Generated by 🚫 Danger

Summary of the differences,

• MethodReturnsInternalArray does not report on static fields anymore -> it probably needs to
  • https://github.com/spring-projects/spring-framework/blob/v5.0.6.RELEASE/spring-core/src/test/java/org/springframework/core/PrioritizedParameterNameDiscovererTests.java#L37
• MethodReturnsInternalArray now reports on this.field -> this is expected
  • https://github.com/spring-projects/spring-framework/blob/v5.0.6.RELEASE/spring-jdbc/src/main/java/org/springframework/jdbc/support/CustomSQLErrorCodesTranslation.java#L51
• ArrayIsStoredDirectly violations moved
• Only one violation is removed for ArrayIsStoredDirectly. I think this is expected as the field is set on another object as this. Otherwise the rule should consider all field stores, which it doesn't.

Yes, ArrayIsStoredDirectly looks fine.

For MethodReturnsInternalArray, I see these two issues:

• false negative: static fields - as you described. Yes, we should report those as well
• false positive: there is one here: https://github.com/spring-projects/spring-framework/blob/v5.0.6.RELEASE/spring-jdbc/src/main/java/org/springframework/jdbc/core/JdbcTemplate.java#L574
  The method here just returns "String" and no array. But it is declared in a class, which is declared inside a method, which returns an array (https://github.com/spring-projects/spring-framework/blob/v5.0.6.RELEASE/spring-jdbc/src/main/java/org/springframework/jdbc/core/JdbcTemplate.java#L518) - we probably got confused by that.

I'll try to look into these two issues now....

@oowekyala I've pushed some commits:

• The two cases are covered. The static field case is a bit special - if the field is declared final, it is considered "constant" and it is ok to return it (see https://sourceforge.net/p/pmd/bugs/1475/). That's of course debatable, because the array is indeed modifiable. But that's how the rule works on pmd6.
• The specific case (https://github.com/spring-projects/spring-framework/blob/v5.0.6.RELEASE/spring-core/src/test/java/org/springframework/core/PrioritizedParameterNameDiscovererTests.java#L37) is actually a false-positive on pmd6 - since the field is final.
• I'm not sure, whether the findBoundries is correct on ASTLocalClassStatement. I've tried to put it e.g. in ASTClassOrInterfaceDeclaration, but that seemed to break more things, so I reverted it again (a1d3610 9bee456)