There is an inconsistency between old and new inference for storing
receivers of resolved calls. In new inference, for captured types,
receiver will be changed and to preserve behavior of the old inference,
we use original one during important checks.
This is more a workaround than a solution and should be revisited.
#KT-31356 Fixed
#KT-29948 Fixed
#KT-31360 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.
Since we skipped trivial constraint with `Any?` from parameter type of
function `equals`, the compiler thought that there is no proper
constraints (upper bounds do not matter here) and marked resolved
call as a failed one, then diagnostic about missing equals was added
Also, tune `TrivialConstraintTypeInferenceOracle` for `Any?`-like
constraints
#KT-30724 Fixed
Add `IntegerLiteralTypeConstructor` that holds types, that can take
integer literal with given value. It has two supertypes
(`Number` and `Comparable<IntegerLiteralType>`) and have
special rules for subtyping, `intersect` and `commonSuperType`
functions with primitive number:
Example (assuming that ILT holds Int type):
* ILT <: Int
* Int :> ILT
* ILT intersect Int = Int
* commonSuperType(ILT, Int) = Int
#KT-30293 Fixed
#KT-30446 Fixed
This commits introduces testdata changes, where NI behaviour strictly
improved, after several previous fixes.
For some tests, just WITH_NEW_INFERENCE directive was added. It
indicates, that some of previous commits first introduced error in that
test, and then some other commit fixed it (netting no overall testdata
change). It is preferrably to keep those annotations until we will
migrate to NI completely, to prevent unexpected regressions.
Otherwise, when completing all the unsuccessfull candiates,
resolution of each lambda-arguments starts repeatedly for each candidate
that leads to exponential time
NB: Changes in `completeArguments` are necessary because otherwise
nested lambdas will be analyzed twice:
once for the main resolved call, and then for all candidates
that again leads to exponential complexity
#KT-16672 Fixed
#KT-19457 Fixed
receiver type with modifiers or annotations should be surrounded in parentheses on rendering: '(@A R).() -> T'
This also fixes stub builder tests (which check that stubs are consistent with rendered descriptors).
(required for 'suspend' on functional types).
TYPE_REFERENCE element now has MODIFIER_LIST child, which hosts annotations and modifiers for the corresponding type reference.
Annotations and modifiers written before an extension function type are now parsed as annotations and modifiers for the functional type, not the receiver type.
So, '@Ann A.(B) -> C' was '(@Ann A).(B) -> C', and became '@Ann (A.(B) -> C)'.
NB: DSL_SCOPE_VIOLATION testData updated accordingly.
Type projection variance modifiers ('in', 'out') belong to a separate modifier list under corresponding type projection (not under a type reference).
'A<in suspend T>' is 'A<(in (suspend T))>', 'A<suspend in T>' is an error.
In stub builder, create a modifier list node to host annotations and modifiers (none so far; TODO properly serialize/deserialize types with modifiers).
There is no need to report an error in case of non-top-level
capture conversion, just don't add relevant capture type in the system instead
If system can be solved successfully without captured type, then it's just fine
(see KT-13950)
In case of contradiction TYPE_INFERENCE_PARAMETER_CONSTRAINT_ERROR is reported
#KT-13950 Fixed
Mikhail Zarechenskiy