[01:30:19] subbu: Is magic links on the roadmap for next 12 months -ish? If so good next step for https://phabricator.wikimedia.org/T145604 would be to identify any stakeholders that you might want to sollicit input from before last call. [01:30:48] (e.g. to inform a possible change in timeline or migration possibilities.) [09:40:24] Uploaded file: https://uploads.kiwiirc.com/files/46dc8ecde3880833b8157a66bff67833/40E38BDD-E990-43E5-838C-19362A6711B0.jpeg [09:40:32] Hello, good day. When I try to install MediaWiki 1.34.1, I get an error message when configuring the database. I have attached a screenshot. [09:43:01] owoqkqkiws: user has no access to the database. Ask your host for permissions, or check that the username and database you introduced are correct [09:44:41] The data is correct. [09:48:32] are you sure [12:07:41] Hi, 2020/04/17 14:03:56 [error] 24964#24964: *9 directory index of "/var/www/html/" is forbidden, client: 192.168.XX.XX, server: wiki.maindomain.ch, request: "GET / HTTP/2. [12:07:42] 0", host: "wiki.maindomain.ch:4141" recieve 403 on Nginx with SSL, PHP is working fine, permissions is ok. I use the sample config from NGINX for Wikimedia: https://www.nginx.com/resources/wiki/start/topics/recipes/mediawiki/ fastcgi is also configured and should be fine. I have to use a different than 443 because on the same VM is an apache2 running. [12:12:01] Hi, maybe a stupid question, but how do I get the name / real name of the user inside a php file? [12:24:38] rinner: Depends somewhat where you are [12:35:46] Reedy I want to change the displayed name in the Flow extension [12:35:58] From? To? [12:36:39] at the moment it displays the username, but the user name is an id. I want to display the real name [12:37:01] but I don't know how to access it from the Flow file [12:37:31] the file is flow_post_handlebars.php [12:38:22] I tried $this->getSkin()->getUser()->getName() but it's not working [12:38:23] I wouldn't advise changing the Flow code directly... But I don't see any way to easily make this change otherwise [12:39:15] do you have a better idea Reedy [12:41:13] In the first instance, potentially post a feature request in phab [12:42:39] you can't just use normal php in the handlebars files AIUI [12:44:39] so what should I do Reedy? Is there a "simple" way to change the displayed name? [12:47:07] I don't think so [12:47:29] You could find where the variable is substituted into the handlebars file, and replace the getName call there, but hacking the code isn't really recommended [12:48:43] hmmm damn, of corse it couldn't be easy :D [12:49:12] *course [15:46:01] Krinkle, we'll chat about it at our team meeting. [17:32:04] Hey all, I'm seeing what would seem to be strange behavior in one of my wikis' job queues. It has $wgJobRunRate set to 0.5 and I'm running a 2-core runJobs.php process against it, but after it's spike to over 120k a couple of days ago, it's been hovering between 67-70k for 24 hours now, though I'm watching the runJobs output and it's been [17:32:04] continually processing jobs. https://imgur.com/R9dEydZ [17:33:29] FWIW they're pretty much all smw.update jobs. [17:55:46] justinl: why have $wgJobRunRate set to more than zero if running jobs extra? [17:56:39] Normally I don't run runJobs.php though I'm moving in that direction. Our wikis have had $wgJobRunRate set for years, the two biggest ones set to 0.1 and three smaller ones to 0.5 [17:57:04] I'm just trying to help clear the queue after some recent template changes created a massive number of jobs. [17:57:25] I do plan on switching to a job runner system in the next couple of months probably. [17:58:15] Just trying to understand why this pattern of jobs is happening and whether it's abnormal or sometimes expected. I've just never seen it this severe, rarely have I seen job queues exceed 80k and even that's relatively rare. [17:59:19] FWIW I also don't work on the content, that's completely left up to our communities, but our big wikis, 4 out of 5 of them, use SMW and layers of templates extensively. [18:01:54] I'd set $wgJobRunRate = 0 for now, then clear the backlog with one job running for a long time. [18:02:18] SMW is broken; it creates jobs for lots of things [18:03:34] We've been using for years and our wikis use it extensively so there's nothing to be done there. I'd like to find ways to improve them from the admin side (finally just started running cron jobs to delete outdated properties and refresh property counts as well, they'd built up over years). [18:04:51] As for the rate, I can change it, and I do plan on letting the current runJobs process continue, but I was just trying to figure out whether there was something wrong given how long the number of jobs has stayed abouthe the same despite the script processing them at a pretty good rate. I know jobs can cause new jobs to be created, this just seemed [18:04:52] a bit excessive. [18:05:40] justinl: https://directory.fsf.org/wiki/Free_Software_Directory:Backlog_active#Monitoring_MediaWiki it took me 14 hours to clear a backlog [18:06:59] My last rebuild of our biggest wiki took over 5 days. Probably better now that I was able to cleanup 4.5 million outdated properties of about 6 million total props. [18:09:13] I am working on setting up a systemd jobrunner service for each of our 5 main wikis, so I do intend to deploy that and disable the $wgJobRunRate for all of them. Just need to get the stuff written and a dedicated instance built and configured to do that work. [18:13:02] this is why I discourage folk from running smw at all [18:15:02] If I'd known like 7 years ago when we were asked to add it to our wikis for our users how painful it can be to deal with, I may have declined, but too late for that now, so I'm just doing my best to provide the best wikis I can to our communities. [18:15:21] Our big non-SMW wiki uses DPL, which is its own can of worms.