[compiz] [PATCH] Water fixes

[Erkin Bahceci]

On 5/14/07, David Reveman <davidr at novell.com> wrote:
> It can give slightly different results depending on if the software or
> hardware water simulation is used but it should be close enough to not
> really matter.

Here the FBO version was actually fine even with 40, but software
rendering needed at most 35.

> However, when I use your patch I see cases where I find
> it to end the water simulation much too early. E.g. when you turn off
> the rain effect and rain_delay is set to 1.

Interesting, both fbo and software work fine here with 35. How is this
machine dependent if the countdown and drawing is not time-based? Do
the floats in the fragment program have different accuracy on
different machines? Does the early ending happen with software
rendering, too?

Regards,
Erkin