https://bugs.kde.org/show_bug.cgi?id=410918
Dmitry Kazakov <dimul...@gmail.com> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |dimul...@gmail.com
--- Comment #2 from Dmitry Kazakov <dimul...@gmail.com> ---
Hi, Agata!
Here is a checklist of what needs to be checked to find out the bug. You will
have to add debugging lines into Qt's code and check what happens there.
1) Go to QOpenGLTextureBlitter::create() and check if the conversion shaders
are actually created.
2) Go to fragment_shader[] in the texture blitter, find the all _TO_BT2020PQ
branches and add explicit color return there, like that:
return vec4(0.8, 0.3, 0.3, 1.0);
Then build and run Krita. If the canvas becomes pink instead of black, then the
problem in the shaders, and not in the Angle's code. If the canvas is still
black, then go on with the following:
3) Check the following functions in ANGLE. All internal system calls should
return E_SUCCESS (or how it is called in DXGL).
* SwapChain11::reset()
** especially this call: mNativeWindow->createSwapChain()
* NativeWindow11Win32::createSwapChain()
Take into account that there are multiple swap chains in Qt. One global, and a
separate swap chain for every QOpenGLWidged.
4) If you still have no clue what happens there, download apitrace and
qapitrace and run it with "api" switch set to dxgi. It will record all the
calls to DXGI/DirectX and you will able to trace/debug it separately.
--
You are receiving this mail because:
You are watching all bug changes.
