This is the mail archive of the libc-alpha@sourceware.org mailing list for the glibc project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Support testing "inexact" exceptions in libm-test.inc


On 05/16/2013 12:35 AM, Joseph S. Myers wrote:
This patch adds support to libm-test.inc for testing whether "inexact"
exceptions are raised.  Unlike other exceptions, there are only
generally requirements on these for functions bound to particular IEEE
754 operations, so the default is that such exceptions are permitted
but not required, and the case where they are forbidden is represented
explicitly with NO_INEXACT_EXCEPTION.

This patch adds expectations for such exceptions only for rint and
nearbyint (where the only difference between the two functions is
whether such exceptions are raised).  The idea is that I'll add such
expectations for other functions bound to IEEE 754 operations in
followups to this patch (and for functions taking and returning a
quiet NaN, since C11 Annex F says "raise no floating-point exception"
for that case).  For the functions where C11 Annex F leaves
unspecified whether "inexact" is raised but DTS 18661-1 says it isn't
(round, ceil, floor, trunc) I propose to make the tests follow the
requirements of DTS 18661-1.

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


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]