XANA
Testing testing
XANA
dr|z3d: what do you think about I2Pd?
orignal
eyedeekay how come Java pubsihes IP for SSU address with introducer?
orignal
wtf?
orignal
this is big issue confusing everybody
XANA
wdym?
orignal
an SSU address in RouterInfo publishes IP address and introducers at the same time
orignal
eyedeekay
orignal
I was under assumtion that if an address published IP I can connect to it
XANA
I do not know much about it
orignal
that's why I need eyedeekay
orignal
to clarify
orignal
it should publish 4 or 6 in caps rather than IP
orignal
eyedeekay if you publish introducers you shouldn't publish IP/port
eyedeekay
I'll go through our code/comments and find out if that is correct
orignal
no it's not
orignal
that's why we introduced '4' and '6' with grandpa
eyedeekay
Then possibly it's just a bug
eyedeekay
I'll have to make sure but if it is we will just fix it
orignal
let me explain
orignal
I assume that adddress is reachable directrly if it publishes IP/port
orignal
and can be IBGW if ipv4
orignal
but Vort found out a lot of addresses with IP and introducers
orignal
that shouldn't be IBGW
eyedeekay
I can already see how that would be a big problem for you
orignal
since I don't relay on router caps but figure it out from addresses
eyedeekay
In the meantime should/can you ban SSU with published introducers from being IBGW on your end?
orignal
that's what I'm going to do
orignal
just didn't know that Java is still doing it
orignal
it came from old code when there were no 4 and 6
orignal
and you had to publish your IP to let others know your address type
eyedeekay
We'll take a look at removing IP addresses from SSU with introducers and as long as we're not using them for anything on our end it shouldn't be hard
orignal
also remember it's a leak
eyedeekay
If we are using them for something, I'll try to find an alternative to publishing them
orignal
e.g. if your user is firewalled he expects that his IP will not appear in netdb
orignal
again, we did it before '4' and '6'
eyedeekay
Yeah sounds more and more like it's probably a bug
orignal
now if you use instroducers you add '4' or '6' to caps and explude IP
eyedeekay
My only non-bug guess right now is that it might be for backward compatibility with some of the really old embedders
orignal
maybe for backward compatibility
orignal
when are you going yo eliminate SSU and start publishing SSU2 only?
eyedeekay
Depends on a few things, but tbh probably soon now that we've solved the stupid Vuze mystery
orignal
how it's possible?
orignal
SSU through introducers and published NTCP2 with i and IP
orignal
eyedeekay nevermind about the original issue
orignal
Vort said it happens only with IP in NTCP2 and introducers in SSU
eyedeekay
Don't know for sure, probably just needs a check
orignal
but this example is interesting how it's possible
orignal
eyedeekay looks like you don't remove i from NTCP2 if Firewalled
orignal
SSU has instruders
orignal
routers has U cap
orignal
but NTCP2 addresses still have IP/port and i
orignal
example
eyedeekay
Yeah I see what you mean
orignal
Vort syays these IP/port are not reachable from outside
eyedeekay
> <orignal> eyedeekay looks like you don't remove i from NTCP2 if Firewalled
eyedeekay
orignal I took a look at our code and it doesn't appear that this is the case
eyedeekay
It looks like there might be a reason, but I'll need more time to investigate that and figure out whether there's a reason we do it
orignal
then what might cause it?
eyedeekay
What you said, sorry, I think we got mixed up in a double negative
eyedeekay
I read our code, it looks like when we become firewalled we don't remove i from NTCP2
eyedeekay
There's this weird thing we do with it though where we use it to determine if we think somebody has NTCP2, NTCP2 Inbound-only, or NTCP1 though
eyedeekay
I don't know if we still need it
eyedeekay
Will investigate
orignal
if you are firewalled you should remove at least 'i'
orignal
to tell other that this NTCP2 is not eligible for incoming connections