IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#i2p-dev
/2025/01/31
@eyedeekay
&eche|on
&kytv
&zzz
+R4SAS
+RN
+StormyCloud
+T3s|4
+acetone
+dr|z3d
+hagen
+hk
+lbt
+postman
+segfault
+wodencafe
Arch
Danny
DeltaOreo
FreefallHeavens_
Irc2PGuest10722
Irc2PGuest16699
Irc2PGuest27222
Irc2PGuest316
Irc2PGuest39482
Irc2PGuest59134
Irc2PGuest95432
Nausicaa
Onn4l7h
Onn4|7h
Sisyphus
Sleepy
Soni
T3s|4_
Teeed
aeiou
aisle1
ardu
b3t4f4c3__
bak83_
boonst
carried6590
cumlord
dickless
dr4wd3_
eyedeekay_bnc
orignal_
poriori
profetikla
qend-irc2p
radakayot_
rapidash
shiver_
solidx66
thetia
u5657
uop23ip_
w8rabbit
x74a6
orignal eyedeekay can we make a fork from your reseed tool?
orignal we need to ajust it for ygg
eyedeekay Go ahead, fine with me
eyedeekay I would accept the changes upstream if you want to send me a patch
orignal no, you don't need it in the upstrem it's for ygg reseeds only
orignal unless Java-I2P starts supprting it
orignal lol. it's old neon's reseed basically
orignal I remeber it was an issue with RSA
eyedeekay I forked an old one from "MDrollette" on github, merged some fixes that were scattered around other repositories, added a boatload of features and set up packaging for it
orignal can someone teel what was used for reseed before neon's implemntation?
zzz dont remember neon, but in the old days, before zip, before su3, it would be just a directory with some RI files, and we'd just fetch the directory listing and then fetch each file
zzz so maybe just a cron job that copied some files over
orignal the one eyedeekay hav forked his reseed from
zzz mattdrollete iitc
orignal his nick was neon
zzz oh, same guy, ok
orignal but SU3 reseed worked when whe was worning on this project
orignal hence it was something beofre
orignal I'm trying to understand where python reseed comes from
zzz are you giving up on python and moving to idk's reseeder?
orignal I'm trying ot understand the scope
orignal see, I run python one because I was told to do it
orignal but go reseed is also douable since we run ygg anyway
dr|z3d I have a fork that does a lot less than eyedeekay's version if you want minimal.
dr|z3d you'd need fail2ban to rate limit, eyedeekay's has that built in, for example.
dr|z3d eyedeekay's will order you pizza if you supply the correct sequence of characters.
zzz orignal, I can't find much on my hard drive for old reseed tools, I really don't know, I didn't manage the reseeds back then, some other guy was doing it
orignal dr|z3d link please
dr|z3d zzz: no.
dr|z3d no need.
dr|z3d at least, no need when running under +
zzz well, orignal might need to know that, and my email this week explains why. I guess if nobody's using it there's no need
orignal we need a tool that: 1. reads netdb folder 2. filters 3. produces random sample 4. make zip file 5. signs it
dr|z3d it does most of that, though as zzz mentioned, it might need tweaking to filter out caps/versions.
zzz sounds right orignal. So your choices are 1) add RI parsing to pyseeder; or 2) switch to idk's reseed-tools
orignal for 2 you need to parse RI file
orignal no my question if if idk's reseed-tool actually parses RI file
zzz yes it does
zzz he pulled in code from his go-router project into reseed-tools
orignal python doesn't it. it just uses regexps
zzz probably wouldn't be that hard for a python guy but might be really annoying if you're not ))
orignal I'm partially a pythin guy
orignal not bad idea to write a simple parsers just for properties and addresses without any crypto
orignal will try
orignal not bad idea
zzz dr|z3d, what flavor of reseeder is coconut using? because:
zzz Test passed for coconut.incognet.io, returned 100 router infos
zzz Router infos included 11 with versions before 0.9.59 and 27 unreachable
dr|z3d shouldn't be happening, those shouldn't be on disk for it to select, will have a look and see wtf.
zzz ok, so it's i2pplus + your fork then, so yeah.
zzz to test: cd $I2P; java -jar lib/router.jar reseeder coconut.incognet.io
dr|z3d ok, thanks, will deal with it later.
eyedeekay orignal the pertinent part of reseed-tools for sorting good and bad routerinfos starts on service.go line 244 github.com/eyedeekay/reseed-tools/blob/be4257c49aa49c2b651c2e0f49f604c3baf9eeb7/reseed/service.go#L244 but the RI parsing logic is all in go-i2p, you'll need to at least read the caps in python to do it
zzz last call for translations, pulling in 10 minutes
zzz eyedeekay, everything look good on your side, checklist-wise, to wrap this thing up?
zzz I updated the tor blocklist
zzz hoping for geoip this weekend as you noted
eyedeekay Yes I put a few things off to the last minute to wait for the last few changes to hit i2p.i2p but I'm ready to go now
eyedeekay On schedule
zzz ok then I'll pull the txes and bump shortly
zzz you have a newsxml entry for me to push to tx?
dr|z3d don't forget to proofread before you push :)
zzz ok you're now on the proofreading team
dr|z3d sure, show me the product.
dr|z3d don't forget harry, he's a good proofreader, too :)
eyedeekay I'll get that newsxml entry up in a few minutes
zzz I already updated the roadmap if you need your memory jogged or want a better summary than history.txt
zzz gah the po4a-translate on the current ubuntu is broken or doesn't like our files anymore
zzz switching to a LTS box to rebuild the RU man pages
zzz maybe somebody can figure it out for next time
dr|z3d I noticed the translation count wasn't working anymore, dunno if that's related.
zzz shouldnt
dr|z3d maybe I've accidentally something.
zzz just for the man pages?
zzz we don't do counts for man pages, only yes/no
dr|z3d tells me everything's 100%.. it's close, but there are a few untranslated strings floating around.
dr|z3d I mean everything.
zzz thats on you, unless it's just rounding up to 100
dr|z3d yeah, on me. like a bad toupe.
zzz eyedeekay, debcheckpatch is failing due to a wrapper.config change, I'm working on it now
zzz ant translationReport for the CLI version
eyedeekay Thanks zzz, not sure why I didn't see it when I ran practice builds the other day...
eyedeekay news is pushed
zzz I put the po4a issue up as ticket #498
zzz ok dr|z3d let's review
dr|z3d what are we reviewing?
eyedeekay i2p.newsxml/data/entries.html right at the top
eyedeekay Not much at all TBH, short summary of what happened this cycle
dr|z3d I think you missed a full stop.
dr|z3d or rather, a forward slash.
dr|z3d or something.
dr|z3d got a valid url for me, eyedeekay? :)
dr|z3d ah, that's shorthand to a repo, perhaps...
eyedeekay Yes it is just a sec I'll get a full URL
dr|z3d ok, first thing, we don't need 2.8.0 referenced in the body text.
dr|z3d it's the title of the news item.
zzz debcheckpatch might only kick in for the release builds, you can look to see if it's in the checklist or not
dr|z3d needless hyphen for un-congested.
zzz change published and updated times to monday
dr|z3d Older leaseSets are now removed from the network database mitigate spam and improve performance.
zzz agreed not just needless but wrong
dr|z3d you're missing a "to"
dr|z3d I was being polite, zzz :)
zzz it's another train-your-dog thing
dr|z3d *** chuckles. ***
dr|z3d right, aside from the obvious grammatical errors and omissions, I think we could make it flow a little better.. let's give it to harry to look at.
zzz 'were made' in 3 sentences out of 4, might try to change it up, or use active voice
eyedeekay revisions posted
dr|z3d ok, we can do better here: Further adjustments were made to how Java I2P routers handle severely congested peers to give them time to become uncongested.
dr|z3d it's a bit clunky.
dr|z3d we're all about Java I2P in the news, no sense in referencing Java here.
zzz We improved handling of congested routers in the network.
dr|z3d yeah, something like that. less is more, when it conveys the same info.
eyedeekay Next revision pushed
dr|z3d ok, next line: Older leaseSets are now removed from the network database to mitigate spam and improve performance.
dr|z3d That's only relevant to floodfills.
dr|z3d Worth mentioning.
zzz not sure
dr|z3d hmm? ok.
zzz I guess you're right, but... we always removed older leasesets
dr|z3d We now have a more aggressive strategy for leaseset removal from the NetDb to improve router performance.
dr|z3d do we need to mention spam? don't think so, it's a word we're familar with here, but it might confuse those who aren't.
zzz now we'll remove them faster when the network is under attack
zzz spam isn't really translatable, better to just refer to 'attacks' or 'congestion'
dr|z3d agree.
dr|z3d And the last line: Other changes were implemented to reduce the observability of events like a router rebooting or shutting down.
zzz or 'overload'
dr|z3d sure, anything like that works. less words, same meaning.
dr|z3d Implementation has started on obfuscating router startup and shutdown timings, specifically to make correlation with locally running services harder.
dr|z3d Or We've started the first phase of..
dr|z3d something to indicate it's WIP.
dr|z3d We anticipate further hardening in the next release.
dr|z3d Something to close the news down with, aside from the last para which is boiler plate.
dr|z3d might be worth referencing anything else there that's scheduled for the next release that sounds suitably interesting. dunno.
zzz we haven't done this in a while, so let's be extra-careful thru the review and release process
zzz let's not make any promises for the next release dr|z3d
dr|z3d not so much promises, more non-specific hints of what's being planned. but up to you.
zzz haven't yet heard an idea that sounds doable
dr|z3d re correlation?
zzz yeah
dr|z3d did eyedeekay mention our chat?
zzz yes and nothing jumped out at me but promised to think about it more which I havent
dr|z3d ok, well, in brief, caching of leasesets, both router and services.
dr|z3d optional deferred publication of service leasesets on startup with some randomness thrown in.
dr|z3d and a longer lifespan for router leasesets.
dr|z3d that's broadly what I suggested to eyedeekay.
eyedeekay I actually gave him a much more thorough breakdown than you just gave and re: routerinfos lifetimes when zzz and I discussed it I found out our assumption was wrong
dr|z3d if you have a cached leases that haven't expired, why create new ones on startup is my thinking. the additional benefit is that services will resume without requiring a new LS after restart.
dr|z3d I was summarizing just to get the ideas out there, eyedeekay, not because I thought you might have omitted things :)
dr|z3d *service will resume for clients..
zzz I got about 7 ideas listed from idk but he didn't list any of those 3 above
zzz but the more the merrier
dr|z3d *** wags his finger at eyedeekay ***
zzz just saying none seem to be obviously good and easy that we should even half-promise, even if you take for granted that it's worth doing at all
dr|z3d also, if a router has been down for longer than the lifespan of the currently cached LS, then we add some extra delay to local service LS publication, just to further decouple router and services.
eyedeekay I most certainly did, but longer lifespan for routerinfos(I didn't use the term "router leasesets") turned out to be irrelevant and I referred to the cached LS strategy in terms of a "schedule" that needed to be maintained across restarts
dr|z3d not trying to start a fight. :)
zzz maybe thats the confusion, please never say 'router leasesets'
dr|z3d yeah, shit phrasing. RIs it is.
zzz we have routerinfos and leasesets
dr|z3d dunno how that crept in. tired.
eyedeekay Apologies if I was unclear I tried to sort and categorize the things we talked when I presented them to zzz, dr|zed
zzz eyedeekay, did you have one more newsxml rev coming or are you done?
dr|z3d no apology necessary, eyedeekay, you're good. :)
eyedeekay Let me check if I pushed it already...
zzz he did mention ri expiration
zzz just not ls expiration ))
eyedeekay One last rev pushed
zzz ok what's the deadline for the translators?
eyedeekay Monday 5UTC work?
zzz sure
dr|z3d Other changes were implemented to reduce the observability of events like a router rebooting or shutting down.
dr|z3d the word "like" seems a bit too slapdash.
zzz I just pushed it to tx
dr|z3d oh well, there we are then.
dr|z3d some improvement there, anyways.
zzz but agree in principle
zzz thanks for the review help
dr|z3d no worries.
zzz announced on tx
zzz 10320 lines of diff, but a lot of it is rrd4j and the new jammy changelog, probably won't be that bad to review
RN does sound like a lot. Y'all been busy! :D
RN nice news release, bummed I missed the proofreading party but it looks good.
RN does this release have a name yet?
zzz deepseek-yrmom
zzz but maybe you have something better?
RN LOL
dr|z3d periscopes up, as they say. *cough*
RN AI-Free-Edition
RN ;)
RN or Personal_Use_Only
dr|z3d your mom?
RN lemme go back to the changelog for some inspiration
RN Winter is Coming (not inspired by history/changelog)
RN or maybe "quidgames-2"
RN heh
RN *squidgames-2
RN or at the risk of stirring the pot....
RN Trump-Card
RN or Full_Pardon
RN is Mu Deng still a 'thing?'
dr|z3d always.
RN "TicTokTroubles"
RN not a lot of ideas yet... but I'll continue to think on it. :)
RN or maybe "Banned Application"
RN perhaps "original Sin" or "Severence2"
dr|z3d I almost had the same though at the same time.
dr|z3d "orignal's new shoes".
RN hahahahahaha