Re: [patch] Remove sysdeps/ieee754/ldbl-128/bits/huge_vall.h

[Carlos O'Donell <carlos_odonell at mentor dot com>]

On 9/27/2012 2:10 PM, Roland McGrath wrote:
>> The only difference between the two is some support that the
>> ieee754 one has for old GCC's that we don't support anymore so I would
>> like to remove sysdeps/ieee754/ldbl-128/bits/huge_vall.h and leave just
>> bits/huge_vall.h
> 
> This statement includes an ambiguity.  There are two kinds of "GCC's we
> support" (or don't any more).  There is a pretty late minimum GCC version
> supported for building libc.  There is a far wider range of GCC versions
> supported for building applications against libc and its headers.  As this
> is a public header, it's the latter set that matters here.
> 
>> Joseph Myers gave a fuller description of why this file isn't needed in
>> glibc-ports (http://sourceware.org/ml/libc-ports/2012-09/msg00088.html)
>> and also mentions that the arch specific versions (x86, m68k, ia64, and sparc)
>> are probably also unneeded but this patch just removes the one in
>> sysdeps/ieee754/ldbl-128/bits/huge_vall.h.
> 
> Joseph's explanation in that libc-ports message clarifies that there is no
> (GNU) compiler that might be used to build applications against any libc
> configuration using ldbl-128/ that would perceive the difference between
> the ldbl-128/ and generic versions of bits/huge_vall.h.  I'm not sure we've
> made any claims about non-GNU compilers, but we don't really care.
> 
> So this change is OK by me.

I've looked through the code and the change is also OK by me.

I'm always kinda paranoid so I'd be happier if you also did a build test
that showed that x86 or x86_64 didn't break. I don't see how they would
break but I also didn't look deeply at the implies ordering for header
inclusion.

Cheers,
Carlos.
-- 
Carlos O'Donell
Mentor Graphics / CodeSourcery
carlos_odonell@mentor.com
carlos@codesourcery.com
+1 (613) 963 1026