#ls2 (irc2p) | #ls2 I2P Protocol Development Discussion - Next Meeting Monday, October 9th 6:30 UTC

orignal I might be later today

orignal hi

eyedeekay Hi everyone, meeting time, who is here today?

eyedeekay Hi orignal

orignal me ))))

RN aloha

eyedeekay Only big topic I wanted to make sure we covered today are SSU2 changes

xeiaso hi

eyedeekay Any other topics for today?

eyedeekay hi RN, hi xeiaso

eyedeekay Last call, topics for the agenda

xeiaso maybe on release? when is next?

eyedeekay Probably late August or early September for 2.4.0

eyedeekay So let's make that 1 and get it out of the way

eyedeekay 1) Release Status 2) SSU2 changes

eyedeekay Java I2P is releasing libraries to maven and downstreams and my subproject(easy bundle) will follow on that, should be done today and that will conclude our 2.3.0 release

orignal release status I guess

eyedeekay you guys have been done for a week or two right orignal?

orignal yes

dr|z3d hihi

orignal we see tunnle creation success rate still low

orignal but looks like it's cause by LU routers

eyedeekay hi dr|zed

orignal also sometimes we pick non-ipv4 routers for OBEP

dr|z3d seems to be a surfeit of LU routers on the network right now.

orignal will be fixed soon

orignal beside this everything else works good

xeiaso so you will not build tunnel through LU?

orignal we will

orignal we pick L for exploratory tunnels only anyway

orignal we need to fix enpoints however

eyedeekay What's wrong with them?

orignal as I said sometimes we pick non-ipv4 routers

orignal that's not right and will be fixed

eyedeekay Oh I see that's all, OK

eyedeekay Glad you're on it

orignal Vort did many tests to see what causes low rate/failures

orignal it's minor issue anyway

eyedeekay Yeah

eyedeekay Our ETBS is not very different from before, too early to say but only one of our changes is likely to have had any effect on that

dr|z3d a casual 72% avg over 24h here.

orignal because you exclude LU

dr|z3d in France it's good luck to tread in dog shit. doesn't mean you need to find some to tread in :)

eyedeekay Sounds like an excuse to not clean it up

RN 85 and 81% here

orignal we prefer to not disciminate LU

eyedeekay I'm simply not ready to accept the numbers I have in front of me yet. Want to see BiglyBT update and everything else first

orignal we tried to exclude LU and the rate becomes 2 times higher

eyedeekay Anyhow, when do you want to do the next release? I'm proposing early September

orignal fine for me

orignal what will it include from your side?

eyedeekay Lots and lots of netDB fixes, and if you're still ready to work on it with me, SSU2 fixes

xeiaso speaking netdb fixes did you see the message i sent in #i2p-chat earlier about local lease sets and receivedAsPublished?

orignal no I mean new features

orignal I need to be aware

orignal no I'm not presented on any echelon's channels

eyedeekay If it's the one I'm thinking of it was for me and yes xeiaso I did

xeiaso yes i was talking to idk and his netdb fixes

eyedeekay RAP handling made it into the release

eyedeekay IIRC we conversed on the subject there

orignal what is RAP handling?

Xeha whats RAP?

eyedeekay receivedAsPublished

xeiaso RAP handling is a netdb hack to prevent linking stores

xeiaso it doesn't really work

eyedeekay Yeah

eyedeekay That's what I mean by lots and lots of netDB fixes, we've got the meta-netDB proposal to work on this release, it's likely to be the bulk of my time

xeiaso the problem is that local LS stores to a destination are stored but overwrite what the router stores as published

orignal so you want to come back to metatLS?

orignal please let me know because I have only stub for it

orignal but the main question is what for

eyedeekay No, meta-netDB, there's not really a thing you'll need because I gather it's fairly simple for you to use separate netDB's for clients and the router, for us the incumbent design doesn't really accommodate that, so we have to do it differently

orignal got it

eyedeekay What it is is an API for organizing access to and from our netDB by offering the ability to manage them by context, in the end every client will be able to get it's own netDB from a sort of "parent" netDB

eyedeekay So it won't have to break literally everything that touches the netDb in our application

eyedeekay But that's a java-specific topic for the most part, and not really a good topic for the meeting, I owe a writeup on our gitlab issue(395 IIRC), will post it here when I finish it

orignal yes

eyedeekay Ready for 2) SSU2 Changes?

dr|z3d shot, eyedeekay

dr|z3d *shoot

orignal yes

orignal but I didn't have time to think about it

eyedeekay Are we still basically settled on the send Alice-sends-Bob his own identifier in the first phase of the handshake, Bob rejects it if Alice is wrong solution?

orignal no

orignal Alice sends Bob's ident hash

orignal right

eyedeekay Yes

orignal and Bob rejects if mismatch

orignal we need two things

orignal 1. new block type

orignal 2. new termination code

eyedeekay Ack makes sense

eyedeekay Can you still write down the changes to the spec so we can both implement it?

eyedeekay For 2.4.0 in the timeframe we proposed?

orignal when I have time

orignal unfortunately I'm too busy

eyedeekay We both have a ton to do, if you don't have time for the SSU2 changes I can focus on our netDB needs until you are ready

orignal fine

eyedeekay Anything else for 2)?

orignal this SSU2 fix is not a big deal

dr|z3d since we're on SSU2, when are you planning on disabling SSU1, eyedeekay?

eyedeekay Hadn't planned on doing it yet, seems like a change I'm not ready to make to lop off the whole mystery-router section

dr|z3d ok

eyedeekay the 0.9.23's specifically, I want to know what those are before we do anything, the explanations I have are conflicting at best

dr|z3d 0.9.23 I suspect is Azureus holdouts.

orignal you should

orignal gradpa promised to deisable it

eyedeekay Gotta be, Vuze is supposed to be using the BiglyBT version of the plugin now

orignal and start publishing it as SSU2

eyedeekay If we find out about the 0.9.23 contingent I'll consider it but for now I'm leaving it on

eyedeekay Anything else for the meeting?

orignal no

eyedeekay OK thanks everyone for coming, I'm circling back to the the last of our release items and will post about meta-netDB as soon as I've got the proposal finished

eyedeekay See you around IRC/in 2 weeks at the meeting

xeiaso ok