mozilla :: #conduit

7 Sep 2017
11:54* zalun had to fix the repository structure and update diffs in Phabricator. That teaches a lot about workflow...
11:57zalunalso - using `origin` as my own remote and `upstream` as the mozilla-conduit one doesn't work well with the `arc`
13:20mcotedkl: are we syncing the BMO sec group crypto-core-security over to Phab?
13:59mcotedavidwalsh: did we never do wildcard support for the security groups to be synced to phab?
14:00mcotelet me put it another way: the current setting is "core-security". are we only syncing that single group?
14:01davidwalshmcote: dkl would have set that setting as I'm not an admin, but I believe core-security is the only current group
14:01davidwalshI have no way to see though
14:01mcoteyeah that's not right :)
14:01mcotebut yes, it's only that
14:02mcotedavidwalsh: so I'm not sure if our "missing-group" functionality is working
14:02mcotedavidwalsh: I thought we were posting a comment when a revision is created against a bug that is in a security group that we aren't mapping over
14:03mcoteglob says that's not the case, looking at the code
14:03globhttps://github.com/mozilla-bteam/bmo/blob/18f98f9c23c427c8a253e8663237fe745eb34538/extensions/PhabBugz/lib/WebService.pm#L104
14:04davidwalshI added a message for groups here: https://github.com/mozilla-bteam/bmo/blob/18f98f9c23c427c8a253e8663237fe745eb34538/extensions/Push/lib/Connector/Phabricator.pm#L76
14:04mcoteI'm not sure if it's working
14:05* glob sees duplicated code
14:05mcoteso today someone tried to post a revision against an nss security group, but the build plan failed
14:06mcoteI assume if this is working correctly the build plan wouldn't actually be marked as failing? or would it?
14:06globlooks like the webservice needs to do the same
14:06mcotedavidwalsh: wouldn't that code you posted only happen when there's a *change* to the bug?
14:06mcotenot on original post?
14:14davidwalshOK, so looking this over
14:14davidwalsh1. Yes, we need to add a message here
14:15davidwalsh2. Push connector only handles changes, not the creation
14:15davidwalsh3. My understanding of our original cards, as was dkl's, was that we'd handle policy updates manually for unknown groups, which is why we're throwing this error
14:16davidwalshMy memory serves that this was one of those "odd cases" that came up as we wrote this; the odd case card
14:16davidwalshI do see that this halts creation of the revision on BMO so we do need a few changes her
14:16davidwalshhere
14:16globdavidwalsh: you should move the group param parsing into a single sub; would make it easier to support globbing
14:18davidwalshSure, sounds like a good idea
14:18* glob likes globs
14:37dklmcote: the current group list is core-security, crypto-core-security now. There are others we should add I am sure. Does NSS have their own security group as well?
14:44mcotehttps://developer.mozilla.org/en-US/docs/Mozilla/Projects/NSS
15:10davidwalsh|pto|afkAs per usual my timing sucks; I requested PTO for today on Tuesday and will be out most of the day
15:11davidwalsh|pto|afkdkl: I think you may need to update your patch with the idea that the Bug ID wont always exist and you wont want to make that revision call if ther'es no bug id
15:11davidwalsh|pto|afkdkl: You can use the `get_bugzilla_bug_id` I created in there
15:14dkldavidwalsh|pto|afk: will do
15:15davidwalsh|pto|afkWe'll also need to make that check and run your new method if the bug ID changes from "" to 328473294, now that I think of it
15:19dkldavidwalsh|pto|afk: yeah. right now i key off of core:create and will need to figure out what to look for for a bug id change
15:22mcotehrm wishing gdocs had an "export as markdown" option
15:23mcotehaving converted it from markdown to proper gdoc formatting for comments, now needing to convert it back to post... blech
15:58mobiwalshdkl: How do you want to split up the new tasks that came up this morning?
16:34davidwalsh|pto|afkdkl: Yeah, I don't think individual fields have a specific event, but in one of my previous PRs, the story object had an oldValue() and newValue() method; that may be a clue to detecting the field
16:34davidwalsh|pto|afkI can help out with that tomorrow
16:34mcotearen't you on PTO? :)
16:35davidwalsh|pto|afkI can't get he problem out of my head
16:36davidwalsh|pto|afkAnd my wife is in the changing room; thinking about this is easier than he pain of knowing impending credit card damage
16:37mcotehaha
16:37dklheh
16:37davidwalsh|pto|afkdkl: can we move some of the "sub revision" logic to the push connector and run it upon initial bug creation to avoid code duplication?
16:38dklbut make sure you pay attention when she asks "how do I look?"
16:38davidwalsh|pto|afk"You look like I better get a second job"
16:38davidwalsh|pto|afkLmao
16:39dkldavidwalsh|pto|afk: not sure how that would work as we dont know the revision id when a new bug is created and phab creates the attachments, not the user
16:40davidwalsh|pto|afkOk
16:43dklok so the NSS issue is because 1) the core-security-release group needed to be added to the sync grouplist, and 2) theBMO couldn't see the revision cause the policy was set custom by the user.
16:55ckolosis anyone using mozphab.dev.mozaws.net any more?
16:58davidwalsh|pto|afkCan we update then BMO sync list with all groups now or find out what wildcard we're using?
17:07dklckolos: im not
17:56dklmcote: ckolos: ops meeting today?
17:57ckolosafaik
17:57ckolosmcote makes the call though
17:58mcotemight as well sync up
18:01smacleodckolos: quite the new camera angle
18:01ckolosyou've been gon a while
18:39mcoteheh so somehow I have a weird binary character at the beginning of a text file (<U+FEFF>)
18:39mcoteas the first three bytes
18:40mcoteit doesn&#39;t show up in vi nor emacs
18:40mcoteand even deleting the first line doesn&#39;t remove this character, as it exists somehow before the edit buffer
18:40mcotehow on earth can I strip this out?
18:41mcotehm maybe with tr
18:45mcoteyeah that did it
18:45mcoteso weird
18:45* mcote blames windows
18:47ckolosdkl and glob&#39;s admin accounts created on prod phabricator
18:49dklckolos: got it thanks
18:50mcoteif anyone is curious, my blog on the phabricator decision-making process and my thoughts on decision making in general is finally up: https://mrcote.info/blog/2017/09/07/decisions/
18:57dklmcote: nice article
18:59mcotethanks :)
19:02mcotereading a related post now by the Open Innovation Team
19:02mcotethis part is funny: &quot;There are over a thousand open source software projects in the world&quot;
19:02mcoteI realize they probably mean polished, well-supported and widely used applications
19:02mcotebut it reads like something written in 1990 heh
19:03mcotethere are 10 million repos on github alone heh
8 Sep 2017
No messages
   
Last message: 14 days and 15 hours ago