[Intel-gfx] [ANNOUNCE] xf86-video-intel 2.8.0

[Clemens Eisserer]

Hi.

> You also said (quote) "File good bug reports and please make use of
> the "regression" keyword as appropriate. We pay attention to that".
> Well, I tried that. The thing is: I think I'm just getting ignored and nobody
> is paying attention.

I experience exactly the same. I filed a few xrender bugs almost a year ago:
- https://bugs.freedesktop.org/show_bug.cgi?id=16917
- https://bugs.freedesktop.org/show_bug.cgi?id=17933
- ( https://bugs.freedesktop.org/show_bug.cgi?id=16913 )

The first two are critical for a project I am working on, and are
still not fixed.
The project is about converting java to use XRender which will be
integrated into mainline soon, so millions of people will stumble over
this. The first even causes corruptions with FireFox.
The RadeonHD team was able to fix the first mentioned issue in their
driver 3 weeks after I reported it.

As well as a performance regressions:
- https://bugs.freedesktop.org/show_bug.cgi?id=18075
- https://bugs.freedesktop.org/show_bug.cgi?id=22127

Both performance bugs are marked as *regression*.
Except from some small-talk nothing happend.

So bottom line is, my Software still looks broken because of XRender
bugs that never got fixed, but now performs worse than ever thanks to
recent driver changes.

And now please tell me I should not complain, that you have many more
important bugs to fix, that I should stay calm and patient ... for
another year?

- Clemens

By the way, the JXRenderMark benchmark is used by a *serious* and
large GPU manufacturer for nightly regression testing. This would not
harm the intel driver at all....
Sure, synthetic testing doesn't reflect real world, but if there's a
regression in a synthetic bench, the real-world results most likely
(if you don't test crap) won't get better.