* [JVM_IR] Limit inner class attributes to types in class file
Inner class attributes should only be recorded for types that
are materialized in the result class file. In particular, we
should not emit inner classes attributes for types that appear
only in fake overrides. We do map these types to track the
fake overrides for JVM signature clashes but they are not
materialized in the class file.
^KT-56104 Fixed
* [JVM_IR] Consistently pass around materialized boolean in mapType.
The original name was still left in the ObjC code for cases
such as:
```
class A(@property:ObjCName("y") @ObjCName("y") val x: Int)
```
which compiled to:
```
@interface FrameworkA : FrameworkBase
- (instancetype)initWithY:(int32_t)x __attribute__((swift_name("init(y:)"))) __attribute__((objc_designated_initializer));
@property (readonly) int32_t y __attribute__((swift_name("y")));
@end
```
This change renames the `x` ocurrence in `initWithY` which is
important in case ObjCName is used to avoid using names that
conflict with ObjC preprocessor macros.
The phase to eliminate redundant function prologue safepoints
was written in Kotlin which meant that it was performing a lot
of jni calls. Moving the simple optimization to C code speeds
things up dramatically.
On a large project this reduces the time for this phase from
11.8 seconds to 1.4 seconds.
Create the field directly and maintain the fact that its corresponding
property is the companion object property. That property has the
deprecated annotation which puts the deprecated marker on the field
when code generating the field.
^KT-54840 Fixed
For non-suspend lambdas annotations are carried over to the
invoke method so that tooling can find the annotation there.
It seems reasonable that annotations are carried over to
the invokeSuspend method on suspend lambdas as well so that
similar tooling can be built and work for suspend lambdas.
They were compared based on the name of the file only disregarding
the module. Therefore, the test framework would only patch
package declarations for one file with a given name even if there
were multiple files with that name in different packages.
Adding -Xbundle-version and -Xbundle-short-version-string
compiler arguments. These can be used in the freeCompilerArgs
in the gradle configuration to control the version strings
in the Info.plist in generated frameworks.
```
version = "1.0"
macosX64("native") {
binaries.framework {
compilation.kotlinOptions.freeCompilerArgs += "-Xbundle-version=$version"
}
}
```
^KT-33117 Fixed.
On Android devices starting SDK 25, ThreadLocalRandom accidentally
started out from the same seed on all application starts. This
means that the same sequence of numbers is returned whenever the
application is started.
This will be fixed in SDK 34. We should avoid using
ThreadLocalRandom until then.
^KT-52618 Fixed
The benchmark constructs a splay tree, measures the time it takes
to performs a number of insertions and removals, and after
measurements validates that the splay tree is valid.
The body of the InlinedLocalFunctionDeclarationInstruction was
not copied. That confuses the information on the exceptional edge
with the information on the normal edge.
^KT-52131 Fixed
Locals introduced in the body of a do-while loop are not
necessarily live at the do-while condition. For example,
if there is a continue in the body before the declaration:
do {
if (shouldContinue(x))
continue
val y = 32 // not always defined in the condition
doSomething(y)
} while (x < 2)
For locals referenced in the condition such code is rejected
by the frontend because a local referenced in the condition
must be always defined when you get there.
However, locals that are not used in the condition were always
put in the local variable table. This leads to invalid locals
information which can trip of debuggers and other build tools
such as the D8 dexer.
This patch only puts in locals information for locals actually
referenced in the local variable table for the condition.
^Fixes KT-51754
If no line number is emitted, the line number of the last line
in the suspend function is used. That combined with the IntelliJ
debuggers handling of breakpoint setting and the D8 dexers
reordering of blocks can lead to cases where breakpoints on
the last line of a suspend function are never hit.
See https://youtrack.jetbrains.com/issue/KT-51936 for details.
^Fixed KT-51936
^KT-51627 Fixed
kotlinc will currently read /tmp/build.txt in an attempt to
determine an IDEA version number for IDEA plugin compatibility
checking. If that file exists and does not match the expected
format kotlinc will fail.
Since standalone kotlinc is not IDEA, this change explicitly
sets the system property for the build number to a fallback
value. That avoids reading /tmp/build.txt.
closes https://github.com/JetBrains/kotlin/pull/4763
When resolving a name that is used as a receiver and all
candidates are classifier symbols, we need to perform
qualifier resolution. The code used to only do that for
regular classes, and would not perform qualifier resolution if
there was a type alias candidate.
^KT-51171 Fixed
In the Android eco-system, there are jars with entries with a
versionNeededToExtract of 0. That should probably have been 10,
but will be hard to fix. This change proposes to just check
that the versionNeededToExtract is between 0 and 20.
The IDE expects the super type calls to resolve to the constructor
invoked and not the class. The frontend itself resolves to the
class, so an extra layer is needed to adjust that to the constructor.
If the class `A` is in a jar `previous.jar`, the following CLI
invocation will take that class instead of the `A` class
defined in `A.java`:
kotlinc -cp previous.jar A.java B.kt
This is problematic for build tools that put the jar for a
previous build on the classpath when recompiling some of the
files.
^KT-51025 Fixed.
This is in addition to supporting the `package` attribute in
AndroidManifest.xml.
In AGP 7 developers can use `namespace` instead of the manifest
attribute. In AGP 8 it will no longer be possible to use the `package`
attribute in AndroidManifest.xml.
^KT-50887 Fixed
Java protobufs generate a lot of inner classes and when compiling
Kotlin wrappers for those Java protobufs the inner classes can
be loaded hundreds of times leading to out of memory. This change
avoids reloading by fixing a bug in the cache implementation that
caused inner classes to not be cached.
Do not coroutine transform static inline class replacements that
forward to a default interface suspend method. No boxing has to
take place on return (as the default interface method always
boxes) so we can simply forward the call.
^KT-49645 Fixed
The inliner uses module descriptors to figure out if it needs to
regenerate objects.
We should avoid the use of descriptors in the inliner, but this
works as a first quick fix.
The coroutine method transformer extends the range of locals
across code where the local is not live when it is safe to do
so. However, it only bailed out for one case of control-flow
merging, namely backwards branches for loops. That is not
sufficient as there can be control flow merges without loops
where the local is only defined on one control-flow path.
This change generalizes the bailout to any control-flow merge.
^KT-49834 Fixed
This shows up in annotation instantiation tests where we need
to make sure to generate a property on the annotation implementation
class for such properties.
- Mangle names for extension receivers in lambdas
- Correctly mark anonymous variables and variables for arguments
for destructuring declaration.
There is one failure remaining which is cause by lambda
type inference differences that leads to FIR having an explicit
return from the lambda whereas old frontend leads to an implicit
return. This difference is visible in debug stepping that the
local variables tests do because the implicit return has the line
number of the closing brace of the lambda. This change adds an
IrText test to make the difference clear.
Putting them in the local variable table means that the debugger
needs to have special handling for parameters with specific names.
That forces us to generate mangled names for these.
Instead of also implementing the name mangling for FIR, this
change gets rid of the parameters from the LVT instead.
The JVM and newer Android runtimes treats that the same as if
there is no enclosing method. However, older Android runtimes
for Android 5 and 6 throw exceptions on reflective access
and even older runtimes have different behavior. To avoid
those issues, exclude <clinit> from enclosing method attributes.
^ KT-48754 Fixed
This fixes a performance problem in the case where the lock object
is a capture and the monitor enter/exit happens directly on
field loads. When the locking happens on field loads instead of a
local, the JVM cannot prove that locking is balanced. That has
the consequence that the code is runs very slow (always in the
interpreter).
^KT-48367 Fixed.
This is already the case for straightline code such as
```
inline fun <R> f(size: Int, block: () -> R): R {
var result: R
result = block()
return result
}
```
However, if the local variable introduction happens at a merge
point as in the following example, we allow the unboxing but
only do it halfway. The initialization of the local is still
done with a null value.
```
inline fun <R> f(size: Int, block: () -> R): R {
var result: R
while (true) {
result = block()
if (size == 0) break
}
return result
}
```
This change disallows unboxing for this move complicated
case as well by bailing out if a local use is with a
TaintedBoxedValue (merge of Object and Integer).
^KT-48394 Fixed.
This ensures that the debugger always has a bytecode offset for
the line number of a break/continue so that you step there and
so that you can set breakpoints there.
The `nop` instruction is optimized out if it has no line number
information.
^KT-46450 Fixed
There used to be code that extended a previous range instead.
However, that does not work as that extension could have the
local cover code where it does not exists. Since we no longer
extend the range of locals, we should always introduce a new
one even if there was another one for a previous range.
This filters it out for use in SAM wrapper generation with
invoke-dynamic.
The intrinsic is not actually used for code generation so
it is added to the list of intrinsics that should have been
lowered away before codegen.
^KT-45779 Fixed
CoroutineTransformermethodVisitor attempts to extend the ranges
of local variables in various situations. Probably in an attempt
to give a better debugging experience. However, all of these
range extensions lead to invalid local variable tables where
something is in the local variable table where nothing is in the
corresponding slot.
The code that extends variables to the next suspension point
instead of ending them when they are no longer live has issues
with loops. When resuming and reentering the loop, the locals
table will mention a local that we did not spill and which
is therefore not restored when resuming.
The code that extends local variable table entries if there
are no suspension points between two entries doesn't work
for code such as:
```
var s: String
if (suspendHere() == "OK") {
s = "OK"
} else {
s = "FAIL"
}
```
If the local variable ranges are collapsed into one, one of
the branches will have the local defined in the local variable
table before the slot is initialized.
FunctionNVarargBridgeLowering checked the name of the method instead
of whether the method overrides an invoke method. That doesn't work
when the name of the function gets mangled because of inline class
arguments.
Fixed KT-45084.
For code such as:
```
try {
var y = "y"
for (i in 0 until 1) {
return y
}
} finally {
println("finally")
}
```
The local variables `y` and `i` ended up covering the finally block as
well in the debugger.
This change splits the range of the locals so that they do
not cover the finally block.
This change does not change the inliner to do similar transformations,
so the range of locals is still wrong win finally blocks when inlined
into an inline function. Added a failing test to that effect.
The coroutine transformation would leave locals in the local
variable table across the code that reloads local variables from
the continuation on reentry. However, when reentering the function
the local has no value until after the reloads from the
continuation.
This change splits the locals in the local variable table to
avoid such uninitialized locals. A local alive across a
suspension point has its range split in two. One that goes
from the original start to the state label for the restart
after the suspension. The other goes from after the local
has been reloaded from the continuation until the previous
end of the local.
Coroutine transform would occasionally *extend* the range of a
local instead of shrinking it (if the next suspension point
is after the end point for the local). That leads to the local
variable table having a local that covers code where it is not
defined. That is invalid and leads to D8 removing the locals
table from the code.
When a suspension point is inlined, the inlining local spans
unspilling code where the local slot has not been initialized.
The transformer already inserted initialization code for the local,
however, it did not split the local variable table. Therefore,
the inlining local is defined on instructions where the local
slot has no value (namely the instructions that initialize
the local slot on the unspilling path).
This change adds splitting of the local variable table as well.
When updating to a new version of D8, the uninitialized local
slot showed up. D8 will repair it in this case, but it is
better to not generate such code.
This breaks from the loop itself which is inconsistent with
what happens for breaks in while conditions.
Also, the frontend will report that code after the loop is
unreachable, which it isn't. :-\
However, those issues are covered in
https://youtrack.jetbrains.com/issue/KT-17728, so for now
we follow the old backend to not "break" anyone. :)
Fixes KT-44412
Mads Ager