The implementation is rather limited. Advanced cases, such as
replacing a script with another one with different import directives,
won't work because of non-trivial relationship between 'FirFile' and
'FirScript'.
Normally, 'DefaultErrorMessages' extensions should only be loaded
from the main compiler JAR. However, in the in-process JPS build,
there are two versions of the compiler classes in the process, and
'ClassLoader's there are not fully isolated.
'ServiceManager' loads extensions from both 'META-INF' locations. If
the JPS plugin bundles newer compiler components, newly appeared
'DefaultErrorMessages' extensions will be loaded from the JPS (parent)
'ClassLoader', causing an exception.
Such a problem appeared with 'DefaultErrorMessagesWasm':
- - - - -
java.util.ServiceConfigurationError:
org.jetbrains.kotlin.diagnostics.rendering.DefaultErrorMessages$Extension: org.jetbrains.kotlin.wasm.resolve.diagnostics.DefaultErrorMessagesWasm not a subtype
- - - - -
In the lowered IR there are often references to type parameters whose
containers are not in the current scope. This is incorrect semantically,
but it works in practice due to erasure, so when the mangler is used on
the lowered IR, we don't want to crash the compiler.
Otherwise, the type parameters of the `kotlin.script.state` property
of type `AggregatedReplStageState` were undefined, which led to
building an ill-formed signature for the script class constructor.
From now on we load script configurations for embedded Kotlin code
snippets. Potentially this might lead to an extra load during a file
analysis.
IDEA side test for the change is here:
`HighlightingTestGenerated.Uncategorized#testKotlinInJavaInjection`
^KT-56632 fixed
A::foo is preferably resolved as an unbound reference to A.foo, however
if A.Companion is a subtype of A and the expected type has one fewer
parameter than the type of the unbound reference, generate a bound
reference with the companion as receiver.
^KT-56519 Fixed
`supportsEnumEntries` in `EnumCLassLowering` is not really
needed, but it was left just in case.
Both K1 and K2 ensure the feature is
enabled before they generate IR.
`supportsEnumEntries` is not enough, because the frontend will not
generate `entries` if `kotlin.enums.EnumEntries` is not available
(linking against an outdated stdlib is considered a valid use-case).
From the resolution point of view,
it's OK to resolve into `entries` while
the feature is off, because then K1 reports an error on the call site.
^KT-56687 Fixed
^KT-55614 Fixed
This change allows to revert adding `WITH_STDLIB` directive
to tests which happened at `a9343aeb`.
Co-authored-by: Alexander Udalov <Alexander.Udalov@jetbrains.com>
The expression needs to be resolved first to determine if there is a
receiver that needs to be extracted to a temporary variable. Also, the
special case for prefix increment/decrement on local variable without
delegates requires resolution to check if the variable is local.
^KT-56771 Fixed
^KT-56659 Fixed
Yan Zhulanow