This was broken in c1ab08c8ce where we started to represent KClassValue
as a ClassId of the referenced class + number of times it's been wrapped
into kotlin.Array. Local classes do not have a sane ClassId, so in this
change we restore the old behavior by representing KClassValue with a
sealed class value instead
#KT-29891 Fixed
Consider the following situation:
```
class Inv<T>
typealias A<K> = Inv<K>
typealias B<V> = Inv<A<K>>
fun <U> materialize(): B<U> = TODO()
```
Type `B<U>` is expanding to `Inv<Inv<U>>` and for this type `B<U>` and
`Inv<A<U>>` are abbreviated types, but due to a bug we forgot to make
substitution for `Inv<A<U>>` and were getting abbreviated type
`Inv<A<K>>` where `K` is a type parameter from the typealias declaration.
This bug didn't affect subtyping anyhow but the incorrect type was
serialized and caused problems during deserialization as there wasn't
`K` in deserialization context.
#KT-24964 Fixed
#KT-20780 Fixed
#KT-20065 Fixed
#KT-28236 Fixed
#KT-21775 Fixed
Augmented assignment operator (e.g., '+=') can be resolved to simple
function call ('plusAssign'). In that case, augmented assignment LHS
can be an arbitrary expression, and may have no associated ResolvedCall.
For example:
(a as MutableList<Int>) += 42
Note that it can happen only in case of augmented assignment operator
convention resolution, because all other forms of assignment-like
operator desugaring require some kind of 'store' operation
(property setter, 'set' operator for array element expression, etc),
and should resolve to some combination of calls.
In that case we simply generate LHS on 'load', and throw assertion on
'store'.
It started to fail with AssertionError in inlineOnly.kt after
d267f1e875 because "$annotations" methods have annotations from the
corresponding property (which can include InlineOnly), but are never
inline. The test will be unmuted as soon as we stop using wrapped
descriptors in the IR backend to determine access flags which should be
generated on a declaration in the bytecode
Introduce MetadataSource as a way to store the original descriptor for
any element (before any lowerings) and maintain it until the end of the
codegen where it's used in generating the metadata. Note that JVM
signatures written to the metadata are formed from the _resulting_
generated elements, not by mapping the original descriptors.
Some corner cases are not supported yet, namely properties declared in
companion objects, synthetic methods for property annotations,
JvmPackageName, etc.
#KT-29119 Fixed
This is needed to get rid of the code that appends "$companion" to
properties moved from companion, because it caused inconsistencies in
the ABI and in JVM signatures stored in the metadata
For comparison intrinsics and for instanceof checks, make
it possible to get the the stack value produced and branch
on that directly instead of materializing a boolean to
branch on from it.
That reduces code such as
```
IF_CMPEQ L1
CONST_0
GOTO L2
L1: CONST_1
L2: IFEQ L3
```
to just one IF_CMP instruction.
Inline function descriptor in derived class represented as FAKE_OVERRIDE.
So we should find it in base class declaration
(not interface cause inline function can't be virtual, but always final)
and then check class version.
#KT-29402 Fixed
The generated code is more inline with java, and we avoid the error of
accessing package-private field outside of the package.
However, this changes semantics a bit. Now, a user should set assertion
status of inline-site's package, instead of inline function's one.
#KT-28317: Fixed
Alexander Udalov