Input and output types are crucial for type variable fixation order and
analysis of postponed arguments (callable references, lambdas).
Specifically, if there is non-fixed type variable inside input types of
a callable reference, then we'll postpone resolution for such callable
reference.
Initial example with the expected type `KMutableProperty1<*, F>` caused
problems because input types were computed incorrectly (while there
aren't input types here)
#KT-25431 Fixed
One might think that it shouldn't be run because of 'if (declaration
!is KtNamedDeclaration) return' check in the 'check'-overload.
However, for default accessors we pass PSI for property, i.e.
KtProperty (see 'DeclarationCheckers.checkAccessors'), which
obviously passes this check
Note that we can't use `DescriptorToSourceUtils` here, because it's
returns `KtProperty` for default accessor too.
This commits adds specific check for that case, to avoid exception in
KT-28385. Ideal solution would be to either don't launch checkers on
such parameters, or explicitly declare semantic of the
'declaration'-parameter, e.g. to rename it to 'reportOn' (see KT-28403
for discussion)
^KT-28385 Fixed
Private function convertToOverloadResults() has two call sites, both:
* inside member functions of the NewResolutionOldInference class
* passing NewResolutionOldInference.languageVersionSettings
Therefore, it is safe to remove this parameter from the signature of
convertToOverloadResults(). The occurence of languageVerstionSetting
in it's return statement will refer (implicitly) to the definition from
its class.
Before this, descriptors for such constructors were created in-place where they needed by the codegens.
However, presence of symbol table in IR backend requires the single instance of constructor descriptor across all compilation to be able to reference it and create a symbol. This support of generics in kotlinx.serialization on Kotlin/Native.
I've put `isIncompatibleEnums` to TypeIntersector because I placed
all of its usages after all of the TypeIntersector::isIntersectionEmpty ones
^KT-28225 Fixed
Extract Java-specific code into module 'frontend.java' and use already
existing JavaAnnotationArgument facilities to determine the default
parameter value in an actual Java annotation class.
Annotation arguments are not yet supported because to create an instance
of AnnotationValue, we need a descriptor, which is not available at this
point.
#KT-22704 In Progress
#KT-28077 Open
Before this commit, we compared property visibility with constructor
visibility only, which is incorrect. Now we compare property visibility
also with class visibility
#KT-19613 Fixed
Return type is not needed for checking overloads, but querying it may
involve resolving function bodies, which usually happens after overload
checking (see LazyTopDownAnalyzer.analyzeDeclarations) and at this point
can lead to incorrect BACKING_FIELD_REQUIRED value being computed for
some properties (see KT-27895)
#KT-27895 Fixed
Only invariant array projections and non-null element types will be
supported soon (see KT-26568), so it makes no sense to store the
complete type in KClassValue. What we need is only the ClassId of the
class, and the number of times it's wrapped into kotlin/Array, which is
exactly what ClassLiteralValue represents.
This change helps in decoupling annotation values from
descriptors/types. The only constant value that depends on descriptors
is now AnnotationValue.
#KT-26582 Fixed
Rename 'doCheckContract' into 'parseContractAndReportErrors',
emphasizing that it works with definitely call to 'contract' from
stdlib.
In particular, it means that it should either return non-null
value or report some errors.
Note that this commit doesn't change behavior of this code (modulo cases
when something in 'parseContract' throws exception), but just refactors
the code to be more clear and easy to reason about.
It conveys meaning of the code more celearly (because that's what we
want to do here: set LazyContractProvider no matter how code finished),
and also defends from several corner-cases (like, make sure that we
properly set LazyContractProvider even if doCheckContract has throw an
exception)
Returning `null` from `doCheckContract` functions means that we
have failed to parse contract function and should report an error, but
if the called function isn't true contract, we shouldn't evaluate that code
at all.
#KT-27758 Fixed
Otherwise, top-level function with the name 'equals' and suitable
signature will be (erroneously) treated as true 'equals' invocation,
leading to further exception (see EA-126602)
Ilya Chernikov