Set TargetBackend.JVM_IR for these tests by default; remove the
generated IR-based test and add a new old-backend-based test. This fixes
the issue where some (3) of these tests were not properly ignored
because of incorrect target backend used in the test generator.
Also update test data for some tests which use local functions, which
are not generated to separate anonymous classes in JVM IR.
Since neither inline function nor super classes affect klib there is no
need for extra passes so compiler doesn't report any dirty files.
Do not check that.
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
JPS does not remove output files for files marked as dirty
for current iteration. This is a problem for adding complementary files
during JPS MPP IC.
If output files are not removed, the compiler might raise an error
for duplicated actual declarations.
Previously IC could go to the next iteration without removing
outputs for the files compiled during the last iteration.
For example, it could happen, when a multifile part is changed
(we add sources for other parts and recompile everything together).
In case of MPP it could lead to compile error, because
the compiler would see the same actual declarations from dirty sources
and dirty outputs from previous iteration (which should have been removed).
That behaviour did not raise an error before dc8240c24e
because actual declarations from binaries (e.g. compile classpath) were ignored.
#KT-35957 Fixed
This fixes the weird cases when a class gets overwritten by an imperfect
copy, reduces the number of classes in the output if an inline function
contains an inline call that causes it to have regenerated anonymous
objects, and makes inlining of same module functions a bit faster in
general. On the other hand, this may increase memory footprint a bit
because classes cannot be flushed to the output jar, as the inliner
would not be able to locate classes for anonymous objects if they have
already been unloaded from memory.
Bug fix `NoConstantValueAttributeForNonConstVals` removes constant attribute from non-constant properties.
Code using such properties does not have to be recompiled, since its value is no longer inlined.
Sometimes IC raises compilation errors when rebuild succeeds.
This happens because IC uses serialized decriptors
for non-dirty files. Serialized descriptors can be different
from source file descriptors. For example, a source file
may contain an implicit return type or an implicit visibility
for overridden methods, but serialized descriptors always
contain explicit return types & methods' visibilities.
These problems can be solved by expanding a scope of incremental compilation
just after the analysis, but before error reporting & code generation.
In other words, we need to compare descriptors before error reporting and code generation.
If there are new dirty files, current round of IC must be aborted,
next round must be performed with new dirty files.
This commit implements IC scope expansion for JS Klib compiler
#KT-13677
#KT-28233
Call checker and declaration checker are used in order to preserve backward compatibility.
Attempt to use classifier usage checker was not good enouth,
since not all errors found with it would actually be reported before.
For example types and constructor calls don't cause supertypes to resolve,
so missing supertypes would not lead to errors in case they are the only use of class name.
Updated tests failing due to missing Java dependencies in superclasses.
Without the `-Xmultifile-parts-inherit` mode for now.
This is implemented as follows: FileClassLowering collects information
about multifile parts and the corresponding facades, which a later
GenerateMultifileFacades phase uses to generate new IrFile instances and
add it to the module fragment that's being compiled.
Note that GenerateMultifileFacades is in the end of lowering phases
because delegates in the facade should be generated for all additional
functions generated by certain lowerings (default arguments,
JvmOverloads, etc.). If GenerateMultifileFacades was right after
FileClassLowering, they would still be generated, but we'd then process
them in lowerings mentioned above, which would result in duplicated
logic in the bytecode. There's a new bytecode text test which checks
that this doesn't happen for functions with default arguments.
Our JPS and Gradle plugins handle
recompilation of multifile classes differently.
JPS plugin delegates handling to the JPS itself,
which tracks dependencies via bytecode,
and marks classes as dirty when they are affected.
So in JPS other parts of multifile classes are recompiled only
when a part's signature is changed.
In Gradle plugin we handle recompilation ourselves in
simpler way: any change in any part leads to a recompilation
of all parts of the same multifile class.
In future we should improve our Gradle plugin, but for now
I'm changing the test so that both JPS and Gradle tests
recompile all parts.
Also the dummy function is added to make sure that we
don't blindly recompile everything when a part is changed.
Aleksei.Cherepanov