fix: use non-deprecated overload of vim.str_utfindex

[MariaSolOs]

The current use of vim.str_utfindex was deprecated in neovim/neovim#30915. Now the encoding parameter is required and return value is an integer (instead of the table from the deprecated version corresponding to the utf-32 and utf-16 encodings).

[leiserfg]

IMHO it would be better to have local function that maps to the old str_utfindex if version is lower than 0.11 or a wrapper of the new one that gets the same args as the old one in >=0.11. Cause we try to keep backward compatibility.

[leiserfg]

Why does bytecol_to_utfcol returns now 3 elements instead of 2?

[MariaSolOs]

Why does bytecol_to_utfcol returns now 3 elements instead of 2?

It should not. Thanks for catching that.

[L3MON4D3]

Hi :)
It seems like our code was a bit misleading, the pre-0.11 str_utfindex returns utf32_index, utf16_index, not the other way around (see here).
Since we only really need the utf32-index in both of these cases, could you extract the added version-check logic into one function in util/util.lua, maybe str_utf32index, and call that in both situations?

I'm also fine with me doing this, just let me know :D

[MariaSolOs]

the pre-0.11 str_utfindex returns utf32_index, utf16_index, not the other way around (see here).

Oh yes, that has been the code I've been referring to but I clearly switched the order 🤦🏻‍♀️

Since we only really need the utf32-index in both of these cases, could you extract the added version-check logic into one function in util/util.lua, maybe str_utf32index, and call that in both situations?

I can do that, but the existing code does seem to use the utf16-index, that's what bytecol_to_utfcol returns. Am I missing something?
nvm, I realized that bytecol_to_utfcol is unused.

I'm also fine with me doing this, just let me know :D

I'm happy to do this, no worries :)

[L3MON4D3]

Alright, looks good :D
Thank You for the addition!

[MariaSolOs]

Alright, looks good :D Thank You for the addition!

Thanks to you for the review!