[12:19:53] Is it safe to have mediawiki backends running a mix of php74 and php82 at the same time [12:21:09] Depends what you mean by safe [12:21:28] It's potentially using different libicu... So your category sorting/collation could get interesting [12:24:30] Reedy: well Miraheze needs to upgrade bullseye to bookworm [12:24:43] And I can't depool half the servers at once [12:24:54] Category sorting is fairly meh [12:25:32] You could install both versions of PHP on the same hosts though [12:26:37] Reedy: that's true, we could be sensible and use backported php packages [12:26:46] I mean that wouldn't be fun though [12:28:30] https://deb.sury.org/ [12:38:01] Reedy: ye that's too sensible [18:14:10] where is the search list saved in the sql? the articles that pop up in the search list when i start typing [18:15:05] it finds articles/pages that are not existent [18:18:08] it may be your browsers autocompletion [18:25:34] no that would look different [18:35:56] If you have CirrusSearch installed, the search index may take a bit to update. Search suggestions may be added to MediaWiki's cache too. Also, if you have a proxy cache or using cloudflare, they may be caching some queries. And there's your browser cache too [18:37:26] If you open the URL that is fetching the suggestion and open it in your browser, you can add a random query string at the end to see if it returns different results. That would bypass your browser and any intermediate cache [18:41:48] Vulpix: is it possible to list the cache in the shell? like in maintenance/ [18:42:58] depending on your cache setup, it may be in database in the objectcache table, or in redis/memcached. Redis has a cli for viewing its contents. I'm not sure about memcached [18:49:55] hm how do i find out my cache setup LocalSettings.php says: $wgMainCacheType = CACHE_NONE; [18:53:26] Well, that should discard mediawiki caching, then [19:33:11] weird, after updating from 1.32.6 to 1.33.4 content seems to be lost, but in mysql when i select page_len from page, it still got the same length as before, it's there but no content [19:34:07] what table is the content for a page_id stored? [19:34:22] text table [19:34:49] revision stores more metadata too [19:35:03] there are only old_id old_text old_flags [19:35:12] mysql> show columns from text; [19:36:14] page -> revision -> slots -> content -> text [19:38:15] or in terms of relations, page is 1:many with revision. revision is 1:many with slots. slots is many:many with content (revisions containing duplicate text is deduplicated via this relationship), content is 1:0-1 with text (actual text content can be stored outside of the main db in non-default setups) [19:38:53] er, slots is many:1 with content. sorry [19:40:14] mysql> select old_text from text where old_id=15780; [19:40:22] weird the content is there [19:40:27] that is the id [19:41:30] with maintenance/view.php "title of that id" i get 'Page has no content' [19:42:52] is it possible to view a page by the pag_id? [19:50:19] w/index.php?curid=15780 is empty too [19:56:30] Bombo: if you have a page id or revision id, you get a text table old_id by following that table chain I mentioned above [19:57:08] or if you have a text table old_id you can get a revision id (potentially many of them) by going through the reverse mapping [20:00:47] moonmoon: so i can't just use the page_id i found in page table to get the content in text table where old_id=the same id? [20:01:21] there is no guarantee of that [20:01:45] in fact that assumption (page_id = old_id) is going to be wrong more than it's right [20:02:09] there are no shortcuts, you have to go through that entire table chain to resolve the proper ids [20:03:58] so [20:04:03] mysql> select rev_id, rev_page from revision where rev_id=15780; [20:04:16] 15780 is the page_id [20:04:32] i get rev_page 12109 [20:06:01] rev_text_id 15627 [20:07:02] with that i go to slots? [20:08:00] i don't get it sorry [20:18:35] so the goal is to get the content of the current revision of the page_id i got from page, so i go to revisions slots content to get an id that leads me to an old_id in text which is the current revisions content? [21:37:18] Hi, its possible to update/upgrade mediawiki on own server without this patch or ta.gz way and copy ...etc? [21:39:13] Client: HexChat 2.16.1 • OS: ArchLinux • CPU: Intel(R) Core(TM) i5-8400 CPU @ 2.80GHz (3.80GHz) • Memory: Physical: 15.1 GiB Total (11.3 GiB Free) Swap: 3.9 GiB Total (3.9 GiB Free) • Storage: 59.8 GB / 479.6 GB (419.9 GB Free) • VGA: Intel Corporation CoffeeLake-S GT2 [UHD Graphics 630] @ Intel Corporation 8th Gen Core Processor Host Bridge/DRAM Registers • Uptime: 57m 46s