In cases when the test has the JVM_TARGET directive, use that one for
Java compilation as well.
Otherwise, for box tests, the corresponding test in `JvmTarget6OnJvm6`
(module `:compiler:tests-different-jdk`) will fail. However, it affects
all codegen tests, so fix a bunch of them which use Java 8+ features to
explicitly compile with JVM target 1.8. In particular, this obsoletes
the SKIP_JDK6 directive in those tests because "JVM_TARGET: 1.8" also
skips it for JDK 6.
The check for IS_SOURCE_6_STILL_SUPPORTED is needed in order to still be
able to run tests in the project while only having a single JDK > 11
installed, and having all of the env vars JDK_16, JDK_17, JDK_18
pointing to that JDK.
If it isn't ANNOTATION_CLASS, the newly added code in
`FunctionCodegen.isDeprecatedHidden` (0e91d3fcb0) ends up transforming
IR annotation constructor calls to annotation descriptors, and an
assertion fails in `IrBasedDeclarationDescriptor.toAnnotationDescriptor`
which checks that the class has kind ANNOTATION_CLASS.
Specifically, this failed in the JVM IR bootstrap on
`CallResolutionInterceptorExtension.interceptCandidates` from module
'frontend'.
Before this commit, we generated fake overrides at last FIR2IR stage,
after having all functions and classes built. This could lead to a
situation when fake override was called before it was generated.
This commit fixes this situation.
We generate @java.lang.Deprecated annotation on methods of $DefaultImpls
classes in compatibility mode. This annotation has RUNTIME retention and
should be visible.
Also, get rid of representing annotations as Class'es (yes we know that
these annotations are in compiler CLASSPATH, but we should not rely on
such classes and associated information).
Before this commit we considered !isOverride as a sign that
function / field / accessor has no overridden symbols.
However, it's false for deserialized, because isOverride
is always false there.
This commit fixes 68 BB tests but breaks 25 BB tests (not yet muted)
To convert smart cast expression, now we just convert
original expression in standard way and wrap it with TYPE_OP.
Before this commit original expression was converted in different way,
that led to errors e.g. for this expression casting.
The only case we may need fake override for private members is
a situation when class refers itself with different type arguments.
So in this commit we forbid such fake overrides when we can prove
that class does not refers itself here.
Library methods such as 'listOf' are resolved
to have the package fragments as their parents,
but JVM expects their containing file classes as parents.
This fix generates those file classes and
uses them as parent replacements for such library methods.
The first change here is swapping the isCompatibility and hasJvmDefault
checks. Otherwise behavior could be different depending on -Xjvm-default
mode even for non-JvmDefault declarations, which makes little sense.
Another change is avoiding generating $default stubs for fake overrides
in interfaces, which replicates the behavior of the current backend.
(Note that this change also fixes the first problem on the newly added
tests, but the first change seems useful anyway.)
Dmitriy Novozhilov