In regards to Int64ToCurrency:
Result := currency((@AVal)^);

In regards to CurrencyToInt64:
Result := ISC_INT64((@AVal)^);

[DCC Error] IB_Utils.pas(1459): E2089 Invalid typecast
[DCC Error] IB_Utils.pas(1595): E2089 Invalid typecast
[DCC Error] SheaLibCompDes.pas(11): F2063 Could not compile used unit
'IB_Utils.pas'

--- In IBObjects@yahoogroups.com, "Martijn Tonies" <m.tonies@...> wrote:
>
> Please be SPECIFIC with your info! Have you ever had co-workers
> coming up to you saying "it doesn't work" and then expecting you to
> answer that?
>
> Now, where exactly does this code fail to compile.
>
> > function Int64ToCurrency( AVal: ISC_INT64; AScale: smallint ):
currency;
> > {$IFNDEF IBO_VCL35}
> > var
> > i: smallint;
> > AFactor: ISC_INT64;
> > {$ENDIF}
> > begin
> > {$IFDEF IBO_VCL35}
> > raise Exception.Create( 'Unsupported' );
> > {$ELSE}
> > // Do not apply rounding - used for reading data as stored in
database.
> > // Originally I was starting by casting AVal to Result and then
scaling
> > // the currency Result variable. This resulted in rounding
> > problems. It
> > // seems that some Delphi maths functions involving currency use
> > extended
> > // for interim values with subsequent problems at the limits of the
> > range.
> > // So now I manipulate the int64 value first, and then assign
the final
> > // value directly to Result. This direct cast produces an implicit
> > scale
> > // of 4 (the fixed integer scale of the currency data type), and
so the
> > // actual scaling value must be adjusted as follows. This
adjustment
> > // avoids overflowing the currency value when scaling adjustment
> > will make
> > // the integer value fit. There is still a limit (14 integer
> > digits), but
> > // this should ensure we accept all possible values within the
limit.
> > AScale := AScale + 4;
> > if (AScale <> 0) and (AVal <> 0) then
> > begin
> > AFactor := 1;
> > for i := 1 to abs(AScale) do
> > AFactor := AFactor * 10;
> > if AScale > 0 then // increasing scale
> > begin
> > {$IFDEF IBO_VCL40_OR_GREATER}
> > if (AVal > (High(ISC_INT64) div AFactor)) or
> > (AVal < (Low (ISC_INT64) div AFactor)) then
> > raise Exception.Create( E_Int64Overflow );
> > {$ENDIF}
> > AVal := AVal * AFactor;
> > end
> > else // decreasing scale
> > begin
> > {$IFDEF IBO_VCL40_OR_GREATER}
> > AVal := AVal div AFactor;
> > {$ELSE}
> > AVal := AVal / AFactor;
> > {$ENDIF}
> > end;
> > end;
> > Result := currency((@AVal)^);
> > {$ENDIF}
> > end;
> >
> >
> > function CurrencyToInt64( AVal: currency; AScale: smallint ):
ISC_INT64;
> > {$IFNDEF IBO_VCL35}
> > var
> > i: smallint;
> > AFactor: ISC_INT64;
> > {$IFDEF IBO_VCL40_OR_GREATER}
> > tmpMod,
> > tmpHalf: ISC_INT64;
> > {$ENDIF}
> > {$ENDIF}
> > begin
> > {$IFDEF IBO_VCL35}
> > raise Exception.Create( 'Unsupported' );
> > {$ELSE}
> > // See notes for Int64ToCurrency for explanation of this code.
> > Result := ISC_INT64((@AVal)^);
> > AScale := AScale - 4;
> > if (AScale <> 0) and (Result <> 0) then
> > begin
> > AFactor := 1;
> > for i := 1 to abs(AScale) do
> > AFactor := AFactor * 10;
> > if AScale > 0 then // increasing scale
> > begin
> > {$IFDEF IBO_VCL40_OR_GREATER}
> > if (Result > (High(ISC_INT64) div AFactor)) or
> > (Result < (Low(ISC_INT64) div AFactor)) then
> > raise Exception.Create( E_Int64Overflow );
> > {$ENDIF}
> > Result := Result * AFactor;
> > end
> > else // decreasing scale
> > begin
> > {$IFDEF IBO_VCL40_OR_GREATER}
> > // Need to calculate the rounding adjustments required for
> > statistical
> > // rounding. Probably should setup asm stuff for this but lets
> > see how
> > // we go without for the moment.
> > tmpMod := Result mod AFactor;
> > tmpHalf := (AFactor div 2);
> > Result := Result div AFactor;
> > // more than half remainder goes up (down if negative)
> > // exactly half remainder goes up (down if negative) only when
> > // the result is odd (even results dont need adjustment)
> > if (abs(tmpMod) > tmpHalf) or
> > ((abs(tmpMod) = tmpHalf) and ((Result mod 2) = 1)) then
> > begin
> > if Result < 0 then
> > Dec(Result,1)
> > else
> > Inc(Result,1);
> > end;
> > {$ELSE}
> > // Sorry Delphi3 people but I will not write
> > // rounding for a system that I cannot check.
> > Result := Result / AFactor;
> > {$ENDIF}
> > end;
> > end;
> > {$ENDIF}
> > end;
> >
> >
> >
> > --- In IBObjects@yahoogroups.com, "Martijn Tonies" <m.tonies@> wrote:
> > >
> > >
> > >
> > > > I don't have much information to give thats the problem.
> > >
> > > Care to post the code at and around those lines?
> > >
> > > > I am trying to install a design time package which contains 2 pas
> > > > files to register the components and properties.
> > > >
> > > > When I build it, D2007 throws this:
> > > >
> > > > [DCC Error] IB_Utils.pas(1455): E2089 Invalid typecast
> > > > [DCC Error] IB_Utils.pas(1591): E2089 Invalid typecast
> > > > [DCC Error] IB_Session.pas(824): F2063 Could not compile used unit
> > > > 'IB_Utils.pas'
>
> Martijn Tonies
> Database Workbench - tool for InterBase, Firebird, MySQL, NexusDB,
Oracle &
> MS SQL Server
> Upscene Productions
> http://www.upscene.com
> My thoughts:
> http://blog.upscene.com/martijn/
> Database development questions? Check the forum!
> http://www.databasedevelopmentforum.com
>