On 2021-01-22 at 18:05 +0100, Erich Eckner via Gnupg-users wrote: > > I was more wondering, why gpg decides to go into "tor mode" on box #2, > when there is actually no tor installed or running. I'm totally happy to > force non-tor mode via config file, but I'm also open to help find the > root for gpg's misjudgement of tor-availability.
The check is in dirmngr.c: > int > dirmngr_use_tor (void) > { > if (tor_mode == TOR_MODE_AUTO) > { > /* Figure out whether Tor is running. */ > assuan_fd_t sock; > > sock = assuan_sock_connect_byname (NULL, 0, 0, NULL, ASSUAN_SOCK_TOR); > if (sock == ASSUAN_INVALID_FD) > tor_mode = TOR_MODE_NO; > else > { > tor_mode = TOR_MODE_YES; > assuan_sock_close (sock); > } That assuan_sock_connect_byname() tests the connection by connecting to both tor port (9050) and the tor browser port (9150). It actually starts negotiating a request (see socks5_connect) > > /* For HOST being NULL we pass an empty string which indicates to > socks5_connect to stop midway during the proxy negotiation. Note > that we can't pass NULL directly as this indicates IP address I don't see how it would automatically treat it as having tor unless you have a socks server on either 9050 or 9150. Best regards _______________________________________________ Gnupg-users mailing list Gnupg-users@gnupg.org http://lists.gnupg.org/mailman/listinfo/gnupg-users