IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#ls2
/2022/03/14
zzz 0) Hi
zzz happy daylight savings
zzz whats on the agenda for today?
zzz how about 1) SSU2
zzz anything else?
eyedeekay Briefly strange/inconsistent Wikipedia blocks
zzz ok that's 2)
zzz anything else we can add at the end
zzz 1) SSU2
zzz I'll go first with status
orignal yes ssu2
zzz went live on the net saturday
zzz or was it yesterday?
zzz anyway, turns out drzed went live a day before I did? so any i2pd crashes were before that
zzz I've been fixing bugs, mostly retransmission stuff
zzz and also related to NOT doing peer test or relay with SSU 2
zzz but everything looks pretty good so far
zzz orignal, what's your status?
orignal so as you see I started publishing SSU2 addresses
zzz yes
orignal very close to finish handshake
zzz great
orignal fixed all issues we discussed last week
zzz so, to let you know: the live net is enforcing tokens, so you'll need to support Retry and Token Request
zzz if you want to test without that, you can test with my test code
zzz I'll put the tester up if you need it, let me know
orignal as we agreed
zzz yes
orignal I send non-zero tokens
orignal beside it I don't handle them yet
zzz I also want to repeat that I would never enable SSU 2 by default until we're done with joint testing. There definitely could be mistakes on my side still
zzz take your time
orignal I'm still far away from it
zzz don't feel any pressure, it will be done when it's done
orignal and SSU2 will have lowest priority in my code
orignal even after yggdrasil
orignal e.g. I try NTCP2, then SSU, then ygg and then SSU2
zzz any other status or any questions for today from anybody?
orignal are you going to publish it as SSU2 later on?
zzz probably not until we turn off SSU 1
zzz unless I find time to completely refactor things
zzz like if I add ygg support then I could switch to SSU2 also
orignal like you did for NTCP2
zzz yeah I know what I need to do, it's just not going to be fun
orignal SSU2 over ygg is a bad idea
orignal UDP doesn't work well there
orignal just FYI
zzz no, not SSU2 over ygg. Just that handling multiple published addresses (besides just v4 and v6) requires a big refactor
orignal well it works but slow
orignal got it
orignal beside it you should
zzz agreed
orignal btw are you alble to find SSU2 address in my RI?
orignal because it goes after ygg address
zzz I see it in the netdb; don't try to connect because no "i" and no host/port
zzz actually that's one bug I fixed and will be checking in soon, I try to relay if there's "i" but no host/port
orignal no I mean are you able to process it of you stop on ygg
orignal "i" will be published just need to restart
zzz I can process it in theory, I don't "stop on ygg"
orignal no host/port is normal sitiuation
orignal if there is "4" or "6" in caps
zzz right, but as I said, I have bugs with trying to do relay and peer test on SSU 2, still fixing those
zzz any other SSU2 status or any questions for today from anybody?
orignal not now
orignal will be back to you in few days
zzz 2) Wikipedia
eyedeekay So one of the redditors posted last week that they had been blocked from editing Wikipedia, apparently for no other reason than having an I2P router on their system
eyedeekay If this is true, that's a really big deal, so I looked into it
eyedeekay I'm not sure what I've seen is conclusive, but I think something weird is going on
eyedeekay It would appear that, at any given time, no less than 15% of my peers are banned from wikipedia, with 55% being the highest I observed
eyedeekay I think this is suspicious, but I think it may be because a bot which draws information from a third party(Referred to in their docs as an API Partner) is accidentally misclassifying us
eyedeekay So I'm planning to share the scripts I wrote to measure the effect, tell them what I think is going on and try to work with them to resolve it
eyedeekay If you'd like to be cc'ed on the email I can add you, I have not contacted them yet
zzz It is interesting. I don't give a crap about editing on wikipedia, but if there's some third party out there scraping the netdb, that's not good
eyedeekay Yeah I thought it was creepy to say the least
zzz to be clear, this is people just running i2p, but accessing wikipedia directly? (not via outproxy?)
eyedeekay Yes that was the first thing I ruled out
eyedeekay And the information is more than a little confusing
zzz so, there's a couple things we could do here
zzz either before or after trying to find out who the "API partner" is, I could ask Lance James for help
zzz he might know who it is or be able to find out
orignal why do we even care about this Pedipedia?
zzz yeah, lets make that clear. what's the goal here eyedeekay ?
orignal excatly
eyedeekay Because getting people banned from sites for running I2P is a terrible precedent. I get it with open proxies and exits, but just enumerating and banning users is not good
orignal Pedipedia doesn't like i2p because they are faggots. so?
zzz I don't care about wikipedia
orignal me too
zzz I do care about somebody making a list of i2p users
orignal bunch of agencies do
zzz can we do something to stop it, either technically, or by asking them to stop
orignal surprised?
eyedeekay I kind of assume that lots of somebodies are attempting it
zzz no, but somebody selling the info to wikipedia isn't great
eyedeekay To be clear, I haven't confirmed that is even what is happening yet
eyedeekay To do that, I would probably have to contact them
zzz for example, is it a group of floodfills that we should ban?
zzz eyedeekay, what's your goal?
orignal I doubt it's because somebody runs an I2P router on theier IP
eyedeekay I'd like to either confirm or rule out my suspicion
zzz and if confirmed, what's the next goal?
orignal I think an offcial confirmation from Pedikedia
orignal why they do it
eyedeekay My first step would probably be to try and get Wikipedia to stop doing it.
eyedeekay But after that, I'm not sure. I don't think that such an approach is widespread, but I also think it's inappropriate and something we should know how to stop
zzz I think why is obvious
zzz they're trying to stop spam via VPN/Tor/etc, and they're misclassifying i2p users as a vpn exit. right eyedeekay ?
eyedeekay I'm not sure it is, my instinct is to go with "by mistake" for exactly that reason
eyedeekay They're misclassifying us as a VPN exit because the classifier is too coarse
zzz ok. you can keep me off the CC. If you can find out who the api provider is and/or want to get help from lance, let me know
eyedeekay Will do
zzz anything else on 2) ?
eyedeekay Not from me
orignal I suspect something else
orignal probably that guy also runs Tor exit node
zzz eyedeekay, have you ruled out tor exit nodes?
orignal it would make sense
orignal if they believe he is an exit node
eyedeekay That's not entirely out of the question, it's true for some of them in fact
eyedeekay Part of the reason the info is confusing is because some of the bans are for shared hosting, some of the bans are for known Tor exits, and a whole bunch of them fall into this P2P VPN classifier which means nothing
zzz he said 15-55% of routers banned? that's way more than just tor exits
eyedeekay Indeed, it doesn't account for all of them
zzz anyway
zzz anything else on 2) ?
zzz anything else for the meeting?
zzz thanks everybody
zzz happy bug hunting
eyedeekay you too
orignal also I think it will be fun with Russian IPs soon