Missing dependency fix for Type Functions

[karl-police]

An alternate version of my fix for #1418

Where instead of changing the positions, I block the dependencies.

As the actual solution concept is just this: #1418 (comment)

The solution is to block all constraints that depend on the pending-expansion that is being expanded, on the resolution of the reduction constraint that is being created.

[karl-police]

Though if the positions can also be forced at the location it would make certain solving faster :person_shrugging:

Let me know if I should change anything 🤷

[karl-police]

welp

[karl-police]

@vegorov-rbx What can I do?

I recently tried to look again into this issue. The method I put here still works.

 

But I made a comparison to a user-defined type function KeyOf with the regular keyof.

And for some reason the user-defined type function doesn't have this issue. But I can't figure out why :(

Why do user-defined type functions not have this issue? I've set so many breakpoints I can't figure out, there's a lot going on in this. And unlike me blocking things. user-defined type functions put the dependency directly inside of the type somehow, I don't know how. I don't know. :(

 

See the difference, right works and is using a user-defined type function
left is without this PR and using regular keyof

https://www.diffchecker.com/0OyWfSsK/

https://paste.ivr.fi/fadymiluwy.json
https://paste.ivr.fi/uhulekebur.json