@Dunqing please read the following Contributor License Agreement(CLA). If you agree with the CLA, please reply with the following information.

@microsoft-github-policy-service agree [company="{your company}"]

Options:

• (default - no company specified) I have sole ownership of intellectual property rights to my Submissions and I am not making Submissions in the course of work for my employer.

@microsoft-github-policy-service agree

• (when company given) I am making Submissions in the course of work for my employer (or my employer has intellectual property rights in my Submissions by contract or applicable law). I have permission from my employer to make Submissions and enter into this Agreement on behalf of my employer. By signing below, the defined term “You” includes me and my employer.

@microsoft-github-policy-service agree company="Microsoft"

Contributor License Agreement

@microsoft-github-policy-service agree

And here I always assumed declaration files worked exactly like source files w.r.t. what’s exported - before today I would have confidently told anyone who asked that A there was internal to the file and not importable, just like it would be if that were a .ts file

Declaration files have really, really weird export semantics. Been trying to augment ts-eslint to handle that in typescript-eslint/typescript-eslint#8611 but it's definitely a battle, so not surprised to see this be wrong.

It hasn't really needed to be something we've needed to think about since our declaration emitter started emitting privates and export {} automatically to work around the issue; it's weird, but it does "just work", back to ancient versions of TS.

This is weird. I am implementing the Isolated Declarations transformation in oxc. When I came across this empty export I thought it was a bug. it doesn't always occur, only non-exports referenced in the export statement generate export {}. so I tried searching in the TypeScript issues to confirm my suspicion.

So I want to ask, when a third-party tool wants to implement Isolated Declarations, does it also have to port this behavior?

Yes. If you're implementing isolated declarations, you should be using transpileDeclaration as your guide to correct output (on passing inputs).

For future implementors of isolated declarations, you need port this "bug" as a feature 😁

We are already using the transpileDeclaration API to guide our implementation, but due to #51338, we assumed it is a bug and hence this PR.

Yeah; if you do find anything weird, please do report it so we can at least make sure it isn't actually a bug. e.g. oxc-project/oxc#3798 is likely a bug on our end is related to strictNullChecks.

Not quite sure where to document export {}, as this has been this way for ages and also applies to DT/hand-written types; I could have sworn this was documented on the website somewhere...