Hi Pritpal, I agree with whole concept. > Let us do it right, right now. Only pointers, > no numeric handles at all. Basically, if we have > to use existing code too, we will be changing the > function calls and there stays the whole effort. > > So, it should be like: > #define wapi_par_HWND( n ) ( ( HWND ) ( HB_PTRDIFF ) hb_parptr( n ) ) > just to avoid the cost of - ( ISNUM( n ) ? ( HB_PTRDIFF ) hb_parnint( n ) > : hb_parptr( n ) ) > OR > the flexible way is better in the long run?
My thinking was that this way, hbwin may work better with other Windows libs or apps, where the developers were still using numeric pointers. This way you can mix these in some occasions. Probably you and Francesco could tell better, I personally don't use any Windows libs, nor the API for anything complicated. If Windows users think it's not needed, it can be dropped, and it would certainly result in more consistent code on both levels, it creates a cleaner state and maybe it will slightly push other Windows libs to use pointers, too. So you have my vote for dropping this, if you think it's not necessary. > BTW, this artical is really thought provoking. I'm glad you liked it. It made me thinking again about the two parallel make systems we have :) Brgds, Viktor
_______________________________________________ Harbour mailing list Harbour@harbour-project.org http://lists.harbour-project.org/mailman/listinfo/harbour
