Until now I've never bothered to look into the internals of keyboard software support and whatnot, but for a long time I somehow put up with #Alpine #Linux (or maybe it was a #KDE / #Wayland issue?) not enabling by default key repeat - or however you call holding a button and have it register as continuous holding instead of a single press.
Lo and behold, on #OpenBSD that just works (I just passed by the conf where key repeat is explicitly defined, so you know real people put real effort into this system). On the other hand, changing the language on my keyboard when using cwm instead of something like KDE?...
setxkbmap -layout ro does not output an error, but still doesn't mean it actually switched me to Romanian (however, something like French actually just works). wsconsctl keyboard.encoding=ro outputs the error that ro is not a valid encoding. According to the documentation, encodings are apparently listed in /usr/include/dev/wscons/wsksymdef.h - and indeed, there seems to be no "ro" in there. Changing locale didn't seem to help either.
Then I took a deep dive into the man page of wsconsctl(8). There it says: "The current mapping can be printed with wsconsctl keyboard.map. The value for each keycode specifies the keysym that is output when each of Key, Shift + Key, AltGr + Key, or Shift + AltGr + Key is pressed" A magic thing then happens... I test wsconsctl keyboard.map+="keycode 15 = l L at" - afterwards, I see in the keyboard mapping "l L at at"; the output is a Polish l=L with slash. I decide to test AltGr with every other key on my keyboard...
I burst into laughter when I realized that I do have now Romanian characters: they were hidden in plain sight, usable with AltGr as modifier. I can't seem to spot them in keyboard.map, where according to the documentation all keysyms should be specified. Maybe setxbkmap did the magic on top? At least I am grateful I can type ฤรขศรฎศ and not have to copy paste the characters.
โ
โ