I had problems getting a DLL function to
return real values when called from excel.
Another function with integer parameters and integer
return value worked just fine. Any good references
on writing DLLs for Excel?

Thanks

Anders
pe...@shh.fi

: Anders Pelli <pe...@shh.fi> wrote:

You must be using 16bit. As far as I can tell this behaviour does not
occur in 32bit.

I would be VERY grateful if you could share your experience in writing
DLLs for Excel. I am about to do the same.

References:
a) The Excel SDK
b) Check Lloyd's help file (available from Borland -> Delphi download
section, near the bottom) why returning a floating point value via a
function to EVERY 16bit Microsoft product (Excel, VC etc.) is a
problem - and learn how it can be fixed. To give you a taste:

{ cut from code, stripped down, not guaranteed to work }
function E(x: double;NearPtr: word): Pointer;
var
  TempPointer: ^double; { make this {*word*99}py behaviour explicit }
begin
  try
      Result := Ptr(SSeg, NearPtr);
      TempPointer := Ptr(SSeg, NearPtr);
      TempPointer^ := 1.225346
  except
    Error;
  end;
end;

c) Quick workaround: do not return a float, but a use a var parameter
to return it:

   procedure(x,y,z: double; var something: double)
--
Stefan.Hoffmeis...@UNI-Passau.DE
http://kakadu.rz.uni-passau.de/~w4hoff01/
University of Passau, Bavaria, Germany