@jsxtools/jsonc-reader

Tokenize JSONC and parse it into a compact span-based AST.

Install

npm install @jsxtools/jsonc-reader

Usage

import { JsonAstKind, TokenKind, parse, tokenize } from "@jsxtools/jsonc-reader";

const source = '{\n  // comment\n  "answer": [42, true, null],\n}';
const tokens = tokenize(source);
const ast = parse(tokens);

if (ast.t === JsonAstKind.Object) {
	console.log(ast.m[0]?.k.t === TokenKind.String);
}

What this package does

• Tokenizes JSONC source into tokens with offsets, line numbers, and columns.
• Parses those tokens into a compact AST that preserves source spans.
• Accepts JSONC comments by default.
• Accepts trailing commas in arrays and objects.

What this package does not do

• It does not materialize JavaScript values like JSON.parse.
• It does not preserve comments in the AST; comments remain tokens.
• It does not allow single-quoted strings.

API

tokenize(input, options?)

Returns an array of tokens plus a non-enumerable text property containing the original source text.

• input: any value, coerced with String(input)
• options.allowComments: defaults to true

parse(tokens)

Consumes a token collection and returns a span-based AST.

Primitive nodes are represented by token kind plus start/end offsets. Object and array nodes expose members via m.

JsoncSyntaxError

Syntax errors include:

• reason
• s start offset
• line
• column

Development

• npm run build
• npm run test
• npm run lint
• npm run format:check