Add observeReadyElements method

[Richienb]

Depends on sindresorhus/type-fest#246

---

I was unable to use any-observable for these reasons:

• If I import it at the top level, the user will be forced to install an observable library even if they don't need it.
• If I import it within the function, I will have to use the import function which will force me to make the observable-returning function a observable-resolving-promise-returning function which would not be user-friendly and a bit annoying to implement.
• If I try to use import at the top-level in a try/catch block, I will need to require Node.js 14.8. ESLint also has trouble parsing this and making it work is quite inconvenient. Enable Top Level Await eslint/eslint#13178
• any-observable doesn't use ESM and the interop layer Node.js uses doesn't seem to allow import 'any-observable/register/zen'; to actually register the observer. ESM any-observable#27

---

When running the unit tests locally, a test for a different part of code that this PR does not change failed:

element-ready/test.js

Line 212 in 32360a0

t.is(promiseStateSync(entireCheck), 'pending', expectation);

Difference:

- 'fulfilled'
+ 'pending'

This error may or may not reappear during runs of the GitHub Action.

---

Fixes #22

---

IssueHunt Summary

Referenced issues

This pull request has been submitted to:

• #22: Subscribe to new elements on the page

---

[sindresorhus]

Wouldn't it be better to use something more platform native and general like async iterable?

[Richienb]

Wouldn't it be better to use something more platform native and general like async iterable?

Yes it would - changed the code.

[Richienb]

// @sindresorhus

[sindresorhus]

Instead of using .stop(), couldn't you just use break to stop the for await loop?

[Richienb]

Instead of using .stop(), couldn't you just use break to stop the for await loop?

Indeed. Looks like as well as a next() method, I can also specify a return() method.

[sindresorhus]

Would be nice to have a test for break usage, and also show usage of break the docs for how to stop the loop, since it might not be immediately obvious to the user.

[Richienb]

Would be nice to have a test for break usage

As far as I can tell, there is no way to test if break; actually propogates to the corresponding internal logic but I'll still add it anyways in case it does and throws an error.

[sindresorhus]

As far as I can tell, there is no way to test if break; actually propogates to the corresponding internal logic but I'll still add it anyways in case it does and throws an error.

A test would at least confirm that it correctly stops and resolves with the correct items.

[Richienb]

CI is failing because of sindresorhus/p-state#6

[sindresorhus]

Regarding #36 (comment). I disagree with your reasoning. The performance difference in real-world scenarios is negligible. But I also don't feel strongly about it and this PR has had enough back-and-forth, so I'm going to leave it.

[sindresorhus]

Thanks for contributing this :)