Is this what we talked through, and if we get to this point we should be safe? Since you can only close something voided or settled and can only go to those states if the queues are empty?

I had momentarily forgotten that, and was about to suggest that ForceVoidMarket take the matching queue but rather than check that it is empty, just clear it and set len to 0. So we could leave this check here.

I think if we're safe at this point it's OK to remove from here. Though somewhat inconsistent keeping order request len check but not matching queue len check.

From a hotfix pov, I do like the idea of the new instruction just clearing the matching queue, so we don't have code changes outside the new instructions.

I do like the idea of the new instruction just clearing the matching queue, so we don't have code changes outside the new instructions.

Happy to do this, what I have done is safe but yes containing hotfix changes to the new instruction is probably better