Tree-sitter writeup

Learn something about the tree-sitter.

Development workflow

1. Create a new parser project mkdir tree-sitter-{language}.

2. Initialize the project with dependencies.

     cd tree-sitter-{language}
     pnpm init
     pnpm add nan node-gyp
     pnpm add tree-sitter-cli -D

3. Initialize global config for tree-sitter, add Paths.

   ./node_modules/.bin/tree-sitter init-config
        ^ this would print out the path to config file
   
   # add the parent folder of `tree-sitter-{language}` to the field "parser-directories"
   # so that tree-sitter can find the parser project

4. Write some grammar in grammar.js. Here is an example.

5. Call ./node_modules/.bin/tree-sitter generate path-to-grammar.js to generate several new files important.

6. Run node-gyp configure && node-gyp build to build the final parser.

Test

1. Add test files in the directory corpus. Check out the example.

2. Run ./node_modules/.bin/tree-sitter test to verify. Add --update option to update the syntax.

Highlight

1. Add highlights.scm in the directory queries. Check out the example.

2. Run ./node_modules/.bin/tree-sitter highlight --html ./examples/source.imp to generate a result in html.

3. Open that html to view the grammar highlight output.

Acknowledgement

Watch the video: Tree-sitter: a new parsing system for programming tools - GitHub Universe 2017

Read some blogs:

• Guide to your first Tree-sitter grammar: A detailed post on writing a parser for tree-sitter.

• Lightweight linting with tree-sitter: A post on using tree-sitter to write a linter. Manipulate the S-expression is interesting but difficult.

  Here is another lint tools with JavaScript bindings.

• How to write a tree-sitter grammar in an afternoon: A good article on how to write a parser in action.