https://bugs.kde.org/show_bug.cgi?id=390698
kde@grau.net changed:
What|Removed |Added
CC||kde@grau.net
--- Comment #30 from
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #29 from Vlad Zagorodniy ---
Well, using render buffers was a bad choice. Apparently, glCopySubTexImaga2D
is used many times throughout KWin..
I've been trying to change GLTexture and GLRenderTarget so it's possible to
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #28 from Vlad Zagorodniy ---
(In reply to Christoph Feck from comment #27)
> Nice work, Vlad. If you think it is ready for review, please post the patch
> to https://phabricator.kde.org/differential/diff/create/
There
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #27 from Christoph Feck ---
Nice work, Vlad. If you think it is ready for review, please post the patch to
https://phabricator.kde.org/differential/diff/create/
--
You are receiving this mail because:
You are watching all
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #26 from Vlad Zagorodniy ---
https://github.com/zzag/kwin/commit/12cc19c69059f2db62ce5e077469ef31f7576099
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #25 from Vlad Zagorodniy ---
(In reply to Fredrik Höglund from comment #22)
> I think finalPrePaintScreen() should create the render buffer and render
> target on-demand, and also destroy them when
https://bugs.kde.org/show_bug.cgi?id=390698
Vlad Zagorodniy changed:
What|Removed |Added
Status|UNCONFIRMED |CONFIRMED
Ever
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #23 from Vlad Zagorodniy ---
(In reply to Fredrik Höglund from comment #22)
> Any transformation that does not just translate and/or scale the window.
> Although scaling can also benefit from this, so I guess any
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #22 from Fredrik Höglund ---
(In reply to Vlad Zagorodniy from comment #16)
> Fredrik, what do you mean by "a complex transformation"?
Any transformation that does not just translate and/or scale the window.
Although
https://bugs.kde.org/show_bug.cgi?id=390698
Vlad Zagorodniy changed:
What|Removed |Added
Severity|wishlist|normal
--
You are
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #21 from Vlad Zagorodniy ---
Created attachment 111747
--> https://bugs.kde.org/attachment.cgi?id=111747=edit
A personal effect + multisampling
--
You are receiving this mail because:
You are watching all bug
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #20 from Vlad Zagorodniy ---
Created attachment 111746
--> https://bugs.kde.org/attachment.cgi?id=111746=edit
Cube + multisampling
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #19 from Vlad Zagorodniy ---
Created attachment 111745
--> https://bugs.kde.org/attachment.cgi?id=111745=edit
Cover switch + multisampling
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #18 from Vlad Zagorodniy ---
Created attachment 111744
--> https://bugs.kde.org/attachment.cgi?id=111744=edit
Flip switch + multisampling
--
You are receiving this mail because:
You are watching all bug changes.
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #17 from Vlad Zagorodniy ---
So, I've managed to get antialiased edges of transformed windows.
https://github.com/zzag/kwin/commits/opengl-multisampling (PoC)
Short summary of changes:
* effects set
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #16 from Vlad Zagorodniy ---
It seems like GLRenderTarget should be refactored in order to use render
buffers. I suggest to use multisampled render buffers because they come from
OpenGL 3.0 while multisampled textures
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #15 from Vlad Zagorodniy ---
(In reply to Fredrik Höglund from comment #14)
> MSAA doesn't blur anything. It works by computing a pixel coverage value in
> the range [0..1] for each rasterized fragment, and multiplying
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #14 from Fredrik Höglund ---
MSAA doesn't blur anything. It works by computing a pixel coverage value in the
range [0..1] for each rasterized fragment, and multiplying that value with the
alpha value prior to blending.
You
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #13 from Vlad Zagorodniy ---
Turning on fullscreen multisampling makes sense with effects like Cube, Cover,
etc.
What about effects which transform individual windows? There could be a short
time when all fonts look
https://bugs.kde.org/show_bug.cgi?id=390698
Fredrik Höglund changed:
What|Removed |Added
CC||fred...@kde.org
--- Comment
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #11 from Vlad Zagorodniy ---
(In reply to Martin Flöser from comment #10)
> The general design of the compositor is described in scene.cpp,
> but I fear it's too high level to be of much help.
Thank you for pointing
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #10 from Martin Flöser ---
(In reply to Vlad Zagorodniy from comment #7)
> Offtopic: Martin, did you give any talks or have articles about KWin
> architecture? I wish I could help somehow but my current level of
https://bugs.kde.org/show_bug.cgi?id=390698
Nate Graham changed:
What|Removed |Added
CC||n...@kde.org
--
You are receiving
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #9 from Vlad Zagorodniy ---
Created attachment 110803
--> https://bugs.kde.org/attachment.cgi?id=110803=edit
There is still present aliasing
This is not a silver bullet. There is still little aliasing.
+ I think it
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #8 from Vlad Zagorodniy ---
Created attachment 110801
--> https://bugs.kde.org/attachment.cgi?id=110801=edit
Comparison of different cases with/without MSAA
Top row: without MSAA
Bottom row: with MSAA
Left column:
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #7 from Vlad Zagorodniy ---
I think this issue should be revised. Windows Vista and Windows 7 had smooth(as
far as I remember) edges in the Flip 3D effect a long time ago.
--
Offtopic: Martin, did you give any
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #6 from Martin Flöser ---
back in the days with flip switch one would need to enable multisampling for
the front window at least. And that's the most prominent one. Imagine the
browser window with all text blurred.
But
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #5 from Vlad Zagorodniy ---
(In reply to Martin Flöser from comment #4)
> Back when I started working on KWin and introduced the first 3d effects I
> also observed this problem and tried to fix it. It failed in a
https://bugs.kde.org/show_bug.cgi?id=390698
Martin Flöser changed:
What|Removed |Added
Severity|normal |wishlist
--- Comment #4
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #3 from Vlad Zagorodniy ---
Created attachment 110792
--> https://bugs.kde.org/attachment.cgi?id=110792=edit
the same with cover switch(also it has problems with window decorations)
--
You are receiving this mail
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #2 from Vlad Zagorodniy ---
Created attachment 110791
--> https://bugs.kde.org/attachment.cgi?id=110791=edit
Top & bottom edges of windows in flip switch are jaggy too
--
You are receiving this mail because:
You are
https://bugs.kde.org/show_bug.cgi?id=390698
--- Comment #1 from Vlad Zagorodniy ---
Created attachment 110790
--> https://bugs.kde.org/attachment.cgi?id=110790=edit
Left edge of google chrome is very jaggy
--
You are receiving this mail because:
You are watching all bug
32 matches
Mail list logo