[02:00:00] If I go to "index.php/MediaWiki:Common.css", I see the current css, but there's no Edit link. I've been running a private wiki for years, with only 1 user. I've double checked it is still in the group Administrators. If I go to "View history", I can see my edits in the past, the most recent of which was on 1.31.1. I've since upgraded to 1.32.0. I am on php7.3, as I thought 1.32.0 was ready for it. [02:01:48] Ahh, if I go to View Source it explains editing CSS was separated from editinterface. Should that have been auto-added to the single administrator user, or is it intended for everyone to fail at this and have to manually give it to everyone? (I did run the update.php.) [02:02:23] yep, it got separated in 1.32 [02:02:38] easiest thing to do is add yourself to the css editors group [02:02:42] (if you have bureaucrat) [02:02:43] jamespharvey20: ^ [02:03:05] or whatever the group is called, should be pretty obvious when you open up Special:UserRights [02:03:27] it is intended that the ability to edit css/js pages is *not* handed out to every existing admin [02:03:43] OK, sounds good. Thanks! [02:04:37] Skizzerz: That did the trick. :-) [02:44:48] jamespharvey20: And yes, it's intentionally not granted to the default wiki user because ideally people wouldn't ever need to use those rights, and shouldn't without understanding the consequences. [03:17:42] If anyone who has expertise in Cargo can help me, I'm trying to figure out how lists work. I know you can query whether a list contains an item with HOLDS, and I know how to store a list by inserting it with the proper delimiter. [03:18:13] But how do you parse each element of the list separately? [03:19:48] Like, with a full cargo query you can use format=template to call a template for each cargo result, how do I call a template for each list item for each cargo result? [04:56:56] Found the answer to my question, it appears I need to use array functions like #arraymap [06:12:06] James_F: interface-admin is granted to the default user on new installs, I'm not quite sure how updating of existing wikis could be handled though, as I don't think there's a way to tell whether the user is updating from a code version which did not have ifadmin yet [06:14:38] the error message links to https://www.mediawiki.org/wiki/MediaWiki_1.32/interface-admin so I think it's not too hard to figure out what's going on [06:52:56] biberao, I put it in ~/public_html/wiki/ [14:51:51] Sveta: can you share your config if youre using nginx? [15:00:28] Even I'm in need of a config for nginx [15:20:01] shreyasminocha: subfolders? [17:50:33] hi, why there is 5000 (chars?) limit on api queries ? [17:52:34] make a POST request and put that stuff in the post body instead if you're making that large of a query [17:53:18] niso: where are you seeing the 5000 limit, out of curiousity [17:53:33] are you making huge queries (really really really long url strings) that are being rejected/truncated? [17:54:19] yep, I'm talking about the result just to make sure, not about my request length [17:54:32] that configurable? [17:55:26] oh [17:55:33] the result is based on your user group [17:55:35] I'm using cargo with table that contains 100k +- and i want to get the data out of mediawiki, I've try to dump the db but some of his column got collation "latin1_swedish_ci" which is hard to get the data out from there [17:55:40] it's configurable as a $wg variable on the wiki side [17:55:44] the default is 500, and for bots 5k [17:55:46] but you don't want to make it too high or you'll hit timeouts [17:56:03] it's for number of records, not number of characters [17:56:08] so 500 records or 5k records [17:56:21] that fast query, and i'm ok with the latency [17:59:01] in any case, the api lets you do paging [17:59:10] so you can use a bot to grab 5k at a time and iterate over the entire result set [18:00:24] yep, but why doing n queries if i can do one? i'm ok with it :) [18:00:26] niso: I spoke incorrectly, the 500/5000 limit is hardcoded [18:00:36] :( [18:01:26] I'm not sure why it's hard to get data out of a latin1_swedish_ci column though... [18:01:54] i think that should be utf-8 (as the other tables in MW) [18:02:15] depends on when it was installed [18:02:34] and what the extension set (if it's an extension table), some aren't well-behaved when it comes to that kind of thing [18:02:43] that cargo [18:03:09] in any case, it'll work with SELECT queries just as well as any other collation, and whatever client is being used to query will process it correctly [18:03:30] or did you try that already and ran into issues? [18:03:52] ofcourse, I'm talking about real problems lol ^_^ [18:04:03] :) [18:04:09] what sorts of problems did you get? [18:05:42] you can override the collation on a per-query basis with the COLLATE keyword -- https://dev.mysql.com/doc/refman/5.7/en/charset-collate.html [18:05:49] I've talk to yaron (cargo guy), and he suggest to use Special:ViewData, which seems to be better solution the problem (crash my server with more than 20k rows limit) - the api problem [18:05:52] or you can alter the table itself to use a different collation [18:06:12] I've got this data and try to get the original one: "\u00d7\u017e\u00d7\u00a9\u00d7\u2014\u00d7\u00a7" [18:12:44] some of this can be parsed with " bytes(data, "cp1252").decode("utf8") - python 3, but some other can not: "\u00d7\u0090\u00d7\u0153\u00d7\u2122\u00d7\u00a2\u00d7\u2013\u00d7\u00a8 \u00d7\u00a9\u00d7\u00a4\u00d7\u2122\u00d7\u2019\u00d7\u0153" [18:12:57] that why i preferred do get this data in another way [18:49:55] Skizzerz: i pasted the config [18:50:00] let me see if i can find it again [18:50:42] Skizzerz: -> https://pastebin.com/7LVmPepq [18:53:43] it looks like you edited more than I told you to [18:53:52] I said redact passwords and secret keys, and *leave everything else as-is* [18:54:47] in any case, nothing looks wrong with it [18:55:00] so either you obscured the issue in your edits, or your problem lies elsewhere [18:55:04] most likely it's the latter [18:56:12] yes [18:56:20] i think the samei dont understand whats happening [18:56:53] the rest of the things is custom for remote authentication [18:58:42] you say that using $wgDBserver works fine [18:58:50] since you only have one server anyway, why not just switch to that? [18:59:07] im doing that [18:59:18] but isnt it weird to have it working and then it doesnt work? [18:59:18] lO