At present, I decided not to support partially JSON expression like below in #27.

Partially JSON expressions such as [notValid, {isValid:true}] ensuring {isValid:true} is transformed.

const data = { bar: invalid_object, foo: 'foo' }
↓
const data = { bar: invalid_object, JSON.parse('{"foo": "foo"}')}

Does anyone need this feature?
If you need this feature, please comment on this issue with your detail 🙇.

Currently, this plugin cannot correctly convert number-keyed properties.

before

const obj = {
  123: "123",
  45: 45
};

console.log(obj[123]); // => "123"

after

const obj = JSON.parse('{"undefined":45}');
console.log(obj[123]); // => undefined

Now, type annotation in src/utils.ts and src/plugin.ts is terrible...

So, we should improve this type annotation.

I stumbled across a JSON.parse suddenly throwing and it seems that some \n are not properly escaped.

Reading the source code, it seems that because the RegExp is not using the global flag, it will only escape the 1st occurence and leaving the following ones as is.

  if (/[\t\f\r\n\b]/g.test(value)) {
    const codes = ['\f', '\r', '\n', '\t', '\b']
    const replaceCodes = ['\\f', '\\r', '\\n', '\\t', '\\b']
    for (let i = 0; i < codes.length; i++) {
      value = value.replace(new RegExp(codes[i]), replaceCodes[i])
    }
  }

For example;

// Input
const foo = ['Test\nthis\nstuff];

// Output
const foo = JSON.parse('["Test\\nthis\nstuff"]');

Using new RegExp(codes[i], 'g') should probably solve it.

issue code

const codes = ['\t', '\f', '\r', '\n', '\t', '\b']

https://github.com/nd-02110114/babel-plugin-object-to-json-parse/blob/master/src/utils.ts#L79

@nd-02110114
Could u clarify my doubt, I think it possible trigger double encode.

This plugin is super noisy, complaining about the vast majority of non-serializable object literals in my apps.

I think running into a non serializable object literal should not be considered an error, and that the message should be removed.

I'm not sure this is a bug, but this behavior is not an ideal one, I think.

This plugin generates a source code that old JavaScript engine cannot parse when object includes U+2028 or U+2029 in a string. cf. https://developer.mozilla.org/en-US/docs/Web/JavaScript/Reference/Global_Objects/JSON#JavaScript_and_JSON_differences

before

const obj = {
  foo: "\u2028\u2029"
};

console.log(obj.foo);

after

const obj = JSON.parse('{"foo":"

"}');
console.log(obj.foo);

When the code after conversion is run in node.js v8:

(function (exports, require, module, __filename, __dirname) { const obj = JSON.parse('{"foo":"
                                                                                     ^^^^^^^^^

SyntaxError: Invalid or unexpected token
    at createScript (vm.js:80:10)
    at Object.runInThisContext (vm.js:139:10)
    at Module._compile (module.js:617:28)
    at Object.Module._extensions..js (module.js:664:10)
    at Module.load (module.js:566:32)
    at tryModuleLoad (module.js:506:12)
    at Function.Module._load (module.js:498:3)
    at Function.Module.runMain (module.js:694:10)
    at startup (bootstrap_node.js:204:16)
    at bootstrap_node.js:625:3

According to MDN, the latest Edge seems to fail too.

Great work!
This repository includes .prettierrc.js but it seems to be unused.
Indeed, contents of this repository is not formatted by prettier.

My editor detects existence of .prettierrc.js and automatically formats the code I'm editing, which made me struggle while working on #1.
Please either enable prettier in this repository, or remove .prettierrc.js.

In addition, automatic linting on commit would be very nice.

Thank you!

This plugin fails to convert following object literal:

const obj = {
  foo: "'abc'"
};

console.log(obj.foo);

result:

At 1 line (start) : The object wasn't converted (Unexpected token, expected "," (1:22) - make sure this is an expression.
> 1 | (JSON.parse('{"foo":"'abc'"}'))
    |