Tree-sitter grammar for comment tags like TODO:, FIXME(user):, etc. Useful to be embedded inside comments.

Check the playground at https://stsewd.dev/tree-sitter-comment/.

Since comment tags aren't a programming language or have a standard, I have chosen to follow popular conventions for the syntax.

• Comment tags can contain:
  • Upper case ascii letters
  • Numbers (can't start with one)
  • -, _ (they can't start or end with these characters)
• Optionally can have an user linked to the tag inside parentheses ()
• The name must be followed by : and a whitespace

• http and https links are recognized

If you think there are other popular conventions this syntax doesn't cover, feel free to open a issue.

TODO: something needs to be done
TODO(stsewd): something needs to be done by @stsewd
XXX: fix something else.
XXX:    extra white spaces.
(NOTE: this works too).
NOTE-BUG (stsewd): tags can be separated by `-`
NOTE_BUG: or by `_`.
This will be recognized as a URI
https://github.com/stsewd/

This grammar doesn't provide a specific token for it, but you can match it with this query:

("text" @todo
 (#eq? @todo "TODO"))

This grammar doesn't provide a specific token for it, but you can match it with this query:

("text" @issue
 (#match? @issue "^#[0-9]+$"))
;; NOTE: This matches `!10` and `! 10`.
("text" @symbol . "text" @issue
 (#eq? @symbol "!")
 (#match? @issue "^[0-9]+$"))

To avoid false positives, Neovim doesn't highlight all tags, but a list of specific ones, see the list at queries/comment/highlights.scm.

If you want your tag highlighted, you can extend the query locally, see :h treesitter-query. Or if you think it's very common, you can suggest it upstream.

Tree-sitter is a LR parser for context-free grammars, that means it works great for grammars that don't require backtracking, or to keep a state for whitespaces (like indentation). For these reasons, parsing languages that need to keep a state or falling back to a general token, it requires some manual parsing in C.

• nvim-treesitter
• helix
• zed-comment
• zed-todo-tree
• Yours?

• tree-sitter-rst: reStructuredText grammar.