Message313145
> What if return 80-bit infinities and NaN as 64-bit infinities and NaN and raise an error in the case of finite numbers overflow?
That's certainly reasonable from a pure floating-point-conversion perspective: it's exactly what I'd expect a general purpose extended-precision to double-precision converter to do. |
|
Date |
User |
Action |
Args |
2018-03-02 17:28:42 | mark.dickinson | set | recipients:
+ mark.dickinson, tim.peters, serhiy.storchaka |
2018-03-02 17:28:41 | mark.dickinson | set | messageid: <1520011721.98.0.467229070634.issue32978@psf.upfronthosting.co.za> |
2018-03-02 17:28:41 | mark.dickinson | link | issue32978 messages |
2018-03-02 17:28:41 | mark.dickinson | create | |
|