We need to supply signatures for private declarations even when their
containing file is not known. In situations where the FirDeclaration in
question is already known (i.e. where we are never going to serach for
the declaation based on that signature), FirDeclaration itself is
sufficient as a key.
The tests fail currently because Klibs are generated by the JS backend,
and signatures differ between JVM and JS (on JVM, return types are
included in mangles for simple functions).
This leads to linkage errors.
Once signatures are reworked and harmonized between backends, the tests
will repair themselves.
This is godugly code, where a flag for file level signatures is passsed
around.
An alternative would be not to create file level signatures for toplevel
private clases, since those still need unique names, at least on JVM.
But that would break binary compatibility.
Signatures are due for overhaul anyway. Hopefully this code can be
reverted at that point.
1. merge(null of type A, null of type B) = null of unknown type;
2. merge(null of type A, something of type B) = merge(unknown null, B).
^KT-52311 Fixed
This is somewhat suboptimal since this results in `::suspendInline`
generating 2 classes while `{ suspendInline() }` only creates 1, but
it's the best allowed by the existing hierarchy of classes in stdlib. At
least it works?
^KT-50832 Fixed
These tests don't work on JVM target 1.6 after 6d664bcd10 because we're
generating `Boolean.compare` which is only available since 1.8. (It is
not a big deal because JVM target 1.6 is prohibited for users now.)
It's incorrect to take the first parameter type from the expression
itself because it can be nullable if smart casts are used. And if it's
nullable, it's mapped to the wrapper type and calling
`comparisonOperandType` for it makes no sense. Instead, take the type
from the callee function, as it's guaranteed to be mapped to a JVM
primitive type.
E.g. in `test1` function in the added test, the problem was that the
dispatch receiver type of the call expression is `Double?`, which is
mapped to `java/lang/Double`, whereas we clearly wanted to obtain the
primitive `D` (double) type.
#KT-52163 Fixed
For negative case, throwing CCE, we already have the issue KT-8135
and dedicated tests:
* genericDelegateUncheckedCast1.kt
* genericDelegateUncheckedCast2.kt
Related issues: KT-8135 KT-39588 KT-48749 KT-49837
`$$forInline` functions do not pass through the state machine generator,
and optimizing `Ref`s before that changes how assignments inside lambdas
passed to `suspendCoroutine`, etc. behave: without a `Ref`, the
assignment is not reflected in the continuation object, so the variable
has old value on resumption.
These functions will be optimized later, after they are inlined
somewhere and the state machine is generated.
^KT-52198 Fixed
E.g. in `x + f()` where `f` is an inline lambda, the instructions for
`+` should have the line number of that expression (while previously
they instead had the line number of the last line of the lambda).
^KT-51738 Fixed
This only affects JVM IR, and this is similar to how this was done in
the old JVM backend in `CallBasedArgumentGenerator.generateVararg`.
#KT-52146 Fixed
In case there are several proxy functions for indy lambdas in the same
container, its names are "...__proxy", "...__proxy-0", "...__proxy-1",
..., yet before this change, only the first one was sanitized. So if
it's happening inside a constructor, `<init>` was left unrenamed which
led to ClassFormatError.
#KT-52040 Fixed
Mikhail Glukhikh