feat(linter): eslint-plugin-next/no-duplicate-head

[IWANABETHATGUY]

implement @next/next/no-typos, part of #1929

Ref

Code
https://github.com/vercel/next.js/blob/canary/packages/eslint-plugin-next/src/rules/no-duplicate-head.ts

[codspeed-hq]

CodSpeed Performance Report

Merging #2438 will not alter performance

_{Comparing IWANABETHATGUY:feat/eslint-plugin-next/no-duplicate-head (02e2f13) with main (a78303d)}

Summary

✅ 27 untouched benchmarks

[Boshen]

Need a error message here.

[Dunqing]

Here we can use AstKind:ImportDefaultSpecifier instead, the parent kind of the AstKind:ImportDefaultSpecifier is AstKind::ImportDeclaration

[Dunqing]

I think we can have a simpler way to find the class, for example via ctx.semantic.classes

[IWANABETHATGUY]

Not sure if this would be faster, you still need to figure out if the class is the ancestor of current ReturnStatement

[Boshen]

Wait ... can't we just find all the read references of Head? We don't need to go over the AST because we have semantic data.

[Dunqing]

Wait ... can't we just find all the read references of Head? We don't need to go over the AST because we have semantic data.

Yes! Sounds reasonable.

[IWANABETHATGUY]

Wait ... can't we just find all the read references of Head? We don't need to go over the AST because we have semantic data.

Good idea, however, you still need to extract the documentImport from ImportDefaultSpecifier, and the ImportDeclaration could be everywhere in a js file. e.g. :

console.log("what ever")
test
Document
import Document, {test} from './test'

So iterating the nodes could not be skipped

[Boshen]

Wait ... can't we just find all the read references of Head? We don't need to go over the AST because we have semantic data.

Good idea, however, you still need to extract the documentImport from ImportDefaultSpecifier, and the ImportDeclaration could be everywhere in a js file. e.g. :

console.log("what ever")
test
Document
import Document, {test} from './test'

So iterating the nodes could not be skipped

Yeah, the import part is still needed.

[Dunqing]

Wait ... can't we just find all the read references of Head? We don't need to go over the AST because we have semantic data.

Good idea, however, you still need to extract the documentImport from ImportDefaultSpecifier, and the ImportDeclaration could be everywhere in a js file. e.g. :

console.log("what ever")
test
Document
import Document, {test} from './test'

So iterating the nodes could not be skipped

You can find the documentImport by symbols. The Document symbol flag will contain SymbolFlags:ImportBinding

https://oxc-project.github.io/oxc/playground/?code=3YCAAIDMgICAgICAgICxG0qZRWtpV52IeeJbT1tdEcK4UeIwEHHwiWefH4zZxP7cwptFUwhrdc4TI795I028%2FXPY7FZcAXXzk10cqHp4LqCsHLj%2FfQlBwA%3D%3D

[Boshen]

Wait ... can't we just find all the read references of Head? We don't need to go over the AST because we have semantic data.

Good idea, however, you still need to extract the documentImport from ImportDefaultSpecifier, and the ImportDeclaration could be everywhere in a js file. e.g. :

console.log("what ever")
test
Document
import Document, {test} from './test'

So iterating the nodes could not be skipped

You can find the documentImport by symbols. The Document symbol flag will contain SymbolFlags:ImportBinding

https://oxc-project.github.io/oxc/playground/?code=3YCAAIDMgICAgICAgICxG0qZRWtpV52IeeJbT1tdEcK4UeIwEHHwiWefH4zZxP7cwptFUwhrdc4TI795I028%2FXPY7FZcAXXzk10cqHp4LqCsHLj%2FfQlBwA%3D%3D

But you don't know if it's imported from next or renamed :-(