Hi @ivanbgd, the last stage doesn’t differ significantly in terms of how your code is executed. All stages are tested using the --resolver flag. The difference is that earlier stages didn’t require you to handle it explicitly.
Suggestion: Focus on completing the last stage first. Once your code handles the --resolver flag properly, it should work seamlessly for all stages.
I was trying to organize code in two different functions: one for standard mode and one for the forwarding mode. Would it make sense to do it that way? I tried to call the standard mode if we don’t receive a CLI argument, but then the previous tests fail, if you know what I mean.
Or does it make more sense if one function handles everything?
I didn’t split the function in two, but what solved the issue for me was to manually change OpCode::InverseQuery to OpCode::Query. There is one test case that sends OpCode::InverseQuery and it seems like resolver had issues with that.