#KT-36247 fixed
A lot of testdata changed because significanly less (error) descriptors
are created for unresolved types, so diagnostics became different.
This problem is only relevant when isTypeRefinementEnabled == true (HMPP projects)
Ambiguity accidentally was removed after 471134d
There, for areCallableDescriptorsEquivalent we stopped assuming
as impossible a situation of having identity-different descriptors
in the same containing declaraton that still might be considered equal
So, before 471134d we were comparing
"fun foo(x: String)" with "[substituted] fun foo(x: String)"
and areCallableDescriptorsEquivalent returned false for such case.
Thus, both overrides were left in the resulting set.
After 471134d, those two descriptors
becamed considered as equal thus having a possibility to remove any of them.
The problem is that "areCallableDescriptorsEquivalent" has kind of
unclear contract. Effectively it checks whether two descriptors match
to the same declaration.
But some of the usages expect that it also makes sure that descriptors
have the same substitution (see org.jetbrains.kotlin.resolve.calls.smartcasts.IdentifierInfo.Variable#equals)
So, the straight solution is using original descriptors for the cases
where we need to make sure that descriptors relates to actually different
declarations
^KT-34027 Fixed
This commit doesn't change behaviour of any inference algorithm, it
introduces opportunity to switch constraint system that is used for
overload resolution and fix problematic cases by changing one enum
entry.
Due to fundamental changes, there are cases where a new inference
algorithm reports overload resolution ambiguity errors (#KT-31670,
#KT-31758), which is correct from its point of view. However, this is
a breaking change and to really make it, we should be very confident
and have enough motivation for it, therefore, we don't change behavior
now in order to collect more examples (if there are any). And if we
find a lot of erroneous examples, we'll be able to change the behavior
quite simply
Expression will be checked against expected type later.
Theoretically, this is not very good, but it aligns with the old
inference, plus it helps avoiding multiple type mismatch diagnostics.
Mikhail Glukhikh