[17:36:04] Vulpix: Specfic usecase [17:36:08] but it may not be obvious that's an image caption in that case. My personal recommendation is to use thumb to put it in a frame and add the caption that way [17:36:30] https://en.wikisource.org/w/index.php?title=Page:Tycho_brahe.djvu/111&action=edit [17:36:59] Here the captioned image has no border [17:37:36] and there have been various "too clever to work" attempts to put captioned images into paragraph runs [17:39:24] so use frame or thumb to add a caption, wrap the entire thing in a div with some css class. Add css rules to Common.css or TemplateStyles to hide the frame/border on child images of that class [17:39:35] ? [17:45:43] Skizzers: That would be the approach I was going to use.. [17:45:56] Are the defined classes documented? [17:46:15] so I know what I need to override? [17:46:57] Another concern arises, in that in the above example the page linked gets transcluded [17:52:07] and the image concerned is placed inside a Paragraph run. [17:52:45] Having images that can be embedded inside a P tag may be an insoluble problem as I don't think HTML5 will let you do that [18:15:44] Hi everyone. I have a question about the return values of the query API and continuation. The API returns a result that requires continuation (i.e. continue is set), but when I do continue, it returns an empty result with no errors. [18:16:06] Here's the query in the Sandbox: [18:16:08] https://uk.wikipedia.org/wiki/%D0%A1%D0%BF%D0%B5%D1%86%D1%96%D0%B0%D0%BB%D1%8C%D0%BD%D0%B0:ApiSandbox?uselang=en#action=query&format=json&maxlag=5&prop=info&generator=categorymembers&redirects=1&formatversion=2&inprop=subjectid&gcmtitle=Category%3A%D0%A1%D1%82%D0%B0%D1%82%D1%82%D1%96%20%D0%BA%D0%BE%D0%BD%D0%BA%D1%83%D1%80%D1%81%D1%83%20%D0%96%D1%96%D0%BD%D0%BA%D0%B8%20%D1%83%20STEM&gcmnamespace=0|1&gcmlimi [18:16:10] t=500 [18:16:25] Um, sorry, the link was too long but I wanted it to be transparent [18:17:14] The same link shortened: https://w.wiki/5q7 [18:18:57] So my question is - how to properly interpret such results? Should I simply silently stop if I see no "query" in the response? [18:24:52] A second question - if I'm querying for properties of pages, do I always get query.pages in the response? Or there are cases when "pages" might be absent yet "query" present when listing pages? [18:31:36] Does anyone know a lightweight way to parse templates? [18:32:03] like that doesn't requiring running mediawiki? [19:04:04] if I have an even split of slave databases (0, 1, 1) is there any reason why load on one slave would average a lot higher than the other slave? and sometimes it's vice versa [19:17:55] groceryheist: "lightweight way to parse templates" like what they produce or how their keys values inside a page or what ? [19:23:07] what they produce would work [19:23:36] I don't think I need to care about transclusions [19:24:04] you can use action=parse from the api.. but templates can have very page specific behavior and effects, so it's not perfect. [19:25:17] yeah I don't think that will work [19:25:48] groceryheist: wikicode sucks for things like this. it's also why VisualEditor is a hell ;) [19:26:13] if only we could go back 20 years to fix the organically grown monster. [19:26:23] for real [19:27:20] yurb: are you sure there are more than 500 results ? If i make you original request, i get 499 results... [19:28:03] hi thedj [19:28:13] Where is class="center" defined? [19:28:40] I'm looking for it so I can see what it does? [19:29:53] Here's a specific case I want to parse offline [19:30:11] ShakespeareFan00: https://github.com/wikimedia/mediawiki/blob/1ce6c003b882301ba978d24a596759bde6d91d20/resources/src/mediawiki.skinning/elements.css#L282 [19:30:15] ShakespeareFan00: there i think... [19:30:16] so I want to produce the undo summary given the user name and revid number [19:30:19] https://en.wikipedia.org/w/index.php?title=MediaWiki:Undo-summary&action=edit&oldid=788679594 [19:30:47] oh, that's even worse, because summaries are a particular subset of wikicode... [19:30:50] this is simple enough to write a simple regex or stack parser [19:31:19] but I wonder if there are any out-of-the-box solutions that might work [19:31:47] thedj: Okay and that's not in Mediawiki namespace is it? [19:31:50] Thanks [19:32:28] ShakespeareFan00: no that's part of core skinning (or the skin itself maybe in some situations) [19:33:12] Oh right [19:33:15] ShakespeareFan00: it's actually mainly defined for centered inline images if i remember correctly, other usages of it are more an accident of history ;) [19:33:39] And theres a reason why the img isn't styled as centered directly? [19:34:10] I have the problem of an inline image here [19:34:39] https://en.wikisource.org/wiki/Tycho_Brahe:_a_picture_of_scientific_life_and_work_in_the_sixteenth_century/Chapter_5 [19:35:03] But it's on it's own on a single PAGE in page namespace [19:35:18] Currently using conventional image syntax would put a DIV inside a SPAN [19:35:23] which is BAD [19:35:39] Is there a way to tweak the image syntax so it can be done inline? [19:36:16] Using ShakespeareFan00: i'm not entirely sure what you want to achieve yet... [19:39:55] thedj: Did you see the linked example? [19:40:26] There is an image that occurs WITHIN the flow of a paragraph [19:40:41] Elsewhere to avoid this problem I've relocated images [19:40:49] outside of a nominal paragraph [19:42:18] Currently on English Wikisource, trying to use Image syntax directly to put images like this in the run of text leads to Linterrors and misnested formatting [19:42:27] because DIV's can't be placed inside P [19:43:07] If an IMG can't be inside a SPAN then as I said having to force a pargraph break where it's inconvenient is unavoidable [19:43:30] and complex to implement user side. [19:43:54] ah wait, once sec... [19:46:43] thedj: I've had related issues with content like sidetitles that need to be DIV based for certain features, but have to be nominally inside P because of where they occur in a text flow [19:46:52] oh.. i remember... this is because of proofread's trick to bind paragraphs split over multiple pages back together into a single paragrpah if you are using [19:47:04] Yep [19:47:07] ah nasty trick [19:48:06] What should really happen is that the 'tidy-up' should force the image wrapper code outside the paragraphs.. [19:48:39] In the way that Mediawiki now relocates malformed insertions in tables to outside of the table [19:48:42] yeah.. but that entire thing doesn't really have knowledge of html and wikicode really. [19:48:50] hi [19:48:56] Does anyone know why there's this issue with this page on mobile? [19:48:57] https://meta.m.miraheze.org/wiki/Terms_of_Use [19:49:13] thedj: Don't get me started on the sidenotes problem ;) [19:49:28] what's the issue? [19:49:30] That one is a limitation of HTML [19:49:31] it's one of those 'hacks' that sort of work, but not always, because well.. it was never designed to do all this... [19:50:02] Proofreadpage is a hack over a hack... [19:50:19] ShakespeareFan00: i have no solution for this, it requires someone going into the code and making adaptations to the parser, specifically for parsing inside [19:50:42] Should I file a Phabricator ticket in the next few days then? [19:50:44] certificate for https://meta.miraheze.org is wrong [19:50:49] and then all existing hacks will break and people will be pissed ;) [19:50:57] Yes.. [19:51:00] (missing intermediate CA maybe) [19:51:05] ShakespeareFan00: i think there already is a ticket about this... somewhere... [19:51:09] which means, DEsktop version is not accessible [19:51:09] Sometimes I wonder if it's worth breaking things big time ONCE [19:51:12] but if not, make a new one. [19:51:46] thedj: Proofread page also has a habit of mangling tables that extend over more than one page [19:52:11] I already put a ticket on Phabricator asking for a proper "continuation" syntax to address this [19:52:45] ShakespeareFan00: the problem is.. no devs to work on this.. like ever [19:53:27] thedj: No devs to wormk on many things.. [19:53:36] Yet the WMF has all that money.. [19:53:39] :( [19:53:50] put it in the Wikimedia runs using 300 people but realistically requires 3000 people to run it, but then ppl get pissed because the WMF spends money on devs and not on content creation. [19:54:43] and when they spend money on both, they become legally responsible for the content so that won't happen either == dead lock. [20:57:12] thedj: that's the thing - there are 499 results yet for some reason it sets the "continue" parameter, i.e. suggests there are more [22:30:01] Emergency [22:30:01] There is something wrong. [22:30:01] I need to talk to the administrators at [22:30:02] higher level of Wikipedia Italia and the Foundation. [22:30:02] You can contact me here. [22:30:02] antolepore AT gmail.com [22:30:03] I can't write a simple important page because a block that constantly banishes the user starts and consequently it is impossible for me to get answers. [22:30:03] "L'indirizzo IP attuale è 93.43.180.184, il numero ID del blocco è #563379 Si prega di specificare tutti i dettagli qui inclusi nel compilare qualsiasi richiesta di chiarimenti" [22:30:03] I wanted to understand why is this happening? [22:30:04] I would like to resolve the issue in the best way. [22:30:04] Thank you [22:32:14] okok, please don't go off-topic in this channel [23:12:13] if I have an even split of slave databases (0, 1, 1) is there any reason why load on one slave would average a lot higher than the other slave? and sometimes it's vice versa [23:52:35] Okay Where's the glitch? [23:52:48] https://en.wikisource.org/w/index.php?title=Page:Handbook_for_Boys.djvu/422&action=submit [23:53:12] Apparently a DIV SPAN combination is the wrong way round, but I can't without more information figure it out [23:53:25] making the output from Linter USELESS [23:53:51] Perhaps someone here when they actually have some time to help resolve an actual problem can provide an explanation? [23:54:11] Calm down and be patient [23:56:11] N3X15: I am calm, but increasingly frustrated at having to play hunt the undocumted minutiae [23:56:23] EVERY single time I am trying to solve a problem [23:56:26] Welcome to old web apps [23:56:39] Now you know why devops get paid a lot [23:57:34] Mediawiki should be better than this [23:57:45] It's a shame it isn't (sigh) [23:58:37] Feel free to update the docs [23:59:13] N3X15: An unhelpful response, because if I knew how to do that, I wouldn't need to be asking for assitance with my concern or issue [23:59:28] If you don't know the answer, at least say so