LLVMConstGEP2 and LLVMConstInBoundsGEP2 were forward declared in
LLVM-11, but not implemented until LLVM-14. This patch adds these along
with the missing LLVMAddAlias2. All three implementations are copied
from llvm branch release/14.x
We don't have "expect as libraries" feature yet. But once we have the
feature, the code in question might start silently behave incorrectly.
Add assertions to make sure that the code will fail
Previously, the validation was that the port is within the [1, 65535) range. Considering the defined range [17001, 18000) for the random port selector, it makes more sense to check that range instead.
A few messages were improved:
* If the daemon process died at the startup, add its last 10 lines of the output to the log and report to the EXCEPTION (ERROR in the case of Gradle) level
* If some exception occurs during connection attempt, add its stacktrace and report to the EXCEPTION (ERROR in the case of Gradle) level
* Added more DEBUG level messages
* Some important messages are moved to the INFO level
* Added a suggestion to report issue to kotl.in/issue
^KT-55322 Fixed
Before this change, some long-running tasks like `run` might cause the Kotlin daemon to go into the LastSession state and stay in that state until the long-running tasks and therefore the Gradle build are finished.
The reason for that behaviour is that if we don't release sessions explicitly, the created "session is alive" marker files are cleaned up by `KotlinGradleFinishBuildHandler.buildFinished` at the end of the build.
^KT-55322 In Progress
Before this change, the logic was to find the most suitable daemon and try connecting to it in a loop, ignoring the fact that it could be dying. Now, instead of trying to connect to the daemon dying daemon in a loop, we will make only 1 attemp and then skip if it's already dying.
^KT-55322 In Progress
Review: https://jetbrains.team/p/kt/reviews/13334/timeline
Thanks to the previous commit, it's now possible to run
DefaultArgumentsInExpectActualizedByFakeOverride on both: frontend (FIR)
and backend (IR).
We aim to perform a thorough examination of checks involving
fake-overrides on both FIR and IR, given their distinct implementations
for handling fake-overrides.
The commit decreases scope of influence of hacky
`shouldCheckDefaultParams` flag.
^KT-63860 Fixed
Review: https://jetbrains.team/p/kt/reviews/13334/timeline
The previous code was nonsense (I wrote it). It doesn't make sense to
subtract actualOverriddenDeclarations from expectOverriddenDeclarations.
Default parameters are mentioned on the expect side. So default params
in expect/actual supertypes won't be subtracted from
expectOverriddenDeclarations (but should be)
There were a number of locals where components of InitialConstraint were
named lower and upper in the FirBuilderInferenceSession.
That is not true for equality constraints, so such naming should be
avoided to avoid misconception about the nature of type relation encoded
in constraint.
^KT-64031 Fixed
It's now impossible to add new classes to metadata, because this
functionality is not implemented in IrGeneratedDeclarationsRegistrar
and FirDeclarationsForMetadataProviderExtension is removed
`IrGeneratedDeclarationsRegistrar` assumes that all generated functions
are correct from a Kotlin point of view. But `writeSelf` method on JVM
is a static method outside any object/companion object
So to properly calculate containing class for this method we should
generate a dispatch receiver parameter, register the method in metadata,
and then remove the parameter (to make function static)
The condition was added to the substitute function, when it was also
re-used in delegate inference.
However, delegate inference no longer uses this function.
It is not very possible to both have variable fixed and being present in
nonFixedToVariablesSubstitutor in builder inference.
^KT-64028 Fixed
Stanislav Ruban