ckohen
ckohen
Alright, this should work. It's probably a much easier diff to read now too :). To summarize, this PR effectively just fixes the over strictness of some parts of the...
The latest refactor brings to light that the change to the listener type is breaking in some very specific scenarios (only for existing generic consumers). In order to help with...
The reason I changed the name is that this no longer has the major point of contention that caused the reversion. I will be putting that in a separate PR,...
I do think a type alias for the "default" map would make sense, though I'd venture to suggest it makes more logical sense to call it something akin to `AnyEventMap`....
As far as I can tell, everything that fails in CI is either unaffected by the changes I made or should've never passed in the first place. Basically every quick...
> CI is fail, fix As mentioned in my previous comment, the CI is failing correctly. I need some guidance from maintainers because those tests should have *never ever* passed...
Ah, of course v22 merges first. I'll rebase and do v20
Well that's very unfortunate. This ultimately seems to be a limitation in the type system currently. Take a look at this simplified example. ```ts class Base {} class Derived extends...
@Renegade334 vlad and I figured out how to let infer do what we want (similarly to what you noticed, as long as it's on the object it can infer from...
One more option we have discovered is an es private property (e.g. `readonly #internalTypeOnlyBrand`) on the `EventEmitter` class (pushed as https://github.com/DefinitelyTyped/DefinitelyTyped/pull/69997/commits/ce5426df2ae914e4f4a7c89d404ee5f132a26e6a since this seems more reasonable). This will work for...