| Subject: | t/reports.t fails (with -Duselongdouble?) |
On some of my smokers the test suite consistently fails:
...
# Failed test 'Correct value for 0.001'
# at t/reports.t line 430.
# +--------+----+--------+
# | GOT | OP | CHECK |
# +--------+----+--------+
# | 0 ms | eq | 1 ms |
# +--------+----+--------+
# Looks like you failed 1 test of 32.
t/reports.t ........
Dubious, test returned 1 (wstat 256, 0x100)
Failed 1/32 subtests
...
Statistical analysis suggests that this happens if perl was compiled with increased floating point accuracy (-Duselongdouble):
****************************************************************
Regression 'conf:uselongdouble'
****************************************************************
Name Theta StdErr T-stat
[0='const'] 0.0000 0.0000 0.00
[1='eq_undef'] 1.0000 0.0000 60645502236045240.00
R^2= 1.000, N= 36, K= 2
****************************************************************