Bug 46793

With compositing? xorg-server-1.11.2 is known to be buggy with compositing after a modeswitch (which can happen just by logging in).

(In reply to comment #1)
> With compositing? xorg-server-1.11.2 is known to be buggy with compositing
> after a modeswitch (which can happen just by logging in).

Without, compositing is still turned off due to bug #37686. I think it happens also on a clean login, but I don't know how to tell if this involves a modeswitch. It certainly happens /after/ a modeswitch (suspend to disk/ram) though. 

Would it help to try this by starting X with 'startx' (since I don't have compositing enabled)?

(In reply to comment #2)
> (In reply to comment #1)
> > With compositing? xorg-server-1.11.2 is known to be buggy with compositing
> > after a modeswitch (which can happen just by logging in).
> 
> Without, compositing is still turned off due to bug #37686. I think it happens
> also on a clean login, but I don't know how to tell if this involves a
> modeswitch. It certainly happens /after/ a modeswitch (suspend to disk/ram)
> though. 
> 
> Would it help to try this by starting X with 'startx' (since I don't have
> compositing enabled)?

I had the opportunity, so I tried this anyway. I booted with the 'nox' option to prevent X from starting automatically, then logged in to a console and executed 'startx'. 
Under these conditions the issue is still present and easily reproduced using the step above.

I suspect this is related to:

commit 6dfe28f978b59ab2fc44cdd8ddf3743d6507588b
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Mar 13 15:47:27 2012 +0000

    sna/gen6: Remove the double application of the render offset
    
    Cut'n'paste error from an older generation.
    
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>

Can you please retest with current xf86-video-intel.git?

(In reply to comment #4)
> I suspect this is related to:
> 
> commit 6dfe28f978b59ab2fc44cdd8ddf3743d6507588b
> Author: Chris Wilson <chris@chris-wilson.co.uk>
> Date:   Tue Mar 13 15:47:27 2012 +0000
> 
>     sna/gen6: Remove the double application of the render offset
> 
>     Cut'n'paste error from an older generation.
> 
>     Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> 
> Can you please retest with current xf86-video-intel.git?

I just tested with fe4ad66aadcfa0c907ba6460a88c49e65225602c and the issue can still be reproduced with the steps described above.

(In reply to comment #5)
> (In reply to comment #4)
> > I suspect this is related to:
> > 
> > commit 6dfe28f978b59ab2fc44cdd8ddf3743d6507588b
> > Author: Chris Wilson <chris@chris-wilson.co.uk>
> > Date:   Tue Mar 13 15:47:27 2012 +0000
> > 
> >     sna/gen6: Remove the double application of the render offset
> > 
> >     Cut'n'paste error from an older generation.
> > 
> >     Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
> > 
> > Can you please retest with current xf86-video-intel.git?
> 
> I just tested with fe4ad66aadcfa0c907ba6460a88c49e65225602c and the issue can
> still be reproduced with the steps described above.

Open-office might also be affected by this bug, and easier to test:
http://www.zimagez.com/zimage/screenshot-04232012-054727pm.php

On the right side, next to the scrollbar you can see some square yellowish artifacts. These seem to be left by the tooltip that pops up when you move the scrollbar by dragging it with the mouse (it shows the current row). It may be necessary to move the mouse quickly to make the artifacts appear.

Tons of little fixes with quite a few in damage tracking that may be of note. Do you still see the unredrawn areas on the screen?

Ping?

(In reply to comment #8)

Just tested with xf86-video-intel 5d2f88fd9972c62c87098ddc7fee7b6f0cea0fdb .
This no longer seems to happen with the OpenOffice scrollbar/tooltip.
The autocomplete popups in omnet however still fail to 'unpaint' properly.

> Ping?
Sorry about the huge ping :)

Does a screenshot capture the omnet popup bug?

(In reply to comment #10)
> Does a screenshot capture the omnet popup bug?

There is a screenshot in comment #1 (the bottom half of the popup has disappeared, but the top half remains). So, yes, it can be captured on a screenshot.

Been a while since I've heard whether this is still an issue.

(In reply to comment #12)
> Been a while since I've heard whether this is still an issue.

Still happens with open office, but it takes a bit of effort:
When quickly scrolling the OOCalc sheet (you can do this with an empty sheet), I can sometimes see a bit of corruption similar to that in the screenshot from Comment 6. This however seems to clear up by the time you stop scrolling. Occasionally a bit of the tooltip still remains though.

The best way I've found to trigger this is to cause the tooltip to move while scrolling. The tool tip will move to stay out from under your cursor. Here's what I mean:
1) Click-hold the scrollbar.
2) Slowly start scrolling.
3) The tooltip will appear.
4) While scrolling up/down, try to move your mouse over the tooltip.
5) The tooltip will move out of the way.

The easiest way to do that quickly is to move your mouse in circles a little bit while scrolling (or use a trackpad ;).
I think the bug is most likely to trigger when you scroll the currently selected cell in/out of view.


Alas I no longer have access to the Omnet software, so I cannot test that (the issue was much more pronounced and hence annoying there though).

I have this theory that by closing this bug will cause 20 people to suddenly spot it - including myself...