Closed vti closed 7 years ago
thanks! it looks good to me, although I don't fully understand the buggy behavior. I pinged @ingydotnet
I think I understand it now. I guess the SvTYPE comparison didn't work correctly before for tied data, that's why the tied.t test is correct now.
thanks again, merged and released as 0.65_001! I also added a test for this.
Released 0.66
It looks like SvPV_nomg adds the string representation for the number and creates side effects.
This breaks tied tests, but it looks like the test was invalid, since generating bare values should be fine there.