ehm cant we fix the herald rule instead ? tags shouldnt have to be removed simply because some automated system does something undesirable.

id call this an OS bug...

The upstream css linting library supports this since 20 days: https://github.com/CSSLint/parser-lib/pull/196

After leaving the edit page, the javascript of mw.confirmCloseWindow no longer knows what the original text was. It therefore assumes that there have always been changes.

Someone added a portlet link to the pageviews api using the save shortcut.
https://uk.wikipedia.org/wiki/MediaWiki:Gadget-Statistics.js

It has nothing to do with deferred loading of buttons. It's because of asynchronous execution and Pageread javascript that needs some serious rewriting..

Right, so @matmarex was right here I think, so incorporated his suggestion and made a patch to Translate extension that as a combination should fix this.

Right, my earlier proposed patch would solve this. I'll try to poke someone for expediting the merge and deploy

I'll first drop a line to the last person to have made several edits to their Common.css file. https://ar.wikipedia.org/w/index.php?title=نقاش_المستخدم:زكريا&diff=18547297&oldid=18496554

We might want to fork this ticket for full support of less, as described in the ticket https://bitbucket.org/birkenfeld/pygments-main/issues/780/add-a-less-lexer ?

Tested a bit. It just takes that long to generate the output. The request times out before the server is able to finish it. On reading this isn't too much of a problem, since the server caches that, but a preview is the first render of something, so here your wait time is simply larger than you browser request time.

Yeah, no luck there. I'm wondering if we should just ask a global interface rights person to intervene to remove those, and then leave their community a message referring to this ticket...

Actually, I should be a bit more nuanced. Less is now an alias for css in upstream. The merged patch was not rolled back, even though this was implied in a comment of the pull request. It's not really ideal however.

Hmm, I didn't see any difference in output so far with this filter in place.

The problem is what to do when you want to keep some item for later which requires follow-up.

It seems so. I guess we will have to look into it a bit deeper.

In T115528#2059159, @Pginer-WMF wrote:

While a single "mark all as read" action can be more effective the more pending notifications you have, in those cases the lack of control (potentially affecting actions out of sight) may be actually discouraging its use.

what the...
It seems that the first option in the languages dropdown has the selected state, but that's weird, since it's aa - Afar, and not en - English, which is the actual selected option visually....

not yet fixed in MediaWiki, which is at 2.1 instead of master (2.2 milestone) right now.

This mapping is now provided by upstream.

The primary issue is now at least consistent and not as visually distracting as before. I note that the transform is still not a recognized css property, but suggest that a separate issue should be filed for that.

To me this sounds like you are missing some of the caching layer optimizations, rather than a problem with this specific extension.

Reverted

Seems to me we need to check all conversions now. Three extensions with problems is a pattern...

Regression in CodeEditor as well, was reverted in:
https://gerrit.wikimedia.org/r/272817

This report is too vague. Please provide more detailed information with a specific example link, so that developers can follow up.

In T87276#2055795, @BBlack wrote:

Fascinating :) Any idea if it also throws an error on the other spelling that the specs have been confusing on, which is origin-when-crossorigin? From the error message, it sounds like it would.

This script uses the function recordUpload2, which is a bit lower level than where UploadComplete resides. While looking however, I notice that this function does have https://www.mediawiki.org/wiki/Manual:Hooks/FileUpload

Pinging @Ladsgroup, maybe he knows why ar.wp does this, or someone else who might know.
It's outside the purview of phabricator really, but keeping it open for just a bit longer won't hurt..

Some browser support dispatchEvent + insertText command
http://help.dottoro.com/ljuecqgv.php

var event = document.createEvent('TextEvent');
event.initTextEvent('textInput', true, true, null, text, 9, "en-US");
textarea.focus();
textarea[0].setSelectionRange(selection.start, selection.end);
textarea[0].dispatchEvent(event);

Nobody has been able to reproduce this as far as I know. provisionally closing for that reason, can be reopened if more information is available.

And this is already documented in case anyone was wondering.

In T126271#2056066, @CKoerner_WMF wrote:

I've got a draft here (with gifs!):

As far as I can tell all WikiEditor issues were dealt with. If similar changes need to be made to VisualEditor, please file a new ticket for VE, linking back to this one.

As far as I can determine this is no longer a problem.

Most of this is solved, other than the potential point of splitting up the bolding per line. Since we have not had any comments about this since 2011 however when this was introduced, I'm guessing it is not too much of a problem, And we should probably just close it.

This was fixed at some point in time. It now switches the status to the external link image after switching.

IE8 is javascript less now, so, calling this Fixed/no longer relevant.

"I have several text areas with WikiEditor toolbars attached on a single page."
Where and how do you have this ? That's not a vanilla MediaWiki situation, so will require special care.

For future reference: http://api.jquery.com/stop/

I see something similar in Chrome 49.0.2623.54 beta (64-bit).

From: https://en.wikipedia.org/w/load.php?debug=true&lang=en&modules=ext.wikiEditor&only=styles&skin=vector&version=e5a8a7e69161

I now get errors in Safari/Webkit console btw. I have therefore opened a feature request ticket with upstream.

In Javascript, you can use mw.config.get( 'wgContentLanguage' ) and mw.config.get( 'wgUserLanguage' ) to determine the language of the content, or the language of user interface respectively.

In my opinion, if this really is a browser bug, we should't invest time into it. Also Windows specific btw, on Mac OS X the navigation behaves like most other mac textfields.

Does that use uploading by url ? The transcode job is a hooked into the https://www.mediawiki.org/wiki/Manual:Hooks/UploadComplete hook if I remember correctly, so if transcodes don't start, then it means the upload path used isn't triggering this specific hook.

BTW. this seems like a change that could VERY easily be tested on English Wikipedia before we apply it everywhere, by using Common.js. Just an idea.

In T125317#2055171, @Danny_B wrote:

In T125317#2055165, @Devirk wrote:

Hello! I've already submitted a patch for this task. Please review it. https://gerrit.wikimedia.org/r/#/c/272708/ .

Weird, that @gerritbot didn't notify it here...

Actually.. without T127311: Echo congratulations thresholds shouldn't be hard-coded, there is no way to turn this off right ? Then I say we should definetly make sure that that array is configurable 'somehow'.

Okay, this is really weird, but I'm actually User:FallingGravity. Somehow I logged into this account when I opened my laptop, even though I never typed in the username and I don't have any clue what the password is (I checked and it's not the same as my actual password). Apparently this account used to be active on the Chinese Wikipedia. Anybody have any clue how this happened? Should I be worried about my account? Actually I think I'll just go ahead and change my user password just in case. 023yangbo (talk) 06:34, 23 February 2016 (UTC)

@Ricordisamoa you are not right, such a tag is only used for extremely high impact issues, or for the implementation of community requests, as mentioned in it's description.

@JMinor , yes i am logged in. Let me try some things.

I still have it today, even if i press refresh. Perhaps the error got cached ?

Grmbl i knew I was gonna regret that change...

Caused by https://gerrit.wikimedia.org/r/#/c/260071 of T108655

This will need to be filed with the Ace project.

I agree that especially as an initial testing ground, the education extension students might be a very good test group. I suggest to draft a more thorough plan about:

Can be built on the {T100106: Replace Kaltura player with Video.js} work that I have in progress (although quite a bit away from finishing that). Anyway, I think it's a nice idea for File pages, but i'm not so sure about transclusions.

The request is to make it so that neither registered users nor anonymous users can interact with the functionality the extension provides. The simplest way to do this, is to simply removing the extension from the setup as far as I know. I that case, all data will remain in the database.

We have a good starting point for this now. Closing, more tests will come as we need them.

I think the only requirement for such a hack on top of mw.collapse, would be firing an mw.hook forwarding the elements which should be collapsible after makeCollapsible has run.

I've closed the RFC.