I've created a video - unfourtunatly quality is very poor and the leftovers caused by the cursors are barely visible.  (sound can be ignored, recorded at a conference talk): http://youtu.be/fkFuzZKKObs

I can't make out what you are trying to show me in the video. Is the artifact in the selection region or the cursor itself?

And of course it helps if we clarify which cursor you mean. The hardware mouse cursor or the keyboard input cursor?

Heh, and we might have to wait a while before eclipse-4.2 hits debian ;-)

Created attachment 67587 [details]
screenshot

sorry for the sloppy bug-report, I thought the screenshot would make it all clear - which I forgot to attach ;)

There are left-overs from the text-input cursor (looks like XOR rendering).

If you are interested, pre-compiled ready-to-run binaries are available at: http://www.eclipse.org/downloads/download.php?file=/technology/epp/downloads/release/juno/R/eclipse-java-juno-linux-gtk-x86_64.tar.gz

Thanks, Clemens

Ok, finally had a chance to being to look at this and running the 4.2 package you pointed me at, thanks.

The selection rendering looks dependent upon the theme, can you tell me which GTK+ theme you are using?

Clemens, can you also check that this isn't fixed already on master? I can hope :)

As I haven't seen this issue after updating eclipse to "4.2.1" (don't think they call it like that), so although I didn't see it with UXA - its now gone under SNA as well :)

Although it could be a magical fix in master as well ;)

As always, thanks a lot

Ok closing as worksforme until we can reproduce again, and just assume that this was fixed along with one of the other bugs, thanks.

Ok, this bug was revived from the dead on my SNB laptop by commit:

[b3d1118bbee1172f72c946163a37ca4ad5feecce] sna: Promote the CopyArea to the GPU if it subsumes the CPU damage

On the positive side, now I might be able to track this down and fix it for real. :(

Is this bug limited to your snb machine?

(And debian is still only on 3.8!)

> Is this bug limited to your snb machine?
I currently do not have another system to test - will re-setup my arrandale machine in a few hours.

> (And debian is still only on 3.8!)
For a good reason ;)
Binary packages are quite ok, however.

I can see an occasional residual cursor in the highlighted section on eclipse-3.8 and SNB. Will rebuild with debugging enabled and see if I can spot what is going on.

just tested - the issue is also reproduceable on my i5 540M.

The easiest was to reproduce it is to position the cursor at the end of a line and slowly move the mouse to the left, therefore slowly selecting more and more of the line from right to left.

Slightly different result:

94cbe7bf7b7acc9a7f2cb56c5e275af028f3cdc8 is the first bad commit
commit 94cbe7bf7b7acc9a7f2cb56c5e275af028f3cdc8
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Thu Jun 20 19:40:44 2013 +0100

    sna: Mark overwriting CopyArea as not needing the dst damage

commit 2e2c448a77ab9dce4807b159708290cd7ad26b5c
Author: Chris Wilson <chris@chris-wilson.co.uk>
Date:   Tue Jun 25 14:22:28 2013 +0100

    sna: Don't flag IGNORE_CPU for partial overwritten blocks
    
    As we set the flags for the migration extents, but only operate upon the
    boxes in the region, we need to be careful to be sure to migrate data
    outside of the boxes overwritten.
    
    Fixes regression from
    commit 94cbe7bf7b7acc9a7f2cb56c5e275af028f3cdc8 [2.21.10]
    Author: Chris Wilson <chris@chris-wilson.co.uk>
    Date:   Thu Jun 20 19:40:44 2013 +0100
    
        sna: Mark overwriting CopyArea as not needing the dst damage
    
    If we start passing around regions rather than extents, we coud do
    finergrained migration decision.
    
    Reported-by: Clemens Eisserer <linuxhippy@gmail.com>
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=55244
    Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>