[08:03:10] woah [08:03:17] g'morning everyone [08:03:42] had a kernel panic last night caused, apparently, by virtualbox :/ [08:03:42] phuedx: morning [08:03:59] phuedx: ohhh, did you succeed in resolving it? [08:04:19] d3r1ck: well i'm back, so sorta ;) [08:04:42] phuedx: ok, thats good to hear. What is sorta? [08:04:52] slang for sort of [08:15:28] phuedx: i don't understand [08:15:57] phuedx: so does if mean soft of? [08:16:03] phuedx: ohhh, i get the poin now. [08:16:06] *point [08:18:23] d3r1ck: should've quoted it propertly: it's slang for "sort of" [08:21:08] ok phuedx i get it now [08:37:19] morning! [08:40:33] morning joakino [09:43:59] hey joakino [10:06:24] gonna go walk the dog, bbiab [10:14:13] joakino: ok, happy walking [10:20:28] phuedx, left a comment on your comment ;) [10:26:11] bmansurov: i've left a comment on your comment ;) [10:26:57] phuedx, right you are [10:28:21] "if content is king, context is god" and all that [10:29:41] context is queen ;) [10:31:02] bmansurov: it's a quote of a quote from the code review culture talk [10:31:55] then you should have put double double quotation marks. I thought it was just a quote and not a quote of a quote. [10:33:40] :P [10:34:18] gotta get to the shop to get lunch stuff for the fam [10:34:18] brb [10:47:34] joakino, yt? [11:50:20] joakino, phuedx|afk do you guys have a permission to create a repo in gerrit? [11:55:51] joakino, phuedx|afk could one of you guys run https://phabricator.wikimedia.org/P2166 ? Thanks [12:00:07] bmansurov|away: annoyingly, i don't have create repo perms [12:00:15] but then that might be a good thing [12:00:25] bmansurov|away: you have to create a task on phab iirc [12:00:33] i'll try and dig out the parent ticket [12:05:55] phuedx, thanks [12:09:20] phuedx, something like https://phabricator.wikimedia.org/T115003 ? [12:27:57] bmansurov|away: found the tag: https://phabricator.wikimedia.org/tag/wikimedia-extension-setup/ [12:28:06] took me a little while, sorry [13:03:19] phuedx, thanks [15:48:37] Why is MF still version 1.0? :P [15:48:38] https://github.com/wikimedia/mediawiki-extensions-MobileFrontend/blob/master/MobileFrontend.php#L32 [15:56:10] Reedy: it transcended versioning a long time ago [15:56:19] lol [15:58:07] that's the nice way of putting it [16:00:36] gonna take a break to have dinner with the fam [16:15:01] \o [16:15:12] Reedy: good old version 1 :) [17:01:53] jhobs: come to standup [17:02:12] jhobs: just kidding, got your update [18:12:46] yo [19:18:02] AWAY!!! [20:04:33] mbinder: don't wait on me, if you were waiting on me [21:02:55] mdholloway: still up for patch talk? [21:02:56] niedzielski: batcave? [21:03:00] niedzielski: yep! [21:03:02] mdholloway: omw [21:07:15] batcave? O_O [21:12:15] nanananana [21:52:06] jdlrobson: qq about textextracts... do we think it would be feasible to add an add'l parameter that adds some post-processing? [21:53:44] dr0ptp4kt: bearND niedzielski ^ [21:53:56] i don't know about you guys, but we're doing some stuff on the client to remove things like IPAs [21:54:25] it'd be nice to get that in PHP land so it can use (some) DOM selectors to do filtering instead of reg-ex stuff [21:54:34] or at least, we can all share (cached) versions of the post-processed output [21:54:44] cc jhobs phuedx|zzZ ^ and up [21:56:03] card / link preview / read more MW API / RESTBase proxy/Node.js service stuff will need to be cache performant to scale for the web [21:56:27] ditto text extracts if that's used [21:56:35] dr0ptp4kt: sorry, computer tanked afterm eeting [21:57:02] bgerstle:^ [21:57:08] dr0ptp4kt: that's why i'm suggesting we just put it under an add'l query param [21:57:37] bgerstle: you want it in php specifically, not restbase? [21:57:50] dr0ptp4kt: for the "card service" are there any talks of using something like a pageprop for this kind of "long description"? [21:57:55] bgerstle: we have post-processing of text extracts in the mobileapps service: https://github.com/wikimedia/mediawiki-services-mobileapps/blob/master/lib/extract.js [21:58:11] bearND: do you think that would be ready for production traffic by end of november? [21:58:57] bgerstle: should be [21:59:07] we'd be hitting it w/ a lot of searches [21:59:41] bearND: also, if all that does is post-processing on text extracts, i see little reason to not move it into the extension [21:59:53] unless JS is _that_ much faster a reg-exing that PHP [21:59:54] than* [21:59:55] bgerstle: ok, this is not a standalone thing. It could be. But right now it's used for the page preview [22:00:08] bearND: right, but we're getting multiple previews at once [22:00:24] so it has to 1) do our search 2) process text extracts [22:01:17] also, putting it into PHP might (depending on how TE is implemented) give you more information to do post-processing on [22:01:53] i don't know how TE actually does extracts, but if it has access to the DOM (which I think it does, since it can return HTML), it would be much easier to strip out IPAs etc. by using selectors [22:02:53] at the end of the day, adding an add'l query param to an already stable(?), or at least highly-trafficked endpoint makes it more appealing for us to use for our next release [22:03:18] we could then gradually switch over to something powered by RESTBase etc. [22:04:01] cool [22:04:01] bgerstle bearND: i think putting stuff in mw php land has definite gains but also definite disadvantages. that's not a free proposition [22:05:10] niedzielski: what are the disadvantages? we're already using TE, it hooks nicely into the generator system we're using to get stuff for the feed [22:06:28] bgerstle: no parsoid in stock install, much larger code base, additional platform (beyond restbase + android) [22:07:57] niedzielski: not sure what you mean by add'l platform. no parsoid included is fine w/ us (we're not using parsoid) and parsoid+MW for WMF wikis is pretty much a given. [22:08:34] anyway, i'm considering extending TE because we already use it in production. mostly looking to avoid risk at this point [22:08:44] Q3, anything goes ;-) [22:08:52] bgerstle: mw is an additional development platform. there is _some_ overhead in switching between dev environments [22:10:29] niedzielski: i think we still have a lot to figure out as far as how to balance responsibilities between MW & the micro services as well as how they relate to each other [22:11:14] that said, i don't know PHP ;-) which is why i was hoping to enlist a couple web-slingers and/or MW peeps [22:12:13] and like i said, am reluctant to put significant effort moving stuff over to RESTBase APIs. probably as a follow-on though. etags & such would be nice :-) [22:22:11] bgerstle: i think you guys should sync up with phuedx|zzZ & jhobs to see what's going on with the card / read more api/service stuff. they may already have something in mind that can fit the several scenarios across channels. [22:23:18] bgerstle: bearND: niedzielski would it be helpful if i arranged a meeting with those guys so we're all on the same page? [22:23:30] dr0ptp4kt: and bmansurov, he claimed the Cards card [22:23:40] task* [22:23:48] jhobs: are you and phuedx|zzZ working on that at all? [22:23:58] cool [22:24:05] jhobs: who's frontend and who's backend on it? [22:24:05] dr0ptp4kt: i probably don't have much anything helpful to add but i'm sure i would learn something from sitting in on it [22:24:21] dr0ptp4kt: phuedx|zzZ, yes, I'm reonboarding with jdlrobson atm and we're determining next steps from there [22:24:39] jhobs: word