- Composite providers may call `mayHaveSyntheticFunctionType` on *all*
of their children if *one* child has
`mayHaveSyntheticFunctionTypes == true`, so
`mayHaveSyntheticFunctionType` should have a proper default
implementation to avoid runtime errors.
- The alternative of requiring to check `mayHaveSyntheticFunctionTypes`
before calling `mayHaveSyntheticFunctionType` at the call-site was
considered, but deemed too error-prone. Given that
`mayHaveSyntheticFunctionType` is an optimization and admits false
positives, an imprecise default implementation of it is allowed.
Thanks to Mikhail Glukhikh for discovering this problem.
for `ACTUAL_ANNOTATIONS_NOT_MATCH_EXPECT` checker near to
other tests of this diagnostic. It is fine to move it out of
`multiplatformCompositeAnalysis` dir, because, actually, it tests
the case with comparison of FirAnnotationCall and FirAnnotationImpl,
which doesn't require composite analysis mode.
^KT-62559
...`ACTUAL_ANNOTATIONS_NOT_MATCH_EXPECT` checker near to
other tests of this diagnostic. This is because now such tests
are run in lazy resolve test suite, so no need to store them
separately.
^KT-62559
It is expected, that in some complex cases
`ACTUAL_ANNOTATIONS_NOT_MATCH_EXPECT` reports in .ll.kt disappeared
at all. This is because FIR checker doesn't catch comlex cases,
which are handled by IR checker.
Also add comments about bugs KT-62886 and KT-63382 found by tests.
^KT-62559
New tests are red in current commit!
All `.ll.kt` files are just copies of `.fir.kt` or `.kt`. This will be
fixed in the next commit to clearly show difference between LL and FIR.
The original purpose of this change was to test
`ACTUAL_ANNOTATIONS_NOT_MATCH_EXPECT` diagnostic, for which FIR checker
will be run only in IDE (to be changed in next MR), so its reports
a lost from usual KMP tests.
But this looks like a useful change for all KMP tests, not only this
diagnostic.
^KT-62559
Properly set up dom-api-compat dependency for JS IR tests. Since this dependency is added
automatically for every Kotlin/JS library, it should be present during tests just as stdlib.
As a result, tests for serializable enums were changed since 1.6.0 runtime does not require enums to be explicitly serializable.
A full mangled name of a declaration is a mangled name that includes
the mangled names of all the declaration's parents.
We don't use full mangled names for building `IdSignature`s. We use them
in two places:
- Generating stable names for JavaScript functions in Kotlin/JS
(see `NameTables.kt`)
- Generating names for LLVM symbols from lowered IR functions in
Kotlin/Native (see `BinaryInterface.kt`)
In both of these places we run the IR mangler, hence we don't need this
functionality in other manglers.
We plan to disable computing full mangled names of declarations in all
manglers except the IR mangler (see the subsequent commits).
From now on, in irText tests we dump only the following mangled names:
- Full mangled names computed using the IR mangler
- Signature mangled names computed using the Descriptor mangler
- Signature mangled names computed using the IR mangler
- Signature mangled names computed using the FIR mangler
Here by a full mangled name we mean the mangled name of a declaration
computed using the `MangleMode.FULL` mode. Those mangled names include
the mangled names of the declaration'a parents.
By a signature mangled name we mean the mangled name of a declaration
computed using the `MangleMode.SIGNATURE` mode.
These mangled names are used to compute an `IdSignature` for
the declaration, hence the name.
Add missing type propagation to getter and setter that was missed due
to copy-pasted logic.
`computeReturnTypeRef` should check the session to avoid infinitive
recursion
^KT-63361 Fixed
- While `collectionInitializationInfo` unwrapped a script's top-level
declarations, `check` forgot to do it, so a script effectively had no
top-level properties in the mind of the checker.
^KT-63286 fixed
- The test data is missing initialization checks and will be fixed with
the subsequent commit.
- The test data is originally taken from
`backingField/TopLevelPropertyInitialization` and has been simplified.
- I didn't add `const val` properties to the test data (compared to the
original) due to KT-20110.
^KT-63286
Marco Pennekamp