounds.  I say 
X is a cleanup of use now and was good on those grounds: we shouldn't have 
code pretending to be used by more languages than it is and so in the 
wrong location and quite likely before Z might be relevant all that code 
will have been rewritten anyway.

> at this meeting that WG14 has no intention to revise C any further.
> (they are willing to produce TRs, but TRs are not normative).
> Concerning WG21, I'm committing most of my resources in trying to
> improve things so I do not really understand that part neither.

The terminology, concepts and logic used in the parts of GCC which compile 
C should be aligned with those in the C standard in order to correspond 
reliably with the language defined by that standard.

The terminology, concepts and logic used in the parts of GCC which compile 
C++ should be aligned with those in the C++ standard in order to 
correspond reliably with the language defined by that standard.

The terminology, concepts and logic used in the standards have been 
diverging for twenty years.  If you are working in WG21 to move towards 
those used for C and towards a C++ standard which defers to the C standard 
for the definition of all meaningfully common areas (as espoused in the 
2001 discussions in e.g. c++std-compat messages 481 and 748), great.  
Otherwise, with effort unilaterally in WG21 and not directed to 
convergence, merged implementations will become *less and less* desirable 
over time as the standards continue to diverge.

-- 
Joseph S. Myers               http://www.srcf.ucam.org/~jsm28/gcc/
    jsm@polyomino.org.uk (personal mail)
    joseph@codesourcery.com (CodeSourcery mail)
    jsm28@gcc.gnu.org (Bugzilla assignments and CCs)


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