I2P dev meeting, May 4, 2021 @ 20:00 UTC

Quick recap

  • Present:

eyedeekay, zzz, zlatinb

Полный журнал IRC

(04:01:04 PM) eyedeekay: Hi everyone, it's time for the May 4th meeting
(04:01:13 PM) zlatinb: hi
(04:01:21 PM) zzz: hello
(04:01:39 PM) eyedeekay: 1) Hi 
(04:01:39 PM) eyedeekay: 2) Mac Launcher Report, Follow-up 
(04:01:39 PM) eyedeekay: 3) Trac Migration Report, Post-op 
(04:01:39 PM) eyedeekay: 4) 0.9.50 release
(04:01:39 PM) eyedeekay: 5) Update Channels Report
(04:01:39 PM) eyedeekay: 6) Docker Improvements
(04:01:39 PM) eyedeekay: 7) Bote Plugin Keys
(04:02:17 PM) eyedeekay: zab are 2) and 5) likely to overlap, should I put them together?
(04:02:26 PM) zlatinb: sure
(04:02:56 PM) eyedeekay: OK so let's swap 3 and 5 from that list above, and do update channels right after Mac Launcher
(04:03:11 PM) eyedeekay: 2) Mac Launcher Report
(04:03:59 PM) zlatinb: so far I've received one positive report by an unknown user, and know at least a few people have tried the .dmg
(04:04:28 PM) zlatinb: so for the installer part I think we're in a very good shape.  I can't think of any changes required that are not related to the update functionality
(04:04:49 PM) zlatinb: s/installer/app bundle/
(04:05:24 PM) zlatinb: that's all on stricly-2) from me
(04:06:10 PM) eyedeekay: Excellent. I don't have anything to add right, so we can move on to 3) Update Channels
(04:06:24 PM) eyedeekay: Unless zzz has something?
(04:06:36 PM) zzz: no
(04:07:00 PM) eyedeekay: Ok then zlatinb update channels are also your topic
(04:07:22 PM) zlatinb: zzz and I did some initial analysis/scoping of what needs to happen to enable update channels
(04:08:05 PM) zlatinb: the consensus (I think) is that there will be some changes to code in i2p.i2p as well as some code residing in the mac-jpackage repo
(04:08:36 PM) zlatinb: we're still enumerating all the corner cases but so far haven't come upon a dealbreaker
(04:09:24 PM) zzz: agreed, sounds pretty straightforward and not too much effort. testing is probably more of the work than coding
(04:09:36 PM) zlatinb: I'm very busy until the release but after that will focus on this.  Can get more technical but it gets very low-level for this meeting
(04:09:39 PM) zlatinb: eot
(04:10:05 PM) eyedeekay: Thanks for the report
(04:10:12 PM) eyedeekay: That brings us to 4) 0.9.50 release
(04:11:08 PM) dr|z3d: you missed Trac migration.
(04:11:26 PM) eyedeekay: I was going to do it as 5, not 4
(04:11:40 PM) dr|z3d: ok, as you were!
(04:11:45 PM) eyedeekay: We're 11 days away from the release now
(04:12:09 PM) eyedeekay: Tags are set to be frozen tomorrow
(04:12:22 PM) eyedeekay: I have no more string changes for i2p.i2p
(04:13:43 PM) eyedeekay: zzz, zlatinb what would you like to add?
(04:14:08 PM) zzz: not much... I'll push the strings to transifex at 4 PM UTC tomorrow
(04:14:26 PM) zlatinb: orignal made an interesting point just 30 minutes ago about NTCP queue capacity, might be worth looking into b4 the release
(04:14:27 PM) zzz: I'm done with 50. already working on the next one
(04:15:18 PM) zzz: I didn't see it, but I'd be reluctant to make any changes now. I am testing some NTCP queue changes for the next release
(04:15:29 PM) zzz: eot
(04:15:38 PM) zlatinb: eot from me too
(04:15:53 PM) eyedeekay: 5) Trac Migration Report, Post-op
(04:16:35 PM) eyedeekay: Trac migration was sticky mostly for the reasons I felt it needed to happen, in particular trac xmlrpc broke on our instance at about the same time as last months meeting
(04:17:34 PM) eyedeekay: After trying and failing to fix it for a couple weeks I just decided it would be easier to (carefully) scrape our trac issues down and migrate them to gitlab using the gitlab API
(04:18:20 PM) eyedeekay: Otherwise, it was successful, and as a by-product created a readable static archive of all our trac tickets at this time
(04:18:32 PM) eyedeekay: eot
(04:18:44 PM) zzz: so whats the status? done?
(04:19:16 PM) eyedeekay: For the purposes of tracking tickets, trac migration is done. Trac still has wiki articles of some interest to back up but the tickets are done.
(04:19:43 PM) zzz: ok. I changed the urls in our code to point to gitlab
(04:20:14 PM) eyedeekay: I changed most of the ones on the website, but am still grepping through .rst files for the last few
(04:20:28 PM) zzz: can you please add notes and links on trac home page and ticket page and login page and wherever else, with new i2p and clearnet links?
(04:20:42 PM) eyedeekay: Sure, will do
(04:21:49 PM) zzz: this now makes us reliant on gitlab (when it was just code, we could always use github) ... do we have any backup admin?
(04:21:49 PM) eyedeekay: I will also go through all the README's and make sure they reference the correct places too
(04:22:50 PM) eyedeekay: echelon has an admin account on gitlab, but no one else has SSH access to the server underneath right now
(04:22:50 PM) mode (+v anonymousmaybe) by ChanServ
(04:23:30 PM) eyedeekay: I can look into syncing the issues with github using a bot, it's not that different than the second half of the migration process
(04:23:39 PM) zzz: ok, you two may want to review who can do what to make sure we're covered
(04:23:45 PM) zzz: good job
(04:24:09 PM) eyedeekay: Thanks
(04:24:22 PM) eyedeekay: That brings us to 6) Docker Improvements
(04:24:42 PM) eyedeekay: zlatinb do you want to fill the people who haven't tried them yet in here :)
(04:25:10 PM) zlatinb: lol yes, the new docker image is smaller and supports persistent volumes for configuration and snark downloads
(04:25:35 PM) zlatinb: documentation is in the source, the Docker.md file.  I would like to add a page to the website with that same content
(04:26:03 PM) zlatinb: that's really it
(04:26:30 PM) eyedeekay: Good call about the site, right now we advertise it but don't document how to use it at all
(04:26:40 PM) zzz: who is in charge of the geti2p docker account and who else has access?
(04:26:48 PM) zzz: or does it not work like that?
(04:27:35 PM) zzz: does it just auto-build every checkin and that's it?
(04:27:37 PM) eyedeekay: I'm in charge of the geti2p docker account, I can grant access to people from gitlab, it was started by Ace Barry or hkparker IIRC but I'm the admin now
(04:28:04 PM) eyedeekay: It builds the `latest` every checkin and builds an image for every tag beginning in `i2p-*`
(04:28:50 PM) zzz: ok so whatever changes zlatinb did are already in there
(04:28:52 PM) zzz: got it
(04:29:00 PM) zlatinb: yes
(04:29:30 PM) zlatinb: eyedeekay: I saw you just dockerized the android build process?
(04:30:50 PM) eyedeekay: Yeah I did, it was a way of bundling up all the release requirements into a re-usable form
(04:31:35 PM) zzz: eyedeekay, speaking of android, I saw something about google adding more rules and bumping requirements effective later this year. You may wish to put aside some time before this release to get ahead of it
(04:33:10 PM) eyedeekay: I'm double-checking all my Android release stuff this week to make sure that all goes smoothly
(04:34:18 PM) zzz: as I said the new rules aren't effective for a few months but wouldn't hurt to address them now
(04:34:41 PM) zzz: or, it might hurt, but better sooner than later
(04:34:42 PM) zzz: eot
(04:35:14 PM) eyedeekay: Well depends on F-Droid, sometimes they lag behind GPlay in requirements in a way which is somewhat mutually-exclusive, but it'll be better to know about it if it's going to happen
(04:36:02 PM) eyedeekay: I think we're ready for number 7) Bote Plugin Keys
(04:36:20 PM) eyedeekay: This one came up for me in conversation with some redditors last week
(04:37:06 PM) eyedeekay: People are trying to use mhatta's fork of Bote but they are not able to do so because they are not able to install the plugin keys easily
(04:37:30 PM) eyedeekay: They also mostly don't know how to intepret the certificate error in the sidebar to troubleshoot the issue
(04:38:17 PM) eyedeekay: s/keys/certificates/
(04:38:41 PM) eyedeekay: I would like us to consider adding mhatta's to the default so people no longer encounter this error
(04:39:17 PM) zzz: 1) he should provide better instructions to his users; 2) he needs to make the request of us
(04:40:22 PM) eyedeekay: Fair enough.
(04:40:46 PM) eyedeekay: That brings us to the end of the listed topics, anything else to add?
(04:41:06 PM) zlatinb: yes, I'd like us to think about making it easier to build testnets
(04:41:08 PM) zzz: and I'd ask that he get .49 into debian, which never happened
(04:41:55 PM) zlatinb: we've had two people build LXC testnets and one person build Docker, all three use quite different approaches
(04:42:14 PM) zlatinb: so is there any interest in figuring out what the pain points are and making things easier?
(04:42:51 PM) zzz: I have interest in finding out if there's interest :)
(04:43:10 PM) eyedeekay: Yes there is from my side, I would like to get a testnet running, preferably a docker one
(04:44:13 PM) zlatinb: cool.. so we should look into it.. of the top of my head initial seeding is the worst part
(04:45:14 PM) eyedeekay: Are there any Docker testnet instructions written down yet or are they all LXC-based?
(04:45:18 PM) zzz: my solution to seeding worked well for me, it's roughly solved for lxc
(04:45:47 PM) zlatinb: LoveIsGrief may have something in his repos on gitlab 
(04:47:55 PM) zlatinb: eot from me
(04:49:07 PM) eyedeekay: I guess then if I want a Docker testnet I should probably look into their work, and fill in whatever blanks I encounter based on the process for LXC
(04:49:43 PM) zzz: I checked all my horrible lxc shell scripts into i2p.scripts
(04:50:29 PM) eyedeekay: Thanks zzz, horrible or not I'm sure they'll tell me what I need to know
(04:51:32 PM) eyedeekay: Anything else for the meeting?
(04:51:50 PM) eyedeekay: timeout 1m
(04:51:50 PM) zzz: no
(04:51:58 PM) zlatinb: not from me
(04:52:19 PM) eyedeekay: OK then thanks everyone for coming
(04:52:44 PM) eyedeekay: I'll post the logs to the site shortly, see you around IRC