There's no point in using `ReflectProperties.lazy` over
`lazy(PUBLICATION)`, especially because of the subtle problems the
former can create, such as KT-27585. This is inspired by the fix in
99b38ccb74, although in this case I couldn't come up with an example
that would lead to a runtime exception.
This task will also automatically support the use case of
a 'carrier' task for the IDE. The IDE can use this task
to execute generic main methods, providing a -P<taskName>.mainClass
property
^KT-58661 Verification Pending
Using 'whenEvaluated' (aka 'afterEvaluate') is not supported
when using lazy APIs (like register). In order to safely
copy user configuration in both use-cases,
'launchInStage(AfterFinaliseDsl)' is used as this will be safe
either way.
^KT-58838 Verification Pending
If `-Xno-unified-null-checks` is specified, do not throw exceptions, but
rather revert to the old behavior, with source code in exception
messages.
Note that `-Xno-unified-null-checks` is a flag which enables behavior of
Kotlin 1.3. It can be removed once we no longer support API version 1.3.
For type parameters of generic properties,
`DeclarationDescriptor#getContainingDeclaration` must return
the property descriptor instead of the accessor function descriptor.
^KT-57436 Fixed
Change the public-facing KtFirDiagnostic to be an interface, not
a class, and change the internal implementations to be classes, not
interfaces. This allows hiding access to the KtPsiDiagnostic underlying
the AA diagnostic class.
Normally during metadata compilation we use
`K2MetadataCompiler` and it serializes
metadata to klibs without generating IR.
Native shared compilation uses the K2Native
compiler, which generates klibs based on both
the fir and fir2ir information.
It would be nice to reuse `K2MetadataCompiler`
for native shared, but currently it has no
support for the commonizer, and also reads
the stdlib metadata from a jar instead of a
klib, and there's no simple idiomatic way
to pass the `distribution.klib` value down
to `KLibResolverHelper` (native compilation
has some commonized dependencies, and they
explicitly depend on `stdlib`, so we must
resolve it this way or strip it away to be
able to use the metadata jar instead). There
may be other issues, I only digged a bit.
Instead, this change modifies the K2Native
in such a way that it can generate klibs
with metadata without fir2ir info.
Note that this change does not fix KT-58139,
because KT-58139 fails due to performing the
constants evaluation directly during the
metadata serialization when the needed info
does (and, as I was told, should) not exist
(namely, valid property initializers).
^KT-58444 Fixed
Review: https://jetbrains.team/p/kt/reviews/9967
1. Fix weird indentation in ValDeferredInit*
2. Use meaningful names in ValDeferredInit*
3. Drop leading indentation in VarDeferredInit*. Not everyone has a wide
monitor
4. Add some NOTICABLE comment that VarDeferredInit* test should be read
as table
Review: https://jetbrains.team/p/kt/reviews/9967
This commit fixes:
Cause 2: java.lang.IllegalStateException: Not reachable case. We can always suggest making `open val` property `final`
at org.jetbrains.kotlin.resolve.DeclarationsChecker.reportMustBeInitialized(DeclarationsChecker.kt:845)
at org.jetbrains.kotlin.resolve.DeclarationsChecker.checkPropertyInitializer(DeclarationsChecker.kt:778)
at org.jetbrains.kotlin.resolve.DeclarationsChecker.checkProperty(DeclarationsChecker.kt:614)
at org.jetbrains.kotlin.resolve.DeclarationsChecker.process(DeclarationsChecker.kt:106)
at org.jetbrains.kotlin.resolve.BodyResolver.resolveBodies(BodyResolver.java:258)
Reproducible in K1 & K2
From user point of view it's an improvement in compilation message.
From technical point of view it's an introduction of new compilation
diagnostic.
Review: https://jetbrains.team/p/kt/reviews/9967
I'm going to deprecate `open val` case in the next few commits KT-57553.
But it is always possible to suggest using `final` for `open val` case.
^KT-58346 Fixed
Review: https://jetbrains.team/p/kt/reviews/9967
BTW this commit accidentaly and partially fixes KT-57553 for K2, because
of a cleaner K2 architecture. I will unify MUST_BE_INITIALIZED behaviour
in K1 and K2 in the next commits
Review: https://jetbrains.team/p/kt/reviews/9967
I'm going to change the logic in this area in the next few commits
KT-57553. So let's cover the current behaviour with tests.
*.fir.kt tests are not properly formatted because of K1 and K2 different
behaviour. I will fix it in the next commits
Sergej Jaskiewicz