[bisq-network/bisq] Fix issue with invalid vote result (#2358)
notifications at github.com
Sun Feb 3 12:39:17 UTC 2019
After some investigation for the reason for the issue in the last voting round I came to that conclusion:
Analysing the data from Devin:
He had published his vote reveal tx at block 1 453 842 (triggered by onNewBlockHeight). He collects all blind votes (BV) he has and creates the hash. The list is filtered so that only valid BV are included. He has received all BV data but those which have not been parsed yet are invalid as the phase cannot be validated. That is why he only included 8e12be8894fcba34f2dc1543c7154a4a9d7b25eccb2858b6446a21e69dbb1999 and bd83b473df48ae577982b38409412cc570659c93bb8a34f3aad0bc401c93c273.
8e12be8894fcba34f2dc1543c7154a4a9d7b25eccb2858b6446a21e69dbb1999 was included in block 1 453 828 so that is an earlier block and was parsed already, so it was considered valid.
bd83b473df48ae577982b38409412cc570659c93bb8a34f3aad0bc401c93c273 was his own tx but should not be parsed at that moment, so it is not clear yet why that was included.
The other 2 BV txs have been included in block 1 453 852 and one in 1 454 047.
Another one in block 1 453 854 (for that never got a vote reveal tx btu it should be included in the hash).
The 2 BV txs at 1 453 852 have been mine and as they had the highest vote stake they created the majority hash. They have been executed before the 2 BV txs at 1 454 047 and 1 453 854, so that is why only 4 txs have been included (the 2 past and the 2 at the same block).
So a main reason for the bug was that the vote reveal tx have been created at the block where the own blind vote tx was included, therefore missing all later BVs.
We trigger the check for vote reveal at each new block (before it is parsed) but use the bsqNode.getChainTipHeight() not the actual height in the daoState, so we are ahead to what has been parsed so far. That seems to be a main issue for the bug. Need more investigation and testing....
You are receiving this because you are subscribed to this thread.
Reply to this email directly or view it on GitHub:
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the bisq-github