IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#i2p-dev
/2023/10/24
zzz thanks for the careful review eyedeekay
zzz all dev build testers please update to -8
zzz this should stomp out the last of the leaks
zzz and is the first rev I'm recommending for wider testing
zzz dr|z3d, RN, eyedeekay, for dests with an alt EdDSA dest, it appears to me the alt LS is not being built
dr|z3d can't confirm or deny, zzz, nothing with dual leasesets here.
zzz the alt dest stays yellow in the sidebar and I don't see the alt LS in either the main or subdb netdb tabs
zzz ok thanks I'm going to file a ticket but would like confirmation from RN or somebody else
dr|z3d dual leasesets were causing postman headaches, so afaik he's nuked 'em.
zzz was he on a dev build?
eche|off he tested a devbuild weeks ago and went to hell, so gone back
zzz ticket #460
zzz eyedeekay reported broken in -1 and fixed in -5 but that sounded like a client-side problem. This is the server side
dr|z3d he's on the latest now, had issues with the anything built before the subdb stuff before the latest round of fixes which I rolled out over the weekend. Apparently it's all looking good now.
dr|z3d *issues with anything built with
zzz lets look at the diff-since-release which has been the fastest way to find stuff for me
zzz usually the recommended fix is revert revert revert
dr|z3d yeah, well, revert in this case wasn't really an option, at least on the dev end of things.
dr|z3d user-side, I have a pre-subd build on on the release update channel, so not difficult to revert there.
zzz hmm where is that code...
zzz nothing jumps out at me. eyedeekay is going to be a lot better equipped to chase this down and it should be easily reproducible
eyedeekay I'm already looking at it but not yet able to reproduce it, I'm setting up dual-key client and a single-key clients and making them talk to eachother to see what logs show
zzz hmm. you shouldn't need clients, it's just not buildling the ls
zzz if you can't reproduce lmk
eyedeekay I can't see it yet but I'm going to make sure I'm running the right build in about 8 more minutes so I'll know more then
zzz in better news, I fixed the gradle build and CI and eyedeekay fixed the docker CI, so our CI is passing for first time in a while
zzz let me try to reproduce on a clean -8, different router
zzz oh interesting
zzz it reproduced initially, but then about 8 minutes later it finally built the EdDSA LS
zzz like the first build didn't work but a rebuild did
zzz and I didn't get the client-side error
eyedeekay I guess I'll ask the simple question, is there a chance that it's just taking a really long time to build the LS?
eyedeekay For some reason we could find hopefully
zzz it should be instant, because it's the same tunnels as the DSA LS
zzz there's no reason to have one but not the other
eyedeekay Weird...
zzz stopped tunnel and restarted, this time it worked
zzz 3rd time, failed again
zzz and again it fixed itself after a few minutes
zzz but the client side still complained
zzz this may be a race
zzz if so it might be my bug
zzz but looking at your change to AliasedTunnelPool
zzz anyway, lmk if you can reproduce it yes/no
eyedeekay I still can't reproduce it here but I see what you mean by AliasedTunnelPool maybe contributing to the race condition
zzz ok thanks. It does seem racey. Let me poke at it a while and see if I can prove it
zzz found it
zzz this is all about the changes you did to CCR. I told you exactly how to do it and you went off slightly different and horked it up
zzz I thought the changes were poorly done but would work. I was wrong
mesh somebody should teach the jabber.ru people how to setup a jabber server on i2p and connect to it over i2p using some popular clients
RN yes mesh, you should.
RN oi, zzz... looking
RN I have both the regular eepsite tunnel and the additional one, both green-light. (On Canon as well as on Plus)
RN Not sure the timing on startup, but I'll pay attention next time...\
zzz thanks RN, I think I have it tracked down
RN c:)
mesh RN: thanks for nothing
RN (oops wrong hat)
RN any time mesh@j6.i2p
dr|z3d mesh: back at you :)
dr|z3d all very well suggesting "someone" do the work. how about you for a change?
RN dr|z3d, for shame! How dare you suggest mesh@j6.i2p do actual work?
dr|z3d I know, RN, it's work enough conducting the crowd from the sofa :)