Rafael Winterhalter

I just checked and I had fixed a similar issue previously where I added annotations to explicitly refer to method names rather then infering from the runtime name. Have you...

The translate method is resolved via reflection as it has different signatures on different VMs. It seems as if the obfuscator breaks those reflective resolution. Thanks for sharing your solution.

Small world! I have not, unfortunately. But I would certainly consider it. Do you have a need for it?

Everything we can automate, I would prefer it automated. Could you create a reproductive test case and add it as a PR draft to the Mockito repository? I am happy...

This class determines what methods to skip: https://github.com/mockito/mockito/blob/f48d794ad14982a134fd14dd2aef03477b699dc6/src/main/java/org/mockito/internal/creation/bytebuddy/BytecodeGenerator.java#L23 Can you check the byte code of the Groovy class' and see if there is any new method that should be added?...

It works already: https://gist.github.com/raphw/e0f6864ddcad69630d839c2eebaad818 But it requires some prework. As far as I know, the Graal team is working on something to ease this.

That's always been a weird limitation in Mockito. But it was necessary behavior where some fields could be copied to establish a working spy. But I agree, the pre-inline behaviour...

That is a bug indeed. The super types are not checked for the field declaration such that it is missed. I will need to fix that in some future version....

Fixed on master.

I'm actually surprised by this as I did not expect this "semi virtual binding" fir static methods. The method graph resolves overloads in the context of generics and such, that's...