> Also, the test cases seem to work by verifying the vblank period time with a
> tolerance of 10ms. Bug 880707 had nothing to do with the period/frequency, the
> period was always correct thanks to CompTimer. Bug 880707 was caused by poor
> _phase_ timing and the tolerance seems to be a max 2-3ms. It's testing the
> wrong code and using the wrong method... :(
Thanks. I'll have a look into fixing it a little later this week then.
As for using GLX to do it, I'm open to that idea, although I wanted to form a testcase for this outside of opengl since there might be other uses for it.
> Also, the test cases seem to work by verifying the vblank period time with a
> tolerance of 10ms. Bug 880707 had nothing to do with the period/frequency, the
> period was always correct thanks to CompTimer. Bug 880707 was caused by poor
> _phase_ timing and the tolerance seems to be a max 2-3ms. It's testing the
> wrong code and using the wrong method... :(
Thanks. I'll have a look into fixing it a little later this week then.
As for using GLX to do it, I'm open to that idea, although I wanted to form a testcase for this outside of opengl since there might be other uses for it.