[00:25:05] mdholloway|afk bearND dbrant: ok, the beta is out. let's keep our eyes peeled for HA reports [00:36:39] \o/ Thanks! [00:44:12] leila: wait, is your IRC lzia or leila? lol [00:45:32] both jhobs. :D [00:47:51] leila: haha ok. Anyways, I updated the task you created because I had a misconception about closing the survey. It'll require a SWAT deploy no matter what, so any day between Mon and Thurs is fine. Please see my comment for more info [00:50:44] great jhobs. responded there. [00:52:12] btw, jhobs, the EL table is not created in the usual database. if it's not there by tomorrow, we should look into it. That won't hurt us in this survey, but it's good to figure out where the data is being stored. [00:53:59] leila: ok. An email to reading-wmf might be helpful for that. Like I said, I'm not too familiar with EL so JK or Baha will probably be able to respond there [00:54:34] got it. I'm trying to send everything through atgomez. I'll ask her about that, too. :-) [00:55:25] ok cool [00:57:26] o/ [00:58:03] hi baha, bye baha lol o/ [00:58:11] hi [00:58:13] bye [00:58:24] freakin' deployments man [00:58:37] what happened? [00:58:52] oh nothing, just online 2 hours past usual 3 days in a row lol [00:59:10] :-\ [00:59:11] oh yeah, sorry [00:59:54] part o the job :) [01:01:28] yeah [09:03:47] hullo [09:43:00] morning! [09:48:55] o/ [09:54:49] hey [10:00:28] bmansuro_: i asked chris about the preferred way of making cards more library-like [10:01:34] and? [10:02:01] he said he had no preference but we should ping deployers (or anyone with more experience with us, i guess) about whether or not to just leave it as an extension [10:02:43] i see. well if security review can be done so, i'm for leaving it as an extension [10:02:49] will we get a security review soon? [10:04:42] soon* [10:06:07] did you go through the code? [10:06:45] joakino: me? [10:06:56] phuedx: can you clarify your 2nd comment in https://gerrit.wikimedia.org/r/#/c/249991/5/resources/models/Card.js ? [10:07:03] no phuedx & csteip [10:07:12] we did [10:07:23] i'm just phabricating and then emailing notes [10:07:36] phuedx: are you saying I shouldn't create the event on the fly? [10:07:51] for chris's approval and the team's santity [10:08:10] there is a security issue that i've just captured here: https://phabricator.wikimedia.org/T118553 [10:09:07] bmansuro_: in that patch, the event that you're raising will /always/ look like { key: "valueOfValue" } [10:09:28] and not "{ nameOfAttribute: "valueOfValue" }" [10:09:35] which is what you've written in the documention [10:09:55] sec [10:11:28] phuedx: ok i'll update the documentation. do you think I should still create the event object and pass it? if so, why? [10:13:07] bmansuro_: try adding this as a test https://www.irccloud.com/pastebin/DzSK9dgl/ [10:21:02] phuedx: ok i see now [10:21:09] f'ing javascript [10:21:17] lol [10:21:29] bmansuro_: i was having a hard time putting it into words, mibad [10:21:43] unit tests > my use of the english language [10:21:45] but your tests were great [10:22:03] speaking about tests, i think i will add them as part of the patch [10:22:08] <3 [10:31:30] phuedx-o what was that useless html in barak obama with 800 nodes? [10:31:33] do you remember? [10:31:45] i'm writing notes of bernd's meeting yesterday to send [10:31:53] ? [10:32:33] i take it we're not talking about the insanely long references list? [10:35:03] no no [10:35:19] phuedx: some i can't recall now [10:35:49] oh yeah [10:35:52] document.querySelectorAll('[typeof="mw:Entity"]') [10:35:53] found it [10:54:36] oh yeah [10:54:46] i forgot about those v.v [10:55:32] there's also a lot of occurrences of the rel attribute that don't affect /rendering/ in the slightest [11:03:21] phuedx: yeah i figured that from the apps team transforms [11:03:28] they only leave rel=nofollow [11:18:50] i'm going to walk the dog for a bit [11:18:51] brb [11:24:16] phuedx: "It's time to update the docs for this extension too ;)" which docs? [11:24:25] lol [11:24:30] exactly ;) [11:24:34] but that should be a different thing [11:25:00] phuedx: you mean docs on the wiki page? [11:25:39] yeah, and the README [11:25:43] all of 'em [11:27:05] ok [11:45:17] phuedx|afk: yt? [11:45:31] yeah-ish [11:45:38] just finishing off cooking lunch for lisa [11:45:43] (she's gone to pick up the kids) [11:47:09] ok [11:47:18] phuedx: i'm fixing your pr, couple of nitpicks [11:47:21] don't pick it up [11:47:23] i'll merge it [11:48:38] lol [12:00:07] phuedx: merged both, gonna update loot-ui [12:01:06] joakino: cool <3 [13:32:15] yurik: around? [13:32:24] dbrant, always ) [13:32:40] yurik: so, i'm messing around with Maps on android... [13:32:52] that's good ) [13:33:03] yurik: do we have vector maps? [13:33:29] dbrant, everything you see is actually vector maps, it just that the server converts it to PNGs for you [13:34:09] alternativelly, you can use webgl and download the vector tiles instead of pngs, but the problem is that some of them are not size-optimized enough, so might be bigger than pngs [13:34:37] yurik: so, is there some kind of url pattern for getting the vector tiles? [13:34:53] also, client side rendering requires that we rebuild our style in the webgl format in addition to mapnik format [13:35:00] dbrant, sure, simply ask for .pbf [13:35:12] but those are protobuf blobs [13:35:27] so you need client side rendering components for them [13:36:15] yurik: so it would be something like https://maps.wikimedia.org/osm-intl/{z}/{x}/{y}.pbf ? [13:37:07] correct, but this is only tip of the iceberg - we also will need to spend some time generating our fonts as blobs for it, and work on the webgl style [13:37:38] there is a tool from mapbox that creates font blobs, need to figure out how to use it, [13:37:46] yurik: i see; so, not really worth pursuing on android yet? [13:38:37] dbrant, well, at some point - definitly, but i would rather concenttrate on getting the actual vector tiles ready first - e.g. have the right data in them, and have them small enough [13:38:56] very good! thanks [13:39:14] the only real benefit from client-side is ability to rotate [13:39:30] and eventually - ability to easily switch languages on the client [13:39:47] ...and seamless zoom, i would think [13:40:09] yes, that too [13:40:45] i mean yes, this is the road we want to pursue, i just don't have enough bandwidth for it ) [13:41:47] that's fine; baby steps. [16:31:30] hi bmansurov. Question: can SWAT deploys happen on Fridays? [16:31:44] leila: hi, no [16:31:58] monday - thursday if i'm not mistaken [16:32:32] got it, bmansurov. There is a good chance we get enough responses for the survey today (the responses are very clean so we really don't need the 5K we estimated we should collect to get 1K clean response). No harm to let it run until Monday morning then. :-) [16:32:38] leila, bmansurov depends on: If the change fixes a highly visible problem, it's possible :) (but you've to check with greg-g) [16:33:02] ok [16:33:03] FlorianSW: got it. Good to know. it's really not a problem. It's a survey. :-) [16:33:23] leila: then I think bmansurov is right, that can wait some days :P [16:33:38] yup. :P [17:19:54] leila, bmansurov, FlorianSW: the last.fm ops team phrased it thus: you can deploy at any time you like but if you deploy outside core hours (10 AM to 4 PM) then you won't live to tell the tale [17:20:27] :) [17:20:29] :P [17:22:30] phuedx: this still excludes Fridays, right? [17:22:46] phuedx: your statement applies to Monday-Thursday? [17:22:53] iirc, it excluded most days ;) [17:22:58] :D [17:23:00] the ops team weren't happy folk [17:23:36] but it seems, they had "persuasive arguments" :P [19:02:17] hey [19:58:32] joakino phuedx|afk jhobs florian hey websperts(?), I'm seeing a gnarly bug on iphone where the zoom is getting fucked up. Just sent an email about it to internal list (should have been public in retrospect) [19:58:54] FlorianSW ^ [20:01:29] joakino FlorianSW were you aware of this? ^ example: https://drive.google.com/a/wikimedia.org/file/d/0B-YjaiFnAja3QVpQYVQzRXZObXVRT2pZd0NRNlpBWUJlQWZr/view?usp=sharing, https://drive.google.com/a/wikimedia.org/file/d/0B-YjaiFnAja3ZmhrY003LTI3TmNyNWJnQTBLTkVPQzROeW9z/view?usp=sharing [20:14:50] JonKatz: No, never seen. It seems that's a problem of the huge timeline element (see section "Timeline of the MacBook family"), which seems to doesn't scale correctly. Would you like to open a task for it? [20:22:15] JonKatz: No, never seen. It seems that's a problem of the huge timeline element (see section "Timeline of the MacBook family"), which seems to doesn't scale correctly. Would you like to open a task for it? [20:33:59] FlorianSW thanks, I will. Stuck in a meeting now.. [20:34:59] JonKatz: thanks, np :) I'll work on a solution, so feel free to assign the task to me :) [20:38:30] great, thanks! [20:42:52] JonKatz: replied on email, although I couldn't reproduce it myself with no iOS device [20:43:06] JonKatz: it definitely does look like an issue with the timeline element though [20:54:59] \n not found: size=8605 error. more details in https://ticket.wikimedia.org/otrs/index.pl?Action=AgentTicketZoom;TicketID=8635433 [20:55:07] can someone take a look please? [21:13:09] JonKatz: I was free and added the task myself (to get the task number): https://phabricator.wikimedia.org/T118601 :] [21:14:03] FlorianSW you are a hero. thanks...still in meeting [21:14:34] :( [21:18:53] np :P :) [21:29:55] dbrant: so i've been using the debug proguard change for a while. for clean builds, i don't understand why but it's quite a bit faster some how. for incremental builds, it seems much much slower. i know we want to test what we fly, but are there any hard requirements to keeping proguard on for debug? [21:31:12] niedzielski: not really. I put it in there because otherwise the app couldn't be debugged at all on the Galaxy Light. But now that the support libraries are updated, we can revert. [21:32:09] dbrant: are you seeing noticeably slower incremental build times? even small changes take quite a while on my machine now [21:32:50] niedzielski: i haven't noticed much of a difference... [21:33:36] dbrant: hm, it's pretty dramatic on my machine but i'd feel more comfortable if someone else was having the same issue [21:35:06] bearND mdholloway: are you guys seeing slow incremental build times after the proguard debug changes? [21:36:28] niedzielski: when did we turn on proguard for debug? [21:36:53] builds have been pretty slow lately [21:37:43] bearND: https://gerrit.wikimedia.org/r/#/c/251981/1/app/build.gradle [21:38:31] ah, thanks [21:38:48] i think that now with the updated support library we could turn it off completely [21:39:26] dbrant: any objection to my removing proguard for debug builds? [21:39:33] niedzielski: nope [21:42:26] Volker_E: I've merged your task into "my", before I saw, that there are comments, sorry :/ [21:42:31] https://phabricator.wikimedia.org/T118600 [21:42:36] FlorianSW: Did you read my inputs there? [21:42:40] yeah, it's fine [21:43:00] Volker_E: yes, I've, now :) [21:43:21] seems iOS 9 specific [21:43:32] FlorianSW: now looking into your patch set [21:44:06] Volker_E: I can reproduce in chrome (Android, Windows 10), too [21:44:37] jhobs: regarding your comment: https://phabricator.wikimedia.org/T118525#1804712 You need to write the "Bug: TXXXX" line after an empty line to make gerrit-bot fetch it and link it in the phabricator task :) [21:44:54] FlorianSW: yeah, just a typo, thanks for correcting [21:45:01] np :) [21:47:31] jhobs: I don't know the QuickSurveys extension very well, but: Where is the difference between "disabling" and "deleting"? I ask, because you would leave a disabled code in wmf-config :) [21:47:50] niedzielski: bearND dbrant my builds have felt slowish for a while, but i haven't noticed one way or the other whether proguard has anything to do with it. i'll pay attention after this patch is merged. [21:48:55] JonKatz: timezones :P I assume you mean this morning pacific standard time? (https://phabricator.wikimedia.org/T118600#1804718) [21:49:09] FlorianSW: on the user end, there's effectively no difference, but I'm leaving the disabled config in there for now in case it needs to be reopened for a second round with a different survey link (one of the reasons we're disabling it is because the survey can only hold a limited number of responses) [21:49:44] since bmansurov is doing the deployment, I'll leave it up to him if he wants to decide to remove it entirely or not [21:50:30] ah, ok :) [22:09:09] FlorianSW: I don't get, how it can just affect iOS 9 devices [22:17:43] FlorianSW: jhobs: who's in charge for +2 the PS? [22:19:05] Volker_E: I don't have +2, so can't be me :/ one of the other mobile web team guys should see it though, they're just all probably offline atm [22:21:34] kaity: 2.134-beta-2015-11-12 does not display three dots menu in view card ? Is it by design? [22:22:45] jhobs: FlorianSW: I'd have +2, but a) I don't have a real device for testing in front of me, b) we should be very clear what we're doing (I'm normally not involved in mobile-web outside of my tasks), don't even have mobile-web codebase locally [22:24:25] anyone, where is more detail about Cards extension? How is it different from Popups? [22:26:50] spagewmf: Cards is an extension to generate uniform cards given a list of pages [22:27:00] spagewmf: it's only used by Read More right now [22:27:07] (and it might not actually be being used by it yet) [22:27:21] Popups is hovercards I think, which is a link preview [22:28:48] jhobs: thanks, yes Extension:Popups=Hovercards. I assume there's some common code with un-hoverCards :), it would be nice if it was shared. Bring back Mantle :) [22:29:57] spagewmf: the ideal is to get Cards to be used by things like RelatedArticles (already planned/maybe already happening), Gather, and yes theoretically even Hovercards [22:30:19] spagewmf: Cards is us trying to pull out some common code into its own extension :) [22:37:34] Volker_E, jhobs I'd have +2, but it's my own change, so ;) But I think, that someone will look in this change on Monday. Like I said already, I can reproduce with Chrome devtools on my desktop pc [22:38:25] FlorianSW: how and where are you able to reproduce it on desktop -> pls add to task as comment [22:39:20] JoshM: 2.134-beta-2015-11-12 does not display three dots menu in view card? Is it by design? [22:42:07] Volker_E: ok :) [22:47:04] Volker_E: updated [22:50:30] hey jhobs [22:50:37] kaity_: yo [22:51:19] jhobs: do you know how we restyle templates into "page issues"? [22:51:37] jhobs: what criteria do we use to distinguish that a template is a page issue? [22:52:06] Volker_E: about your comment on https://gerrit.wikimedia.org/r/#/c/253035/1 Where do you find a "*-container" class in the Timeline extension? :/ [22:52:13] kaity_: I just looked into this recently but can't recall, one sec lemme find it again [22:52:38] jhobs kaity_ we choose any template that have an ambox or mbox class iirc [22:52:49] FlorianSW: hmm, I can't reproduce it on Chrome, must have a setting wrong :\ [22:53:14] FlorianSW: to my comment: I'm talking about all DOM, not just timeline extension [22:53:40] FlorianSW: could you search mobile codebase for '-container' vs. '-wrapper' [22:53:43] FlorianSW: ? [22:54:27] thanks jhobs and FlorianSW ! [22:55:06] kaity_ jhobs sorry, ".mbox-text, .ambox-text" are the classes, which needs to be wrapped into table.ambox, table.tmbox, table.cmbox :) [22:55:23] Volker_E: let me look :) [22:55:53] Volker_E: what do you do :D Maybe the screen resolution is to big? Do you see the complete timeline or only a part of it? [22:55:55] FlorianSW: jhobs thinking about a more descriptive name than "page issues" [22:56:11] kaity_: what FlorianSW said [22:56:49] FlorianSW: "page issues" probably covers the most, but sounds vague [22:57:01] FlorianSW: does "Content Issues" make sense? [22:57:17] Volker_E: in MF we mostly use *-container, but sometimes *-wrapper :P But I've no strong preference, if you feel better with *-conatiner, I'll change it :) [22:58:06] kaity_: let me look, what templates use these classes, iirc they are used very widely, not only for "page/content issues" at all. [22:59:46] FlorianSW: I'd go for one way, if majority is now *-container, let's stay with it, alright? [23:00:27] thanks FlorianSW. Yes there's probably tons of other uses. Wondering what the most popular usage is. If 90% are for "Content Issues" that label might work [23:00:35] kaity_: ambox is used on article pages, tmbox on talk pages, cmbox on category pages (*yay*, now I know, where these class names come from :D). I'm not sure, if "content issues" is the right thing for category pages. Neither I think that page issues is the right name ;) [23:01:35] Volker_E: remember, that the change is in another extension :) And I don't know the preference in all mediawiki {core|extensions} code :) [23:01:43] FlorianSW: right, seems like most correct name would be "Article message" "Talk message" and so on [23:02:45] kaity_: I'm not sure, but it feels, like we had this discussion already somewhere :P Let me look for a task, ok? :) I think, that page issues is not the best one, but at lest the best we had for now. We could use different names for different types, sure [23:03:20] FlorianSW: thanks for the help! [23:08:13] kaity_: we already show different messages for different pages (e.g. by https://gerrit.wikimedia.org/r/#/c/182124/): [23:08:44] good to know [23:08:52] kaity_: "Page issues" for article pages [23:08:59] FlorianSW: well, my take: if core has got a preference, extensions should follow [23:09:12] kaity_: "About this talk page" for talk pages [23:09:25] kaity_: "About this category" for category page [23:09:27] *page [23:09:29] *pages [23:09:31] argh! [23:09:44] FlorianSW: haha [23:10:01] FlorianSW: what about for article pages, do you think most deal with content issues? [23:10:09] Volker_E: yeah, you're right (you, by chance, don't know the preference in core? :P) [23:12:25] kaity_: Template:Underlinked isn't really a content issue, but uses ambox :/ So, I'm not sure, I think "content issue" is more (and far to) specific and doesn't cover all page related issues (like underlinked), while "page issues" at least covers most of these things :) [23:13:15] FlorianSW: will have a look now [23:13:40] Volker_E: i can't see a strong preference for one of these, for my short look, -wrapper is used as many times as -container :D [23:15:45] FlorianSW: right. and the template for current events for example, isn't really an "issue" more of a "notification" hmm [23:15:57] which one? [23:16:00] kaity_: ^ [23:17:39] FlorianSW: https://en.wikipedia.org/wiki/Template:Current [23:19:52] kaity_: hmpfh. It's difficult to find the correct name :/ [23:21:18] jhobs: do you think putting the timeline styles into the top queue makes sense? [23:23:06] FlorianSW: ok, seems arbitrary [23:23:11] FlorianSW: no clear preference [23:23:21] FlorianSW: 30 results here, 40 there [23:23:25] let's stay with wrapper [23:23:35] FlorianSW: I think bottom is the ideal, I just wanted to bring up the issue. I think we should push it as bottom but create a task for a spike to see if it's a problem [23:23:50] FlorianSW: we did the same thing with the menu code [23:23:52] FlorianSW: who is mainly behind Timeline extension? [23:24:19] jhobs: ok! :) [23:24:28] Volker_E: I have no idea :D Let's find it our [23:24:30] *out [23:26:48] hmm, it seems, that no one really contribute to it. The extension was created by https://meta.wikimedia.org/wiki/User:Erik_Zachte (who actually seems to work for the WMF) [23:26:53] Volker_E: ^ [23:27:15] Volker_E: and some generic contributions of Chad, MatmaRex, ori and some other people [23:27:43] FlorianSW: yeah, just looked at it myself :) [23:27:54] FlorianSW: not too much activity [23:32:39] FlorianSW: I've updated the commit message, you could +2 it now