#saltr (irc2p) | I2P+ 2.10.0-21+ >> i2pplus.github.io | skank.i2p | git.skank.i2p | purokishi.i2p | vituperative.github.io/i2pchat | notbob.i2p | ramble.i2p | drop.i2p | shinobi.i2p | wall.i2p | /msg an op for voice

not_bob 0 hop does not work with NAT.

not_bob 1 hop resolves firewall traversal while still giving you much of the speed.

onon_ dr|z3d as for streaming configs, try the i2cp options input.

onon_ Can it only apply the parameters described here? i2p-projekt.i2p/ru/docs/protocol/i2cp

onon_ Or is this list of available options longer? And where can you watch it in this case?

dr|z3d here's your first git lesson: cd ~/i2pplus && git grep i2cp *.java | grep snark

dr|z3d as for optional tweaks, anything related to i2cp should be listed there, but maybe there's more, not sure. there's also some snark specific configs, configured in snark.config

dr|z3d git grep PROP_ *.java |grep snark

onon_ git grep PROP_ *.java | grep snark | grep ack

onon_ nothing about AckDelay =(

dr|z3d git grep ack *.java | grep snark

dr|z3d PROP_ is for properties, usually configuration options.

onon_ git grep AckDelay *.java

onon_ options.getSendAckDelay() from now (1000 ms)

onon_ 1000 ms ???

dr|z3d when you get a return on your grep, open the referenced file to review the code block and context.

dr|z3d /**

dr|z3d * If the next send time is currently >= 0 (i.e. not "never"),

dr|z3d * this may make the next time sooner but will not make it later.

dr|z3d * If the next send time is currently < 0 (i.e. "never"),

dr|z3d * this will set it to the time specified, but not later than

dr|z3d * options.getSendAckDelay() from now (1000 ms)

dr|z3d */

dr|z3d public void setNextSendTime(long when) {

dr|z3d synchronized(_nextSendLock) {

dr|z3d if (_nextSendTime >= 0) {

dr|z3d if (when < _nextSendTime) {_nextSendTime = when;}

dr|z3d } else {_nextSendTime = when;}

dr|z3d if (_nextSendTime >= 0) {

dr|z3d long max = _context.clock().now() + _options.getSendAckDelay();

dr|z3d if (max < _nextSendTime) {_nextSendTime = max;}

dr|z3d }

onon_ private static final int DEFAULT_INITIAL_ACK_DELAY = 500;

onon_ Too many

onon_ must be 100

onon_ for 3in 3out hops

onon_ Or you're better off requesting the ACK explicitly on the sender's side, in this case.

dr|z3d you've obviously read the RFC that's indicated in the comment above that value.

dr|z3d (not)

dr|z3d A TCP SHOULD implement a delayed ACK, but an ACK should not

dr|z3d be excessively delayed; in particular, the delay MUST be

dr|z3d less than 0.5 seconds, and in a stream of full-sized

dr|z3d segments there SHOULD be an ACK for at least every second

dr|z3d segment.

cumlord we doin lots of pings now looks like

onon_ Set it to 1/10 RTT

dr|z3d sadly the ticket reference will be dead.

dr|z3d well observed, cumlord :)

dr|z3d public static final int DEFAULT_INITIAL_RTT = 8*1000;

dr|z3d private static final int MAX_RTT = 60*1000;

cumlord I like what’s going on with susidns, seems like a lot of wasted tunnels though? Or no big deal?

dr|z3d no big deal, tunnels are short lived.

dr|z3d you also win some cover traffic.

cumlord lol thanks

onon_ Ok AckDelay = 1/10 RTT but if > 100 ms set to 100 ms

onon_ simple

dr|z3d if you clean out the cruft in your addressbook once it's done a full hostchecker run, subsequent runs will be faster.

cumlord but I need all 20k of those dead domains

dr|z3d sure you do.

onon_ The problem is that the more AckDelay you have, the more packets it simultaneously acknowledges to the sender. And accordingly, the more packets it sends at the same time, because space is freed up in the window

dr|z3d if they're not truly dead, they'll return when your subscriptions update.

dr|z3d looking at it.

onon_ And the more packets the sender sends at the same time, if he does not use pacing, the higher the probability of packet loss due to peak traffic.

onon_ Do I really have to explain such simple things to you?

dr|z3d no, you don't. did I ask you to?

dr|z3d but it doesn't hurt to state your thinking, I'm not the only person in the channel. :)

onon_ > dr|z3d: no, you don't. did I ask you to?

onon_ Can I refer to this comment every time I am asked why java-router is not working well?

dr|z3d if you want to get a better idea of what the HostChecker is up to, cumlord, add net.i2p.addressbook=INFO to /configlogging and keep an eye on your logs. set refresh to 5s.

dr|z3d log refresh interval can be set by hovering over the file location on /routerlogs ...

dr|z3d onon_: you seem to be forgetting that a lot of the code you're reviewing isn't mine per se, it's inherited from I2P.

onon_ If you claim to be the developer of an “improved” version of i2p, then you should at least know how it works...

dr|z3d are you going to continue being an antagonistic ass, or can we just keep it impersonal and technical?

onon_ You guys get paid to develop. I don't get anything. And when I try to give you free advice on how to improve the operation of i2p-router, I get back:

onon_ > dr|z3d: no, you don't. did I ask you to?

dr|z3d I get paid to develop? *laughs*

dr|z3d where did you read that?

dr|z3d *** laughs some more. ***

orignal drozd is not paid

dr|z3d you want to know how much I2P has rewarded me since I started contributing 15 years ago? I'll tell you. zero, nada, nothing. not a cent.

orignal ann all money were stolen by Nazi

cumlord cool I’ll add it in :)

orignal afaik drozd has been kicked out from the project

onon_ Who the fuck is Nazi

orignal echelon

dr|z3d this is super old news, orignal *chuckle*

orignal right but it happened

orignal and str4d commited your code later

dr|z3d in as much as I lost my commit privs when mtn was a thing, sure, won't argue with that. that was maybe 14 years ago.

dr|z3d yup, that's also true.

dr|z3d str4d committed a year or more of my work renovating the console UI, and got paid for it.

dr|z3d easiest job ever, review some code, check it in, get paid. *laughs*

onon_ How confusing everything is...

onon_ Now I even less want to participate in this in any way.

onon_ What does zzz say about this?

orignal about what?

dr|z3d this is old news, historical. we've moved on.

orignal ofc Java-I2P has been acquired by stormy ))))

orignal tell us why zzz.i2p is read-only?

StormyCloud Good thing Java-I2P is open source nobody can own it

dr|z3d :)

orignal tor is open source too

dr|z3d zzz is read only because he got tired of zab trolling him on the site, and also presumably because he didn't want the maintenance burden.

orignal so?

orignal zab is not online for a long time

orignal and zzz.i2p is still read only

onon_ zab is zlatnib?

orignal yes

orignal another faggot )))

dr|z3d any more info relating to zzz.i2p you'll have to query him about.

orignal let me tell you somthing

orignal 10 years ago zzz blamed me that I post my text on habr rather than zzz.i2p )))

orignal *texts

dr|z3d what is this? a walk down memory avenue?

dr|z3d *** chuckles. ***

orignal ?

orignal habr.com/ru/users/orignal/articles

dr|z3d 10 years ago, 12 years ago...

dr|z3d WE'VE MOVED ON!!!

dr|z3d even if your ui hasn't.

onon_ Nothing funny. dr|z3d

onon_ Oh, thats funny =)

dr|z3d :)

onon_ Good joke

StormyCloud speaking of @onon_ I moved web irc over to i2pd and much more stable irc connection. idk will look at this after 2.11

orignal * [orignal] idle 00:05:46, signon: Sat Nov 29 16:02:20

orignal that's how irc stability should look like

orignal it's ilita ofc

orignal and it's not a local connection

orignal * [orignal] is connecting from *@er7pg2v2l67ab3pd4n6ux2b2imqrdofc2evkl55xe55gvkz65c3q.b32.i2p 127.36.126.243

onon_ In this case, it’s not entirely clear to me why, if dr|z3d is in no way connected with the “canon” i2p router, why does it have such problems with making changes to the code

orignal dr|z3d do you happen to know what has happened to that Hamilton alcoholic?

dr|z3d I have no issues whatsoever making code changes.

dr|z3d clue me in, orignal

orignal idk

orignal don't see her for a long time

dr|z3d oh. you mean Sadie?

orignal yes

orignal do you know that Hamilton is a shithole?

dr|z3d maybe the funding dried up? dunno.

dr|z3d I wasn't aware, no.

dr|z3d there you go, onon_, let's see how she blows: git.skank.i2p/i2pplus/I2P.Plus/commit/47486ef979fc4f9b95bdae0eb08f6de079984e53

orignal that's what I thought

dr|z3d don't get me started on Sadie, please.

dr|z3d *** chuckles. ***

orignal no money and such people disappecr

onon_ private static final int MAX_RTT = 30*1000;

onon_ is it 30 seconds?

dr|z3d it is.

dr|z3d down from 60.

onon_ =\

dr|z3d still too high, or?

onon_ if you overload tunnel, real RTT can be more than 1 minute

onon_ In this case, if RTT > MAX_RTT do you change the route?

onon_ I don't know the logic behind your router, that's why I'm asking.

dr|z3d private void setRTT(int ms) {

dr|z3d synchronized(this) {

dr|z3d _rtt = ms;

dr|z3d if (_rtt > MAX_RTT) {_rtt = MAX_RTT;}

dr|z3d }

onon_ What is PROP_INITIAL_RESEND_DELAY

onon_ resend what?

dr|z3d look at the full file, review.

dr|z3d and remember, PROP_ is normally a user-configurable value.

onon_ It's too late to figure this out today. Perhaps tomorrow.

uop23ip onon_, does i2pd have a feature to rekey the http proxy for a new client b32 (even auto rekey after x minutes) while running the router? java-i2p has that feature

dr|z3d we'll meet in the middle, I'm setting MAX_RTT to 45s for now. see how that works.

onon_ uop23ip, no

onon_ http proxy have static key now, changes after reboot or button in webconsole

uop23ip "button in webconsole" there is a button for rekeying? you know where? onon_

onon_ Router commands->Reload tunnels configuration

dr|z3d which presumably terminates all active tunnels?

onon_ If the key is not specified or specified as transient, then it will be updated

uop23ip before that any changes to tunnel.conf and then reload tunnel configuration? or not needed?

uop23ip ok thx

onon_ If nothing has changed in the tunnel configuration, it will not reboot.

orignal not yet

dr|z3d close on inactivity, new key on open.

orignal good idea, no stream or leasests ten rekey