[Solved] Integer Dynamic Array

[jcmontherock]

Is "TIntegerArray" a dynamic array ?

[CM630]

Type TIntegerArray anywhere in your code, click it and press Alt+Up_Arrow.
Then you will get:

Code: Pascal  [Select][+][-]

1. {$ifdef CPU16}
2.        IntegerArray  = array[0..(32768 div SizeOf(Integer))-2] of Integer;
3. {$else CPU16}
4.        IntegerArray  = array[0..$effffff] of Integer;
5. {$endif CPU16}
6.        TIntegerArray = IntegerArray;

[cdbc]

Hi
NO

Code: Pascal  [Select][+][-]

1. IntegerArray  = array[0..$effffff] of Integer;

Regards Benny

eta: To allocate it dynamically, you're meant to do so via

Code: Pascal  [Select][+][-]

1. PIntegerArray = ^IntegerArray;

The reason it's such a big range, is to /not/ run into "Index out of bounds" errors. (which you could if you defined it like: intarry = array[0..0] of integer;)

[jcmontherock]

So, I works for me if I define it like this:

Code: Pascal  [Select][+][-]

1. Type
2. TintArray = Array of Integer;
3. ...
4. private
5. IntArray: TIntArray;
6. // IntArray: TIntegerArray; // this definition is not a dynamic array.
7.  

[cdbc]

Hi
If you want a dynamic array of integer (managed), then YES that would be the way to go...
Regards Benny

[Thaddy]

What you want is defined in the types unit: TIntegerDynArray.

[H₂SO₄]

I consider this definition to be legacy cruft and usually just overshadow it with a proper dynamic array definition.
TStringArray  for example is exactly what you'd expect, so why not make things consistent with  TIntegerArray?:

Code: Pascal  [Select][+][-]

1. program Example;
2.  
3. {$mode objfpc} // <-- Implicitly `uses` the objpas unit
4.  
5. uses
6.   SomeOtherUnit;
7.  
8. // Here `TIntegerArray` = array[0..$EFFFFFF] of Integer
9.  
10. type
11.   TIntegerArray = array of Integer;
12.  
13. // Here `TIntegerArray` = array of Integer
14.  
15. // Original definition can be accessed using `objpas.TIntegerArray`
16.  
17. begin
18. end.

With the amount of legacy identifiers that FPC needs to support for compatibility's sake, it's a blessing that identifiers can be shadowed/redefined so easily. 

[Thaddy]

That would in this case be a breaking change, hence the types unit that is often overlooked.