[00:00:22] <Platonides>	 your change *plus* lowering $wgRejectedAccountMaxAge should do
[00:00:28] <johnywhy_>	 in ConfirmAccount.config.php i see  # When configuring globals, set them at LocalSettings.php instead
[00:00:30] <johnywhy_>	 so that's good
[00:01:52] <johnywhy_>	 this might do. Set wgRejectedAccountMaxAge (How long to store rejected requests) = 0. Then prune. Should remove the rejected request immediately. I'll try it
[00:02:03] <johnywhy_>	 The other switch seems unrelated
[00:07:02] <johnywhy_>	 Platonides still getting "Username is already in use in a pending account request."
[00:07:17] <johnywhy_>	 Do i need to run the wiki update process first?
[00:07:53] <Platonides>	 no
[00:08:09] <Platonides>	 what did you set Print_antifeature_intro-sandbox
[00:08:15] <Platonides>	 what did you set wgRejectedAccountMaxAge to?
[00:08:20] <Platonides>	 1
[00:09:55] <johnywhy_>	 Print_antifeature_intro-sandbox does not exist in LocalSettings.php
[00:10:06] <johnywhy_>	 i set wgRejectedAccountMaxAge  = 0
[00:10:30] <johnywhy_>	 some progress, i see that previous remnant rejects are gone, and they can re-request
[00:11:00] <johnywhy_>	 so they got pruned. Question is whether they got pruned by my new code, or by the default code
[00:11:24] <johnywhy_>	 that is, the one new request that i just tested is still there, cannot re-request
[00:11:31] <johnywhy_>	 only previous ones are gone
[00:12:18] <johnywhy_>	 maybe the prune has to happen after existing code runs, in a separate process. Caching issue?
[00:13:21] <johnywhy_>	 possible fix: run prune at beginning of new request
[00:13:24] <johnywhy_>	 i'll try
[00:13:34] <johnywhy_>	 (instead of at end of reject action)
[00:17:30] <johnywhy_>	 have to find php that runs at beginning of new request....
[01:03:12] <johnywhy_>	 Platonides, looks like i solved it
[01:03:51] <johnywhy_>	 in /ConfirmAccount/frontend/specialpages/actions/RequestAccount_body.php
[01:05:23] <johnywhy_>	 function showForm
[01:06:00] <johnywhy_>	 add very last command in the function
[01:06:00] <johnywhy_>	 ConfirmAccount::runAutoMaintenance();
[01:06:17] <johnywhy_>	 thx for help!
[14:27:18] <Wurgl>	 Hello!
[14:28:11] <Wurgl>	 https://tools.wmflabs.org/persondata/test.php <-- what is wrong here with connecting to the database (the website shows its source code)
[14:32:28] <bawolff>	 Wurgl: your error handling looks a little weird, but I don't actually see a connection error
[14:32:44] <jynus>	 should dewiki.labsdb work?
[14:33:25] <bawolff>	 $dbi->connect_error is null when I loaded the page, which suggests that there's no connection error
[14:34:57] <Wurgl>	 it works when started on the command line
[14:35:20] <jynus>	 yeah, but on command line maybe there is extra magic :-)
[14:35:37] <jynus>	 I am checking docs
[14:36:33] <jynus>	 https://wikitech.wikimedia.org/wiki/Help:Toolforge/Database#PHP recommends a longer alias
[14:36:45] <jynus>	 maybe the one you use works, not sure
[14:37:32] <jynus>	 dewiki.analytics.db.svc.eqiad.wmflabs or dewiki.web.db.svc.eqiad.wmflabs
[14:37:43] <jynus>	 but please don't take may word for it
[14:37:49] <jynus>	 just looking at the wiki
[14:39:18] <Wurgl>	 I tried with dewiki.analytics.db.svc.eqiad.wmflabs before (changed the code now) does not work either :-(
[14:39:36] <jynus>	 then it is something else, sorry
[14:39:59] <bawolff>	 Wurgl: What are you expecting to happen instead?
[14:41:26] <Wurgl>	 On the commandline I see  public $client_info =>  string(79) "mysqlnd 5.0.11-dev - 20120503 - $Id: bf9ad53b11c9a57efdb1057292d73b928b8c5c77 $"
[14:41:41] <Wurgl>	 Just one of the lines of var_dump
[14:42:18] <Wurgl>	 error.log does not show anything
[14:45:48] <bawolff>	 I would suggest trying to do a query, and see if that triggers any error, or succedes
[14:47:37] <Wurgl>	 Hmm … does var_dump behave different when called from the webserver? I added a line "print $dbi->client_info;" and this one shows some data?
[14:48:22] <Wurgl>	 strange
[14:49:42] <bawolff>	 Its possible its different versions of php
[14:50:20] <bawolff>	 And its not a normal php object, so lots of magic happening behind the scene. Could be lazy loaded in a way that var_dump() doesn't load the field or something
[15:10:53] <Wurgl>	 … or var_dump behaves different on objects for security reasons?
[19:24:08] <TK-999>	 hi, I found a weird schema situation: https://gerrit.wikimedia.org/r/plugins/gitiles/mediawiki/core/+/4124558d7b40df818337f1d7eeae10c9ee9688e2%5E%21/#F7 changed the primary key of revision table to just rev_id rather than (rev_page,rev_id), but I see no corresponding updater step
[19:24:49] <TK-999>	 so it seems old wikis are stuck with the old revision schema... however it seems like it's not causing issues (at least on mysql)
[19:28:00] <TK-999>	 however at first glance this might mean the rev_page,rev_id index might not need to be a UNIQUE one
[19:28:44] <TK-999>	 maybe it could be a simple index on rev_page
[19:35:55] <TK-999>	 I'll need to go now but I'll catch up on any replies
[19:40:58] <bawolff>	 TK-999: if you mean for wikimedia. On large wikis we use a non-standard schema for revision in order to do paritioning for performance reasons (I believe)
[19:41:05] <bawolff>	 I guess he left
[19:42:37] <Debenben>	 Hi, does anyone have an idea what happened to my post: https://www.mediawiki.org/wiki/Talk:Reading/Web/PDF_Functionality#What%20happened%20to%20my%20post???
[19:42:53] <Debenben>	 I asked a steward and he couldn't tell me
[19:43:11] <Debenben>	 Seems to be some bug with flow
[19:46:37] <bawolff>	 Wow, following that flow history page is hard
[19:49:19] <Debenben>	 thanks for having a look
[19:49:44] <Debenben>	 Stryn and I could see it here: https://www.mediawiki.org/wiki/Special:Contributions/Debenben
[19:50:01] <Debenben>	 but not here: https://www.mediawiki.org/w/index.php?title=Topic:U7yyjlehi3rvc4b7&action=history
[19:50:15] <Debenben>	  and it is not shown on the page anymore
[19:51:10] <bawolff>	 yeah, I totally don't understand flow
[19:51:17] <bawolff>	 so I have no idea
[19:51:29] <Debenben>	 me neither
[19:51:46] <Debenben>	 everyone I asked said he doesn't understand flow