Improve clog, clog10 handling of values with real or imaginary part slightly above 1 (bug 13629)

[Carlos O'Donell]

On 8/4/2012 12:08 PM, Joseph S. Myers wrote:
> On Sat, 4 Aug 2012, Allan McRae wrote:
> 
>> I have not formally checked, but I assume this patch is the reason I
>> started getting these errors on i686:
> 
> As log1pl uses x87 logarithm instructions, it's quite plausible that the 
> exact results may vary between x86 implementations and so require more 
> than one person to update ulps.
> 

These ULPs should get updated *immediately*, Allan could you please checkin the update?

I don't want to see the testsuite regress at any point in time.

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