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: fixincludes for glibc 'inline' non-C99 conformance

Geoffrey Keating wrote:

I don't see how backing out the patch will permit anyone else to make progress. Users cannot update their code to be C99-compliant if they have no compiler that will accept the resulting program.

That's a false dichotomy: you could always provide an additional option to use the C99 semantics, e.g., -fc99-extern-inline, or -std=notgnuc99.

I'm big on standards compliance, and I'm not at all big on extensions, but I don't think your change is appropriate without more warning, discussion, etc. It's just too abrupt.

In the recent discussion about whether to allow conversions between vectors with different base types, people (including Apple people) went back and checked what older compilers did, because they were concerned about breaking existing code -- even though any such code would be non-comforming relative to the Altivec PIM, etc.

The bottom line is that right now there is controversy. You feel your patch is appropriate; some other experienced maintainers don't. Why not back it out (or conditionalize it) and work to build consensus?

Mark Mitchell
(650) 331-3385 x713

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