- 21 1月, 2021 20 次提交
-
-
由 Mikhail Glukhikh 提交于
-
由 Jinseong Jeon 提交于
and fix CONFLICTING_OVERLOADS to use it
-
由 Igor Yakovlev 提交于
[FIR] Fix invalid diagnostic fir node sites and improved invalid type parameters count diagnostic report
-
由 Simon Ogorodnik 提交于
-
由 Igor Yakovlev 提交于
-
由 Ilmir Usmanov 提交于
accesses it. #KT-44131 Fixed
-
由 Mikhael Bogdanov 提交于
Generate linenumber for inline call site in case of implicit iterator/hasNext/next calls in for loop Otherwise -1 is generated in smap
-
由 Sergey Igushkin 提交于
This leads to cleaner error messages in Gradle variant-aware dependency resolution failures. Gradle has deprecated those configurations since long ago, and we didn't see much use of them as variant-aware dependency resolution entities either. So this commits sets `canBeConsumed` to false on these configuratons: * compile (+ testCompile, fooCompile in MPP) * runtime (+ testRuntime, fooRuntime, ...) * compileOnly (+ testCompileOnly, fooCompileOnly, ...) * default (+ fooDefault in MPP) This change replaces the PR #3995
-
由 Sergey Igushkin 提交于
-
由 Sergey Igushkin 提交于
-
由 Sergey Igushkin 提交于
When dependencies are added to our corresponding configurations, report errors, as there's no proper way to add those dependencies. Issue #KT-43944 Fixed
-
由 Andrei Klunnyi 提交于
-
由 Dmitriy Novozhilov 提交于
- argument type is flexible - supertype has flexible type argument - type of expression is more specific than bare type
-
由 Jinseong Jeon 提交于
Also, update error inputs to conform to old FE's These should have been part of commit 39df3e2b
-
由 Jinseong Jeon 提交于
Also, update error inputs to conform to old FE's This should have been part of commit 03cb0c3c
-
由 Jinseong Jeon 提交于
Also, update error inputs to conform to old FE's These should have been part of commit 5167d69b
-
由 Ilmir Usmanov 提交于
parameter, since in this case, the compiler generates a bridge, where the result is unboxed.
-
由 Ilmir Usmanov 提交于
since the bridge unboxes it. #KT-44141 Fixed
-
由 Ilya Goncharov 提交于
-
由 Steven Schäfer 提交于
When resolving inline class methods in binary dependencies we look for methods matching both the new and the old mangling scheme. On the IR backend the method signature has to be computed for the inline class replacement, since the logic for signature mangling is not contained in the method signature mapping, unlike in the old backend.
-
- 20 1月, 2021 10 次提交
-
-
由 Mikhael Bogdanov 提交于
-
由 Mikhael Bogdanov 提交于
-
由 Mikhael Bogdanov 提交于
#KT-44429 Fixed
-
由 Ilya Goncharov 提交于
^KT-42176 fixed
-
由 Pavel Kirpichenkov 提交于
Take local/anonymous classes and their inner classes into account. Simplify approximation: use first available super classifier instead of first super class. This approximation should only happen for private declarations that were not previously approximated by frontend. So basically the only requirement for the approximated types is to be denotable. Note that this only works if the types are not used later. JVM uses a different string table implementatin as it needs exact types of private members for reflection. ^KT-20996 Fixed
-
由 Pavel Kirpichenkov 提交于
They used to use the same approximation logic anyway ^KT-20996
-
由 Pavel Kirpichenkov 提交于
Extract the logic of approximating to denotable class supertype from JS/KLIB and use it in metadata compiler. ^KT-20996 In Progress
-
由 Pavel Kirpichenkov 提交于
-
由 Ilmir Usmanov 提交于
#KT-44140 Fixed
-
由 Sergey Shanshin 提交于
Fixes Kotlin/kotlinx.serialization#1265
-
- 19 1月, 2021 9 次提交
-
-
由 Ilmir Usmanov 提交于
-
由 Ilmir Usmanov 提交于
Since in 1.5 JVM_IR is enabled by default and supports suspend functions in fun interfaces, it is safe to enable the feature by default as well.
-
由 Yaroslav Chernyshev 提交于
-
由 Svyatoslav Kuzmich 提交于
Dynamically calculating a year is problematic, especially in cases like FIR where generation happens in every build. When you are working on a last year's commit, each build creates a bunch of uncommited changes of updating the year in copyrights.
-
由 Svyatoslav Kuzmich 提交于
-
由 Svyatoslav Kuzmich 提交于
-
由 Ilya Kirillov 提交于
This list is backed by mutable list and during iterating on it we can resolve supertypes of that class via IDE light classes as IJ Java resolve may resolve a lot of stuff by fir light classes This causes ConcurrentModificationException
-
由 Vyacheslav Gerasimov 提交于
Until `KTI-464 Test tasks with failed tests are cached` is resolved
-
由 Sergey Bogolepov 提交于
Because it is more intuitive for end-users.
-
- 18 1月, 2021 1 次提交
-
-
由 pyos 提交于
Type parameters do not necessarily match one-to-one, or preserve order.
-