Browse Source
This routine was entemplatificated from the integer version, which was designed to be perfect across representable ranges of points. The floating point version suffered from a rounding problem.master
1 changed files with 20 additions and 4 deletions
Loading…
Reference in new issue