[02:45:12] Okay anyone able to explain something? [02:45:24] https://en.wikisource.org/wiki/User:ShakespeareFan00/Sandbox/Firstletter [02:46:01] and [02:46:02] https://en.wikisource.org/wiki/Page:Ruffhead_-_The_Statutes_at_Large,_1763.djvu/84 [02:46:10] The former works but the latter doesn't [02:46:33] Define works [02:46:36] Is it too much to expect identical coding to produce CONSISTENT renderings? [02:46:51] AntiComposite: In the former a custom CSS class is defined [02:47:15] https://en.wikisource.org/wiki/User:ShakespeareFan00/common.css [02:47:33] I was working out drop-cap formatting [02:48:01] In the sandbox defining the class on a paragraph works [02:48:13] In the actual page where I wanted to use the approach it does not [02:48:21] This is inconsistent [02:49:30] and I get slightly annoyed that I am again having to run-around playing hunt the pedantic combination of exactly the right sequence to use [02:49:39] rather than actually being able to code up what I want to [02:50:47] It's not unreasonable to expect the same rendering output for the SAME input markup [02:51:10] Did you bypass your cache? [02:51:15] Yes [02:51:55] First thing on my list when I update things [02:52:09] Works for me, could not reproduce. [02:52:19] Which browser [02:52:25] and Screenshot please? [02:52:49] https://en.wikisource.org/wiki/Page:Ruffhead_-_The_Statutes_at_Large,_1763.djvu/84 [02:52:54] being the page that won't [02:52:58] Chromium Version 72.0.3626.96 (Official Build) Arch Linux (64-bit) [02:53:21] AntiComposite: I use a Firefox Nightly build [02:53:27] so it should also be working here [02:54:11] The last paragrpah on the page should have a drop-caps I [02:54:40] https://i.imgur.com/MMwlgBk.png [02:55:07] Yes... That is what it SHOULD be displaying [02:55:18] So I'm going to start screaming [02:55:26] because it's not doing that for me [02:55:39] Check your cache again. [02:55:50] I've done that MANY times [02:55:57] Even re-loaded the page [02:56:10] Try also using the style in the Firefox devtools. Couldn't tell you how exactly on Firefox though [02:57:00] Well if it works for you it's clearly not my code that is wrong [02:57:25] The I is still displaced though [06:55:03] Is it expected for clients to be limited to about 2MB/s for downloading from dumps.wikimedia.org? [06:56:46] IIRC, yes [06:57:25] https://dumps.wikimedia.org/mirrors.html [06:57:45] ^ phuzion, yeah, it's generally suggest to use a mirror [06:57:55] Ok cool [06:58:12] Or a download manager that can pull from multiple mirrors ;) [06:58:30] https://tools.wmflabs.org/dump-torrents/ is a bit out of date unfortunately [06:58:33] your.org is maxing out my downstream so that works :) [06:58:55] Or close enough to it to matter. [06:59:50] heh [07:02:24] Reedy: Do you have a recommendation for a multi-source download manager by chance? [07:04:02] Not offhand. I think they've grown out of fashion somewhat [07:04:30] IIRC flashget could do it [07:04:51] Alright. I'll see if JDownloader can do it, and if not, I'll just deal with "only" getting 95mbps lol [07:05:19] I think if you're almost maxing out your connection, it's probably not worth it [07:08:56] Yeah I spent like 1-2 minutes poking around JDownloader (already had it installed) and it doesn't appear to easily be able to do that, so screw it. [10:33:35] https://en.wikisource.org/wiki/Page:Ruffhead_-_The_Statutes_at_Large,_1763.djvu/84 [10:33:41] Somthing is clearly NOT working [10:33:54] I defined a custom CSS class to implement a first letter behaviour [10:34:03] Except it's not consistently applied [10:34:11] Compare the above with [10:34:22] https://en.wikisource.org/wiki/User:ShakespeareFan00/Ruffhead [10:34:53] Where it is [10:35:07] As the latter transcludes the former there shouldn't be a difference [10:35:16] What is getting in the way please? [10:40:27] Its not obvious to me what I'm supposed to be seeing [10:41:07] Are we talking about the "N" being a drop capital? [10:41:11] No [10:41:23] It's the drop capital I on the subsequent paragraph [10:41:48] I know the drop capital I works because it's using an existing template [10:41:56] Ah ok. For me it is not present on both the linked pages [10:42:01] Okay [10:42:43] So we have a completely inconsistent and random behaviour then [10:42:46] :( [10:42:47] Sigh [10:43:03] or you just have a different cache [10:43:21] I've tried purging and reloading BOTH pages MULTIPLE timesa [10:43:54] I should NOT have to repeatedly reload pages to get a consistent rendering [10:44:11] If Mediawiki can't give the same rendering for the same code [10:44:16] then it's broken [10:44:34] Where is the CSS you defined live? [10:44:49] https://en.wikisource.org/wiki/User:ShakespeareFan00/common.css [10:45:12] Well i mean, that's probably why i'm not seeing it [10:45:16] I was trying to come up with a cleaner way of handling the drop-capital/initials [10:45:18] If its in your personal css [10:45:40] the firstletter behaviour can't be done inline [10:45:56] because it's a selector [10:46:21] The next approach would be to use a TemplateStyle [10:46:45] Unless you know of another way of implementing the custom CSS so ALL can see it [10:47:05] without needing an interface admin to implment it globally [10:47:38] I know, I just mean, if you don't tell me its in your personal css, then I can't see it during testing, as i'd have to add it to my personal css too [10:48:58] Okay... [10:49:15] Do template styles only work in Template namespace? [10:49:24] Or does the tag work anywhere? [10:51:09] https://en.wikisource.org/wiki/Page:Ruffhead_-_The_Statutes_at_Large,_1763.djvu/84 [10:51:22] So I create a seperate CSS file for my experimental code [10:51:32] and use a TemplateStyles tag to import [10:51:40] only to get an error about content models.. [10:52:28] Well I tried to come up with something that worked [10:52:40] What a shame Mediawiki seems determined to be pedantic [10:52:42] :( [10:53:15] Maybe, you can't float: a pseudo-element [10:53:22] No.. You can [10:53:29] because elsewhere the example works [10:53:52] The issue about content-models aside [10:54:44] When i imported your css, it seemed to work [10:54:58] Hmm [10:55:09] Then I perhaps should just the ditch the whole effort [10:55:17] as it still doesn't work for me [10:55:34] You can change the content model via https://en.wikisource.org/wiki/Special:ChangeContentModel/User:ShakespeareFan00/experimental.css [10:55:45] No I can't [10:55:49] I'm not an admin [10:56:20] and I thought the whole point of TemplateStyles was so that experimental stuff like this DID NOT NEED admins to change stuff in the first place [10:56:25] (Sigh) [10:56:41] It's things like this that make me somewhat frustrated [10:57:05] What if you try to create the page as https://en.wikisource.org/wiki/User:ShakespeareFan00/experimental2.css?action=edit&model=sanitized-css&format=text/css [10:57:32] Nope [10:57:36] Doesn't work [10:57:46] Still get a permissions error [10:58:04] Hmm, blame enwiki. It was originally just autoconfirmed and then they got angry :P [10:58:19] Ok, i guess you can only do it by creating a subpage in template namespace [10:58:50] Well thank you for convincing me that my approach is impossible because no-one can ever actually implement something usable XD [10:59:30] Anyways, with your css imported, i get the drop cap, but its vertically much further down then where its supposed to be [10:59:40] Yes [10:59:51] That's a known issue, and is to do with float handling [11:00:38] An apparent limitation of CSS/HTML in their current forms [11:01:08] Having sensible dropcaps is planned for the next version of CSS [11:01:37] but it isn't implemented in any major browsers yet, so I am stuck with the broken version for now [11:02:57] bawolff: Do you get the drop cap on bothe the Page: namespace and it's transcluded version? [11:03:08] Because I only see it on the trnascluded version [11:03:20] yes [11:03:32] Well then I don't understand something [11:03:35] because I don't [11:03:56] Maybe it's just time to say I'm too stupid to code this [11:04:06] and admit defeat [11:04:43] Do you know how to use the inspector in the web developer tools? [11:05:15] You could use it to check if the sanity class is there, and which css rules are applying to that paragraph [11:05:55] It would be weird to be a css caching issue, as the other page works [11:06:13] bawolff: I've spent far too long on this to debug it further [11:06:37] and yes checking if various classes were being applied was something I thought I had checked [11:07:03] Maybe somehow you are getting old html. But that's be weird too. Maybe if you have non-canonical parser output that for some reason wasn't purged. Although that would also be a very weird thing [11:07:51] Well according to the CSS inspector the CSS class is loaded [11:07:56] but nothing happens [11:08:23] So I am now frustrated because of some pedanticly specifc interaction that seem impervious [11:08:30] to attempts to figur it out [11:09:34] As I've said repeatedly elsewhere [11:09:52] It would be nice if I did not repeatedly have to rely on playing "hunt the glitch" [11:10:14] or random obscurity when trying to develop consitently rendered content [11:10:28] It would also help that when i've repeatedly pointed this out [11:10:49] that people actuallly LISTENED and fixed the problem [11:11:01] *consistently [11:11:22] You've been very helpful [11:11:49] in helping diagnoise what might have gone wrong. [11:12:21] but it doesn't solve the fundamental problems that currently drop initials are something that Mediawiki can't fully support [11:12:29] Time for a different approach... [11:12:32] * ShakespeareFan00 out [11:12:45] Really more of the browsers fault on that one [11:12:54] Or W3C perhaps [11:13:06] * bawolff passes the buck [11:18:46] bawolff: agreed [11:19:08] As a developer, passing the buck is my most important skillset! [11:20:08] Raised here - https://phabricator.wikimedia.org/T215750 [11:20:25] because I'd like to rule out it being a mediawiki glitch [11:20:38] before I start yelling at browser vendors [11:53:16] VE running on IE - is it supported? [11:53:45] I don't know the version number because the reporter did not specify the version (nor they seem to know how to fetch it) [11:56:25] found the answer: https://www.mediawiki.org/wiki/Help:VisualEditor/FAQ "IE 10 or 11" [12:13:04] anomie: could you ping me later today if you have a few minutes to chat about using the comment table’s comment_data field in an extension? [12:13:11] (otherwise I’ll ask around elsewhere) [14:03:21] Lucas_WMDE: I'd be happy to chat, although I really do only have a few minutes right now. The field was added to support things like Wikibase's i18n of edit summaries, but support for actually using it in core beyond CommentStore saving/loading it is probably somewhat lacking at the moment. Patches to make it more usable would be welcome. [14:37:10] anomie: that’s exactly what I suspected, just wanted to confirm the impression was true, thank you :) [14:37:27] any estimate on how much work it is would be nice, but otherwise I’ll just try to make my way through it for now [14:37:45] currently trying to update RecentChange.php, and then I guess the methods to render it (Linker?) [14:41:32] mainly just wanted to confirm there’s not some hidden layer of support beyond CommentStore + CommentStoreComment that I’m missing :D [14:48:01] This is now getting silly [14:48:01] https://en.wikisource.org/wiki/The_Statutes_at_Large_(Ruffhead)/Volume_9/Doncaster:_Small_Debts,_Lighting,_etc._Act_1763 [14:48:06] I update a template [14:48:14] so that a particular fix is implemented [14:48:31] and the template code seems unwilling to even notice the fix [14:48:47] Can someone else please take a look at figure out what the **** is wrong with my code [14:49:10] and why Mediawiki seems determined to force me into playing hunt the bug EVERY single time i try to make a useful improvment? [14:49:14] *improvment [14:49:29] *improvement [14:49:57] Diff? [14:50:30] https://en.wikisource.org/w/index.php?title=Template%3AOutside2&type=revision&diff=9100921&oldid=6706763 [14:50:40] It's a simple parser function [14:50:54] More diffs here [14:50:54] https://en.wikisource.org/wiki/Wikisource:Scriptorium/Help#Sidenotes_(ongoing) [14:51:15] Somehow the relevant parameter isn't getting passed down a template chain properly [14:51:48] Either that or clear doesn't APPARENTLY work on spans [14:53:08] The sidenotes should not overlap [14:53:35] And adding a 'clear' in the CSS is the approach taken elsewhere [14:53:59] such as with {{MarginNote}} where the clear: approach means they do not overlap [14:56:15] Lucas_WMDE: It might be that all you'd need to do would be to allow RecentChange to accept a CommentStoreComment object instead of only accepting a string, and whatever else (I think nothing?) is in between RecentChange and CommentStore. Then pass it a CommentStoreComment that was created using a Message rather than a string. [14:57:10] anomie: well, I assume I’ll also have to adapt Linker::formatComment() to accept a CommentStoreObject instead of (well, in addition to) a string [14:57:23] and then figure out how to even turn the comment_data into a string to display (a hook, I guess) [14:57:40] (because comments are on the revision level, not the slot level, so we can’t ask a content handler to render it) [14:58:25] btw, thanks for CommentStoreComment::newUnsavedComment() – the fact that it also accepts CommentStoreComment objects makes it very convenient to use :) [14:58:31] Lucas_WMDE: ... Re Linker, maybe. Re turning comment_data into a string, hopefully not as long as you originally created the CommentStoreComment by passing a Message rather than a string. [14:59:19] If the CommentStoreComment contains a Message, CommentStore should be saving that Message into comment_data, and pulling it back out on load. [14:59:30] okay, so we (extension) shouldn’t do anything with comment_data directly? [14:59:37] and instead just use the Message (if possible)? [15:00:38] Lucas_WMDE: Definitely not with comment_data directly. Access it via CommentStoreComment's ->data member, and set it via the $data parameter to the relevant methods. [15:00:44] * anomie has a meeting now [15:01:01] okay, thank you! [15:01:10] that definitely clears things up [15:01:13] I’ll look into it further then [15:05:50] AntiComposite: Why I don't understand, what's going on is that in all the templates I should be using near identical coding [15:06:05] As I've said in the past consistency would be appreciated [15:06:07] :( [17:03:41] I'm not sure if I'm in the right channel for this, but a portion of upload.wikimedia.org seems to be down. [17:04:06] Not a single image on Wikipedia seems to load for me. [17:04:30] hi [17:04:39] ops are aware [17:05:05] Alright, good to know. Thanks. [17:05:50] Looks like images are redirecting to themselves [17:06:08] Gaelan yup, ops are aware. [17:06:17] great [17:08:28] Paladox - https://en.wikipedia.org/wiki/Wikipedia:Teahouse#Can_anyone_help? - This affected??? [17:08:44] anything that touches upload will be affected [17:08:49] Would it be best to put out watchlist notifications? [17:09:03] though yes, that looks like the same problem. [17:09:09] maybe [17:09:41] Where would be best to request watchlist notifications, or get an alert out. [17:09:55] Im not sure. [17:29:14] Paladox - Any update? [17:29:29] yup, things recovered. [17:30:13] Good.That was quick