This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] PowerPC - check for finite and Infinity parameters inIBM Long Double 128 fmal().
On 9/7/2011 1:16 PM, Jakub Jelinek wrote:
> On Wed, Sep 07, 2011 at 10:49:27AM -0500, Ryan S. Arnold wrote:
>> On Tue, Aug 23, 2011 at 6:40 AM, Dave Flaherty
>> <flaherty@linux.vnet.ibm.com> wrote:
>>> Check for finite/infinity parms in IBM Long Double 128 fmal( )
>>>
>>> This patch addresses some IBM Long Double 128 fmal () test-ldouble.out
>>> and test-ildoubl.out failures. If the âxâ and âyâ parameters are
>>> finite values and âzâ is infinity, the result of fmal () should be âzâ
>>> not NaN.
>>>
>>> This patch is checked into ibm/2.13/master and can be
>>> cherry-picked from commit 9cb11a4f1e562490f02ef634946ec015bf6e75e4
>>>
>>>
>>> 2011-08-18 David Flaherty <flaherty@linux.vnet.ibm.com>
>>>
>>> * sysdeps/ieee754/ldbl-128ibm/s_fmal.c: New file which checks
>>> for finite and infinity parameters.
>>
>> Ack!
>>
>> Please consider committing this patch. I've been running with it in my
>> ibm/2.13 branch for a few weeks with no problems.
>
> Well, all the patch does is work around a testsuite failure, the double
> double fmal still doesn't do what fmal is supposed to do (though, the
> question is what exactly it should do with the weirdo floating point format,
> as the precision is variable).
> That said, I guess the patch is a (tiny) improvement on what glibc does
> right now.
>
> Jakub
Jakub,
I'm looking into the fmal() not calculating as a ternary operation issue.
Dave