[11:36:04] 6Project-Creators, 6Reading-Admin: Clarify purpose of the #Reading-Admin project - https://phabricator.wikimedia.org/T112111#1880465 (10dr0ptp4kt) Done. [13:58:22] 6Project-Creators: Create project/tag WebPageTest - https://phabricator.wikimedia.org/T120916#1880675 (10Aklapper) 5Open>3stalled [18:00:26] Erm, I just got a notif about an "Unknown object (task)" that was closed invalid. [18:00:45] Clearly I must've been subscribed at some point? [18:01:15] It's the one blocking https://phabricator.wikimedia.org/T119891, but I dunno what it is. [18:03:47] * ostriches goes spelunking [18:03:53] some task taht was deleted I guess [18:06:37] Deleted? [18:07:22] Ah found it [18:10:32] Ah, I think it was move to a space. [18:11:13] MariaDB [phabricator_maniphest]> select id,phid,status,title,viewPolicy,editPolicy,spacePHID from maniphest_task where phid in (select dst from edge where src = 'PHID-TASK-rntcje2gdj2ijqv3y3xj' or dst = 'PHID-TASK-rntcje2gdj2ijqv3y3xj'); [18:11:18] chasemp: Give that one a shot ^ [18:11:23] Interesting stuff. [18:11:51] that seems like a bug..? [18:12:30] Well getting a notif on being move to a space you can't access seems like a bug, yeah. [18:12:39] But I'm more confused about the space usage here. [18:12:48] I meant the left behind bad object ref [18:12:51] but yes [18:18:50] 6Phabricator, 6Release-Engineering-Team: "Discovery" project does not show up in the list of projects in "Projects" field - https://phabricator.wikimedia.org/T121551#1881514 (10Smalyshev) 3NEW [18:23:37] ostriches, did you get a notification on a task that had a blocker (in another space which you can't view) closed as invalid perhaps? [18:23:56] Yep, that's what we're talking about [18:24:00] if so, I thought this was expected [18:25:25] I surely didn't expect it :) [18:33:30] 6Phabricator: "Discovery" project does not show up in the list of projects in "Projects" field - https://phabricator.wikimedia.org/T121551#1881595 (10Aklapper) [18:34:57] 6Project-Creators: Create user-Luke081515 project - https://phabricator.wikimedia.org/T121362#1881605 (10Aklapper) [18:41:25] 6Phabricator: "Discovery" project does not show up in the list of projects in "Projects" field - https://phabricator.wikimedia.org/T121551#1881614 (10Aklapper) [18:41:43] 6Phabricator: "Discovery" project does not show up in the list of projects in "Projects" field - https://phabricator.wikimedia.org/T121551#1881514 (10Aklapper) See T76732 [20:55:49] 6Phabricator: Rename "Access Request" security dropdown option - https://phabricator.wikimedia.org/T101999#1882551 (10Aklapper) Would fix / kill {T100100} [20:56:10] 6Phabricator: Refreshing an "open tasks" query results into "unknown object" - https://phabricator.wikimedia.org/T109258#1882554 (10Aklapper) 5stalled>3Invalid a:3Aklapper >>! In T109258#1841562, @mmodell wrote: > Propose to close this as invalid since it appears to be caused by `noscript`/`some other brow... [21:46:15] i was just thinking that, if we get the "create tickets via mail but still stay private" thing fixed, doesn't that mean we could technically replace the entire OTRS with phab? [21:46:32] ssssshhhh! [21:46:51] "Would it scale?" - "No." Ah well. :) [21:47:16] not atm as they have crazy custom stuff we can't / won't replicate [21:47:36] there is an app on the horizon for phab that is very anonymous call in like for support [21:47:41] but it's a ways off [21:48:00] i am saying that because.. on en.wp we tell people to report bugs ... [21:48:03] in OTRS [21:48:06] so we have random tech tickets there [21:48:17] that a completely different set of people look at [21:48:19] if any [21:48:22] who is the we there [21:48:29] en.wp general community? [21:48:58] my understanding of otrs use case is for like mobile apps calling in crash reports etc [21:49:01] there are quite some queues in there. not only "en". [21:49:07] the editors of https://en.wikipedia.org/w/index.php?title=Wikipedia:Contact_us&action=history [21:49:14] that's a very prominent page [21:49:26] and it sends people to mail stuff to info-en [21:49:37] then if somebody copies tickets from there to phab [21:49:44] they complain because it's public [21:50:06] but there was no reason to keep it private in the first place, like asking for db corruption and regular bugs [21:50:51] at the same time we have https://www.mediawiki.org/wiki/Phabricator/Help#Using_e-mail which mentions task@phabricator [21:50:52] we have such complaints? [21:51:02] I don't know why they chose otrs over bugzilla to begin with I guess [21:51:04] that made me think "why cant we tell them to mail phab" [21:51:24] that made me think "somebody will say we need the private option" etc [21:51:36] and that is like the hardware maintenace queue from RT [21:51:42] andre__: yes, Reedy just got one [21:51:42] "mailing in" was horrible in BZ and BZ isn't a support blah thingy [21:51:54] andre__: if Reedy doesnt respond i think there arent many others who look [21:52:05] oh really bz wasn't considered a valid support channel? [21:52:22] i dont know where you draw the line between reporting bugs and "support" [21:52:37] users will mix this anyways [21:52:44] I was mentally drawing the line at like automated bug reporting and crash stuffs [21:52:49] which was all I knew otrs to be used for [21:53:06] That is definitely not the primary use of OTRS [21:53:06] this smacks of historical things for which I don't know the reasoning [21:53:17] what is? [21:53:28] that stuff does/did go to OTRS, but it's an exception to the norm [21:53:29] (that's the only use case I've seen proposed to be ported) [21:53:53] OTRS is basically group email inboxes [21:54:07] OTRS is a random support desk for lots of things, email-based. [21:54:11] respond from the group address, etc. [21:54:13] ticket based [21:54:23] yeah [21:54:24] "Automated bug reports" already feels like an abuse of OTRS to me, srsly [21:54:51] I don't know if those android app crash reports still go there [21:55:02] so it's an email client that threads an issue by some report id that people can colloborativly respond to issues from [21:55:02] (automated stuff should be dealt automatically with, not by poor humans. In FOSS, Mozilla and RedHat have systems for exactly that.) [21:55:19] usually only one person for each ticket, but essentially yes chasemp [21:55:24] oh yes that's why I've maintained crash reports should not go to phab [21:55:45] that was in response to andre__ [21:56:18] I mean, other than phab is probably far, far more oriented to technical planning and machinations than end user support for which it is not at all suited [21:56:29] especially anonymous users which it doesn't support well at all [21:56:30] if at all [21:56:35] yeah that's similar [21:56:41] often somebody will have a question and they don't know yet if it's support or a bug, that's a typical pattern "am i doing it wrong or is it a bug" [21:56:56] so that distinction might not be helpful [21:57:04] [21:56:51] Reedy, I remember a change in gerrit where this was changed, that the crash reports will go to a third party provider, but I'm not sure :([21:56:51] Reedy, I remember a change in gerrit where this was changed, that the crash reports will go to a third party provider, but I'm not sure :( [21:57:09] having to send them to a different tool in between etc [21:57:29] Phab isn't made for automated crash reports. That's all. (GNOME had some automated crash reporter tool that created bugs in Bugzilla. It was mostly useless as it wasn't good in identifying duplicates and wasted a lot of human triager time.) [21:57:39] right at some point we were goign to send ccrash reports to some company who got was bought by a big fish, MS I think [21:57:43] I have no idea if that happened [21:57:56] andre__: releng/tgr was looking at that with sentry [21:57:57] I was pretty shock and awed by the idea considering the lengths we go to in other arenas [21:58:35] different tool in between: yes. needed. like http://crash-stats.mozilla.com/ or https://retrace.fedoraproject.org/ [21:59:00] mutante: idk man, it's an interesting question [21:59:13] what's the big difference between phabricator and bugzilla when it comes to automatic tickets [21:59:25] bugzilla had none either afaik [21:59:37] neither is setup to handle it [21:59:41] ok [22:00:25] this is just about the number of tickets, andre? [22:00:45] duplicates etc [22:01:32] mutante, on the server-side, basically yes. but then gathering useful info, like which versions are affected, and graphically displaying that is the very obvious next request. [22:01:50] not to mention phabricator cannot handle anonymous users [22:02:03] and OTRS does that thing? [22:02:06] yeah [22:02:10] ah [22:02:11] (interesting but useless knowledge: Mozilla throws away 90% of their crash reports immediately because it's already significant enough to analyze just 10%) [22:02:17] it's about the only thing it does do :P [22:03:14] so if i mail OTRS it hides my email address from others? [22:03:22] re: anonymous users [22:03:45] fwiw that's not at all what I meant by phab isn't prepared for anon users [22:04:13] mutante: Nope [22:04:26] so it's about automatic account creation on mail? [22:04:42] "account creation" [22:04:46] because you can only use task@phab if you already exist? [22:04:47] mutante, no. I do see the email as "CustomerID" in OTRS [22:04:53] you can't login [22:05:32] Reedy: https://www.mediawiki.org/wiki/Phabricator/Help#Using_e-mail does not mention i have to use the web ui first [22:05:46] I wonder if it works :) [22:05:47] Reedy: you can't log in to what? [22:05:51] OTRS [22:05:56] Reedy, sure you can [22:05:57] after you've just sent an email to it [22:06:02] Ah. [22:06:04] Yeah. [22:06:14] Sending an email to it does not create an account. That's true. [22:06:22] that's the "automatic account creation", like RT [22:06:26] gotcha [22:06:36] mutante: well, except with RT, you can get a password so you can login etc [22:07:07] rt had both types of user objects, pure email and registered [22:07:27] phab fails at it more because it's hugely normalized and relational [22:07:34] so you make a comment, add a project, make a task etc [22:07:40] it's all linked back to a user linked back to other objects [22:07:54] when you add anonymous users the wheels fall off [22:08:00] they tried a brief hack for it [22:08:09] and it has so many edge cases as to be useless [22:08:16] phab is literally not made to handle anonymous users [22:08:46] pretty much everything you do in phab counts on one tied in user account that is persisent and email is just one of the secondary mechanisms for notification [22:08:50] is it about users being anonymous? [22:08:57] i thought it's just about creating new users [22:09:15] that would be like any other user after they got created [22:09:23] only that they were created because somebody sent a mail [22:09:33] that mechanism does not exist [22:09:36] so it's a moot point [22:09:49] that's what the other systems did [22:10:00] yes but why does that make any difference? [22:10:08] the feature was not that people are anonymous [22:10:16] phabricator barely handles anonymous reading access [22:10:31] so i'm confused why we are talking about anonymous [22:10:33] Does https://secure.phabricator.com/w/projects/nuance/ solve that problem? [22:10:43] https://phabricator.wikimedia.org/T107014 [22:10:43] that's the idea Reedy [22:10:50] but it's super super early [22:11:04] So it's really, a "probably, just not now"? [22:11:08] well [22:11:09] we keep finding silly things that are broken for anonymous users because of a bunch of things that assume it to be impossible [22:11:18] nuance is meant for their own purposesl ike many of their apps [22:11:25] they are running a "we host your phab" business [22:11:33] and they have their own support view point [22:11:36] 'clearly the least restrictive policy is to restrict it to all registered users' etc. [22:11:40] nuance seeks to fill that zenddesk type space [22:11:49] we are comparing apples and oranges if we talk about "anonymous" and then compare it to OTRS and RT [22:11:57] that's just a different thing [22:12:04] maybe we want that too, not sure [22:14:43] they have the same problem to solve with database relations and they just create new users that are users like the others [22:16:22] yes but the crux the conversation is phab does not do that [22:17:34] i disagree that this is because phab is relational. the others are relational as well. [22:18:05] that's fine your misunderstanding what I was saying, they tried to have some version of what you are suggesting [22:18:16] and it relied on not honoring many of the relatinoal aspects because [22:18:19] emails are no user id's [22:18:31] so they have to decide is bob@yahoo.com bob or are emails now user id's [22:18:35] and they went a quasi route [22:18:38] and it ended up as a mess [22:18:57] so like I said, the way phab is setup does not lend itself to auto user creation [22:19:17] ok, earlier i asked if it's "about auto user creation" [22:19:19] for that reason [22:19:45] sure and I thought I answered but I don't kwno what details you are after [22:19:51] phab does not do this [22:21:26] the nuance bit I think goes down teh road of emails are user id's if we don't already know you [22:21:42] my impression is they are realizing it's a lot of work [22:22:39] ok, sounds like we can't do anything about it. i just wanted to bring up how random tickets are in OTRS [22:22:50] and how that might not be optimal [22:23:13] I think some people and places use phab for versions of on-wiki support [22:23:20] thanks for explaining. "anonymous users" was just still something different to me [22:23:27] but the ideological split between otrs and phab I don't understand [22:23:28] than "a user gets created for me when i mail [22:28:08] have a good day. (https://en.wikipedia.org/wiki/Serenity_Prayer)