This should fix a rare zapwallettxes failure on travis, but also avoids having init operations (re-adding wallet transactions to mempool) running after RPC is free'd.
I believe this was the failure at https://travis-ci.org/bitcoin/bitcoin/jobs/311747844 (from #11605).
This should fix a very rare travis failure in zapwallettxes, but
is also more correct, as you can currently race
ReacceptWalletTransactions with stop RPC calls to get bitcoind to
(IMO) eroneously return a non-0 exit code.
ACK 07c4838.
Shutdown is handled after AppInitMain(). If the initialisation succeeded or not shouldn't depend on fRequestShutdown.
utACK 07c483895d219d4c44bd8fec5167a53f4701c364
If the initialisation succeeded or not shouldn't depend on fRequestShutdown.
"Terminated during initialization" is a valid reason for initialization to fail. The reason for returning false here is that initialization didn't fully finish, which is used in the GUI code to (not) spin up other things.
Yes but only if something was interrupted in the initialisation, that's not the case here, where everything was initialised.
Yes but only if something was interrupted in the initialisation, that's not the case here, where everything was initialised.
So in the case something was actually interrupted, it still early-outs with false? Then this change is fine with me.
Indeed, if initialization did not finish because of fRequestShutdown, false is returned after block index loading fails to complete.