Inserting a two byte UTF-8 character moves the cursor two characters

[MoritzKn]

Inserting a two byte UTF-8 character like ä moves the cursor two characters.

Given a buffer like this with the pipe being the cursor:

|1234

Inserting ä will result in the cursor being placed like this:

ä1|234

I've tried to fix it myself but couldn't find the right place in the code. With a few hints I may get further... how is the cursor represented in code? where are insertions handled?

[nathansobo]

Interesting. A decent place to start might be drawCursors in text_plane.js. See if we're relaying the correct column but measuring it wrong or if the column is off. We currently represent text as UTF-16 internally (something we should eventually fix) so it's a bit surprising that a 2-byte character is causing issues.

[MoritzKn]

See if we're relaying the correct column but measuring it wrong or if the column is off.

The next character is inserted in the position the cursor is displayed. So I assume it has to do with the internal logic rather than the rendering.

We currently represent text as UTF-16 internally

Maybe the problem is somewhere else then. Perhaps when converting the input event into a change to the buffer.

And thats where I think I located the issue. The edit() function in buffer_view.rs still uses a rust string slice (&str) which is AFSIK utf8. In the function text.len() is used to get the length of the string. len() of a string however returns the the byte length. I'm going to try and fix this.