`no-unresolved` is not aware of `exports` definition in `package.json`

[ignatiusreza]

From this announcement

Package entry points and the package.json “exports” field

There are now two fields that can define entry points for a package: “main” and“exports”. The “main” field is supported in all versions of Node.js, but its capabilities are limited: it only defines the main entry point of the package. A new package.json field “exports” can also define the main entry point, along with subpaths. It also provides encapsulation, where only the paths explicitly defined in “exports” are available for importing. “exports” applies to both CommonJS and ES module packages, whether used via require or import.

Might need to be fix in resolve package..

[conartist6]

Could you clarify: why do really old node versions need to be able to resolve exports though? Or is there some other use case I'm missing?

[ljharb]

The use cases for resolve are far more plentiful than for eslint. resolve needs to resolve, in any node version, a path as if it would be resolved in any other node version. A pre-exports node may want to emulate “exports”. A post-exports node may want to emulate pre-exports. A modern node may simply want to know what a pre-exports node user would see.

[piranna]

I think the sane solution here is to don't worry at all about any unmaintained Node.js version, it's said, don't provide support for anything older than Node.js v18.

[conartist6]

I really don't buy that the only way to have good compatibility is if there is a single implementation of the algorithm. Tons of projects are already using enhanced-resolve anyway, and putting a new implementation out in ESLint will be an amazing way to safely get feedback on whether there is behavioral parity.

[ljharb]

@piranna ”sane” isn’t an appropriate word to be using in this context; and almost all of my packages support down to node 0.4, and will continue to do so for the foreseeable future. Platform support has no correlation to what a package should support.

[ljharb]

@conartist6 thats fine, your agreement isn’t required.

[conartist6]

I think this is great because now we're unblocked. It's clear from this conversation that my beliefs about the problem are different enough from yours that I should build a new package, and that it will be clearly differentiated in the marketplace.

[conartist6]

Everyone who just wants to use the new stuff can stop bugging you, which is a win for you as well.

[ljharb]

Go for it. As you said tho, there’s already packages that do this, so i don’t think there’s much value in yet another one; the value is in resolve itself doing it.

[conartist6]

I would agree that all that was off-topic if this were a conversation on the resolve package, but as this is eslint-plugin-import parts of that conversation seem quite relevant.

To recap the now-hidden conversation:

• @ljharb is the sole maintainer of the resolve package
• @ljharb's mission for the resolve package is to support node >= 0.4.0, meaning he can only work in es3
• ESLint has required node >= 4.0.0 for the last six years (at least), and can use es6
• resolve favors speed and compatibility over readability and correctness

I believe that in sum this means that eslint-plugin-import should seek out a resolve implementation which is better aligned with its needs and the needs of its users.

[ljharb]

It won’t; using resolve - the most common implementation - is what is most aligned with it and its users needs.

Again, your agreement isn’t required, and to be clear, is also irrelevant.

[conartist6]

I would appreciate if you would allow my condensed argument to stand as a matter of public record. Please?

[ljharb]

Your summary would be fine; your conclusion would only distract.

[conartist6]

The comment may be deleted now, but I'm still calling it out. That was rude and inappropriate. We can compete with and disagree with other people and still respect them.

[ljharb]

@conartist6 no comments were deleted afaict, so I’m not sure what you’re referring to.

[conartist6]

Not deleted by you then, but it was emailed to anyone who subscribes to this thread.

[ljharb]

ah, yes, i found it in my email :-) some people have nothing better to do, i guess.

[conartist6]

I feel that you have abused your position as a moderator to protect your image in the face of reasoned dissent, so I can't say I'm that thrilled with you at the moment.

[conartist6]

Is there an ombudsman or another moderator who could assist in untangling this?

[conartist6]

I'm not wedded to any particular outcome, but it's unacceptable to me that all discussion on the topic should be summarily silenced.

[ljharb]

It’s just me, and what’s acceptable is only up to collaborators. I’m going to lock the thread for a bit.

Additionally, hiding comments isn’t silencing anything - that’d be called “deleting comments”. I’m just making them hidden by default so they don’t distract new readers of the issue.