[00:55:17] If I am getting " Main slot of revision 41 not found in database!" in my vagrant, what does that mean? [01:17:50] SMalyshev: maybe MediaWiki thinks the database was migrated to the MCR schema but it actually wasn't? [06:10:09] tgr|away: this is a distinct possibility. Do you know any way to fix it? [08:19:37] SMalyshev: `vagrant destroy; vagrant up` ;) [08:46:23] SMalyshev: change $wgMultiContentRevisionSchemaMigrationStage or run populateContentTables.php [12:30:28] d-day for EP [14:22:32] legoktm: Re https://gerrit.wikimedia.org/r/c/mediawiki/core/+/461696, does the proposal of keeping AutoloadGenerator as it is in that patch and adding a CI test to check for mixed directories seem ok to you? [15:15:03] RIP EducationProgram [15:38:06] * greg-g pours one out [15:38:11] Reedy: Can I mark https://phabricator.wikimedia.org/T125618 as Resolved? [15:40:08] It's mostly just cleanup left to be done [15:40:15] Will do the user groups later this aftenoon [17:00:42] anomie: are you on board with the general idea of having a PageTypeHandler, roughly in the shape proposed, with the prospect of adding more stuff to it later, like action overrides? [17:01:11] if so, i'd perhaps pull that out and introduce it first, before SlotRoleHandler. [17:16:00] tgr|away: populateContentTables didn't help... but setting the migration stage to READ_OLD seems to make it go away [18:21:48] anomie: yes, I think that proposal seems fine. The current implementation of AutoloaderStructureTest does verify in testPSR4Completeness that there are no non PSR4 files in the directory, so I think that part is OK [18:23:56] legoktm: So is there anything else for me to do in that patch? [18:24:08] DanielK_WMDE_: I'll have to give it a look. [18:24:57] anomie: likely not, but I didn't actually read the code yet (only the commit message), I'll try to take a proper look soon [18:25:14] ok [18:27:45] anomie: ok. note that i'm currently in the process of transitioning all my subscriptions and stuff from my wmde accounts to the new wmf accounts. some stuff may fall between the cracks. also, it's quite distracting [18:31:37] DanielK_WMDE_: it's not going to be official until your IRC nick changes :p [18:32:02] * James_F grins. [18:32:43] if anyone has a second, I bet someone in this channel could tell me about https://gerrit.wikimedia.org/r/#/c/operations/mediawiki-config/+/463294/ it's posted for SWAT. I don't know the user or the area of code, but it's reversing something from 2008 so I'm not comfortable with SWATting it without someone else taking a look. [18:33:34] thcipriani: I looked and felt a little uneasy but I don't know enough. [18:34:00] yeah, I can't grock enough from the limited context [18:34:30] -c [18:34:31] It means they won't be able to have different per-language Main Pages, I think. [18:34:44] So… is that good? [18:34:48] DanielK_WMF: Gosh. [18:35:13] Translate gives them the per-language Main Pages [18:35:28] Oh, right, yeah. OK, I guess if the community want it then "fine". [18:35:39] James_F: Community want many things [18:35:42] Most of them are not fine [18:35:50] Context. :-P [18:35:53] basically $wgForceUIMsgAsContentMsg makes the message use the UI language instead of the content language [18:36:22] undoing that will have it use the content language, which makes sense because then it gets passed on to Translate. I think. [18:36:48] Right. [18:37:07] https://phabricator.wikimedia.org/T18701#208739 <-- they're running into what brion warned about in 2008 [18:37:19] so sounds reasonable [18:37:28] Surprise, b.rion actually knows what he's talking about. [18:37:42] cool, thanks for the overview! I'll got ahead and merge. [18:39:01] the patch author and +1'ers are meta-wiki regulars so it looks fine on that front too [18:41:02] * dues is trying to fit in with the crowd [21:38:57] legoktm: Is it me, or is that variable just grossly misnamed? [21:39:25] "Force ui-message as content msg" tells me that something that would use user-interface language normally (e.g. all messages) implicitly get inContentLanguage() called. [21:39:32] In actuality, it does the opposite. [21:40:15] Right? What it does is make inContentLanguage() a no-op. [22:43:09] Krinkle: agreed [22:51:28] Krinkle: I think the idea is that with inContentLanguage you ask for a "content message" (a message in the contentlang) and that settings forces a UI message to be used when the code asks for a content message [22:51:35] but yeah, terrible name [22:52:06] Interesting. [22:52:16] A bit inception-style, but I get it. [22:52:22] Thanks :) [23:19:41] DanielK_WMDE: If you're OK with the naming in https://gerrit.wikimedia.org/r/c/mediawiki/core/+/461975/16 (c.scott's deprecation of unTidy addWikiText methods) I'll merge it.