Phillip Berndt

I need to reopen this: Turns out the fix broke #143 / wasn't the right one. My current understanding is that some video drivers will not automatically reset the mode...

There's a feature for wildcarding EDID's per output, but the inverse doesn't exist yet. The simplest way to add this, I guess, would be to extend `XrandrOutput.from_config_file` to support finding...

Try adding print(find_edid) print(options["output"]) print(inverse_edid_map) below the first chunk of changes and paste the output, please.

See #145.

Thanks! The segfaulting process likely is xrandr, and I guess that the issue is related to #87, but as long as this isn't reproducible and doesn't occur frequently I doubt...

The xrandr output from the time when the "Parsing XRandR output failed.." message appears is what'd be interesting. (Parsing the output you've pasted above works as expected.) Try adding to...

Since at `800x512` there's at least the 208 bytes from that mode's verbose data, but autorandr claims that there's 91 bytes left unmatched, the output must be truncated. Which means...

IIRC, in the past, we couldn't use this since this only worked once - the second change wouldn't trigger the service execution anymore. #61, iirc.

This behaviour is documented in systemd.service(5), I'm quite certain it's intentional.

The key expectancy is that if you then connect the screen that you just unplugged again, autorandr should run again. Maybe they changed the behaviour - this used to be...