Tool/software: Linux
Customer working to try to optimize their html page load times, driving two displays LCD and HDMI (with different content). Using current Linux PSDK v5.02 on AM572x EVM. They use multiple QWebEngineViews to "preload" several different control view, then the idea is they can rapidly switch between them. This works well once the pages are initial rendered, but they been trying to improve what happens the first time the page is loaded. As it is, when the page is shown for the first time, there is a fairly ugly white flash that is shown for maybe a couple hundred ms, and then their page is shown fully rendered. Our pages do take some time to draw so the delay is understandable, but they need a solution to avoid the white flash foir their product.
One odd thing is that is noticed when they first load a web page, Qt emits a debug message:
Non-blocking swap buffers not supported. Subsurface rendering can be affected.
What does this mean?
Might have something to do with the fact that the version of Qt which comes with the board kit has openGL enabled. I pass --disable-gpu when I start the QT application, and I also use ensure my WebEngineView sets the settings to QWebEngineSettings::WebGLEnabled, false and QWebEngineSettings::Accelerated2dCanvasEnabled, false, but they still get the following debug messages from Qt:
Using Wayland-EGL
wlpvr: PVR Services Initialised
Non-blocking swap buffers not supported. Subsurface rendering can be affected.
Is there's an easy way to completely remove open GL support? They do not use it, and they believe it could be limiting thier performance. Would this mean they will need to rebuild Qt without openGL support? They tried not starting pvr in /etc/init.d/rc.pvr, but then the application does not work at all (nor does matrix browser).
Any suggestion/ideas on how to force QtWebEngineView to render the web page before it is shown somehow. They have tried everything what thye can think of