Add new use-ember-data-rfc-395-imports rule

[dcyriller]

Description

This PR introduces a new rule to ease ember-data migration to new packages: use-ember-data-rfc-395-imports.

Examples of incorrect code for this rule:

import Model from "ember-data/model";
import attr from "ember-data/attr";

import DS from "ember-data";
const { Model } = "ember-data";

Examples of correct code for this rule:

import Model, { attr } from "@ember-data/model";

import Model from "@ember-data/model";

Note

Split into 5 commits that may be reviewed one after the other.

It leverages what have been done in the past with Ember globals and ember-cli-shims.

A few discussion are still in progress around https://github.com/ember-data/ember-data-rfc395-data/ (the data set). Hence the WIP!

Links

ember-data RFC
a related codemod

EDIT: I've updated this description to reflect the changes made following ember-data/ember-data-rfc395-data#7

[dcyriller]

cc @runspired @rwjblue

[dcyriller]

This is ready now.

[snewcomer]

👍 Nice work!

[runspired]

I believe there should be a new entry for this rule in README.md and in recommended-rules.js (with a value of "off" for now). To achieve this use npm run update

[rwjblue]

@bmish - Mind doing one more pass on this one before we land / release?

[runspired]

I dislike being pedantic especially when I feel the name has been changed a few times, but is no-ember-data-global-imports really the best name for a rule that:

• lints against older module import locations
• AND lints against an older "single-export-object" style import
• AND lints against global syntax

There was a suggestion to name the rule after the rfc, I think it's probably the way to go.

use-ember-data-rfc-395-imports

[dcyriller]

There was a suggestion to name the rule after the rfc, I think it's probably the way to go. use-ember-data-rfc-395-imports

@rwjblue @bmish if you agree, I'll change that. Let me know! (I want to be sure everyone agrees before changing it.)

[rwjblue]

Yeah, that seems fine to me. My only real objection was the "new" in the original name, I'm sorry for the run around / bike shedding here.

[dcyriller]

No worries, I prefer to make it right!

[bmish]

@dcyriller don't worry, I'm not going to require you to refactor existing code if you're just moving it around. I commented about the excessive length and complexity of some of the functions because it makes the review difficult, and it could be a good time to clean this up. But we can merge the PR soon if you're done with your changes.

[dcyriller]

@bmish should be good now.

[dcyriller]

I've made fun splitting utils functions into smaller ones (I really did 😀). I'm not 100% convinced it makes the code better though.

[bendemboski]

@dcyriller @bmish @rwjblue this caused a crashing regression. Let me know if there's any way I can help get this resolved as it's blocking me updating to the latest eslint-plugin-ember (and causing a number of my addon builds with floating dependencies to fail).