[09:59:52] mhurd: coreyfloyd i've got swift integrated, but having problems w/ WMFLogo image for some reason [10:00:09] something about adding swift to the project is breaking image asset copying—just for that image [13:03:52] jzerebecki: hello [14:35:43] https://gerrit.wikimedia.org/r/#/c/212936/ [14:57:45] jzerebecki: jdlrobson2 where the browser tests at? [14:58:07] jzerebecki: want to hack some functional tests for iOS? [15:20:16] FlorianSW: mw.config.get( 'wgMFTocEnabled' ) === true for some reason [15:22:18] jdlrobson2: not in alpha or beta :/ [15:24:21] FlorianSW: i'm getting the error on this patch - https://gerrit.wikimedia.org/r/212945 [15:27:07] jdlrobson2: i don't get it, can you try another page with only one section or run ?action=purge ? [15:27:16] mediawiki [15:29:38] gilles: https://gerrit.wikimedia.org/r/212945 [15:29:44] i'm just sending a new version now [15:29:46] sent [15:29:54] added more encapsulation [15:30:12] jdlrobson2: added the optimisation of loading everything but the lead section [15:30:33] okay taking a look [15:30:42] (and then stealing) [15:31:33] infinite scroll coming soon.. [15:32:01] for some reason my stuff doesn't work in chrome in device mode [15:32:09] what's moushira's username? [15:33:04] jzerebecki: https://github.com/watir/watir-webdriver/issues/238 [15:33:36] ah, it's because I'm getting the no JS site [15:46:26] ori moushira [15:46:33] gilles: another thing we could explore is some kind of horizontal browsing.. [15:46:54] e.g. click item in table of contents loads that section [15:47:01] and replaces lead section [15:47:05] (obviously a talk with a designer) [15:47:56] gilles: https://gerrit.wikimedia.org/r/212945 infinite scroll not quite working but loads after 2s [15:48:07] (without toggling) [15:56:31] gilles: your change works for me in dev tools with iPhone 5 [15:57:17] on regular 3G simulation pointed at my VM firstPaint is halved [15:57:59] there is a penalty in terms of when things all appear, though, I think around 3 seconds [15:58:16] which makes sense because it takes time to load the initial pageload and we only request the rest then [15:58:32] instead of requesting it all from the get go with the status quo [15:59:10] we could use prefetching/preloader hints to improve that [16:00:09] i.e. the browser would find a pointing to the "allbutlead" mobileview API call for that page and start loading it as soon as it encounters it on pageload, before the actual mw.api call in JS is triggered [16:00:34] gilles: is this when you load instantly? [16:00:53] jdlrobson2: yes, at the bottom of mobile.startup