IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#i2p-dev
/2023/02/03
eyedeekay Opicaak yes
Opicaak Alright, I wanted your opinion on the guys's browser issue, it's resolved now.
Opicaak Sorry for bothering you.
eyedeekay NBD, I was already reading the thread. Holy cow, I refresh the page and it goes from 23 to 58 comments
eyedeekay And all top-level comments
Opicaak Haha, yeah, not counting PMs.
eyedeekay I am deeply sorry to hear that lol
eyedeekay reddit has been a nightmare lately
Opicaak Oh yeah, the amount of PMs I'm getting daily, it's insane.
eyedeekay They PM me like mad too
eyedeekay I just give them a canned "I won't answer any questions in private that I wouldn't also answer in public"
eyedeekay Sometimes they shut up
Opicaak But that will just move it to another place and I will end up being the one answering it anyways.
eyedeekay Yeah but then you can copypasta the old thread when the next one comes along
Opicaak Right, and it could act as a knowledge base, too.
eyedeekay Until the recent influx it was a pretty good strategy. These new people don't bother to use the search
Opicaak Not only that, but most people don't even read. Often, the answer they are looking for is right infront of them.
dr|z3d eyedeekay: I think I know how to short circuit a bunch of the more inane Reddit support queries.
dr|z3d or at the very least, significantly reduce the burden of supporting them.
Opicaak By adding a bot?
dr|z3d what reddit, and the project needs, is a support wiki that answers the most obvious questions in a checklist type format. maybe a bot could be programmed to recognize questions relating to wiki topics, but I don't know how reddit bots work.
dr|z3d wiki would be hosted on .i2p and clearnet.
Opicaak There can be a wiki page directly on the subreddit.
Opicaak But there is still the same issue, people don't read.
dr|z3d and the content/db should be made available so anyone wanting to mirror it can.
Opicaak They would rather create a new post and expect someone else to answer/look it up for them.
dr|z3d no. you're wrong, sorry.
dr|z3d have you looked at the reddit wiki?
Opicaak Not really, no.
Opicaak I just saw some tabs called wiki.
dr|z3d take a look and then tell me we have a functional, useful, vibrant wiki on reddit. and I'll laugh at you.
dr|z3d so imagine you're a new user..
dr|z3d "oh, they have a wiki, maybe my question's answered there, let me take a look..."
Opicaak Seems like a simple markdown
dr|z3d it's crap. new users won't be getting their support issues answered there any time soon.
dr|z3d either a community-run wiki or a readthedocs site, or both. with the content/db available for mirroring.
dr|z3d the emphasis here is on community-run. no one person should be providing all the content.
Opicaak Or, you know, ML/AI bot could solve all of this.
dr|z3d could it? I don't know enough about what's possible with reddit, but given some training and reinforcement, maybe it could.
dr|z3d maybe a bot could create the wiki content, and then a human could refine it. also possible.
dr|z3d really the point I'm making, the sub-text, is that it's untenable for eyedeekay to keep answering questions because he feels obligated. he has much better things to do with his time.
Opicaak Same over here.
Opicaak Ok, maybe the automod could be somewhat useful, it can match specific words in title, for example, and then reply.
eyedeekay This used to be the thing that served as the reddit wiki: github.com/J-tt/r-i2p-wiki I think the goal was for it to be forkable and clonable
eyedeekay That was many years ago
dr|z3d yeah, inadequate and abandoned.
dr|z3d media wiki and/or read the docs probably best.
eyedeekay there's wiki.i2p-projekt.i2p but I don't think there's a clearnet end is there?
dr|z3d wiki.i2p.net / wiki.geti2p.net
eyedeekay Not there for me
dr|z3d there's probably around 1/2 dozen questions that cover a fairly big chunk of reddit support questions.
eyedeekay That's about right. A little less than half of the questions fall into less than 6 categories, used to be the biggest chunk was browser config but that has been shrinking and I have been learning to be careful what I wish for
eyedeekay Now the biggest category is "doesn't refresh"
dr|z3d so, readthedocs, perhaps, as the first port of call to cover those oft-repeated questions. and then maybe think about a wiki.
dr|z3d really, reddit should be all about philosophical questions, opsec, technical details. the support stuff should be offloaded, but I think we both know that :)
eyedeekay I'm open to it continuing to be support but support like "how do I expose the SAM API to my Docker bridge network" or "How do I play Freeciv over I2P" or something even remotely interesting
eyedeekay Maybe, I've not used readthedocs extensively but I'm concerned about the workflow for contributions?
dr|z3d looks like it interfaces with github, so not difficult for contributors to send merge requests. readthedocs.org
dr|z3d and that would also cover the redistribution/reuse angle.
eyedeekay Yeah I like stuff to be clonable
eyedeekay everybody gets to be a backup
dr|z3d *nods and smiles*
dr|z3d reverse lookups for router, off by default in I2P+, enabled with a config. figured it would be potentially useful when looking for abusive routers etc, and just generally.
eyedeekay Interesting, am I reading this correctly that the domain corresponds to the router and that some of them are showing github.com?
dr|z3d correct.
dr|z3d though why they're not showing ip addresses but are showing a reverse lookup, I don't know.
dr|z3d usually it's the other way around.
eyedeekay Does that mean they're running on something github-owned then? like CI abuse?
dr|z3d cdn-185-199-111-153.github.com
dr|z3d your guess is as good as mine.
eyedeekay My mind immediately leaps to running commands to build I2P during the build phase of some CI, starting the router, then blocking and waiting for the CI to time out as a cheap way to spin up routers on OPIP
eyedeekay or it could be non-malicious, somebody could be spinning up routers in CI to test a SAM app
eyedeekay Still a bad idea though
eyedeekay I think somebody's got a mock SAM bridge somewhere they should be using instead
dr|z3d yup, that thought had occurred to me, too. if there's no timeout on the CI, you've just won some free router hosting :)
eyedeekay There's a timeout :)
dr|z3d ah, sensible.
eyedeekay I've hit it a couple times
eyedeekay Longest you can stretch it out is a little over an hour and 12 minutes
eyedeekay Probably a useful thing to know if you're watching them
dr|z3d yeah, I vaguely recall cryptomining abuse on github a while back which is possibly when they hardened things.
zzz would somebody with at least several thousand part. tunnels and a few days uptime please report your tunnel.participatingMessageCountAvgPerTunnel lifetime average value stat?
postman will do zzz
eyedeekay I've got between 1000-4000 tunnels on this router usually, 19 days uptime, 0.9.57
eyedeekay Lifetime average value: 126.729 (34,280 events)
dr|z3d 357.916 (8,844 events)
zzz thanks guys, just double checking some stuff while I work on congestion caps
wodencafe What in tarnation is this BiglyBT thing?
wodencafe I don't understand why BiglyBT or whatever is a dependency for X-Seeding. I use CLI commands with `transmission-cli` to strip other trackers and add the postman tracker.
wodencafe I wrote it up as a script.
wodencafe Also, what is "http://w7tpbzncbcocrqtwwm3nezhnnsw4ozadvi2hmvzdhrqzfxfum7wa.b32.i2p/a" ?
Opicaak It says it right there in the document, it's a bittorrent client that supports both clearnet and i2p.
Opicaak Apparently, the only one in existence.
wodencafe So here's a use case I've been facing. I want to share torrents and preserve the existing torrent hash, to allow others to X-Seed over clearnet, but I don't want to install BiglyBT and I don't personally want to X-Seed.
wodencafe I'm happy to seed over I2P, and I'm happy for others to X-Seed over clearnet and I2P, but I personally don't want to seed to clearnet. But I upload a lot of torrents.
wodencafe Right now, I don't involve BiglyBT. I use `transmission-cli` to remove the trackers from the torrent and add postman's tracker.
Opicaak So, you want to only seed over i2p?
wodencafe I want to only seed over I2P, but preserve the torrent hash so that others can seed over clearnet.
Opicaak Shouldn't XD/XD-Cli work just fine?
Opicaak Well, the hash is based on the files inside
wodencafe What is that?
wodencafe Yeah but if I recreate a torrent on I2P, even using the same files from clearnet, I get a different hash - unless I take the original torrent file, strip the trackers, and add an I2P tracker.
wodencafe Doesn't matter if it's the exact same files, if I recreate it, it's a different hash.
Opicaak XD is a torrent client for i2p.
Opicaak Written in Go
wodencafe I guess what I'm getting at is, I want to share torrents from clearnet to I2P with minimal dependencies. I don't want to depend on some other torrent I2P client when I have I2Psnark. I don't want to depend on BiglyBT.
Opicaak Hm, I'm not sure actually, sorry
wodencafe No, I know the answer. I'm following a process to make this work. My concern is that others in the future will encounter this same obstacle, and have similar objections to mine.
wodencafe I solved the use case, but I had to figure it out on my own, and it took a lot of time to figure out.
wodencafe So I guess what I'm saying is, where is the document for my use case? Does it not exist? Should I write it?
Opicaak You could document it for others then.
Opicaak Probably doesn't exist, yeah.
wodencafe Ok. I'll write something up. :)
RN :D
zzz the hash will only stay the same if the piece size remains the same wodencafe
wodencafe zzz: As in, the torrent files themselves don't change? I'm not disputing that.
wodencafe For my use case, I have an existing torrent from clearnet and I want to put it on I2P, without using BiglyBT or some other dependencies, but I want to preserve the torrent hash so that others can cross stream it between clearnet and I2P.
wodencafe I've found a way to do that, the point I'm raising is that there is no documentation for that use case, and I had to figure it out myself.
zzz right.
zzz the reason the hash doesn't stay the same when you re-create it in snark is it picks a different piece size
zzz yeah I get it, was just explaining why the info hash isn't always preserved unless you edit the torrent file
zzz if you already knew that, great :)
RN
RN Lives!