Merge pull request #260 from nojaf/dict-syntax

Add support for dict syntax

Author: nojaf

.github/workflows/ci.yml

@@ -29,6 +29,11 @@ jobs:
       - name: Checkout repository
         uses: actions/checkout@v4
 
+      - name: Set up Node.js
+        uses: actions/setup-node@v4
+        with:
+          node-version-file: '.nvmrc'
+
       - name: Set up tree-sitter
         uses: tree-sitter/setup-action/cli@v2
 

.nvmrc

@@ -1 +1 @@
-v23.7.0
+v24.11.1

AGENTS.md

@@ -0,0 +1,80 @@
+# AGENTS.md
+
+This file provides guidance to AI/LLM when working with code in this repository.
+
+## Build and Test Commands
+
+```bash
+# Generate parser from grammar (required after grammar.js changes)
+npx tree-sitter generate
+
+# Run all corpus tests
+npx tree-sitter test
+
+# Run tests for a specific grammar rule (filter by name)
+npx tree-sitter test --include "switch"
+
+# Parse a file to see syntax tree (useful for debugging)
+npm run parse path/to/file.res
+
+# Compare with ReScript compiler's AST (useful for debugging grammar issues)
+npx bsc -dparsetree -only-parse -ignore-parse-errors path/to/file.res
+
+# Launch interactive playground (builds WASM first)
+npm start
+
+# Install npm dependencies
+npm install
+```
+
+## Architecture
+
+### Grammar Definition
+
+**grammar.js** - The main grammar definition using Tree-sitter's JavaScript DSL. Key sections:
+- `externals` - Tokens handled by the custom scanner (newlines, comments, template strings, decorators, parentheses)
+- `precedences` - Operator precedence rules for expressions and module paths
+- `conflicts` - Shift/reduce conflict resolutions for ambiguous constructs
+- `rules` - All grammar rules (~80+ covering statements, expressions, types, modules, JSX)
+
+### External Scanner
+
+**src/scanner.c** - Custom C scanner for context-sensitive tokens that the grammar alone cannot handle:
+- Significant vs insignificant newlines (statement-ending vs formatting)
+- Nested multiline comments (`/* /* */ */`)
+- Template string interpolation (`\`hello ${name}\``)
+- Parenthesis nesting tracking (affects newline significance)
+- `list{` and `dict{` constructor detection
+- Decorator parsing (`@decorator` vs `@decorator(args)`)
+
+The scanner maintains state (`ScannerState`) tracking parenthesis nesting depth, whether inside quotes/backticks, and EOF reporting.
+
+### Generated Files (do not edit manually)
+
+- `src/parser.c` - Generated LR parser from grammar.json
+- `src/grammar.json` - Intermediate grammar representation
+- `src/node-types.json` - AST node type definitions
+
+### Query Files
+
+**queries/** - TreeSitter query files in S-expression syntax:
+- `highlights.scm` - Syntax highlighting rules mapping AST nodes to semantic scopes
+- `injections.scm` - Language injection boundaries
+- `locals.scm` - Variable scope definitions
+- `textobjects.scm` - Editor text object definitions
+
+### Test Corpus
+
+**test/corpus/*.txt** - Test cases in Tree-sitter's corpus format. Each test has a name, ReScript code, and expected parse tree. Files cover: comments, decorators, expressions, functions, JSX, let bindings, literals, modules, type declarations.
+
+## Development Workflow
+
+1. Modify `grammar.js` to add/fix grammar rules
+2. Run `tree-sitter generate` to regenerate the parser
+3. Add test cases in `test/corpus/*.txt` with expected parse trees
+4. Run `tree-sitter test` to verify
+5. For scanner changes, edit `src/scanner.c` and rebuild
+
+## Language Bindings
+
+Bindings are provided for Node.js, Python, Rust, Go, Swift, and C in the `bindings/` directory. CI tests Rust, Python, and Go bindings automatically.

CLAUDE.md

@@ -0,0 +1 @@
+AGENTS.md