Have a GPaste 45 or 50 owning the name re-execute, which it does on SIGUSR1: a package update has replaced its binary, so what comes back is this version's daemon.
Two short calls to the bus, made synchronously.
Only that daemon is signalled: one serving this interface has Reexecute, and may be settling its storage behind a prompt the signal would tear down; and the pid file names the daemon that wrote it and outlives it, so only when it names the process owning the name does the signal reach a daemon rather than whatever process got that pid since. Anything else fails with g_paste_error_failed, and a question the bus could not answer with g_dbus_error or g_io_error, as does a platform with no signals.
| client |
a Client |
|
whether the daemon was signalled |