Floats in For/ Next?

[Amiga5k]

Trying out PB4b1 and I'm surprised to see that I still can't do this:

Code: Select all

For a.f = 0 To 40 Step .5
  PrintN(StrF(a))
Next

Yes, there are other ways to do the above, but it would be nice if For Next supported all "core" number types.

Russell

[remi_meier]

For now you can do something like that:

Code: Select all

Macro FFor(var, ffrom, fto, fstep = 1.0)
	var = ffrom - fstep
	While var <= fto - fstep
		var + fstep
EndMacro


x.f = 7.0
FFor(x, 11.5, 13.0, 0.5)
	Debug x
	
Wend

[chris319]

Trying out PB4b1 and I'm surprised to see that I still can't do this:

Code: Select all

For a.f = 0 To 40 Step .5
  PrintN(StrF(a))
Next

Yes, there are other ways to do the above, but it would be nice if For Next supported all "core" number types.

Russell

I would elevate this to "necessity" from "nicety".

[blueznl]

personally, i think coding for / next loops with anything but ints is wrong, i'd always suggest to use

Code: Select all

n.f = 0
while n.f < 100
  ...
  n.f = n.f+1.7
wend

but i agree, it's a matter of taste...

[Amiga5k]

...Except that every other BASIC dialect currently available has supported floats in For/Next loops for a long time now. (Can't remember if Commodore Basic did or not...)

But yeah, a Float can be faked using Macros, functions or what have you. But those miss the point: It probably should have already been in there a while ago. Probably a simple oversight, which is understandable. This is really the only core language functionality I have a problem with, especially now that we finally have the new types for character (.c), double (.d) and quad (.q). And fixed length strings (yay!). And default values for function parameters. And... (Yes, overall PB is looking VERY good now!)

But still no double precision floats Maybe next time?

Russell

[blueznl]

.d ?

[PB]

> every other BASIC dialect currently available has supported floats in
> For/Next loops for a long time now. (Can't remember if Commodore
> Basic did or not...)

It sure did:





Although, I note with interest that it doesn't show "3" as the final print...?

[freedimension]

Well, "3.something" is larger than 3

[walker]

2.8 + 0.2 = 3.0 :roll:

[blueznl]

walker, you're right, but in floats is 2.8 + 0.2 not the same as 3.0 as in float 2.8 might be 2.80000000000000001 and 0.2 might be actually 0.20000000000000001 so the total might be 3.00000000000000002 and that's more than 3...

that's why i never use for next for floats

[chris319]

personally, i think coding for / next loops with anything but ints is wrong

...Except that every other BASIC dialect currently available has supported floats in For/Next loops for a long time now.

As do C, Perl, and a bunch of languages I never program in. People coming to PB from other languages or dialects of BASIC may see this as a shortcoming and that could hinder the advancement of PB. As for 3.00000000000000002 being > 3, I say leave that to the programmer to deal with, not the language developer.

[Psychophanta]

walker, you're right, but in floats is 2.8 + 0.2 not the same as 3.0 as in float 2.8 might be 2.80000000000000001 and 0.2 might be actually 0.20000000000000001 so the total might be 3.00000000000000002 and that's more than 3...

that's why i never use for next for floats

Each basic do things different.
MSX version of the same thing:

[PB]

If the C64 isn't going to show "3" then it shouldn't show "1" or "2" either.

[Shannara]

personally, i think coding for / next loops with anything but ints is wrong, i'd always suggest to use

Code: Select all

n.f = 0
while n.f < 100
  ...
  n.f = n.f+1.7
wend

but i agree, it's a matter of taste...

Why is that? Especially when using Longs in For..Next rather then Int, results in faster code.

[netmaestro]

Look at the Commodore example. Mostly, it hit it on and the loop worked. But in one case, that of 3.0, it missed. It didn't print the 3.0 because it added 0.2 to 2.8 and got 3.000001 or something and decided it was >3.0 rather than =3.0 and considered the loop finished.