if it is called using parens and not by calling 'invoke' method.
Use underlying type when calling continuation constructor if suspend
function is method inside inline class.
#KT-43505 Fixed
#KT-39437 Fixed
inside 'create' if 'create' overrides 'create' from
BaseContinuationImpl. In other words, unbox the parameter if 'create'
accepts only one parameter.
#KT-43249 Fixed
#KT-43533 Fixed
inside 'invoke' if 'create' does not override 'create' from
BaseContinuationImpl. In other words, when suspend lambda accepts more
than one parameter (including receiver).
Do that only if we do not generate bridge 'invoke' method, since
inline classes are unboxed in the bridge.
Use mangled name for 'create' function in this case inside 'invoke'.
#KT-43249 In progress
#KT-39847 Fixed
#KT-38937 Fixed
Like function arguments, they are context-dependent, but unlike function
arguments, callable references should be resolved eagerly as if they are
explicit receivers.
- rename fileToPurenessInitializers onto fileToInitializerPureness
- remove redundant check on top-level property
[JS IR] Rename initialis* to initializ* for consistency
[JS IR] Move propertyLazyInitialization property to context from configuration
[JS IR] Add test on lazy initialization properties order
[JS IR] Add multi module for lazy initialization of properties
[JS IR] Move tests onto js.translator
[JS IR] Rename fileToInitializerPureness according to context name
^KT-43222 fixed
The debug experiece of destructuring patterns in lambdas is different
across the two backends due to the IR backend moving local variables
to fields.
However, since the destructuring variable is never actually visible in
the debugger (no linenumbers in the live range of the variable), and
the variable is never used for anything other than hiding it from the
debugger, we propose that it is not actually necessary to include it
in the LVT (and in fact, could be left out of the LVT on the old
backend).
1. Use 'x' for each parameter, which is not an inline class, every
possible clash is handled by signature rather than name. This change
makes more API changes binary-compatible. So, the changes are in line
with the vision of inline classes are value classes, like primitives.
2. Take return type into account when mangling a function if the return
type is inline class. Otherwise, boxing bridge will not be generated,
which leads to CCE at runtime.
Do not destructively update the @JvmStatic function, instead
create a copy on first access, and replace the original with
the copy in the jvm static lowering. This ensures that the original
function is seen in other lowerings independently of file lowering
order.
In particular, `@Deprecated` is not lost anymore and accessors of
deprecated property are generated with ACC_DEPRECATED just as in the old
backend.
#KT-43327 Fixed
When calling vararg methods with a generic vararg type without
passing explicit parameters, we have to allocate an empty array
of the right type. We failed to do so previously, as we did
not take the type arguments for the dispatch receiver into
account.
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'.
Iaroslav Postovalov