tree-sitter-r

An R grammar for tree-sitter.

R package

This grammar is available as an R package.

You'll also want the R package providing bindings to tree-sitter itself.

Rust bindings

This grammar is available as a Rust crate on crates.io.

References

• The R Draft Spec
• gram.y

Known deviations

This section describes known deviations from the R grammar.

]] as a literal token

The following is valid R syntax, note how ]] has been split over multiple lines.

x[["a"]
]

This applies to ]], but not to [[, for example, this is not valid R syntax:

x[
["a"]]

The technical reason for this is that in the grammar R treats [[ as a single token, but ]] is treated as two individual ] tokens. Treating ]] as two individual ] tokens allows whitespace, newlines, and even comments to appear between the two ] tokens:

x[["a"] # comment
]

While we'd like to precisely support the R grammar, it is also extremely useful to treat all of (, ), [, ], [[, and ]] as literal tokens when using the tree-sitter grammar. This allows you to treat call, subset, and subset2 nodes in the same way, since they all have exactly the same node structure.

Because treating ]] as a literal token is so useful, and because we've never seen any R code "in the wild" written this way, this grammar does not allow whitespace, newlines, or comments between the two ] tokens.