Subject: | to_ieee754-mbf.t failing in 1.999817 |

I recently updated blead to 1.999817 but somehow failed to run tests before pushing, and I've now noticed that one is failing. Verbose output attached.
We have a blead release on Sun 20th. It would be great if this could be fixed by then. If not then I will just revert the update.

Subject: | output.txt |

D:\Dev\Git\perl\t>.\perl harness -v ../cpan/Math-BigInt/t/to_ieee754-mbf.t
../cpan/Math-BigInt/t/to_ieee754-mbf.t ..
1..60
#
# Computing test data for k = 16 ...
#
#
# smallest positive subnormal number (k = 16)
#
ok 1
#
# largest subnormal number (k = 16)
#
ok 2
#
# smallest positive normal number (k = 16)
#
ok 3
#
# largest normal number (k = 16)
#
ok 4
#
# largest number less than one (k = 16)
#
ok 5
#
# smallest number larger than one (k = 16)
#
ok 6
#
# second smallest number larger than one (k = 16)
#
ok 7
#
# one (k = 16)
#
ok 8
#
# minus one (k = 16)
#
ok 9
#
# two (k = 16)
#
ok 10
#
# minus two (k = 16)
#
ok 11
#
# positive zero (k = 16)
#
ok 12
#
# positive infinity (k = 16)
#
ok 13
#
# negative infinity (k = 16)
#
ok 14
#
# NaN (encoding used by Perl on Cygwin) (k = 16)
#
ok 15
#
# Computing test data for k = 32 ...
#
#
# smallest positive subnormal number (k = 32)
#
ok 16
#
# largest subnormal number (k = 32)
#
ok 17
#
# smallest positive normal number (k = 32)
#
ok 18
#
# largest normal number (k = 32)
#
ok 19
#
# largest number less than one (k = 32)
#
ok 20
#
# smallest number larger than one (k = 32)
#
ok 21
#
# second smallest number larger than one (k = 32)
#
ok 22
#
# one (k = 32)
#
ok 23
#
# minus one (k = 32)
#
ok 24
#
# two (k = 32)
#
ok 25
#
# minus two (k = 32)
#
ok 26
#
# positive zero (k = 32)
#
ok 27
#
# positive infinity (k = 32)
#
ok 28
#
# negative infinity (k = 32)
#
ok 29
#
# NaN (encoding used by Perl on Cygwin) (k = 32)
#
ok 30
#
# Computing test data for k = 64 ...
#
#
# smallest positive subnormal number (k = 64)
#
ok 31
#
# largest subnormal number (k = 64)
#
ok 32
#
# smallest positive normal number (k = 64)
#
ok 33
#
# largest normal number (k = 64)
#
ok 34
#
# largest number less than one (k = 64)
#
ok 35
#
# smallest number larger than one (k = 64)
#
ok 36
#
# second smallest number larger than one (k = 64)
#
ok 37
#
# one (k = 64)
#
ok 38
#
# minus one (k = 64)
#
ok 39
#
# two (k = 64)
#
ok 40
#
# minus two (k = 64)
#
ok 41
#
# positive zero (k = 64)
#
ok 42
#
# positive infinity (k = 64)
#
ok 43
#
# negative infinity (k = 64)
#
ok 44
#
# NaN (encoding used by Perl on Cygwin) (k = 64)
#
ok 45
#
# Computing test data for k = 128 ...
#
#
# smallest positive subnormal number (k = 128)
#
Can't take log of 0 at ../../lib/Math/BigFloat.pm line 4474.
# Looks like your test exited with 34 just after 45.
Dubious, test returned 34 (wstat 8704, 0x2200)
Failed 15/60 subtests
Test Summary Report
-------------------
../cpan/Math-BigInt/t/to_ieee754-mbf.t (Wstat: 8704 Tests: 45 Failed: 0)
Non-zero exit status: 34
Parse errors: Bad plan. You planned 60 tests but ran 45.
Files=1, Tests=45, 1 wallclock secs ( 0.00 usr + 0.00 sys = 0.00 CPU)
Result: FAIL
D:\Dev\Git\perl\t>