Open jducoeur opened 5 years ago
I think what happened is I took the algo for next double from ScalaCheck, without understanding it. I wanted to back later and understand it. So I will do that. Here's the one from ScalaCheck:
And here's ours:
They look pretty similar. I'll study up on IEEE 754 and document it.
See the TODO on this line, which suggests that we aren't quite sure about the behavior at these extreme edge conditions, and somebody should investigate that.