it was an oversight in a special case of block validity checking. I had all the complicated pax deposit/withdraw totals matched up with pending cross chain swaps but there was a test for a special 2 vout coinbase as we use the second vout for the price feed data.
So it was testing for this special case properly, but the default path of the 2 vout coinbase (which cant be any pax tx) was set to ok, instead of error.
totally my oversight, and the case where you spend all your effort on the difficult part and dont check for the simplest exeptions to the simplest case. I will take a 1 million KMD responsibility for this and contribute that much to a cryptox make good fund. I believe the long delay from our request to stop deposits and when it was finally done was a contributing factor that made it go as big as 3.5 mil, but with 1 million KMD, that means at least one third of all the cryptox KMD can be made real.
this combined with the 25 BTC to make cryptox whole and I think we have most of the solution
I will use personal KMD for the 1 million contribution to the makegood fund.
All things considered I think you handled this situation very well James, well done!
Yesterday when I first heard of the strange 100K blocks mentioned in slack my impression was there was going to be a rollback to correct an attack involving 250k-1M coins, and this would be done while most of Europe and North America was asleep. Under those assumptions I thought invoking a rollback without at least considering other options was crazy, and likely to cause a loss in KMD confidence of a magnitude great enough to destroy KMD & SuperNET in an instant, voiding all your work since 2014, and the holdings of all the investors.
I can see now with more information about the nature of the attack that a rollback (or whatever technical name best describes the fix) was the best solution under the circumstances. You acted quickly, took responsibility for the bug, and fixed it asap, and nobody can really ask for more than that when investing in experimental projects.
