dr|z3d
squashed the disable ssu1 issue with the latest commit, zzz?
dr|z3d
also, regarding the offending router, not entirely sure it's a BOB/retroshare 3, given its a 0.9.56 router and orignal added some limits to client tunnel requests iirc.. 500? so maybe something more sinister?
dr|z3d
s/3/thing
zzz
dr|z3d, yes
dr|z3d
yes it's fixed now?
dr|z3d
ok, thanks, zzz. good work.
zzz
yes fixed
zzz
orignal, new problem for today:
zzz
Peer test, I am Alice, Bob is i2pd
zzz
Bob picks a charlie that is firewalled and is not publishing IP/port
zzz
so I give up
orignal
I will check
orignal
do you know if Charlie was good before?
orignal
wait for sec. what's wrong with such Charlie?
orignal
he can still send msg 5 to alice
orignal
6 should also come due to hole punch made by 5
orignal
hence when I pick Charlie I only check if Chralie can reach Alice. Not oposite
zzz
well, a couple problems:
zzz
first, here's the ssu 1 spec:
zzz
When Bob receives a request from Alice via IPv4, Bob must select a Charlie that advertises an IPv4 address. When Bob receives a request from Alice via IPv6, Bob must select a Charlie that advertises an IPv6 address.
zzz
second, I want to send msg 6 to the IP/port in the RI
orignal
are you sure it's right?
orignal
we ecxlude too many routers from Charlie
zzz
no, we can talk about it
zzz
but right now I fail if there's no IP/port in charlie's RI
orignal
I can change the behavious
orignal
but I'm afaid we will have to produce error "no Charlie" too often
orignal
maybe we should discuss on Monday
zzz
ok, think about it, I'll do more research, and put in on the list for monday
orignal
I don't see any reason to use IP/port from RI for msg 6
orignal
because it's for symmetric NAT only
zzz
one reason is for security, "know your charlie", that's why we wanted to send the RI and signed response from charlie
zzz
so we're not just sending msg 6 to wherever
zzz
that's not necessarily a good enough reason, but it's one reason
orignal
then let's discuss
zzz
on the list