Re: Pointer/screen problems, part two

On Wed, 2 Oct 2002, Owen Taylor wrote:

> > in light of the long term plan to simply move screen into the
> > event structure for 2.4, datasets are good and easy enough
> > to use for migration (in the end, that's what they exist
> > for and provide a hash table for you already),
> You mean, for 3.0? 3.0 is a long time off...

well, that doesn't mean it sould be out of consideration at this point ;)

> Question on gdk_event_new()
>  GdkEvent *gdk_event_new (void);
> or:
>  GdkEvent *gdk_event_new (GdkEventType type);
> ? Advantages of the second:
>  - We could theoretically allocate the event short
>    for smaller events.
>  - We can guarantee that the event is initialized
>    properly for the particular event type.
>  - It corresponds better to the idea that events
>    "derive" from GdkEvent. Changing the type of a
>    GdkEvent is conceptually like turning a GtkButton
>    into an a GtkEntry.
> Still, it would cause problems for fill-in interfaces
> that work like gdk_event_translate() (inside GDK we
> could clearly use an internal interface). So on the 
> balance, preference is the simple first form.

really? after you outlined three good reasons for the second
form? you have my vote for gdk_event_new (GdkEventType type)
since i'd like to conceptually think of it as a typed object.

> Regards,
>                                         Owen


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]