Prevent duplicate operationids

As per #7 (closed), this resolves the duplicate operationIds. It does so by doing a check against a hash of generated operationIds; while I originally tried to change the operationId generation algorithm to avoid the duplicates, this caused most tests to start failing as they were based on the previous operationId generation algorithm. Rather than trying to correct all 80 failed tests, I did this check instead. However, there is one spec that needs to be corrected because it uses a shared context that leads to duplicate operationIds.