Mirror of forum.nim-lang.org

10470 :: why float casting is not working ?

• maverk (orginal) [2023-09-08T20:54:05+02:00] view original

  
  var casting : float = cast[float](1)
  
  echo casting
  Run

  result :

  
  4.940656458412465e-324
  Run

  why is this happening ?

  how does the cast keyword work on which types ?

---

• juancarlospaco (orginal) [2023-09-08T20:56:26+02:00] view original

  Use 1'f or 1.float.

---

• maverk (orginal) [2023-09-08T20:58:50+02:00] view original

  can you please explain to me why the suffix is needed ? because i already gave the cast keyword the desired type i want to make of 1

---

• dlesnoff (orginal) [2023-09-08T20:58:57+02:00] view original

  cast never does what you want. It should reinterpret bit by bit your number but it is unsafe.

---

• maverk (orginal) [2023-09-08T21:03:32+02:00] view original

  so it just translate the binary of the given value to the desired type ?

---

• namisboss (orginal) [2023-09-08T21:05:53+02:00] view original

  I've seen talk of people suggesting renaming cast to bitcast, because that's what it does. You want a conversion which is done via float(1) or any of the variants juancarlospaco suggested (as they all are pretty much the same thing). In general you should never be using cast.

---

• Chronos (orginal) [2023-09-08T21:08:57+02:00] view original

  cast is essentially telling the compiler that X is of type Y, but because floats are represented differently in the compiler (due to how binary works), casting an integer to a float gives outputs like that, when you do 1.float or 1'f, it tells the compiler that you want to represent the integer value 1, as a float

---

• maverk (orginal) [2023-09-08T21:09:03+02:00] view original

  yes i was confused since there are string and char and other built in functions for doing the same thing well i will avoid cast thanks brother much appreciated

---

• cblake (orginal) [2023-09-08T21:12:50+02:00] view original

  Also, you don't need the apostrophe ' for float and uint (and maybe others):

  echo type(1f)," ",type(1u)
  Run

  And if you put the word "Nim" after the triple open quotes it will syntax-highlight your code on The Forum (if you did not know).

---

• maverk (orginal) [2023-09-08T21:23:49+02:00] view original

  what is the difference between type() and typeof() ???

---

• cblake (orginal) [2023-09-08T21:40:39+02:00] view original

  I think typeof is better in an expression context like above -- sorry -- and type use there is "deprecated" (but works obviously if you compile & run it). I have no idea what the timeline for really deprecating this use is or how much of the nimbleverse will break if it actually goes away.

---

• cblake (orginal) [2023-09-08T22:11:00+02:00] view original

  A quick grep suggests about 210 packages in ~560 files use expression type(). So, I don't know.. that's like 10% of Nimble packages. Of course, there are definitely many packages not in the nimble index, and also dark code that is not even public. I am aware of at least order(100) subcommand "dark use cases" of cligen, for example.

  I think the idea for deprecation is that the same word two different ways (another way is this which uses both on the same line of code) is "confusing". Personally, I always liked the brevity of type() and there are certainly other cases of the same word being used two different ways (like of in case vs. in a type expression like type Animal = ref object of RootObj), but it's not much of a hill worth dying on, as the old saying goes. ;-)

---

• PMunch (orginal) [2023-09-08T22:30:16+02:00] view original

  Just for completeness sake you could also just type out a floating point number:

  var x = 1.0
  Run

---