Overview and Logs for the tini2p Dev Meeting Held on 2019-09-19

Gönderen: el00ruobuob / oneiric

Logs

<tini2p_gitlab> time of the meet
<tini2p_gitlab> 0: Greetings
<tini2p_gitlab> h
<tini2p_gitlab> 1: What's been done
<tini2p_gitlab> A lot
<tini2p_gitlab> Finished tunnels impl, and merged into master
<tini2p_gitlab> (with helpful review from @DavidBurkett)
<tini2p_gitlab> I'm still used to being the only person working on the project, so have to habituate myself to a workflow that includes other people
<tini2p_gitlab> started work on the router context (what ties all the pieces together)
<DavidBurkett> Don't change anything because of me. I'm slow and unreliable
<DavidBurkett> I'll catch up eventually, just keep doing what you're doing :)
<tini2p_gitlab> in the process, came across some impl bugs in NT*****2 (fixed), updated NT*****2 to include a testnet v. mainnet netID, and started working on ECIES updates
<tini2p_gitlab> no worries, I just meant that you're the first other than me to review my code :)
<DavidBurkett> Haha, got it
<tini2p_gitlab> for the ECIES updates: there has been a lot of them
<tini2p_gitlab> zzz pushed updates to move us to something much closer to standard Noise, and I love the changes
<tini2p_gitlab> it's a bit of work to change my impl, but all the changes are for the better
<tini2p_gitlab> the code is looking much cleaner now
<tini2p_gitlab> the cryptographic guarantees are stronger and more clear as well, which is the bigger win imo
<DavidBurkett> :rocket: Do you have a link to zzz's updates?
<tini2p_gitlab> you can follow my work in my context branch: https://gitlab.com/tini2p/tini2p/tree/context
<DavidBurkett> Ah perfect
<tini2p_gitlab> yeah, zzz's updates are pushed to the main 144 proposal: https://geti2p.net/spec/proposals/144-ecies-x25519-aead-ratchet
<tini2p_gitlab> we're still working out some kinks in the handshake, and will need to update the ratchet KDFs afterwards
<tini2p_gitlab> we are getting **much** closer to finalizing the spec though, which is awesome
<tini2p_gitlab> so, my impl may deviate from the spec a bit until we get everything figured out
<tini2p_gitlab> we're trying to get as close to standard Noise as possible, while maintaining the necessary properties for ECIES-X25519 stated goals
<DavidBurkett> That's great news
<tini2p_gitlab> definitely
<tini2p_gitlab> :smile_cat:
<tini2p_gitlab> the downside to all that is I've spent a good amount of time refactoring, so don't have code that is ready for a release candidate
<tini2p_gitlab> long-term it's for the best, in the short-term it's a slog to get to alpha release
<tini2p_gitlab> which leads to the next item
<tini2p_gitlab> 2: What's next
<tini2p_gitlab> continue the refactors for ECIES-X25519
<tini2p_gitlab> get something working for end-to-end crypto sessions, and work with zzz on fixing spec bugs (if/when they come up)
<tini2p_gitlab> finish the context impl
<tini2p_gitlab> create the docker testnet for local integration testing
<tini2p_gitlab> idk how long the docker testnet will take, but I'm hoping something minimal will be fairly easy to setup
<tini2p_gitlab> if it takes more than a day or two, I'll push an alpha release candidate without the testnet
<tini2p_gitlab> the biggest hurdle I see for the testnet is populating the NetDB of the routers with the RouterInfos of the other routers in the testnet
<tini2p_gitlab> I don't have reseed impled, so I'll have to do something manual
<zlatinb> you may find this useful: https://zzz.i2p/topics/2756-testnet-with-lxc-instructions
<tini2p_gitlab> really hoping this is the last time I'll have to say this: pushing the alpha release date back another two weeks
<tini2p_gitlab> awesome! thank you @zlatinb :)
<zlatinb> also talk to "obscuratus" on Irc2P, he has a mixed i2pd/java i2p testnet, also with lxc
<tini2p_gitlab> right on, I'm hoping to do it with docker for eventual windows support
<tini2p_gitlab> right now, tini2p only builds on linux. will be working on cross-platform builds post-alpha
<tini2p_gitlab> there's probably a lot of crossover in approach to lxc, so I appreciate the resources. will definitely be reaching out to obscuratus
<DavidBurkett> Builds on mac also
<DavidBurkett> But of course, that's based on linux
<DavidBurkett> I nearly had it building on windows, but I got frustrated and quit :)
<tini2p_gitlab> haha wow, thanks for build testing
<tini2p_gitlab> yeah, I imagine the windows experience is pretty hellish atm
<DavidBurkett> np, wasn't intentional. I'm the 0.1% of devs who prefers windows
<DavidBurkett> So I tried there first
<DavidBurkett> It wasn't quite as bad as you might think, but 1 of the dependencies was pissing me off lol
<tini2p_gitlab> noise-c?
<DavidBurkett> Yes, that's the one
<tini2p_gitlab> yup
<tini2p_gitlab> that's a bear
<tini2p_gitlab> I'll be moving all the noise stuff into the project proper, and remove the noise-c lib
<tini2p_gitlab> it's a great reference, but making it cross-platform is such a pain
<DavidBurkett> Excellent. Let me know when you want me to try building again
<tini2p_gitlab> will do, I appreciate the offer
<DavidBurkett> :thumbsup:
<tini2p_gitlab> long-term I hope to contribute a CMake build system to noise-c
<tini2p_gitlab> atm that's low priority though
<tini2p_gitlab> that will be extremely time-consuming, but I want to give something back to the project
<tini2p_gitlab> rweather did a lot to help me understand Noise :)
<tini2p_gitlab> so, that's more-or-less what I'll be working on: router context, ECIES updates, some NetDB stuff, docker testnet
<tini2p_gitlab> 3: Questions/comments
<tini2p_gitlab> alright, so that's the meeting
<tini2p_gitlab> will do my best to get a release candidate out by next Thursday, and alpha release the following week
<DavidBurkett> Looking forward to it
<tini2p_gitlab> :+1:
<tini2p_gitlab> 4: Next meeting
<tini2p_gitlab> 2019-10-03 18:00 UTC
<tini2p_gitlab> holy shit, we're already almost in October
<tini2p_gitlab> thanks all for attending
<tini2p_gitlab> @tini2p_gitlab twirls the gavel


Post tags : Dev Diaries, Kovri I2P Router, Cryptography