zzz
0) Hi
zzz
hi
zlatinb
hi
orignal
hi
zzz
what's on the agenda for today?
orignal
SSU2 status update I guess
zzz
ok that's 1)
zzz
anything else for the list?
orignal
release
zzz
ok that's 2)
zzz
1) SSU2 status update
zlatinb
interop testing, although that's part of 2) really
zzz
I'll go first
orignal
oh we have some question about Bob
orignal
if you still remeber how it works
zzz
ok bob is 3)
zzz
SSU2: I've done a little work on hole punch and getting ready for the signed messages
orignal
hole punch message?
zzz
nothing checked in, but it should go smoothly
zzz
yes
orignal
what do you think we should do in SSU1?
orignal
stop using relays there?
orignal
because you have veru good point
zzz
we can't really turn off SSU 1 until we've moved to SSU2
zzz
what's my very good point? :)
orignal
about zero lnegth hole punch
orignal
not turn off
orignal
but don't connect trough interoducers there
zzz
yeah, zero length isn't great
zzz
how are you doing? any progress on peer test?
orignal
ok. should we mention with SSU2?
zzz
I think I added something to the SSU2 spec about it already
orignal
so I started implemnting peer test
zzz
great
orignal
partially
orignal
my question is what is "role" for since we have "msg" field
zzz
one sec let me look
zzz
ok. I put the 'role' in there because it's part of the signature. The 'msg' field is not in the signature
orignal
yes but how an advesary can forge it ?
orignal
if it was not a part of signature?
orignal
also what if msg and role mismatch?
zzz
I think I was worried about alice or bob saving the signed data and using it for something else
zzz
but maybe that's not realistic
orignal
but they can also save with role
orignal
basically there are two possible values
orignal
is nonce part of signature?
zzz
if you don't think it makes any sense we can get rid of it
zzz
yes nonce is part of the sig
orignal
yes, role is notthing but confusion
orignal
initially I thought it's a real role
zzz
but without the "role", could alice or bob take charlie's response and turn it into a request?
orignal
but then found msg
orignal
with the same nonce?
zzz
yes
orignal
well it might make sense
zzz
but alice and bob hashes are part of the sig
zzz
so maybe not
orignal
not they are not
zzz
here's from the spec of what's in the sig:
zzz
prologue: 16 bytes "PeerTestValidate", not null-terminated (not included in the message)
zzz
bhash: Bob's 32-byte router hash (not included in the message)
zzz
ahash: Alice's 32-byte router hash (for messages 3 and 4 only; not included in the message)
zzz
role: 1 byte role of the signer
zzz
ver: 1 byte SSU version
zzz
nonce: 4 byte test nonce
zzz
timestamp: 4 byte timestamp (seconds)
zzz
asz: 1 byte endpoint (port + IP) size (6 or 18)
zzz
AlicePort: 2 byte Alice's port number
zzz
Alice IP: (asz - 2) byte Alice IP address
orignal
then yes I forgot
orignal
then we don't need role
zzz
ok, agreed
orignal
futhermore should be the same format as for relay
zzz
anything else on 1) ?
orignal
no
zzz
2) 0.9.54 release
orignal
and ubuntu 22.04
zzz
go ahead
orignal
these moron swuthed to openssl 3 and didn't include openssl 1.1.1 to repo
orignal
while it's still supported for many years
zzz
what's the impact for you?
orignal
i2pd is well tested with 1.1.1
orignal
and ssumes it
eyedeekay
Hi, sorry I'm late
orignal
altough I have changed to make it buildabled with 3 but nobody seriously tested with it
orignal
that's the issue
zzz
was it removed from debian also?
orignal
and remeber they had regression with siphash in 3
orignal
idk how many other bugs they might have
orignal
no, afaik in ubuntu 22.04
orignal
debian is fine
orignal
and other distros
zzz
interesting. usually ubuntu just takes everything from debian
orignal
even if they use 3 they also have 1.1.1 as alternative
orignal
not reall because ubuntu uses "newer" packages
zzz
so, will this affect or delay your release?
orignal
no, I'm running with 3 for tests
orignal
works good
orignal
but still not sure if there are any regressions
zzz
ok, hopefully you can get some more testers this week
orignal
also compilation produces bunch of deprecation warnings
orignal
yes, people test
zzz
we plan to release next monday
orignal
yes, we are fine
orignal
no SSU2 yes, right?
orignal
*yet
zzz
correct, disabled by default
zzz
zlatinb, your topic 2a) interop testing
zlatinb
yes, basically I need to know when the code for i2pd will be frozen
zlatinb
also the testnet is running bionic containers so I won't be able to test with openssl 3 I think?
zlatinb
after the code is frozen I need a day, although if everyting is fine it should be less
zlatinb
eot
zzz
I assume there's a build option for 1.1.1 or 3?
orignal
it's almost forzen now
orignal
they only upcoming change might be Bob
orignal
but Java doesn't care about Bob anyway
zlatinb
right I"m only testing SSU1 and NTCP2
orignal
no we build we system openssl
orignal
*with
zzz
ok so you handle 1.1.1 or 3 automatically then in the build?
orignal
if I need to build with something else I change Makefile manually
orignal
yes
orignal
and even 1.0.2
orignal
centos 7 still uses 1.0.2
zzz
I'm sure if zlatinb has any problems he knows where to find you ))
zzz
oh I want to add a 2b) topic
zzz
anything else on 2a) ?
zlatinb
no, thanks
zzz
2b) outproxy
zzz
false.i2p appears to have died completely?
zzz
eyedeekay, you have any insight?
orignal
purokishi.i2p works
orignal
even prisoners use it ))
orignal
to get access to clearnet
orignal
because Tor is also blocked
zzz
sure, but that's at the wish of dr|z3d. We can't put it in as the default
orignal
we are not going either
eyedeekay
I don't have any yet but I'll bang on his door so to speak and see what I can find out
zzz
we are working with a group to support a default outproxy. Maybe in 3 months it will be ready
zzz
acetone has been helping them apparently
zzz
nothing more to report now. I guess I'll leave false.i2p as our default for the release, in case it comes back
dr|z3d
haha, orignal. I like that. "even prisoners use it". maybe a new byline :)
orignal
acetone's is Tor's. No?
orignal
dr|z3d seriously
zzz
if I have any more info on the new outproxy I'll let everybody know
orignal
they couldn;t go to clearnet trhough Tor
zzz
all I know is that acetone was helping. I don't know if it's the same as acetone's
dr|z3d
orignal: navalny? :)
zzz
I believe it will be direct, not through tor
zzz
eyedeekay, we never got a bandwidth estimate from mikal either, that would still be helpful
orignal
dr|z3d no
zzz
anything else on 2b) ?
orignal
another drug vendor
zzz
3) BOB
orignal
is there a way to not publish local destination like in SAM?
zzz
the i2cp option should work
orignal
which option?
dr|z3d
there's an option to make the dest session-only, not sure if that helps?
zzz
one sec
orignal
i2cp.dontPublishLeaseSet ?
orignal
thanks
zzz
i2cp.dontPublishLeaseSet=true
orignal
that's waht I wanted to know
zzz
at least it probably worked for us, don't know about i2pd :)
orignal
it's fine
orignal
even if doesn't work we will fix it
zzz
what application is using bob on i2pd?
orignal
reg.i2p
orignal
as I meantioned before
orignal
lookup
zzz
old brain, forgot
zzz
anything else on 3) ?
orignal
to check if as address is online
orignal
we try to find leseset
zzz
yup
eyedeekay
At least when used with SAM, a simple test to connect an unpublished client to an unpublished client does not seem to work
orignal
inr.i2p did it the same way
orignal
eyedeekay it definitly works in SAM
eyedeekay
Yeah the option seems to have the correct behavior there
orignal
how would it work if it's unpublished
orignal
it measn you can't find leaseset
orignal
unless you have a leaseset alreadt
zzz
anything else on 3) ?
orignal
no
zzz
anything else for the meeting?
zzz
everybody please test test test this week, both i2pd and java
eyedeekay
Not much but since I promised doing go-i2p weeklies I'm actually slightly ahead of the schedule I set last week
zzz
great
zzz
forgot to hold your feet to the fire today
zzz
thanks everybody
eyedeekay
I should have common structures done entirely by Monday and be on the road to crypto refactor