tweak key reporting

[willmcgugan]

Some tweaks to how keys are reported (with the Kitty protocol).

• If "shift" is held with another modifier it is reported as "shift", rather than a capital letter. Previously if you press shift + ctrl + e, it is reported as "ctrl+E". Now it is reported as "ctrl+shft+e".

I think this makes sense, as such a combination is unlikely intended to be a printable key.

• Removed caps_lock and num_lock modifiers. This resulted in shift + key, and caps_lock + key being considered as different keys -- which I don't think makes sense. If this is ever a requirement, I think we should add booleans to the Key event.

[darrenburns]

Will this break existing apps with bindings for ctrl+E for example? Will those apps need to update their Bindings to ctrl+shift+e?

[willmcgugan]

It would. I've added a note in the changelog.

[darrenburns]

Just had a thought about this - isn't it a bit strange to have this reported like this but not have "E" be reported as "shift+e"? I'm not sure which is better, but I'd definitely expect it to be consistent.

[willmcgugan]

Just had a thought about this - isn't it a bit strange to have this reported like this but not have "E" be reported as "shift+e"? I'm not sure which is better, but I'd definitely expect it to be consistent.

There isn't a simple one to one mapping of shift + letter. It depends on your keyboard and locale. e.g. shift+3 will give a £ in GB layout or a # in US layout. I think that's also true of letter keys with some international layouts. There may also be multiple ways of producing a key. For instance, if capslock is on, you can produce capital letters without holding shift.

I think its safer to trust that when the keyboard produces a key, that's what the user intended, and we don't want to make assumptions about how it was produced.

[willmcgugan]

Actually, I think I just talked myself out of part of this update. When the keyboard reports an alphanumeric letter, Kitty does not report that the shift modifier was pressed.

It does report the shift modifier for non alphanumeric keys. i.e. shift+space reports that the shift modifier was held.

So I don't think Textual should be trying to deduce that shift was pressed at all. So ctrl+shift+E should be reported as ctrl+E.

[darrenburns]

Does this need updating to match the latest state of the PR?