[foreign-jextract] RFR: 8263018: Improve API for lifecycle of native resources [v2]

[Maurizio Cimadamore]

On Wed, 24 Mar 2021 13:53:32 GMT, Maurizio Cimadamore <mcimadamore at openjdk.org> wrote:

>> src/jdk.incubator.jextract/share/classes/jdk/internal/jextract/impl/HeaderFileBuilder.java line 99:
>> 
>>> 97:                 emitFunctionWrapperNoAllocatorOverload(javaName, functionInfo);
>>> 98:                 emitFunctionWrapperScopedOverload(javaName, functionInfo);
>>> 99:             }
>> 
>> With the emitFunctionWrapper call above this, don't we get a publicly accessible function wrapper that will always throw a WrongMethodTypeException because it's missing an allocator argument? Maybe it should be conditionally made private? (since the other 2 functions reference it, it's still needed).
>
> I see what you mean.

Actually, I think the code is correct, as the generated "main" method will either have no allocator argument (in case return is not a struct) or will have allocator parameter (in case return is struct). The issue with the code is that, when no  allocator argument is present, we should pass DEFAULT_ALLOCATOR to the underlying MH.

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

PR: https://git.openjdk.java.net/panama-foreign/pull/474