Comment # 4 on bug 1030250 from Adam Majer

This failure is GCC related for this version of gsl.

gcc6-6.3.1+r245113-1.5 from Tumbleweed yields success.

gcc7-7.1.1+r247574-21.2 from Staging test error. The tests fail not by a tiny
margin either,


> [  209s] ======================================
> [  209s]    gsl 2.3: specfunc/test-suite.log
> [  209s] ======================================
> [  209s] 
> [  209s] # TOTAL: 1
> [  209s] # PASS:  0
> [  209s] # SKIP:  0
> [  209s] # XFAIL: 0
> [  209s] # FAIL:  1
> [  209s] # XPASS: 0
> [  209s] # ERROR: 0
> [  209s] 
> [  209s] .. contents:: :depth: 2
> [  209s] 
> [  209s] FAIL: test
> [  209s] ==========
> [  209s] 
> [  209s] FAIL: schmidt csphase=1 deriv2 i=102 (-1.42436215874661529e-16 observed vs -1.42460074537287091e-16 expected) [9186711]
> [  209s] FAIL: schmidt csphase=1 deriv2 i=776 (1.85098628876862506e-16 observed vs 1.85127865747637551e-16 expected) [9187385]
> [  209s] FAIL: schmidt csphase=-1 deriv2 i=102 (-1.42436215874661529e-16 observed vs -1.42460074537287091e-16 expected) [15218755]
> [  209s] FAIL: schmidt csphase=-1 deriv2 i=776 (1.85098628876862506e-16 observed vs 1.85127865747637551e-16 expected) [15219429]
> [  209s] FAIL test (exit status: 1)
> [  209s] 
> [  209s] ./specfunc/test.log
> [  209s] FAIL: schmidt csphase=1 deriv2 i=102 (-1.42436215874661529e-16 observed vs -1.42460074537287091e-16 expected) [9186711]
> [  209s] FAIL: schmidt csphase=1 deriv2 i=776 (1.85098628876862506e-16 observed vs 1.85127865747637551e-16 expected) [9187385]
> [  209s] FAIL: schmidt csphase=-1 deriv2 i=102 (-1.42436215874661529e-16 observed vs -1.42460074537287091e-16 expected) [15218755]
> [  209s] FAIL: schmidt csphase=-1 deriv2 i=776 (1.85098628876862506e-16 observed vs 1.85127865747637551e-16 expected) [15219429]
> [  209s] FAIL test (exit status: 1)

The expected deviance is <1e-10 for relative error and this is closer to 1e-4

Since GSL didn't change its code, there must be some change in GCC floating
point optimization?

Also, disabling optimization causes tests to fail in some other case :/

Michael, Richard, do you have any clues as to what changed in GCC that could
result in this? There are no problems with any tests on x86_64. Disabling
optimization causes failures in other tests due to error accumulation.