[07:23:09] load.php should be cached fully for logged out users if mediawiki is setting the headers properly [07:23:34] We don't have wiki replicas for Miraheze wikis [07:23:45] And in any case it's not take effect for months [07:24:22] The chance that there's anyone in here that message is relevant to that doesn't get it anyway is almost zero [07:27:31] It's of no relevance to anyone here until we start debugging the next mediawiki release [07:27:47] girl we're all in the mediawiki server [07:27:51] In which case there's a chance a missed patch could be a problem but we'll of forgot your message in 6 months time [07:27:59] if not on the mailing list itself [08:35:05] `load.php` should be cached for everyone, even logged in users [08:38:07] We follow mediawikis cache headers [08:38:13] So it'll do whatever they do [09:04:36] I already patched it for MirahezeMagic (the only place it is used in our extensions) lol. So it doesn't even effect us at all anymore. [09:21:18] Relax! Breath! Don't speak for everyone! If you want to read it, read it! If you don't, skip it! Don't make such drama over a post I make ! [09:57:36] You've been told countless times that it's pointless to post random things you see on the mailing lists though [09:57:42] Tech have more than one brain cell [09:58:12] I mean both were decerprated before we even existed [09:58:47] If Tech tell you that we don't find your content relevant to the #tech channel and sometimes confusing for users, you should stop posting it [10:05:03] nvm. I'm tired of your tirades against everything You seem irrelevant. That's all I will respond to. So don't come back with another long explanation like usual. [11:23:20] @atxatx i was looking at one of the failures between php 8.1 and 8.3 [11:23:30] i found an issue with gzip [11:23:51] e.g. [11:23:52] > $value = mb_substr( SMWInfolink::encodeCompactLink( '-5B-5BProperty%3A%2B-5D-5D-20-5B-5BCategory%3ALorem-20ipsum-5D-5D/-3FHas-20description%3DDescription/-3FHas-20type/mainlabel=/format=table/class=datatable/sort=/order=asc/offset=100/limit=50' ), 3 ); [11:23:59] > $value = base64_decode(str_pad( strtr( str_replace( '.', '__', $value ), '-_', '+/' ), strlen( $value ) % 4, '=', STR_PAD_RIGHT )); [11:24:06] returned [11:24:10] [1/3] > string(139) "c:E?1?0 [11:24:10] [2/3] EO?%1z(EU?^???"%8?݁?ʀ????wm???-\Hl??{U??V??4??q`???X??ϛw??˙t?X,?Z5}?u????ׄ#%? [11:24:10] [3/3] KF?1???Œ?'+??g?Iu?,J?|?9??" [11:24:13] under php 8.1 [11:24:45] under 8.3, it returns differently and thus gzinflate would fail [11:25:07] e.g. under 8.3 it would be C:E10 not the one above [11:25:31] @cosmicalpha maybe you will know ^? [11:35:17] oh [11:35:34] was just my console cuz doing it under php 8.1 i get [11:35:38] [1/4] ```> var_dump($value); [11:35:38] [2/4] string(139) "c:E10 [11:35:38] [3/4] EO%1z(EU^"%8݁ʀwm-\Hl{U?V4q`Xϛw˙tX,Z5}uׄ#% [11:35:39] [4/4] KF1'+gIu,J|9"``` in docker [11:49:44] [1/4] > > var_dump(mb_substr( $value, 2 )); [11:49:45] [2/4] > string(137) "E?1?0 [11:49:45] [3/4] > EO?%1z(EU?^???"%8?݁?ʀ????wm???-\Hl??{U??V??4??q`???X??ϛw??˙t?X,?Z5}?u????ׄ#%? [11:49:45] [4/4] > KF?1???Œ?'+??g?Iu?,J?|?9??" [11:49:46] php 8.1 [11:49:59] [1/4] > > var_dump(mb_substr( $value, 2 )); [11:49:59] [2/4] > string(133) "E?1?0 [11:50:00] [3/4] > EO?%1z(EU?^???"%8?݁?ʀ????wm???-\Hl??{U??V??4??q`???X??ϛw??˙t?X,?Z5}?u????ׄ#%? [11:50:00] [4/4] > KF?1????'+??g?Iu?,J?|?9??" [11:50:03] php 8.3 [11:54:42] guess it's `base64_decode(str_pad( strtr( str_replace( '.', '__', $value ), '-_', '+/' ), strlen( $value ) % 4, '=', STR_PAD_RIGHT ));` [12:05:03] oh! [12:05:05] maybe i have a fix [12:17:04] seems mb_substr handles things differently in php 8.3 vs 8.1 [12:17:08] using substr fixed it [12:27:59] I marked Cargo and SMW as tested on Phorge, though those should be taken with a grain of salt since I did not test them extensively: if they don't crash on Mirabeta I'm going to assume that CI did its job properly. [12:30:49] @atxatx https://github.com/SemanticMediaWiki/SemanticMediaWiki/pull/6206 [13:50:55] Error 503 Backend fetch failed via cp38 [13:51:07] Again! [13:51:49] nvm. back up again [14:59:36] @atxatx https://github.com/SemanticMediaWiki/SemanticMediaWiki/pull/6207 another fix [16:54:07] i'm wondering if @cosmicalpha you know how to fix https://github.com/SemanticMediaWiki/SemanticMediaWiki/actions/runs/17069619250/job/48394940081?pr=6206 [16:54:13] [1/2] > Not dropping new table, as 'sunittest_user' is name of both the old and the new table. [16:54:13] [2/2] > .......EScript php ${MW_INSTALL_PATH:-../..}/tests/phpunit/phpunit.php -c phpunit.xml.dist --bootstrap tests/bootstrap.php handling the phpunit event returned with error code 2 [17:34:53] [1/2] @blankeclair I was looking through open tasks with no assignee, do those still need to be done or looked into? [17:34:54] [2/2] https://cdn.discordapp.com/attachments/1006789349498699827/1407417553713893537/image.png?ex=68a6073d&is=68a4b5bd&hm=75c770d7cf04628aaa0830677e1f319de8ee89b3aea2cbddfc12180684122d33& [17:39:21] royalwiki has something fucked [20:25:18] i forgot 💀 [20:37:21] gg [20:53:13] [1/2] 1. you're welcome [20:53:13] [2/2] 2. do not look at https://issue-tracker.miraheze.org/maniphest/query/yF9.mncLTnIx/#R [21:06:19] 😭 [22:51:59] [1/2] https://discord.com/channels/407504499280707585/1407390933095551109/1407497003516166194 [22:52:00] [2/2] Hopefully I didn't mislead other users. I do believe this is a MH configuration issue because `$wgAdvancedSearchDeepcatEnabled` defaults to true but it needs to be set to false. [22:52:55] deep :3