Subject: | 'Input minimal float for %time' fails with -Duselongdouble |
If perl was compiled with increased floting point accuracy (i.e. using -Duselongdouble while configuring perl), then the test suite fails:
...
# Failed test 'Input minimal float for %time'
# at t/verysmall.t line 37.
# got: '00:00.000'
# expected: '00:00.001'
# Looks like you failed 1 test of 6.
t/verysmall.t ....
Dubious, test returned 1 (wstat 256, 0x100)
Failed 1/6 subtests
...
The problem is also visible on CPAN Testers Matrix: http://matrix.cpantesters.org/?dist=Time-Format;reports=1#sl=0,0
There are only FAILs for all configurations with "-ld" in the archname.