Previously, type substitution, which is critical for matching generic
header/impl members with each other, was only performed when
checkImplementationHasHeaderDeclaration was called for impl class
(areCompatibleClassifiers creates the correct substitutor). This was
done in areCompatibleClassifiers: a substitutor which maps type
parameters of the header class to type parameters of the impl class was
created.
Now we create the same substitutor when
checkImplementationHasHeaderDeclaration is called for an impl member of
an impl class as well, manually.
#KT-15230 Fixed
Currently this is achieved with several hacks:
- Postpone computation of argument type info when there is no candidate resolver. We have to do this, because we don't have expected type and therefore we could write wrong information to trace
- Presume that for annotation calls there is only one candidate resolver and then resolve arguments with expected type (see `getArgumentTypeInfo`), otherwise because of quadratic complexity of the algorithm resolve would be slow
Previously there were three LanguageFeature instances -- Coroutines,
DoNotWarnOnCoroutines and ErrorOnCoroutines -- which were handled very
awkwardly in the compiler and in the IDE to basically support a language
feature with a more complex state: not just enabled/disabled, but also
enabled with warning and enabled with error. Introduce a new enum
LanguageFeature.State for this and allow LanguageVersionSettings to get
the state of any language feature with 'getFeatureSupport'.
One noticeable drawback of this approach is that looking at the API, one
may assume that any language feature can be in one of the four states
(enabled, warning, error, disabled). This is not true however; there's
only one language feature at the moment (coroutines) for which these
intermediate states (warning, error) are handled in any way. This may be
refactored further by abstracting the logic that checks the language
feature availability so that it would work exactly the same for any
feature.
Another issue is that the difference among ENABLED_WITH_ERROR and
DISABLED is not clear. They are left as separate states because at the
moment, different diagnostics are reported in these two cases and
quick-fixes in IDE rely on that
The problem was that when resolving super-calls we used known substitutor
when creating a type alias constructor, thus its original return itself,
while it's expected that it should return the descriptor before substitution
The main idea of the fix that `createIfAvailable` should always return
unsubstituted constructor.
Note that known substitutor for type alias constructor should be based
on abbreviation.
The test change seems to be correct as PROJECTION_IN_IMMEDIATE_ARGUMENT_TO_SUPERTYPE
is already reported.
Beside this, resolution behavior isn't expected to be changed dramatically
See how we translate raw types to Kotlin model:
RawType(A) = A<ErasedUpperBound(T1), ...>
ErasedUpperBound(T : G<t>) = G<*> // UpperBound(T) is a type G<t> with arguments
ErasedUpperBound(T : A) = A // UpperBound(T) is a type A without arguments
ErasedUpperBound(T : F) = UpperBound(F) // UB(T) is another type parameter F
Stack overflow happens with the following classes:
class A<X extends B> // NB: raw type B in upper bound
class B<Y extends A> // NB: raw type A in upper bound
when calculating raw type for A, we start calculate ErasedUpperBound(Y),
thus starting calculating raw type for B => ErasedUpperBound(X) => RawType(A),
so we have SOE here.
The problem is that we calculating the arguments for these raw types eagerly,
while from the definition of ErasedUpperBound(Y) we only need a type constructor
of raw type B (and the number of parameters), we don't use its arguments.
The solution is to make arguments calculating for raw types lazy
#KT-16528 Fixed
Members declared in interface or overriding members declared in super-interfaces
can be implemented by delegation even if they override members declared in super-class
(NB for interface this can be only 'kotlin.Any').
When synthetic member comes not from the receiver type itself,
but from one of its supertypes it doesn't make sense to subsitute
the member with receiver type, we should obtain relevant supertype
and use it instead.
#KT-16578 Fixed
When there is unsuccessful (e.g invisible) result of one kind (static/non-static)
and there is a successful candidate for another kind, choose the latter one.
Note, that we have to postpone commiting trace until we choose one of the results,
otherwise errors of unsuccessful results are reported
TODO: Maybe it makes sense to report all results when all of them are
unsuccessful (NONE_APPLICABLE or something like this)
#KT-16278 Fixed
The problem was that LanguageVersionSettingsImpl.DEFAULT did not have
"WarnOnCoroutines" as a feature and so it was manually added to the settings,
but only in two places: in the compiler and in the IDE
In most of the cases it was used together with CLASS_ONLY and vice versa
They only case when CLASS_ONLY was used without INNER_CLASS
is possibleParentTargetMap.COMPANION_KEYWORD.
But after diagnostic NESTED_OBJECT_NOT_ALLOWED has been introduced,
there's no sense in the restriction of exactly the companion's parent
For clarification see test data changed
Usually, we have only CapturedType with such constructor.
We should prevent creation such types in the future (KT-16147).
Also added test for KT-14740 where this problem originally appears.
A lot of problem arise with current solution
(loading them with lowpriority annotation + additional call checkers):
- We errorneously treated ArrayList.stream as an existing method, while
it's just a fake override from List
- The same problem arises when creating a class delegating to List.
Also the latter case is failing with codegen internal error
(see issue KT-16171)
The negative side of this solution is that instead of reporting meaningful
diagnostic, there will be UNRESOLVED_REFERENCE.
But it seems to be better than having strange problems like ones described above.
#KT-16073 Fixed
#KT-16171 Fixed
mglukhikh