[02:04:10] So I found out today that special page aliases can be done in JSON now! I somehow missed that in 1.42 lol [02:06:00] Which is cool lol [02:06:28] I like the JSON format a lot better (and it may come to TranslateWiki eventually also lol) [02:07:03] Already converted CreateWiki to it lol [02:07:06] https://github.com/miraheze/CreateWiki/commit/6e1dd27aa915ad37b03895dfb397d89a7c19c21f [02:21:13] @cosmicalpha https://github.com/JeroenDeDauw/Validator/issues/40#issuecomment-2469301409 [02:37:50] I'm really confused about this, am I just reading it wrong? [02:38:18] If they are on mw 1.35+ there should be no reason to call the Validatot.php file and they wouldn't get the error? [02:55:02] [1/2] Hi hello just quickly forwarding an apparently important report: [02:55:03] [2/2] https://cdn.discordapp.com/attachments/1006789349498699827/1305727530514255952/image.png?ex=67341506&is=6732c386&hm=bfe1f81def399cf7ca2b5915fcc3426d04eade701a582b7f8117a58ff18f8294& [02:56:23] (https://discord.com/channels/407504499280707585/808001911868489748/1305609749924151318) [03:07:46] I have reverted back to hcaptcha until we can fix that. The change will go live shortly which will fix it. [03:08:05] @candlesign.4539 ^ [07:04:43] oh no problem, so its workin now id assume? [07:05:17] i cant find it on [this list](https://dev.miraheze.org/wiki/CSP_Whitelist) [07:05:36] but its on [this list](https://github.com/miraheze/puppet/blob/master/modules/mediawiki/data/csp.yaml) [07:18:59] oh lol devwiki is not kept up to date only the puppet one matters. I did add it now yeah. [07:20:00] okay sweet time to test [07:20:37] If it doesn't work let me know [07:21:15] If it doesn't, can you open the browser console and tell me the CSP error it gives. I may have missed something. [07:34:09] https://cdn.discordapp.com/attachments/1006789349498699827/1305797769675542631/image.png?ex=67345670&is=673304f0&hm=e0c58b9dbc3c4ff98aa0609ccf5ff5086761dfebcebc5859b482c2e497ed358a& [07:34:16] use.typekit.net is still in the CSP [07:35:02] I guess I need to add to script-src and style-src also... [07:35:37] oh just style-src [07:35:45] script-src is a different domain [07:35:53] [1/2] pretty much everything is set too [07:35:53] [2/2] https://cdn.discordapp.com/attachments/1006789349498699827/1305798205275111454/image.png?ex=673456d8&is=67330558&hm=036eb15f62a39b95d159de95636770b53d074fc9a98b4bc745d9392171055118& [07:36:07] as evil as adobe is their font selection is cool [07:37:34] https://github.com/miraheze/puppet/commit/2379eaa39d102a022d1c0125b3151cc60650ae3f [07:38:05] It will be deployed ~30 minutes maybe less [07:38:12] okay i'll check back then [07:38:31] Ping me if you have another error in say 31 minutes [07:38:39] okie [07:39:44] simpleanalyticscdn is not something I think we can add btw. If that becomes a requirement from adobe I'm afraid we won't be able to do it. [07:39:57] but that might be something else [07:40:05] I haven't looked into that one at all [07:46:44] its not thats my own analytics thing that i tried adding [07:47:02] i just use it for my other wiki sites thoughtd itd be nice [07:47:21] cuz i dont really like Special:Analytics [07:48:13] You aren't allowed to collect your own analytics via third party services if they'd have to process user pii [07:48:32] [1/2] its not really doing anything right now i can just remove it [07:48:32] [2/2] whats user pii [07:49:06] Personally identifiable information [07:49:13] oh it doesnt do that [07:49:17] Which includes IP addresses [07:49:32] IP + User Agent is included in that [07:49:38] https://cdn.discordapp.com/attachments/1006789349498699827/1305801667924594729/image.png?ex=67345a12&is=67330892&hm=7a7c8cd7a7d722d651e11162cb93e4bc668c58e4d43b927d6706d067e78618b0& [07:49:42] Which given it's a web request, it's doing [07:49:49] I said process not store [07:49:53] oh process [07:50:01] oh right it does see locations [07:50:12] You have to process an IP address to handle a web request [07:50:32] So it has to go through the CSP process to be approved [07:50:41] And we've never allowed external analytics [07:51:17] yeah ive had this convo before [07:51:22] oh one thing i wanted to ask [07:52:02] Section 3 of the privacy policy doesn't allow it to be shared in a non anonymised fashion [07:53:18] can Special:ManageWiki add an option for changing the favicon based on the user's color scheme [07:54:15] Erm not sure [07:54:23] okay sweet @cosmicalpha its loading now [07:54:26] Cause favicon is probably cached [07:54:43] So we'd need to deal with that cause it's always got to be at favicon.ico [07:54:50] because i do remember seeing somewhere you could put in something that detects the users colours scheme and serves a different icon [07:54:56] Nice! [07:55:13] Where did you see that [07:55:20] congratulations for the 5 people who use adobe fonts [07:55:26] let me search [07:56:57] https://cdn.discordapp.com/attachments/1006789349498699827/1305803506736955412/crying-emoji-dies.gif?ex=67345bc8&is=67330a48&hm=4917bbf0f9e484f6d2dfbae3882a396c4702316a64d4030167eda8257af2fbe7& [07:57:00] banned [07:57:06] was doing something like `` [07:57:54] its def gotten better but whyyy is it public [07:58:17] like i can see untitled tag games one right here https://utg.miraheze.org/Special:Analytics [07:58:38] not sure about that actually [07:58:57] the initial extension author didn't think it was important to implement the ability to hide the page [07:59:02] Maybe we should add a right allowing wikis to restrict it but public by default... [07:59:09] I was about to ask [07:59:14] It is five lines to add [07:59:16] should we implement a new right that controls that? [07:59:18] lol yeah [07:59:18] [1/3] or how tables like this have the tinest bars [07:59:19] [2/3] its like you gotta hover over that correct pixel [07:59:19] [3/3] https://cdn.discordapp.com/attachments/1006789349498699827/1305804101560303637/image.png?ex=67345c56&is=67330ad6&hm=fa529fe717978566496fd82dd0d58dec343b94fb732d0ae3a6ab854993e76d3e& [07:59:25] yes please [07:59:38] managewiki option [07:59:45] nothing a little bit of css can't fix [07:59:46] @agentisai yes there is absolutely no reason not to add a right [07:59:57] very well [07:59:59] it shall be done [08:00:00] Just assign to * by default [08:00:05] in extension.json [08:00:10] yep [08:00:53] name the right viewanalytics [08:01:01] [1/2] imo there should be a selector for first x options otherwise it because totally unusable [08:01:02] [2/2] last time i checked there were no charts though so it is improvin [08:01:56] you can tell we aren't graphic designers [08:02:09] lol yes [08:02:52] also do yall track edits made/day [08:03:01] you could add an event for that and show that as a stat [08:03:37] that'd be tracked outside of Matomo but I did think about that [08:03:50] not too sure how matomo works [08:03:56] @agentisai I'll give a PR review when you add the right if you want I trust you for a trivial change though if you want to just merge you can feel free. Just as long as it is called viewanalytics [08:04:00] but im sure they have some events tracking thing right [08:04:23] Thanks for the heads up. I was about to name the right `matomoanalytics-view` lol [08:04:43] It's a generic analytics service so it can't tell and log MediaWiki edits [08:04:49] ah [08:04:52] lol yeah I think that is to long personally simplicity is better [08:05:01] but it shouldn't be too hard to get MediaWiki to output those stats [08:05:03] and it goes better with the noanalytics right lol [08:05:08] heh yeah [08:05:49] the noanalytics right was a smart idea but why did no one think of adding a view right? lol [08:06:01] [1/2] maybe the recentchanges api? [08:06:01] [2/2] surely its somewhere [08:06:30] who wrote the extension initially? [08:06:34] I'm guessing John? [08:07:01] you can probably filter using an SQL query to look at the revision table and aggregate results that way [08:08:11] now thats wayyy outta my knowledge [08:25:01] Thanks for the patch @agentisai [14:43:43] Hm, Another wiki crat approached me with a valid reason but it seemed pretty specific. Just wondering why you’d rather It’d:ThinkerMH: be private [14:47:01] Anyways, @agentisai @cosmicalpha I’m assuming the right should be added to the appropriate global groups? Tech and steward are a given, idk if there’d be a point in adding to GA or WM [14:49:24] Um [14:49:39] [1/3] @agentisai friend, I believe you may have made a typo? [14:49:40] [2/3] > You do not have permission to view Matomo-collected analytics, for the following reason: [14:49:40] [3/3] The action you have requested is limited to users in the group: Ajaxpoll.json. [14:50:59] Isnt that how it should be by default? [14:51:17] What reason is there for it to be private by default [14:51:18] Why would a random person need to see detailed information about your website and its visitors [14:51:37] What’s the harm in having so [14:51:43] The good thing now is that its now a setting [14:51:57] hmm? [14:52:23] On the analytics page [14:52:33] The group is assigned to is a bit uh [14:52:35] Funky [14:52:57] on what wiki? [14:53:06] https://phighting.wiki/Special:Analytics [14:53:22] And https://rainverse.wiki/wiki/Special:Analytics [14:53:40] [1/2] Theres so much data there about users [14:53:40] [2/2] It really should only be sysop only [14:54:20] all of the data is aggregated so it doesn’t give information about any one user [14:55:17] It’s been like that for 5 years but I guess the page was so boring no one checked it out [14:55:18] [1/2] that aside imo it still is just weird that its public [14:55:18] [2/2] who else just has their site analytics lying around for you to look at [14:55:26] i mean an option allows you to [14:55:28] have that on [14:55:30] or off [14:55:32] easy [14:55:56] Really just a memetic sleep agent [14:56:10] funny enough fandom has a analytics page that tells you to keep things private but it has a channel in its discord to share analytics [14:56:34] I know Wikimedia has views public [14:56:36] Idk bout the rest [14:56:43] Always cracks me up [14:56:44] They don’t care about privacy though. That’s very clearly to stop competition lol [14:57:06] Anti fun [14:57:47] @cosmicalpha https://github.com/miraheze/MirahezeMagic/commit/e26fb1bc73f299f210580198c836b072a9d608fd [14:58:35] How did I get 13 emails in [14:58:53] 2 hours? [15:36:20] I think I missed something, what's up with templates not working on Meta? [15:37:05] [1/2] as in... [15:37:06] [2/2] https://cdn.discordapp.com/attachments/1006789349498699827/1305919305917136906/image.png?ex=6734c7a1&is=67337621&hm=444ba022ade1164558e99fac8207ef79777385de0e31a5ce1a1c7e6cc3637e59& [15:40:58] Hmm, is it my internet or is MH loading terribly? [15:41:05] (3888ms) [15:59:21] Yeah, claire was able to track it to "Warning: Template include size is too large. Some templates will not be included.", not sure what's gone funky. [16:13:47] I can reproduce as well, and SR at least seems to load very slowly [16:22:35] Currently stuck on P0 work project so can't dig in further, but two avenues I'd check would be any settings changes around expensive parser function limits + any changes to key templates/modules [16:38:47] yeah, the farm is unusually slow [16:43:27] We need more tractors! [17:03:35] Grafana show anything? [17:04:13] From now on tractors is the official name of a mw* server [17:13:27] it would've been funny if we used names of tractor brands and models as our server names 😄 [17:14:22] And always break down [17:48:14] FYI, multiple users have reported issues with receiving emails related to account verification [18:00:32] fixed the template issue - https://meta.miraheze.org/wiki/Steward_requests/Miscellaneous?action=history - wakibot made this clearly unintentional edit which made the page really big [18:01:13] err https://meta.miraheze.org/wiki/Steward_requests/Miscellaneous?diff=prev&oldid=441834 [18:01:40] Simply reverting that fixed everything [18:03:18] it somehow added {{status}} after every single character [18:03:26] @link [18:03:26] http://enwp.org/Template:status [18:03:56] Value miraheze was stored into default-link-wiki in the config [18:03:56] @configure default-link-wiki=miraheze [18:04:00] @link [18:04:00] https://meta.miraheze.org/wiki/Template:status [18:06:40] how did that break so many pages though? [18:06:50] that really shouldn't happen [18:07:36] status was transcluded 100s possibly 1000s of times and so the page size adds up, but not sure how that branched out to non sr pages [18:09:07] oh wait, Requests dashboard has to read the sr pages to get open status count, so that would have affected any page it was transcluded on [18:09:37] {{status}} was transcluded 32,886 times... [18:09:38] [18:10:33] @suzuneu just want to be sure that you are aware of what happend [18:10:38] That would do it [18:12:28] ahhhh [18:12:31] going by https://issue-tracker.miraheze.org/T12874#258193 over 65k times [18:13:23] 65,971 [18:52:12] ^ [18:56:44] @suzuneu the bot did it again thus, I am disabling it until you can fix [19:34:54] From which provider? [19:35:32] (Which is an important info when talking about email deliverability) [19:37:46] I think some were gmail [19:37:50] I’ll check [19:40:07] Don’t recall what emails though [19:40:36] Also it would probably help to state what kind of issues; is it… just delayed? Or sent to Junk? Or just not at all arriving? [19:40:46] Wait, oh there’s an email I completely forgot to make an account for whoops [19:41:15] If you are gmail, you can add + or just use the . magic to make it look like different address [19:41:31] (ie. revi.wiki and reviwiki@gmail is identical address) [19:41:36] (IIRC, that is) [19:41:38] Tali said they checked junk iirc [19:42:04] Huh, my school account has a dot in it, I’ll try that [19:42:43] This sounds like a good time to check if my reviwiki@ address forwarding is still working :-p [19:44:31] `04:43:50 metawiki  [[Special:Log/newusers]] byemail  Revi *  created new account User:Revi-test-20241113: Testing email deliverability issue` let's see… [19:45:47] Yeah, not arriving at my main inbox, gotta log in to gmail… [19:49:40] Lol, it was my gmail forwarding address error [19:51:55] [1/19] ```eml [19:51:56] [2/19] Received: from mail-sor-f97.google.com (mail-sor-f97.google.com. [209.85.220.97]) [19:51:56] [3/19] by mx.google.com with SMTPS id ca18e2360f4ac-83df512b5bdsor736197639f.10.2024.11.12.11.43.49 [19:51:56] [4/19] for [19:51:57] [5/19] (Google Transport Security); [19:51:57] [6/19] Tue, 12 Nov 2024 11:43:50 -0800 (PST) [19:51:57] [7/19] Received-SPF: pass (google.com: domain of noreply@miraheze.org designates 209.85.220.97 as permitted sender) client-ip=209.85.220.97; [19:51:58] [8/19] Authentication-Results: mx.google.com; [19:51:58] [9/19] dkim=pass header.i=@miraheze.org header.s=mail header.b=aqO9cfTr; [19:51:58] [10/19] spf=pass (google.com: domain of noreply@miraheze.org designates 209.85.220.97 as permitted sender) smtp.mailfrom=noreply@miraheze.org; [19:51:58] [11/19] dmarc=pass (p=REJECT sp=NONE dis=NONE) header.from=miraheze.org; [19:51:59] [12/19] dara=pass header.i=@gmail.com [19:51:59] [13/19] Received: from ::1 ([2602:294:0:c8::113]) [19:52:00] [14/19] by smtp-relay.gmail.com with ESMTPS id 8926c6da1cb9f-4de78800020sm620539173.60.2024.11.12.11.43.48 [19:52:00] [15/19] for [19:52:01] [16/19] (version=TLS1_3 cipher=TLS_AES_256_GCM_SHA384 bits=256/256); [19:52:01] [17/19] Tue, 12 Nov 2024 11:43:49 -0800 (PST) [19:52:02] [18/19] ``` [19:52:02] [19/19] Hmm, it seems like I'm getting it well @ inbox [19:53:51] Oh. I have 'from:`*.miraheze.org` > never report as spam' settings [20:15:59] At least I don't think miraheze is hitting the rate limit for smtp-relay.gmail? https://support.google.com/a/answer/2956491?hl=en [20:16:27] [1/3] That is [20:16:27] [2/3] > 4.6 million in a 24-hour period [20:16:27] [3/3] > 319,444 per 10-minute window [20:17:31] Yeeeaaaaah I doubt it [20:18:22] But they say 'we can adjust that value based on your sending practice' so it is possible for them to lower the limit [20:19:00] But then I don't know how MH mail integration works nowadays; I resigned in 2018 [20:19:29] When we used to have our own mail server [20:23:03] If noreply@ is a… functioning user account, it might be hit by 10K mail limit by Google [20:23:31] But then why would you use smtp-relay… so I'm lost without proper access :-p [20:29:13] Oh, that makes sense [20:29:26] I’d be lost regardless so props [20:29:44] It has a quite of gmail limits [20:30:15] But I feel we’d feel if more, or at least be alerted if the limit hit [20:30:16] @Infrastructure Specialists [20:30:38] [1/3] 1. 2K total mails sent [20:30:39] [2/3] 2. 10K total recipients [20:30:39] [3/3] 3. 3K unique recipients [20:30:48] For a regular google workspace gmail account. [20:31:16] But then I received the email just fine, so it is not the limits being hit [20:32:35] (Personally I don't think using gmail is not the best way to send automated emails but) [20:35:06] I think it was likely because we’re already managing mail with google workspaces for mailing lists? Or maybe vice versa, I only joined in 23 and started really volunteering in tech and community in 24 so this is before I was around to hear the reasoning [20:35:44] we weren't [20:35:46] Most likely quick decision from migrating away from self-hosted emails, I'd assume for occam's razor [20:36:12] yeah, basically [20:36:32] WikiTide used Workspace as it got it via non-profit status so we kept it when we acquired Miraheze [20:36:51] As expected [20:36:56] Did we only move off of hosting email in 24? [20:37:11] yeah, some time in January [20:37:21] Wow [20:37:27] So not before my time after all [20:37:28] Sometimes after 2023 for sure yeah [20:37:47] Because I used to have my old filters working around dec 2023 [20:38:15] Which relied on the fact that miraheze sends email from `*.miraheze.org` [20:38:26] Oh, makes sense [20:39:08] So if noreply@ is a regular account in gsuite, it is possible that you are hitting rate limits [20:39:26] And intermittent limits maybe because it could be a rolling rate limit? [20:39:39] @agentisai ? [20:39:42] (Gsuite is shorter to type in mobile) [20:39:44] noreply@ doesn't exist as a user [20:39:50] Good, then it's a relay. [20:39:53] yep [20:40:06] At least we denied one possibility [20:40:36] I'll check logs in a bit [20:40:42] shouldn't be a mystery to figure out [20:40:56] Yup, it's a mystery for me who has no log access [20:41:11] At least, 4.3M daily (and 3nnK per 1minute) is quite a large limit for MH to hit so [20:41:24] I doubt we’re hitting it [20:41:31] Yup [20:41:33] Me neither [20:41:53] nothing in logs [20:41:54] email works [20:41:58] I just confirmed it myself [20:42:00] If so that deserves a plaque [20:42:04] It’s not all emails [20:42:14] It’s some specifically [20:42:15] Try sending yourself an email via Special:EmailUser. It works there at least [20:42:30] Can we see logs of individual users [20:42:34] Well [20:42:35] Hmm. Next time at least collect which email provider they use [20:42:37] Emails sent to [20:42:47] So we can get which servers to blame [20:42:51] I can see the email log and I see emails being sent every second [20:42:56] @tali64 probably remembers which emails, we have them somewhere [20:43:10] It may be the receiving end silently dropping [20:43:33] Or check dmarc reports? DMARC policy is reject [20:44:27] Smart [20:44:57] At least fastmail does ignore dmarc policy reject but I don't think fastmail is a majority [20:45:13] (They just treat p=reject as p=quarantine, /me shrugs) [20:45:41] dmarc is passing, just verified [20:46:03] Then we probably need which provider to pinpoint and test against. [20:46:32] A majority of the affected addresses were definitely Gmail [20:46:32] A majority of the affected email addresses are definitely Gmail [20:46:35] Hmm. [20:46:44] Do you have the rest? [20:46:59] And this [20:47:18] [1/3] 1. Just delayed [20:47:18] [2/3] 2. Arrives at 'Junk' [20:47:18] [3/3] 3. Never arrives [20:47:38] The users reported never getting an email whatsoever [20:47:39] According to agent, not likely to be 3? [20:47:51] what're their emails? [20:47:55] Hmm, ^ [20:48:00] I can look them up in the logs [20:49:13] Can someone try sending email to [[mh:revi:Special:Contact]]? (If you are fine with me having your email address) [20:49:14] [20:49:28] (Just trying to see headers) [20:50:12] Sure [20:51:02] Contact or EmailUser? [20:51:10] Anywhere is fine [20:51:26] (I just typed out shortest to type — mobile typing sucks) [20:51:56] Done [20:51:56] Good [20:52:03] SMTP 250 OK [20:52:07] amen… [20:53:42] [1/19] ```eml [20:53:43] [2/19] Authentication-Results: phl-mx-06.messagingengine.com; [20:53:43] [3/19] bimi=skipped (DMARC Policy is not at enforcement) [20:53:43] [4/19] Authentication-Results: phl-mx-06.messagingengine.com; [20:53:43] [5/19] arc=none (no signatures found) [20:53:44] [6/19] Authentication-Results: phl-mx-06.messagingengine.com; [20:53:44] [7/19] dkim=pass (2048-bit rsa key sha256) header.d=miraheze.org [20:53:44] [8/19] header.i=@miraheze.org header.b=U+sKNiBp header.a=rsa-sha256 [20:53:45] [9/19] header.s=mail; [20:53:45] [10/19] dmarc=pass policy.published-domain-policy=reject [20:53:45] [11/19] policy.published-subdomain-policy=none policy.applied-disposition=none [20:53:46] [12/19] policy.evaluated-disposition=none [20:53:46] [13/19] (p=reject,sp=none,d=none,d.eval=none) policy.policy-from=p [20:53:47] [14/19] header.from=miraheze.org; [20:53:47] [15/19] iprev=pass smtp.remote-ip=209.85.219.228 (mail-yb1-f228.google.com); [20:53:48] [16/19] spf=pass smtp.mailfrom=noreply@miraheze.org [20:53:48] [17/19] smtp.helo=mail-yb1-f228.google.com [20:53:49] [18/19] ``` [20:53:49] [19/19] DMARC policy is not at enforcement? What's wrong with you fastmail lol [20:55:20] [1/7] `X-Spam-hits: BAYES_20 -0.001, DMARC_PASS -0.001, FREEMAIL_FORGED_REPLYTO 2.095, [20:55:20] [2/7] ME_HAS_VSSU 0.001, ME_SC_SENDERREP -100, ME_SENDERREP_ALLOW -4, [20:55:20] [3/7] RCVD_IN_DNSWL_NONE -0.0001, RCVD_IN_MSPIKE_H2 -0.671, [20:55:21] [4/7] SHORTCIRCUIT -0.0001, SPF_HELO_NONE 0.001, SPF_PASS -0.001, [20:55:21] [5/7] UNRESOLVED_TEMPLATE 1.252, LANGUAGES unknown, BAYES_USED user, [20:55:21] [6/7] SA_VERSION 4.0.0` [20:55:22] [7/7] I was initially WTF'ed at FREEMAIL_FORGED_REPLYTO but then realized it is normal because Reply-To is indeed supposed to be gmail so [20:56:57] Hmm on an unrelated note phorge mails are routed from wikitide.org [20:58:45] Anyway I believe agent is in better position to comment on that stuff because I've got no log access :-p [21:06:17] for now \:P [21:12:28] Yup [23:47:19] I just woke up, what the hell really happened to cause this? [23:48:44] :you: [23:48:49] also [23:48:49] :you: :3 [23:48:56] morning [23:49:02] mornyan [23:57:19] okay fixed [23:57:45] safe to unblock? [23:57:56] it was not blocked [23:58:15] [[User:Waki285-Bot/status]] was false [23:58:16] [23:58:34] oh