CLDR-16604 kbd: complete scancode data

[srl295]

• Also adds 'ks' as per CLDR-16873

CLDR-16604

• This PR completes the ticket.

ALLOW_MANY_COMMITS=true

[miloush]

Note that if you map keys from the left, iso=abnt2=jis and us=ks, with the only difference being 56 next to left shift. So currently we have two physical arrangements, either with or without that key.

[srl295]

Note that if you map keys from the left, iso=abnt2=jis and us=ks, with the only difference being 56 next to left shift. So currently we have two physical arrangements, either with or without that key.

• iso=abnt2=jis - jis has the extra 7d on the top row, iso doesn't have 73 on the fourth row
• us = ks: ks only has the extra 7d vs what us has.

[miloush]

@srl295 what I mean is that they behave identical when the keys are not used, so technically there doesn't need to be that many forms defined. For example, you could have a form with 14 scancodes defined in the first row (13 on second, 12 on third and fourth), and <row> for the first row will have either 13 keys or 14 keys, it doesn't matter. (And to push it further, we could define a blank key and then have everyone use one prescribed form that includes 56 and not worry about forms, and/or make them think what a keyboard that has the key should do.)

It is important to have overridable/custom forms though, so nothing needs to change, it is just an observation. There is no harm in specifying them all individually explicitly.

[srl295]

Jan, that makes sense. It could be worth noting, since keys are optional on the right edges of any row.