[07:19:43] 3Phabricator: Search on phabricator is utterly broken and unusable - https://phabricator.wikimedia.org/T88514#1013813 (10Joe) 3NEW [08:13:45] 3Code-Review, Engineering-Community: How to prioritize code review of patches submitted by volunteers - https://phabricator.wikimedia.org/T78768#1013878 (10Qgil) [08:25:24] 3Code-Review, Engineering-Community: How to prioritize code review of patches submitted by volunteers - https://phabricator.wikimedia.org/T78768#1013886 (10Qgil) What about organizing a Gerrit Cleanup Day? * A full day of code review. * Focusing on oldest changesets first ** waiting for a first review ** +1 wai... [09:19:33] 3Code-Review, Engineering-Community: How to prioritize code review of patches submitted by volunteers - https://phabricator.wikimedia.org/T78768#1013963 (10jayvdb) +1 To prioritise changesets by volunteers, I suggest that we build a list (in advance) of patches ordered by people with the least number of patches... [09:27:23] 3Code-Review, Engineering-Community: How to prioritize code review of patches submitted by volunteers - https://phabricator.wikimedia.org/T78768#1013975 (10scfc) I don't think increasing the review work would have a positive impact on volunteers ("you only need to wait one quarter for a review!"). If WMF wants... [09:35:24] 3Code-Review, Engineering-Community: How to prioritize code review of patches submitted by volunteers - https://phabricator.wikimedia.org/T78768#1013989 (10Reedy) >>! In T78768#1013975, @scfc wrote: > I don't think increasing the review work would have a positive impact on volunteers ("you only need to wait one... [09:55:34] 3Code-Review, Engineering-Community: How to prioritize code review of patches submitted by volunteers - https://phabricator.wikimedia.org/T78768#1014013 (10scfc) Something unreviewable can get a -1 immediately (or WMF could just use their discretion to not pay people trying to game the system). Looking at the W... [10:39:09] 3Code-Review, Engineering-Community: How to prioritize code review of patches submitted by volunteers - https://phabricator.wikimedia.org/T78768#1014053 (10jayvdb) >>! In T78768#1013975, @scfc wrote: > I don't think increasing the review work would have a positive impact on volunteers ("you only need to wait one... [10:41:11] qgil: Where did this idea of paying people come from? ugh :/ [10:41:27] paying people? [10:42:43] where? [10:42:48] https://phabricator.wikimedia.org/T78768#1013975 [10:42:53] Not saying you suggested it :P [10:42:59] "If WMF wants to throw some money at it, I'd prefer that each volunteer gets 1 $/d/unreviewed patch out of the corresponding team's budget. Then you have an incentive for WMF and a consolation price for volunteers." [10:45:13] Reedy, I'm just as puzzled. I though scfc was joking, but he seems to be actually serious about it? [10:45:49] Reedy, anyway, I don't think the WMF wants to "throw money at" code review beyond a share from the WMF salaries [10:46:52] heh [10:48:12] Certainly triaging/abandoning old patches with "useful" comments will go a long way [10:48:29] -1 for more than a year but no sign of any activity etc [11:02:23] indeed [11:02:44] I will create an own task for the Gerrit Cleanup Day proposal [11:45:54] 3Code-Review, Engineering-Community: Organize a Gerrit Cleanup Day - https://phabricator.wikimedia.org/T88531#1014128 (10Qgil) 3NEW a:3Qgil [11:46:55] 3Code-Review, Engineering-Community: Organize a Gerrit Cleanup Day - https://phabricator.wikimedia.org/T88531#1014128 (10Qgil) [12:35:43] 3Phabricator: Default search should be limited to open tasks. - https://phabricator.wikimedia.org/T88533#1014179 (10daniel) 3NEW [13:11:43] 3Phabricator: Default search should be limited to open tasks. - https://phabricator.wikimedia.org/T88533#1014275 (10Qgil) See T76273#794511 [13:11:58] 3Phabricator.org: Default search should be limited to open tasks. - https://phabricator.wikimedia.org/T88533#1014277 (10Qgil) p:5Triage>3Volunteer? [14:43:19] 3Phabricator: Enabling Herald - https://phabricator.wikimedia.org/T630#1014570 (10Qgil) I didn't think that we would be without Herald still in February... What about a next step enabling Herald for a #Herald-Users group? wmf-nda and individually approved guests? [15:34:54] 3Phabricator: Enabling Herald - https://phabricator.wikimedia.org/T630#1014897 (10Negative24) I would like to help test Herald. I've taken a look at some of its code before. Was T493 resolved with a Phabricator release? The latest update (T86772) doesn't mention it and I don't have any contacts to ask about it. [15:35:17] at $previous_job we had a cron that ran and calculated a kind of activity points based on review activity and diff submissions etc [15:35:33] and then weekly in our all hands call (there) it was put up on the board [15:35:53] it was a decent example of, encourage the good is more important than punishing the bad [15:36:09] now someone could be crazy prolific in reviewing and who would know? [15:36:30] not saying rewards are necessary but it's a mostly thankless and pretty invisible work load [15:37:35] I think twentyafterfour wrote it actually [15:40:55] no that was after I left [15:41:37] well nvmd then :) maybe it was forno [15:42:31] actually I think I kinda remember it being developed but not actually in use... wasn't it sasha's baby? [15:43:45] he definitely was in love with it and read it off every week [15:43:49] so probably [15:43:53] I don't remember exactly but I'm pretty sure it was 99% done in js and used conduit api but didn't actually involve modifications to phabricator other than maybe 1 tiny hack to expose data that was missing from phab api [15:44:05] yes basically [15:44:42] yeah I think sasha wrote it and it ran in client-side js but queried conduit [15:44:55] sounds cheesy but it was mostly a positive [15:45:08] in that prolific and dedicated reviewers were recognized [15:45:26] it was a two top 10 lists iirc [15:45:37] he kept the code secret so he could avoid people gaming the system. and his algorithm for assigning the points involved some dark magic [15:45:58] (they gave out dA points which were eventually sort of monetized but pretty small scale) [15:46:43] if I wrote it I would just go the community metrics route [15:46:45] <^d> We used to have stats about patches in SVN. The math was hard so I settled on using mt_rand($min, $max) [15:46:49] I remember it more clearly now ;) [15:46:50] and use db directly [15:46:58] <^d> It made for a fun guessing game as to how many patches you actually had to review :p [15:47:06] heh [15:47:42] I'm not a community genius here but it seems like to first step to encouraging good behavior is recognizing it [15:47:53] that's what I have seen work [15:47:59] another metric it considered was tasks filed, another was task comments submitted to other people's tasks [15:48:11] yes that was the non-review top 10 portion tho [15:48:17] it was like, review and general activity [15:49:33] yeah and a weekly most prolific reviewer, and most active bug fixer.. as I remember he actually attempted to figure out lines of code reviewed for each person [15:49:55] no one ever reverse engineered his formula I think :D [15:50:04] my guess at the time was he weighted code review comments [15:50:20] because the many-times-winner was just that a very prolific commenter in diff [15:50:22] (woody) [15:51:07] which kind of makes sense, otherwise the "sure go ahead" TSA style code review overshadows [15:51:32] quality was assumed basically I guess [15:51:40] otherwise you would have been fired [15:51:51] yeah and I think he counted the length of comment text, as well as inline comments being worth even more [15:52:00] agreed [15:54:20] you could look for inline comments followed by a new revision which implies quality and/or relevant comment [15:54:23] anyways [15:54:28] should float it to qgil [15:54:45] the reason that stuff worked so well at dA was because the "points" are actually kinda valuable but also everyone wanted to impress sasha, and it was just fun competition across teams [15:54:59] yeah I tend to think mainly it was fun [15:55:05] bragging points more than anything [15:55:23] and also, it wasn't like important really and truly but you noticed who was kicking ass [15:55:30] and appreciated them [15:55:53] I couldn't tell you a single person here who does a lot of code review [15:56:05] I remember the biggest competitive motivation was having the best demo, getting the teams to compete rather than the individuals only [15:56:21] yes I'm the one who actually introduced the vote for that :D [15:56:30] which sasha then took over and would have fun with [15:56:31] nice [15:57:03] honestly that would work here well but monthly [15:57:06] imo [15:57:30] show off what you did, and if what you did amounts to nothing ppl will think you are lame [15:58:07] <^d> https://www.mediawiki.org/wiki/User_talk:%5Edemon#Random_acts_of_CR_barnstar [15:58:19] <^d> :) [15:58:33] ^d: vacation time [15:58:42] relevant portion there ^d "that we know of! " :D [15:58:52] damn those anons doing CR [15:59:15] <^d> Reedy: I'm done CR. I'm going to take a break until everyone's reviewed as much as me :D [15:59:25] does anyone here have a good view of how much code review anyone else does? can I see all things I've +1'd? [15:59:47] <^d> You can coax it out of Gerrit's API and database. [15:59:57] <^d> But gerrit sucks, so it would never show you something so interesting or useful. [16:00:03] <^d> On its own :p [16:00:22] that is super lame to not be able to see things you've reviwed easily [16:01:21] <^d> Oh, you can see that. [16:01:28] <^d> https://gerrit.wikimedia.org/r/#/q/reviewer:self,n,z [16:01:55] that seems to show things I am a reviewer on [16:02:03] but not necessarily things I've reviwed [16:02:31] I think? [16:02:44] <^d> Yeah [16:03:13] <^d> You can search based on CR scores, and by who's on the review list, but not based on who gave what scores. [16:03:26] <^d> That requires DB access or a ton of API calls and some post-processing. [16:03:32] gotcha [16:03:44] the old no cc / reviewer distinction [16:03:55] <^d> Something like that [17:12:08] 3Phabricator, Wikimedia-Git-or-Gerrit: Task/Job 7092771 (PhabricatorRepositoryCommitHeraldWorker) has over 1600 failures (Importing revision) - https://phabricator.wikimedia.org/T87549#1015185 (10Se4598) new one for documentation: [[https://phabricator.wikimedia.org/daemon/task/7250335/|Task 7250335]] with 233 f... [17:58:41] 3Code-Review, Engineering-Community: How to prioritize code review of patches submitted by volunteers - https://phabricator.wikimedia.org/T78768#1015372 (10ggellerman) +1 tom prioritizing work not just by 'oldest' first [18:59:19] 3Phabricator.org, Phabricator: "Resolved" should be the first (default) option in "Change Status" action - https://phabricator.wikimedia.org/T548#1015536 (10Qgil) [19:00:54] 3Phabricator.org: Content field should have more emphasis in advanced search form - https://phabricator.wikimedia.org/T77708#1015550 (10Qgil) [19:05:21] 3Phabricator.org: Project icons missing in Firefox 34 on some tasklists in "Activity" dashboard - https://phabricator.wikimedia.org/T86523#1015562 (10Qgil) I see all the icons in Firefox now. The only problem is the separate icon (as shown in the screenshot of the description). I wonder whether Phabricator can... [19:07:33] 3Phabricator.org: Let users configure date format in account settings - https://phabricator.wikimedia.org/T110#1015572 (10Qgil) p:5Low>3Volunteer? [19:08:28] 3Phabricator.org: autocomplete fields don't work in Firefox 32 on Android - https://phabricator.wikimedia.org/T455#1015573 (10Qgil) p:5Low>3Volunteer? [19:13:07] 3Phabricator.org: Notification flyout lacks timestamps, hard to understand when actions happend - https://phabricator.wikimedia.org/T78198#1015607 (10Qgil) 5Open>3declined a:3Qgil >>! In T78198#841358, @Qgil wrote: > In my opinion, we should decline this task. [19:30:41] 3Wikibugs: Match on usage of Additional Hashtags, so that project renames don't break the bot - https://phabricator.wikimedia.org/T87825#1015676 (10Quiddity) >>! In T87825#1000112, @Legoktm wrote: > Yes, it would be appreciated if you could fix the config. > > We had previously discussed storing immutable PHID... [19:34:36] 3Wikibugs: wikibugs project renames - https://phabricator.wikimedia.org/T87846#1015682 (10valhallasw) 5Open>3Resolved a:3valhallasw Resolved by @legoktm in I151c1fb1fe2620ef665c308f089af17882c53544 [19:36:49] 3Phabricator.org: Shortcut keys for description fields text formatting are needed. - https://phabricator.wikimedia.org/T816#1015690 (10Qgil) [19:36:50] 3Phabricator.org: Phabricator does not have any (useful) keyboard shortcuts? - https://phabricator.wikimedia.org/T1053#1015689 (10Qgil) [19:37:32] 3Phabricator.org: Phabricator does not have any (useful) keyboard shortcuts? - https://phabricator.wikimedia.org/T1053#18305 (10Qgil) This task didn't specify any desired keyboard shortcut, so I merged it with {T816}. [19:39:58] 3Phabricator.org: Uninstalled applications are still listed - https://phabricator.wikimedia.org/T1381#1015698 (10Qgil) 5Open>3Resolved a:3Qgil This problem was resolved upstream a few weeks ago (Jan 6). I believe this change is already deployed here. https://phabricator.wikimedia.org/applications/ looks good. [19:48:28] 3Phabricator.org: User profile pages require login - https://phabricator.wikimedia.org/T639#1015719 (10Qgil) [19:50:19] 3Phabricator.org: User profile pages require login - https://phabricator.wikimedia.org/T639#10395 (10Qgil) Upstream keeps thinking that user profiles should not be publicly accessible, and the Wikimedia Phabricator team keeps thinking that it is not worth maintaining a local patch for this. I propose to decline... [19:58:36] 3Wikibugs: wikibugs project renames - https://phabricator.wikimedia.org/T87846#1015738 (10Quiddity) [20:00:51] 3Phabricator.org: Email notification for "edited the task description" does not contain the actual content changes (diff) - https://phabricator.wikimedia.org/T75851#1015747 (10Qgil) For what is worth, it is possible to configure the option for sending diffs of code patches via email: `metamta.diffusion.inline-pa... [20:10:43] 3Phabricator.org: Send a regular email with a saved search (like Bugzilla's "whining" for saved seaches) - https://phabricator.wikimedia.org/T76825#1015784 (10Qgil) Back to the use case: > I especially like to have regular emails containing a list of all tasks modified on a project in the last 24 hours, like th... [20:16:40] 3Phabricator: Task creation by email is unclear - https://phabricator.wikimedia.org/T865#1015791 (10Qgil) [20:19:23] 3Phabricator: Task creation by email is unclear - https://phabricator.wikimedia.org/T865#1015796 (10ori) IMO, we should simply drop the feature. Having a single, canonical workflow that can be concisely expressed seems vastly more desirable. See . [20:20:12] 3Code-Review: GitHub->Phabricator bridge for new contributors - https://phabricator.wikimedia.org/T173#1015797 (10Qgil) [20:20:15] 3Code-Review: Arcanist installer for Windows - https://phabricator.wikimedia.org/T172#1015801 (10Qgil) [20:20:16] 3Code-Review: Running `arc diff` twice on the same commit creates another redundant diff within the same differential - https://phabricator.wikimedia.org/T137#1015802 (10Qgil) [20:20:17] 3Code-Review: Editing files and contributing changes via web - https://phabricator.wikimedia.org/T409#1015799 (10Qgil) [20:20:19] 3Code-Review: Arcanist isn't packaged in Fedora, Debian, or Ubuntu - https://phabricator.wikimedia.org/T133#1015800 (10Qgil) [20:20:20] 3Code-Review: Landing a patch with arc currently will sometimes strip author information - https://phabricator.wikimedia.org/T612#1015804 (10Qgil) [20:20:21] 3Code-Review: Allow viewing diffs of single commits within a Differential - https://phabricator.wikimedia.org/T125#1015803 (10Qgil) [20:21:47] 3Phabricator.org: In the CC field, Phabricator should link to the list of project members and watchers - https://phabricator.wikimedia.org/T78604#1015805 (10Qgil) 5Open>3declined a:3Qgil [20:22:44] 3WMF-Design, Phabricator: Migration of the Design team to Phabricator - https://phabricator.wikimedia.org/T832#1015811 (10Qgil) [20:22:59] 3Phabricator.org: Phabricator should have two different notification levels - https://phabricator.wikimedia.org/T78357#1015813 (10Qgil) 5Open>3declined a:3Qgil [20:24:54] 3Phabricator.org: New task comment dropdown is confusing - https://phabricator.wikimedia.org/T238#1015830 (10Qgil) 5Open>3declined a:3Qgil >>! In T238#852784, @Qgil wrote: > Proposing to decline this task since we are not going to put own resources in improving this area. [20:25:57] 3Phabricator: Add IRC name and MediaWiki Username as alternate auto-complete lookups - https://phabricator.wikimedia.org/T877#1015833 (10Qgil) [20:56:53] 3Phabricator: Task creation by email is unclear - https://phabricator.wikimedia.org/T865#1015905 (10Qgil) As far as I know, task creation via email is part of a feature required by Operations, in relation to Procurement. Besides, I don't think we can drop the feature via configuration. Honestly, I don't see an... [21:31:33] 3Phabricator: Task creation by email is unclear - https://phabricator.wikimedia.org/T865#1015976 (10ori) >>! In T865#1015905, @Qgil wrote: > As far as I know, task creation via email is part of a feature required by Operations, in relation to Procurement. > > Besides, I don't think we can drop the feature via c... [21:52:23] 3Phabricator.org: Send a regular email with a saved search (like Bugzilla's "whining" for saved seaches) - https://phabricator.wikimedia.org/T76825#1016029 (10Nemo_bis) > I don't think sending an email is more efficient than this I do. > I'm quite certain that this will be a Wontfix if submitted upstream. Up... [21:54:03] 3Phabricator.org: User profile pages require login - https://phabricator.wikimedia.org/T639#1016031 (10Nemo_bis) > not worth maintaining a local patch for this How about a local customisation that unlinks the profiles, then? [21:56:12] 3Phabricator.org: User profile pages require login - https://phabricator.wikimedia.org/T639#1016038 (10Qgil) Another local patch to maintain, with no benefit that I can see. [23:04:34] 3Code-Review, Engineering-Community: Organize a Gerrit Cleanup Day - https://phabricator.wikimedia.org/T88531#1016290 (10brion) +1 [23:07:30] 3Code-Review, Engineering-Community: Organize a Gerrit Cleanup Day - https://phabricator.wikimedia.org/T88531#1016301 (10Reedy) Some basic rules should be devised here. Such as, if a patch is -1 for X time, with no response, or work from the author (or any interested parties), it should be abandoned, with a mes... [23:33:31] 3Legalpad, WMF-Legal, Phabricator: remove legalpad.wm.org - https://phabricator.wikimedia.org/T87688#1016366 (10Dzahn) @chase can i merge? [23:42:35] 3Legalpad, WMF-Legal, Phabricator: remove legalpad.wm.org - https://phabricator.wikimedia.org/T87688#1016394 (10chasemp) >>! In T87688#1016366, @Dzahn wrote: > @chasemp can i merge? I'm wondering if we make this box the phab failover, but cleaning out legalpad configuration is of no concern either way I suppose... [23:50:43] 3Code-Review, Engineering-Community: Organize a Gerrit Cleanup Day - https://phabricator.wikimedia.org/T88531#1016423 (10jayvdb) >>! In T88531#1016301, @Reedy wrote: > Some basic rules should be devised here. > > Such as, if a patch is -1 for X time, with no response, or work from the author (or any interested... [23:54:18] 3Code-Review, Engineering-Community: Organize a Gerrit Cleanup Day - https://phabricator.wikimedia.org/T88531#1016436 (10Dzahn) What jayvdb said, often volunteers uploading something, it doesn't get attention, then quite some time later an employee says he would like a little bit to be changed and -1's it, and t...