Incorrect decompilation for anonymous class with parameter about cfr

[leibnitz27]

Oh, that's nice.

The synthetic inner class is actually static, it's supposed to have a static marker, but it doesn't, which is entirely valid.

I wonder what compiler generated that.....

[nharrand]

Oh, I incorrectly reported that is was compiled with javac, but it was actually compiled with eclipse jdt. (More precisely with maven plugin plexus-compiler-eclipse version 2.8.4)

[leibnitz27]

Ah, nice. JDT does all sorts of horrible and weird things (and outputs unsafe code in places).

Still, I really need to compile up my entire test set using JDT - one to add to https://github.com/leibnitz27/cfr_tests/blob/master/pom.xml

[leibnitz27]

Oh, I forgot - jdk also does this since java 9. (doesn't mark static anonymous inners as static).
But my test to infer this doesn't kick in until java 9, and this is java <6.

Makes sense. The fix as committed is useful to have anyway!