Given I've not been able to replicate the underlying issue I've essentially taken a best guess approach to what the problem is based on the evidence and written a fix that should resolve that. It's not totally foolproof in that it assumes that the duplicate submissions are identical but I don't understand how they cannot be.
It doesn't throw up an error, just silently throws away the identical data and carries on which I think is fine as then end result will be the same for the user and this way they don't get an error message they can't do anything about.
Given I've not been able to replicate the underlying issue I've essentially taken a best guess approach to what the problem is based on the evidence and written a fix that should resolve that. It's not totally foolproof in that it assumes that the duplicate submissions are identical but I don't understand how they cannot be.
It doesn't throw up an error, just silently throws away the identical data and carries on which I think is fine as then end result will be the same for the user and this way they don't get an error message they can't do anything about.
Fixes #40