Open kdy1 opened 2 years ago
Maybe I'll make own repo with non critical lints based on swc crate
I think time to create swc_linters
(i.e. swc_ecma_lint
/swc_css_lint
and maybe swc_html_lint
in future), so it will be more scalable and will not distract the main swc
repo and team (in future).
In long term I think we should replace eslint
/stylelint
, our parser and AST are great, perf is better, it will happen sooner or later anyway
I'll not remove rules from the repository, I'll just move it to @swc/lints
I'll review your PRs soon
Is swc cureently bundling its own wasm runtime? Does that mean after seperation user need to keep two identical wasm runtime on disk?
No we don't bundle wasm
Any plan to support code fix/code suggestion like eslint?
@IWANABETHATGUY lints will not have any formatting checks
https://github.com/swc-project/swc/pull/3888#issuecomment-1061515898
@IWANABETHATGUY lints will not have any formatting checks
Nope, code fix not equals to formatting checks.
@IWANABETHATGUY lints will not have any formatting checks
But thanks for explaining.
I mean something like this.
No we don't bundle wasm
Would #5350 result in final binary?
No, but we may replace native bindings for some platforms with wasm runtime.
Until stabilizing it, we will bundle wasm runtime (@swc/wasm
) for those platforms.
(Note: We bundle wasmer
, which is also a wasm runtime. So we may need to distinguish them)
Oh, if you mean wasmer
(for wasm runtime), the linter does not use wasm runtime at the moment.
But if we're going to make @swc/lint
a general purpose linter, plugin system is required. That would make it relies on wasmer too I guess
Yeah, but when we have enough resources for providing general plugins apis for linter, we may move it back to core Alternatively we can use a shared dylib, though
We cannot replace eslint because of plugins and the conclusion of the discussion was to drop it