Roger Dingledine transcribed 1.0K bytes:
Isis: I'd like to highlight this change for you, since it means that the bridge authority's networkstatus files are now going to have HSDir flags on the bridge status lines.
I don't know if this is going to be a problem for any of your parsing code in any way (hopefully not), but I figured now's a great time to let you know it's coming.
--Roger
On Thu, Jul 16, 2015 at 07:47:05PM +0000, nickm@torproject.org wrote:
--- a/src/or/networkstatus.c +++ b/src/or/networkstatus.c @@ -1678,7 +1678,7 @@ networkstatus_getinfo_by_purpose(const char *purpose_string, time_t now) if (bridge_auth && ri->purpose == ROUTER_PURPOSE_BRIDGE) dirserv_set_router_is_running(ri, now); /* then generate and write out status lines for each of them */
- set_routerstatus_from_routerinfo(&rs, node, ri, now, 0, 0);
- set_routerstatus_from_routerinfo(&rs, node, ri, now, 0); smartlist_add(statuses, networkstatus_getinfo_helper_single(&rs)); } SMARTLIST_FOREACH_END(ri);
Hey Roger,
Thanks for the heads up! I've added unittests to BridgeDB to test how its parsers handle the HSDir flag (and make-believe flags like "Unicorn" for good measure). [0]
Does this mean that, if I were a client using Bridges, and I were also running a Hidden Service, that there is some chance I might use one of my Bridges to store my HS descriptor? If so, wouldn't that be harmful to the HS's anonymity?
[0]: https://bugs.torproject.org/16616