Subject: | (round_down_pow(3) 3^64 + 1.25 exp) t/bigfloat.t test fails with Math-BigInt-1.999720 |
After upgrading Math-BigInt from 1.999717 to 1.999720, t/bigfloat.t test fails like this:
$ prove -b t/bigfloat.t
t/bigfloat.t .. # Math::BigFloat version 1.999722
# Math::BigInt version 1.999722
t/bigfloat.t .. 1/59 # +inf is inf
# -inf is -inf
# nan is NaN
t/bigfloat.t .. 54/59 # Test 56 got: "64.0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000" (t/bigfloat.t at line 316)
# Expected: "64" (round_down_pow(3) 3^64 + 1.25 exp)
# t/bigfloat.t line 316 is: ok ($exp, 64, "round_down_pow(3) 3^64 + 1.25 exp");
# Test 59 got: "64.0000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000000" (t/bigfloat.t at line 328)
# Expected: "64" (round_down_pow(3) 3^64 + 1.25 exp)
# t/bigfloat.t line 328 is: ok ($exp, 64, "round_down_pow(3) 3^64 + 1.25 exp");
t/bigfloat.t .. Failed 2/59 subtests
I tries to reproduce it with Math::BigFloat only by running the test against local copy of round_down_pow(), but it passed. So maybe there is some weird interaction between Math-PlanePath modules and the Math::BigFloat.