RE: gsl-1.6 make check failure on Suse 9.0

["J.Underwood" <J dot Underwood at open dot ac dot uk>]

Oh dear, that'll teach me to not build on a new system at gone midnight. The errors from make check I reported below appear only with -O3 optimization, and are absent with no optimization or -O2. As using -O3 is a risky business at the best of times, this should probably not be regarded as a problem.

Sorry for the noise.

Jonathan.

-----Original Message-----
From:	gsl-discuss-owner@sources.redhat.com on behalf of J.Underwood
Sent:	Mon 07/03/2005 1:25 AM
To:	gsl-discuss
Cc:	
Subject:	gsl-1.6 make check failure on Suse 9.0
Hi

When building the latest release (1.6) on a Suse 9.0 machine, I get a number of failures for make check - they seem to be tolerance issues.  This was with no optimization at all. gcc version 3.3.1 (SuSE Linux). Let me know if I can provide anything else of use.

grep'ing the output of make check for FAIL: gives

FAIL: symm(10), direct eigenvalue(1), unsorted (1.75621604502731543 observed vs 1.75621604502731588 expected)
FAIL: symm(10), direct eigenvalue(2), unsorted (-1.55367382353584205 observed vs -1.55367382353584249 expected)
FAIL: symm(10), direct eigenvalue(6), unsorted (0.281591917636177524 observed vs 0.281591917636177635 expected)
FAIL: symm(10), direct eigenvalue(7), unsorted (-0.07036176334030321 observed vs -0.0703617633403031684 expected)
FAIL: herm(10), direct eigenvalue(3), unsorted (2.1422064843154156 observed vs 2.14220648431541472 expected)
FAIL: herm(10), direct eigenvalue(4), unsorted (-0.955285146892125514 observed vs -0.955285146892125292 expected)
FAIL: herm(10), direct eigenvalue(6), unsorted (1.41142172899002949 observed vs 1.41142172899002993 expected)
FAIL: herm(10), direct eigenvalue(7), unsorted (-0.0545196117924557286 observed vs -0.0545196117924557147 expected)
FAIL: test

Jonathan.