Function `apiVersionIsAtLeast` was introduced to be able to have
different inline function content inlined to user code call sites
depending on their api version settings. Thus, it should not be
compile-time evaluated when being called in the body of inline stdlib
function.
^KT-59452: Fixed
Mangle invocations of functions with value classes in signature which
override (directly or indirectly) a method declared in Kotlin code.
Otherwise, NoSuchMethodError is being thrown.
^KT-55945: Fixed
In this commit we have a lot of change in test data. This was caused
by the way where we evaluate constants. We split constant evaluation
into two distinct parts: only necessary evaluations for `fir2ir`
(like const val and annotations) and optimizations for lowering.
Now we don't do all constant evaluation on `fir2ir`, but IR
dump is executed after this phase, so test data changed.
#KT-58923
This way we achieve faster compilation time. We want to traverse
IR tree as little as possible. If we add new checker than the time
to evaluate constants basically doubles.
#KT-58923
Variables in `IrInlinedFunctionBlock` declared before the composite
blocks with arguments evaluation may lead to error on codegen while
processing `IrInlinedFunctionBlock`.
^KT-58779: Fixed
This commit fixes the receiver of these method in FIR2IR code
by removing a particular hack.
However, as the consequence of it 2 black box tests become broken in K2.
The reason is KT-54952 which is not yet fixed
Related to KT-54887
This reverts commit 0387ce0365.
This change was not needed because the 'IGNORE_BACKEND_K1: JVM'
directive makes the test generated as ignored, in this case it is the
method named `ignoreSignedToUnsignedConversions` which is not even
recognized as a test.
Local delegated property may origin from a body of inlined function.
In that case, they contain no metatada - which is ok, as the original
local delegated properties contain it and will be serialized.
^KT-58780: Fixed
There are some cases for fake-override actualization which become
prohibited in the new expect/actual model, so few tests start to fail,
despite the fact that they are checking entirely different things
This is needed because of mutable nature of receiver values: implicit
receiver values can be modified because of smartcasts, but in candidate
we need to store snapshot of receiver in the form it was at the beginning
of the resolution
^KT-58823 Fixed
The issue appeared after code refactoring. Originally we didn't
save generated accessor for symbols in `IrFunctionReference`. These
symbols will be processed in their own turn.
#KT-59079 Fixed
It was happening because for MyClass.foo we didn't set overriddenSymbols
properly because in ClassMemberGenerator.convertFunctionContent we
used incorrect containingFirClass that was pointing to anonymous class
instead of MyClass.
^KT-58902 Fixed
Disabling of this language feature (see KT-57570) was not supported in
the (default) light tree mode, and after this change it will not be
supported in the PSI mode as well. This makes the behavior more
consistent and allows us to remove the directive
IGNORE_BACKEND_K2_LIGHT_TREE in the future (KT-56757).
This only applies to JVM and fq-names in declaration references
in IR dumps.
This enables us to run more irText tests on platforms other than JVM
(see KT-58605).
Mikhail Glukhikh