make check failures

[Andrew Walrond]

Richard Henderson wrote:
> On Sat, Mar 08, 2003 at 11:59:19AM -0800, Ulrich Drepper wrote:
> 
>>>Searching the archives reveals -march=pentium4 might be to blame? Should
>>>I stick to i686 and -O2?
>>
>>Add -mno-sse2.
> 
> 
> Err... Unless you also added -mfpmath, -march shouldn't affect
> how fp code is generated.
> 
> 
> r~
> 

Well, adding -mno-sse2 worked for me, and I don't have -mfpmath, so I 
think you must be mistaken.

Andrew Walrond