18:00:41 <ahf> #startmeeting weekly network team meeting, 29, jan 2018
18:00:41 <MeetBot> Meeting started Mon Jan 29 18:00:41 2018 UTC.  The chair is ahf. Information about MeetBot at http://wiki.debian.org/MeetBot.
18:00:41 <MeetBot> Useful Commands: #action #agreed #help #info #idea #link #topic.
18:00:44 <ahf> hello
18:00:51 <catalyst> hi
18:01:21 <isis> hi
18:01:23 <isabela> oi
18:01:38 <ahf> the pad is https://pad.riseup.net/p/W6ZrtvaaUSO9
18:05:51 <ahf> nothing marked in bold on the lists that needs to be discussed?
18:08:05 <ahf> should we begin with the discussion points?
18:08:09 <dgoulet> maybe point out to everyone to picka week for the hackfest in June
18:08:16 <dgoulet> haah yes that
18:08:45 <ffmancera> Can I get the link please? I cannot find the correct email hehe
18:08:48 <ahf> yes, teor has written out a document that works a bit like a "doodle" with which dates and location for the upcoming summer hackfest - could everyone please write if they can make it some of the dates there?
18:08:49 <isabela> i think others should be prioritized over me - so i will let others pick first
18:09:29 <isis> this is the hackfest that is maybe in hawaii, would now be in june instead of april?
18:09:40 <ahf> ffmancera: hey, the pad is here: https://pad.riseup.net/p/W6ZrtvaaUSO9
18:09:48 <ahf> isis: yep, location is being discussed too
18:09:53 <ffmancera> Thanks!
18:10:39 <ahf> i don't think the hackfest document is public, but network team people can see it on the team list. i think teor would really like us to decide on a date so he can book his traveling
18:10:40 <isis> ahf: cool thanks
18:10:59 <isabela> (btw, point 1 of discussion is important as well, 200 tickets on 0.3.3.x )
18:11:16 <ahf> isabela: yep, i think we have taken them in out-of-order right now :-P
18:11:28 <isabela> :)
18:11:40 <dgoulet> on my todo very short term to triage many things out of 033
18:11:48 <ahf> but i guess there is not much more to this hackfest stuff: it looks there is people who wants to help teor with it and i guess we can pick a date by majority?
18:12:02 <isabela> isis: from my update, I am working on the email nick asked last meeting. you should receive it soon
18:12:25 <ahf> dgoulet: ok, so  we move to:     * There are over 200 open tickets in 0.3.3.x.  Some are bugs we should fix.  Some we should defer to 0.3.4.x. Some should get tossed into the aether. Suggested process?
18:12:26 <dgoulet> ahf: yah most important is date and then see if budget
18:12:36 <isis> isabela: cool, thanks! is there anything you need me to do?
18:12:54 <isabela> isis: it will be on the email
18:13:23 <isabela> re: 200tickets
18:13:40 <isabela> i would ask people to please prioritize what is there that is related to the team roadmap
18:14:09 <isis> isabela: also… there's been a problem that google shut down my access to the compute/appengine server being used as the reflector, and i need someone from tor org to set the account in order (i can't use my personal credit card to pay for the reflector)
18:14:09 <isabela> or in other words to make sure previous commitments are included on the tickets you pick to stay on 0.3.4.x
18:14:35 <isabela> isis: np, if you can email me the information i can make sure that is done
18:14:46 <isis> ok great, will do
18:16:28 <ahf> isabela: so it would be good if we all could walk over our tickets both in 0.3.2.x and 0.3.3.x and figure out what to do with them, but ensure to prioritize the roadmap?
18:16:45 <ahf> as in, tickets that belong to tasks on the roadmap
18:17:21 <isabela> i think so, if folks want twe can start now or set up a session to do it
18:18:46 <isabela> what does folks prefer?
18:18:48 <ahf> should we do a session later this week where maybe nick can be around too?
18:18:49 <isabela> should we have nick around
18:18:53 <isabela> hehe
18:18:55 <ahf> :-S
18:19:00 <isabela> i think might be better to have nick around
18:19:09 <isabela> we can try one on wed or thursday
18:19:19 <dgoulet> I think we can triage ticket each of us individually quite a bit
18:19:21 * isabela prefers thursday cuz wed i have ux meetings
18:19:34 <dgoulet> triaging 200+ tickets in a meeting will take us 5h
18:19:38 <isabela> hahahaha
18:19:41 <ahf> yep, friday is a bit bad for me and thursday evening too (UTC from 19 to 21)
18:19:49 <dgoulet> but the last 20 tickest "should be in 033 or not" meeting makes sense
18:19:58 <isabela> ok so lets do this, i can send an email to the list asking folks to dedicate some time this week triaging things
18:20:01 <isabela> and give some orientation
18:20:08 <dgoulet> +1
18:20:12 <isabela> and next week we see if the result was good or not, if the 200 tickets went down
18:20:15 <isabela> lol
18:20:15 <isis> does anyone have opinions about the proposal meetings w.r.t. grouping them together (like discussing two proposals at the same meeting)?
18:20:49 <ahf> isabela: +1
18:20:52 <isis> yay for triaging
18:21:38 <isis> i think i'm leaning towards making all the meeting be separate so we can (hopefully!) keep them shorter
18:21:57 <dgoulet> sure
18:21:57 <isis> the proposal meeting pad is https://pad.riseup.net/p/Pxo2fQiiaSWo
18:22:23 <ahf> isis: agreed
18:22:33 <isis> ok, i will make some doodles
18:23:24 <ahf> cool
18:23:58 <isis> mikeperry: does prop#247 need/want a discussion meeting?
18:24:29 <mikeperry> not yet
18:24:42 <isis> ack
18:24:50 <mikeperry> I am going to rework it based on what we're converging on
18:25:10 <mikeperry> it has a lot of options that we are not going to implement. so mostly just cutting thosre to make it clearer and easier to understand
18:27:34 <isis> oh right, how is the github thing going?
18:27:47 <isis> i saw someone made a tor-test repo
18:28:10 <catalyst> yeah i'm trying to figure out if we can usefully use the "protected branch" feature of github
18:28:26 <ahf> what's the end goal with all of it? we should allow people to contribute via PR's, but we don't "merge" them on github but do the usual merging that nickm does?
18:28:41 <dgoulet> (ah! I was just about to ask this question ^)
18:28:47 <isis> same
18:29:09 <catalyst> right, we do like another project i used to work on, and have github be a read-only-ish mirror. don't press the big shiny merge button ;)
18:29:10 <ahf> (irssi went from being a dying project to a quite active project and got a ton of new contributors after we moved from "home" run svn to a github org, so i think it's a positive move to allow people to do github PR's even with our "custom" merge plans)
18:29:16 <isis> (also i think i don't really understand what a "protected branch" is?)
18:30:00 <Hello71> https://help.github.com/articles/about-protected-branches/
18:30:06 <isis> the "protected branch" thing would prohibit us from pressing the big shiny merge button?
18:30:08 <catalyst> "protected branch" covers a lot of different features. at a minimum it prevents people from force-pushing it or deleting it; there are additional options like requiring approved review or CI passing
18:30:11 <ahf> catalyst: if i submit a PR with some patches to torproject/tor.git on GH, and nickm pulls these commits in with a merge/rebase, wont GH automagically close to the PR as well?
18:30:21 <Hello71> github has very good documentation
18:30:25 <isis> Hello71: thanks!
18:30:37 <catalyst> isis: only if we make it so none of us regularly logs into an account with owner/admin rights, which seems annoying to set up
18:31:35 <catalyst> ahf: github will automagically close the PR when a merge or rebase of the PR commits gets pushed, assuming they are the exact commits pushed, i think
18:31:37 <ahf> i created an admin user there for org stuff
18:31:51 <ahf> exact commit == same hash or?
18:32:24 <isis> i think same hash yes? like rebase-merges don't count
18:32:28 <catalyst> ahf: yeah, i think same hash. (hopefully nobody is minting SHA-1 preimages at will yet)
18:32:33 <ahf> XD
18:32:35 <ahf> yet :-P
18:33:00 <catalyst> isis: i think a "rebase" that is a fast-forward counts (i.e., commit hashes don't change)
18:33:43 <mikeperry> isis: (if you were asking about my github thing, I am planning on trying to make that usable this week. it is just a collection of python scripts that only asn and I know how to use atm)
18:34:02 <isis> do we want to have a rule that trac.tpo tickets must be set up to track PRs?
18:34:21 <isis> mikeperry: wait what is your python script thing?
18:34:24 <catalyst> isis: i'm not sure what that means?
18:35:08 <ahf> if we want something with github pr must have a trac commit don't we run into some of the things we discussed a bit during one of the "let's evaluate gitlab" discussions where we wanted to see how far we could get without any scripting of any "sync mechanism" ?
18:35:55 <isis> like if i make a PR on github, and you review it and approve it, we then need to get nickm to merge… do we go make a trac ticket saying "hey please merge X from PR#foo"
18:35:57 <catalyst> i think if we wait until a sync mechanism gets implemented we might stall useful progress :-/
18:36:03 * ahf nods
18:36:20 <mikeperry> isis: https://github.com/mikeperry-tor/vanguards
18:36:39 <catalyst> isis: yeah, i think any PR on github would need a Trac ticket. whether we create an automated process is an interesting question but i'm hoping the volume will be low enough we can handle it manually
18:36:49 <isis> mikeperry: "Vanguards help guard you from getting vanned..." lmao
18:37:04 <ahf> haha
18:38:22 <isis> catalyst: ok, if not then we can pause taking PRs for a bit and make a bot
18:38:38 <catalyst> isis: sounds like a plan!
18:39:02 <isis> ok, do we have anything else important to talk about?
18:39:13 <ahf> i think we are done with the list of things
18:39:21 <catalyst> so for locking things down so none of us accidentally hits the big shiny merge button, it's cumbersome, etc.
18:39:55 <catalyst> do we want to live with the risk that it might happen and people will just deal if they have to rebase stuff when the "official" tpo repo mirroring happens and blows away the github-generated merge commit?
18:39:57 <ahf> but let's all remember to comment on dates for the summer hackfest - and isabela would send out an email with mass bug triaging :-)
18:41:51 <isis> i don't actually see a doodle/poll for the summer hackfest dates
18:42:01 <isis> we're just responding in the email thread?
18:42:05 <dgoulet> Pad
18:42:07 <ahf> isis: it's not a doodle - it's a "free form doodle" as a etherpad
18:42:15 <isis> oooh the pad
18:43:58 <ahf> catalyst: it's a terrible solution, but assuming something ever goes wrong with the "official" tor hosted tor.git, shouldn't the sync between that and the github/gitlab/<...> one always do a "force push" (overruling whatever ended up in the GH one that is not in the official one) ?
18:43:59 <catalyst> oh, also who's willing to help triage github PRs?
18:44:39 <catalyst> ahf: i don't know whether tor-bot does a force-push or not (or who's maintaining it for that matter)
18:45:24 <ahf> i don't even know what tor-bot is :-S
18:45:39 <catalyst> it's the bot account that pushes to torproject/tor to do mirroring from git.tpo
18:45:49 <isis> i believe weasel is maintaining tor-bot
18:45:52 <ahf> ah. i assume weasel runs it then
18:46:03 <catalyst> isis: ah, thanks
18:47:37 <dgoulet> catalyst: my two cents, I would recommend emailing tor-dev@ with the Github idea because adding another place to look at for code means a change in development flow, it means more ticket work, it means keeping in sync trac and Github ... all this could imo benefit from feedback
18:48:19 <catalyst> dgoulet: that's one reason i'm asking for volunteers :)
18:48:28 <isis> maybe some other project has already made a ticket-sync bot that we could repurpose
18:49:05 <isis> i can help write a thing to explain the idea of what we want to do
18:49:08 <dgoulet> catalyst: volunteer for what?
18:49:18 <catalyst> dgoulet: triaging new github PRs
18:49:47 <catalyst> isis: that would be a useful thing to write, thanks!
18:49:51 <dgoulet> well this is why I'm proposing an email thread because for instance I personally have issues with this workflow ;)
18:50:17 <isis> oh, i'm down to triage too
18:51:14 <catalyst> i'm hoping we will have enough regular github users volunteer that people who don't want to interact with github can just look at trac. (commenting on a github PR still needs a github account though)
18:51:24 <catalyst> isis: thanks!
18:52:24 <ahf> i think i visit github once a week at most after i've started doing tor stuff to check up on irssi things, so i don't think i'd be that useful there
18:52:26 <catalyst> anyway i think it's really important to run at least some CI on every change before we merge it and github pull requests will help a lot with that
18:52:41 <isis> yeah, i think it should be fine if some people don't want to be logged into github all the time
18:53:32 <isis> conversely, i have to be logged in all the time for all the other stuff i've worked on, so i don't really mind
18:53:51 <ahf> we have ~5 min. left now - what should we decide on this? discuss on ML like dgoulet proposed?
18:53:57 <isis> ok, we are starting to run out of time
18:54:34 <isis> +1 ML?
18:54:34 <dgoulet> please yes, lets discuss this Github idea because it is something new when people start using Github instead of Trac ticket because we offered them Github
18:54:49 <dgoulet> but our whole flow is Trac
18:54:51 <catalyst> yeah let's take it to the list
18:54:56 <ahf> cool
18:55:05 <ahf> anything else we are missing or should we end the meeting?
18:55:20 <isis> i think we're done?
18:55:30 <isabela> yes, i think it needs discussion with more ppl beyond this meeting
18:55:54 <ahf> ok!
18:55:58 <ahf> #stopmeeting
18:56:04 <ahf> #endmeeting