@eyedeekay
&eche|on
&zzz
+R4SAS
+RN
+RN_
+StormyCloud
+T3s|4
+acetone
+dr|z3d
+eche|off
+not_bob_afk
+orignal
+postman
+snex
+wodencafe
Arch
BubbRubb
Chrono
Danny
DeltaOreo
FreefallHeavens
Irc2PGuest16752
Irc2PGuest33667
Irc2PGuest3752
Irc2PGuest51880
Onn4l7h
Over1
Sisyphus
Sleepy_
Teeed
aargh2
ac9f_
b3t4f4c3__
b4dab00m
bak83_
duanin2
duck
john231
leopold
makoto
mareki2p_
nilbog-
poriori_
profetik1
r00tobo
rapidash
shiver_
solidx66
thetia
u5657
uop23ip
vivid_reader56
x74a6
xHarr
theoddone
Apparently there are clients out there passing negative port numbers in their dht messages
snex
how is that not unsigned
theoddone
I dunno which client it is.
theoddone
Gonna see if they respond to the positive version, but generating a node id w/ the negative number does produce a "valid" node id.
theoddone
"valid" in the sense that if I generate a node id locally, i get the same values for bytes 5,6, invalid if I try to validate that this is false (nodeId[4] ^ (port >> 8)) !== sha256Hash[4]
theoddone
Well so far all the negative ones aren't responding if I just Math.abs(port), so gonna assume their client has done something really wrong and the peer should be discarded.
snex
or maybe theyre just using security through obscurity
theoddone
more like security through unreachability
snex
maybe they are still using outbound connections
theoddone
I'm sure they are for somethings, but these ports are specific to dht and are query/respond ports
snex
any common thing among the IPs? same country, etc
theoddone
They wouldn't be able to get any responses to their queries, so no need to connect to the dht at all
theoddone
No idea, just have the port/b32 address
theoddone
I suppose it's possible that I'm decoding the compact node info incorrectly, but I don't think so since generating a node id locally matches.
zzz
theoddone, derp, it's an unsigned number to begin with, treat it that way, don't do math.abs()
zzz
eyedeekay, re: samsocks on github, why is the 7 MB binary checked in, and there's no build instructions?
eyedeekay
Re: the binary can't remember, probably a mistake. Re: build instructions also can't remember but 'go build' should do it
zzz
ok
zzz
re: torsocks, I added support to our socks proxy a couple years back
zzz
it would be hard for you to do it in samsocks since you're using a 3rd party socks lib
eyedeekay
I actually have some plans to switch libraries from armon/socks5 to txthinking/socks which should enable me to add that
zzz
didn't know you were a socks/torsocks guy, or remember if I involved you in my torsocks hackery in late 2022, but you can peek at our SOCKS5Server code to see how I did it
altonen
thanks for sharing your thoughts on the rate limit issue, very valuable info, i'll take a closer look once i'm done with ssu2
zzz
torsocks and outproxy support are the two big adders to standard socks
zzz
np altonen. Is SSU2 what you're actively working on now? or?
altonen
i'll start it next week
eyedeekay
IIRC samsocks: I actually did it to show that you could just replace the standard 'Dial' function in a Go application to I2P-ify the entire business, SOCKS was a neat example because it could all fit in one short file
zzz
ok
orignal
zzz, how is your fix going along?
zzz
altonen, it's wild that you're already getting PRs. Who are these unicorns that know both i2p internals and rust, and where are they coming from?
orignal
of NSR
zzz
orignal, I made some good progress, will be testing later today
zzz
always takes me some time to context-switch my head back into ratchets
altonen
i have no idea who they are but it's exciting that people are interested in i2p, maybe it's the magic of RiiR like tor folks mentioned as rationale for arti
zzz
RiiR?
altonen
rewrite it in rust
zzz
lol
orignal
bunch of people are inetersted in i2pd because RiiR ))
orignal
someone is not in sense
orignal
zzz, I'm asking beacuase I'm about to start esting crypto type 6
zzz
which has set tor back several years by some measures, because they all stopped working on C tor
orignal
because it was a condition of US gorvnment funding
orignal
for me Rust sounds like DEI
orignal
and since i2pd doesn't receive any US funding we have the luxury to keep writing C++ )))
zzz
altonen, I just always assumed that everybody in the world who know about i2p internals were here in this channel, I was very wrong
altonen
maybe they are here, "hiding" behind irc2pguest
orignal
you are always wrong about this
orignal
many people form ilita never visited this channel
zzz
sure, I'm exaggerating
zzz
also ofc all the chinese "researchers"
altonen
lol
zzz
more type 5 progress, found a major bug
orignal
great
zzz
I'm not sending NSR at all, I'm sending a new NS in response to NS (((((
orignal
then how I was able to connect to you?
orignal
oh, never mind
orignal
that's why I noticed slowness
zzz
yeah it's really stupid. I assumed if it was working, it was working RIGHT
zzz
bug is when I'm trying to match the static key with the LS key but I'm looking for type 4 in the LS and it doesn't match, so it's "unvalidated"
zzz
I am getting decrypt fails on your NSR, but I have to fix my problems first before we can get to that...
orignal
sure
orignal
we will check