| Subject: | Tests fail (with Config::Model 2.080 and later?) |
Currently there's only one fail report at CPAN Testers ( http://www.cpantesters.org/cpan/report/657dfb86-e307-11e5-8917-374a6e9158c7 ), but more will be available in a few hours.
Statistical analysis of test reports generated on my smokers suggest that the problem happens only with Config::Model 2.080 and later (negative theta is bad):
****************************************************************
Regression 'mod:Config::Model'
****************************************************************
Name Theta StdErr T-stat
[0='const'] 1.0000 0.0000 20638086192200216.00
[1='eq_2.056'] 0.0000 0.0000 1.59
[2='eq_2.060'] 0.0000 0.0000 0.23
[3='eq_2.061'] -0.0000 0.0000 -0.44
[4='eq_2.064'] 0.0000 0.0000 0.83
[5='eq_2.067'] 0.0000 0.0000 0.17
[6='eq_2.068'] 0.0000 0.0000 1.49
[7='eq_2.073'] 0.0000 0.0000 0.33
[8='eq_2.074'] -0.0000 0.0000 -0.44
[9='eq_2.078'] -0.0000 0.0000 -0.33
[10='eq_2.080'] -1.0000 0.0000 -16850926812823184.00
[11='eq_2.081'] -1.0000 0.0000 -14593330697217224.00
R^2= 1.000, N= 40, K= 12
****************************************************************