Artem

`ts_node_child_containing_descendant()` is in tree-sitter's master, but not yet available in any release.

The ancestor checking part of the predicate can be moved into C, which would reduce the time down to 1.5ms from 7.5ms (measured on a different computer). ```lua ['has-ancestor?'] =...

`:help` for 'help' contains this paragraph: ```vimdoc {subject} can include wildcards such as "*", "?" and "[a-z]": :help z? jump to help for any "z" command :help z. jump to...

I think this might work for all characters, although I couldn't test it for non-ascii: ```lua local lang = 'en' local subject= '[==]' local code_points = vim.fn.str2list(subject) local escaped_str =...

> couldn't test it for non-ascii I tested it and it works for multibyte characters. I also wrote a test to see which other help tags are affected: Results (for...

The second commit removes the duplicate checking. These checks prevent scenarios where the user wants to open help for one tag, but the help is opened for a duplicate tag...

> is there any reason not to squash these commits before merging? Either way is fine for me.

I added `child_containing_descendant()` calls to the end of `test_parent_of_zero_width_node()` in my pr (#3214), so the original 'test.patch' no longer has all the necessary changes to pass on c19f23f. https://github.com/tree-sitter/tree-sitter/blob/c19f23f42046d68801c6a99fca34308e4985337e/cli/src/tests/node_test.rs#L287-L291 Updated...

I updated the logic to remove '\r' from the end of lines (on all platforms). Although I checked if this was an issue before and it seems like it wasn't,...

[`buf_get_text()`](https://github.com/neovim/neovim/blob/c7958356bef304320d86cd541d0de8db968c6cc8/src/nvim/api/private/helpers.c#L517) doesn't clamp `start_col` after subtracting it from the line length.