mozilla :: #conduit

18 May 2017
10:53zalunhttps://www.irccloud.com/pastebin/znKUU3cK/
10:53zalundocker-compose -f docker-compose.yml -f docker-compose/test.yml run phabricator test-ext
10:54zalunone issue though
10:54zalunI need to liberate every time https://pastebin.com/SxFTzM2r
12:12zalundkl|afk: https://github.com/mozilla-services/mozphab/pull/9
12:34dklzalun: nice. will take a look
12:40zalundkl: this as well, but I need to double check if that's working without mozphab modification as before I've been playing with conduit-demo
12:43zalunhttps://github.com/mozilla-conduit/demo/pull/7
12:47zalunone ../ needs to be removed before arc and it should be working, but not the tests
12:47zalunliberate is working but then lowercase is killing it
12:55zalunI will need to leave in a while - is there a reason that after I run a command (arc liberate) maps is not remembered when I later test (run arc unit) ???
13:00dklnot yet familiar with that. davidwalsh might know
13:22imaduemejabba's posted some interesting points here: https://github.com/mozilla/iam-project-backlog/issues/144
13:22imaduemeaffecting how we access users scm levels
13:30imaduemeAlso I was wondering - if the plan is to let anyone autoland code if they are part of the group for that module/area of code - are scm levels the right way to do that? Or are we just trying to get things to work as is right now
13:43marsdkl: the conduit-demo cleanup PR looks good to merge
13:43dklcool
13:43dklfeel free
13:49imaduemefixing that other build failure now david :) https://circleci.com/gh/mozilla-conduit/lando-ui/26?utm_campaign=vcs-integration-link&utm_medium=referral&utm_source=github-build-link. (My bad for not checking tests in the original PR review)
14:10davidwalshimadueme: woops, didn't know we had tests
14:10davidwalshMy bad
14:25imaduemedavidwalsh / mars: https://github.com/mozilla-conduit/lando-ui/pull/10 (I think the Auth0 card is done anyway, this is just fixing the tests)
14:26imaduememars: You might have opinions on the tasks.py changes so please do share them!
14:26imaduemeI can explain my rationale further if it helps
14:41imaduememcote: https://github.com/mozilla-conduit/demo
14:43mcoteright
14:43mcotefound it :)
14:45mcotemars: your sound appears to be dead
14:45mcotewe can hear you
14:49zalunhttps://www.dropbox.com/s/rzx20b1ov38hap7/Screenshot%202017-05-18%2016.49.22.png?dl=0
15:00globimadueme: fwiw jabba's proposal having scm levels in two fields would totally work for us
15:00* imadueme
15:00globhappy to go over the differences between the two if you need clarification about what's going on there
15:02imaduemethat would be helpful ya, so we can give him a good response. My only concern is whether he thinks the cron job is the right idea or not - I'm assuming so since he proposed it. I just don't want him to make something he's uncomfortable with because another team asked for it :P
15:03globi don't think we should think too much on the implementation details; we have a requirement which they agree is valid and should be implemented
15:03globso..
15:03imaduemetrue :) and it is pretty important
15:04globthe reason why there's two 'groups' is your ability to push code to hg.m.o will expire if you don't use it for a while
15:04globhowever
15:04globyou aren't removed from the scm group
15:04globso there's really two bits of information that's important from an IT perspective
15:04glob1. is this person a member of the group?
15:04glob2. can this person use that group membership to push?
15:05globas jabba pointed out, they can't predict which of those is important to consumers
15:06globfor conduit we should probably only care of a user is a member of the group
15:06globbecause i suspect using autoland doesn't refresh the `last used` timer
15:06globso someone only using autoland may have their direct push access lapse. they still should be able to use autoland in this case
15:07imaduemethat makes sense
15:14imaduemeI'm glad it is simpler than I thought it would be :)
15:30mcotejust have to run to the washroom during this technical chat :)
15:30mcotebrb
15:37mcoterobots are not to be trusted
15:43globmars: the devops code for servo is at https://github.com/edunham/ansible-rust-infra (homu's in there somewhere)
15:52zalunbattery going out - see you later
15:53* imadueme
16:03imaduemeWell glob based on what you said I'm comfortable asking jabba to go ahead with his plan. (For anyone else reading that plan is here: https://github.com/mozilla/iam-project-backlog/issues/144)
16:03glob+1
16:03imaduememcote don't know if you have opinions on ^
16:03globoh, wait..
16:04* imadueme
16:04glob
16:04globthat's right. emoji
16:04* imadueme
16:05imaduemeI was a terrible offender of overusing emoji at my last workplace
16:08globnever would have guessed that
16:12davidwalshdkl: Do you have any PHP I can look at for what you're trying to do?
16:13dkli threw away the little bit i did before and had not had a chance to try the wrapper class you mentioned so not yet. I will try to get something together this afternoon unless you beat me to it
16:39imaduemeI created this card so we don't forget about it: https://trello.com/c/NkMu5oDR/378-determine-how-we-want-to-handle-security-problems-in-github
18:01mcotezalun|afk: you able to come to the ops meeting?
18:28zalunsorry have issues with connection
19 May 2017
No messages
   
Last message: 36 days and 21 hours ago