Support expand on concrete tuple values.

[csyonghe]

Closes #8061.

Along with the fix, also enhanced coercion/overload resolution to filter candidates based on the target type, allowing tests\language-feature\higher-order-functions\overloaded.slang to pass.

[juliusikkala]

I had also started on this yesterday (minus the overload part), but got preoccupied with something else and luckily didn't get too far yet. I had stopped when I realized I'd need a function like the maybeExpandArgList you implemented 😄

It's good that we can remove isInVariadicGenerics. isKeywordAvailable is in general a really nice solution IMO, since now we can add new keywords without breaking changes.

One more thing I'd like to see tested, is this case:

Tuple<int, float> x = (2, 3.0f);

float arr[2];
int i = 0;
expand arr[i++] = float(each x);

or in general, expanding concrete tuples outside of function call parameters. I encountered a type-related error with that code even after allowing expand-each outside of variadic generics. I'm not on a PC with slangc currently, so can't check what it was or if this PR has already fixed it. Either way, I'd feel more comfortable with something like that being in the tests.

[csyonghe]

@juliusikkala I don't think your code is valid though, because there is no overload of float.init that will take a generic type like each T without any constraints.

[csyonghe]

Again, even with concrete types, the expression inside expand is checked before expansion, similar to other generic cases.

[juliusikkala]

Ah ok, that's a bit unfortunate :( but if that's the desired behaviour, then LGTM