[04:04:30] legoktm, https://gerrit.wikimedia.org/r/#/c/430026/ should be a quick review. [05:31:48] Reedy: didn't notice you were converting GWToolset, repeated your work:P [11:30:22] MaxSem: You duplicated a few of my patches it seems :P [18:36:54] Reedy: I was actually thinking of doing something like https://gist.github.com/MaxSem/1f64df4f9448eeb32be548fbddb52be1 in config [18:37:35] https://gerrit.wikimedia.org/r/#/c/392217/11/includes/Config.php [18:37:45] Which I think I stole from your hack to prod before [18:38:07] yeah, and I don't like my solutionn anymore XD [18:38:30] Just undeploy it [18:38:32] Much easier [18:39:32] do we have usage stats? [18:40:11] * MaxSem fires up Hadoop [18:40:32] I'd be surprised if it's actually highly used [18:43:08] http://tools.wmflabs.org/pageviews/?project=commons.wikimedia.org&platform=all-access&agent=user&range=latest-90&pages=Special:GWToolset [18:43:38] how dare they? I just waanted to undeploy sumthin'! [18:44:59] Can we do better than that? [18:45:06] ie more than just casual browsing [18:45:41] https://commons.wikimedia.org/wiki/Special:Log/gwtoolset [18:45:43] * hoo hides [18:46:02] behave :P [18:46:18] Looks like the answer is that it's well used [18:46:24] Indeed [18:46:37] At least, on and off [18:46:43] lol [18:46:54] 5000 log entries still only goes back to 26 April 2018 [18:47:08] I suspect we should get WMDE to build something better ontop of FileImporter ;) [18:48:14] Sounds trivial… How hard can it? [18:48:44] I guess it's just dealing with the import format thing [18:49:16] Yeah, it's some crazy description XML and then pulls stuff from a remote host [18:50:46] I wonder if it's worth filing a task [18:51:05] * Reedy writes one [18:52:51] https://phabricator.wikimedia.org/T193526 [19:17:55] does anyone know why are there records on the logging table with log_type = 'block' and log_action = null ? [19:18:01] is this a bug? [19:18:22] Recent entries? [19:18:33] those records also have log_title, log_params, log_page as null [19:18:35] yes [19:18:37] last 30 days [19:20:01] last one happened on '20180421214430' [22:12:52] anomie: Hm.. I'm curious if it would make ssense to allow both titles= and pageids= to be used on an API query. [22:13:22] I thought maybe we don't do that because the output format differs (e.g. pages[title]={} vs pages[id]={}), but we actually use pages[id] in both cases. [22:13:33] Seems like they're compatible/mergeable