init_send_tx locks outputs only after a successful Tor send. During the request, parallel sends can select the same outputs.
We should lock them before calling Tor and keep them locked on failure, since the slate can still be sent manually. cancel_tx can unlock them if needed.
init_send_tx locks outputs only after a successful Tor send. During the request, parallel sends can select the same outputs.
We should lock them before calling Tor and keep them locked on failure, since the slate can still be sent manually. cancel_tx can unlock them if needed.