Unicode support in terminals is kind of a mess for TUI software developers. There is no good way to know how many columns a Unicode character will take, because terminals use different grapheme clustering algorithms.

For a text editor I'm making I ended up with:

> Unicode support is limited to what common modern terminal emulators support. In Hat, "character" is a Unicode codepoint. So no grapheme clustering, no variation selection, no ZWJs, no terminal emulator-specific logic. As long as every buffer byte is displayed and editable, we're ok.

https://github.com/ivanjermakov/hat/blob/master/CONTRIBUTING...

And I do basic codepoint range check to find its approximate width: https://github.com/ivanjermakov/hat/blob/00782dbaee1e1a814ce...