key.go does not respect current codepage leading to utf8.RuneError

[Vinschni]

Bubbletea does not respect the current codepage while parsing the console input from a user: https://github.com/charmbracelet/bubbletea/blob/624343388b02eb2f5199ee8fcd29961ba3b17581/key.go#L515

It assumes the input encoding is utf8. I'm using Windows Terminal. And chcp

PS C:\Windows\system32> chcp
Active Codepage: 850.

shows default encoding is Codepage 850 == DOS-Latin-1. If I'm entering characters (in a textarea widget) on my german keyboard like ß, ö, ä, ü this leads to utf8.RuneError.

I could fix key.go just for my system by using:

var decoder = charmap.CodePage850.NewDecoder()
utf8bytes, err := decoder.Bytes(b[i:])
r, width := utf8.DecodeRune(utf8bytes)

instead of

r, width := utf8.DecodeRune(b[i:])

but this is no good nor general solution. I assume a good solution would include using something like GetConsoleCP()

Futhermore pressing buttons on my keyboard like POS1 / END inserts a bracket to the textarea which suggests parsing of control characters does not work.

Is this a known problem?

[muesli]

Futhermore pressing buttons on my keyboard like POS1 / END inserts a bracket to the textarea which suggests parsing of control characters does not work.

Is this a known problem?

This should be fixed by 2fe4aad5e60ed5ea53ee1fbe546db429da748832.

[muesli]

r, width := utf8.DecodeRune(b[i:])

Shouldn't this have been converted to utf8 by the localereader above?

Probably related: #153 and #343.