Well, I tried the new mouse code in Allegro 4.1.17 with Sunny Ball. It's better than the old code, but I think it's still too fast.

Attached is a modified 4.1.16 version that James Lohr kindly sent me. The mouse movement is slower, but feels much smoother to me. I'd rather use this, but will then of course lose the 4.1.17 advantages.

Is there any chance that new versions of Allegro will have an acceleration setting so we can set it to suit? Or even get the raw mouse mickeys to scale in our own way?

How hard is it to just track the movements and accelerate them yourself? Also, isn't mouse acceleration OS provided?

That's strange, since his original patch made it more or less unmodified into 4.1.17 (where there are still some problems with it that have been fixed in CVS)

Not to mention those of all future versions. By the way, may I say that you've picked a horrible time to upgrade to 4.1.17?

I've never used set_mouse_speed() personally, but I suppose you could use that.

No plans. The main issue with mickeys being that reporting them reliably in a portable way is rather tricky (read: impossible).

Again, the lack of active developers using Windows doesn't help much either...

The problem is that the mickeys are accelerated and accelerated incorrectly.

Well actually I had a look at wmouse.c in 4.1.17 . It looks like my patch has been tampered with in a way as to completely remove the improvement that I made. Why on earth has:

static int last_was_x = 0;
.
.
.
            if (last_was_x)
               last_mickeyy = 0;

been added? It undoes exactly what my patch added.

Furthermore I seem to recall that my patch was not applied at all to 4.1.16 (but i don't have a 4.1.16 handy to check this).

At the moment I'm having to rewrite some of wmouse.c every time I get a new version of the WIP 'cause the mouse acceleration has gone back to being absolutely awful.

As I recall, because messages being send to the message handler aren't guarenteed to be alternating x/y pairs. I'm not sure what the differences were, but it should be functionally similar.
As I recall, the problem was incorrect scaling (by accelerating only on the amount of x/y values change instead of by the total distance crossed). That should have been fixed.

Also, check out current CVS for some more changes that were made. I don't recall off the top of my head what they were, but the sourceforge tracker should help there.

Wether for 4.1.16 or 4.1.17, I did apply something equivalent to the patch you posted at some point (manually, the patch didn't apply through the patch program for some reason).

Rather than hacking about your copy of Allegro, it'd be really nice if you bring things like these up after a WIP is released - either on the forums, or through private message or e-mail. Because things that are broken don't get fixed in the next release unless we know they're broken.

EDIT:
The CVS revision that James' patch was applied in is 1.38, the 4.1.16 release is revision 1.36. Elias made two changes to this, one in revision 1.40 (which is also the 4.1.17 release) and again in 1.41.
The diff from 1.37 to 1.38 (ie, the acceleration patch) is

(Or check the coloured diff [url http://cvs.sourceforge.net/viewcvs.py/alleg/allegro/src/win/wmouse.c?r1=1.37&r2=1.38]here[/url])
Elias' first fix (that is in 4.1.17) is

(or [url http://cvs.sourceforge.net/viewcvs.py/alleg/allegro/src/win/wmouse.c?r1=1.39&r2=1.40]colour diff[/url])
and the second (in current CVS) is

([url http://cvs.sourceforge.net/viewcvs.py/alleg/allegro/src/win/wmouse.c?r1=1.40&r2=1.41]colour diff[/url]).

Oh sorry, I just remembered what's going on .

1.40, 1.41 look fine, the reason why the one I sent Richard was more smooth was because I wrote two patches - a simple one (the one implemented and improved on above) and a second one which is a bit more complicated. The second one is the one I use myself and is the one I sent Richard.

What a mess .

Sure, I'll write another patch this time for the wmouse.c in 4.1.17 with "added smoothness". I'll also include the changes made during CVS to my original patch (assuming they do not have a negative affect).

Great, thanks James! Although if 4.1.18 is going to be released in the next few days it might be best to wait..

If you release 4.1.19 in a few months time I can wait..

Doh!

I know this is off topic but, is there any way to turn mouse acceleration off so it just acts normal? I notice it's fine unless I run my allegro apps in windows and fullscreen.

This app is using allegro 4.0.3