Unknown CSI inputs result in garbage key events

[knz]

When a terminal CSI sequence is not recognized by bubbletea, it is not completely ignored; instead, bubbletea will match the initial escape byte and the character immediately afterwards as a single key with the alt modifier, then interpret the rest as regular key inputs. For example, ESC [ 1003 h gets parsed as alt+1, then 0, then 0, then 3, then h.

Since the CSI sequences all have a regular structure, they could be parsed and ignored as a single unit.

(The definition is strict: a CSI sequence starts with ESC [ and terminates with a byte value in the range 0x40 to 0x7E.)

[muesli]

Could you elaborate how such a sequence would be triggered?

[knz]

Could you elaborate how such a sequence would be triggered?

For example until #396 gets merged, if you press Shift+Home in xterm you get:

• 1 key message for Alt+1
• 1 separate message for a single ; character
• 1 separate message for 2
• 1 final message for H.