[08:55:12] <Mirghon>	 pywikibot-room is not as responsive as this one, huh :)
[12:51:54] <hexnewbie>	 Hello. I would imagine that https://www.mediawiki.org/wiki/Special:MyLanguage/Download listing 1.36.2 as “LTS” is a mistake, and the LTS is still 1.35.4 as per topic, that would be correct, no? Is there an official source other than the Wiki where I can confirm these?
[12:54:38] <Reedy>	 That page and how we manage it is stupid
[12:54:45] <Reedy>	 Yes, 1.35 is still a supported LTS
[12:55:44] <Reedy>	 https://www.mediawiki.org/wiki/Template:MW_lts_branch_number
[12:55:50] * Reedy scratches his head
[12:57:51] <Reedy>	 yeah, it is
[12:57:52] <Reedy>	 * MediaWiki {{MW legacy release number}} ({{MW legacy release link zip|1=<translate><!--T:40--> download <tvar name=1>.zip</tvar></translate>}}, {{MW legacy release link|1=<translate><!--T:41--> download <tvar name=1>.tar.gz</tvar></translate>}}) - <translate><!--T:49--> current long-term support (LTS)</translate>
[12:57:57] <Reedy>	 legacy release does not make it an LTS
[13:04:03] <Reedy>	 hexnewbie: fixed
[13:04:36] <hexnewbie>	 Thanks
[13:04:53] <Reedy>	 For future reference, https://www.mediawiki.org/wiki/Version_lifecycle is probably the page you were looking for
[16:33:24] <RhinosF1>	 Does anyone know why https://github.com/wikimedia/mediawiki/blob/23d5d441c9965ab21870e6977baa8c83001c01ee/maintenance/nukeNS.php#L93 can't be moved to after the loop?
[16:34:05] <RhinosF1>	 That line seems to causing awful performance
[16:36:41] <Vulpix>	 And potential data loss
[16:37:34] <Reedy>	 I mean, it's deleting potentially thousands of rows...
[16:41:39] <Vulpix>	 purgeRedundantText has a lot of issues if you search for it in https://phabricator.wikimedia.org/tag/mediawiki-maintenance-system/
[16:43:03] * Reedy pretends to be surprised
[19:35:47] <RhinosF1>	 Reedy: https://integration.wikimedia.org/ci/job/wmf-quibble-selenium-php72-docker/123773/console will fix the crappy script performance
[19:36:00] <RhinosF1>	 https://gerrit.wikimedia.org/r/c/mediawiki/core/+/740165
[19:36:12] <RhinosF1>	 CI is failing though due to a GE issue
[19:59:46] <wctaiwan>	 Reedy: Thanks for looking at T296103 and T295737. I'm happy to try to work on a patch for either if you can give me some pointers on what the behavior should be.
[19:59:47] <stashbot>	 T295737: Install script doesn't validate that provided MySQL user has necessary privileges - https://phabricator.wikimedia.org/T295737
[19:59:47] <stashbot>	 T296103: Error message truncated in install script - https://phabricator.wikimedia.org/T296103
[20:04:16] <wctaiwan>	 I guess for T296103 we can just catch the error and emit it manually instead of relying on MWExceptionHandler::handleUncaughtException?
[20:05:38] <wctaiwan>	 (i.e. wrap a try/catch around `performInstallation`)
[20:45:22] <SIlverSurfer>	 Hello, I have a quick upgrade question: Can I upgrade from MediaWiki 1.31.0 to 1.37.0 or do I have to upgrade to 1.35.4 first and then from 1.35.4 to 1.37.0? Or do I even have to upgrade from 1.35.4 to 1.36 and then to 1.37.0?
[20:46:31] <RhinosF1>	 I believe the release note indicates which version to use
[20:47:38] <RhinosF1>	 Oldest you can upgrade from is 1.27
[20:48:33] <RhinosF1>	 SIlverSurfer: ^
[20:49:02] <SIlverSurfer>	 Ah, okay, thank you! The UPGRADES file says "Upgrading from releases older than two LTS release is not supported anymore." I wasn't sure what exactly counts as LTS.
[20:49:12] <SIlverSurfer>	 UPGRADE*
[20:49:46] <RhinosF1>	 LTS is every 4 releases
[20:50:52] <SIlverSurfer>	 Cool, that's good to know! Thank you for your kind help and have a nice day!
[20:52:27] <RhinosF1>	 Looks like it's mentioned on the line below
[20:59:04] <SIlverSurfer>	 Oh, yeah, sorry, my bad. You're right! I totally missed that line. Looks like I'm a little tired from working all day.
[20:59:30] <SIlverSurfer>	 Take care & stay healthy! Bye!
[21:00:23] <AntiComposite>	 see also https://www.mediawiki.org/wiki/Version_lifecycle
[21:18:11] <SIlverSurfer>	 Thank you!
[22:34:15] <RhinosF1>	 Reedy: with https://gerrit.wikimedia.org/r/c/mediawiki/core/+/740165 applied, it made more progress in 1 minute than the 4 days without that change
[22:35:39] <Reedy>	 I wonder if it's more the transaction or the loop
[22:36:01] <RhinosF1>	 Reedy: no idea but that's an insane change
[22:36:12] <RhinosF1>	 From that one line patch
[22:36:34] <RhinosF1>	 That's 5760x faster
[22:37:00] <Reedy>	 The code looks weird
[22:37:17] <Reedy>	 Why is $this->beginTransaction( $dbw, __METHOD__ ) before the loop
[22:37:42] <Reedy>	 But there's a $this->commitTransaction( $dbw, __METHOD__ ) in the nested delete... But never starting it again
[22:38:29] <RhinosF1>	 That makes 0 sense
[22:39:19] <Reedy>	 It looks to have been like that for 13 years
[22:39:26] <Reedy>	 Not that it makes it right. Probably just that it's not used much
[22:39:44] <RhinosF1>	 Yeah I can't imagine it used much
[22:39:54] <RhinosF1>	 I need sleep
[22:40:26] <RhinosF1>	 But I'm sure if you comment on the patch more can be added or upload a new PS
[22:44:03] <Reedy>	 meh
[22:44:06] <Reedy>	 seems we do this elswhere
[22:46:21] <Reedy>	 but...
[22:51:40] <Reedy>	 I can't decide if this is wrong, write, weird...
[22:51:44] <Reedy>	 *right