fix: bug in typings after recent PR

[ffMathy]

• I have added automatic tests where applicable
• The PR title is suitable as a release note
• The PR contains a description of what has been changed
• The description contains manual test instructions

Changes

I removed the Unknown event type. It ruins the type discrimination it seems. Not sure how this was not caught by me earlier. Apologies. Didn't discover this until I had to consume the event myself.

Test

Test that the types are still OK.

[gismya]

The problem is that there are other valid event types than the ones currently typed, and they are unknown. Neither solution is great, and I'm now sure which one is worst. If only TypeScript had negated types.

We could examine if the new const in typescript 5 would allow anything useful here. But if we're stuck with the imperfect way it currently works I'd love arguments on this solution rather than type-casting.

And if we go with this we should probably add some comments about how any custom events would have to have their own interface created for them when used.

Does @jimmycallin have any thoughts? I know you've struggled with the lack of negated types before.

[ffMathy]

I think the way forward is to define all of the event types to be honest. And I don't think the current state of things in this PR is very bad. I think the alternative is worse.

[gismya]

I'm inclined to agree that unknown types breaking the type discrimination might be worse. I have some ideas on how to make it more easily extendable and versatile, but that would have to wait until the event typing overhaul I mentioned in the previous PR.

Added a comment suggestion for informing about the incompleteness of the type definitions.

[ffMathy]

That's a great idea.

How many event types are there in Ftrack, and does it take too long to make them all? Can they somehow be auto-generated?

[gismya]

The event hub forwards any event sent to it, with arbitrary topics and arbitrary data. They have been added when they were needed from whatever part of the stack needed them. That's why it takes some research to find them all, document what data they contain, and whether they are public events or just events we use internally. So I'm not sure how many there are that are missing, but from the previous investigation I started there's about a handful.

[ffMathy]

I am an idiot. I think I forgot to bump the version in package.json. So it was not released.

[gismya]

We don't manually bump, that is done through the release scripts. I'll check what went wrong.

edit: I accidentally named the version v.1.4.6 instead of v1.4.6. Now it should be working.

[ffMathy]

Thanks a lot!