[frostschutz]

you can see what it returns if you set debug=1

after some trial and error with dd and hexdump - it worked yet didn't. short reads missing the event we were interested in, delay due to unwanted pipe buffereing or waiting for additional block when read wasn't short, and added to that the general cumbersomeness of doing this for more than one device in parallel, etc. etc.

I caved in and wrote a little C helper to just poll across devices and just dump whatever events there are. The interpretation of those still happens on the shell side

A few things fell under the bus while doing so (whoops, touchscreen= button= update= is currently uneffective) I'll re-implement those later.

Right now it's experimental and the main question is, does it drain the battery or not, and does it work otherwise too.