| Re: [AD] Using system mouse cursor |
[ Thread Index |
Date Index
| More lists.liballeg.org/allegro-developers Archives
]
On Sunday 05 September 2004 02:20, Chris wrote:
> Nope, changes nothing. :/
Ok, I'm going to revert in my local tree then. It seems to cause some
erratic behavior in some of my own programs.
> I did, however, completely rebuild AllegroGL (since that's what my
> program uses) and it cleared up some of the problems, but the mouse
> still seems to be eratic when reentering the game window. It'll stutter
> around, sometimes fixing itself after a few moments and other times not.
Are you using the mickeys at all? Or just with reading the mouse position
and displaying it on the screen?
> Me and Elias were discussing, before you had a chance to get on IRC..
> and we think the best thing to do is remove the mouse warping hack. I
> think what we should do is require the user to periodically call
> position_mouse if they want continued movement, and either leave
> get_mouse_mickeys undefined when it goes past the screen edge or force
> it to not get movement when it does (for platforms that do get continued
> movement).
I think this makes sense for windowed mode at least. I'd like the mouse to
have two modes: a mickey mode and a normal mode. Mickey mode disables the
hardware cursor and re-enables the mouse warping.
As far as I can tell, this is what SDL does too. The problem is that it
breaks the API somewhat. On the plus side, it eliminates a lot of problems
caused by features someone may not want to use at all.
> The only places get_mouse_mickeys has ever worked right,
> AFAIK, is DOS, fullscreen modes in Windows (windowed modes have problems
> when hitting the edge of the desktop), and possibly Linux console.
> Everywhere else has either not worked right or required hacks to do so.
I think that's right.
> PS. BTW, what would happen if a program tracked the mouse via
> get_mouse_mickeys and drew it manually? Like if they want to implement
> some kind of mouse smoothing feature (like many FPSs do nowadays).
> AFAICS, the mouse would enter "mickey mode" and then get stuck in one
> spot. The warp hack assumes the user won't be calling get_mouse_mickeys
> if they're drawing it manually.
The mouse_x and mouse_y variables (positioning Allegro's mouse) continue to
be updated normally. I'm not sure if that's what you meant?
> Plus, it seems to have a nasty habit of
> making Allegro switch to drawing it in the timer and back again if
> conditions are "just right". IMO, the only thing that should cause it to
> switch back to timered drawing is an invalid (to the system) image..
> having it possibly do it literally at any time is very ugly.
Yes, it is.
The mouse routines could actually benefit from a rewrite too, I guess...
Evert