Fix spurious "inexact" exceptions from dbl-64 sqrt (bug 15631)

[Andreas Jaeger]

On 06/14/2013 11:12 PM, Joseph S. Myers wrote:
> Bug 15631 is spurious "inexact" exceptions from the dbl-64 sqrt
> implementation, which show up as testsuite failures for architectures
> not using hardware square root instructions.  This patch fixes these
> by saving and restoring exception state, which then requires an
> explicit check for whether the computed result is the exact square
> root so "inexact" can be raised when required.  Tested MIPS64 (n32
> ABI).  (MIPS is in fact an architecture that*should*  be using
> hardware square root instructions, but as a result of sysdeps
> directory ordering it isn't at present; I've filed bug 15632 for
> that.)

ok, thanks,

Andreas
-- 
  Andreas Jaeger aj@{suse.com,opensuse.org} Twitter/Identica: jaegerandi
   SUSE LINUX Products GmbH, Maxfeldstr. 5, 90409 Nürnberg, Germany
    GF: Jeff Hawn,Jennifer Guild,Felix Imendörffer,HRB16746 (AG Nürnberg)
     GPG fingerprint = 93A3 365E CE47 B889 DF7F  FED1 389A 563C C272 A126