Hi
Viktor Szakáts wrote:
>
> I've just checked further and the fix is quite simple
> and obvious, so I can make it.
>
> The other comment is that maybe there is no need to
> make any sort of hack to check for "Q" or "HB" in the
> beginning of class name, but simply check whether
> :pPtr member exists and retrieve it if it does. Low
> level code check for required object pointer type
> anyway, so there is no need to add these slow and
> error-prone checks on the higher level.
>
> Here is a non-optimal/pseudo version of this method:
> (optimal could be done in C probably)
>
> ---
> FUNCTION hbqt_ptr( xParam )
>
> IF hb_isObject( xParam ) .AND. ;
> __objHasMsg( xParam, "PPTR" ) // we need something which doesn't RTE
> on missing msg and which check for variable.
> RETURN xParam:pPtr
> ENDIF
>
> RETURN xParam
> ---
>
> (I assumed XHB stuff is already removed from here)
>
> Same thing could be rewritten in C probably, for
> smoother/faster results.
>
(I assumed XHB stuff is already removed from here)
But why we do not want to include XBP stuff here ?
This does not interacts with XBP anyway, only XBP interacts
with it, and this poses no harm.
The final function skeleton may be like this
//-------------------//
FUNCTION hbqt_ptr( xParam )
IF hb_isObject( xParam )
if __objHasMsg( xParam, "PPTR" )
RETURN xParam:pPtr
elseif __objHasMsg( xParam, "OWIDGET" )
RETURN xParam:oWidget:pPtr
endif
ENDIF
RETURN xParam
//---------------------//
It will allow to mix HBQT and HBXBP use the same syntax
with interchangele way.
Of course, C level call is much more efficient, I will give it a try.
I have more changes to commit, so please do not commit
anything here, I will do thw spadework, just suggest for
some time.
Regards
Pritpal Bedi
--
View this message in context:
http://old.nabble.com/SF.net-SVN%3A-harbour-project%3A-13275--trunk-harbour-tp26826400p26834186.html
Sent from the Harbour - Dev mailing list archive at Nabble.com.
_______________________________________________
Harbour mailing list (attachment size limit: 40KB)
[email protected]
http://lists.harbour-project.org/mailman/listinfo/harbour
