~dr|z3d
@RN
@RN_
@StormyCloud
@T3s|4
@T3s|4_
@eyedeekay
@postman
@zzz
%Liorar
+FreefallHeavens
+Xeha
+bak83
+cancername
+cumlord
+hk
+profetikla
+r00tobo_BNC
+uop23ip
An0nm0n
Arch
Danny
DeltaOreo
Irc2PGuest26185
Irc2PGuest52850
Irc2PGuest53061
Meow
Nausicaa
Onn4l7h
Onn4|7h
Over1
acetone_
anon2
anu3
boonst
mareki2pb
not_bob_afk
orignal
poriori_
shiver_
simprelay
solidx66
thetia
u5657
weko_
T3s|4
dr|z3d: thanks for that highly disturbing CCP surveillance link
dr|z3d
always a pleasure, T3s|4 :)
orignal
dr|z3d how would I2P+ react to a router without R or U cap?
dr|z3d
that's a good question, orignal. I don't think we've got an explicit policy right now.
orignal
I want to get rid of them
dr|z3d
why?
orignal
because they tell nothing
orignal
ipv4 publushes IP and ipv6 publishes introducers
orignal
what's the code?
orignal
I meant they worked in jrandom's days when it was ipv4 only
dr|z3d
R in that case.
orignal
what if opposite?
orignal
introducers for ipv4 and ip for ipv6?
orignal
do you understand that R is the worst thing in this case?
orignal
right now I publish these code based on ipv4 only
orignal
but it's also wrong in my opinion
orignal
furthermore, why can't you figure out based on addresses?
dr|z3d
R or U is a handy filter when you're searching for routers.
orignal
searching where?
dr|z3d
it's also handy when you want to reject requests based on reachability. searching in the netdb, for example.
orignal
serching where? in your code?
dr|z3d
I can search for LU in the netdb for example. wouldn't be possible without the U.
orignal
how do you search?
dr|z3d
not in the code, in the netdb search facility we have in java i2p
orignal
doesn't it read routers first?
orignal
when you read RouterInfo you always know if it's R4, U4, R6 or U6
dr|z3d
yeah, sure.
orignal
if you don't read RI how you know if it's valid?
orignal
you must verify signature
dr|z3d
but when you perform a search for LU, for example, the U is necessary, otherwise you're not going to get the results you're looking for.
orignal
when you read RI you can assign these codes
orignal
you don't need to receive in from netwrok
dr|z3d
sure, you could.
orignal
that's why I think to get rid of them
orignal
publish like Xf
obscuratus
This is a problem for me if I enable IPv6. I can only advertise as R, but I'm really R4 and U6.
orignal
oppisite is worse
orignal
you publish R but you are U4 and R6
orignal
and this R confuses everybidy
orignal
for me I'm not sure what to publish in case of ygg only
orignal
I don't publish anything
dr|z3d
ygg would mostly be U
orignal
also what about completely unrechable router?
dr|z3d
at least, from the network's perspective.
orignal
ygg should be R because it can be reached from outside directly
orignal
completely unreachable router is neither R no U
orignal
ah well, there is another supidity called H
orignal
another useless code
dr|z3d
if a router's completely unreachable with no transports, then it's banned until it's contactable.
orignal
banned where?
dr|z3d
in java i2p.
obscuratus
Isn't that what Hidden mode does?
orignal
no I think in I2P+ ionly
orignal
I mean banned in transports or where?
obscuratus
I don't think we ban it if the RI has introducers.
dr|z3d
no, I2P and I2P+ both ban routers if they don't have any published transports. short term ban, assumption being router is starting up.
orignal
obscuratus I meant if it doesn't
orignal
wrong assumtpion
orignal
it might work though a proxy
orignal
it might be NTCP2 only and firewalled
obscuratus
It it's NTCP2 only and firewalled, the RI should have introducers, no?
orignal
no SSU2 at all
orignal
hence no introducers
obscuratus
Ah, right.
orignal
what I think
orignal
to not flood such routers
orignal
because they can't participate any tunnels but their own
orignal
btw, is there a place where dev is being dicussed now?
orignal
besicde our #dev that's limited to i2pd things only
obscuratus
I'd still been using #i2p-dev as the need arises. But activity is lower. Do we need a channel with an owner?
orignal
no, I need a dev channel without echelon
eyedeekay
LS2 is still available
dr|z3d
here, #ls2, #i2p-dev. wherever suits.
dr|z3d
here or #ls2, orignal. either's fine.
orignal
eyedeekay if we should discuss dev on ls2 let's do it
orignal
I though might be another channel
orignal
eyedeekay your opinion about R and U codes?
eyedeekay
I came here to join/participate in dev discussions for I2P/I2P specifically and be around other people writing Java, LS2 should still be the home of multi-router/general protocol discussions IMO
orignal
ls2 seems inactive
eyedeekay
That's true, I've been busy, but if you check the topic I do plan to resume meetings on the 10th
orignal
i2pd doesn't use R and U at all and publishes them for compatibility with Java
obscuratus
I think Java I2P didn't used to use it very much at all, but it's received moer consideration since the last round of network attacks in January and February.
dr|z3d
U cap routers aren't used for local service tunnel builds at all in I2P+, an very occasionally for exploratory tunnels in I2P.
eyedeekay
re: publishing R or U from i2pd, not sure it makes much difference. Maybe it means something if we see a router publish some caps and determine that it's probably lying about them, but one way or the other it entails some kind of test
obscuratus
I think we quit using U routers for exploratory tunnels. But if the RI is reachable, and another router asks us to build a tunnel with them. I think Java I2p will attempt to comply.
dr|z3d
I2P+ also temp bans slower tier routers that are U and older than current.
dr|z3d
zzz stopped using U cap for exploratory tunnels, and then added them back in, minimally.
obscuratus
dr|z3d: Right, I recall seeing that patch now.
obscuratus
What about [char]fU routers. Didn't we stop using them as floodfills?
dr|z3d
they get downgraded.
dr|z3d
in I2P+ they get banned.
dr|z3d
see FloodfillPeerSelector
obscuratus
So what's the proposal on this issue? Quit publishing R/U?
dr|z3d
> Temp banning Router [XXXXXX] for 120m ➜ 0.9.57 / LU
orignal
yes, eliminate R nd U completely
orignal
no publishing nio processing
orignal
for fU routers I just make them as non-FF
obscuratus
orignal: Bring this issue up at the next meeting.
obscuratus
dr|z3d: What do we currently do when we receive an RI that is lacking either "R" or "U". My vague recollection is that we assume they're starting up, and assume "R".
dr|z3d
I don't think we make any assumptions.
dr|z3d
I'd have to look, but it would make sense to check transports and see what the RI is reporting for IPs.
obscuratus
I agree, but I have a queezy feeling about auditing the code base to find the unexpected places where we use the "R" and "U".
dr|z3d
git grep "\"U\"" will give you a good idea.