[00:00:39] * robla begins editing [00:15:16] I'm thinking of putting Collaboration at the end. We'll probably be pretty exhausted, and in some respects more ready for a meta conversation than at the beginning. [00:16:42] we'll also probably have (unfortunately) plenty of examples of poor collaboration behavior to reflect on. [00:17:22] I hereby issue the challenge to prove me wrong! :-) [00:34:21] done futzing with Monday for a little bit...now Tuesday [01:08:52] * robla got distracted talking to YuviPanda [01:22:35] robla: I was proposing to merge the "content format" section in the api ticket with "content format" in general [01:23:01] not all RFCs in both tickets [01:24:13] the "Content structure & APIs" section part in https://phabricator.wikimedia.org/T119029#1866851, with T105845, T99088, T106099, T114072, T111588, T114542 [02:42:53] legoktm: I fixed your legend sort order bug in grafana. [03:58:58] * robla takes a closer look at gwicke's comment [06:06:32] bd808: woohoo, thank you :DDDD [07:55:36] [[Tech]]; ArchiverBot; Bot: Archiving 1 thread (older than 30 days) to [[Tech/Archives/2015]].; https://meta.wikimedia.org/w/index.php?diff=15057504&oldid=15038866&rcid=7116025 [13:45:37] Nemo_bis: Need approval for https://codein.withgoogle.com/dashboard/tasks/4980248990973952/ I think. [14:00:23] Niharika, Nemo_bis: I've edited the task to actually include a description (let's avoid pasting only links as people are lazy to click) and published it. Thanks! [14:00:40] andre__: Okay, thanks. :) [18:44:21] if a page protection has expired, it shouldn't be in the API response, right? https://en.wikipedia.org/w/api.php?action=query&prop=info|flagged|revisions&inprop=protection&rvprop=user&rvlimit=1&titles=Alicia%20Vikander [18:44:32] guess there's some caching [19:19:33] I don't see it, I get "protection": [] [19:19:51] Expired protections are removed from the DB but not immediately [19:20:13] I think every 10th time we look at protections, we purge expired ones or something [19:20:29] In any case, the API should probably not show you expired protections even if they are still in the DB [19:20:34] That's worth a bug report [19:23:03] yeah it went away after about an hour or so [19:23:40] not a big deal, scripts/bots can just compare the timestamp with the current time [22:19:42] I'm hoping someone responds to my question about https://www.mediawiki.org/w/index.php?title=User:RobLa-WMF/Blog&diff=1960827&oldid=1960788 on #mediawiki-i18n .... otherwise, I'll probably bite the bullet and file a Phab ticket [22:20:07] * robla wants instant gratification! :-P [23:22:08] hello, something weird I'm noticing on enwiki that I think might be a technical issue. See https://en.wikipedia.org/wiki/User:Datapublish [23:22:27] at the bottom it clearly says it's in the Category:Candidates_for_speedy_deletion_as_spam [23:22:37] yet over here it is missing https://en.wikipedia.org/wiki/Category:Candidates_for_speedy_deletion_as_spam [23:23:15] the tag was added to that user page over a month ago [23:24:24] do hidden categories show up in the API? I see it is missing here as well https://en.wikipedia.org/w/api.php?action=query&titles=User:Datapublish&prop=categories [23:24:31] It doesn't clearly say it [23:25:13] no? it is a "hidden category". Those might only be visible for logged in users I think [23:26:22] There's been issues with tracking things like this [23:27:03] so this a known issue? categories not populating properly? [23:27:49] I see that other pages that are in the speedy category do show such categories in the API, e.g. https://en.wikipedia.org/w/api.php?action=query&titles=Amy%20Shira%20teitel&prop=categories [23:28:56] ah ha, if I remove the tag, save, and add it back it is now in the category [23:29:00] https://en.wikipedia.org/w/api.php?action=query&titles=User:Datapublish&prop=categories [23:29:25] maybe there was an issue with the jobs starting up when the tag was first placed? [23:36:32] search phabricator and I don't immediately see a report of this, I guess I'll try to file one [23:36:58] this is a big issue as there may be more serious things like attack pages or copyright infringement that don't end up getting deleted [23:37:13] https://phabricator.wikimedia.org/T117332 [23:40:29] Reedy: are we sure that's the same issue? I can comment with our specific problem if it helps [23:40:59] Yes [23:41:00] aaron merged 1 task(s): T118670: Items not showing up in category they have been added to. [23:41:08] "links" is a vague term in MW [23:41:19] ok cool [23:41:20] thank you! [23:41:52] FYI, the fix is starting to go out [23:41:59] /should be out [23:42:03] it's in 8 [23:42:24] in 8? [23:42:33] 1.27.0-wmf.8 [23:46:20] do we know if we're going to be able to repopulate all of those lost links? [23:46:26] Yeah [23:46:30] Null edit everywhere :) [23:47:00] automated? [23:51:46] well I guess I'll monitor the phab for updates, many thanks! [23:51:54] There probably won't be many more [23:52:23] Certainly, feel free to null edit any affected pages [23:53:07] I only know of a handful, there could be a lot more? [23:53:24] There's probably a lot more [23:53:50] so the null edits are all on the humans? [23:54:11] It could be done with a bot [23:54:22] Or a bot doing API purges with forcelinkupdate [23:54:22] finding all effected pages it the problem [23:54:30] for the sake of speedy deletion I think it should [23:54:35] since they aren't in the links table by definition [23:54:43] yeah [23:54:45] :/ [23:54:51] oh boy [23:55:22] Shit happens, unfortunately [23:55:22] the usually wisdom here is that they will be edited again "soon" and that will fix things [23:55:47] I wonder if searching for the pages with the tags would yield useful results [23:55:48] that's optimistic, but I'll take it [23:55:48] very few things that happen on wiki are life or death matters luckly [23:56:12] hmm.. yeah it may be possible to grep for the tags [23:56:16] yeah, I was thinking I'd go through the contribs of our known new page patrollers and see what we find [23:56:29] Though, chances are most of the speedy deletions are gonna be new articles [23:56:35] You could use that to limit scope [23:56:43] or user pages [23:56:59] articles/pages/whatever [23:57:08] fortunately if they used Twinkle or Huggle and tagged as an attack page, it gets blanked [23:57:20] and any subsequent edit to restore the offensive material would put it back in the category [23:57:25] assuming the timing was right [23:57:45] err, I guess that subsequent edit could have removed the tag [23:57:54] If you find it's a really big issue that needs a technical nudge, I'm sure we can help out with something [23:58:16] I only worry about the attack pages, frankly, and copyright vios [23:58:27] a non-notable entry here and there is normal [23:58:42] copyvio are most likely tagged by bots [23:58:58] so like you say, you could just null edit any existing pages left in a bot that does it