azalea/kotlin-fork
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
83,829 Commits 1 Branch 0 Tags
8317daa00ba55c07e44e67f3e3e010616c981967
Commit Graph

113 Commits

Author SHA1 Message Date
Ilmir Usmanov c4ddf3530d Use GETFIELD instead of unbox-impl inside inline classes 
#KT-36783 Fixed
 2021-05-12 06:50:06 +00:00
Steven Schäfer 84d78035e4 JVM IR: Avoid double mangling of function reference invoke methods 2021-04-29 13:10:43 +02:00
Mikhael Bogdanov 3d8e8dd3ba Fail on compilation errors in AbstractBytecodeTextTest 2021-02-17 18:43:23 +01:00
Mikhail Glukhikh 465261e611 FIR: don't create intersection override for duplicating symbols 2021-02-02 16:25:08 +03:00
Alexander Udalov d022bb0248 Switch default JVM target to 1.8 
#KT-29405 Fixed
 2021-02-01 11:54:04 +01:00
Georgy Bronnikov 076272f7ca FIR2IR: avoid descriptors computing hashCode 
When synthesizing the hashCode function for data classes, descriptors
were used, in partcular, memberScope for primitive classes.
IrBasedDescriptors have no member scope, so we compute the hashCode
function based on IR structures.
 2020-12-23 18:50:52 +03:00
Ilmir Usmanov 0dc5f3ac00 IC: call JvmDefault method of inline class using boxed receiver 
#KT-43698 Fixed
 2020-12-08 04:33:36 +01:00
Dmitry Petrov e6a3e38c4d JVM_IR no static inline class members for Kotlin JvmDefault methods 
KT-43698 KT-43051
 2020-12-02 20:04:13 +03:00
Mikhail Glukhikh 1c71e64f58 [FIR] Create string interpolating call even for single argument 
Before this commit, questionable optimization existed which
unwrapped string interpolating call with single argument to this argument.
However, this led to source element loss and the necessity of sub-hacks.
In this commit we dropped this optimization (anyway user can remove
this single-expression string template in code if needed) to keep
source elements intact.
 2020-11-26 08:37:50 +03:00
Ilmir Usmanov b33774e5f2 IC mangling: Use empty list as a separator in the new mangling scheme 2020-11-19 17:39:28 +01:00
Ilmir Usmanov 488d4ab018 IC mangling: Use '_' instead of 'x' as a placeholder before hashing 2020-11-19 17:39:27 +01:00
Ilmir Usmanov c62093f54c IC mangling: Change mangling rules 
1. Use 'x' for each parameter, which is not an inline class, every
possible clash is handled by signature rather than name. This change
makes more API changes binary-compatible. So, the changes are in line
with the vision of inline classes are value classes, like primitives.

2. Take return type into account when mangling a function if the return
type is inline class. Otherwise, boxing bridge will not be generated,
which leads to CCE at runtime.
 2020-11-19 17:39:24 +01:00
Ilmir Usmanov 7b315a8b52 JVM_IR: Do not box inline class in methods 2020-11-02 10:50:00 +01:00
Ilmir Usmanov 28e148d35a Old JVM: support init blocks in inline classes 
#KT-28055 Fixed
 2020-10-29 20:38:03 +01:00
Mark Punzalan 238cc7c257 [FIR] Enable BytecodeText tests for FIR. 
143 out of 767 tests (18.6%) are currently failing.
 2020-09-29 10:21:21 +03:00
pyos bd6ead0467 JVM_IR: check for null when converting unboxed inline classes to strings 
`C?` can be unboxed into `T?` if if wraps a reference type `T!!`, but in
this case `null` is not a valid value to pass to `toString-impl`.

 #KT-42005 Fixed
 2020-09-25 16:50:34 +02:00
Steven Schäfer 78ab957bfe JVM IR: Fix mangling for suspend functions with inline class params (KT-41374) 2020-08-31 18:43:29 +02:00
Steven Schäfer 097e0fb46c JVM IR: Mangle delegated properties with inline class type (KT-40125) 2020-07-10 17:06:44 +02:00
Dmitry Petrov 202bbdf8dd Forward compatibility hacks for Result.{success, failure} 
Don't mangled functions annotated with @JvmName.
Annotate 'Result.success' and 'Result.failure' with @JvmName and
@Suppress("INAPPLICABLE_JVM_NAME").
NB this would require bootstrap.
 2020-06-04 12:16:27 +03:00
Steven Schäfer 7ea71a17f0 JVM IR: Use language feature for inline class mangling rules 2020-05-29 19:54:09 +03:00
Dmitry Petrov a270ee094c Language feature for new inline class mangling rules (since 1.4) 2020-05-29 00:53:01 +03:00
Steven Schäfer dc0ef996b7 JVM IR: Implement the new inline class ABI 2020-05-28 18:00:35 +03:00
Dmitry Petrov c165b8d55c JVM: Update IR bytecode text and signature tests 2020-05-20 07:19:30 +03:00
Dmitry Petrov cf70c83ab7 JVM: Update tests 2020-05-20 07:19:29 +03:00
Dmitry Petrov 9615b20e5d Box inline class return value in lambdas (including suspend lambdas) 2020-04-08 19:56:26 +03:00
Steven Schäfer 6a1e35389c JVM IR: Mangle interface implementation methods in inline classes 2020-03-16 18:48:16 +01:00
Steven Schäfer 465e9f2d68 JVM IR: Always produce stubs for interface methods in inline classes 
This matches the behavior of the JVM backend, with the exception of
@JvmDefault methods, which are currently broken on the JVM backend.
 2020-02-28 14:48:17 +01:00
Dmitry Petrov 94d20d9176 Update bytecode text tests in JVM_IR 2020-02-17 23:52:48 +03:00
Steven Schäfer 5760c0be9c JVM IR: Avoid redundant coercions in ExpressionCodegen 2020-02-15 22:32:23 +03:00
Alexander Udalov 6a90dc2efe Fix bytecode text tests for language version 1.4 
New null check assertions are generated a bit more efficiently, with a
call to `checkNotNull` instead of IFNONNULL+jump.
 2020-01-20 19:12:59 +01:00
Mikhail Zarechenskiy 5c5635ce20 Fix codegen & bytecode tests after unifying exceptions in JVM backend 
See KT-22275 for details
 2020-01-20 16:36:03 +03:00
Alexander Udalov e9946b21b5 JVM IR: do not hide constructor with inline class types and defaults 
This corresponds to the ABI generated by the old backend. Moreover, when
compiling the module 'ir.tree' with JVM IR (could not reproduce on a
small sample), this led to a codegen crash when trying to reassign
parameter value to the default stub, which is an error expression and
can't be generated by ExpressionCodegen.
 2019-12-02 10:19:55 +01:00
Steven Schäfer c905209504 JVM IR: Lower IrStringConcatenation 2019-10-18 17:24:53 +02:00
Steven Schäfer 21af7dfbe1 Add tests for constructors taking inline class arguments 2019-10-18 17:20:22 +02:00
pyos 06c00f4d9e Unmute some JVM_IR inlining tests 2019-10-08 17:19:41 +02:00
Steven Schäfer ce3ef4e4d0 Add more tests for inline class equality with nullable arguments 2019-09-18 18:52:58 +02:00
Steven Schäfer 49efa5fbc4 Split up Result API boxing tests 2019-09-18 18:52:58 +02:00
Steven Schäfer b85b2d9af8 Add more tests for inline class equality 2019-09-18 18:52:58 +02:00
Steven Schäfer cdc5e1347b JVM: Avoid boxing in inline class equality 2019-09-18 18:52:58 +02:00
Steven Schäfer f53b28e8a3 JVM: Generate equals-impl0 method for inline classes 2019-09-18 18:52:58 +02:00
Steven Schäfer 475079611d JVM IR: Avoid parameter null-checks in inline class impl methods 2019-09-18 18:52:58 +02:00
Steven Schäfer cdd3f82396 JVM IR: Use correct scope owner in JvmInlineClassLowering 2019-09-18 18:52:58 +02:00
Dmitry Petrov 7d49b72b79 JVM_IR: Support 'CHECK_NOT_NULL' intrinsic 
TODO: migrate to changes related to KT-22275
 2019-08-14 11:16:10 +03:00
Steven Schäfer ac667403ef (Un)mute and add tests for vararg codegen 2019-08-13 14:24:55 +02:00
Alexander Udalov 2baddb029c Use Intrinsics.checkNotNullParameter to throw NPE in parameter null checks 
Similarly to previous commits, this method was unused, so we're changing
its semantics in API version >= 1.4.

 #KT-22275 In Progress
 2019-08-12 16:09:23 +02:00
Steven Schäfer 9182fe887e Fix inline class handling in DefaultParameterInjector 2019-08-05 16:48:18 +03:00
Steven Schäfer 79f71f61d5 JVM BE: Add more tests for default arguments 
- Inline class parameters should not be boxed.
- Empty vararg parameters should not allocate an empty array.
 2019-08-05 16:48:18 +03:00
Steven Schäfer 078ccbf077 JVM_IR: Implement name mangling exceptions for Result class. 2019-07-01 15:49:12 +02:00
Steven Schäfer 917ef250cf Add and (un)mute inline class tests 2019-06-13 12:25:06 +02:00
pyos 6b2d874ccc JVM_IR: generate non-null assertions for arguments 2019-05-21 18:30:27 +03:00