This is the mail archive of the
libc-alpha@sourceware.org
mailing list for the glibc project.
Re: Improve tgamma accuracy (bugs 2546, 2560, 5159, 15426)
- From: Andreas Jaeger <aj at suse dot com>
- To: libc-alpha at sourceware dot org
- Date: Wed, 08 May 2013 08:22:50 +0200
- Subject: Re: Improve tgamma accuracy (bugs 2546, 2560, 5159, 15426)
- References: <Pine dot LNX dot 4 dot 64 dot 1305022233490 dot 12072 at digraph dot polyomino dot org dot uk>
On 05/03/2013 12:35 AM, Joseph S. Myers wrote:
This patch fixes various bugs relating to inaccuracy of results from
tgamma. Bugs 2546, 2560 and 5159 are different cases of inaccurate
finite results, arising from the use of exp (lgamma) which inevitably
results in large errors for results with large (positive or negative)
exponent. Bug 15426, found by the new testcases prepared for this
patch, is wrong sign of infinity on overflow for negative arguments
close to 0. (Bugs 15425 and 15427 were also found by testing unfixed
glibc with the testcases included in this patch; this patch does
nothing about those bugs in lgammaf and exp, although they no longer
affect the results of tgamma after the tgamma implementation changes
in this patch.)
This is fine,
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