[00:00:13] there's no documentation as to what the role's purpose is on wiki [00:02:33] There's an obvious attempt to replicate wikimedia yet there's a rather lack of people to fill roles, despite that there are proposals to invent more roles. [00:04:19] My concern is that consolidating some of the roles will raise the bar for appointing [00:05:44] this is for only technical reasons, see [[Special:Log?logid=1110409]] [00:05:44] https://meta.miraheze.org/wiki/Special:Log%3flogid%3d1110409 [00:05:45] [00:05:48] oh [00:06:01] [[Special:Redirect/logid/1110409]] [00:06:01] https://meta.miraheze.org/wiki/Special:Redirect/logid/1110409 [00:06:03] [00:06:31] why not just limit it showing up for testwiki instead of at meta? [00:07:54] idk [00:09:55] No it is not useful [00:09:55] [1/4] let's start with the concept of a trusted user, what is the bar to be considered trusted? [00:09:56] [2/4] autopatrolled grants both autopatrol and skip captcha [00:09:56] [3/4] - since this already grants skip captcha, the proposal just invents a role without a good justification. all the proposal does is split this role in 2. [00:09:56] [4/4] ip block exempt grants ipblock-exempt, StopForumSpam bypass, tor unblocked [00:10:09] Ok sure you can add that to the proposal [00:10:21] already added as a comment [00:11:04] as note, from special:listgrouprights that autopatrolled grants both autopatrol and skipcaptcha so the proposed role doesn't add anything of value compared to existing roles [00:12:00] there's also a confirmed role that grants skip captcha [00:12:05] [1/3] We have found the cause. This is a wiki farm where I can use my ManageWiki, but even though this betawiki has a group named test-group, the group does not show up when I try to operate it from metawiki. This means that if I try to change a user group from metawiki, I can only change the user group that exists on metawiki, regardless of what user groups exist on betaw [00:12:05] [2/3] iki [00:12:06] [3/3] https://cdn.discordapp.com/attachments/407537962553966603/1244442970271518730/image.png?ex=66552154&is=6653cfd4&hm=3a5812090ba876baff78864fc3804069e7c2d21c154f8d201504d17854af8011& [00:12:38] This is a global, not local, group [00:12:43] This means that if you want to manipulate testwiki permissions from metawiki, a group with the same name must exist on metawiki [00:13:14] That specific autopatrolled role exists on metawiki, you currently cannot be given that role on every wiki [00:17:13] yes its automatic based on certain conditions. as the oppose in the proposal. instead of earning autopatrol or even an explicit grant of confirmed; overruling local communities...what for? [00:17:57] [1/4] > overruling local communities [00:17:57] [2/4] Skipping captchas is not overruling [00:17:57] [3/4] > what for [00:17:58] [4/4] For the purposes stated in the RfC [00:19:22] the proposal is for a new global role that grants skip captcha across all wikis instead of earning it on per wiki basis, that's called overruling the conditions to meet autoconfirmed status which is automatically granted [00:19:59] [1/2] A local community can always opt out of global roles [00:19:59] [2/2] Skipping captchas is not an advanced permission [00:20:18] ok, now we getting somewhere [00:20:44] effectively a global role that combines autopatrolled, confirmed will not give any advanced permissions [00:20:51] Global sysops and Stewards can already skip captchas on every wiki regardless of a wiki's configuration and nobody has complained about that even though skipping captchas is irrelevant to their job [00:21:09] Well yes the RfC never said it granted advanced permissions [00:21:17] I'm gonna go now talk later [00:21:23] stewards and global sysops are the highest authority across miraheze [00:22:54] not even a comparison [00:24:21] I don't think global CAPTCHA skip is required. We have IPBE for those who contribute to constructively and are stuck in a "bad hood" sharing a range with those who do noncostructive shit on MH [00:25:02] +global rollbacker, well I don't think it's appropriate to compare it to other role [00:25:17] [1/2] I wasn't able to get that group even though I'm a constructive user and want to use Tor [00:25:18] [2/2] And furthermore that is for a different purpose than this one [00:26:52] If you have to find all the fire hydrants for 10 edits and 4 days it's not a major issue, but I do think that requirements for GIPBE have to be lowered [00:27:02] it's relevant in the sense that a steward will not nearly be able to do their job as well if they have to fuss about captcha when intervening to enforce a global policy, smacking vandals with their nuclear buttons or otherwise messaging/acting in an elevated capacity tbh [00:27:13] ditto gs [00:27:47] True but they could still do that if they experienced captcha, it'd just be annoying [00:28:10] captcha exempt in this scenario lacks the responsibility element that justifies bypass of the usual steps although skipping captcha is harmless and that's something I wouldn't mind seeing as automatic for users past a certain level of global edits or something, it's purely an antispam feature [00:28:17] That is very annoying when trying to revert vandalism or fix a wiki's templates and etc [00:28:54] really the trouble with gce is that there'd be so many people who nominally would be just fine with the exception that it would take all day many days to cover everyone with a right to that call [00:28:55] Each captcha takes a few minutes, I have had this several times where I spend a few minutes on each trivial edit [00:29:12] the underlying idea behind this RfC is my belief that Captchas are only for anti-spam and that non-robot users do not need Captchas [00:29:15] Take en.wiki as an example, a lot of edits from IPs don't even need the verification, unless links are being inserted or etc [00:29:29] it should be automated, manual assignment is incredibly clunky [00:29:36] Miraheze has a bigger spam problem [00:29:53] Because our filters aren't as good and we don't have as many patrollers [00:30:02] So captcha is very needed [00:31:01] That would be good [00:31:33] The CAPCHA does also tend to remember the last time you clicked on all the traffic lights and the next time you click the box it might not require another solution, this is especially true for Cloudflare Turnstile, the friction for users is very low but bots tend to have a hard time. [00:32:19] I've found cloudflare to be better than the others with that [00:32:36] [1/2] Another concern is that the Global groups reform, currently in draft form, would eliminate or raise the appointing standards for GR, a position that can skip Captcha for reverting vandalisms [00:32:36] [2/2] (In this aspect, perhaps this RfC should have been included in the Global groups reform) [00:35:41] the autopatroll role which is granted automatically provides skip captcha permission [00:40:22] @suzuneu I would advise to first outline what particular roles are needed then design the roles with the relevant permissions instead of trying to literally replicate wikimedia which isn't going to be possible [00:53:19] Hi. Do you know if there is a mediawiki special page for "longest page title". To know which pages have the longest titles? [00:59:07] there is not, you can find all special pages listed at Special:SpecialPages [00:59:37] Ok. Thanks [01:03:11] MediaWiki does not have that functionality and I am not aware of any extensions that do, but your best is going to Special:AllPages, copying all the titles into something like Notepad++ and sorting by line length. [01:04:16] It requires edits on each individual wiki, provided rationale above for why I have an issue with the lack of a global role [01:08:19] as I mentioned to waki and in the past, evaluating all existing roles and perhaps redesigning any that are replace-able might be a more useful rfc [01:10:14] This is true but I think this RfC can be a starting place, it doesn't preclude getting rid of other roles after or before it passes [01:11:19] #Language embedding to Wikipedia article [01:12:44] the main issue is later, would need an rfc to basically kill this proposed role and merge all users part of these into such new role for redundancy. one goal is to do this in a smarter way and reduce redundant actions for later. [01:13:05] It’s due to appear in red [01:13:17] Like there’s not page related [01:13:36] looks like a interlanguage issue [01:13:42] Yup [01:13:56] if you don't have access, notify someone who does [01:14:05] I need to appear in blue [01:16:33] I don't see why we should kill this role [01:16:46] If we do need to merge it into something else I expect that to take weeks to months [01:23:53] [1/2] I was saying later when the global roles are reformed, we may need to kill this as redundant. [01:23:53] [2/2] work smarter, not harder [01:24:46] what are we killing as redundant? [01:24:58] [1/2] That’s what I’m mean about [01:24:58] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1244461309706109120/IMG_3568.png?ex=66553269&is=6653e0e9&hm=d779e59c8c8f6f0a509c7981c39939fd3a83162a077faed1a3d0c32f85d0b3c9& [01:25:13] And I don’t know [01:25:16] you're linking to a wikipedia article? [01:25:53] read the discussion please instead of randomly jumping in [01:26:42] I only want at the time of using the {{native name | lang |}} the name of the language appear on blue, not red. @pixldev [01:26:42] https://meta.miraheze.org/wiki/Template:native_name_ [01:27:03] What [01:27:10] It’s like there’s not page linked to that language [01:27:43] I have no idea how this works so I can't help since you provided no link [01:27:59] https://furrnations.miraheze.org/wiki/Draganian_Republic_of_Turisa?mobileaction=toggle_view_mobile# [01:28:17] user needs someone with interlanguage permissions [01:29:28] I only want use the {{native name | lang |}} correctly [01:29:28] https://meta.miraheze.org/wiki/Template:native_name_ [01:29:43] asking a question, really no need to be hostile [01:30:15] [1/2] Yeah but that's later [01:30:15] [2/2] It's not going to save much time to just not make this because at some point we will need to reform all global roles, at some point we may need to kill global captcha exempt but for the time being reforming all the global roles is going to take a while and this can be useful until it is done [01:31:47] i'm not being hostile, its a very basic expectation when joining a discussion [01:33:37] Brandon, see DM [01:34:18] Asked a question relating to context on the topic, because backreading an entire conversation is difficult in a clogged channel. You may not have intended to be hostile but you certainly came across that way and I’d advise you to try and check for that in the future. [01:35:52] Both calm a bit please. Don't need to start an argument over something that small. [01:37:28] [1/2] By example, on other of the wiki entry appears correctly [01:37:28] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1244464456805646367/IMG_3569.png?ex=66553557&is=6653e3d7&hm=36fa1f382b906e49186e245cca4c39ee76508656d74a531ce9f382b6deaf95dd& [01:42:12] I don’t understand [01:48:57] Done for translation (and following current RfC for today) iirc all the important policy and functionary pages are now up to date and translated [01:57:09] I know that’s not a solution but [01:58:13] [1/2] I fixed with: [01:58:14] [2/2] {{lang | iso code | text}} ([[w: name language|name]]) [01:58:14] https://meta.miraheze.org/wiki/Template:lang_ https://en.wikipedia.org/wiki/_name_language [02:05:04] But doesn’t look good [03:48:05] How do i set this up in my server? [03:48:09] https://cdn.discordapp.com/attachments/407537962553966603/1244497345425575998/image.png?ex=665553f8&is=66540278&hm=466e4eee4f82c0bdd2173f5b66ebb5b1caa5c0a7b44d7dc18aa6fb47da8acf86& [03:48:33] agent set it up ig when he owned lgballt [04:03:01] Is he Agent? [04:03:27] And I wanted to know what's the progress of my request, to add .p7m and .p7s into my wiki [04:03:50] you was offered alternatives [04:04:08] May I know what are the alternatives [04:07:25] [1/2] please scroll back to your previous messages, encryption isn't my field of interest [04:07:25] [2/2] also things like new file extensions and other tech stuff has to be requested in Phorge [04:07:56] Okay, thanks. Whenever necessary I will file a request in there [04:12:14] that would be extension:DiscordNotifications [04:12:23] all you need is a discord webhook [04:12:45] ??? [04:12:49] thanks [04:13:57] You said this user is Agent so I was wondering if he became active again [04:14:25] agent set up the thing for lgballt [04:14:30] the user who edited was me [04:15:35] Ohh got it [04:16:39] So you mean the Discord bot🤣 I just assume if he has set up something making changes on your wiki so the bot posted a notification [04:16:54] yeah [04:18:13] Oh I have no idea, I added this wiki bot in my server but has never seen it posting anything [04:42:11] Not a bot, it's automatically pulling from recentchanges via discord webhook [04:43:14] Is this periodically checked or automatically updated once changes have made? [04:52:52] automatically updated [05:02:44] I've read #announcements and stuff but they say the issue's fixed— my wiki seems to be loading every page significantly slower than it used to; is it okay if I create pages right now or is my wiki very poorly optimised? [05:05:09] [1/2] There’s a way to remove this category? [05:05:09] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1244516723189874708/IMG_3578.png?ex=66556604&is=66541484&hm=d727fdb4d6c8d59b4f76873d415f6382edf23985c911cc9bfbf09875baa83fb7& [05:11:23] Adding `HIDDENCAT` to the category page will make that category a hidden category. [05:12:06] That won't remove it though; my theory is that a template has the category in its markup, causing it to show [05:12:24] The problem is it generates automatically when the language isn’t available [05:13:00] Just I used the {{native name|lang|text}} thing [05:13:00] https://meta.miraheze.org/wiki/Template:native_name [05:15:56] That may be what's causing the category to appear; the only issue is that all the template does is invoke a module, so the real culprit is a module that has to be edited to remove the category (which can break things if you're not accustomed to Lua) [05:17:16] There’s a way to import the languages like on Wikipedia? [05:18:24] https://furrnations.miraheze.org/wiki/Module:Lang - At line 1013 is where the module inserts the category; in theory, all you have to do is remove that line, but I don't really know much about Lua (the good thing is that any breaking changes can easily be reverted) [05:18:35] I'm not quite sure what you mean by that [05:27:15] Same issue as mine [05:29:31] I presume it's a server side issue then(?) [05:30:51] Most likely; I myself have received a few errors in the past hour, but it appears to have gone back to normal-ish [05:31:20] fair enough [05:31:58] I've only gotten the bad gateway error once, but pages are still somewhat slow— I'd understand if it's my poor optimisation, since most pages use multiple chained templates, and I don't know how much that could impact performance [05:33:37] It's causing the math extension to not work properly on the wiki [05:33:51] And sometimes bad gateway errors [05:35:10] wait so my chained template stuff is what's causing the occasional bad gateway error? [05:36:55] Probably not; bad gateway errors are almost always server-side (trying to import something that's too large is the only client-side cause of bad gateways that I know of) [05:37:45] ahh alrighty [05:38:20] thank you for answering- I'm normally really scared to ask questions like that due to my limited knowledge on computers [05:40:18] advice - don't import from wikipedia, their templates do a lot of stuff the majority of small wikis don't need, it's very redundant [05:41:53] out of nowhere question but- is there a way to get some form of custom language to be recognised within my wiki? [05:42:20] I haven't looked through the menus for that yet because I've been working on other things but I have been wondering, since my wiki is about fictional languages [05:42:55] the content of the wiki, perhaps, but interface if mediawiki itself - questionable [05:43:31] mediawiki localisation is based on translatewiki, it's like a separate project [05:43:32] fair enough- I'm not sure if I really need the whole interface translated considering that'll probably be way too much [05:43:39] I see [05:43:40] makes sense [05:44:27] if you need like custom scripts I guess custom fonts is the way to go [05:44:45] nah just Latin text should do [05:44:59] though looking into fonts might also be an option for further reference [05:53:39] 502 again? [05:54:06] yeah it's not loading for me either [05:54:07] Nevermind, thought it was another outage haha [05:54:18] it's just being really slow [05:54:21] took a while for it to load [05:54:34] It is being slow [05:54:50] I thought it was my internet [06:14:14] Is there a way to use visual editor on mobile view? [06:14:59] it's just there, unless it decided to be buggy [06:15:25] What I get is a code editor [06:15:25] actually scratch that, I never used VE in mobile frontend lol [06:15:59] maybe it's for the better lol [06:16:02] The problem is I don't know a way to efficiently make code edits with the code editor [06:16:24] I don't know exactly if the code editor itself is hard to use or this is my problem [06:16:25] I don't use MFE either so my wikis w/ VE have no issues on mobile [06:17:02] So previously I changed the default screen for mobile to be Vector-2022 [06:17:13] idk man, source editing pretty much has no cons [06:17:23] I also wanted to get the same experience with desktop [06:17:31] visual editor constantly messes up the code/formattings [06:18:13] just disable mobile frontend extension and set vector 2022 as default skin, it's mobile adaptive after all [06:18:54] The problem I'm facing is, how to efficiently insert code blocks including page display title, category, etc. to the source code without manual inputs [06:19:05] copy paste? [06:19:08] May I know how to do this thing? [06:19:23] If on mobile, it's relatively difficult for me [06:19:29] I edit on my phone quite often, and in source [06:19:37] oh well [06:20:11] For example to invoke the copy/paste menu on Android browsers I need to click on the selection drop while being in single mode (the one with cursoe) [06:20:15] cursor* [06:20:30] And wait for a second or two until it appears [06:20:40] [1/3] that's just it [06:20:40] [2/3] go to extensions admin menu, disable mobile frontend [06:20:41] [3/3] then go to additional wiki settings, styling tab, and check if vector 2022 is default skin, it's the first setting in the list [06:20:56] Thanks. But may I know what does this extension do? [06:21:19] Like to create an responsive view for mobile browsers for those non-responsive skins? [06:21:22] mobile frontend? it creates that mobile/desktop mode switch [06:21:48] So this is what this extension actually do? [06:22:07] which means you can use two skins for two modes but it has limitations [06:22:42] I know, so yes I don't need it [06:22:58] there are several skins which are mobile responsive by themselves, like Timeless, and thus don't need mobile mode -> no problems w/ styling and additional settings [06:23:37] Ohh [06:24:39] mobile mode comes w/ Minerva skin by default, which is very very nerfed, and it throws away all styling and js from Common and skin specific pages, which means if you want to keep the same look you'll have to copy paste same code to Mobile.css/js [06:25:07] and from my experience it's slow [06:25:22] there's very noticeable flash when the page loads in mobile mode [06:25:37] so yeah, I just use mobile adaptive skins [06:25:43] So what will happen if disabling this extension [06:26:06] Like only the desktop mode will apply to both desktop and mobile browsers? [06:26:20] you won't have mobile/desktop switch anymore, just one skin [06:26:42] pretty much, but if the skin is mobile adaptive it will look good on mobile [06:26:55] imitating mobile mode [06:26:58] I know the skin will be unique, but my question is will settings for desktop only apply to mobile browsers [06:27:36] For example I can't use the visual editor on mobile view, despite manually added ?veaction=edit on the address bar [06:27:36] [1/2] yes [06:27:36] [2/2] you have one skin, browser has no other choices and will use settings/styling of that skj [06:27:48] Oh [06:27:50] I know [06:27:51] you'll be able to use VE [06:28:16] https://cdn.discordapp.com/attachments/407537962553966603/1244537640460619836/image.png?ex=6655797f&is=665427ff&hm=796c07de804b3a93bd6b323179519356c8d57fcf47f5ee56203f002329984b0f& [06:28:27] https://cdn.discordapp.com/attachments/407537962553966603/1244537685712834560/image.png?ex=6655798a&is=6654280a&hm=dfa3783a6a147f302d3ace30ababa4b584a3bf1f8e6a851b43fa6825e2e31b16& [06:28:36] The above one is while in mobile mode [06:28:43] The below one is while in desktop mode [06:29:38] [1/2] https://cdn.discordapp.com/attachments/407537962553966603/1244537982900240504/IMG_20240527_092859.jpg?ex=665579d1&is=66542851&hm=060b9384fc9363c6fd5e7cefa9be83352d34be578ffd25a978365c2ff36cbeb7& [06:29:38] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1244537983281926196/IMG_20240527_092924.jpg?ex=665579d1&is=66542851&hm=3b20f4edd8004fc17e60ccfb456aaf05316c7aa2fa2d95f425ab664cf7a4001b& [06:30:00] Timeless, no mobile mode, on my phone, VE enabled [06:30:09] My wiki on mobile mode will only give you an option to edit codes [06:30:37] I'm telling you - try disable mobile frontend [06:30:38] I know, this screen is more like FANDOM's default [06:31:06] I showed you visual editor on my phone [06:31:12] no mobile mode [06:31:18] it works [06:31:27] no switching to desktop or anything [06:31:31] it's just one skin [06:31:54] My question is does this extension affects this behaviour, like under a same skin being in different modes will affect some functionalities [06:33:00] I don't know which other extensions don't work w/ mobile frontend [06:33:38] whatever works on normal desktop view will work if you'll disable mobile frontend [06:33:40] Ohh okay, I will check with documentations then [06:34:58] I checked with the documentation, as it says this creates a completely separate site for mobile browsers [06:35:16] So this suggests that the extension can dictate which extensions should be used for mobile views [06:36:18] And this controls whether the visual editor is available or not, so yes simply disabling it will result in a unique configuration for desktop and mobile browsers [06:42:19] i just had a 502 error- it went away instantly but I had flashbacks [06:59:05] Thanks Legroon for the help, I just disabled this extension as it makes the visual editor unsuable [07:24:23] sometimes 502 [07:25:04] yeah I'm getting the 502 again myself [07:25:07] oh its back now [07:29:46] slowness as well, but it's like in old days lol [07:29:52] certainly not like yesterday [08:09:59] Again, where to submit config changes request? [08:10:56] wiki is slow today [08:11:19] I experence no slowness at all, possibly because I'm already using a slow connection [08:19:21] [[phorge]] [08:19:21] https://meta.miraheze.org/wiki/phorge [08:19:23] [08:24:33] Thanks [09:27:56] My wiki is being vandalized by people who hates me [09:28:10] And thinks that I’m a pedo [09:35:18] #cvt for vandalism report and help [12:44:56] <_______________________________d> miraheze is slow again!!!! [12:45:15] <_______________________________d, replying to sovieticdragon> enable the moderation extension in [[Special:ManageWiki/extensions]] [12:45:15] https://meta.miraheze.org/wiki/Special:ManageWiki/extensions [12:45:19] [12:50:11] I enabled it [13:35:48] quick Q - I've got a graph I made in Excel and want to put on my wiki. the easiest way would be to just make it into an image and upload that, right? [13:36:07] Likely [13:38:33] alright, thank you 👍 [14:05:57] how long till a link colour change in the css goes into effect? I've purged the css page twice now and my intention is to make external links match the in-wiki links (on the timeless skin specifically) but they're currently still blue. [14:08:28] You may need to purge the page it appears on [14:10:26] I did do that as well, but I'll purge, refresh then purge again [14:11:34] hi everyone [14:11:44] how do I import images?? [14:11:57] last time I tried to do that I made a request and it ended up being ignored [14:12:07] What’s the task [14:12:13] are you sure the code is correct? [14:12:18] give me a second [14:12:23] external links are a bit fubky [14:12:28] Purging on the browser or Miraheze [14:12:47] https://issue-tracker.miraheze.org/T11480 [14:13:15] [1/4] >>> Hi, [14:13:15] [2/4] Since no response to the latest question has been received in a while, we must assume that this task is no longer needed and will now be closing it as Declined. (This is simply to indicate to us that the task was not done) [14:13:16] [3/4] If you still need this task and are able to respond, please feel free to reopen this task and indicate that it is still needed and we will start working on it. I would like to apologize once again for any inconvenience caused by this task not being resolved within a reasonable time. [14:13:16] [4/4] Reception123 [14:13:22] maybe I am wrong but images did not get imported [14:13:27] Well there you are [14:13:28] yeah I am wrong [14:14:03] do I really need to make a new task?? [14:14:30] [1/3] https://herotvdatabase.miraheze.org/wiki/MediaWiki:Timeless.css [14:14:30] [2/3] Default skin is currently Cosmos though since I wanna get this link colour stuff sorted out before I make it the default. [14:14:30] [3/3] I think I put everything in correctly here. I used the purge command on the pages each time. [14:14:58] [1/2] Link colour examples rn [14:14:58] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1244655088832614440/image.png?ex=6655e6e1&is=66549561&hm=2eff137c8dac1acfd13e88fd39bdd7b30a243cc8b3496c6da35fccc7d7248bd8& [14:15:43] As explicitly said in the response, no [14:15:56] You are free to simply mark the task as open [14:16:11] okay [14:28:41] performance should be improving significantly [14:33:22] please direct any complaints to OpenAi [14:33:50] With the upmost pleasure [15:10:05] Cc @suzuneu https://issue-tracker.miraheze.org/T12161#243593 [15:42:51] [1/2] a while ago I came here and asked a question and then this morning I still didn't have a working solution so I went back here but searched in chat before sending my question again, and it turned out someone happened to have provided the exact answer I needed in the past day or however long since I asked, and it worked [15:42:51] [2/2] I love this server [15:43:06] we all have the same problems or have had in the past so we can all help each other [16:31:19] Except anything related to dark mode, apparently. Everyone wants to use dark mode but hates working with it [16:33:23] [1/2] "Where do I put the '.mw-no-invert' part to make stuff not invert" and you have 15 people replying "idk" [16:33:23] [2/2] Currently my fight with dark mode is why it's reverting the link colours to default in timeless [16:40:00] @freosam might know [16:40:16] I'm sure DarkMode is ex comm tech [16:41:01] https://dev.miraheze.org/wiki/Disable_Dark_Mode_invert [16:41:48] Or @canyada, thanks! [16:46:46] Testing that in inspect style edtior and that just undoes the entire dark mode (if you put in everything, anyway) [16:50:52] YES I FIGURED OUT HOW TO FIX THE HUE SHIFT FOR SOME IMAGES IT WAS SCREWING WITH [16:51:10] BUT AT A COST [16:52:32] [1/6] if u just want links to not invert try this in the js? [16:52:33] [2/6] ``` [16:52:33] [3/6] $(document).ready(function{ [16:52:33] [4/6] $("a").addClass("mw-no-invert"); [16:52:34] [5/6] }); [16:52:34] [6/6] ``` [17:19:03] gotta stop staying up this late I tried putting this in the style sheet inspect element as a test [17:19:25] could try pasting it in the console [17:19:36] that tab in the inspect element thats mad all the time [17:21:07] hello, I want to collapse by default a summary. How can be done that in one page? [17:26:53] [1/3] ```
[17:26:53] [2/3] text [17:26:53] [3/3] /div``` in the most basic form [17:27:18] but I mean the standar one, although that code will be usefull too thank you! [17:27:43] not sure what that means but yw [17:27:45] like here: https://tuscriaturas.miraheze.org/wiki/Juegos:C%C3%B3dice_de_Ashenbach [17:28:12] when you press "ocultar", that means hide [17:28:22] that appears like that by default but just in that page [17:28:46] oh do you mean the table of contents? [17:28:49] yes! [17:30:26] [1/2] the more I look at dark mode through inspect element, the more I'm like "Uh-huh. Here's the problem: too many toasters" [17:30:26] [2/2] And by toasters I mean effects. [17:32:22] Dark modes need to prevent certain things from inverting like images, but in the process things like hue can get altered or inverted anyway. Most dark modes people want are just "I want white text on black bg and a toggle to remove that", but alas, it cannot be that simple... [17:34:06] not sure if there's a straightforward way to do that but [here]()'s a workaround i found that hopefully works for you [17:35:15] instead of MediaWiki:Vector.css you can use MediaWiki:Common.css [17:36:12] which is why its probably easier to get rid of what the dark mode does entirely and set the colors yourself! [17:37:37] there's themetoggle extension which is practically make your own dark mode, also other themes if you wanna, but it was declined from miraheze [17:37:40] [1/2] Thing is I need a toggle between light and dark. I can't just set one colour. [17:37:41] [2/2] I've had some chats with people with vision issues and I wish i was joking when I said half prefer the standard black on white, the other half prefer the reverse. [17:38:19] make dark grey on light orange [17:38:23] well yeah thats what this does, get rid of the inherent dark mode invert so u can set colors yourself [17:41:20] Ahh right. Gotta figure out how to make the bg black and the text white, then [17:41:35] also it's past 3am this is worse than last night I cant keep doing this [17:42:20] feel ya [17:42:26] go sleep [17:43:37] [1/5] but I did find something interesting: [17:43:37] [2/5] In inpect style sheet, the code recognises these links need to be changed but...it doesn't change them. [17:43:38] [3/5] why are external links so damn stubborn... [17:43:38] [4/5] https://cdn.discordapp.com/attachments/407537962553966603/1244707596464689211/image.png?ex=665617c8&is=6654c648&hm=0ffb0e18f0c1fa6f177aefab3251def6cc44fcf5c919d3020f55b76113926e2d& [17:43:38] [5/5] https://cdn.discordapp.com/attachments/407537962553966603/1244707596833919077/image.png?ex=665617c8&is=6654c648&hm=a88f71b4f09cb4f8af1e1fc33719b6c103b2bd77cc1961540a7ed67ca1d4ecbb& [17:46:41] `!important` [17:50:05] I did try that [17:51:06] I can relate yes, but I don't think we have to many toasters [17:56:22] [1/6] the. [17:56:22] [2/6] THE MOMENT> [17:56:22] [3/6] I LEAVE THE WINDOW TO [17:56:22] [4/6] THEY CHANGE [17:56:23] [5/6] I AM GONNA KILL THESE LINKS [17:56:23] [6/6] *conks out on laptop and dies [18:11:55] Hellooo [18:12:41] I have a really quick question regarding AI generated content on wikis. I'm using chatGPT to come up with new ideas or writing points to be expanded upon later and was wondering if this is ok or not [18:13:55] we've got a disclaimer here if it helps https://grimtown.miraheze.org/wiki/Disclaimer#Ethical_use_of_generative_AI_and_chatGPT [18:15:08] @luasucks just a note, connecting to Miraheze from OpenAI won't work so it won't be able to read content from Miraheze [18:15:23] Unless you copy it into a prompt [18:15:36] Thats ok, I usually supply relevant context into the prompt [18:16:01] I'm just wondering if theres any miraheze policies on it [18:16:06] Nope [18:18:24] Cool, just in case I'm going to continue using chatGPT and rewrite its output or indicate if it is used directly [18:19:02] that sounds good [18:19:16] There's no real test yet on GenAI and copyright law afaik [18:19:50] @owenrb / @serverlessharej might know different [18:21:21] If the community wants a policy on AI then that’s reasonable but I’m not sure there’s a specific law to target. We’re in the regulatory uncertainty stage. [18:22:23] I don’t think any Miraheze global policy necessarily disallows the use of generative AI tools. It may make more sense on some wikis than others. [18:22:38] We are in the community uncertainty stage too [18:22:53] Our policies have mostly been reactive [18:23:05] Most policies are 😁 [18:23:29] @notaracham: maybe you could add GenAI to the next #meetings [18:23:56] There is no active policies about AI. However you do have to remember that LLMs (such as GPT-3.5 and GPT-4 that power OpenAI's ChatGPT) regergitate the prompt and their output may be untrue and must be caarefully checked [18:24:53] The tests we did with CodeRabbit went awfully [18:27:32] I use ChatGPT a lot (in fact I have purchased licenses for the four members in the org, it's useful) but at the same time I am absoulutely agaisnt AI generated code in any codebases I oversee. It's a great learning tool and it's pretty good to use to bounce ideas off and to automate some tasks but not much more. [18:28:01] Yeah that makes sense. I'm purely just using the outputs as inspiration and for expanding on points later. Its quite good at suggesting what to write next or to provide a base layer to be rewritten. I forgot to mention that I mostly use it for creative writing [18:29:13] Yeah, my wiki is mostly about creative writing [18:29:39] It could not produce useful analysis of code in many situations, some of its suggestions were just wrong. It also had a pretty awful UX. [18:30:49] If you use Lua or Java it's suprisingly good, but if you throw something like C or Lisp (famous for it's snytax) at it, you are likely going to very bad results [18:31:22] Thats one thing I'll never use it for, which is code. I mostly know how to code and I'm familiar with documentation [18:31:52] PHP was god awful [18:32:35] Even its suggestions that were kinda acceptable were nothing static analysis couldn't pick up [18:33:19] It suggested things like SQLis that many tools can find but missed the fact that the input was trusted [18:33:43] It also had some weird stuff where it utterly misunderstood code being moved around [18:33:59] And hallucinated the fact that parameters were not what it thought [18:34:05] Then argued with me about it [18:34:06] Yeah I remember trying to show it some python and it got very confused [18:42:55] If I can't get to the meeting I would like to put the suggestion out there that we provide a managewiki option that adds the block chatgpt thing to robots.txt [18:43:40] Uh how do I put this [18:43:46] too little too late :) [18:44:06] What else can you do against AI lmao [18:44:16] Oh no [18:44:27] i mean we already pummeled it into submission [18:44:34] Uh [18:44:38] Idk wym but ok [18:44:51] OpenAI thought it would be really funny to account for 20% of requests to the site [18:45:06] so rhinos said fuck it and blocked its UA [18:45:14] Oh LMAO [18:45:37] lmaooooo [18:45:43] nother one bites the dust [18:46:09] Well there are other AI user agents too [18:46:16] Fair [18:48:43] It is not going in robots.txt [18:48:47] It is way too high [18:49:08] It can't be delegated to wiki level [18:49:31] There can be wiki specific robots.txt [18:49:38] There is now [18:49:49] I mean the wiki specific one [18:50:06] Unless you want all wikis down then we can't allow ChatGPT [18:50:21] Oh ik [18:50:25] But there are other AI [18:50:29] It's request rate is far too excessive [18:50:37] Other AI can do what they like [18:50:45] But OpenAI is not going near us [18:50:54] Some wikis might want to block all AI though [18:50:56] 12 million requests a day is not a risk I'm taking [18:51:01] Then they can do that now [18:51:09] You can edit robots.txt [18:51:24] But I'm not adding OpenAI too it, it's staying blocked at the edge [18:51:33] Sigh you are missing the point [18:51:36] about ~2 million of those were not caches and were reaching our mw servers [18:51:38] I never said unblock OpenAI [18:51:56] Most people don't know that [18:52:04] https://discord.com/channels/407504499280707585/407537962553966603/1244722519420960828 [18:52:13] Then teach them that [18:53:33] Which is about a 16% increase [18:53:45] Ok so where should we be telling people "add this to robots.txt to block AI from training on your wiki" and why is that superior to a simple managewiki option [18:54:11] When there are already like a hundred managewiki options that can be done through a weird css or js back [18:54:35] I meant blocking AI in general [18:54:49] How do you block AI in general? [18:56:10] You came up with adding OpenAI to robots.txt, I've said no. Where's the magic block all AI? [18:57:16] is it possible in the first place? if yes, I'd like to protect the wiki, and it sure would be nice to have as familiar managewiki option [18:57:29] No [18:57:49] here you go [18:58:04] [1/3] ... Most major AI scrapers respect robots.txt, including Google Bard and others [18:58:04] [2/3] https://www.eff.org/deeplinks/2023/12/no-robotstxt-how-ask-chatgpt-and-google-bard-not-use-your-website-training [18:58:05] [3/3] https://www.cyberciti.biz/web-developer/block-openai-bard-bing-ai-crawler-bots-using-robots-txt-file/ [18:58:08] Misinformation [18:58:26] [1/2] Stop talking about things you don't know about [18:58:26] [2/2] "I haven't heard of it" ≠ "It doesn't exist" [18:58:35] That's not all AI [18:58:44] Thanks captain obvious [18:58:45] That's blocking individual AI [18:58:48] Yes [18:58:49] collei [18:58:51] dude [18:58:52] Most major AI [18:58:55] And we are not adding ChatGPT [18:59:17] If you have a problem, you can escalate it to the director of the technology team [18:59:34] [1/3] Quit obsessing over that [18:59:35] [2/3] You don't need to add chatgpt to robots.txt [18:59:35] [3/3] I said probably 10 times that's not what I meant [18:59:45] Let's remember to be civil and respectful here @koreirose [18:59:54] Then linked me an article that talks about adding it to robots.txt [19:00:16] If you want to block other AI, you can use Mediawiki:Robots.txt [19:00:55] But any scraper that is likely to or has degraded the experience of users as a whole will get blocked. [19:01:37] Sigh [19:02:29] Rhinos is trying to "correct" me on something I misspoke about, I don't think anyone thinks you can add something to your wiki to block every AI that exists or ever will exist, I didn't mean to say to unblock ChatGPT [19:02:36] > Ok so where should we be telling people "add this to robots.txt to block AI from training on your wiki" and why is that superior to a simple managewiki option [19:02:46] We have dev wiki [19:02:49] Or metawiki [19:03:01] I never knew [19:03:03] Please feel free to create documentation on either for blocking AI [19:03:08] Ok [19:03:18] It's a wiki. Anyone can edit. [19:04:35] is MediaWiki:Robots.txt empty by default? [19:04:57] @theoneandonlylegroom the custom one is, we add our own rules on top of whatever you put in there [19:05:11] The code should add it on automatically [19:05:40] I just tried to check on mw.org and it only has comment line [19:05:47] https://github.com/miraheze/mw-config/blob/master/robots.txt are our global rules [19:09:23] Although Bytespider is ignoring that [19:09:42] The hilarity that comes with assuming that I’m going to ban you for being uncivil….and then immediately being uncivil and assuming bad faith. [19:09:59] It's made around 54M requests last 24 hours [19:10:13] Anyway, Reception123 has warned you and I strongly suggest you adhere to that. All these conversations can be kept civil. [19:10:19] yikes [19:10:39] That's perfectly acceptable [19:10:39] what if we made an AI to counter the AI :ThinkingHardMH: [19:10:50] that seems like a lot lol [19:11:13] Wait [19:11:17] Can I add up [19:11:40] I'm tried [19:11:41] Tired [19:11:47] 54M is total of everyone [19:11:52] It's 112K [19:12:16] The ideal 86.4K a day [19:12:34] But 112K is acceptable [19:13:04] [1/2] which category at Miraheze should I make the wiki for my Interesting Facts? [19:13:04] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1244730108539834388/Combustos_Interesting_Facts_40.png?ex=66562cbf&is=6654db3f&hm=71e52c305026de59b1d26d5f51aeb82f1a0c47f577148229821b3319d27605f6& [19:13:30] Google and Yandex bot are about 200K [19:13:34] I am happy with that [19:15:31] Our highest individual IP is 208K [19:15:41] Which is someone scraping static earlier today [19:15:54] _doesn't really care about static much though_ [19:17:26] Then I will put it in Uncategorized [19:19:14] my Interesting Facts are safe in as no porn facts ok? [19:20:32] yes [19:20:45] Ok sorry I got a bit too angry [19:21:27] I'm angry too [19:22:56] why [19:25:36] I would appreciate an apology [19:26:17] Sorry [19:27:51] it's long story. it's just some people at Deviantart think I'm autistic, but I'm not! so I delete my account from there for good. it's why I decided to make wikis for my Interesting Facts and others like SoulCalibur. I promise I will update my wikis often this time. [19:29:35] Whether you're autistic or not, there is no shame in it. [19:29:58] I am in the Autism spectrum and I accept it as a value [19:30:38] If they think you're autistic... then they assume things they do not know [19:30:44] Ignore that [19:30:56] Coming from a probably autistic person, that sounds like something many of my neurodiverse friends would say [19:31:16] And a fair number of our users and volunteers are neurodiverse [19:31:21] I would like to say that I'm standing for that [19:31:36] <.labster> Technically everyone is on the autism spectrum because a spectrum includes all values. [19:31:41] @everyobe shut up! I'm not autistic!! [19:31:56] Please don't use @everyone although it won't work anyway [19:32:02] I didn't say you were [19:32:25] If they think you're autistic... then they assume things they do not know. Just ignore them [19:32:29] Well done [19:33:02] <.labster, replying to rhinosf1> believe it or not I'm just quoting [[w:Larry Wall]] [19:33:02] https://en.wikipedia.org/wiki/Larry_Wall [19:33:03] [19:33:03] AI can get general grasp but it for sure cannot get the fine details, which in computer programming, is where most of the key details are. [19:33:44] Well put [19:35:24] +1 hugs for my fellow autists. [19:35:25] :pupCoffeeMH: [19:36:22] [1/2] there goes my new wikis then people will make fun of my things like are you guys going to say this "he looks like he saw Mommy kiss Santa Claus" but he does not! [19:36:23] [2/2] https://cdn.discordapp.com/attachments/407537962553966603/1244735974219448341/Combustos_Interesting_Facts_37.png?ex=66563236&is=6654e0b6&hm=e0a1012a23905d4eaa918426e0cebdcccde982584bf6f6dae29bd288104edbe2& [19:36:33] what [19:36:57] I can compare it to asking a thirteen year old to read a academic paper and describe what they read, they will get the general idea but the specifics will go over their head as they are likely to lack the academic and practical knowledge and skills to understand, interpret and work that information. [19:38:05] @felenov I just turned 42 years old a month ago (April 6) [19:38:45] he's talking about AI, different conversation [19:42:48] I don't remember how to disable comments from each pages in my wikis [19:43:00] sorry I forget [19:43:55] [1/2] if you mean talk pages you can't disable them but can at least hide the tab [19:43:55] [2/2] of you mean actual comments under articles that's most likely some extension that you can disable in admin settings [19:45:05] Steven Universe wiki has comments disabled because of the spam [19:46:33] [1/2] yeah, I mean comments isn't default feature of any wiki, it's an extension, an addition [19:46:33] [2/2] on Miraheze wikis as well, you can just disable this addition [19:48:01] about the deaf kid in my facts actually happens in Feb 1974, not Dec. because December is Christmas month right? [19:48:58] Does Miraheze have a collection of quotes from chats. IIRC WMF and GNU have those [19:49:25] oh no even Disqus? [19:49:27] there was a link somewhere [19:49:39] discussed kinda recently [19:49:57] but it was randomised? [19:50:01] https://en.wikipedia.org/wiki/Disqus [19:50:36] https://bash.toolforge.org/top Found it. [19:50:43] Your discussion rather interrupted the discussion of Artificial Intelligence [19:51:35] https://cdn.discordapp.com/attachments/407537962553966603/1244739801102422167/image.png?ex=665635c6&is=6654e446&hm=0d250a5adc499a18292106c186c30b029dcf02f4ef35cdc10e98ea102ba7b04c& [19:51:37] That is my favourite tool [19:51:42] Sorry, my bad. Thought you meant "not even interested discussing the boy?" [19:52:18] Why? [19:52:53] I misread your comment [20:42:40] so, captchas basically [20:43:07] lol [20:57:49] I mean we do that [20:58:33] We serve 180,000 bot challenges a day [21:00:59] The majority of our more loose cloudflare protections are challenges [21:02:05] Excluding ChatGPT, about 14% of our cloudflare efforts are using managed challenge [21:02:27] About 50% are blocks on individual IPs [21:03:47] Another 40% is various other UAs of scrapers blocked [21:03:55] I would say blocking from CF side is too much but a challenge with a strict filter is enough to handle the issue [21:04:17] The remaining being banned is vuln scraping where I haven't whacked the IP yet [21:04:41] Very very few users who get a challenge pass [21:05:26] We use the challenges for pretty much anything where there's a change it's not 99.99% malicious traffic [21:06:04] Our friends from captcha farms have already found a way to crack Cloudflare turnstile, but they charge a hefty premium for that so I would call that effective. [21:06:43] The challenge is pretty effective [21:07:00] They are very few requests get past it [21:08:04] One of the ways I have to fuck with scrapers is return random codes. If a user gets the page requested with a 404 or 502 code they will not notice, a scraper will since they mostly operate on codes. So I serve everything with randomized codes, inc. 404s, 200 and 500 codes, anything goes. [21:08:37] Blocks and failed challenges return a 403 [21:09:03] I ain't really bothered if they are wasting Cloudflare's resources [21:10:57] We served 43 million requests in the last day from cloudflare [21:11:01] I would just randomize the codes, even if a scraper gets what is needed, most of the scrapers are run by skids who don't know how to get around that. Simple solution server-side that is relatively low effort for admins, filters the 90% of attackers and highlights the remaining 10% since they serve more of a threat than [21:12:07] The latest attack they mitigated was 70 million a second [21:13:12] Why do I need to care though [21:13:23] Cloudflare can handle our daily traffic in a second [21:13:29] They can keep hitting it [21:15:25] not very nice to legit tools that wanna read a page unless you wanna lock down fully [21:15:25] [1/2] I find it funnier to mess with those who operate scrapers rather then to just lock them out, since if they know what locks them out they will eventually get aound with with enough determination, if they are being fucked with it could be a long time before they realize. [21:15:25] [2/2] Those who come to just drop the site, yea, those can do the challenge [21:16:53] I'm not really interested in messing with them [21:17:01] Just that they don't impact the usability of the site [21:22:55] funny is all well and good but not operationally necessary as miraheze goes [21:57:41] I am 100% sure it can be bypassed because it appears to just be a client-side check but it's highly obfuscated so I'd need to put a lot of time into trying to crack it [22:07:36] wonders if "bypass" is @koreirose ´s middle name :ThinkerMH: [22:08:14] /me wonders if "bypass" is @koreirose ´s middle name :ThinkerMH: [23:34:56] Are there plans to have TSPortal translated?