Sporadic dead transactions if receiving and spending at the same time

[GoogleCodeExporter]

I am using bitcoinj-0.7 plus my patches plus the patch for issue 277 (race 
condition in saving a wallet). See my repo branch patched-0.7 commit df9fffe. I 
have reset my blockchain to get rid of any cruft. I'm on testnet3 with address 
mySD89iqpmptrK3PhHFW9fa7BXiP7ANy3Y.

http://blockexplorer.com/testnet/address/mySD89iqpmptrK3PhHFW9fa7BXiP7ANy3Y

I am mass-sending ~0.01 BTC transaction to my wallet with WalletTool, about 15 
tx per test run.

While these 15 tx are processed, I am spending 1 BTC from my wallet back to 
WalletTool. Every now and then (it happened now for the fourth time), these 
outgoing tx first do not get confirmed with the next block and then be marked 
as dead.

One of these tx has id 0014902fa6... it doesn't show up on 
blockexplorer.com/testnet, which has by the way, been reactivated by theymos 
recently. I have attached my wallet and a dump of it, the dead tx is sitting at 
the very bottom. (You are welcome use this wallet for testing, but if you spend 
please co-ordinate with me so no unintended double spending happens)

It is my understanding that dead tx should never happen if I don't share my 
key(s) with other clients.

Original issue reported on code.google.com by andreas....@gmail.com on 12 Jan 2013 at 6:00

Attachments:

• wallet-dump.txt
• wallet-protobuf-testnet

[GoogleCodeExporter]

And I should add that I had this issue already with bitcoinj 0.6.1 plus my 
patched (branch: patched-0.6) so it's likely not a regression of 0.7.

Original comment by andreas....@gmail.com on 12 Jan 2013 at 6:06

[GoogleCodeExporter]

Original comment by hearn@google.com on 5 Feb 2013 at 2:41

• Changed title: Sporadic dead transactions if receiving and spending at the same time
• Changed state: Accepted
• Added labels: Component-Wallet