Hi,

Using fpc 2.1.5 on Mac Intel, I am getting a wierd error on this line:

  Buffers[CurrentBuffer].DrawTo(MainForm.Canvas.Handle, 0,0);

With the following declarations:

  public
    { Public declarations }
    CurrentBuffer : Boolean;
    Buffers: array [Boolean] of TBitmap32;
....

This exception:

TApplication.HandleException Invalid type cast

And this backtrace:

Breakpoint 1, 0x0000e0a6 in fpc_raiseexception ()
(gdb) bt
#0 0x0000e0a6 in fpc_raiseexception ()
#1 0x00056682 in SYSUTILS_RUNERRORTOEXCEPT$LONGINT$POINTER$POINTER () #2 0x00010fd7 in SYSTEM_HANDLEERRORADDRFRAME$LONGINT$POINTER$POINTER () #3 0x0000cf71 in fpc_class_as_intf ()
#4 0x00027ebf in TMAINFORM__TRANSFORMFRAME (SENDER=0xfb3020, DONE=true, this=0xf2c810) at MainUnit.pas:359
#5 0x00022bf4 in TAPPLICATION__IDLE (WAIT=true, this=0xfb3020) at application.inc:299
.....

What kind of typecast is that? I don't see any on the code. And why is fpc_class_as_intf being executed? I don't see any interfaces being used.

This code comes from a delphi example application, and I'm a little lost about what is wrong or how to fix it.

Maybe the sintax array [Boolean] doesn't work and I should substitute it by something like [0..1]

thanks,

Without compilable example code, that is impossible to say.

Jonas

Jonas Maebe wrote

Without compilable example code, that is impossible to say.

Sorry for the noise, I found the problem =)

Inside the DrawTo function code there is a cast to an interface which isn't implemented.

I just suposed that on the backtrace gdb would show exactly where the exception occured. There were more functions called at the point where the exception really happens ... I don't know why he skiped some. Debuging with the IDE step-by-step showed the correct location of the exception.

thanks,

Maybe optimizations were turned on.

Jonas