If serialized metadata version has evidently incorrect size the module
should be marked for rebuild prior to reading serialized version.
#KT-42194 Fixed
There seems to be no point in configuring the compiler argument per
project. This argument will be deleted soon anyway, when we remove
support for JDK 1.6 & 1.7.
Also remove `disableDeprecatedJvmTargetWarning`. It didn't have any
effect in all modules where it was applied because these modules
reassign `freeCompilerArgs` anyway, with
`-Xsuppress-deprecated-jvm-target-warning` in it.
Optimizations may remove the `x$delegate` field, in which case there
should be a `getX$delegate` method to reconstruct the value if needed
due to a call to `KProperty{0,1,2}.getDelegate`. We need to know the
signature of this method somehow.
This will be used in the compiler starting from 1.5.20 instead of the
currently used approach of looking for the single value parameter of the
primary constructor. The problem with the current approach is that
primary constructor can be private (since 1.4.30) and the property could
always be private. Relying on private declarations from metadata is
dangerous; for example lazy IR doesn't usually create stubs for private
declarations, and it didn't create stubs for private inline class
constructors before b5f9b1df, which led to the problem reported in
KT-44723.
Move it to build-common where it's still used in incremental compilation
caches, and simplify a bit. In the future, it'll make sense to remove it
completely and to avoid writing it to caches. In this commit, I don't do
that to prevent the IC cache version to be updated, causing rebuilds for
all JPS projects.
#KT-41758
Improve the test which checks that we use correct metadata version if
`-language-version` is passed by checking all supported language
versions.
The change in libraries/reflect/build.gradle.kts is needed because
kotlinx-metadata-jvm of version 0.1.0 is based on pre-1.4 Kotlin, which
doesn't support the new module file metadata generated with metadata
version 1.4 and later, and module files need to be readable there to be
able to transform them for the shadow plugin.
Similarly override dependency on kotlinx-metadata-jvm in the
binary-compatibility-validator module.
Instead of generating these annotation classes as package-private on
JVM, serialize their metadata to the .kotlin_module file, and load it
when compiling dependent multiplatform modules.
The problem with generating them as package-private was that
kotlin-stdlib for JVM would end up declaring symbols from other
platforms, which would include some annotations from package
kotlin.native. But using that package is discouraged by some tools
because it has a Java keyword in its name. In particular, jlink refused
to work with such artifact altogether (KT-21266).
#KT-38652 Fixed
When we inline an anonymous object which captures something such as
crossinline values or reified parameters, we copy and transform its
metadata in `AnonymousObjectTransformer.transformMetadata`. Basically we
read the metadata of the original class, add a minor protobuf extension
and write it to the new class.
This also includes copying the string table. We read the string table
into `JvmNameResolver` (a representation of string table used in
deserialization), then construct a `JvmStringTable` (a representation
used in _serialization_) and then write it back.
There's a few optimizations in the string table representation in JVM
metadata which allow to store less strings and thus take less space. See
`StringTableTypes.Record` in `jvm_metadata.proto` for more information.
One of the optimizations `Record.range` allows to avoid storing the same
record many times in a sequence. For example, if we have N different
strings in the string table but none of them require any operation (such
as substring, char replacement, etc.), then we only store the record
with all default values (no operation, no predefined string, etc.) and
set its `range` to N. Upon reading such optimized record list in
`JvmNameResolver`, we "expand" it back to normal, so that we could index
it quickly and figure out what operation needs to be performed on each
string from the string table.
The problem was that when we expanded this list, we didn't set the range
of the expanded record entry to 1. So each record in
`JvmNameResolver.records` still has its original range. It doesn't cause
any problems most of the time because the range in this expanded list is
almost unused. However, when copying/transforming metadata for anonymous
objects, we mistakenly passed this expanded list with incorrect ranges
to `JvmStringTable`. So the metadata in the copied anonymous object
ended up being incorrect: each record now was present the number of
times equal to its range. Copying such metadata once again led to
another multiplication of the record list size. Multiple copies resulted
in exponential increase in memory consumption and quickly led to OOM.
For the fix, we now take the original, unexpanded list of records when
creating `JvmStringTable` out of `JvmNameResolver` for transformation of
anonymous object metadata.
Note that another possible fix would be to make range for each record in
`JvmNameResolver.records` equal to 1. This is undesirable though, since
then we'd need to copy each `JvmProtoBuf.StringTableTypes.Record`
instance, of which there could be many, and use some memory for no
apparent gain (since ranges in that expanded list are now not used at
all).
#KT-38197 Fixed
Before Kotlin 1.3-M2 we didn't write `has_field` flag for constants
inside multifile classes. Now we write and rely on this when
trying to load constant initializers, which is totally fine for
binaries that were compiled with the 1.3-M2 or newer version.
Unfortunately, constant initializers will not be loaded for old binaries.
One way is to avoid relying on this flag, but then we'll get other
problems (e.g. 3345dc81fd).
Therefore, for binaries that were compiled with at least 1.3-M2 version,
we'll rely on the flag, otherwise, we won't.
Preface: Kotlin 1.3 will be able to read metadata of .class files
produced by Kotlin 1.4 (see KT-25972). Also, to simplify implementation
and to improve diagnostic messages, we're going to advance JVM metadata
version to 1.4.0 in Kotlin 1.4, and would like to keep it in sync with
the compiler version thereafter. This presents a problem: in an unlikely
event that before releasing 1.4, we find out that the metadata-reading
implementation in 1.3 was incorrect, we'd like to be able to fix the bug
in that implementation and _forbid_ 1.3 from reading metadata of 1.4.
But prior to this commit the only way to do this was to advance the
metadata version, in this case to 1.5, and that breaks the
metadata/compiler version equivalence we'd like to keep.
The solution is to add another boolean flag to the class file, called
"strict metadata version semantics", which signifies that if this class
file has metadata version 1.X, then it can only be read by the compilers
of versions 1.X and greater. This flag effectively disables the smooth
migration scenario proposed in KT-25972 (as does increasing metadata
version by 2), and will be used only in hopeless situations as in the
case described above.
Note that this change brings an incompatibility: `Array<Foo>::class`
will be seen as `Foo::class` by the old deserializer. We consider this
OK because the compiler never had any logic that relied on reading class
literal arguments correctly (otherwise it wouldn't have worked because
it could only see `Array<*>::class` before this commit), and the support
of annotations on types in JVM reflection is only available in the
upcoming 1.3 release (KT-16795)
#KT-22069 Fixed
This commit reverts 59e2101a25 partially,
leaving only the implementation of KT-25972 for JVM. The reason is that
we can't fully commit to stabilizing JS (and .kotlin_metadata) binary
metadata formats so much as to postpone any changes done to it for a
whole release year time. It's likely that we will need to update JS
metadata format incompatibly pretty soon, and with the scheme where we
can read the "current + 1" version, it'd require advancing the metadata
version by 2, which would break the nice property that the metadata
version (since Kotlin 1.4) is equal to the version of the compiler that
produced it.
See KT-25972
JVM versions are increased in order to differentiate pre-1.3-M2 .class
files where signatures mentioning inline classes were not mangled. Other
versions are increased in case something similar will need to be
detected
Already existing tests testRequireKotlinInNestedClassesAgainst14{,Js}
now check that there's no error when loading a module/class with
metadata version 1.4.0
#KT-25972 Fixed
Andrey Uskov