nullptr in 3.0 due to result type definition.

[sjaakd]

Internal error in the mapping processor: java.lang.NullPointerException at org.mapstruct.ap.internal.model.BeanMappingMethod$Builder.build(BeanMappingMethod.java:201) at org.mapstruct.ap.internal.processor.MapperCreationProcessor.getMappingMethods(MapperCreationProcessor.java:373) at org.mapstruct.ap.internal.processor.MapperCreationProcessor.getMapper(MapperCreationProcessor.java:149) at org.mapstruct.ap.internal.processor.MapperCreationProcessor.process(MapperCreationProcessor.java:120) at org.mapstruct.ap.internal.processor.MapperCreationProcessor.process(MapperCreationProcessor.java:74) at org.mapstruct.ap.MappingProcessor.process(MappingProcessor.java:283) at org.mapstruct.ap.MappingProcessor.processMapperTypeElement(MappingProcessor.java:263) at org.mapstruct.ap.MappingProcessor.processMapperElements(MappingProcessor.java:221) at org.mapstruct.ap.MappingProcessor.process(MappingProcessor.java:152)

[sjaakd]

Reproducer in attached PR.

[Captain1653]

@sjaakd I have reproduced this exception. The problem is here
. Our result types aren't assignable. If we will remove this bad line BeanMappingPrism.getInstanceOn( method.getExecutable() ).mirror then we will receive message from compiler something like that:

org.mapstruct.ap.test.mypackage.Issue1821Mapper.Target not assignable to: org.mapstruct.ap.test.mypackage.Issue1821Mapper.ExtendedTarget.

Is it acceptable for us?

[sjaakd]

While many parameters of @BeanMapping can be inherited, it does not make sense to inherit the result type.

Your proposal does not take away the cause of the problem, it solves the symptoms. The solution is not to copy the resultType when inheriting. I guess that's somewhere done in MappingOptions I guess.

[sjaakd]

done.. resolved by PR #1822