Original problem is that lowered ir closures doesn't meet inliner expectations
about captured variable position in inlining method.
E.g.: Call 'foo(valueParam) { capturedParam }' to
inline function 'foo' with declaration
inline fun foo(valueParam: Foo, inlineParamWithCaptured: Bar.() ->) ....
is reorganized through inlining to equivalent call foo(valueParam, capturedParam1, cp2 ...).
But lowered closure for lambda parameter has totally different parameters order:
fun loweredLambda$x(extensionReceiver, captured1, cp2..., valueParam1, vp2...)
So before inlining lowered closure should be transformed to
fun loweredLambda$x(extensionReceiver, valueParam1, vp2..., captured1, cp2..)
#KT-28547 Fixed
This patch mutes the following test categories:
* Tests with java dependencies (System class,
java stdlib, jvm-oriented annotations etc).
* Coroutines tests.
* Reflection tests.
* Tests with an inheritance from the standard
collections.
Most of modifier diagnostic is expressed by REDUNDANT_MODIFIER, INCOMPATIBLE_MODIFIERS, REPEATED_MODIFIER, WRONG_MODIFIER_TARGET, WRONG_MODIFIER_PARENT.
A set of modifier diagnostics is not in use now (but not deleted yet).
And String.length as well.
This is done for JVM interoperability: java.lang.CharSequence is an open class
and has a function 'length()' which should be implemented in subclasses
somehow.
A minor unexpected effect of this is that String.length() is now a compile-time
constant (it wasn't such as a property because properties are not supported in
compile-time constant evaluation)
#KT-3571 Fixed
Mikhael Bogdanov