This works in many cases, however, it is incomplete since there
are cases where classes are extracted to top-level and therefore
reparented. Therefore, we lose the information about the function
class are nested inside.
Two known issues with generateNonPartClassDeclarations that was here
before were the fact that we didn't sort sealed classes and its
subclasses which led to NoSuchMethodError (KT-27097), and the fact that
we didn't skip expect classes which led to incorrect duplicate JVM class
name diagnostic (KT-30843)
#KT-27097 Fixed
#KT-30843 Fixed
The KotlinType.toIrType function sometimes produces unbound symbols
and was only used in IrIntrinsicFunction and SharedVariablesManager.
This change removes the broken function and changes SharedVariableManager
to use IrType consistently and IrIntrinsicFunction to use KotlinType
consistently.
- drop calculatable from 'inlineFunctionSymbol' fields: 'sourceFileSymbol' and 'sourceFileName' and
replace then with extension values
- drop unused constructors.
This change let backend calculate and bind right function to the code block on generation debug information
e.g. for Kotlin/Native backend this let codegnerator to produce DILocation(..., inlinedAt:..) with correct value
for inlinedAt parameter, thus signifiacntly improve behaviour of debugger.
We would like DeclarationDescriptor.getOriginal() to always point to a
(preferably unique) unsubstituted descriptors for the given possibly
substituted descriptor. In case of LazySubstitutingClassDescriptor
(which can be observed for nested generic Java classes), this invariant
was broken, because 'getOriginal()' for constructors of substituted
class returned a copy created for this particular substituted class.
Instead of materializing an invokable object and storing it in an
instance of the implementation, create a specialized class that calls
the lambda directly, resulting in one less nesting layer and one less
class.
HeaderInfo object, and modifying the operator in the loop condition.
The "additional emptiness condition" is no longer necessary with this.
The open/closed property was removed from HeaderInfo in an earlier
commit, but bringing it back in to simplify the loop building makes
more sense.
Also expanded tests for evaluation order of range bounds.
do-while with enclosing "not empty" check).
Also do not add additional "not empty" condition for `until` loops when
the given bound is a constant != MIN_VALUE.
Mads Ager