[14:08:17] <Guest49>	 I need help with a MYSQL query for media wik
[14:08:33] <Guest49>	 Is there somebody good with MySQL?
[14:08:45] <taavi>	 don't ask to ask
[14:09:13] <Guest49>	 I need to get these info from my Wiki:
[14:09:14] <Guest49>	 list of pages in the wiki with username of who created the page, username of who made the last change, timestamps for these two actions, and text length.
[14:09:14] <Guest49>	 I can't figure out in which way to join the users and the timestamps to the page_title and page_len columns.
[14:09:15] <Guest49>	 Could anybody help me in building the correct query please?
[14:09:47] <Reedy>	 Have you looked at the revision table?
[14:10:21] <Guest49>	 yes, my problem is the JOIN query in MySQL
[14:10:32] <Guest49>	 because I need all that info with just one query if possible
[14:11:01] <Reedy>	 You might want to do two queries... One for the first and one for the last... But that's basically the same uery
[14:11:34] <Guest49>	 problem is that then I won't have the user name of the person that made the last change on page_title
[14:11:55] <Reedy>	 Why won't you?
[14:12:26] <Guest49>	 because it's in another table than mw_page, right?
[14:12:49] <Reedy>	 Yeah... But that's a simple join?
[14:13:44] <Guest49>	 yes exactly, I'm asking for help in building that Join query because I'm good at MySQL but not enough for this
[14:19:32] <Guest49>	 could you help me with the join query?
[14:39:20] <Guest49>	 nobody?
[16:25:26] <Penelope>	 Hi Guest49.
[16:25:41] <Reedy>	 They left hours ago
[16:25:56] <Penelope>	 Their loss, really.
[16:50:17] <addshore>	 Reedy: im quite surprised that the patch i proposed uses so much more memory apparently =o
[16:50:50] <Reedy>	 addshore: heh. Something definitely feels odd, but it could be benchmark weirdness too
[16:51:11] <Reedy>	 Definitely happy to try and help dig into it
[16:51:26] <Reedy>	 But why I wanted to create the benchmark so we have some idea what we're actually doing
[16:51:30] <addshore>	 oh, haha, today is the day it gets deployed to candidate wikis
[16:51:40] <Reedy>	 yeah, you're a bit late to the party :P
[16:51:52] <Reedy>	 I mean, I know an ipset is much more compact, for large ranges and consecutive ips
[16:52:35] <Reedy>	 Feel free to double check my benchmarking tests
[16:52:47] <addshore>	 I saw your patch adding the json serialization stuff, and thought you might be going to add some caching :P
[16:52:54] <Reedy>	 :D
[16:53:34] <addshore>	 *stares at the banchmark*
[16:53:57] <Reedy>	 Feel free to merge my other patches if you want so we've got a merged baseline etc
[16:54:15] <addshore>	 other patches?>
[16:54:25] <addshore>	 oh, the benchmarks?
[16:55:35] <Reedy>	 some other cleanup and stuff
[17:20:49] <addshore>	 I'm not sure i trust your benchmarks, but also not sure why yet
[17:21:39] <Reedy>	 well, I didn't write the code :P
[17:21:50] <Reedy>	 but it's possible the 10 loops isn't having the desired effect
[17:33:58] <addshore>	 the benchmark of old code is doing getIpDenyList which avoids the IPSet? and the benchmark of my new code does the IPSet?
[17:34:14] <Reedy>	 Entirely possible
[17:34:20] <Reedy>	 Entry points are... limited :)
[17:35:35] <Reedy>	 https://doc.wikimedia.org/cover-extensions/StopForumSpam/
[17:35:39] <Reedy>	 look at the awesome test coverage
[17:44:31] <addshore>	 Reedy: I don't think I can reliably run the benchmark on my shoddy internet
[17:44:42] <Reedy>	 heh
[17:44:51] <addshore>	 but could you do https://gerrit.wikimedia.org/r/c/mediawiki/extensions/StopForumSpam/+/824249 PS1 vs https://gerrit.wikimedia.org/r/c/mediawiki/extensions/StopForumSpam/+/822727 PS11 ?
[17:45:00] <Reedy>	 I mean, that's the other thing... It's not actually diffentiating for "time to download" (which is hard to do)
[17:45:12] <Reedy>	 Maybe we increase testability, and use a file on disk etc
[17:45:26] <addshore>	 yeah, i think that should be possible by just configuring SFS
[17:45:36] <Reedy>	 can you just stick a file in? I cannae remember
[17:45:42] <addshore>	 I think so
[17:45:45] <addshore>	 *reads the docs*
[17:45:55] <addshore>	 $wgSFSIPListLocation - location of an unzipped IP address blacklist
[17:46:29] <Reedy>	 		if ( is_file( $wgSFSIPListLocation ) ) {
[17:46:29] <Reedy>	 			$ipList = $this->fetchFlatDenyListHexIpsLocal( $wgSFSIPListLocation );
[17:46:31] <Reedy>	 ah, yeah
[17:46:39] <Reedy>	 So that probably helps your local testability if you can grab it :)
[17:46:40] <addshore>	 *looks at which list it normally uses*
[17:47:03] <addshore>	 https://www.stopforumspam.com/downloads/listed_ip_30_ipv46_all.gz
[17:47:28] <Reedy>	 We're using the 90 one in prod, fwiw
[17:47:50] <Reedy>	 moar bigger
[18:04:31] <addshore>	 Reedy: "you've spelled recache wrong :P" lets pretend i didnt spend 15 mins investigating that ....
[18:05:34] * Reedy invoices addshore for his time
[18:05:50] <Reedy>	 I've just +2 the two patches to get them merged
[18:07:42] <addshore>	 cool, then ill rebase my newer code and profile them both
[18:07:45] <addshore>	  / benchmark
[18:22:20] <addshore>	 Reedy: loooks like the profiling makes more sense now, and the new code i wrote = more memory, but less time
[18:41:37] <addshore>	 Reedy: also benchmarked with the list your using on wmf sites now, and I think the patch is good ! :)
[18:41:50] <Reedy>	 Sweet
[18:41:55] <Reedy>	 I'll have another look myself :)