- This commit moves all session invalidation tests from the IDE which
are non-cyclic to the Analysis API. Cyclic session invalidation tests
cannot be moved because the Kotlin test infrastructure doesn't support
cyclic module dependencies.
- In fact, these new tests cover more cases, because the IDE tests only
covered global module state and module out-of-block modification
events. Crucially, global source modification events are now
supported, which require libraries to be left alone.
- Module out-of-block modification for libraries is technically
undefined, but the test still checks this behavior. The current
session invalidation implementation behaves the same for any kind of
module-level modification event, so it is possible to publish a module
OOBM event for libraries. If this changes in the future, we can
disable those tests. But for now, it's good that these tests run,
because they will alert us to (accidental) changes to this behavior.
^KT-56288 fixed
^KT-65265 fixed
- Rewrite `TestModule.publishModificationEventByDirective`: we can get
the `KtModule` to publish an event for via the test module. The
directive can now be made optional, which is necessary when we want to
specify it in potentially multiple modules.
- Introduce the `WILDCARD_MODIFICATION_EVENT` directive, which allows
specifying THAT some modification event should be raised for a module,
but not WHICH one. This allows generating multiple tests which raise
different modification events over the same test data.
- Add various convenience functions for publishing wildcard modification
events. `publishWildcardModificationEventsByDirective` for the test
module structure is quite opinionated, but takes a lot of work from
test implementations.
- Because `ModificationEventDirectives` are part of the analysis test
framework, we register them as a general Analysis API test directive.
^KT-56288
Instead of always looking up smart casts on `lastNode` and separately
tracking the receiver state, simply track which flow the smart casts
belong to right now. This tracked flow is auto-advanced when new
`lastNode`s are created, but can be manually rolled back for things like
KT-63709.
- Perform all calls to xcrun in XcodeValueSource.
- Use the fully computed Xcode instance as the current Xcode instance.
- Hide all of this inside a Gradle plugin and apply it to kotlin-native
project.
- Unified and optimized visitor instead of transformer with many
redundant operations.
- `LLFirAnnotationArgumentsTargetResolver` now has the correct state
keeper for nested declarations – e.g., property accessors or value
parameters.
- `FirAllLazyAnnotationCalculatorVisitor` in tests now almost the same
as in the production as now it uses the same visitor without recursive
visiting
^KT-63606 Fixed
^KT-64000 Fixed
^KT-64579 Fixed
We don't have to resolve `FirFileAnnotationsContainer` before each
declaration.
This became possible after KT-65345 as now `FirDesignation` has `FirFile`
in its path so, `PersistenceContextCollector#collectContext` resolves
it as well as we don't have problems with unresolved annotations
during diagnostics processing
^KT-65876
Checker for conflicting declarations will now check for the following
scenarios too
- two expect declarations in different modules
- actual and non-expect declarations in different modules
^KT-63826 Fixed
Code fragment is a class that extends from Any, so technically it can
have fake overrides for equals/hashCode/toString. However, K1 does not
build them, and it's already checked in IR text tests in
`analysis/analysis-api/testData/components/compilerFacility/compilation/codeFragments`,
so don't build them in K2 as well. Also, the code in
`KtFirCompilerFacility.createJvmIrCodegenFactory` looks for a single
function defined in the code fragment.
This change fixes FIR tests with JVM IR serialization enabled, some of
which start to fail after enabling IR fake override builder mode by
default.
In this case, even accessing the function body with `IrFunction.body`
does not work in some cases because it tries to load IR and ends up with
unbound symbols for some reason (most likely related to KT-63509). So we
avoid accessing the body by copying the `IrFunction.acceptChildren`
implementation and removing the `body.accept` call.
Using deep-copy to create a fake override is questionable here, but it
will need to be refactored separately.
Yahor Berdnikau