[01:06:11] RhinosF1: Yes that's why I set it up like that, because of DefaultSettings not yet being initialised, and running it in PhpAutoPrepend makes the profiling more accurate (or seems to at least). I can take look on our setup later also, yes. Also it does work to setup profiler for the forceprofile=1 but not otherwise populating so I'm not certain but will look. [01:36:06] hi folks. there's something with charsets i totally don't understand what's going on in MW. [01:36:28] https://de.wikipedia.org/wiki/Sanctuary_%E2%80%93_W%C3%A4chter_der_Kreaturen/Episodenliste#%C3%9Cbersicht [01:37:07] the 2nd header. in the source view it has a ­ and the browser nicely splits the word there if needed. [01:37:31] but when i curl the page on that position i just have a 0xC2. [01:38:26] but according to wikipedia ­ is AD or C2 AD, and i remember C2 to be a UTF8 prefix. [01:38:53] now what would be the content REALLY stored in the db? [02:05:16] ede: whatever you see in the "view source" / "edit" box is what is stored in the db [02:05:22] so in this case it is the string "­" [02:06:04] moonmoon: but why do i see 0xC2 when i curl the file and then feed it do hd? [02:06:11] blame curl [02:06:24] https://de.wikipedia.org/w/index.php?title=Sanctuary_%E2%80%93_W%C3%A4chter_der_Kreaturen/Episodenliste&action=edit§ion=1 has ! rowspan="2"| Episoden­anzahl [02:08:25] hm, i curl'd again (restarted the whole thing twice at least), ­ now [02:09:36] oh, seems the C2 comes from firefox oO [02:12:07] when I inspect it in dev tools (using a chromium-based browser) I see ­ in the markup. When I view html source in the same browser, it shows ­ [02:12:25] which lends weight to the browser (and possibly also curl) transforming things on your behalf [02:14:19] the source in firefox shows ­, same as curl. but the html editor in the treeview totally breaks it. [02:14:51] and ­ is in latin already, so nearly stoneage. [02:17:32] anyway it's the literal string "­" in the db, that much I'm 100% certain of [02:18:53] wondering what firefox is doing there, i'll check it tomorrow with some other versions. i really don't feel like installing chrome :) [02:22:02] heh, even pandoc replaces the entity with 0xAD when converted to markdown. [10:24:35] Hi. Where does MediaWiki import jquery? I know it happens via a ResourceLoader, but where in the code does this happen? [10:24:35] I am trying to move away from using jquery and the first step would be to not import it and see what breaks, etc. [10:28:28] Equinox76: ResourceLoaderStartUpModule::getBaseModules() [10:28:53] I'll check that out. Thank you! [10:28:57] I appreciate it [10:28:59] Equinox76: i'm afraid everything will break, as almost everything in MediaWiki uses jQuery for all the mundane stuff :) [10:29:04] but good luck experimenting [10:29:05] :/ [10:30:20] Equinox76: you might be interesting in https://phabricator.wikimedia.org/T200868 , you're not the only person who wants it gone [10:30:23] interested* [10:30:33] but it would require a lot of work [10:30:52] That is certainly very useful. Thanks for sharing [10:37:27] MatmaRex, that works btw (I can disable jquery loading from that method) [10:37:35] you're also right about almost everything breaking [10:38:26] Equinox76: is there are particular reason you want to store using it? [10:40:57] trying to find out any page speed improvement techniques by cutting down on meaty packages [10:41:01] I found this: https://github.blog/2018-09-06-removing-jquery-from-github-frontend/ [10:41:04] which is what got me thinking [10:41:14] so, just experimented on a dev server:) [10:41:19] experimenting* [10:42:21] I would highly recommend hacking away at core mediawiki files (unless you are planning to submit patches upstream), because otherwise essentially you are just making it really hard to upgrade in the future [10:59:11] *recommend against [10:59:47] that makes sense [10:59:58] just some dev experimentation is all I'm upto [11:00:14] I appreciate the advice. thanks p858snake