This is already the case for straightline code such as
```
inline fun <R> f(size: Int, block: () -> R): R {
var result: R
result = block()
return result
}
```
However, if the local variable introduction happens at a merge
point as in the following example, we allow the unboxing but
only do it halfway. The initialization of the local is still
done with a null value.
```
inline fun <R> f(size: Int, block: () -> R): R {
var result: R
while (true) {
result = block()
if (size == 0) break
}
return result
}
```
This change disallows unboxing for this move complicated
case as well by bailing out if a local use is with a
TaintedBoxedValue (merge of Object and Integer).
^KT-48394 Fixed.
Looking for the primary constructor manually doesn't work if it's
private in the other module on JVM, because private declarations are
skipped in IrLazyClass.
Otherwise, R8 does not transform kotlin-reflect, failing bootstrap.
Leaving end label the same is safe, since we do not remove labels during
transformation.
Do not check, that all Unit predecessors are POPs. This is safe for the
same reason, as it is safe to allow some of ARETURN sources not be
suspension point results.
To elaborate, before Unit, the stack is empty. This is because if there
are multiple paths to Unit and at least one of them comes from POP after
suspension point (we are interested in this case only - otherwise, the
call is not tail-call), in path from said POP the stack is empty, since
after suspension point the stack contains only one element. Thus, the
stack in other paths leading to Unit has to be empty, otherwise, merge
operation is not possible and ASM will report error during analysis.
Since the stack is empty in all paths, we can hoist Unit and following
ARETURN to predecessors, effectively turning path from suspension point
to tail-call.
The problem in the test case was that `JImpl.entrySet` was detected by
ReplaceDefaultImplsOverriddenSymbols as a class fake override, which
overrides non-abstract interface method. Thus, overriddenSymbols of
`MyMap.entrySet` were changed in
`ReplaceDefaultImplsOverriddenSymbols.visitSimpleFunction`.
Later, BridgeLowering tried to determine which special bridges to
generate for `MyMap.<get-entries>`, and for that it inspected existing
methods and their overrides.
Normally we would generate the following special bridge:
getEntries()Ljava/util/Set;
invokespecial JImpl.entrySet()Ljava/util/Set;
However, because of incorrect overrides, the generated class method was
selected here instead of the expected `Map.<get-entries>`:
https://github.com/JetBrains/kotlin/blob/06001fc0919c814e757caa494891619882fae15f/compiler/ir/backend.jvm/src/org/jetbrains/kotlin/backend/jvm/lower/BridgeLowering.kt#L282
and since the JVM signature of the generated class method is the same as
that of the fake override in MyMap, we never got to generating the
special bridge.
The solution is to skip Java classes when looking for class methods
which override non-abstract interface methods. This logic only makes
sense for Kotlin code, because only Kotlin has DefaultImpls, and the
requirement to generate non-abstract fake overrides of interface methods
as actual methods in the bytecode, delegating to DefaultImpls.
#KT-48167 Fixed
For some reasons, lambdas and function expressions are represented
slightly differently in psi2ir. Lambdas are translated to a block with a
function of origin LOCAL_FUNCTION_FOR_LAMBDA and name "<anonymous>", but
function expressions are translated to a block with DEFINED function
"<no name provided>".
Tweak the condition for detecting local functions a bit, to avoid
similar situation in the future if we add some other origins for local
functions.
#KT-48207 Fixed
In case internal inline function references private top level function
after inline such function (T.L.P.) couldn't be referenced
in klib and IC cache. So create internally visible accessors for P.T.L.
function similar to what JVM backend does.
- add box test
In order to make resolution still work for members not available from
`Nothing`, we track the type without `Nothing?` and use that for
resolution instead.
Mads Ager