Buffer related input event bug? -- 1.4.1~git20080131

[Rohan Lean]

On Tue, 2008-07-29 at 23:30 +0200, Simon Thum wrote: 
> Rohan Lean wrote:
> > Is anyone looking into the issue? Do you need more information?
> I'm not, but to me it looks very much like the game's problem, which 
> Xorg merely triggers. You should report it there, maybe make mention of 
> the input event rework on xorg's side.

As this happens in all games that would mean a problem in SDL, which was
my first guess. Of course, I did not consider that even if a downgrade
to the SDL version that still worked could still mean a bug there if it
made faulty assumptions that just happened to be true in the past.
I do still find that interpretation rather unlikely however. What
exactly does xorg do different but correct since the input event rework?

> You might want to try another mouse, gaming devices often exhibit 
> extreme reporting frequencies to minimize delay. This of course 
> maximizes processing time, which is your problem. Sweet irony.

The only other mouse with which I can actually reach the critical speed
uses 1000Hz polling, too, and has even slightly more CPI, 2000.
All other mice I have will not be able to reach the value because of bad
sensors, internal buffering and (!) low polling frequency.
But yes, lowering the CPI and increasing the sensitivity in the game of
course hides the problem at the cost of accuracy. That was not necessary
before however and I would prefer a bugfix to a dirty workaround.


--
Rohan