whaack: http://btcbase.org/log/2017-03-16#1628076 A q regarding: "because their parent coinbase gets zapped from the index illegitimately" . When the coinbase gets nix'd, are all traces of the txn removed from the blockchain? You say that there may be txns that can't be walked back to their coinbase in existence, but I don't see how a txn can be completely missing if my node needs to verify the merkle

whaack: root when syncing

whaack: Does trb keep a separate index for O(1) lookups of txn's by hash? And you're saying that the hash table becomes corrupted, but perhaps there is another location where the txn / block data is being stored?

whaack: Related, I just created a prototype walk-back function for my block explorer. It takes a txn and traces the tx's of all used inputs back to their coinbases.

whaack: Also, I am only on block 134,280 ; but so far I have only seen two coinbase txn's with matching hashes. Neither of them were ever spent. Are there more known ones in the wild? Also, is it possible to create a matching txid with a coinbase from a previous block reward cycle? (I guess this boils down to whether the value of the coinbase is explicit in the transaction or implicit by the block number,

whaack: which I don't know) If it's impossible to create poison coinbase after the block reward changes, then I guess one could argue that the real confirmation time is 210,006 blocks lulz