This issue continues but took a very unexpected turn.
It's so strange that of the software I use frequently, it's only affected RStudio.
I woke up the other morning convinced I'd sorted it out. It must be opening on a phantom screen, or it's opening on the laptop screen despite it being closed, but after coffee realized it wasn't showing up in the taskbar, the task switcher or the task manager. Swing! And a miss.
The equipment in question comprises a Gigabyte Aero Classic laptop, a Dell D6000 USB-C dock and an LG multiple-input monitor. The laptop sits in a little rack with it's power supplies and dock. The dock connects to the monitor via Display Port out to HDMI in, and is connected to the wired network and numerous USB cables cluttering things up. It's almost never opened, just sits in it's little space and is used via the peripherals (welcome to month 15 of Morbidly Stationary 2020).
I'd conceded defeat, backed everything up, removed the laptop from it's spot and disconnected from the dock. Rebooted to try and catch the recovery screen for a full refresh, but missed the critical millisecond and went to a full boot. There was the RStudio icon staring at me, taunting me. So I clicked it. And suddenly, it worked.
What. The. Hell?
Now what? We test. And test and test and test and discover new tools and information, like the Windows Reliability tool.
(Other than a couple of specific-incident data points, the failures have the same information. )
I tested the laptop on it's own, connected to the dock and connected directly to the monitor via USB-C as opposed to via dock's display port. I tried multiple combinations of resolutions, scaling, duplicating, extending and assigning "main" status. The results:
The laptop on it's own works exactly as one would hope. It functions flawlessly.
If connected to the dock, the laptop must be open and the laptop screen must be the main monitor, whether it's the laptop only, in duplicated mode or extended mode. The resolution and scaling play no part (as anticipated).
Leaving it connected to the dock (network, peripherals), but driving the monitor via the 2nd USB-C port (on the laptop), it again functions exactly as one would expect. With the laptop screen, without the laptop screen, duplicated, extended, main, sub, whatever you want to throw at it, it works.
So, I have a work-around at least. but it's not a solution. RStudio remains the only program affected, at least in so much as I've tested my "normally used", there may be others that I did not test. It's clearly related to a) the Dell dock and b) the dock display output. Comparing the timelines of Windows driver updates and the sudden failure, there is a definite correlation.
BEX64 is often either a video driver issue, or a USB-driver issue which sort of narrows it down, but not really. the laptop has the latest video drivers and there hasn't been an update in the Dell software for the dock in a long time. I'll keep picking away at it and we'll see what we find. Perhaps it's as simple as rolling back the Windows driver update.