John Novak

I'm not in favour of using system notifications for this purpose. In the case of Munt it's actually a good solution because usually you're using it in conjunction with an...

> It looks EVEN BETTER than the Boxer App example above! Thanks for the mockup, @johnnovak! ...and it will become reality, unless I'm hit by an evil meteor! 

Do we still want to get rid of `surface` @kcgen? I think the answer is yes, just double checking.

> Absolutely. Even though `surface` might be the only output mode usable on obscure platforms (say, like OpenIndiana or some BSD virtual machines), practically these are often poor performing configuration...

> Because the logic doesn't keep a running tab for these 'sub pixel' dx, dy values, the accumulated position isn't considered, and the behaviour you mentioned is quite common (fully...

> @johnnovak Are you sure this is an emulator issue, and not the game issue? I could reproduce the problem: > > * with DOSBox Staging built-in mouse driver >...

> gives a clue: the game does not like high mouse sampling rates. And I confirm this by experimenting with various DOS mouse drivers and enforced/hardcoded settings. > BTW, the...

Just to be clear, this is *game* issue; the weird mouse behaviour is 100% reproducible on real hardware. There's nothing to fix here at the emulator level. When I raised...

Yeah, I've also noticed that in many games the pointer moves _much_ smoother in DOSBox than on real hardware. CPU speed doesn't really matter here, I'm using a fast Pentium...

Cheers, I've added these to the top description.