@Xaosflux it's a problem because it has to be implemented in the parser and it requires a two pass to do so. So first the entire wikicode needs to be parsed and converted into html (which is currently a string concatenation effort mostly) and then the whole thing needs to be losslessly (so with whitespace and comment preservation) loaded into a DOM tree, manipulated an serialised back to HTML text. While Parsoid in theory can do that latter step, Parsoid is not yet part of our core and I have no idea about the performance of such a setup, but I doubt it will be trivial. Lastly, someone has to reimplement 1300 lines of table sorting code from JS to PHP of course.

@Aklapper support closing this tracking ticket

TIL: PHP had array and string offset access with curly braces

FYI: I use an iPhone X with Safari 13.0.3 (iOS 13.2.3) and it's not a problem for me.

this can definitely be marked as solved yes..

I think the audio player will be a problem. Native players are pretty big and we often have very restrictive spacing. Don’t know a good design either though..

Was it considered to serialise additional information inside HTML comments ? Those wouldn't interfere with the render output.

Original introduction of mw-content-DIR

TMH has not used any plugins for a while, at least for Wikimedia. I guess in theory you could configure it to make use of the VLC plugin, but VLC doesn't support NPAPI either any longer and when we switch from Kaltura to Videojs.

Testing this requires you to make use of the Spanish input method in Keyboard Settings -> input sources -> add Spanish

"syntax highlight enabled" so this is about the CodeMirror editor syntaxhighlighting and not about the wikicode content highlighting with <syntaxhighlight> feature.

Do we want to reuse wgMinimumVideoPlayerSize with this new lazy load setup ?
https://gerrit.wikimedia.org/r/550926

@Soda yes, this is absolutely fine, although i'm not sure if that is gonna be much easier that way, due to different loading methodologies of the UI components. But give it a try.

@Soda yes, this task is still open and you can work on it if you want.

I think it is. As a matter of fact, maybe it should even go into core (with <pre> being a core element).

The only effect of overflow-x:auto is to create a new 'block formatting context'. Practically this means that these kinds of blocks avoid drawing their background behind floating elements like inboxes. (The H2 headers have the same setting, to avoid their underline runs behind the infobox).

With new notifications there’s another issue, apparently the “time stamp” is always hidden outside the right window border; however, here at least the screen size isn’t affected, so you can scroll sideways to see all.

So the problem here is that two players share a 'caption screen'. When you pause a track, at some point it hides the captions. It doesn't remove the current caption, just in case you continue playback later on.

Nicely discovered !

<syntaxhighlight> on en.wp already does wrapping... I see no objection to making that more universal...

Thank you @Florian !

Yeah if someone can give a list of a few images that triggered this, then this can be looked into. I do remember there being a limitation with our Coordinate parsings in Exif being left in there... we might have to trawl through some old commit logs / tickets to check what that issue was.

I think that adding a 'pager query' category to wgRateLimits that is used for these types of requests would be the most ideal.

the most likely cause for this, is simply a broken request that got cached in the browser. Try to bypass the cache: https://en.wikipedia.org/wiki/Wikipedia:Bypass_your_cache

In T236743#5614758, @WanderingWanda wrote:

I see this error: "Failed to set referrer policy: The value 'origin-when-crossorigin' is not one of 'no-referrer', 'no-referrer-when-downgrade', 'same-origin', 'origin', 'strict-origin', 'origin-when-cross-origin', 'strict-origin-when-cross-origin' or 'unsafe-url'."

FYI, the blob is used to load a JS worker. JS workers cannot be loaded cross domain (no support for CORS). We used to have the JS hosted on the static domain and it required this method to do the loading... (it was actually added to Ace specifically for us, by me). It seems that since Ace updated this worker loading to ALWAYS use a blob.. Not sure why, though commit logs seem to imply to avoid issues on IE..

Small note. Generally we had much longer periods with deprecation messages being emitted and the admins would clean up the scripts they used themselves. This limited impact.

Because IE6 support (and it's content-analyser) is now dropped (T232563), this ticket might now be implementable for those who want to have a go at it.

I played a bit with bootstrap's grid layout and templatestyles: https://www.mediawiki.org/wiki/User:TheDJ/bootstrap

I think we can conclude that we are never going to test IE/Edge behavior for the old spec for this right ? :)

Another reason the en.wp method is better than forcing whitespace with padding, is because using actual spaces causes the screenreader to see a word break between the element content and the content of the ::after element.

PS. i'm pretty sure the usage of border-bottom over text-decoration: underline is intentional here, but it's not documented in the CSS sourcecode.....

just noting that en.wp uses:

.hlist dd:after, .hlist li:after {
  content: " · ";
  font-weight: bold;
}

Not sure why minerva uses something else.

@Theklan that seems more in the scope of the themes T234087: Theme: Local Development and Onboarding and T234091: Theme: People and Processes. Deploying and hosting is about anything after the code (including an integration) has already been written from my point of view.

Maybe we can discuss about an incentive system to review code. I feel with the current setup of teams it is very hard to incentivize anyone to review code outside of their own responsibility. It's basically a favor that people are doing to eachother and while nice at the human level (and that will probably always be the primary motivator to use), maybe there is more we can do to encourage people to reach across team/volunteer/3rd party lines..

I'd be interested in code health and it would be lovely if we could see about getting some statistics from the past year for instance, see if we can also find some empirical evidence next to the anecdotal evidence about progress made etc.

Not to be blunt but... this mostly seems like having to attend a homework session for WMF/WMDE :)
I'm not opposed to the concept of this, as a volunteer i'd love a more predictable method for me to interact with teams as well, but..... this doesn't seem like a session i'd join at the end of a long day...

Something that I'm missing here honestly, is the dedication to educate and teach new developers/team members as a core principle, before tossing them into the deep end... This is honestly more of an internal WMF thing to decide on, but work in that area can easily benefit any contributor and I think WMF is rather bad at this area (possibly because of their history of people who 'grew up with the codebase'..

Upstream states: "I can't reproduce. Could you try Canary?" and was then merged into https://bugs.chromium.org/p/chromium/issues/detail?id=1002442 and marked as fixed.

I really like this idea. Being more flexible in experimentation and rapid deployment seems important to me.

So we would potentially discuss things like red/blue, checksumming, a better maintenance mode for upgrades, database changes rollback, the web installer and a web configurator ?

Good points being raised. Personally I see beta as actually two things.

• shadow service (use the latest software against existing data for verification), which it is not, since it uses separate data structures
• pre-production staging of significant software configuration changes and features. Which it is not, since it doesn't really reproduce production

so.. some of this theme seems rather WMF focused.. Have we considered taking lessons and input from 3rd party MediaWiki hosters within this theme ? Maybe even have a separate session to learn about what they are doing ? Or at least validating the WMF ideas against their usecases ?

What is the problem that led to this session proposal ?

I'll just say: https://laravel.com/docs/6.x again ;)

The description seems clear yet when I read it, I still have no idea what we would be doing in this session.. Identify pain points ? Throw ideas at the wall and see what sticks ? These things tends to get incredibly broad and vague ('better documentation')

For newcomers getting started in a situation that is easy to setup and has proper defaults (config and extension) is most important. To me that means as much composer driven installation and config as possible (because its the php ecosystem default). Where composer is not possible (due to services in different ecosystem etc), composer should inform users what to do next.

Hi, just a quick note, that due to a very busy and long week for me where I am at an exhibit for work, I won't have time to read any of this before the 11th, when I return from this event. I'm happy to spend some time during the weekend, to still go through everything and provide commentary and you can do with it as you desire.

Note: Without contentanalyzer, we probably can also support KML uploads.

The size of the original was too large to upload to phabricator I think (just over 10MB). Saved at half size:

@matmarex i think at the very least, when we trip the limit, we would want to have that proactively logged, regardless of the solution ?

Pretty sure now this was limited to an older Testflight build I had installed.

wow. nice sleuthing !!!

so.... turns out we are missing the XMP data from a significant number of files most likely. I suspect that Photoshop actually triggers this bug, so ;)

i had the user field selected and it did not fill the fields :(

https://commons.wikimedia.org/wiki/File:Ford_Grand_Torino_(29725737128).jpg
RuntimeException: Expected <rdf:Bag> but got http://www.w3.org/1999/02/22-rdf-syntax-ns# Seq.

i guess so... ;)

One problem, the current database structure only allows one set of 2fa credentials per user.... :(

Note from webkit:

• [WebAuthn] Enable WebAuthn by default for MobileSafari and SafariViewService

I like this improvement a lot ! The messages do indeed still need some work.

works fine for me with Chrome 77.0.3865.70 (Official Build) beta (64-bit) MacOS. What exact version do you have @SerDIDG ?

@Samwilson i think that is not a bad idea...