IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#saltr
/2024/11/15
dr|z3d which, in theory, means they'll mostly try NTCP first and only use SSU2 inbound once you've exhausted your NTCP slots. If I'm wrong, I'm sure zzz will let me know :)
uop23ip ok, thanks. i thought it would be the opposite, ssu2 newer/better/lighter and more used with new routers :)
dr|z3d all about the "cost" of connection to a specific transport.
dr|z3d if you hover over the NTCP2 / SSU2 labels in your netdb section where routerinfos are displayed, it should show you the cost of the transport.
orignal peers prefer NTCP2 because there is an issue with SSU2
orignal i2pd connects through NTCP2 to a new routers
orignal and uses SSU2 only is router has proven itself before
orignal sorry guys I have decided to give up with this issue
RN is SSU2 that unreliable?
orignal RN it can be not that router your are connecting to
orignal you are connecting to router with ident xxx, xxx in RI copies SSU2 address from router yyy
orignal you think you connect to xxx but actually you connect to yyy
RN oh, because of attack attempts?
orignal yes, it's a big hole in the protocol
orignal that's why i2pd connects to xxx through NTCP2 first
RN yeah, I think I recall a bit of discussion about this
orignal and one it see that xxx publishes real IPs can use SSU2
orignal yes, it was a long dicussion
RN does the ntcp2 connection give the ssu2 address?
orignal but since Java huys doens't want to do anything with it for undisclosed reason I gave up
RN I mean, by using ntcp2 can you get the ssu address from the target router?
orignal when you connect through NTCP2 you cannt's connect to anpther router
orignal because identhash is part of SessionRequest
RN right, so you know you have the right router, can you ask it through ntcp2 to give you its ssu2?
orignal when I sucessuffylly connect though NTCP2 I confirm that a router is real
RN instead of what is published?
orignal I don't need to
orignal I see bpth addresses in RI
RN ahhh.
orignal the attack was that a router I was trying to connect to never existed
orignal they just used IP from real router
orignal however if I can connect through NTCP2 it means it's real
orignal that's the answer why more NTCP2 than SSU2
orignal also i2pd doesn't care about cost
RN ok, that makes sense
orignal why Java guys doesn't want to fix this problem it's mistery for me
orignal although it could be fixed many months ago
orignal as zzz suggested to add one more round to MixHash with remote ident hash
orignal and that's all
orignal futhermore NTCP2 has too much noticable overhead now
orignal say my NTCP2 thread consume 2 times more CPU than SSU2
uop23ip mmmm, i got 3k+ transit tunnels, wouldn't i at least get some ssu2 inbounds and not 0?
orignal how many transport sessions do you have?
orignal can someone explain why I always get disconnected from here
orignal looks like LeaseSet update issues
orignal maybe it comes too late
uop23ip sorry forgot to mentioned that i am on plus, orignal
cumlord Is there a difference in how i2pd and Java handle that?
orignal idk how Java handles it
cumlord I originally put my irc server on i2pd then switched to multi homed i2p+ and trying to see if difference in disconnects
orignal I have 7500 tranbsits
orignal around 2000 SSU2 sessions
orignal and more than half are incoming
uop23ip if i2pd prefers ntcp2 (or complete no ssu2), my outbound ssu2 can only be java routers which doesn't use me because of costs, or so maybe
orignal wrong. i2pd doesn't prefer NTCP2
orignal it prefers NTCP2 for unknown routers only
uop23ip and then switch if trusted?
orignal you should run irc on i2pd
orignal because I get disconnected from here every few hours
orignal and never from Ilita
orignal it chooses randomly SSU2 or NTCP2
orignal * [orignal] idle 00:38:18, signon: Mon Nov 11 19:43:39
orignal see at Ilita
orignal * [orignal] idle 00:00:21, signon: Thu Nov 14 21:48:36
orignal and here
cumlord I switched back to i2pd last night, looks like less disconnects so far from the logs
orignal that's how irc on i2pd works
RN orignal, is your irc client's ping timeout set too low? it could be your irc-client disconnecting...
orignal znc and irc.ilita.i2p are on different routers
orignal RN the same config of znc for both
RN for this server it is reccomended to set it to 320
RN yeah, not the bouncer, but the client...
RN I don't rmember if there is a ping timeout setting in znc
RN for hexchat for example: /set net_ping_timeout 320
orignal but my bouncer keeps disconnecting not client
orignal will check in znc
RN and your bouncer doesn't try to reconnect more than 2 times in 2 minutes?
orignal not sure
orignal but see another issue
cumlord My relays had issues with reconnecting too fast to postman
orignal you keep disconnecting from Ilita frequently
orignal no. RN
RN hmm...
orignal seems Java has some issue with long-live connections
cumlord I don’t think I get many disconnects, client side I use Java
orignal last time you reconnected around 3 pm
orignal I mean RN
RN I didn't consider the re-connections "frequent" but I see that with most irc servers
orignal the picture looks similar to me here
RN and I am on 2 that use i2pd and two that use Canon, and one that I'm not sure which
RN no miscounted
cumlord see anything on simprelay on ilita
RN 3 Canon 2 i2pd 1 unknown
orignal <orignal> * [orignal] idle 00:38:18, signon: Mon Nov 11 19:43:39
orignal see how it should be
RN all behave about the same as far as connection longevity
orignal why monday? Just because I rebuilt i2pd for that router
cumlord I wonder if it’s to do with multihoming and delay with ping timeout
orignal RN but your client is always Java
RN yes
RN both flavors
RN (at least when you see two of me)
orignal seems only i2pd-i2pd connection are stable
orignal and can stay for months
orignal and any connections with Java produce disconnects every few hours
RN I kinda got used to it from back in the day when netsplits between the (then) two leaves and main ircd were often daily
orignal I can add logging but I'm pretty sure it will be no LeaseSet
RN so I was used to much worse, and it has improved a lot over the last 20 years or so
RN my expectations perhaps are too low
orignal I know how it was
orignal i2pd became that stable maybe like 6 months ago
RN I did not find any ping timeout setting in the webui for znc
orignal when Vort has implemented new logic to switch between remore leases
RN maybe something in a file that can be changed, but not exposed in the ui
orignal yes, but it doesn't affert Ilita
orignal I think it's about leasesets not about ping timeout
cumlord maybe id give i2pd a try for irc connections to compare but i haven’t had issues with java for client side
orignal * [cumlord] idle 00:00:28, signon: Thu Nov 14 03:48:18
orignal not impressive really
RN might be, though ping timout does reduce the frequency of disconnects because some (stupid) irc-clients send disconnec when they think they timed out
cumlord try simprelay or pisslord
cumlord I did some restarting anyway though so it probably wouldn’t be long now that I think of it
orignal * [pisslord] idle 04:28:02, signon: Thu Nov 14 03:48:18
orignal the same
orignal and see this
orignal * [R4SAS] idle 53:57:19, signon: Tue Jul 9 18:05:23
orignal he is connected to another server
RN do both of mine disconnect? the one with and the one without a tail on the name?
orignal * [AreEnn_] idle 00:59:20, signon: Thu Nov 14 21:21:25
RN 4 months? and that is not a server on his localhost?
orignal not long uptime
orignal right, maybe
orignal probably his is localhost
orignal while my znc is not
RN neither are either of my znc
orignal * [un] idle 33:12:05, signon: Sun Nov 10 04:05:24
orignal another example
RN but one is in same lan
orignal seems nonody here stays more than a day without disconnect
orignal * [dr|z3d] idle 04:16:13, signon: Thu Nov 14 01:34:23
dr|z3d that's normal for me, orignal
RN yeah, postie is probably a local host connection too
dr|z3d router restarts.
RN you restart all the time dr.
orignal do you remeber if you stayed here without disconnects for longer time?
RN we have someone from nov 12
orignal like a week?
RN [Liorar] idle 57:01:59, signon: Tue Nov 12 10:22:58
RN sorry for the highlite Liorar
orignal local?
orignal well nev 12 is not week
RN no. just poking around to see who has longest aside from our host
orignal * [zzz] idle 07:41:49, signon: Wed Nov 13 21:10:07
orignal but would be nice to see an example with uptime like a week
dr|z3d I can't get excited about random disconnects, bigger fish to fry.
orignal I'm trying to say that Java code has some issue with sending leaseset updates
dr|z3d possible. zzz fixed, I sometimes still see leaseset updates not being published in a timely manner, but a multihomed setup should ensure that leasesets are available in the event one of the multihomes is late in sending an update, no?
orignal not for client
orignal cleint doesn't publish
orignal it must send with stream data
dr|z3d I thought we were talking about IRC server?
orignal RN gets disconnected from Ilita server becuase no update leaseset from her
orignal we are talking about IRC connection in general
orignal RN has the same issue at Ilita as I have here
orignal I think the same cause
dr|z3d ok, more refinements to /advancedconfig in advanced mode in latest + build, dark theme for now, others to follow, please test and report any issues. thanks!
dr|z3d configs now limited to viewport, so no scrolling to get to the save/cancel buttons.
zzz <orignal> but since Java huys doens't want to do anything with it for undisclosed reason I gave up
zzz <orignal> why Java guys doesn't want to fix this problem it's mistery for me
zzz 165 / i2pdSSU2 hash in handshake?Proposal 165 reviewed late 2024-04, awaiting update from authors, no ETA Additional suggestions contributed in May
orignal zzz. and?
orignal we agreed about one more MixHash
zzz Proposal 165 "awaiting update from authors"
zzz standing by for 6 months and counting
orignal looks like you are not interested
zzz it's on my roadmap
orignal then when?
zzz I gave you two rounds of feedback, in late april and early may
orignal but be realistic
orignal nobody is going to do it
orignal I mean update proposal
zzz I'm standing by until you guys give us an updated proposal
orignal that mens never
zzz ok then, but don't blame "java guys"
zzz or pretend it's a "mystery"
orignal you are intersted to fix this hole in the protocol so much that wait for proposal update from some random guys
orignal who migh not be on board anymore
orignal that's my point
zzz you're not a random guy. I am interested
zzz but you think people fart out ideas and then I do all the work? no
orignal well I'm not a random guy, but I'm a non-writing proposals guy ))
zzz then find somebody on your team who can help
orignal I think you consider this issue as non impprtant
orignal that's all
orignal the problem is it's not an idea
zzz it's on my list, and I spent a lot of time reviewing it and giving feedback. but there's nothing more I can do until you guys have updated it
orignal it's bug in the protocol
orignal what if it will never happen?
zzz sure but until we agree on how to fix it, there's nothing I can do. It's on my list for 6 months, I'm standing by
orignal you want this bug last forever?
orignal we already agreed
zzz if it's important to _you_ then you will incorporate my review comments so we can keep moving
orignal can I just request to close this proposal as wrong?
zzz maybe we agreed, maybe we didn't, but it's not written down so I don't remember what we agreed to
orignal no, for me it's not important anymore becuase we have good woraround
zzz if you want the proposal closed and you want to start over, that's fine, doesn't matter to me
zzz if it's not important to you, stop blaming "java guys"
orignal I really can't understand your point
orignal I'm not blaming
zzz if you want the protocol fixed, then write down what you want to do about it
orignal the question was why less SSU2 sessions than NTCP2
orignal I naswered why
zzz I agree we made a mistake, we should fix it
orignal so, you actually want me to write a proposal instead have the bug fixed?
orignal then maybe I2P needs own DOGE )))
zzz you already wrote the proposal: 165
zzz I gave you extensive feedback after review
zzz and I'm standing by awaiting your update
orignal do I understand you right than we agreed how to fix it but it matters who wtites a proposals
orignal not me, weko did
orignal we did it because you were not there to push idk to do something
orignal instead taking it seriously and think
zzz we discussed it and I gave you feedback. I don't know if we had 100% agreement on everything
orignal he started prodducing some bullshit
orignal let me finish
orignal after that we came to conclusion that nothing can be achieved this way and implemented that workaround
orignal and give up with the proposals
orignal when you came back I raised it again
orignal and you offered right solution
orignal but this proposals is not something we plan to come back
orignal because
orignal 1. weko is busy in somthing else and have different prioritiis on his live
orignal 2. I'm nuts in writing any text iincluding proposals
orignal simply speaking it's not because I don't want to I just can't
zzz ok, but the proposal is dated 2024-01-19, I was already back for 4 months
zzz this is complicated, it needs to be written down. please find somebody to help you
orignal remeber I tryid to write a proposal abouyt something you found 3+ mistakes in every sentence
orignal idk why it was data in 2024
orignal it was wrtiteen in 2023
zzz who cares about english mistakes, the tech part has to be right. I'm not an english teacher
zzz I understand how we got here, the question is how to move forward
orignal nobody wants to read a proposal with so many grammar mistakes
orignal definitly we need to come to final agreement manybe at i2p-dev, record the logs and someone should write a proposals
zzz if it's not written down, I don't know what we agreed to, and nobody can implement it. I'm standing by waiting for an update and a 2nd review
orignal why not idk since he is native englsh speaker?
orignal we agreed about one more MixHash
orignal in SessionRequest
zzz it needs to be documentd so it can go into the ssu2 spec so idk and hk can implement it
orignal how about to setup a meeting at i2p-dev and discusss and come to solution?
zzz your idea is that i2pd comes up with all the ideas and java does all the work to document them? no
zzz you think we have a solution, it's just not written down? so write it down
zzz you're a technical and professional person. I don't have to explain why specs are important. Go find somebody to help you if you need it
orignal how can I write it down?
orignal let me forumate ths problem
orignal i2pd has dicovered the problem during an attack
orignal investigated and fouund the source of the problem
orignal assume we didn't have a good solution yet
orignal sending ident hash block is bad one
orignal it was just a dumb idea
orignal first thing that came to my mind
orignal simply speaking this proposal is just wrong
zzz so now it's time to fix the proposal and make it right
zzz I spent a lot of time reviewing it and giving feedback. Now please do us a favor, do the work, and incorporate that feedback into the document
orignal yes, it's time to fix
zzz super
orignal how can I make a change in the proposal? make a PR at github?
zzz better to just email me a patch, or a PR on our gitlab
zzz I don't have permissions on our github
orignal I prefer PR at github
orignal huh? why?
zzz lol if you're going to do it your way, then don't ask me )))
orignal no I'm asking aboutthe proper way
orignal so you want me to send an e--mail to you?
orignal and you are going to review the changes manually?
zzz that's the easiest for me, but we can make it work whatever
zzz I'll check it in and then we can all review it. I'm not going to review it before checking in
orignal what is the proposal # again?
zzz 165
orignal will make an update and send e-mail
zzz super
orignal and grammr mistakes is your reposibility )))
zzz I am just a humble reviewer. It's everybody's responsibility to review
orignal can you remind your idea how to handle SessionRequests from older versions of Alice?
orignal because you don't know if it's older version when receive SessionRequest
orignal I know you had something but I forgot
zzz please review the IRC logs with my comments in late April - mid May
zzz because I don't remember all the details either
zzz I do remember that my idea was two flag bits
orignal thanks, that's enough
orignal I'm not sure if I have logs
zzz major has them ))
orignal zzz in mean time please think about IRC disconnects
orignal what causes it
dr|z3d orignal: as I've mentioned previously, you can write proposals with ai assistance, in Russian, and then have the ai engine translate the proposal after you're happy with it.
orignal it will produce shit
dr|z3d no, it won't.
dr|z3d *** chuckles. ***
not_bob Using AI to write something for you will get you at worst something average in quality. Not great, but quite passable.
not_bob I use it often "How could I phase this better?"
not_bob That does not mean I just take what it spits out as best though.
dr|z3d depends.
not_bob *** nods. ***
dr|z3d it's a tool.
not_bob It's a tool that can be quite useful if used right.
not_bob hahaha
dr|z3d if you're a tool, it'll spit out shit.
not_bob Well, there is that.
not_bob But, I don't think orignal is a tool.
dr|z3d I wasn't suggesting a proposal from scratch. I was suggesting it as an assist in editing what we already have.
not_bob *nod*
not_bob Normally I stay out of things like this. But, I was just trying to suggest that orignal use a tool to make his life simpler.
not_bob As others have suggested.
hk not_bob: speaking of AI, have you considered using AI as an auto-moderation tool?
dr|z3d others haven't suggested. I did :) that's singular.
not_bob Either way ;)
hk for the great you've done, I think you would benefit quite a lot; at the very least it can flag things for you
hk great work*
not_bob hk: I have not! I have a few simple word filters and htat's it. It works well enough.
hk also i wanna say thanks for notbob.i2p :D
hk ah i see
not_bob hk: No problem. It's been a project of love.
hk yeah man the plushie and the mercedes van... pure SOUL
not_bob The number of comments I normally get is a couple a day. So, automation is not much of an issue.
hk gotcha gotcha
not_bob I have had issues in the past where someone was abusing the comment system. That was another issue.
not_bob But, it didn't last long.
hk yeah it's inevitable
hk i've heard horror stories from fb moderators, people get ptsd
not_bob And, nothing like waking up to death threats ;)
hk the AI hype is exhausting, but it has great utility
hk shit... yeah... it's tough
not_bob Very much so. I use it all the time.
hk but it's almost (with a big asterisk) flattering since you need to be high on the totem pole to get death threats, people like streamers, big developers, people that are recognized etc
hk im not saying it's a good thing, absolutely not... but it's inevitable with the internet if you know what I mean
not_bob *** nods ***
not_bob It's to be expected at some point.
hk not_bob: yeah AI for me personally has been a huge boon
hk yep
not_bob ut, as dr|z3d says, it's a tool. It should be used as needed, and not when not needed.
hk amen
not_bob s/ut/but
not_bob How long have you been using I2P?
hk when it comes to code, you could become alien to your own code and have to "catch up" to what you "wrote", I've had things like that where I have o1-preview spout out code and I have take hours analyzing it
hk hm
hk I lurked briefly 2 years ago under a different pseudonym
hk but only now have I come and began to develop for go-i2p
not_bob I mostly use AI for snippits of code. "How could i write this better?" Or "What else could I do here instead?"
hk I was learning go and saw go-i2p was in the works, I thought why not play my hand to learn go and put stuff out publically
hk yes, that's the best; boilerplate
hk it's good when you hit brick walls as well
not_bob I was aware of that project, but had never tried it.
hk ah it's in the works, we just figured out router info
not_bob And, at times it does produce total garbage.
hk or directs you down a non-working rabbithole XD
not_bob "WHY WOULD YOU DO THAT!?"
hk but the scary part... it can sometimes pull itself out of that rabbit hole
not_bob Exactly.
hk LOL yup
not_bob In the end, it's the future. Being able to use it is so much better than not.
hk yeah, it's here to stay
hk better to adapt than be stubborn and not use it even if you could benefit
not_bob Anyway, I've been working on the underpinnings of my scanner the last few days.
not_bob Just now put the one I've been working on into production.
hk ooo tell us about it
not_bob Before graphs were generated by the scanner. But, I also had a tool to generate graphs manually (or, via a script).
not_bob Now the scanner does not generate graphs. But, when it's done it calls the second program to do the job.
not_bob So, a single graph codebase instead of two.
hk nice, when things are modular it makes things easier
not_bob I also wrote a nifty tool to monitor how many scans are being run in real time easily.
hk you know that kind of shit is useful but also super nice as eye candy
not_bob Which lets me know how long scans take (avg and max).
not_bob There it goes.
hk nicee, you could also program for anomalies
hk spikes and whatnot
not_bob I need to rewrite the code that handles the scans.
hk what language do you use?
hk HOLY HSIT
hk I WAS
hk I swear to fucking god
hk I was going to guess that
hk you cant verify it but I swear to god holy shit
not_bob As of right now I start x number of scans per min blindly.
not_bob It's worked for years.
hk nice nice
not_bob I can scan all 800ish sites on the network in about 10 min.
hk not bad!
not_bob But, best to not do that as it causes load issues.
not_bob So, I spawn it out over an hour.
hk fair enough
not_bob But, still blindly spawning processes.
cumlord oh wow
not_bob The next version will be smarter about that.
hk good good, improvements all the way
cumlord how many times normally you ping a site in an hour?
not_bob cumlord: Once
not_bob That's a lot of traffic.
not_bob And I also ping dead sites too as they may come back to life.
not_bob But, I don't ping them as often. I've got a routine for that where it backs off pings the longer a site has been dead.
cumlord nice 👌 figured it wasn’t super high per hour
not_bob No, don't want to bother anyone.
not_bob During testing I may ping some more often.
not_bob But, I tend to use sites I control for that, again. So I don't bother anyone.
not_bob Extra data does not cause any problems.
not_bob hk: What lead you to think I used perl? :)
cumlord thats a good way to do it. Wanted to make sure my throttling isn’t too aggressive
hk not_bob: honestly, just a hunch; first thought was python -> then settled with perl in microseconds
not_bob I know python too :)
hk haha yeah they're pretty adjacent if I recall correctly
hk I couldnt say what the edge perl provides over python though, I would assume greater performance?
hk but whatever tool works, works
not_bob Somewhat.
not_bob Я также говорю по-русски, плохо.
not_bob If that helps :)
not_bob Honestly? I know perl much better.
not_bob Performance isn't so much of a worry. I've got a pretty beefy server.
hk nice nice
hk let me guess again...
hk THREADRIPPER
not_bob I got and visit it once in a while to dust it out.
not_bob I wish :(
hk aww
hk xeon then :P
not_bob xeon.
hk yeah I wish I could splurge on a threadripper too but I dont have much use out of one besides compiling linux for fun or some shit
hk nice
not_bob Still works well enough.
not_bob Mine XMR on one :)
hk yeah honsetly hardware has gotten so far, we're reaching points where it's not exactly diminishing returns but the software is not that demanding for it; maybe except GPUs now
hk nicee
hk oh I mean i know
hk people in industry use CAD
not_bob Right, unless you need more a basic entry level system is more than enough for anyone.
hk and simulation software, that can demand a threadripper easy
hk or massive datacenters
hk yup yup
not_bob Yes, depending on tasks.
not_bob Even machines 10 years old can still be very useful.
hk im building a home server right now that has a xeon from '15
not_bob Exactly.
hk an hp z240, but right now
hk the fucking.. BRACKETS, I cant fit this fucking $15 gpu to get basic terminal
not_bob Mine is a custom build, but yeah. It's not super recent.
hk i got a low-profile gt 710 thing, fanless and super tiny; so comfy
hk nice yeah it should be great enough
hk I need to get some sort of saw to get the brackets out, but im gonna just bring it into a shop which is embarrassing because I've done computer stuff myself this entire time but this is a physical issue and I cant get the bracket thing off, even with the strange tabs on it; there are no tutorials online on how to do it
hk it's alright, no biggie lol; just excited to get this thing working
not_bob That's odd.
hk I was gonna do an auto setup with debian to get ssh working on boot but it's too much headache
cumlord lol my nas is literally spare parts held together with duct tape and hopes and dreams
hk yeah it is, and I dont want to break it
not_bob I've used a hacksaw to modify a few boards over the years ;)
hk cumlord: lolz, pure soul
not_bob So long as you don't cut into anything that matters it's fine!
hk hacksaw is definitely an unexpected tool with hardware like this
hk oh dude im like
hk you have no idea
hk I dont have OCD, but I am so so so fucking careful with components it's unreal
hk I literally sit my laptop on a soft case because I dont want to "slap" it against the desk and have unnecessary shock
not_bob I had one where I couldn't get it it due to the heatsink. I took a dremmel to the heatsink. Shaved off where it was catching and it was perfect.
not_bob I throw my laptops around.
hk and I wipe my hands all the time to make sure there's no water anywhere, and when I was fiddling with the gpu; I was getting so nervous that it was tackling the gold constantly, im just gonna bring it to a repair shop and have them install it lolz
hk hahahah
hk not_bob: re: heatsink, BRUTAL
hk guess it worked out
not_bob It had to be done!
not_bob That's for my dash visor computer.
cumlord lmao going at it with the hacksaw ;) that's brave
not_bob Raspberry pi with passive cooling. Strapped to the visor so I can see things when I'm driving :)
hk hahahah super nice
not_bob You know, watch IRC or logs.
cumlord that's awesome
not_bob I almost wreked my van last week because of it though...
not_bob Some asshole in front of my suddely stopped on the innerstate.
hk oh shit D:
cumlord been wanting to make like a portable pi setup thing
cumlord settled for my "i2pphone"
hk not_bob: dude you have a dash cam right?
hk just in case
not_bob I do.
hk nice
hk cumlord: yeah it's doable with a battery
not_bob I record everything.
hk good shit good shit
not_bob Alas, I need to get to doing other things. Good chatting with both of you.
hk same here yeah gotta write unit tests
hk good talk not_bob, and keep at it with the great work man seriously
not_bob Thank you.
cumlord for real stay safe out there
orignal zzz, check your e-mail