R support

[jimhester]

R is a widely used, growing language often used in Data Science and Statistics.

While it does not have a published formal specification, there is a draft specification that describes lexing and parsing the language.

In the most widely used implementation the parsing is done with a bison parser defined in gram.y.

The lexing rules for R are somewhat complex, but the parsing is relatively straightforward, as generally everything is an expression.

It would very beneficial to the R community to have support for R in semantic!

[XVilka]

A first step would be creating a tree sitter grammar for R. See more in their documentation on creating new parsers.
After that, the hardest part begins - implementing the semantic part: https://github.com/github/semantic/blob/master/docs/adding-new-languages.md

[tclem]

Definitely start with a tree-sitter parser.

After that, the hardest part begins - implementing the semantic part

This is about to get much, much easier as we are almost entirely automating the semantic part. My advise would be to get the tree-sitter grammar in good shape and then check in again with the team so see if we can use the new path for supporting languages in semantic.

[2yz]

Definitely start with a tree-sitter parser.

After that, the hardest part begins - implementing the semantic part

This is about to get much, much easier as we are almost entirely automating the semantic part. My advise would be to get the tree-sitter grammar in good shape and then check in again with the team so see if we can use the new path for supporting languages in semantic.

Is the new path for supporting languages available soon?

[tclem]

Is the new path for supporting languages available soon?

I don't have a specific timeline to give you, but you can see an example for java and python of what's required to generate code from the new node-types.json. Obviously part of the work here is to better surface our documentation.