Using `ReflectionProperties.lazy` is incorrect because it allows several
threads to observe different resulting values if they're computing it
simultaneously (unlike `lazy(PUBLICATION)`, which always returns the
value that "won the race").
In the case of property delegates, for example, if we're invoking
`isAccessible = true` and then `getDelegate()` concurrently, it might
happen that when some thread invokes `getDelegate()`, it gets the
underlying Field object which was written by another thread and which
has not yet been made accessible, leading to
IllegalPropertyDelegateAccessException.
#KT-27585 Fixed
Also minor cleanup. Remove the comment about the issue jansi#35 because
although the issue is fixed, the behavior is correct right now: we
enable colors by default iff stderr is a TTY (and the platform is not
Windows), and to determine that we need to call `CLibrary.isatty`.
#KT-55784
The file fingerprints are going to be used for
a cache invalidation in JS IR incremental cache.
It should reduce an incremental rebuild time in JS BE.
CityHash128 is used as IrFile fingerprint.
Before, the only way of getting analyzable elements was to create an
analyzable file by using 'createAnalyzableFile()'. So made all utilities
available in 'KtPsiFactory' useless as they delegate to 'createFile()'
that always set the 'doNotAnalyze' flag.
The new behavior is to pass the 'analysisContext' instead if it is
passed to the 'KtPsiFactory' constructor.
The newly appeared API is going to be used in the Kotlin's UAST
implementation.
Review: https://jetbrains.team/p/kt/reviews/7418
`State` enum contains `DISABLED` item which is redundant for
`LanguageFeature`. It's redundant because `sinceVersion = null` is used
to express the same semantic.
I had to reorder some items in the enum because otherwise
`testLanguageFeatureOrder` fails. I could keep the same logic in
`testLanguageFeatureOrder` but I would need to make
`isEnabledWithWarning` public which I don't want to. Anyway, it's not
obvious what is more readable: keep `isEnabledWithWarning = true` items
at the end or at the beginning.
Also all effectively disabled features (`sinceVersion = null` or
`defaultState = State.DISABLED`) are now moved to the "Experimental"
section at the end which is obviously better.
Since KT-53308, we started to cache results of typeOf invocation
in reflection. The cache uses the origin 'KClassifier' as a key with
an optional list of 'KTypeProjection' in type argument position,
and computed result as a value.
Without classifier check, the caching produces incorrect execution in
a very specific classloaders usage which is leveraged
by IDEA source-tree:
* All Kotlin stdlib and reflect classes are always loaded by the same
classloader
* Other classes that depend on Kotlin are loaded by separate
classloaders
The reproducer:
* Attempt to use typeOf<kotlin.List<Foo>> from one classloader caches
the resulting KType
* Attempt to use typeOf<kotlin.List<Foo>> from another classloader
for the same 'Foo' that differs only in classloader reuses the
computed KType, meaning that KType type argument classifier
is simply incorrect and points to the 'Foo' from the first classloader
#KT-54611 Fixed
#KT-54629 Fixed
Merge-request: KT-MR-7470
Merged-by: Vsevolod Tolstopyatov <qwwdfsad@gmail.com>
[JS] Remove incremental js jps test
[JS] Remove test of maven js archetype
[JS] Fix another one test
[JS] Fix tests for jps
[JS] Try to fix maven test data
[JS] Fix test data
[JS] Fix test data for ant
[JS] Fix jsExtraHelp test
[JS] Fix test run from not-Gradle build tools
[JS] Set flag without compiler error
[JS] Disable warnings and errors in legacy compiler called from Gradle
[JS] Proofread messages
[JS] Not proofread messages
KT-42326
Now it could be only be applied inside settings file pluginManagement
block. This will update both settings repositories and all projects in
the repo with cache redirection settings.
- just legacy - report warning about deprecation
- both - report warning about deprecation of legacy
- no compiler explicitly chosen - error about explicit setting compiler
warn from cli legacy compiler
nowarn flag - kotlin.js.compiler.nowarn
KT-42326
KT-53074
This is a partial revert of commit bc1c2ada
We have to use old type intersector for emptiness check because
KT-53656 were found. We return to old type intersector usage also
for type parameter bounds to prevent other possible problems.
* Replace pcollections with ClassValue/ConcurrentHashMap-based caches
* Do not store weak references, instead cache strong references and count on ClassValue to unload the corresponding classloader if necessary
* ConcurrentHashMap does not rely on WeakReference as it's only selected on Android where classloader leaks don't exist
* Update reflect/scripting JDK requirement to Java 8 in order to proceed
#KT-53454
#KT-50705
Merge-request: KT-MR-6788
Merged-by: Vsevolod Tolstopyatov <qwwdfsad@gmail.com>
discrepancy after switching to the new REPL internals was
discovered also on IJ plugin tests, so added a test emulating it.
also some test infra refactoring
Ilya Chernikov