[00:09:53] http://blog.ajf.me/2015-02-06-php-predictions-for-2015 [00:19:40] self-fulfilling prophesies? [00:26:01] maybe [04:52:51] 13MediaWiki-Core-Team, 2MediaWiki-extensions-SecurePoll, and 1 other: Fix SecurePoll_BallotStatus for Status/StatusValue changes - https://phabricator.wikimedia.org/T89475#1040770 (10bd808) @deskana asked for this fix to be backported to the 1.25wmf17 production branch which was the first branch to include t... [04:57:47] 13MediaWiki-Core-Team, 2MediaWiki-extensions-SecurePoll, and 1 other: Fix SecurePoll_BallotStatus for Status/StatusValue changes - https://phabricator.wikimedia.org/T89475#1040771 (10bd808) >>! In T89475#1037244, @Aklapper wrote: > Is T89425 a dup / related? I merged this in as a dup because it sounded the... [10:21:26] 13MediaWiki-Core-Team, 2Librarization: Make MWException handle non-MW exceptions better - https://phabricator.wikimedia.org/T76652#1041042 (10Aklapper) @aaron: Please reply (and reset the task status to open). [10:29:41] 13MediaWiki-Core-Team, 2Code-Review: Import all gerrit.wikimedia.org repositories with Diffusion - https://phabricator.wikimedia.org/T616#1041057 (10Aklapper) [11:09:51] 13MediaWiki-Core-Team, 2MediaWiki-Page-editing, and 1 other: Long edit comments get entirely removed instead of truncated (error in cutting multibyte chars?) - https://phabricator.wikimedia.org/T85700#1041131 (10Aklapper) Not sure if this is still Core or already DB territory. Probably the former if MW does... [11:10:23] 13MediaWiki-Core-Team, 2MediaWiki-ResourceLoader, and 1 other: Fix "PHP Notice: Uninitialized string offset: 1 in includes/libs/jsminplus.php on line 1968" - https://phabricator.wikimedia.org/T77169#1041133 (10Aklapper) p:5Triage>3Volunteer? [15:12:36] 3Services, Datasets-General-or-Unknown, MediaWiki-Core-Team: Improve Wikimedia dumping infrustructure - https://phabricator.wikimedia.org/T88728#1041566 (10ArielGlenn) Here are a few stories I've run across: A company wants to mirror English Wikipedia with relatively up to the minute changes (or at least up to... [15:22:38] 3Services, Datasets-General-or-Unknown, MediaWiki-Core-Team: Improve Wikimedia dumping infrustructure - https://phabricator.wikimedia.org/T88728#1041575 (10ArielGlenn) actually can we start a wiki page for this someplace? [15:53:37] 3Services, Datasets-General-or-Unknown, MediaWiki-Core-Team: Improve Wikimedia dumping infrustructure - https://phabricator.wikimedia.org/T88728#1041632 (10bd808) >>! In T88728#1041575, @ArielGlenn wrote: > actually can we start a wiki page for this someplace? Sure. I started a page: 3Services, Datasets-General-or-Unknown, MediaWiki-Core-Team: Improve Wikimedia dumping infrustructure - https://phabricator.wikimedia.org/T88728#1041633 (10bd808) [15:57:27] 3Services, Datasets-General-or-Unknown, MediaWiki-Core-Team: Improve Wikimedia dumping infrastructure - https://phabricator.wikimedia.org/T88728#1041644 (10bd808) [16:34:29] 3operations, MediaWiki-Core-Team, Wikimedia-Logstash, Incident-20150205-SiteOutage: Prototype Monolog and rsyslog configuration to ship log events from MediaWiki to Logstash - https://phabricator.wikimedia.org/T88870#1041691 (10bd808) This code path is running in beta now and logs are showing up in Kibana as exp... [16:37:10] 3operations, MediaWiki-Core-Team, Wikimedia-Logstash, Incident-20150205-SiteOutage: Prototype Monolog and rsyslog configuration to ship log events from MediaWiki to Logstash - https://phabricator.wikimedia.org/T88870#1041701 (10bd808) The full set of changes requires several changes that will be included in 1.25... [17:43:22] anomie|holiday: I apparently don't have the gerrit rights to force https://gerrit.wikimedia.org/r/#/c/190770/ past the failing Jenkins test :( [17:43:54] * bd808 will backport the test fix [17:45:03] bd808: Remove Jenkins as a reviewer, then you should be able to force it [17:45:51] Ok. will do next time. I just backported the test fix and rebased [17:45:58] easy enough [18:32:17] bd808: So, #contint is blocked on #mediawiki-core to fix concurrency with sqlite. TimStarling found a way with the PDO lib https://phabricator.wikimedia.org/T89180#1033549 Not sure if that has any side-effects. Know someone who could help with that? I can give it a shot too, but might take longer. It'll be... educational. [18:33:50] Krinkle: hmm... off the top of my head, ^d and Aaron would be folks I'd ask. [18:34:12] And since ^d is in releng now he'd be my first target [18:34:31] * bd808 protects his resources as a good PM will [18:35:27] p for product, project, problem? All of the above I'd say :) [18:35:51] I certainly have all three [18:54:44] Krinkle: isn't Tim's fix just a matter of adding a couple of lines somewhere? [18:55:50] ori: It should yes, but I wouldn't really know what I"m doing or have a way to test it. [18:55:58] I'll write the patch though, someone can review it that way [20:24:35] legoktm: Hey, around? [20:24:46] hoo: ish [20:25:11] Whoops... I read my calender wrong, everything is ok [20:25:18] Thought I couldn't make it to SWAT tomorrow [20:25:19] but I can [20:25:21] ok :P [20:25:41] I'll be on a train during that time [20:26:12] I thought we could deploy that today... but turns out today is a holiday in the US [20:26:22] :D [21:36:32] 3wikidata-query-service, MediaWiki-Core-Team: Investigate BigData for WDQ - https://phabricator.wikimedia.org/T88717#1042321 (10Haasepeter) Nik, which dates/time would be be good for you to talk? I can arrange a web conference. [22:45:26] legoktm: Keegan: If Stewards rename through the new queue special page... will that be logged locally to each wiki? Seems like it's not in some cases [22:45:40] I didn't look myself, but that's what others tell me [22:45:51] seems like it's not logging locally for local only renames [22:45:52] that's bad [22:52:58] hoo: for a non-global rename the current logic just fires a LocalRenameUserJob -- https://github.com/wikimedia/mediawiki-extensions-CentralAuth/blob/master/includes/specials/SpecialGlobalRenameQueue.php#L399-L415 [22:53:27] I see... but I don't think that one does any logging [22:53:52] yeah, it doesn't [22:53:53] :S [22:55:05] This is why we can't have nice things [22:55:35] So if the job doesn't log and the job is async... where should the logging happen? [22:55:55] Probably in another job [22:56:25] Under what circumstances should the local rename not be logged? [22:58:41] No idea [22:58:48] how do we handle that for global renames? [22:58:54] yeah, me neither [22:58:55] I know we changed that at some point [22:59:04] specials/SpecialRenameuser.php: $logEntry = new ManualLogEntry( 'renameuser', 'renameuser' ); [22:59:20] the special page does the logging for the Renameuser extension [23:01:19] the global action does it after submitting the jobs -- https://github.com/wikimedia/mediawiki-extensions-CentralAuth/blob/master/includes/GlobalRename/GlobalRenameUser.php#L123-L127 [23:01:39] So I guess the same thing should happen in the special page for the local only case [23:02:27] That wont work... it's cross wiki, in the end [23:02:32] so it would need to be a job [23:02:48] *nod* so it logged on the correct wiki [23:03:14] so in the global case I guess it's only logged on loginwiki? [23:04:12] "Log a global rename into the local log" -- https://github.com/wikimedia/mediawiki-extensions-CentralAuth/blob/master/includes/GlobalRename/GlobalRenameUserLogger.php [23:04:42] So it's logged wherever the specialpage is used? [23:04:47] yeah [23:04:52] which is ... meta? [23:04:54] like all of CentralAuth... messy [23:05:01] yeah [23:05:03] yep, meta... but in theory people could do it on any wiki [23:05:08] right [23:05:13] even messier [23:05:17] (they'd need to change global rgiths first) [23:06:02] I see the name of the Marius guy all over this code. I bet he can figure out how to fix it. ;) [23:06:10] The log entry should occur when the local account is actually renamed, imo [23:06:41] every time or only if it is a local only rename? [23:07:04] I think for starters it would be nice to have *any* logging [23:07:14] maybe we can change how renameuser itself logs [23:07:47] move that from the special page into the RenameuserSQL class (through which all renames go) [23:07:50] bd808: Every time, I think... [23:08:01] * Deskana ponders [23:08:16] As a bureaucrat I always find it annoying when some things change in one log and others in another [23:08:20] The rename log is a total mess [23:08:41] * Keegan reads scrollback and sobs quietly [23:08:48] I like hoo's idea if we want to log all the time [23:08:57] * bd808 hugs Keegan [23:09:07] we have history in the rename queue table [23:09:13] we can fix things if needed [23:09:29] it is "logged" just in a weird way in a weord place [23:09:44] bd808: No one out there will find theses logs [23:10:06] understood, but we have data. That's the most important part [23:10:08] Even I would probably forget to look there :P [23:10:15] True taht [23:10:39] so if we figure out what should have happened we can update other tables appropriately [23:11:22] IMO: Log all renames locally + Log global ones on meta + history in the special page [23:12:50] One time at $DAYJOB-2 we lost the table that tied State income tax payments to individuals. I ended up with $1.7M that I wasn't absolutely sure which account to debit. [23:12:59] This is no where near that crappy [23:13:39] That's perspective. [23:13:45] I once had an intern in a $governmentThing (no idea how to translate that) [23:14:25] and their SQL servers died hard... they lost 3 days of license plates they already gave out physically [23:14:25] * internship [23:14:33] MS SQL server, that is [23:15:05] Title and registration systems. /me shudders [23:15:16] +1 to hoo's logging principles [23:15:27] yup [23:16:30] so basically we should add logging to RenameuserSQL [23:16:57] Identify (potential) side effects, then do that, yes [23:18:59] Keegan: Can you write that up in a tiny little phab task? [23:19:12] Sure [23:19:17] Awesome :) [23:20:35] The RenameUserComplete hook fires when the rename is done too. We could hook that if we really had to. [23:20:58] bd808: We could... but :S [23:21:20] yeah I'd rather see it centralized in the code too [23:22:29] https://phabricator.wikimedia.org/T89681?workflow=create [23:22:32] Phabulous [23:23:01] Well, https://phabricator.wikimedia.org/T89681 for posterity :P [23:24:00] Nice :) [23:55:39] TimStarling: thanks!