IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#saltr
/2026/03/19
@RN_
@T3s|4_
@eyedeekay
@orignal
@postman
@zzz
%Liorar
%acetone
%mareki2p
%snex
+FreefallHeavens
+Onn4l7h
+Onn4|7h
+altec_lansing
+leopold
+nyaa2pguy
+poriori
+profetikla
+qend-irc2p
+r00tobo
+uop23ip
Arch
Danny
Holmes
Irc2PGuest28384
Irc2PGuest63277
Irc2PGuest71791
Irc2PGuest84270
NiceBoat_
OfficialCIA_
Over1
SilentWave
U1F642
Wikk_0
ahiru
anontor
cims
dr|z3d
eyedeekay_
floatyfloatilla
mahlay
makoto
mesh
n2
nilbog
not_bob_afk2
pory
r00tobo[2]
red
sektorchef
thetia
user_
utp
vivid_reader56
dr|z3d ok, python script now available for (most) I2P+ sessionbans -> iptables (run as root via crontab) -> github.com/I2PPlus/i2pplus/commit/a912a75b0fb16c19613a97fdc1d716e429e3e0ba
dr|z3d couple of headsups, zzz...
dr|z3d firstly, init.d is dead in trixie. and likely to soon be in ubuntu.
dr|z3d so that means rejigging i2prouter install.
dr|z3d (and probably the debian installer, too)
zzz hmph, any pointers?
dr|z3d also, wrapper update.. I've got the win64 build happening on Linux, you might want to take a look: github.com/I2PPlus/i2pplus/tree/master/installer/lib/wrapper
zzz nice
dr|z3d there's some instructions for win64 wrapper in the README.md and a build script.. you may need to manually download the source and point the build script at it.
zzz git pull is easier ))
zzz I see you got rid of all the crap finally like javadoc files and spotbugs and god knows what
dr|z3d indeed it is. just letting you know where the files are @ :)
zzz thx
dr|z3d I think I may have caught the tail end of a conversation you were having with orignal about the state of my repo and the accumulated cruft..
dr|z3d yes indeed. and I took the opportunity to tidy a few things up, clean up root, installer/resources, convert a few docs to MD so they display nicer via git web ui.
zzz it was way out of hand but you do you
dr|z3d out of hand.
dr|z3d yes mom.
dr|z3d anyways, there are some compiled win64 binaries there which you've probably noticed by now, together with the latest stable wrapper files for everything supported.
dr|z3d it should all be synced on 3.6.4
zzz I also noted you yeeted 6 weeks / 500 commits / 50K lines of diff and rebased back to Jan. 30, not noticed here, mom would not approve, but I had mentioned here it seems like you had painted yourself into a corner
dr|z3d that was a rabbit hole I'd rather not have gone down, but hey.
dr|z3d you may or may not want to pickup the other i2prouter tweaks, mostly a spinner for starting and some info for status.
dr|z3d I2P+ Service is running
dr|z3d • PID: 2752583
dr|z3d • Build: 49bbd968 / 2.11.0-3+
dr|z3d • Wrapper: STARTED (3.6.4)
dr|z3d • Java: STARTED
dr|z3d • Started: Thu Mar 19 foo 2026
dr|z3d • Uptime: 999m
zzz bigger than a rabbit
zzz init.d is going to be gone in resolute 26.04? or later?
onon_ Hi there
onon_ When will you make java-router start working quickly?
dr|z3d well, given resolute is likely synced with trixie, yeah, probably, but I can't say definitively, you know how Ubuntu likes to diverge from time to time.
onon_ People still complain and turn away from the project when they try java-router
dr|z3d any idea what he's on about, zzz?
dr|z3d resolute or not, no harm in getting in early, shouldn't break anything with older ubuntus.
zzz not so much diverge as delayed verge
zzz re: start, no
zzz have any details onon_ ?
zzz drz what happens to our installed package on upgrade to trixie? does init.d get migrated somehow and big fat warnings, or crapout, or?
onon_ details about what?
onon_ How to make java-router fast and stable?
dr|z3d > also maybe of interest, zzz, there's now native support for zen3 for jbigi and updates to GMP 6.3.0
dr|z3d > not sure that buys us much, but it's something :)
onon_ I have already written here many times how to do this.
dr|z3d and, minor thing, gradlew is current.
dr|z3d are you running it, onon_?
zzz drz what happens to our installed package on upgrade to trixie? does init.d get migrated somehow and big fat warnings, or crapout, or?
dr|z3d either canon i2p or i2p+ ?
dr|z3d crapout, zzz
dr|z3d service will just fail to start.
dr|z3d no migration
onon_ Yes, of course I tried both canon and i2p+
dr|z3d no, but are you *running* it?
onon_ not now
dr|z3d trying either for 5m won't tell you much.
onon_ i2p+ really works a little faster both for sending and receiving, but it still loses connection, just like canon
zzz onon_, by 'details' I mean by what metrics do you mean 'slow' or 'unstable'. Time from what to what? Or CPU time? Bugs?
onon_ transfer speed
onon_ sending and recieving
zzz ok, that's not 'java-router start working quickly', that sounded like some startup time complaint
zzz so apparently your complaint is about streaming?
onon_ You see it, dr|z3d disconnecting all time
onon_ I never disconnecting from ilita
zzz So, your question is, when will the java router meet your standards for speed and stability. There's no schedule for that, and don't think those standards are defined enough to answer the question anyway
onon_ It's sad
onon_ Это печально
zzz yeah I'm over here crying
onon_ It is very sad to realize that the main developer of i2p is slowing down the development of the network. Precisely at a time when all people so need an alternative to clearnet
zzz doing all I can over here, sorry to disppoint you
zzz dr|z3d, does your init.d fix work on bullseye/jammy? Since the jetty12 package is broken in trixie/resolute we're using the jammy build for everything currently
dr|z3d onon_: be nice, you can phrase things a little differently and get much better reception.
dr|z3d probably, zzz. should work with anything that has systemd.
zzz hmm does canon have any non-systemd downstreams I wonder
zzz so you didn't leave the init.d support in at all? is it possible to support both in one package?
dr|z3d I think both should be supported.
T3s|4 onon_: back way off with your unsupported and grossly unfair allegations, especially if you care if folks ever want to take you seriously
zzz drz 'should' as in you think you do, or you think it's possible?
onon_ "unsupported and grossly unfair allegations"
dr|z3d fallback to init.d in the absence of systemd, zzz.
zzz right, q is, is that what you did or you're speculating it's possible?
dr|z3d that's what I gone done.
dr|z3d 245: if [ "$HAS_SYSTEMD" = "1" ]; then
dr|z3d 246: install_systemd
dr|z3d 248: install_debian_initd
dr|z3d 252: if [ "$HAS_SYSTEMD" = "1" ]; then
onon_ T3s|4 , i2pd can send and receive a stream at a speed of 1 MB/sec. java-router can't do that.
onon_ This is a verifiable fact.
zzz ok, thx. I put it on the list
zzz you said it's dead in trixie but I haven't seen any complaints?
zzz so the first link says it's gone in systemd 260. The 2nd link says it's gone in trixie, but trixie has systemd 257
zzz so I;m confused, is it dead in trixie or not?
dr|z3d $systemctl --version
dr|z3d systemd 260 (260~rc4-1)
dr|z3d maybe I'm ahead of the curve, when I said trixie, I meant forky.
dr|z3d $grep VERSION_CODENAME /etc/os-release
dr|z3d VERSION_CODENAME=forky
zzz that would explain my confusion ))
dr|z3d yeah, sorry, I was about to berate you for disbelief :)
zzz that pushes out the issue two years
zzz and explains why no complaints
zzz so the 2nd article is just wrong
zzz dr|z3d, turning the tables on the you-might-want-this discussion, you haven't cherrypicked or manually merged much if anything from canon since last summer? just got too hard, or new policy to not do it?
dr|z3d last summer? nah, much more recently than that.
dr|z3d mostly it's manually merged these days, so it won't show up as you in the commit logs.
dr|z3d I do try and credit upstream when I manually merge.
dr|z3d the pulling / adding your history.txt entries I may be a bit slack on.
zzz ok, just wondering whether to cancel zzz's heads-up service
dr|z3d which one's that? :)
zzz heads up you should grab this
dr|z3d no, don't do that. :)
zzz current example:
dr|z3d would *never* be my policy to not pull from upstream.
dr|z3d I just got caught in the weeds the last couple of months, so I'm a bit behind the times until I catch up.
zzz heads up I just pushed a real doozie fixing a 2009 stats problem
dr|z3d can you summarize the issue you fixorated?
zzz Author: zzz <zzz@i2pmail.org>
zzz Date: Thu Mar 19 10:25:12 2026
zzz Stats: Fix 60s stat getAverage() returning 20% high
zzz due to CoalesceStatsEvent being on a 50s timer
zzz - Caused by rounding in Rate.coalesce()
zzz - Affected data on /stats, Prometheus plugin, console sidebar, and internal use
zzz - Did not affect graphs
zzz - Apparently dates back to original jrandom code
zzz basically, COALESCE_TIME was not an integer divisor of the stat time, so rounding happened
dr|z3d ok, so if I understand correctly, 50s was essentially rounding up to 60s in the stats and inflating the numbers?
dr|z3d or maybe the opposite.
dr|z3d anyways, sounds like a good catch if it's been lurking undiscovered for so long.
zzz there was a 60/50 division cast to an int, so 1.2 -> 1
zzz to confirm if you have the bug, compare the current value on any 60s graph to the sidebar value or average on /stats
zzz there was a 60/50 division cast to an int, so 1.2 -> 1
zzz to confirm if you have the bug, compare the current value on any 60s graph to the sidebar value or average on /stats
dr|z3d looks fine here on casual inspection. b/w stats and the graph stats correspond.
dr|z3d but i'll snag that commit anyways, and all the outstanding commits. aside from jetty12, I'm up to date with 2.11 I think.
dr|z3d (thx)
zzz look at router.maxMemory, router.knownPeers
zzz *router.memoryUsed
dr|z3d graph says 6.6K, sidebar says 6675K
zzz hmph. how about on /stats ?
dr|z3d (known)
dr|z3d I don't think I'm doing any fancy with the graph numbers, just truncating to 1 decimal place. so they look about right.
zzz the graphs were right, it's anything that called getAverage() that was high
zzz actually a lot of those sidebar values are current values, not from the stat, so they wouldn't be affected
dr|z3d known in stats looks a bit different than sidebar known.
dr|z3d 7.7K vs 6.6
zzz sounds like 20% to me
dr|z3d Blinded message
zzz touche
zzz think I'm going to have to rip out izpack4, it's probably not fixable for java 17
zzz for the sidebar, it would only affect ones that are from a 60s rate and a getAverageValue() or getAvgOrLifetimeAvg() call on that rate
dr|z3d I vaguely remember seeing something about it not working with later javas, still the case?
dr|z3d doesn't stop you from explicitly building with java8, though, does it?
dr|z3d compile with java17 or whatever, izpack4 with java8.
zzz didn't think of that, will try it
zzz if you're caught up to 2.11, I'll offer two more headsups:
zzz 1) the ISJ/DLM fixes
zzz 2) Banlist max size enforcement to prevent blowups
zzz the boys have also been busy on docker recently
zzz if you're caught up to 2.11, I'll offer two more headsups:
zzz 1) the ISJ/DLM fixes
zzz 2) Banlist max size enforcement to prevent blowups
zzz the boys have also been busy on docker recently
dr|z3d aight, thanks, will get on those soonest.
zzz 1) affects mostly lowpowered routers which probably isn't your userbase
zzz 2) ditto, maybe
Gully hi :-)
Gully i2p is still under attack?
zzz hi Gully the worst is over, what remains is manageable
Gully ok good
dr|z3d new attack?
dr|z3d even with all the blocks applied in i2ptables, I see 13K known peers.
dr|z3d mostly L tier.
StormyCloud I see a total of 27k unique routers
dr|z3d yeah, that's not normal.