This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: [PATCH] Fix cbrtl for ldbl-96
- From: "Joseph S. Myers" <joseph at codesourcery dot com>
- To: Andreas Schwab <schwab at suse dot de>
- Cc: <libc-alpha at sourceware dot org>
- Date: Wed, 17 Jul 2013 13:32:46 +0000
- Subject: Re: [PATCH] Fix cbrtl for ldbl-96
- References: <mvmfvvde46g dot fsf at hawking dot suse dot de> <Pine dot LNX dot 4 dot 64 dot 1307171316450 dot 23307 at digraph dot polyomino dot org dot uk> <mvmbo61e2ug dot fsf at hawking dot suse dot de>
On Wed, 17 Jul 2013, Andreas Schwab wrote:
> "Joseph S. Myers" <joseph@codesourcery.com> writes:
>
> > This patch is missing the testcase additions for libm-test.inc, which
> > should cover both overflow and underflow cases.
>
> I'm leaving that to you. I don't know how to compute accurate numbers.
For this bug, testing cbrtl (0x1p16383L) == 0x1p5461L and similarly for
0x1p-16383L would suffice - examples where the correct result is an exact
hex float value.
(Since cbrtl might still produce a slightly inexact result, of course the
patch shouldn't go in until after the 2.18 release, to avoid architecture
maintainers needing to update libm-test-ulps again for 2.18.)
--
Joseph S. Myers
joseph@codesourcery.com