Local lateinit var differs in behavior from a simple local var,
and logic that relies of 'instanceof Local' checks assuming that all
instances of Local are simple local vars can produce faulty code
(as in KT-23260, where a local lateinit var was not explicitly put on
stack when passed as an argument to an inline function, thus causing
null propagation).
#KT-23260 Fixed Target versions 1.2.50
* add separate run function for each test class to simplify and deduplicate generated code
* move the code that process IGNORE_BACKEND directive to the separate function, outside of generated code.
** it reduces size and complexity of a generated code
** it allows to mute and unmute tests w/o regenerate tests
* add an ability to generate IGNORE_BACKEND directive automatically, it's disabled by default
* add an ability to remove IGNORE_BACKEND directive automatically, it's enabled by default
* remove whitelists
Hierarchy of specially wrapped suspend function may have inconsistent
overridden descriptors: some of them may be wrapped, some unwrapped.
To mitigate this problem, currently it's proposed to render only
unwrapped functions, which are always have stable overridden descriptors
#KT-23809 Fixed
There was an exponential complexity in case we have a chain
of nested LexicalScopeWrapper instances:
we were walking through the chain and on each step running
the algorithm recursively.
Such a case is possible in scripts where each line contains LexicalScopeWrapper
The fix looks safe to me because delegate was used there
for checking !is ImportingScope && !is LexicalChainedScope
In all other means, being recursively run the algorithm does the same job.
#KT-22740 Fixed
Generate continuation type as kotlin.coroutines.Continuaion. This code will
fail at runtime since there is no stdlib backing this change yet.
However, in order to generate compatible stdlib we need a compiler, which
generates continuation type as kotlin.coroutines.Continuation.
Thus, firstly we support the change in the compiler, make it bootstrap
compiler and only then change stdlib and tests accordingly.
#KT-23362
This can happen e.g. if all supertypes were on the cycle (A : B, B : A)
-- then SupertypesLoopChecker will cut all supertypes that are on cycle,
leaving set of supertypes empty.
Some clients in the compiler are not ready to work with empty list of
supertypes -- notably, CommonSupertypes.
^KT-23029 Fixed
Private suspend functions need to be generated as package-local, since
they are called from their continuations.
However, this means that they can be called from Java, which breaks
their private visibility.
Adding synthetic to them fixes the issue.
#KT-17584: Fixed
- Do not longer call hashCode on java/lang/Object but do this
on statically checked type.
- Do not longer call Intrinsics.areEqual but use directly
equals method on statically checked type.
Fix of https://youtrack.jetbrains.com/issue/KT-12330
Treat them like lambdas. This means:
1) Like local suspend lambdas, which become "non-suspend" after creating
jvmSuspendView, they do this as well
2) They both are generated with the same codegen.
3) They are treated differently only on front-end side.
#KT-20364: Fixed
Writing a Jigsaw-modular Kotlin program which doesn't require
kotlin.stdlib doesn't make sense because it most likely will fail at
runtime, on access to anything from the standard library. Previously, we
checked that kotlin.stdlib was in the module graph, but that's not
enough, we should also check that the source module requires it.
'-Xallow-kotlin-package' can be used to disable the error.
Add a test checking that an indirect (transitive) dependency is also OK
The standard way of loading resources with built-ins metadata from the
current class loader no longer works in the modular mode on Java 9
because the current class loader is for module 'kotlin.reflect', but the
metadata is located in the module 'kotlin.stdlib'. On Java 9, we now use
the class loader of 'kotlin.stdlib' to load these resources.
#KT-21266 Fixed
Using the new multi-release jar feature, store compiled
module-info.class files into META-INF/versions/9 instead of the artifact
root. Hopefully, this will break fewer tools which do not support
module-info.class files because any sane tool should not do anything
with files in META-INF because before Java 9 that directory only
contained resources.
Upgrade some Maven plugins to newer versions which do not fail on
module-info.class files
#KT-21266 In Progress
Sometimes instead of {POP, GETSTATIC Unit.INSTANCE, ARETURN} sequence
the codegen emits {CHECKCAST Unit, ARETURN} sequence, which breaks tail
call optimization. By replacing CHECKCAST with ARETURN we eliminate
this issue.
#KT-19790: Fixed
Consider call chain `foo.bar.baz()` which has expected type `Int`
Elements `foo` and `bar` doesn't have expected type `Int`,
this is especially important in NI, because we use expected type for
candidate applicability
Before this change, the codegen used to generate beforeInlineMarker
in order to spill variables later. However, since the tailrec call
is replaced with jump, no afterInlineMarker is generated, leading to
inconsistency of these markers. Thus, the spilling fails.
This change disables generating of beforeInlineMarker.
#KT-21521: Fixed
Unlike ordinary lambdas suspend lambdas has create method which invokes
the constructor of the lambda object (continuation).
The inliner could not cope with this.
The previous change fixed the case of suspend lambda without receiver.
This change adds support of suspend lambdas with receiver.
#KT-21605: Fixed
Alexander Udalov