Size of an Integer

General
1

Today, an Integer is the same as Int32. When Xojo moves to 64-bit, an Integer will be Int64, yes? So how can I tell the size of an Integer, for example, if I want to create a library that return the correct result on either 32- or 64-bit apps?

I ask because I’m introducing a function that will return the maximum value of an Integer. Right now, that’s &h7FFFFFFF, but eventually it will be &h7FFFFFFFFFFFFFFF.

2

You could always use #if RBVersion <= xxxx and just update that with every release.

Alternatively, coerce Integer -1 to UInt64 -1 and see what you end up with.

3

I believe Xojo does not have a sizeof(), but it’s documented that on 32 systems Integer is 32 bit signed and in the future 64 bit systems it will be 64 bit signed, as a workaround you could set a constant based on #if Target64bit

4

Yes, I could certainly update it manually in the future. I would just rather not have to remember, hence the test.

I guess calculating the max Int64 then seeing if an Integer can hold it is my best bet.

5

“The Integer type is an alias to the platform’s native signed integer type.
On 32-bit builds, this ends up being SInt32 and on 64-bit builds this ends up being SInt64.”.

http://documentation.xojo.com/index.php/Integer_datatypes

So I believe you can safely decide at compile time. :wink:

6

[quote=33212:@Kem Tekinay]I guess calculating the max Int64 then seeing if an Integer can hold it is my best bet.
[/quote]

[code]dim x as Int64
dim y as integer

x = &h7FFFFFFFFFFFFFFF
y = x
if x = y then it’s 64 bits.[/code]

7

[quote=33215:@Brad Hutchings][code]dim x as Int64
dim y as integer

x = &h7FFFFFFFFFFFFFFF
y = x
if x = y then it’s 64 bits.[/code][/quote]

Or just use Target32Bit/Target64Bit…

8

Holy smoke, I didn’t know Target64Bit was there. When was that introduced?

9

Thanks to all.

This is a case where there are really two answers. I’d have used Target64Bit, but that was introduced after the earliest supported version of Real Studio (this is for MacOSLib), so I ended up with a version of Brad’s code:

Protected Function NSIntegerMax() As Integer
  static r as Integer
  static needsCalc as boolean = true
  if needsCalc then
    dim n64 as Int64 = &h7FFFFFFFFFFFFFFF
    r = n64
    if n64 <> r then
      r = &h7FFFFFFF
    end if
    
    needsCalc = false
  end if
  
  return r
  
  
  // This is a method and not a constant because it will be a different value under 32-bit and 64-bit.
  // Would use Target64Bit here but that is not supported in 2011r1. Otherwise, the code would look like this:
  
  '#if Target64Bit
  'return &h7FFFFFFFFFFFFFFF
  '#else
  'return &h7FFFFFFFF
  '#endif
  
End Function
10

You could use something like:

#if RBVersion >= 2011.02
  #if Target64Bit
    return &h7FFFFFFFFFFFFFFF
  #else
    return &h7FFFFFFFF
  #endif
#else
  return &h7FFFFFFFF
#endif

I’m not sure on the version information, but I’d expect this to work.

11

Well that’s true. I was so fixated on a mathematical solution, I missed the obvious one. Thanks.

12

2011 r2 or so

13

Yes, I changed my code to something very close to what Joe recommended. I just used RBVersion >= 2013.01 to be on the safe side.