On Mon, 11 Jan 2010, Mindaugas Kavaliauskas wrote:
Hi,
> From Xavi's sample:
> >HANDLE wapi_par_HANDLE( int iParam )
> >{
> > void ** ph = ( void ** ) hb_parptrGC( NULL, iParam );
> >
> > return (ph ? ( HANDLE ) *ph : ( HANDLE ) hb_parptr( iParam ));
> >}
> I guess, he is trying to avoid logic of GC pointers. Or maybe, some
> code is not ported to GC pointers and he wants to mix both versions
> of code. The problem of plain hb_parptr() is that in case of GC
> pointers it returns GC block (but not HWND; one more dereference is
> required), in case of simple pointers it returns raw HWND.
Thank you for clarification. I missed it.
> I guess it would be better if Xavi learn GC idea, and adopt some old
> or his own libraries to GC pointer style instead of trying to hack
> GC pointers code.
I also think so.
Xavi if you need it as temporary solution (i.e. for time necessary
to update some code) then simply create new file with hb_itemIsGC()
you send for your own use. You only have to include hbvmint.h before
hbapi.h, i.e.:
#include "hbvmint.h"
#include "hbapi.h"
BOOL hb_itemIsGC( PHB_ITEM pItem )
{
return pItem && HB_IS_POINTER( pItem ) && pItem->item.asPointer.collect;
}
but such function should not be part of core code because it unusable in
code which correctly validates all pointer item types adding protection
against GPF or even worse internal memory corruption when wrong pointer
item is passed as parameter.
best regards,
Przemek
_______________________________________________
Harbour mailing list (attachment size limit: 40KB)
Harbour@harbour-project.org
http://lists.harbour-project.org/mailman/listinfo/harbour
