The previous changes to the disposer registration changes introduced a
weird problem exposed in the Gradle (test?) environment, that caused
a performance regression - it seems that keep alive system property
in that setting was cleared at the time the disposer was called, causing
app env disposal when keeping it alive was expected.
This fix captures property value at the moment of disposer registration,
making it similar in that respect to the previous variant and avoiding
the performance degradation.
#KT-56992 fixed
WasmImport annotated functions:
* Restricted to top-level external functions
* Only supports primitive numbers, booleans and Unit
(as return type only) in its signature
* Can't have default parameter values
* Can't have vararg parameters
@JsExport is only allowed on top-level functions in K/Wasm
Restrictions for nested function are handled by reusing K/JS diagnostics
Restrictions for classes and properties are handled by specifying
annotation target in stdlib
Prohibit:
- external enum class
- external tailrec fun
- external suspend fun
- external lateinit var
Add tests for other external diagnostics inherited from K/JS
Now all tests with `Fir` in name are named accordingly to parser which
is used in them -- `FirPsi` or `FirLightTree`. This is needed to keep
consistency between different types of tests, because there is no
single default in parser mode between different scenarios of using FIR
during tail-call optimization.
There can be code, where all next instructions are non-meaningful and
there is a back-edge, for example, while(true){}. Previously, analyzer
incorrectly assumed, that this cannot happen. Now, it keeps track of
visited instructions and says, that there is no meaningful instruction
in such case.
#KT-56815 Fixed
CoreJrtFileSystem uses JrtFileSystemProvider provider to read contents
of jrt-fs from JDK
Implementation of FileSystems.newFileSystem causes metaspace memory leak
that wasn't fixed until JDK 17, see
https://bugs.openjdk.java.net/browse/JDK-8260621
When FileSystems.newFileSystem used to create jrt-fs on JDK9 with
provided java.home value it creates new ClassLoader under-the-hood,
which subsequently leaks due to aforementioned bug
Remove conditional usage of `java.home` + FileSystems.newFileSystem and
switch to use jrt-fs classloader cache regardless of
compiler runtime JDK to reduce classloader leaks
^KT-56789
Trick codegen into generating getfield from the anonymous class instead
of its supertype (e.g. kotlin.jvm.internal.FunctionReference,
PropertyReference or AdaptedFunctionReference). This gets rid of
unnecessary accessors in some cases, and will help in the subsequent
commit that changes logic around access to fields from Java
superclasses.
It was added in 66da676b9e to avoid dependency on descriptors, but now
it doesn't use descriptors anyway, so it's better to inline it than have
it in the interface.
Dmitriy Novozhilov