weird optimization in sin+cos, x86 backend

[Andrew Haley]

On 02/16/2012 04:21 PM, Joseph S. Myers wrote:
> On Thu, 16 Feb 2012, Andrew Haley wrote:
>
>> On 02/16/2012 02:42 PM, Vincent Lefevre wrote:
>>
>>> C99 doesn't require directed rounding modes, but as long as they
>>> are claimed to be supported by<fenv.h>, they should work:
>>
>> Ah, I see.  So, we could bring gcc+glibc into compliance by not
>> defining the rounding mode macros.
>
> I don't think that's a good approach; I think we should fix the GCC bugs
> with -frounding-math etc. (and implement the standard pragmas).

For avoidance of doubt: I don't think that *anyone* thinks that's a good
approach.

Andrew.