Question

Below is a simple example of using pointer in delphi.

Type

TRecord1 = Record
field1 : String;

end; 

procedure TForm1.Button2Click(Sender: TObject);
var
   Rec : TRecord1;
   Ptr: ^TRecord1;

begin
   Rec.field1:= 'field1';
   Ptr := @Rec;
   memo1.Lines.Add (Ptr^.field1); 
   memo1.Lines.Add (Ptr.field1); // it also works.

 end;

In such case, Ptr^ and Ptr both work. It seems delphi is to allow user more flexibility in pointing to the value. But just by reading the two lines, they are syntactically different and may mean differently. In such case both work. But my question is:

1. how can a user know in other situations where ^ can or can not be
   omitted or, where with ^ or without ^ means the same or differently?
2. What are those situations ? Examples will be appreciated.
3. Why? (Optional)

Thanks a lot in advance.

Answer 1

how can a user know in other situations where ^ can or can not be omitted or, where with ^ or without ^ means the same or differently?

What are those situations ? Examples will be appreciated.

A plain Pointer doesn’t have any fields or properties, so ignoring Delphi’s smarts, the Pointer.Field syntax doesn’t make sense. Because of that there can’t be a conflict between Pointer^.Field and Pointer.Field, simply because the plain . syntax is meaningless if you don’t dereference the pointer.

If the type pointed to by the pointer doesn’t have any fields you have to use the ^ syntax. That is, when the pointer is a pointer to a basic type, or is an untyped pointer.

Why? (Optional)

Class instance references (what most people would call "objects") are also pointers in Delphi, I assume the syntax was introduced to make working with Pointers less verbose and more like using classes. It’s also harmless, because as mentioned above, the compiler can’t get it wrong.

I personally prefer the ^. syntax, because it makes it clear I’m working with a pointer and not a record or class.