JVM_IR has consistent stepping behavior for secondary constructors
and init blocks. This change fixes the expectations for tests that
expect the less consistent JVM backend behavior.
Temporary variable loads in when expressions had the offsets
of the variable declaration. That leads to hacks during codegen
for line number generation.
Instead of those hacks, give the variable loads the offsets of
the context in which they occur. That avoids the codegen hacks
and fixes stepping behavior for more when expressions.
Now, FirTypeScope returns only direct overridden and a base scope
where a client might look for deeper overridden
It's necessary to make the API fast and simple
when only direct overridden are needed
At the same time, while this API looks a bit complicated for most of the clients
there are simple extension helpers that obtain just the direct or all overridden
NB some cases such as captured extension receiver for an extension
lambda are not supported yet; to be discussed, to what extent should we
actually follow JVM code shape here.
Report the input module name and size once before any measurements
instead of duplicating it in all of them. Also, render measurements in a
table to improve readability.
Before:
info: PERF: INIT: Compiler initialized in 467 ms
info: PERF: ANALYZE: 1 files (2 lines) target main-java-production in 277 ms - 7.220 loc/s
info: PERF: IR: Translation 1 files (2 lines) target main-java-production in 291 ms - 6.873 loc/s
info: PERF: GENERATE: 1 files (2 lines) target main-java-production in 513 ms - 3.899 loc/s
info: PERF: IR: Generation 1 files (2 lines) target main-java-production in 142 ms - 14.085 loc/s
After:
info: PERF: main-java-production, 1 files (2 lines)
info: PERF: INIT: Compiler initialized in 421 ms
info: PERF: ANALYZE 342 ms 5.848 loc/s
info: PERF: IR TRANSLATION 296 ms 6.757 loc/s
info: PERF: GENERATE 453 ms 4.415 loc/s
info: PERF: IR GENERATION 137 ms 14.599 loc/s
Roman Artemev