fix(jest-types): make ConfigGlobal an interface

[G-Rath]

As discussed here.

Summary

Makes ConfigGlobal an interface to allow for declaration merging:

// ts-jest/index.d.ts
declare module '@jest/types' {
  namespace Config {
    interface ConfigGlobals {
      'ts-jest': TsJestGlobalOptions;
    }
  }
}

// jest.config.js
/** @type {import('@jest/types').Config.InitialOptions} */
const config= {
  globals: {
    'ts-jest': {
      isolatedModules: true,
      myProperty: 1 // error, "myProperty" is not defined!
    }
  }
}

module.exports = config;

Test plan

None 😬

[SimenB]

Suggested change

	[K: string]: any;
	[K: string]: unknown;

?

[G-Rath]

That'll be the ideal; let's commit and see if anything breaks :D

[SimenB]

I also wonder if we should call it global instead of K?

[G-Rath]

K is typically the convention for this pattern, as it's a generic that represents the property name.

i.e

/**
 * Construct a type with a set of properties K of type T
 */
type Record<K extends keyof any, T> = {
    [P in K]: T;
};

I'm not fussed, but the actual name won't show up much anyway.

[SimenB]

It's not really generic though, it's a string. but it doesn't matter like you say 🙂

---

seems this introduced a type error in normalize

[G-Rath]

Ah so the proper type should be object, due to:

const mergeGlobalsWithPreset = (
  options: Config.InitialOptions,
  preset: Config.InitialOptions,
) => {
  if (options['globals'] && preset['globals']) {
    for (const p in preset['globals']) {
      options['globals'][p] = {
        ...preset['globals'][p],
        ...options['globals'][p],
      };
    }
  }
};

[G-Rath]

@SimenB so we've got a couple of options, depending on how you're feeling:

• we stick with any,
• we adjust mergeGlobalsWithPreset to do a typecheck to ensure options['globals'][p] is an object
• we refactor this away: https://github.com/facebook/jest/blob/f47ade4edb112261e14be459185e2396d9c8cbfc/packages/jest-config/src/ValidConfig.ts#L54
• we use a cast or ignore

Looking at the docs, I think any maybe the "correct" type?

A set of global variables that need to be available in all test environments.

For example, the following would create a global __DEV__ variable set to true in all test environments:

Poking around a bit in normalize, I can't actually find what stops jest from crashing with an error about ... ?

Today I learned that you can actually spread a boolean without erroring O.o

[SimenB]

To me it looks like the code is broken, and mergeGlobalsWithPreset will discard e.g. the __DEV__: true example from the docs? It should use a proper recursive merge, not blindly spread into one another

[G-Rath]

Agreed. I'll adjust the implementation.

[G-Rath]

Ah, finally made it give bad output:

[SimenB]

I think unknown is correct, we just need to properly merge the config provided rather than blindly spreading

[G-Rath]

Yeah for sure - I've got a test and everything; just need to come up with a merging solution.

One of the jest packages already has one coded but not exported - I could move that into jest-utils (iirc that's a package that exists?) a shared package.

Otherwise deepMerge is used by a few packages, so I'll probably just stick that there 🤷‍♂

[codecov-io]

Codecov Report

Merging #9570 into master will decrease coverage by <.01%.
The diff coverage is 100%.

@@            Coverage Diff             @@
##           master    #9570      +/-   ##
==========================================
- Coverage   65.09%   65.09%   -0.01%     
==========================================
  Files         287      287              
  Lines       12145    12144       -1     
  Branches     3007     3009       +2     
==========================================
- Hits         7906     7905       -1     
  Misses       3604     3604              
  Partials      635      635

Impacted Files	Coverage Δ
packages/jest-config/src/normalize.ts	76.66% <100%> (-0.07%)	⬇️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4f4513b...75407fe. Read the comment docs.

[SimenB]

ended up being a bug fix, nice 😀 any sucks

[G-Rath]

any sucks

It does indeed :/

[github-actions]

This pull request has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.
Please note this issue tracker is not a help forum. We recommend using StackOverflow or our discord channel for questions.