Using egui and TreeSitter - a problem of byte- and character indices

[the-drunk-coder]

Hi,

I'm using egui for a text editor for my music coding language (which is LISP-like in syntax).

I wanted to make the editor more flexible, so instead of ad-hoc solutions for finding s-expressions etc, I created a TreeSitter grammar to navigate the code.

But, alas, TreeSitter gives you positions exclusively in bytes, while egui's text cursors all work with character indices ...

There's a helper function byte_index_from_char_index but not vice versa, also it seems inefficient to iterate (in the worst case) the entire text content ...

Is there a good way to bring the two together, or would there be any alternatives to TreeSitter?

Best,
N

[the-drunk-coder]

To answer my own question, using a text rope (i.e. the ropey crate) helps translating between the two.

I replaced the TextBuffer in my TextEdit widget implementation (I'm using a custom one) with ropey::Rope and it facilitates working with the byte positions provided by TreeSitter