#i2p-dev (irc2p) | 2.7.0-2

eyedeekay Electrical storm just killed a surge protector and a printer. I'm probably going to take some stuff offline in the next hour and not bring it back up until the weather clears. Gitlab will be unaffected will be unaffected, everything else under *.idk.i2p will go down

zzz eyedeekay, can you check the reverse proxy for the changes tab on MRs? I just get the spinny circle forever

eyedeekay Yeah I'll take a look

zzz I have to do code->download->plain diff as a workaround

zzz thanks

eyedeekay I think I see it, CSP thing, there's a script that's getting denied on the browser side

eyedeekay Should be fixed soon

eyedeekay Looks like fixing it will require me to edit both the gitlab config and the nginx config so it will require a restart, I do a scheduled restart every night so it will be deployed by morning

eyedeekay Hi everybody, welcome to the July 5 Meeting

eyedeekay 1. Hi

eyedeekay 2. 1.9.0 development status

eyedeekay 3. Support for Apple Silicon in the Mac easy-install bundle

eyedeekay 4. Windows easy-install bundle - out of beta?

zzz hi

zlatinb hi

eyedeekay Hi guys, anybody else here today?

eyedeekay Moving right into 2. 1.9.0 development status

eyedeekay We are at a little less than 6 weeks in I think, approx. 7 to go

eyedeekay zzz and orignal have been working hard on the implementation of SSU2 in i2p.i2p and i2pd

eyedeekay It won't be activated in 1.9.0 but if I understand correctly it's nearly done

eyedeekay I've been working on UDP tunnels again, mostly there but something's broken still, I'll probably need to ask zzz for help this week

eyedeekay zzz zlatinb anything else to add

zzz that's right, ssu2 is mostly working

zzz still fixing small bugs

zzz right now I'm working on making tunnel peer selection more efficient (unrelated to ssu2)

zzz other than that, just doing the usual bug fixing

zzz targeting a late august release

zzz EOT

eyedeekay Anyone interested in helping test should visit zzz's forum for instructions and to give feedback: zzz.i2p/topics/3314-how-to-enable-ssu2-in-i2p-1-8-0

eyedeekay 3. Support for Apple Silicon in the Mac easy-install bundle

eyedeekay zlatinb this is your topic, take your time

zlatinb hi, the big issue was jbigi but that is now solved via backporting the x18 register patch

eyedeekay So the easy-install bundle no longer needs to run on the emulated x86_64 mode?

zlatinb I would like to put up an official and notarized beta around end of july

zlatinb it has to be a separate download

zzz this would be a second bundle. we'd have two, one for x86 and one for arm

zzz I'm in favor, as it appears the extra dev effort would be small, and the speedup is large. Right?

zlatinb there is a caveat with updates when switching between bundle4s

zzz sure, as long as there's instructions on how to switch, even if it's as simple as uninstall and reinstall, that should be fine

zlatinb i'll put them on /download/mac

zlatinb the news url needs to be changed manually

zzz the dev effort is small and the speedup is large, correct?

zlatinb speedup is there, also looks good to keep up with the times :) dev effort except for notarization is small

zzz great. +1

eyedeekay +1

zlatinb thx EOT

eyedeekay Thanks zlatinb

eyedeekay 4. Windows easy-install bundle - out of beta?

eyedeekay I said ~3 months ago that I would be ready to move the easy-install bundle out-of-beta when I had a successful end-to-end update

eyedeekay I got that at 1.7.2 IIRC, but between 1.7.2 and 1.8.0 there were a number of bugs in how it worked when used with an external I2P router(i.e. a non-jpackaged router) in this situation the bundle is intended to work as a firefox-launcher only but this functionality was broken from 1.7.3-1.7.6

eyedeekay So I pushed it back again, but now I'm sure that all of the issues with running in both bundled and non-bundled modes are resolved

eyedeekay So I think it's finally ready to move out of beta, all the core functionality I thought was necessary has been established

zzz here's some things that may still need to be resolved:

zzz - is it just you that is successfully updating or are there other users that are also?

zzz - license requirements all met now?

zzz - are the windows bundle and the mac bundle now on the same JRE and roughly same release schedules?

zzz -- i.e. prepared for the 3-monthly JRE releases

zzz - and the usual question, do we have the resources to support this as an official product

zzz - also, are we removing the non-bundle download like we did for mac? or not?

zzz eot

eyedeekay License requirements are now fixed

eyedeekay I don't get a lot of feedback but at least 2 non-me updaters

eyedeekay I am sticking to LTS JRE, I think zlatinb is doing 18, but yes we are on the same schedule more-or-less

eyedeekay I do not think the resources required to work on the bundle are extensive, the hardest part is getting the build set up and it's not that hard, small investment now that it all works

eyedeekay I kind of want to remove the non-bundled download or possibly make it an "advanced installation" procedure

zzz - I don't think 2 non-you updaters is enough testers to declare it non-beta. Can you pimp it more on reddit or something?

eyedeekay Sure can do

eyedeekay I'm not in a hurry to rush it out

zzz - I'd like to see win and mac on the same JRE. Let's consolidate. I don't know which of you is right, but one of you is

zzz can you two come to an agreement?

zzz or are there reasons to be different?

zzz to be clear, not just the same version, but the same JRE supplier

zzz this will reduce the CVE review required every 3 months

eyedeekay I'm not married to my decision, I chose LTS strictly because of the statement "LTS"

eyedeekay We're both using Oracle right now to my knowledge

eyedeekay The only vendor with a similar update cycle is Amazon

eyedeekay I'd rather use Oracle than Amazon I think

zzz you two should be making these decisions together and in sync. I don't know why you're not and it sounds like you don't know either? :)

eyedeekay No I don't know. We did discuss vendors and update cycles at a couple points which is when we landed on Oracle due to releases being very quick compared to CVE's, but I don't know why I'm on 17 and zlatinb's on 18 now

zlatinb I don't remember the exact reason either

zlatinb maybe I wanted to test the ram reductions

zzz ok. I'm in favor of it coming out of beta, but let's get on the same JRE, get some more testing first and come back to us in a month or two

zzz I also want to think more about whether to remove the standard installer, maybe even worth a separate meeting about that

zzz eot

eyedeekay Re: actual reasons, perhaps a discussion for another time but I know the reason I'm on Oracle/17 is because I needed to pick a vendor who would respond to CVE's in less than 24 hours and because I wanted to ensure that I would have stability in terms of what I could expect from the API's and the JVM. I didn't really consider other reasons

eyedeekay Based on a sample of 1 event, I assessed that Oracle and Amazon were the fastest to release an update.

eyedeekay I considered Amazon because they are packaged in chocolatey which makes life on Windows a lot easier, but decided on Oracle instead

eyedeekay So that's how I landed on Oracle/17, basically out of caution

zlatinb I can go down to 17 but really prefer to stick to oracle

eyedeekay eot

eyedeekay So would I, so we're agreed

zlatinb on a related note i'll be afk until a week after the next jre release

eyedeekay Will we/can we make any arrangement for signing the bundles?

zzz I think you need 18 for the best apple arm support? if so then let's drag windows up to 18 also. But again, you two shouldn't be doing JRE vendor research and selection in separate silos. work together and decide together

zlatinb hopefully there won't be any urgent fixes

zzz we have limited resources, let's not do stuff twice for no reason

zlatinb signing requires deanon

eyedeekay I can push an unsigned bundle and tell people that they'll need to click through the pop-up in the news if need be

eyedeekay Still signed `su3` just not signed `exe`

zlatinb can't "click-through" on mac

zlatinb oh the exe signing is fine

zlatinb we'\ll do it as usual

eyedeekay OK then. Let me know if there's anything I can do(less rapidly deanoning) to help

zlatinb Nothing really other than take down the mac download page in the worst case

eyedeekay Well you have my signal number if you need me to do it

zlatinb ok. it would be for a week at most

zzz eyedeekay, you're not setting the news URL in the feed, search for CHANGEME_URL_HERE :)

eyedeekay OMG I can't believe I missed that

eyedeekay Will do

zzz i put that in there a while back so you'd remember, I guess it didn't work (((

eyedeekay Re: Java 18 and Apple hardware, zzz just mentioned that 18 may have better support? If that's the case than that would be a thing I didn't know before and a reason to use 18 instead, can you confirm that?

zlatinb haven't done my homework on that sorry

zzz maybe a wild guess on my part, but don't need to decide in this meeting

zlatinb but if linux arm support is any indication it's very likely

zlatinb history of*

zzz but in general, newer is better, so isn't the point of bundling the JRE in a "easy bundle" is to have the latest?

eyedeekay I don't know, sometimes stable is better, and sometimes "easy" is somewhere in between

zzz that's why putting a LTS in a bundle doesn't make sense to me.

eyedeekay Probably won't settle it today, but maybe I'm wrong about 17. zlatinb do you have a time that would be good this week to hammer out the 17/18 debate?

zlatinb generally any time is good and also bad because of my injury unless you want to get on a call

eyedeekay I'm happy to get on a call, we can figure out a time in private

zlatinb ack

eyedeekay Anything else for 4?

eyedeekay Or for the meeting?

eyedeekay Timeout 1m

eyedeekay All right thanks everybody for coming

eyedeekay Same time next month

zzz thanks eyedeekay

zzz also FYI I'm working on a fix so that the news doesn't pop back up in the console everytime you rebuild it, only when a new news entry appears

zzz because there's all these separate feeds plus rebuilding for translations, it gets annoying

zzz it's all coded, just haven't gotten around to finishing the testing

SilicaRice why does the SAM bridge time out without ever sending a PING?

zzz shouldn't

zzz afaik

SilicaRice sending PING to it keeps it alive, but it never seems to send a PING of its own

SilicaRice but I don't remember it sending a STREAM STATUS when disconnecting, so maybe it's an OS thing where it kills idle sockets?

zzz are you asking for 3.2?

SilicaRice yes

SilicaRice MIN=3.2

SilicaRice Blinded message

eyedeekay I'm not sure I understand "sending PING to it keeps it alive, but it never seems to send a PING of its own" in the context of PING/PONG?

SilicaRice trying to read the control socket blocks forever until it suddenly closes, without ever reading a single "PING"

SilicaRice it's not a big deal (can just write instead) but it does seem like a bug? but this is on 1.8.0 rather than git HEAD

zzz if you think it's a bug, open a ticket on gitlab with all the info

zzz when you send a ping do you get a pong?

eyedeekay quick test with telnet seems to work on my side

SilicaRice yes

zzz works for me:

zzz > HELLO VERSION MIN=3.2

zzz < HELLO REPLY RESULT=OK VERSION=3.3

zzz > SESSION CREATE ID=x STYLE=STREAM DESTINATION=TRANSIENT

zzz < SESSION STATUS RESULT=OK DESTINATION=...

zzz < PING 1657057368888

zzz < SESSION STATUS RESULT=I2P_ERROR MESSAGE="PONG timeout"

zzz Connection closed by foreign host.

SilicaRice hmm...

SilicaRice will do further testing, maybe I forgot a newline somewhere