| Re: [AD] Using system mouse cursor |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
On Sunday 03 October 2004 00:43, Chris wrote:
> If Allegro can't set a system cursor, no it shouldn't switch back to a
> software cursor. It should switch to a custom hardware cursor, using
> Allegro's default cursors.
Not all drivers support hardware cursors - and most drivers that do also
support system cursors. So switching to a custom cursor when there is no
system cursor isn't that useful if it should only be done for hardware
cursors IMO.
> Why are these seperate at the API level? The only way to get a system
> cursor is to pass NULL to set_mouse_cursor, but this can still give you
> a custom cursor if a system one isn't available. How does/why should the
> API differ past that?
Ok, maybe not API. I meant from the user's perspective.
The user decides if he wants a system or a custom cursor. Allegro decides
if it uses a hard or soft cursor to display it. That's what I meant (I
think - too late to think clearly).
> I think you're over-analyzing it. Custom cursors are done in software,
> unless the mouse driver can put them into hardware/the OS (and the
> appropriate function has been called). If a NULL cursor bitmap has been
> set to use, it selects a system cursor if available, or else it uses
> Allegro's default (in hardware).
The latter part is what can't be guarenteed (hardware cursors aren't
supported by all drivers).
> * If bitmap is NULL, try to set a system cursor (done by the OS/
hardware).
> * If that fails/isn't available/bitmap isn't NULL, try to set a custom
> hardware cursor using bitmap (if not NULL) or Allegro's default.
> * If that fails (shouldn't if bitmap is NULL), set a software cursor.
Yes. This is what the code I posted does (although it could be cleaned up).
At this point, I think the discussion is mostly about semantics on what
under which circumstances should be called a system/hardware or software
cursor.
Evert