Donny/강동윤
Donny/강동윤
No. I don't have any idea as I don't know what `Deno.emit` does.
Using `never` type requires a good enough type checker We can use a 'known list' of expression with never type alternatively
@bartlomieju Of course. The rules below can be merged using a simple trait. - no-undef - no-import-assign - no-ex-assign (uses `scopes.rs`) - no-func-assign (uses `scopes.rs`) - no-class-assign (uses `scopes.rs`) -...
I tried working on this, but I found that the `no-const-assign` rule already exists. See https://github.com/denoland/deno_lint/blob/07e0ff827e85efb06661844599fbddc89b394c53/src/rules/no_const_assign.rs
I think it can be implemented with match statement in `visit_member_expr` if you mean lexical analysis. If you mean non-lexical analysis, full control flow graph like typechecker is required.
I think the intent of the user is important. I mean, if ecmascript syntax forces the user to place a node on the specific position, like the return type in...
It's okay to exit with failure on parsing error in `deno` because the parser does not do any linting - all errors are critical issue for compiler and runtime. I...
You are correct. I recently added some documents for the resolver at https://swc.rs/rustdoc/swc_ecma_transforms/resolver/fn.resolver_with_mark.html
Currently, it's not supported.
[swc](http://swc.rs/) is a faster alternative for babel