[08:54:13] <wm-bb>	 [telegram] <Angel> Hello everyone❤️
[09:08:40] <wm-bb>	 [telegram] <Thecladis> Hi, are you interested in a MediaWiki and Wikimedia related topics? (re @Angel: Hello everyone❤️)
[09:09:18] <wm-bb>	 [telegram] <Angel> What's the meaning? (re @Thecladis: Hi, are you interested in a MediaWiki and Wikimedia related topics?)
[13:34:02] <wm-bb>	 [telegram] <bd808> https://mediawiki.org/wiki/Extension:QuickInstantCommons (re @HikeAndMap: is there any progress on MediaWiki for this weird thing that larger galleries & tables with lots of images are not shown properly - but needs to be ?action=purge once or multiple times to actually render all the pictures?)
[16:32:54] <legoktm[m]>	 I think the problem might be https://phabricator.wikimedia.org/T266155
[16:34:04] <legoktm[m]>	 thumbor rate limits how many thumbnails a single user can generate in a certain time period, so if you encounter or create a page with lots of thumbnails, it'll render as much as you can in the limit, and then you have to wait it out, or another user can visit it and generate more thumbnails, ...
[19:53:37] <wm-bb>	 [telegram] <fuzheado> legoktm - isn't the state of Thumbor pretty sad in general? Old Python 2 code, running on a legacy system?
[19:54:32] <legoktm[m]>	 very sad :(
[19:54:53] <legoktm[m]>	 the biggest problem is that no one is assigned to work on it or even be barely responsible for it
[19:55:40] <legoktm[m]>	 https://lists.wikimedia.org/hyperkitty/list/wikimedia-l@lists.wikimedia.org/message/WMPZHMXSLQJ6GONAVTFLDFFMPNJDVORS/ is a good summary
[20:03:27] <wm-bb>	 [telegram] <chicocvenancio> that is a depressing thread (re @wmtelegram_bot: [irc] <legoktm[m]> https://lists.wikimedia.org/hyperkitty/list/wikimedia-l@lists.wikimedia.org/message/WMPZHMXSLQJ6GONAVTFLDFFMPNJDVORS/ is a good summary)
[20:05:23] <legoktm[m]>	 https://phabricator.wikimedia.org/T294484 is probably where to subscribe to get updates
[20:05:34] <legoktm[m]>	 but yeah...
[20:05:58] <legoktm[m]>	 thumbor is really just the tip of the iceberg, the rest of the media stack is also mostly unmaintained
[20:06:11] <wm-bb>	 [telegram] <chicocvenancio> legoktm, I made the mistake of looking on the whole thread. It gets a lot worse.
[20:11:12] <wm-bb>	 [telegram] <HikeAndMap> I was just about to write that exact same text.
[20:11:13] <wm-bb>	 [telegram] <HikeAndMap> But sadly it's the truth. The fact that modern servers have problems with media being used by wiki articles is the sad reality. And it actually forces me to split a summary page into many pages so at least users see everything, but not as a summary to scroll but instead like a book where you have to go to the next page and thus lookup is not possible on a quick view (re @chicocvenancio: that is a depressing 
[22:58:19] <wm-bb>	 [telegram] <jhsoby> I updated a gadget in no.wikipedia, and want to verify that my changes did what they should. However, when I load a page with ?debug=1, it always loads the script with &version=ztntf at the end (which I assume is a hash). Is there a way to tell ResourceLoader to not do this (temporarily, via a URL param or something)?
[23:06:48] <wm-bb>	 [telegram] <lucaswerkmeister> I don’t think so, but that hash should ensure you get the latest version anyways, if I understand correctly
[23:09:25] <wm-bb>	 [telegram] <jhsoby> not what's happening for me unfortunately. compare https://no.wikipedia.org/w/load.php?debug=1&lang=nb&mohttps://no.wikipedia.org/w/load.php?debug=1&lang=nb&modules=ext.gadget.verifiable&only=scripts&skin=vector&version=mysuperrandomhashs=ext.gadget.verifiable&only=scripts&skin=vector&version=ztntf (cached version which is what loads for me) with this (current version)
[23:10:58] <wm-bb>	 [telegram] <lucaswerkmeister> I see no difference between the two
[23:11:19] <wm-bb>	 [telegram] <jhsoby> really? oh.
[23:13:30] <wm-bb>	 [telegram] <jhsoby> yeah, now it's working for me too. I went directly to the URL I linked and did Ctrl+F5, and then it updated.
[23:14:03] <wm-bb>	 [telegram] <jhsoby> I thought Ctrl+F5 (in Firefox) would do that for all resources that are loaded on a page, but maybe not?
[23:14:48] <legoktm[m]>	 site-wide JS is cached in varnish for give or take 5 minutes
[23:16:17] <wm-bb>	 [telegram] <jhsoby> this was longer than 5 minutes, i made the change 30 minutes ago and it just worked now. But it only worked after I did Ctrl+F5 on the resource itself though.
[23:19:29] <wm-bb>	 [telegram] <chicocvenancio> Local cache? (re @jhsoby: this was longer than 5 minutes, i made the change 30 minutes ago and it just worked now. But it only worked after I did Ctrl+F5 on the resource itself though.)
[23:21:07] <wm-bb>	 [telegram] <jhsoby> yeah, must've been anyway, problem solved (for now 😆). thanks folks
[23:22:40] <wm-bb>	 [telegram] <lucaswerkmeister> I thought ResourceLoader would change the hash with each edit to make sure it’s not a cached version
[23:24:00] <wm-bb>	 [telegram] <jhsoby> what do you mean by "edit" here? All the various resources (which were loaded one-by-one because of ?debug=1) had the exact same hash
[23:24:40] <wm-bb>	 [telegram] <lucaswerkmeister> oh really
[23:24:55] <wm-bb>	 [telegram] <lucaswerkmeister> then I guess the hash only applies to on-disk code, not gadgets?
[23:25:06] <wm-bb>	 [telegram] <lucaswerkmeister> (It must be good for something, surely?)