Re: [AD] _al_current_display

[Trent G. <tr...@ga...>]

On Sat, August 9, 2008 4:13 pm, Peter Hull said:
> Bit of a whinge, but can we deprecate the use of
> _al_current_display?
> I think it's misleading.
> It's used like this: (in lots of places)
> ALLEGRO_BITMAP *al_get_backbuffer(void)
> {
>    ASSERT(_al_current_display);
>
>    return
> _al_current_display->vt->get_backbuffer(_al_current_display);
> }
> which would be OK if _al_current_display were a global but it's
> actually a macro
> #define _al_current_display al_get_current_display()
> so that function gets called 3 separate times (all right, 2 in
> release
> builds) .  al_get_current_display() is non-trivial because it
> accesses
> the thread-local storage stuff, and on OS X that means a call into
> pthreads.
> It may not be a problem but I think it would be clearer to write
> ALLEGRO_BITMAP *al_get_backbuffer(void)
> {
>   ALLEGRO_DISPLAY* current = al_get_current_display();
>   ASSERT(current);
>
>    return current->vt->get_backbuffer(current);
> }

Yes, I had the same thoughts. The only reason it is the way it is
now is because it  /used/ to be a global, and it was easier to make
a define rather than change all the code. We should definately phase
it out.

Trent :{)>