There is added a new service named `SubstitutingScopeProvider`, that
provides factory that creates captured types and approximator for them.
In OI they are the same as before commit, for NI they are empty, because
that approximation interferes with NI algorithm
That service is injected into function descriptors and property descriptors
and used for creating `SubstitutingScope` with correct services
Also there is changed time when we approximate captured types in NI
(after all call checkers)
#KT-25290 Fixed
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.
If property accessor rendering is disabled in a test, render annotations
on accessors as use-site-targeted, as was done with
`@setparam:`-annotations. Otherwise they were lost
Note, that this change potentially has some other effects in corner cases
(like the changed test data that is rather sensible because `bar`
in the example is not effectively projected out and can be called
with nulls)
Probably, we need to consider rewriting all other isSomeType methods
in KotlinBuiltins, but now it seems to be a rather dangerous change
#KT-16424 Fixed
Prior to the 1.1.4, nullability related annotations were stored
in types that became hard to maintain at some moment and
we got rid of it (see 57b7b91444)
But enhancement for properties overrides stopped working
because there were effectively no annotations in the
resulting descriptor
#KT-19409 Fixed
#KT-19409 Fixed
Except AnnotationDescriptorImpl, which is refactored in the subsequent
commit.
Note that we no longer check the presence of parameters with the
corresponding names in the annotation class in
LazyJavaAnnotationDescriptor, this is why test data changed
The regression appeared after
b5a8ffaddc
when we started trying both static and member methods until
first success and when there is no successful
we were just leaving the last one (e.g. private member)
But the actual problem is that we were commiting the trace
in case of single (but incorrect) result in resolution mode of
SHAPE_FUNCTION_ARGUMENTS when we couldn't yet choose the
correct static method
Also we shouldn't choose a shape for callable reference
using only the knowledge that result is single:
it may lead to the wrong inference result
(see test with Pattern::compile)
#KT-17597 Fixed
Denis Zharkov