[07:01:02] Hi [07:01:30] How so they resert login credential on toolforge? [07:02:53] Ive created an LDAP account but when I try to login it says the user/passw is not correct [07:07:17] So, I want to recreate a new account but it says the ```Wikimedia unified account is already in use. ``` [07:55:29] hi all! What should I add to $wgWhitelistReadRegexp in order to use api.php? Right now I got "readapidenied" in response to api.php?action=query&format=json&prop=info with $wgRevokePermissions['guest']['read'] = true; and a single rule in $wgWhitelistReadRegexp (which works as expected, I can only read single page, but can't edit it in VisualEditor because of that error) [07:56:37] when I comment out #$wgRevokePermissions['guest']['read'] = true;, that request returns this: {"batchcomplete":""}, so it works, I believe [07:57:26] so, I presume, api.php somehow ignores whitelistening, right? [09:10:41] Hi all [09:13:00] I want to do a simple querry on petscan tool but after reading the man i can't still do it [09:13:08] here is want i want: [09:13:59] find Wikipedia articles that are related to a city (e.g. New York) and have fewer than 10 links in them [10:16:49] hello, i discovered today after upgrading to 1.30 a couple of weeks back that MobileFrontend is barfing [10:16:49] https://wiki.amahi.org/index.php?title=Main_Page&mobileaction=toggle_view_mobile [10:17:14] even though the version installed is supposed to be for 1.30 [10:18:22] looking for advice on how to fix this [10:21:17] hmm, just commenting that line seems to make things (mostly/all?) ork [10:21:19] work* [10:54:57] Hello there. I want to implement ldap/ad login for mediawiki. Which is the best way to go to implement this? Tried Simplesaml, but that isn't working very well for me or i don't have some settings the right way. [10:55:27] Is there someone here who can recommend me a spefic extension to use? [10:55:58] And if possible if there is an example config point me to that? I do not seem to be able to find it. [12:07:15] Hi! [12:07:56] hi bam_ [12:07:59] How do I do to download the json output of a PetsCap querry? [12:08:21] what's PetsCap? [12:09:24] I mean PetCan tool [12:09:48] This is my PSID is 3311599. [12:10:15] So I want to download it using JavaScript and print titles [12:24:50] https://meta.wikimedia.org/wiki/Talk:PetScan ? [12:26:57] Thank you for tihs link. [12:27:32] After I have checked, it seems as if it has no helpfull info for me [13:33:33] Hi [13:34:39] Hi andre__ . I'm gradually triaging the enormous i18n workboard. Is it possible to make it so that when you click the i18n tag, the workboard will be shown? [13:35:04] Currently it shows the board description, which is less useful. [13:35:09] I want to send an ajax request to PetScan tool of this PSID 3311599. How do I do it§ [13:36:28] aharoni: {{done}} [13:36:50] p858snake: thanks. Can you please tell me how? [13:37:22] Manage > Edit Menu > then click on the pin icon to change the default [13:38:36] p858snake: thanks again [16:52:10] Guys, anyone knows a freelancer that can install VisualEditor on my Wiki? [16:52:56] WikiBrazil, https://www.mediawiki.org/wiki/Professional_development_and_consulting [17:04:03] thanks andre! [17:13:46] andre, it's a little outdated, thought. i'll still try! ;-) [17:31:09] WikiBrazil: I'm on there and still get queries. I know that doesn't say much, but I know it is still being used. [17:33:20] hex, is VisualEditor something complicated to install? I have 3 wikis on the same Vultr server (one wiki being used, the other are brand new). [17:33:37] a heavy editor is asking me for it [17:34:20] WikiBrazil: it isn't too hard, but it isn't the same as a PHP app. Is this a dedicated server where you have root? [17:41:24] yes, it's on vultr [17:42:32] i'll actually get a new droplet to migrate the mediawikis there. separate them for wordpress sites [17:43:25] but everyone i talk about visual editor, said they couldnt install [17:54:01] WikiBrazil: it should work. How comfortable are you on the command line? [18:00:21] I'm fine with it. Although sometimes I have no clue on what it's doing. I'm an ti consultant with low knowledge of linux. [18:01:14] low = almost zero. i google when i need any command like search for a file... i often forget vi is for editing... etc... lol [18:01:36] does it take too much of server resources? [18:07:30] WikiBrazil: hard to say w/o knowing your set up. Is your server heavily used right now? [18:07:54] If not, you should add it to your current server and see what happens [18:08:05] (That's what I would do.) [18:11:55] the problem is that my server have a ok amout os users on normal dates... but sometimes i might get like 1000 online users. and than 30 the next day. [18:12:05] i get users on specific dates [18:12:56] If they are logged out users who just read (but dont edit) you should look into varnish cache [18:13:21] its very good at dealing with bursts of readers access pattern [18:14:16] i'll take a look. since i'm not a linux guy, i'm using runcloud to admin my server [18:14:47] it works with redis and varnish. i used to use varnish in the past [18:16:53] Do you have application (wgMainCacheType) caching setup (i.e. apcu or memcached caching. You can also use redis for this but its less common)? [18:17:22] As thats also quite important when dealing with large bursts of users [18:18:28] bawolff: I just discovered that BerkelyDB has an SQL front end. I hear it supports concurrency better than SQLite. Have you heard anything about it? [18:19:24] Not really. I know that mysql used to support a mode where it used berkelyDB behind the scenes [18:19:59] baw, i read that wikipedia is using redis. isnt that right? [18:20:17] Well, and then there is PHP support. I'm interested b/c WikiFundi is using SQLite. [18:20:25] WikiBrazil: Right [18:20:46] I was generally under the impression that berkelydb was primarily key/value store (like persistent memcached) [18:21:00] me too [18:21:06] wikiBrazil: only for sessions and job queue [18:21:12] then I saw a SO hint [18:21:20] and maybe lock manager, cant remember [18:21:26] and then this: https://packages.debian.org/stretch/libdb5.3-sql [18:22:01] generally wikipedia primarily uses memcached, with redis used in some specialized cases [18:22:18] https://packages.debian.org/stretch/db5.3-sql-util [18:22:19] although we are moving away from it for job queue afaik [18:22:30] away from Redis? [18:22:45] Isn't that the primary use case ATM? [18:23:46] Its appearently having scaling issues [18:24:00] interesting [18:25:03] I think they are replacing it with something called ChangeProp or something. I dont really know the details [18:26:25] https://www.mediawiki.org/wiki/Change_propagation [18:28:29] nice to know [18:29:37] WikiBrazil: what is your timeline for trying VE? [18:30:28] hexmode: anyways, if its just an sql api layered over a key/value db, i have trouble imagining it has good performance for range/other complex queries. I would also look carefully at what sort of consistency garuntees it offers [18:31:03] hexmode: I would definitely say moving to mariadb is probably the best option [18:31:18] bawolff_: On a RaspberryPi? [18:31:25] Oh [18:32:01] Well that probably has different trade offs :) [18:32:20] they are also have time issues b/c Pi don't have a battery backed clock [18:32:30] SQLite ? [18:32:39] so attributing edits to people... [18:32:49] Ivy: That's what they're using now. [18:33:21] How much concurrency can a RaspberryPi be getting? [18:33:22] Are they expecting a lot of different people editing at once? [18:33:25] Oracle is evidently taking over BDB development [18:33:34] For reads, SQLite should be fine. [18:33:56] if its low edit concurrency i would say sqlite in that case [18:34:04] WikiFundi takes a Pi to remote areas and teaches a class how to edit [18:34:12] I guess I'm a bit confused how you're high enough traffic that concurrency is an issue, but you're running on a RaspberryPi? [18:34:18] Oh, hmm. [18:34:22] so it is all offline and not too much concurrency7 [18:34:43] but.... it has VE [18:34:45] still, i guess a class of 30 people [18:34:48] Concurrency is only an issue for SQLite with writes, AFAIK. [18:35:01] hex, it's a basic website that i run alone. it has a banner, but just to pay for the server. so i dont actually have a timeline. i can do it or not... maybe i could just forget about it :-) [18:35:05] Though sometimes there are unintended writes. [18:35:16] WikiBrazil: so, VE can run on a RasberryPi, I think you'd be fine. [18:35:37] It has to be a RaspberryPi? There are other machines with better specs that could be remote. [18:35:38] hex, but it actually looks nice. i was also thinking of migrating to wordpress. but i still have my doubts [18:35:41] Like any basic laptop. [18:36:08] hex, niiice [18:36:54] Ivy: That is just what they use. Yeah, a laptop would work--and have a clock--but the cost is more [18:36:56] This is the sort of thing where i think one what have to actually go and measure how much contention there is due to sqlite table locking before making an informed decesion [18:37:38] I'm not familiar it with much beyond talking with the woman who runs WikiFundi about her frustrations. [18:37:58] unless contention is an actual issue, id stick to sqlite [18:38:07] Maybe they have a really compelling reason to use a Pi over a laptop [18:38:51] contention wasn't one of the issues she mentioned. Which doesn't mean it isn't an issue... [18:38:58] bawolff_: Same. [18:39:09] if it is an actual issue, id maybe look at other sql based dbs designed for embedded usage, and see if anything exists in that space that makes sense to use (and could be made to work in that space) [18:39:45] (Which is why I said "BDB has an SQL interface?") [18:40:26] also mariadb is quite configurable. Perhaps it could be made to work well in such an environment with a tweaked config. I dont know enough about it to say [18:40:47] Seems unlikely. [18:41:07] WikiBrazil: if the site is only edited by you, then WordPress may be better. If you want more people to contribute, then I'd stick with MW. [18:41:10] Even running on MariaDB on the same host as the application code can be a bit much... [18:41:26] I wouldn't put MariaDB on a RaspberryPi unless it was doing nothing else. And even then. [18:41:46] For the real time clock issue, i would probably solve that by requiring the instructor to enter time at boot before mediawiki starts [18:41:57] Right, exactly [18:42:19] Manually entering the time? Lordy. [18:42:26] It doesn't have a small battery? [18:42:44] https://learn.adafruit.com/adding-a-real-time-clock-to-raspberry-pi [18:43:05] Again, a cheap laptop has all this solved already... [18:43:24] I guess it depends how much your time is worth. [18:43:37] But trying to rebuild a server from the most basic hardware isn't cheap. [18:43:50] Lol at time puns [18:44:08] Heh. [18:44:32] Sure, but then you can build a lot more Pis easily once youy've done the leg work [18:44:55] Maybe. [18:45:17] Still, 10x cost for 1/2 the work might make sense. [18:45:33] (No clue what the actual ratios would be.) [18:45:49] Yeah. [18:46:02] Or if you dont care about actual time, all you really have to do is make sure it monotonically increases [18:46:07] I just mean when you're at the point that you can't keep time, in 2018, that seems like a weird path to be taking. [18:46:23] Since the most basic digital watch dates back like 40 years now or something? [18:46:31] or something [18:47:03] I'm just amused that my dad had slide rules still when I was growing up [18:47:51] anyway, enough of this... thanks for the ideas, Ivy, bawolff_ [18:48:04] Sure. [20:19:29] Ivy, bawolff: just heard back from Florence: people have asked about installing WikiFundi on their own laptops. I'll probably write up how to do that. [20:56:34] Vagrant maybe? [20:56:40] Or Docker? Or something. [20:56:41] Depends on the OS, really. [20:56:55] And what they want to do on top of base MediaWiki. [21:08:23] Ivy: so they are already pre-loading a kiwix subset into the wiki [21:08:36] and they have VE working [21:08:52] Docker on windows? [21:09:13] OS image for VirtualBox [21:09:23] or just use WAMP [21:16:16] I don't hear great things about Docker on Windows. [21:16:21] Are they really running Windows? [21:16:56] if they're asking "Can I put this on my laptop?" then that is most likely. [21:19:31] WikiFundi write up: https://blog.wikimedia.org/2017/03/10/digest-wiki-fundi/ [22:33:46] hi! any advices on api.php read access vs $wgWhitelistReadRegexp? [22:34:15] !cms [22:34:15] Wikis are designed for openness, to be readable and editable by all. If you want a forum, a blog, a web authoring toolkit or corporate content management system, perhaps don't use wiki software. There is a nice overview of free tools available at including the possibility to try each system. For ways to restrict access in MediaWiki, see !access. [22:34:33] :) [22:34:39] nice one [22:34:46] thanks [22:35:23] btw, I share the same thought