Delphi Generic Template classes Usage Compilation Error - delphi

Question

I am new to Delphi Generic Classes. I don't get it of how to use the generic classes in the implementation code.
Here is the code:
Type TDataElement = class(TObject)
protected
Procedure SetName(sNewValue:String); virtual;
private
m_sName:String;
published
property sName:String read m_sName write SetName;
end;
Type TDataArray<T : TDataElement> = class(TObject)
public
function Find(dtElement:T):integer;
Procedure Add(dtElement:T);
private
m_vContainer : array of T;
protected
Function GetData(Index:integer):T; virtual;
Procedure SetData(Index:integer; NewValue:T); virtual;
public
property vData[Index: Integer]: T read GetData write SetData;
end;
implementation
function TDataArray<T>.Find(dtElement:T):integer;
var i:integer;
begin
Result:=-1;
for i := 0 to high(m_vContainer) do
if (m_vContainer[i] <> NIL)and(m_vContainer[i] = dtElement) then
begin
Result:=i;
exit;
end;
end;
.....
When I try to create instances of the Generic Classes like in the following code:
Method1)
var z:TDataArray<TDataElement>;
z:=TDataArray<TDataElement>.Create();
I get the following error:
E2010 Incompatible types: 'TDataElement' and 'class of TDataElement'
If I do this 2nd method I get another strange error:
Method 2)
type TDataElementClass = class of TDataElement;
var z:TDataArray<TDataElementClass>;
F2084 Internal Error : I8230
What I am doing wrong?
Entire source code in one file
System.SysUtils,Classes,
dtArray_unit in 'D:\VisionBot\Software\VisionBot\GUI\Units\dtArray_unit.pas';
Type TDataElement = class(TObject)
protected
Procedure SetName(sNewValue:String); virtual;
private
m_sName:String;
published
property sName:String read m_sName write SetName;
end;
Type TDataArray<T : TDataElement> = class(TObject)
public
function Find(dtElement:T):integer; overload;
Procedure Add(dtElement:T);
private
m_vContainer : array of T;
protected
Function GetData(Index:integer):T; virtual;
Procedure SetData(Index:integer; NewValue:T); virtual;
public
property vData[Index: Integer]: T read GetData write SetData;
end;
type
TDerivedDataElement = class(TDataElement)
end;
var
z2: TDataArray<TDerivedDataElement>;
//------------------------------------------------------------------------------
Procedure TDataElement.SetName(sNewValue:String);
begin
self.m_sName:=sNewValue;
end;
//------------------------------------------------------------------------------
//------------------------------------------------------------------------------
//------------------------------------------------------------------------------
function TDataArray<T>.Find(dtElement:T):integer;
var i:integer;
begin
Result:=-1;
for i := 0 to high(m_vContainer) do
if (m_vContainer[i] <> NIL)and(m_vContainer[i] = dtElement) then
begin
Result:=i;
exit;
end;
end;
//------------------------------------------------------------------------------
Function TDataArray<T>.GetData(Index:integer):T;
begin
Result:=NIL;
if Index < 0 then exit else
if Index > high(Index) then exit else
Result:=self.m_vContainer[Index];
end;
//------------------------------------------------------------------------------
Procedure TDataArray<T>.Add(dtElement:T);
begin
SetLength(self.m_vContainer,Length(m_vContainer)+1);
m_vContainer[High(m_vContainer)]:=T;
end;
//------------------------------------------------------------------------------
Procedure TDataArray<T>.SetData(Index:integer; NewValue:T);
begin
if Index < 0 then exit else
if Index > high(Index) then exit else
self.m_vContainer[Index]:=T;
end;
//------------------------------------------------------------------------------
begin
try
z2:= TDataArray<TDerivedDataElement>.Create();
readln;
except
on E: Exception do
Writeln(E.ClassName, ': ', E.Message);
end;
end.

Answer 1

var
z: TDataArray<TDataElementClass>;
The problem is that TDataElementClass is not a class derived from TDataElement.
The following would be valid:
var
z: TDataArray<TDataElement>;
Or this:
type
TDerivedDataElement = class(TDataElement)
end;
var
z: TDataArray<TDerivedDataElement>;
In your code you have
type
TDataElementClass = class of TDataElement;
Now, TDataElementClass is a metaclass.
A variable of type TDataElement can hold an instance of the type TDataElement, or an instance of any class derived from TDataElement.
A variable of type TDataElementClass can hold a type, which must be TDataElement, or any class derived from TDataElement.
You claim in the question that using TDataArray<TDataElement> leads to a compiler error, but that is not true. Consider this compiling program:
type
TDataElement = class
end;
type
TDataArray<T: TDataElement> = class
public
function Find(dtElement: T): Integer;
private
m_vContainer: array of T;
end;
function TDataArray<T>.Find(dtElement: T): Integer;
begin
for Result := 0 to high(m_vContainer) do
if (m_vContainer[Result] <> nil) and (m_vContainer[Result] = dtElement) then
exit;
Result := -1;
end;
var
arr: TDataArray<TDataElement>;
begin
arr := TDataArray<TDataElement>.Create;
end.
In your edit you show this code:
Procedure TDataArray<T>.Add(dtElement:T);
begin
SetLength(self.m_vContainer,Length(m_vContainer)+1);
m_vContainer[High(m_vContainer)]:=T;
end;
The erroneous line is here:
m_vContainer[High(m_vContainer)]:=T;
This fails because T is a type rather than an instance. I think you mean:
m_vContainer[High(m_vContainer)]:=dtElement;