[01:02:51] Can somebody run CI again plz đŸ¥º [06:07:58] Fails but only 1.47 stuff that you didn't touch [06:08:01] Cc @cosmicalpha [06:08:04] So you're fine [06:08:53] I'm already aware of that. Plan to fix phan for 1.47 on MirahezeMagic and MatomoAnalytics tomorrow (the only two that fail) [06:09:09] Perfect [06:09:19] We can probably merge @ccxtwf's then [06:09:26] I assume we want someone to run it though [06:10:18] Not yet, need to add some change to ManageWiki to get it to remove when realname is added to hiddenprefs. ManageWiki does not yet support this but will soon. [06:10:41] I mean no harm merging but no need to run it. [06:10:45] Ok [06:12:10] I will merge now [06:13:14] Merged @rhinosf1 @ccxtwf [06:13:34] Thanks for the work with that! [06:14:00] I just pressed big buttons [06:14:06] @ccxtwf did all the work [06:14:12] btw, I decided to make phan failing early this time also because there is hardly any real breaking changes and I wanted to see where we really are with 1.47. [06:14:34] Cool [06:14:40] It should be a super easy upgrade it seems like. [06:15:05] Hardly any SQL or maint scripts also. Extremely uncommon for an LTS. [06:16:21] I am just glad we finally got all the links migrations done with 1.46 after 11 upgrades of doing actor then links migrations back to back. [06:17:45] We still have the file table migration after 1.47 but that shouldn't be hard. [06:20:54] Never say that [06:22:54] Saying things are easy or quiet is banned [06:25:54] That's true lol [07:34:10] [1/3] backlogs for bot access: [07:34:10] [2/3] - https://issue-tracker.miraheze.org/T15998 [07:34:11] [3/3] - https://issue-tracker.miraheze.org/T16068 [07:35:37] the first one includes ua string and should be set private [07:40:45] and should we add tips on emailing tech on the page [[Tech:User Agents]]? [07:40:46] [09:00:07] tbh I think parsoid is the only interesting thing in 1.47 [09:00:27] That is the case yes. [09:00:38] I assume that’s the thing for lts [09:01:13] Yeah most of 1.47 focused on that it seems like [09:01:25] I have issues with parsoid because it is slower and literally just exists as a pet project for the wmf wanting to seem useful pretty much [09:02:03] They have been working on it for 14 years. [09:02:54] it breaks a bunch of things still though [09:02:58] portableinfobox [09:03:09] it’s just worse for everything except visualeditor [09:03:13] PI will be fixed before 1.47 is stable. Or at least should be. [09:03:41] my only other complaint is that it’s slower but [09:04:13] It actually isn't. I measured it being 60% faster in the parses I benchmarked for us. [09:04:39] I thought it was slower [09:05:20] I thought that was the finding at the WMF also but that wasn't the case when I last benchmarked it. I only remember the 60% faster though I don't remember exact numbers or exactly when I did it. [09:05:43] I think it was about a year ago. [09:06:16] tbh I might go send seer to find if the profiles we have recorded show that [09:06:17] They have also improved its performance quite a lot now also. [09:06:44] I will do more benchmarking on 1.47 Parsoid soon. [09:07:15] I need to go get 1.47 on the WO cluster for testing [09:07:23] but using prod is a pita [09:08:05] I installed it on our beta cluster 30 minutes after it was branched lol. I finally actually upgraded our exttest wiki to it today though. [09:08:56] did you fix cve-2026-99999 btw [09:09:13] I realised that was a thing 2 days ago and had to go reboot the hypervisors [09:11:17] my testing concerns shoving 1.47 onto the prod cluster and then upgrading testwiki and routing it to the staging host [09:11:31] more so for development speed than practicality [11:38:52] @cosmicalpha Cosmos is looking really good so far on https://cosmos.mirabeta.org/wiki/Special:Version, although I'm not a fan of the search bar spanning the entire top bar. [11:39:07] But I know that's been like that for a long time in Cosmos' history. [11:39:45] Do you think you'll add the ability for people to add custom links and dropdowns in that spot and shrink the search bar? [11:42:53] Also an interesting thing, when you try to "login" to the cosmos test wiki with the cosmos skin loaded, it doesn't load the proper font, instead falling back to Times New Roman [12:51:49] weird gloop found it to be about 2.5x slower, I think because it emits loads of metadata into the html [16:15:01] [1/2] something funny going on at : I see a flash of the avatar for a split second, then something overrides it and it disappears. I see nothing in a cursory check of managewiki, common.css, no js to see, applies regardless of my safemode/filters, and the culprit invoking 'display:none!important' as far as the inspector is concerned [16:15:01] [2/2] is a random string, and I don't quite have the knowledge/time to dive into what that's supposed to be. And it seems to only occur on this wiki and not other wikis I'm visiting with SocialProfile. Curious if this has any quick thoughts or needs to elevate to phorge [16:24:29] doesn't happen for me on [16:24:49] or when viewing normally or when viewing on private mode [16:52:48] does it happen at all for you? happens for me exactly under that link which is how I typically browse, and it was reported by a local user to me in dm [16:52:55] cc @bloo_j [17:16:39] I need a way to fix this bug [17:22:43] apparently its not fully certain that everyone actually sees it so more input there is necessary [17:23:14] @bloo_j do you use ublock origin? [17:23:46] I am unsure [17:24:22] check your extensions because ubo was the root for me [17:24:41] a filter might be catching it under this specific domain for some reason [17:24:59] I would rn I am away at my college but I could be checking that when I get back from it [17:25:18] I need to move to other things but will take a very quick look in the logger [17:26:12] and now with a proper recheck the problem is gone [17:35:02] I don't see any existing examples within `ManageWikiSettings.php`, but how do I define two settings that need to be disabled for one setting to be enabled. [17:36:40] [1/10] I was thinking maybe this: [17:36:40] [2/10] ```php [17:36:40] [3/10] 'requires' => [ [17:36:41] [4/10] 'settings' => [ [17:36:41] [5/10] 'setting' => [ 'wmfTestVariable', 'wgTestVariable' ], [17:36:41] [6/10] 'value' => false, [17:36:42] [7/10] ], [17:36:42] [8/10] ], [17:36:42] [9/10] ``` [17:36:43] [10/10] But I'm not sure if I can just stuff `'settings'` within its own array and have two arrays, each with their own settings array inside of them? [20:27:28] anyone know why this happens? [20:34:55] [1/3] ``` [20:34:56] [2/3] Found: 5751 [20:34:56] [3/3] Failed: 5751``` [21:36:24] @posix_memalign @karmaskeleton I don't recall anything else blocking it aside from the untested extensions? DarkMode and Cosmos were removed, and TableProgressTracking is blocked on [#15](https://github.com/Telepedia/TableProgressTracking/pull/15), Cargo was not checked iirc [22:48:21] This was blocked by Trust and Safety clearance on access to temporary account data (e.g. country, ASN, and full IP). It has been cleared now in https://issue-tracker.miraheze.org/T15150#326158. Someone just needs to implement it.