Partial record key type checking failure

[AntonOfTheWoods]

Bug Report

🔎 Search Terms

typescript partial record key type checking failure

🕗 Version & Regression Information

• This is the behavior in every version I tried, and I reviewed the FAQ for entries about "Common "Bugs" That Aren't Bugs"

⏯ Playground Link

https://www.typescriptlang.org/play?ts=4.6.2#code/C4TwDgpgBAsiDSEQBVzQLxQOQEMtQB9sAjfIrAYywG4AoUSKAZQFdiA1HAGxYgGdUjTAG9aUKADduvPgC4ofYACcAlgDsA5gG0AunQC+demlggAwgHs1wHOohLBGKKPHEcAE049+8gAo4lYBVuAB4AJQgKCyV3ELhEFDQAGmY2LxlHAD5MgyMAMxY1CiCrKBwCopK1AAocAGskeXikRwBKFyguCGAoBpAoTHqkOnEAelG+gewqaihxqEKogFsliGsoVbK1dzK+PhUNNS3+izyylOIoaKgKLZ3lsBUuHCqoPNsuPjEbq0UoNQgAHcACIvHBNcxWGx2BwmERQNyeaQ+ZxQLR9HTyYTHYAAC3UGgAkgB1HDWeS6KD6KlUgxAA

💻 Code

type MyKeyType = 'a' | 'b' | 'c';
type SubValuesType = {
  values: string[];
};

type MyContainerType = {
  badValues: Partial<Record<MyKeyType, SubValuesType>>;
};

function afunction(akey: MyKeyType){
  let key = akey;
  //key = 'c'; // uncomment me and assign any of a, b or c and compilation fails
  const newData: MyContainerType = { badValues: { [key]: { anythingIWant: [] } } };
}

🙁 Actual behavior

There is no type checking compilation error. I am able to assign an invalid type to newData

🙂 Expected behavior

Checking error. At the very least I would call this a MASSIVE "principle of least surprise" failure. If this is the expected behaviour, it is extremely unintuitive.

[RyanCavanaugh]

Sort of a duplicate of #38663

I think we could get away with raising an error for t = { [k]: v } when k is a literal union type "s1" | "s2" | "s3" ... and v isn't assignable to the union t["s1"] | t["s2"] | t["s3"] | ... (intersection would be more sound, but impractical due to producing never a lot of the time). It'd be a breaking change but it's hard to see how that code wouldn't have an actual bug.

[AntonOfTheWoods]

Sorry I missed the duplicate @RyanCavanaugh . Looking at the issue you refer to looks close enough to me. If the fix you think might be viable here would work there also, I'm definitely happy to close this.

[RyanCavanaugh]

On second read, the issue I linked is definitely the same. I'll transfer the metadata