[00:03:31] FYI [00:03:32] 00:03 < twentyaft> !log Phabricator upgrade starting momentarily. Service will be offline for a short time, most likely less than 5 minutes. [12:14:26] 10Phabricator, 06Project-Admins: Requests for addition to the #acl*Project-Admins group (in comments) - https://phabricator.wikimedia.org/T706#2618899 (10Aklapper) >>! In T706#2610784, @ovasileva wrote: > Hello. I am the new PO for the reading web team and would like to have permissions to create sprints and p... [15:13:44] greg-g: Is it you who made the "Days since last incident" thing on Wikitech? It's, umm, not happy: https://wikitech.wikimedia.org/wiki/Incident_documentation [15:16:03] James_F: not me, actually, I think it was Krinkle (I'd have to check the edit history to be sure) [15:16:10] Its comming from Module:Days_since_incidents [15:16:15] holy [15:16:18] just opened [15:16:24] Yeah. [15:16:25] that's an amazing failure state [15:16:30] It really is. [15:16:30] And Krinkle last edited these things from sometime erly march [15:16:32] Go go Lua. [15:16:41] I'll remove, fix later :) [15:16:46] Kk. Thanks! :-) [15:16:52] Contemporary web design. :P [15:17:40] (removed) [15:17:45] Lua error in Module:Days_since_incidents/since at line 11: attempt to compare nil with number. [15:17:55] ^^ probaly it coulden compare a number [15:17:56] lol [15:17:59] going into a 1:1 shortly, will look later, but not high prio for me, obviously ;) [15:18:54] Ok [15:27:36] 10Diffusion, 10Phabricator (Upstream), 07Upstream: "Unhandled Exception" error in Diffusion when displaying a repository with only one empty commit - https://phabricator.wikimedia.org/T140236#2619473 (10Paladox) Was there a task filled upstream? [15:29:18] 10Diffusion, 10Phabricator (Upstream), 07Upstream: "Unhandled Exception" error in Diffusion when displaying a repository with only one empty commit - https://phabricator.wikimedia.org/T140236#2619478 (10Paladox) It errors out here https://github.com/phacility/phabricator/blob/121e68e3adae4cd21731b79c07ca8967... [15:37:29] 10Diffusion, 10Phabricator (Upstream), 07Upstream: "Unhandled Exception" error in Diffusion when displaying a repository with only one empty commit - https://phabricator.wikimedia.org/T140236#2619494 (10Aklapper) >>! In T140236#2619473, @Paladox wrote: > Was there a task filled upstream? See https://phabric... [15:38:08] 10Diffusion, 10Phabricator (Upstream), 07Upstream: "Unhandled Exception" error in Diffusion when displaying a repository with only one empty commit - https://phabricator.wikimedia.org/T140236#2619495 (10Paladox) Oh thanks. [15:47:06] 10Differential, 06Developer-Relations: Add a welcome bot to Differential for first time contributors - https://phabricator.wikimedia.org/T135186#2291145 (10Paladox) We could use herald to do this. [15:52:08] 10Phabricator, 06Project-Admins, 06Developer-Relations (Jul-Sep-2016), 07RfC: Abandon (or at least strongly simplify) project creation policy - https://phabricator.wikimedia.org/T139210#2619523 (10Paladox) I guess we could open the project creation part to all Wikimedia employees and have a whitelist for v... [16:04:30] 10Differential, 06Developer-Relations: Add a welcome bot to Differential for first time contributors - https://phabricator.wikimedia.org/T135186#2619598 (10Aklapper) @Paladox: How **exactly**? Please share exact steps in comments. I haven't seen any condition in the dropdown options that could target first tim... [16:06:05] 10Differential, 06Developer-Relations: Add a welcome bot to Differential for first time contributors - https://phabricator.wikimedia.org/T135186#2619602 (10Paladox) Well I'm not really sure, but I said you could I didn't say use it. But anyways it can be added to new patches but yeh I guess we carnt use it ju... [16:20:24] 10Differential, 06Developer-Relations: Add a welcome bot to Differential for first time contributors - https://phabricator.wikimedia.org/T135186#2619617 (10Aklapper) @Paladox: If you write that "we could user Herald to do this", you imply that it is technically possible. Hence I asked you **how exactly** it is... [16:20:48] 10Differential, 06Developer-Relations: Add a welcome bot to Differential for first time contributors - https://phabricator.wikimedia.org/T135186#2619618 (10Paladox) Ok sorry. [17:30:10] hiya! i'm thinking about moving a github repo into diffusion [17:30:14] never done this before, have some qs [17:31:12] Oh you can import it [17:31:29] its currently in my own github, needs to be moved to wikimedia scope [17:31:31] just create a task under the diffusion tag and a repo admin should be able to import it [17:31:43] are diffusion repos mirrored to github like gerrit ones are? [17:32:00] yup [17:32:03] Yep [17:32:04] awesome [17:33:09] 10Phabricator, 06Project-Admins, 10Wiki-Loves-Monuments: Migrate Wiki Loves Monuments projects to subprojects under one umbrella - https://phabricator.wikimedia.org/T138952#2619846 (10JeanFred) Can we move forward with that? Is it blocked on something? [17:37:37] 10Diffusion: Import Kasocki repo from github to diffusion - https://phabricator.wikimedia.org/T145097#2619853 (10Ottomata) [17:39:34] 10Differential, 10Diffusion, 06Repository-Admins: Import Kasocki repo from github to diffusion - https://phabricator.wikimedia.org/T145097#2619871 (10Paladox) [17:39:36] 10Differential, 10Diffusion, 06Repository-Admins: Import Kasocki repo from github to diffusion - https://phabricator.wikimedia.org/T145097#2619873 (10greg) Just to be clear (as names are hard): you want to use Differential for the code review system, yes? Not Gerrit? (it'll be fun!) [17:53:24] 10Differential, 10Diffusion, 06Repository-Admins: Import Kasocki repo from github to diffusion - https://phabricator.wikimedia.org/T145097#2619906 (10Ottomata) Yes indeed! I want to use Differential! [17:54:04] ottomata: :) :) [17:55:30] :) [17:55:39] new repo, new review system :) [17:55:56] never used differential so this will be a learning experience [17:57:21] ottomata you will need php, arcanist and git :) [18:00:32] i installed arcanist [18:00:36] have git [18:00:40] have php but what do I need that for?! [18:00:53] ottomata since arcanist needs php [18:00:54] libphutil you mean? [18:01:00] and yes you need that too [18:01:13] hm ok [18:01:21] $ which php [18:01:21] /usr/bin/php [18:01:23] okey doke [18:01:29] PHP 5.5.30 [18:01:45] Yep :) [18:01:46] thanks [18:02:21] ottomata you will also need to do in the git repo this arc install-certificate [18:02:32] in the git repo? [18:02:54] i did that as part of the arcanist install instructions [18:02:56] but not in my repo [18:03:13] You need to be in a repo that is a git one, and has the .arcconfig file there with https://phabricator.wikimedia.org/ in there [18:04:22] ottomata: Whatever might confuse you on the way, https://www.mediawiki.org/wiki/Phabricator/Differential potentially welcomes improvements. ;) [18:04:52] * andre__ trying that "gentle nudging" concept on guinea pigs ;) [18:05:04] +1, the number of migrations to Differential is still <10, so docs improvements will be natural :) [18:05:32] hm k [18:06:55] so there should be a section in Phabricator/Arcanist for setting up arcanist for a repo? [18:07:08] You will need [18:07:09] https://phabricator.wikimedia.org/diffusion/PHAB/browse/wmf%252Fstable/.arcconfig;dbeb7d006398e88d09c81b627dd7873f6a5252f1$2 [18:07:12] and that arc install-certificate should be run there, not just wherever? [18:07:16] and https://phabricator.wikimedia.org/diffusion/PHAB/browse/wmf%252Fstable/.arcconfig;dbeb7d006398e88d09c81b627dd7873f6a5252f1$3 [18:07:17] optionally [18:07:22] and yes [18:07:22] or, does that need to be run once, and then again later in a repo [18:07:24] i thinkso [18:07:39] It needs to be run where the .arcconfig file is set [18:07:44] greg-g: should I create a .arcconfig before you create the diffusion repo? [18:08:17] ottomata: either way, but it's easier to do so now probably [18:08:54] and commit to the github repo? [18:08:54] ok [18:10:57] yah [18:11:12] you can add it post migration as well, it's just cleaner to do it this way [18:11:30] ottomata https://github.com/ottomata/kasocki/pull/1 [18:12:17] nice, thanks paladox was in the process of writing the exact same thing :) [18:12:26] Oh, your welcome :) [18:12:55] greg-g you can create ^^ now :) [18:14:28] One thing though you need someone to approve commits through differential [18:15:11] i think we realy need to allow self accept7 [18:15:13] accept [18:15:14] I'm not going to do it right now, sorry ottomata, mostly because I haven't done it before and I'll let those with more experience do it (mukunda etc) [18:19:33] ... [18:20:51] `arc install-certificate` sets up your home directory, should only be needed once. [18:21:53] the .arcconfig just points to the phabricator url and repo callsign... [18:22:21] ottomata: the file paladox linked should be fine. so you need a repo created in diffusion? [18:24:28] Yep i think he does :) [18:24:48] twentyafterfour see https://phabricator.wikimedia.org/T145097 please :) [18:31:13] paladox: thanks [18:31:37] your welcome [18:35:24] twentyafterfour: yup! [18:35:24] thank you [18:35:33] want to move work on this over to phab/diffusion [18:59:54] twentyafterfour: ehhhhhh ?:D [19:04:40] 10Diffusion, 10Phabricator: Prevent imported commits from upstream phabricator to close random tasks - https://phabricator.wikimedia.org/T116528#2620433 (10Ciencia_Al_Poder) 05Resolved>03Open It happened again, associating commits to unrelated tasks and triggering a bunch of email notifications: rPHDEP81dc... [19:21:28] legoktm Hi, did you manage to do https://gerrit.wikimedia.org/r/#/c/302416/ please? [19:21:42] wrong place [19:21:48] anyways should be -core [19:39:45] qchris Hi, there is a strange Microsoft Edge bug https://bugs.chromium.org/p/gerrit/issues/detail?id=4526 [19:41:47] Ouch. [19:43:03] Yep [19:43:25] And strange part is microsoft edge is newer then internet explorer yet microsoft edge is hitting the bug and ie isent [19:45:52] ottomata: I'm on it [19:49:00] qchris strange they are applying display none [19:49:07] on microsoft edge [20:07:49] 10Differential, 10Diffusion, 06Repository-Admins: Import Kasocki repo from github to diffusion - https://phabricator.wikimedia.org/T145097#2619853 (10mmodell) {rWKSK} [20:44:53] paladox: I am a little unsure about how reviews are done [20:45:02] how are differential revisions merged? [20:45:11] only via arc land, or are there GUI ways? [20:45:12] ottomata ah, [20:45:13] arc land [20:45:37] Currently it is a little triky to do it through gui so for now we have to do arc land [20:45:49] but you wont be able to do arc land unless the diff is accepted [20:45:51] ok [20:46:04] how do you accept it? [20:46:04] By accepted i mean someone else. [20:46:44] ottomata in the box at the bottom of the differential diff there should be a box you select and you can select what to do [20:47:12] accept wont show for you because you are the onwner and we need to switch that on, but i can accept it for you if you want [20:47:55] ottomata did you upload the diff to differential, arc diff? [20:49:19] yes [20:49:22] https://phabricator.wikimedia.org/D340 [20:49:27] milimetric accepted it [20:49:32] now i'm trying to run arc land [20:49:39] so, i added the diffusion remote [20:49:41] as phab [20:49:42] to my git [20:49:44] so [20:49:46] i'm trying to do [20:49:50] arc land --remote phab D340 [20:49:51] D340: README update - https://phabricator.wikimedia.org/D340 [20:50:02] Yep [20:50:04] getting [20:50:05] Exception [20:50:05] Branch "D340" does not exist in the local working copy. [20:50:05] D340: README update - https://phabricator.wikimedia.org/D340 [20:50:07] Oh wait [20:50:21] ottomata see arc land D340 [20:50:21] D340: README update - https://phabricator.wikimedia.org/D340 [20:50:25] if tat works [20:50:26] same [20:50:28] tat = that [20:50:29] oh [20:50:33] i did that too, but it tells me its going to try to use origin [20:50:35] which is still github [20:50:46] and even if i say yes [20:50:51] it give me the branch missing thing [20:50:54] what do you have set in your .arcrc file which would be where ever arc install-certificate saved it [20:51:19] { [20:51:19] "config": { [20:51:19] "default": "https://phabricator.wikimedia.org/" [20:51:19] }, [20:51:19] "hosts": { [20:51:19] "https://phabricator.wikimedia.org/api/": { [20:51:19] "token": "XXX" [20:51:20] } [20:51:20] } [20:51:21] } [20:52:02] That looks correct [20:52:22] ottomata did you arc patch D340 [20:52:23] D340: README update - https://phabricator.wikimedia.org/D340 [20:52:27] and then [20:52:28] arc land master [20:52:40] that's why arc land needs a branch [20:52:46] Doing ^^ should work [20:52:48] ottomata ^^ [20:52:55] arc patch? [20:52:58] why does it need arc patch? [20:53:04] Yep, that should download the patch [20:53:06] Or [20:53:10] i thought that was if i was downloading someone elses patch [20:53:15] No [20:53:27] You can use that for your own, but try [20:53:30] arc land master [20:53:32] first [20:53:36] ottomata ^^ [20:54:19] 10Diffusion, 10Phabricator: Prevent imported commits from upstream phabricator to close random tasks - https://phabricator.wikimedia.org/T116528#2620934 (10mmodell) 05Open>03Resolved I reset all of the tasks' status' that were affected and fixed the autoclose setting on yet another repository. [20:54:22] Exception: This repository is read-only over SSH. [20:54:28] oh i should add the http url as the remote? [20:55:24] hm [20:55:24] fatal: unable to access 'https://phabricator.wikimedia.org/diffusion/WKSK/kasocki.git/': The requested URL returned error: 403 [20:55:32] Ah [20:55:38] twentyafterfour ^^ [20:55:50] Could you add him as able to upload and merge patches please? [20:56:45] ottomata you need to add your ssh key to [20:56:47] phabricator too [20:56:52] or you can try over http [20:56:57] and set your own password [20:57:32] and also you need to be added to the group that can upload to that repo [20:58:05] the repository is read-only [20:59:35] twentyafterfour could you set it so he can upload and merge patches please? [21:00:52] ottomata: what is the relationship to github? I imported the repo from github, therefore github is the authoritative repository. I can switch it off (stop replicating from github) and then you'll be able to push directly to diffusion [21:01:02] is that how you want it set up? [21:01:56] Yes i think thats how he wants it [21:02:18] oh, yeah twentyafterfour [21:02:23] we are migrating it into wmf domain [21:02:27] so either gerrit or diffusion [21:02:34] thought diffusion is the new way so [21:02:40] i want diffusion to be master [21:02:54] and for github.com/wikimedia/kasocki to be mirror [21:03:06] github://ottomata/kasocki will be deprecated and point to diffusion&wikimedia github [21:03:09] right, the default when you import a repo is phabricator treats it as read-only. I just have to remove the github url from the repo and it'll be read/write [21:03:26] ok cool, and we'll get a mirror of it in github://wikimedia/ [21:03:27] ? [21:03:36] github.com/wikimedia/kasocki [21:03:37] ? [21:03:39] You can [21:03:57] The repo needs setting up there though first, then we can mirror [21:04:00] that last part is possible but we'll need to create the repo in github and I don't have the access to do that [21:04:09] ostriches may? [21:05:12] i think i can do it [21:05:20] shoudl i just push kasocki as is there? [21:05:25] or have it be blank? [21:05:41] op [21:05:41] nope [21:05:42] i don't have perms [21:05:43] nm [21:06:08] hah, yall can tell me if this maybe isn't ready too...i can use gerrit :p [21:06:09] Oh [21:06:30] Yeh it is ready, just not many people using it yet [21:06:35] k [21:06:53] twentyafterfour arc land master [21:06:57] woops [21:06:59] wrong thing [21:07:00] https://phabricator.wikimedia.org/diffusion/WKSK/manage/uris/ [21:07:01] nevermind I do have access to github create repo [21:07:06] Oh [21:07:08] :) [21:07:22] ottomata: it's read/write in diffusion now [21:07:44] ottomata: welcome to the party :) [21:08:07] nope I was wrong I don't have permission to create a github repo [21:08:19] Oh [21:08:20] but everything is set on our side [21:08:25] :) [21:08:33] I can make a github repo [21:08:43] greg-g: I can has access to github.com/wikimedia/ ? [21:08:51] do I have access? [21:08:52] Krenair: cool, thank you [21:09:02] greg-g: I don't know who to ask [21:09:17] what's your github username twentyafterfour? [21:09:18] Krenair: github.com/wikimedia/kasocki [21:09:21] chad or Krenair or bd.808 are my go tos for that [21:09:22] ottomata You can also https://phab-01.wmflabs.org/ use to test differential and diffusion [21:09:25] Krenair: my username is 20after4 [21:10:01] done: https://github.com/wikimedia/kasocki [21:10:21] nice ok, so, should I push there? or do we need diffusion to do it? [21:10:28] DIffusion will d [21:10:30] do that [21:10:32] k [21:10:41] twentyafterfour has setup the mirroring for you [21:10:42] :) [21:10:48] twentyafterfour, I converted you from a member to an owner [21:11:06] you should be able to do it too now [21:11:16] ottomata try arc land master now? [21:11:38] ok, so just master, not the D340 thing [21:11:39] D340: README update - https://phabricator.wikimedia.org/D340 [21:11:42] and --remote phab [21:11:48] Nope [21:11:50] yep [21:11:51] just [21:11:54] arc land master [21:11:57] i should have the http URI, not the vcs ssh one [21:12:04] You can have either [21:12:08] ok [21:12:18] with http uri you will need to setup password in your settings on phabricator [21:12:18] RESET Local "master" landed into remote "origin/master", resetting local branch to remote state. [21:12:19] hmm [21:12:34] i think i need --remote phab [21:12:46] Nope [21:12:54] arc land master is the command [21:12:59] well [21:13:08] but my 'origin' is still github ottomata [21:13:12] phab is phabricator [21:13:15] Oh ah [21:13:29] phab https://phabricator.wikimedia.org/diffusion/WKSK/kasocki.git (push) [21:13:29] I guess you could delete the repo and re clone from phabricator [21:13:36] i can change the remotes if i need to [21:13:40] but let's try it iwth --remote first [21:13:40] Yep [21:13:50] hehe, arc land master [21:13:52] change the remotes, i have no idea if --remote will work [21:13:53] with my origin as github [21:13:57] just pushed to github ottomata [21:13:59] oh lol [21:14:11] but, with --remote phab [21:14:15] Yep [21:14:15] it asked me for username and password? [21:14:19] Yep [21:14:20] but then said [21:14:20] fatal: unable to access 'https://phabricator.wikimedia.org/diffusion/WKSK/kasocki.git/': The requested URL returned error: 403 [21:14:23] You need to set your password [21:14:26] in phabricator [21:14:30] it doesn't use key? [21:14:34] oh maybe i ned to use ssh url [21:14:39] You [21:14:46] you should be able to push over https, I think [21:14:48] also need to add your ssh key to phabricator for ssh url [21:14:55] Krenair yeh but you need a password [21:14:57] i did that [21:15:00] that you set in phabricator [21:15:03] i mean [21:15:06] http password [21:15:07] right, a separate http password [21:15:08] like in gerrit [21:15:18] yep, but this one is customisable [21:15:26] BUILDS PASSED Harbormaster builds for the active diff completed successfully. [21:15:26] PUSHING Pushing changes to "phab/master". [21:15:26] Exception: You do not have permission to push to this repository. [21:15:26] fatal: Could not read from remote repository. [21:15:26] Please make sure you have the correct access rights [21:15:26] and the repository exists. [21:15:27] Usage Exception: Push failed! Fix the error and run "arc land" again. [21:15:28] for example [21:15:28] https://phabricator.wikimedia.org/settings/user/Paladox/page/vcspassword/ [21:15:32] ottomata ^^ [21:15:40] i'd prefer to use ssh key [21:15:41] replace paladox with your username on phab [21:15:44] and using the ssh remote [21:15:45] ottomata ok [21:15:46] seems to have worked [21:15:51] but [21:15:54] i don't have push rights? [21:15:54] https://phabricator.wikimedia.org/settings/user/Paladox/page/ssh/ [21:16:03] Oh wait you doint have push rights either [21:16:05] yeah i added a key there [21:16:17] twentyafterfour ^^ could you add ottomata as a pusher please? [21:18:40] will i be able to manage repo access too? once i have more rights? [21:18:52] Nope you will need edit rights [21:18:53] like, if i want my teammates to be able to arc land? [21:19:17] and im not sure if releng will allow us to manage our own phab repos like we can in gerrit [21:19:24] :( [21:19:59] ottomata but you can ask releng such as greg-g, twentyafterfour if you want to add more team mates [21:20:23] But i think we should allow authors of the repos to access the who is allowed access part like we do in gerrit [21:20:24] :) [21:22:07] yeah [21:23:46] ottomata: do you have an acl group that represents who you want to be able to push? [21:24:06] eg, we have #acl*releng for releng people that we used on eg the scap repo [21:24:10] eg eg eg [21:25:23] ottomata: fixed the can-push to include ottomata and repository-admins, but yeah it would be better to use a group [21:25:33] +1 too [21:26:01] ottomata: I can give you control of the access [21:26:13] :) [21:26:18] thanks! [21:26:21] acl group where? [21:26:22] in gerrit? [21:26:28] ottomata: no, in phab [21:26:31] hm [21:26:40] https://phabricator.wikimedia.org/diffusion/WKSK/manage/policies/ [21:26:42] eg: https://phabricator.wikimedia.org/tag/acl*releng/ [21:26:43] * ottomata gets confused with all the different types of phab group sand projects [21:28:10] twentyafterfour: we have an analytics tag [21:28:14] but i think it is not a team tag [21:28:19] sorry [21:28:20] greg-g: * [21:28:22] and not restricted membership, either, probably [21:28:39] https://phabricator.wikimedia.org/project/members/11/ [21:28:46] ottomata you should probaly include a group that has restricted access, ie no one can join without a member adding them [21:28:57] yeah that's too many folks [21:29:02] otherwise any user can join and do horrible things [21:29:04] can i just create a new group? [21:29:20] acl groups should be logged at least [21:29:22] lemme see [21:29:23] Well that is restricted to project admins [21:30:10] ACL projects are used to enforce policy restrictions, especially for Spaces. This type should be used instead of locking down normal projects such as Teams, so that anyone may still join and watch such Team projects without that giving them restricted access. Creation of such a project or a related Space must be proposed and discussed before being created. See T90491. Icon+Color: Policy+Red. [21:30:10] T90491: Create policy projects and convert people projects to open - https://phabricator.wikimedia.org/T90491 [21:30:24] from: https://www.mediawiki.org/wiki/Phabricator/Project_management#Types_of_Projects [21:31:03] ottomata try arc land master? [21:31:10] again, since you can now land [21:31:13] hopefully :) [21:31:27] DONE Landed changes. [21:31:28] yeehaw [21:31:47] :) [21:31:48] so, greg-g i should request an acl*analytics policy creation? [21:32:01] Yep i thinkso [21:32:30] ottomata: probably, so you can manage (more easily) who has access to the repo, since you'll have control over the membership of that group [21:32:53] ottomata we also have https://phabricator.wikimedia.org/T139210 to allow everyone to create projects without requesting them be created :) [21:33:14] ottomata: since you are on the operations team I added acl*operations-team [21:33:53] greg-g: I set it so that ottomata can manage the policies on the repo [21:34:00] twentyafterfour: perfect [21:34:15] ottomata: it might be a good idea to make #acl*analytics [21:34:33] twentyafterfour: should rmsca have a specific person as well other than (or in addition to) #repo-admins? [21:34:42] (I was using that as my "known good example" to work from) [21:34:42] to make it easier to manage who can push (currently you can add individuals to the can-push policy without using a group-project) [21:35:11] great thanks [21:35:12] https://phabricator.wikimedia.org/T145135 [21:35:37] greg-g: no need to have a specific person named as long as the people who matter are in the groups that matter ;) [21:36:12] fair :) [21:39:16] greg-g: want to block a spammer? -- https://phabricator.wikimedia.org/T145136 [21:39:33] so hmmm [21:39:37] ok after I arc landed master [21:39:40] what happens to https://phabricator.wikimedia.org/D340 [21:40:03] You just click i think resolve or the button that sounds like it closes it [21:40:08] ottomata: it will get marked as merged in a bit [21:40:15] Or that ^^ [21:40:18] greg-g: did we ever decide about acl* prefixes? was it decided to just leave that bit off? [21:40:24] god I love seeing more differential [21:40:34] Close revision? [21:40:40] yes [21:40:44] that sounds like the one [21:40:50] ottomata: it should close the revision automatically [21:40:53] its compliated chasemp (but gerrit was too when I first joined WMF) [21:41:01] k [21:41:08] how does arc land master [21:41:11] know to close D340 [21:41:11] D340: README update - https://phabricator.wikimedia.org/D340 [21:41:12] ? [21:41:35] ottomata: phabricator sees the commit landed and it closes things on the server side [21:41:40] after parsing the commit that got pushed [21:41:45] twentyafterfour could you block https://phabricator.wikimedia.org/p/Haryy123wpr/ please [21:41:50] He or she is a spammer [21:41:52] hm [21:41:56] per bd808 [21:42:06] what does 'master' refer to in arc land master [21:42:10] the head of my local master branch? [21:42:21] what if i had more commits there then what was in the arc revision? [21:42:22] ottomata thats the branch [21:42:24] done [21:42:31] twentyafterfour, you thought we should drop the acl* prefix? [21:42:32] thanks [21:42:41] for example in mw it will be arc land REL1_27 [21:42:43] ottomata: `arc land` is good enough - no need for the master [21:42:45] and so on [21:42:49] ottomata %% [21:42:55] woops thats meant to be ^^ [21:43:06] Krenair: it was discussed I couldn't remember the outcome [21:43:06] twentyafterfour thanks [21:43:07] :) [21:43:38] the outcome I think was to keep status quo [21:43:39] ottomata: 'arc land' will land whatever branch you tell it, or HEAD if you don't specify [21:43:39] paladox: 'the branch'? where? [21:43:39] in my local working clone? [21:43:49] Yes [21:43:56] or the branch on the diff detail in differential? [21:44:02] Yes [21:44:06] branch in diffusion [21:44:10] or working copy [21:44:13] is the branch in the diff detail kinda like gerrit's topic branches? [21:44:14] ottomata: the default workflow assumes you work with feature branches and it'll land the whole branch by squashing it onto master [21:44:16] since you can create branches [21:44:23] ottomata: but it's configurable [21:44:28] ok great [21:44:32] that's gerrit workflow, i can figure that out [21:44:44] But you carnt create branches in diffusion gui [21:44:45] squashing it? [21:44:56] right, but if i had a local working branch [21:44:57] and you carnt land from gui either [21:44:59] and i did arc diff [21:45:01] ottomata: it does a squash merge if there is more than one commit [21:45:17] it would create a diffusion revision with my-custom-branch as the feature branch [21:45:21] interesting [21:45:27] so [21:45:33] multiple commits in my-custom-branch [21:45:34] then arc landed [21:45:36] or you can make it immutable if you want to always do a regular merge instead of squash [21:45:38] woudl be squashed as one into master [21:45:51] ottomata: right [21:46:00] but those multiple commits would always only live in my local clone, and also differential [21:46:05] but if someone just cloned repo from diffusion [21:46:05] unless you customize it in .arcconfig - there are several possible workflows [21:46:10] they souldn't have those individual commits [21:46:12] hm [21:46:19] ottomata: right [21:46:34] i like the grouping of work into one differential revision [21:46:34] someone can always use 'arc patch' to grab those commits from differential [21:46:38] aye [21:46:50] buut, i think it'd be better to not squash them onto master [21:46:57] so that's configurable? [21:46:59] in the .arconfig? [21:47:16] what's the advantage of squashing onto master? [21:47:20] just to make git log look real nice? [21:47:24] ottomata: https://secure.phabricator.com/book/phabricator/article/arcanist_new_project/ [21:47:36] see: history.immutable [21:48:03] ottomata: right, it squashes when you land, in order to have one commit with all the work instead of a bunch of messy intermediate commits [21:48:19] gerrit enforces one commit, arcanist just transforms it into one commit at the end [21:49:01] hmmm, i see [21:49:05] but history.immutable: true will disable the squash - it works more like github pull requests then and merges the feature branch [21:49:22] with immutable, a differential review that has multiple commits [21:49:27] will look pretty messy when merged, right? [21:49:31] unlike gerrit [21:49:35] where we do git commit --amend all the time [21:49:45] so there are multiple patchecs in gerrit, but only the final one is merged [21:49:47] https://secure.phabricator.com/book/phabricator/article/arcanist_new_project/#history-mutability [21:49:52] ja am readin gthat [21:50:04] ottomata: right [21:50:41] I've gotten used to doing it the gerrit way, so I always tend to git commit --amend and keep just one commit locally [21:50:48] and arc handles that just fine [21:50:55] but that's only before you do arc diff [21:50:56] right? [21:51:03] once it is in differential [21:51:04] you just amend the commit and run `arc diff` [21:51:06] can you still --amend? [21:51:07] oh [21:51:08] hm. [21:51:26] every time you run `arc diff` it updates differential with a new revision [21:51:28] arc doesn't care how you do git things usually, it only knows hash1->hash2 diff [21:51:28] it will just overwrite the patch in differential? [21:51:42] it will add the new one [21:51:48] hm [21:51:54] just like gerrit when you upload a new patchset [21:51:55] but the old one doesn't really exist anymore in my history [21:52:05] so what happens to the first one during the sqaush? [21:52:11] ottomata: right, it only exists in the differential history [21:52:19] (just like gerrit I think?) [21:52:21] when you amend it gets thrown away [21:52:22] not squashed [21:52:23] oh ok [21:52:36] what if there are multiple patches in the differential revision [21:52:38] but you could still look it up in the web interface if you wanted to see the history [21:52:41] and i ammend one of them? [21:52:45] or, the latest one of them? [21:52:59] multiple patches? [21:53:00] it just overrwites the latest [21:53:03] my advice IMO is not to do --amend [21:53:07] hehe [21:53:09] I hate that we are forced into that workflow here [21:53:12] it's dumb [21:53:14] yeah its pretty weird i guess [21:53:18] and that soles this problem among many others [21:53:22] solves even [21:53:23] sqaushing does help there [21:53:26] you can always [21:53:30] do git add -A --all [21:53:31] git commit [21:53:32] so, yeah, multiple patches in the differential revision before it is landed [21:53:33] arc diff [21:53:38] then if you want to amend [21:53:46] multiple commits* [21:53:46] git add -A --all [21:53:49] git commit -a --amend [21:53:52] arc diff [21:53:55] and it updates the diff [21:53:55] ottomata: differential is just a series of `git diff`'s and eventually you merge the most recent one (the one that gets accepted) [21:54:06] oh [21:54:09] so [21:54:14] but what actually merges is the code you have locally [21:54:17] if i do git diff 3 times [21:54:20] Sha A B C [21:54:23] whatever code is present when you run arc land [21:54:28] what will show in differntial is not 3 commits [21:54:32] right [21:54:34] but just the total diff of those 3 together [21:54:39] three revisions with 1 commit [21:54:41] each [21:54:44] so locally i'll have my feature branch dev history [21:54:51] and differential will show what would be squashed during land [21:55:08] every time you run arc diff it does `git diff` on the merge-base of master and HEAD [21:55:13] then uploads that whole thing [21:55:27] interesting [21:55:28] ok [21:55:36] differential shows all the commits, whether it's just one or several [21:55:39] that makes more sense then [21:55:40] ok [21:55:43] and then landing it results in the squash [21:55:46] kinda like github PR there [21:55:47] ok [21:55:50] right [21:55:53] actually [21:55:59] upstream are trying to implement [21:56:00] that [21:56:00] it's a hybrid between github pull request and gerrit's workflow, really... [21:56:07] ha, yeah [21:56:08] but they say it will be 80+ hrs of work [21:56:10] differential shows like PR [21:56:14] but merge is like gerrit (sorta) [21:56:27] sounds fine :) [21:56:34] +except the merge happens client-side instead of server-side [21:56:37] so i get how --amend works then, because differnetial doesn't really care i gues [21:56:45] land is client side? [21:56:45] right [21:56:56] arc land is client side, then pushes to the server [21:56:58] squash locally, then push master? [21:57:10] yes [21:57:11] hm [21:57:12] k [21:57:22] there is a server-side merge feature but we haven't set that up yet [21:57:52] we have to implement harbormaster for that and it's not trivial [21:58:33] We are also [21:58:36] testing other things [21:58:41] such as the arcrd [21:59:08] twentyafterfour ottomata ^^ [21:59:24] also that reminds me twentyafterfour will you be trying out arcrd soon? [21:59:37] Since it needs alot of updating to work with the latest phabricator update [22:00:09] arcrd [22:00:19] paladox: you mean arcyd? [22:00:19] ? [22:00:26] Yes [22:00:34] that sorry i got the names mixed up [22:00:39] I don't think we will use that but will probably do something similar [22:00:46] Oh [22:00:55] Do you know any other ones [22:01:04] the one I haven't written yet [22:01:11] Oh [22:01:15] your building one [22:01:17] it's really not that complicated [22:01:21] oh [22:01:28] Is it in python [22:01:35] we could test it on phab-01 [22:01:41] the thing is we haven't decided for sure what we are doing yet [22:01:46] before deploying it on production [22:01:50] oh [22:01:50] so for now it's just speculation [22:01:56] what do you mean? [22:02:19] twentyafterfour ^^ [22:02:50] I'm not sure if arcyd is the way to go, but it does seem close to what we want. But we aren't ready for that yet [22:02:58] right now it's kinda paused [22:03:02] ok [22:03:35] I wonder can we get the inline editing and the gui merging before we move to differential please? [22:04:54] ok, thanks guys, looking good over here, i think we got everything we need except for acl*analytics, but that is not a hurry [22:09:30] ottomata: I'm gonna go ahead and set up your acl project, I can't think of any reason it should be prevented [22:10:56] ok cool., thanks! [22:12:51] ottomata: who else should be added to acl*analytics [22:14:11] joal, milimetric, nuria, mforns elukey [22:14:17] twentyafterfour one thing [22:14:34] though can we setup a new project for jenkins testing to use the new jessie instances [22:14:35] for now [22:14:41] until we find a way to support nodepool [22:14:51] it's to support https://phabricator.wikimedia.org/T145140 [22:15:05] ceremonial +1 on the task [22:15:46] 10Differential, 10Analytics, 10Continuous-Integration-Infrastructure, 10EventBus, 10Wikimedia-Stream: Run Kasocki tests in Jenkins via Differential commits - https://phabricator.wikimedia.org/T145140#2621158 (10Paladox) [22:16:53] 10Differential, 10Analytics, 10Continuous-Integration-Infrastructure, 10EventBus, 10Wikimedia-Stream: Run Kasocki tests in Jenkins via Differential commits - https://phabricator.wikimedia.org/T145140#2621160 (10Ottomata) mocha or maybe better, npm test [22:17:03] I don't know anything about the jessie instances [22:18:24] 10Differential, 10Analytics, 10Continuous-Integration-Infrastructure, 10EventBus, 10Wikimedia-Stream: Run Kasocki tests in Jenkins via Differential commits - https://phabricator.wikimedia.org/T145140#2621167 (10Paladox) Yep, we can do npm test, but just to let you know jenkins is currently broken in diff... [22:18:36] twentyafterfour those were setup by legoktm [22:18:46] due to the slowness of nodepool [22:18:49] a month ago [22:18:59] and recently hashar has migrated tests back to nodepool [22:19:51] twentyafterfour https://wikitech.wikimedia.org/wiki/Nova_Resource:Integration-slave-jessie-1003.integration.eqiad.wmflabs [22:19:53] :) [22:20:19] https://wikitech.wikimedia.org/wiki/Nova_Resource:Integration-slave-jessie-1002.integration.eqiad.wmflabs [22:23:32] paladox: I can't keep up with the nodepool changes [22:23:42] Yep [22:23:56] Those should be instance and not nodepool [22:24:00] the links i gave [22:40:24] 10Differential, 10Analytics, 10Continuous-Integration-Infrastructure, 10EventBus, 10Wikimedia-Stream: Run Kasocki tests in Jenkins via Differential commits - https://phabricator.wikimedia.org/T145140#2621287 (10mmodell) it should now be set up to run `npm test` [22:41:40] 10Differential, 10Analytics, 10Continuous-Integration-Infrastructure, 10EventBus, 10Wikimedia-Stream: Run Kasocki tests in Jenkins via Differential commits - https://phabricator.wikimedia.org/T145140#2621295 (10Paladox) Thanks @mmodell [22:45:56] 10Differential, 10Analytics, 10Continuous-Integration-Infrastructure, 10EventBus, 10Wikimedia-Stream: Run Kasocki tests in Jenkins via Differential commits - https://phabricator.wikimedia.org/T145140#2621308 (10mmodell) @ottomata: running `npm test` locally on a checkout of this repo throws a bunch of er... [22:51:10] 10Differential, 10Analytics, 10Continuous-Integration-Infrastructure, 10EventBus, 10Wikimedia-Stream: Run Kasocki tests in Jenkins via Differential commits - https://phabricator.wikimedia.org/T145140#2621117 (10mobrovac) How about using Travis for now since the repo is on GitHub? [23:03:12] 10Differential, 10Diffusion, 06Repository-Admins: Import Kasocki repo from github to diffusion - https://phabricator.wikimedia.org/T145097#2621426 (10Paladox) Can we close this as resolved now? [23:29:40] 10Differential, 10Analytics, 10Continuous-Integration-Infrastructure, 10EventBus, 10Wikimedia-Stream: Run Kasocki tests in Jenkins via Differential commits - https://phabricator.wikimedia.org/T145140#2621501 (10mmodell) @mobrovac I think that is the status quo for now. This task, however, is about differ... [23:30:05] 10Differential, 10Diffusion, 06Repository-Admins: Import Kasocki repo from github to diffusion - https://phabricator.wikimedia.org/T145097#2621503 (10mmodell) 05Open>03Resolved a:03mmodell [23:49:28] 10Differential, 10Analytics, 10Continuous-Integration-Infrastructure, 10EventBus, 10Wikimedia-Stream: Run Kasocki tests in Jenkins via Differential commits - https://phabricator.wikimedia.org/T145140#2621577 (10Ottomata) Yeah, npm test 'works' in travis now because `.travis.yml` sets up the KAFKA_HOME bi...