- 19 6月, 2020 16 次提交
-
-
由 Jinseong Jeon 提交于
-
由 Jinseong Jeon 提交于
-
由 Ilya Kirillov 提交于
#KT-39700 fixed
-
由 Ilya Kirillov 提交于
-
由 Ilya Kirillov 提交于
-
由 Abduqodiri Qurbonzoda 提交于
-
由 Abduqodiri Qurbonzoda 提交于
-
由 Abduqodiri Qurbonzoda 提交于
-
由 Toshiaki Kameyama 提交于
#KT-39200 Fixed
-
由 Abduqodiri Qurbonzoda 提交于
-
由 Abduqodiri Qurbonzoda 提交于
-
由 Abduqodiri Qurbonzoda 提交于
-
由 Abduqodiri Qurbonzoda 提交于
-
由 simon.ogorodnik 提交于
-
由 Vladimir Ilmov 提交于
-
由 Vyacheslav Gerasimov 提交于
-
- 18 6月, 2020 24 次提交
-
-
由 Ilmir Usmanov 提交于
-
由 Ilya Goncharov 提交于
#KT-38331 fixed
-
由 Ilya Goncharov 提交于
#KT-38331 fixed
-
由 Ilya Goncharov 提交于
#KT-38331 fixed
-
由 Ilya Goncharov 提交于
#KT-38331 fixed
-
由 Ilya Gorbunov 提交于
Leave these methods as hidden in order to preserve binary compatibility. #KT-39504
-
由 Pavel Kirpichenkov 提交于
Without forced clean stale references to declarations from source-dependent libraries can be accessed after invalidation, i.e. resolved references can point to incorrect PSI. TBD: more granular invalidation. KT-24309
-
由 Alexander Udalov 提交于
-
由 Ilya Matveev 提交于
This commit allows parallel in-process execution of the K/N compiler that was prohibited by 254a978a. Issue #KT-38991 fixed
-
由 Ilya Matveev 提交于
Calling FileSystems.newFileSystem(URI, ...) throws a FileSystemAlreadyExistsException if a ZipFileSystem for this URI is already created. We still can use a single instance of ZipFileSystem by calling FileSystems.getFileSystem. In this case we use reference counting to determine when this instance can be safely closed. But we cannot count references if the same ZipFileSystem is used from different class loaders. This patch fixes this issue by creating a file system from Path instead of an URI. Contract of FileSystemProvider.newFileSystem(Path, ...) doesn't imply throwing FileSystemAlreadyExistsException. Issue #KT-37443 fixed
-
由 Natalia Selezneva 提交于
-
由 Natalia Selezneva 提交于
Wrong gradle version when wrapper is used fixed in 201
-
由 Natalia Selezneva 提交于
GradleProjectSettings that were used, are updated after KotlinDslListener is called, so it isn't correct to check if kotlinDslModels are supported using it. Also GradleScriptDefinitionsContributor should use gradle home from BuildModel, not from settings for the same reasons. ^KT-39104 Fixed
-
由 Victor Petukhov 提交于
^KT-39660 Fixed
-
由 Vyacheslav Gerasimov 提交于
-
由 Igor Chevdar 提交于
-
由 Yan Zhulanow 提交于
Review KT-CR-1301.
-
由 Yan Zhulanow 提交于
-
由 Pavel Kirpichenkov 提交于
Resolution facade should be used consistently with direct usages of frontend components. Otherwise they can start processing descriptors from foreign resolvers which leads to memory leaks. Plain resolution API with provided facade is not suitable as-is for compiled declarations in KotlinIndicesHelper though. Resolution facade for module sources contained in helper can't handle decompiled sources from PSI indices (leads to "ModuleInfo not contained in resolver" errors). That's why "hacked" resolve via import references should be used there. #KT-39642 Fixed
-
由 Jinseong Jeon 提交于
-
由 Mikhail Glukhikh 提交于
Recently second javaType function appeared in reflect and now FIR can't correctly resolve the code :(. Will be fixed after annotation deserialization full support
-
由 Mikhail Glukhikh 提交于
This reverts commit 3768af4f
-
由 Mikhail Glukhikh 提交于
Recently second javaType function appeared in reflect and now FIR can't correctly resolve the code :(. Will be fixed after annotation deserialization full support
-
由 Mikhail Zarechenskiy 提交于
The issue was fixed in df1595e4 #KT-39630 Fixed
-