[foreign-jextract] RFR: 8261906: Improve jextract support for virtual functions [v6]

[Maurizio Cimadamore]

On Thu, 4 Mar 2021 17:09:01 GMT, Maurizio Cimadamore <mcimadamore at openjdk.org> wrote:

>> src/jdk.incubator.jextract/share/classes/jdk/internal/jextract/impl/OutputFactory.java line 408:
>> 
>>> 406:                 varInfo = VarInfo.ofFunctionalPointerVar(clazz, layout, Utils.javaSafeIdentifier(funcTypedef.get()));
>>> 407:             }
>>> 408:         }
>> 
>> I see what you mean here. I guess we expect these functional interface classes to be generated when visiting the typedef?
>
> yeah - not great - as the functional interface name might differ significantly from the typedef name in case of clashes...

Can we rely on typedef names being unique? If so, we could build a map String -> FunctionBuilder when we construct a functional interface for a typedef.

In a header file this should perhaps always be the case - but I've seen that you can have suspicious stuff like

{
typedef int a;
}

{
typedef float a;
}
But maybe we shouldn't worry about this with C headers?

-------------

PR: https://git.openjdk.java.net/panama-foreign/pull/456