[08:41:42] 10Phabricator: Delete @MarcoAurelio-Bot - https://phabricator.wikimedia.org/T154651#2918892 (10MarcoAurelio) [08:42:55] 10Phabricator: Delete @MarcoAurelio-Bot - https://phabricator.wikimedia.org/T154651#2918907 (10MarcoAurelio) [09:13:35] Hello, I was wondering about custom fields in maniphest (vs projects specifically) [09:17:30] I was specifically wondering how much custom fields can behave like projects and what the trade offs are in anyone has experience with this [09:18:21] In the old interface, there were fields for things like component, os etc in the prototype I just converted these to projects but people say they liked the prompt about which fields to add to tickets [09:41:50] mpickering: I'm confused. Where did you convert them to projects? [09:42:13] "the old interface" = "Bugzilla" in the context of Wikimedia. [09:42:25] Ah yes sorry, I am asking for advice for another project [09:42:29] we were using trac [09:42:42] mpickering, https://secure.phabricator.com/book/phabricator/article/custom_fields/ and #phabricator [09:42:52] ...as this is downstream :) [09:42:56] I know.. but I want actual user experience [09:43:07] If it is not welcomed then I will leave and never bother you again [09:43:31] but you are the people with the greatest experience of actually doing a migration [09:43:35] ...and https://secure.phabricator.com/book/phabricator/article/forms/ for having custom forms when people enter stuff. [09:43:58] (e.g. setting certain fields to required.) [09:44:12] that's all I can offer from the top of my head :) [10:25:24] 10Phabricator: Delete @MarcoAurelio-Bot - https://phabricator.wikimedia.org/T154651#2919033 (10Aklapper) 05Open>03Resolved a:03Aklapper (Followup to T153191) ``` Destroying objects... Destroying PhabricatorUser @MarcoAurelio-Bot... Permanently destroyed 1 object. ``` [10:53:02] 06Project-Admins: Archive #WikiCon-Francophone-2016 after sorting out its open tasks - https://phabricator.wikimedia.org/T153084#2919110 (10Pyb) >>! In T153084#2918007, @Aklapper wrote: > @Pyb: Could you please provide feedback within the next two or three weeks? Because otherwise I'll just mass-close the tasks... [11:00:08] 06Project-Admins: Archive #WikiCon-Francophone-2016 after sorting out its open tasks - https://phabricator.wikimedia.org/T153084#2919148 (10Aklapper) Heh, thanks! Looks like only https://phabricator.wikimedia.org/T139172 is left now being open. :) I'd rather create a new project for 2017 though if Phabricator is... [12:10:50] vandal in phab: https://phabricator.wikimedia.org/T154656 [13:14:33] 10Differential, 05Gerrit-Migration: Consider documenting the Arcanist installer for Windows and Mac that Paladox created - https://phabricator.wikimedia.org/T131939#2919398 (10Paladox) @greg-g Windows users can do the same except from they need to add it to the path by going through the control panel or export... [15:36:03] 10Phabricator: Task description diffs sometimes don't show all changes - https://phabricator.wikimedia.org/T154667#2919587 (10Jay8g) [17:09:22] mpickering: My own experience with phab is that it is opinionated, and that its orientation is toward minimalism. For us, that has resulted in a fair amount of compromise, where we adjust our ways of working to accommodate how phab thinks. That's neither good nor bad (or is both good and bad), but can be uncomfortable. [18:45:06] meeple27: Are you specifically referring to anything? we already do all code review on phab but still use trac for tickets [19:57:02] Technically I don't really see "minimalism" and I wonder who "us" is. [19:58:29] Upstream might be reluctant to accept stuff that creates maintenance costs. And I might be reluctant to add things downstream. If those two count as "minimalism". :) [20:22:21] mpickering: I was mostly referring to issue tracking in phab [20:24:42] andre__: examples of minimalism would be: Not supporting distinct "types" of issues, and treating blockers and subtasks as equivalent [20:25:12] devs use the argument of not wanting more maintenance costs, which I understand, but they seem to push that to a pretty extreme degree, even if it causes some pain for users [20:25:36] "us" is myself and perhaps a dozen heavy phab users in my close circle [20:26:10] what is the difference between a subtask and a blocker in your view meeple27? [20:26:18] genuinely curious [20:26:25] I don't think I distingiush [20:26:27] My saying is: All issue trackers suck. Phab just sucks in different ways from all the rest. It's no more frustrating to me than the other systems I have used [20:26:59] chasemp: the biggest distinction comes with estimation. You really want to estimate all the subtasks, but not the umbrella. But you want to estimate all the blockers/blockeds [20:27:24] but...isn't something a blocker by definition if it's a subtask? [20:27:28] that's more what I mean [20:27:34] how could you ahve a subtask that wasn't a blocker [20:27:43] or so I'm seeing it [20:29:35] it's how you think about it. If I have a task that is real work, and then something that has to happen before I can do the work, that's a blocker [20:30:02] in our mind it's hiearachical and all subtasks are blockers but not all blockers are subtasks [20:30:03] is taht right? [20:30:07] in your mind even [20:30:11] If I have some big amorphous bag of stuff I want to do, that's a supertask (which itself doesn't have any actual work) and then I'll create subtasks for the actual work [20:30:46] mathematically subtasks are blockers, but many of us never think of them that way, so the term is confusing [20:31:06] * meeple27 heads to a meeting [20:31:24] * chasemp waves [20:31:40] I disagree that the problem there is the system since yes subtasks are blockers [20:31:48] and any model where subtasks are not blockers is just...wrong [20:31:52] then they are not subtasks [20:31:57] but I think I kind of grok [20:35:31] You could set up a custom field for "types" in an installation. Not adding random fields in upstream is sanity, not minimalism. [20:36:07] There's a priority field by default, but no "urgency" and "severity" and "risk level" fields for example (yes, we've discussed that once) [20:36:56] I see "whiteboard entries", "keywords" and "tags" by default in the Bugzilla UI when it comes to upstream feature creep. So I prefer minimalism by default and being very expandable and flexible when needed, and I'd say that about Phab. [20:38:51] (I'm actually still a bit surprised that upstream added a "Story Points" field.) [20:38:58] they gave in :D [20:39:05] sellouts :) [20:39:52] The art of product management is to say no. Phab maintainers have impressed me so far. :) [20:57:25] Custom fields are not very good though compared to projects? [20:57:39] I suppose you can set herald rules which do similar things [20:58:35] they do two separate things and then custom forms kind of wrap that up [20:59:01] chasemp: just because subtasks are internally blockers doesn't mean the ui has to call them that. But at some point they switched, so now we don't have blockers---we only have subtasks [20:59:46] my thinking is more like, the nature of a thing that is a subtask is to be a blocker [20:59:51] vernacular is secondary [21:00:05] so when you have two tasks which conceptually aren't sub/super of each other, but one is a blocker, you have to mentally remember whether subtask is blocker or blocked, which is unnatural (at least for many of us) [21:00:05] if subtasks are part of a larger whole then teh whole isn't whole until the subtasks are closed [21:00:18] fair, that's where we trip up I think [21:00:29] I don't think things can be hiearachical sub/super w/o being blockers :) [21:00:42] that's just nonsensical to me and I have a hard time working through it [21:01:00] teh nature of being a sub portion of a larger thing is to be a blocking portion [21:01:01] honestly I have never thought of subtasks as being blockers of the supertask/story/epic, even though mathematically they are [21:01:26] i guess blocker to me has always had a very specific, sequential definition [21:01:39] "do a before b", whereas subtasks are "do a as part of doing b" [21:01:45] I think linear and gant chart thinking is not aligned her at all [21:01:51] so that would make sense why it's confusing [21:02:34] I meant not aligned as in not on upstreams mind, it makes sense if that's your perspective [21:03:14] gantt charts were the bane of my existence for a few semesters of college [21:03:29] That's when I decided I would never be a manager! [21:05:55] fortunately most managers don't actually have to do gantt charts [21:06:49] meeple27: I think your perspective makes some sense but sans a gant chart or otherwise real PM-y presentation the subtask as part of a whole is I thik the canonical form or reference [21:07:03] thanks for relating that actually, it's curious [21:07:39] * greg-g appreciates as well [21:07:50] meeple27: Luckily, I've had to do exactly zero of them since graduating :) [21:08:57] " the subtask as part of a whole is I thik the canonical form or reference " I find it fascinating to try to wedge sequential tasks into that framework ;) [21:09:34] you can make a series of subtasks to do it meeple27 [21:09:41] that basically becomes untenable a few layers in tho [21:09:51] I'm not entirely sure if that's intended [21:10:27] yeah, and it breaks the "supertask is the big thing that we're breaking down; subtasks are little bits of work" that I'm familiar with [21:11:33] if you could slice and dice the supertask to allow ordering subtasks where even if the ordering was convention and not enforcment [21:11:45] you could generate a gant chart if you had story points or time estimates [21:11:52] but man it gets deep quick :) [21:12:23] I think I may be convinced now ordered subtask presentation is valuable [21:12:25] jira (as an example) had super/subtasks, and then also had blockers/blocked-by [21:12:36] and also related [21:12:38] iirc [21:12:49] i think they were distinct mechanisms [21:14:37] it's been years but yeah pretty sure [21:14:50] I once lobbied for a related in phab [21:15:07] and I think they may have given in and treated "mentioned" as more first class to appease [21:17:20] Phab allows easy renaming of items in the UI. We could always call a subtask a blocker or vice versa or or. [21:17:57] I think the actual treatment of such things is in question here and not necessarily label [21:18:03] that being said no thanks :D [21:18:11] +1 :P [21:18:31] andre__: you are going to be in SF? [21:18:39] Yepp [21:18:58] I expect a beer on the rooftop with you. [21:19:26] ok plan [21:24:32] ok thanks for the small discussion guys [21:30:30] andre__: Renaming wouldn't help. It was named one way, which was confusing in some cases. Now it's named the other way, which is confusing in the other cases. The problem is fitting 2 concepts into a single mechanism...it works mechnically, but not conceptually [22:43:59] I tried setting up MediaWiki OAuth in my Phabricator instance today, however, I get an error (https://pastebin.com/NiAqUHyy) in Phabricator after allowing it access to MediaWiki.