https://github.com/jayfoad commented: This looks OK but I'm not too familiar with operand parsing. Is there a cleaner way to implement this, so that we don't have two different places where "lit" is parsed? https://github.com/llvm/llvm-project/pull/69527