Graphics device reset/loss scenarios need to be tested more thoroughly, on Windows at least. Live driver updates, ability to disable devices, unplug external devices, getting a device reset (and losing graphics resources) upon specific events, etc. has been a thing for at least a decade or more on platforms with a reasonable driver model. Qt should be better prepared for this in the future.
- Must verify if a Quick app (with default settings, no transient images etc.) can survive a graphics driver update when running with the D3D11 backend of QRhi (which is the absolute minimum expectation for modern applications)
- VkDevice loss testing for Vulkan, Windows at least - first test resulted in an infinite wait somewhere, instead of recovering so needs further investigation.