Because those test are single Java files, many of them use a top-level
class as a container for multiple classes to be tested. Such tests do
almost nothing if those nested classes aren't handled.
- Add utilities to add new attribute to ConeAttributes
- Get rid of FlexibleNullability attribute (it can be easily inferred
for any flexible type at any moment)
- Fix determining of EnhancedNullability attribute
Previoisly, there were two places where mapping had happened:
- toConeKotlinTypeWithNullability
- enhancePossiblyFlexible
The first one was used for supertypes and bounds and the second one
was used for other signature parts
The main idea is to perform type mapping once to a flexible type,
and then use it as it's needed (it's lower bound, or for the further ehnancement)
Also, this commit fixes flexibility for type arguments, see the tests
Really, this commit implements early J2K mapping for all Java types.
It's questionable and probably wrong at least for super-types,
because, for example, we cannot resolve spliterator() in classes
derived from java.lang.Iterable
Before this commit, we created type parameter symbols each time
when type parameter was referenced or created.
In this commit, we introduced class-bound Java type parameter stack
and use it to find referenced type parameter symbol.
So now they are created only when Java type parameter is created
Basically, this commit splits test data from the from java-txt-kt to two pairs java-txt and kt-txt.
This commit leads to some duplication in test data.
This is temporary: in the platform types branch the test data for LoadJava tests will be changed dramatically, so duplication will go away
pyos