Mirror of forum.nim-lang.org

9239 :: implicit conversions from float literals to float32?

• flint (orginal) [2022-06-16T23:03:39+02:00] view original

  Is it possible to make float literals be considered as float32 ? Or to have implicit conversion from float literals to float32 ? If there's any chance to get rid of all the "f" that clutters the numbers without wasting cpu+mem, well then... I'm asking.

---

• treeform (orginal) [2022-06-16T23:16:47+02:00] view original

  I also wanted some thing like, a pragma at the top of the file that says:

  {.onlyFloat32allowed.}
  Run

  Or

  type
    float = float32
    float64 = {.error.}
  Run

  So that no float64 sneaks in. In some operation like a * 1 / 3 or me using float by accident instead of float32 and code just keeps working but slower.

  Things we wish for...

---

• DavideGalilei (orginal) [2022-06-16T23:38:00+02:00] view original

  I don't think it would be possible, since float is an alias for float64: https://github.com/nim-lang/Nim/blob/devel/lib/system.nim#L30

---

• ElegantBeef (orginal) [2022-06-17T00:18:17+02:00] view original

  Welcome to the world of term rewriting macros

  template noFloats{f}(f: float) = {.error: "No 'float64's allowed!".}
  
  var a = 10 / 20
  Run

---