Sadly there's nothing we can do about that specific issue, at least not unless we rewrite the parsing code.
I'll flag this as Won't Fix because it's not something we can immediately fix, and it's a part of a bigger problem. We will take steps to try to solve this in the future