Incremental compilation scope (set of files to be compiled)
could be expanded, even if the frontend reported some errors
to a message collector.
The next iteration of IC could produce a different set of
compiler messages, so we want to report them only after
scope expansion converged.
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
Previously JS IR versions of stdlib and kotlin-test were build
by default using compiler previously built on a buildserver.
It had some issues:
- This required us to advance bootstrap every time we made any
incompatible IR changes. This happens often since IR ABI is
not quite stable yet.
- We never tested the exact combination of compiler and stdlib we publish
We tested:
- new compiler with new stdlib build by new compiler (in box tests)
- old compiler with new stdlib build by old compiler (in stdlib tests)
We published:
- new compiler with new stdlib build by old compiler
After this change JS IR compiler tests, builds and publishes
single configuration:
new compiler with new stdlib build by new compiler
JS IR stdlib and kotlin-test are now built using JavaExec of CLI instead
of Gradle plugin to avoid troubles of loading a freshly built plugin.
This also allows to have a granular dependencies: we don't rebuild klib
if we changed a lowering in a compiler backend, but we do rebuild it if
we changed IR serialization algorithm.
Revert "[JS IR] Build hybrid versions of stdlib and kotlin.test"
This reverts commit b9f88350dd.
Revert "[JS IR] Add gradle plugin integration tests"
This reverts commit d872b27663.
Revert "Update bootstrap"
This reverts commit bc47594c7a.
Revert "[JS IR] Support generating both IR and pre-IR libraries"
This reverts commit 1b8df45bfe.
Pass computed list of changed classpath names to KAPT instead
of relying on the history files to be computed by stub generation.
Also, stop generating classpath history changes during the stub generation.
This commit does not compute the actual changed classpath entries,
and that will be done in the following commits.
#KT-23880
`-Xbuild-file` argument allows the compiler to run without
passing any Kotlin source file in arguments.
We have been using this property in
Kotlin Gradle plugin for a few important cases:
1. incremental compilation (to update caches when there are only removed files);
2. for KAPT (Kotlin sources don't make sense in context
of running APs).
We want to stop using `-Xbuild-file` in Kotlin Gradle plugin,
and avoid breaking the Gradle plugin or IC in other build-systems.
This change adds an argument to explicitly run
the compiler without specifying any Kotlin source file.
Add support for incremental annotation processors in KAPT. These
processors conform to https://docs.gradle.org/current/userguide/java_plugin.html#sec:incremental_annotation_processing
specification.
Support is provided by using javac compiler APIs and
recording the source file structure. At runtime, processors
are instrumented with custom Filer that is used to keep track of generated
files. In order to support classpath changes, stub generation task is
used to generated a list of changed FQCNs, and this is simply used by KAPT.
Both worker and non-worker mode are supported.
#KT-23880
Nikolay Krasko