Functionality merged, design implementation missing.
Tue, Nov 21
Mon, Nov 20
Sorry, I don't understand the question.
Fri, Nov 17
Did some amends on the last patch, have a look @Niedzielski and merge if you want
Thu, Nov 16
Wed, Nov 15
Initially we should set just one, like with the logo, and we'll have to discuss later how we swap them when the language changes.
I think most of the failures are due to the removal of the deprecated signature of stub stub(obj, 'meth', fn).
Tue, Nov 14
Mon, Nov 13
Fri, Nov 10
Thu, Nov 9
Wed, Nov 8
Is this a duplicate of T179833: PHP Exception: Call to a member function isMainPage() on a non-object (null)?
Maybe related to the banner with an image?:
For measuring the CSS & JS sizes you can use mw.loader.inspect, or in an incognito window open the network tab and see the total sizes at the bottom after loading the mentioned page (something like https://m.mediawiki.org/wiki/Special:MobileOptions?mobileaction=beta)
Tue, Nov 7
Mon, Nov 6
Thu, Nov 2
We're missing the design specs so let's assign to @Nirzar
@Fjalapeno @bearND this tix is/was mostly about the UI part for us, but it has a lot of conversation. I'm happy to make a new one for us and unlink it this one from our board if you want to preserve the convo here! Let me know what you decide.
@Niedzielski It is what it is if we want an explicit PageModule typing. It seems to me we should use default exports instead of a named page export, so:
Wed, Nov 1
Added the details from that article to the description grouped by source
This https://infrequently.org/2017/10/can-you-afford-it-real-world-web-performance-budgets/ has some interesting data, and it mentions the Moto G4 for last year and the Moto G5 for this year:
Quick note that section collapsing on mobile viewports is what we do now on mobile web, but it is very possible that the model will change in the future, based on product requirements from user tests. We have already seen that the apps are using a different TOC and section UX so that may be the direction that mobile web is asked to take in the future.
Tue, Oct 31
It seems like QUnit has nothing to do with the problem.
Should this be added to npm test to run on CI? Why does it need to be a new job?
Lots of good points. Wrapping the content is a nice convenience for reads but we could maybe get away by other means if it is a big problem.
Oct 20 2017
Thanks for the through explanation.
Some info I've found:
Oct 18 2017
Oct 17 2017
We should have good interaction data to know how quickly users interact with links from the event logging no?
It may be helpful to consider the personas and information from the new readers reports, and other profiles from the developed countries to have a comprehensive list.
Oct 16 2017
Oct 6 2017
Oct 5 2017
Thanks for the insights @ksmith
@Peter thanks for commenting! We just ended up planning the epics for the quarter so we're ready now to start with this.
Some related links to add to the RFC:
Oct 4 2017
As mentioned in the description, removing the @nomin banner could also work well giving the sourcemap comment on debug=true and having jsmin remove it on debug=false.
For reference, this involves unsetting https://github.com/wikimedia/mediawiki-extensions-Popups/blob/master/webpack.config.js#L35 when compiling for production.
@Jdlrobson To be clear I don't think we should use padding-start, since it is not well supported. We should use padding-left and which will flip to padding right on rtl languages. padding-left + janus is like the polyfill for padding-start. It doesn't seem like a hack to me, it is like a polyfill until we have padding-start supported everywhere.
See T177367: Initial version of Marvin user interface chrome for the chrome epic
👍 Good catch.
Oct 3 2017
It is not faking, it is how browsers do it by default! 😃