[09:00:35] hi! [09:27:51] hullo [10:08:08] phuedx: y u no gitter [11:37:07] phuedx, hey [11:37:31] phuedx, do you get a value when you run mw.cookie.get('GeoIP') in the browser's console? [11:37:55] phuedx, $.cookie('GeoIP') returns a value but not the above function call [11:38:18] bmansurov: arg [11:38:19] sorry [11:38:24] bad arguments i gave you [11:38:38] speak like yoda i will? [11:38:57] sure [11:39:07] on go ;) [11:39:45] bmansurov: commented on the patch for posterity [11:39:50] thanks [11:39:56] but mw.cookie.get uses the wiki's cookie prefix by default [11:40:00] so you need to cancel it out [11:40:03] (second parameter) [11:40:18] i see [12:03:40] breaking for a little lunch [14:48:35] jdlrobson: gtg take the boys out [14:48:44] i've left the ticket open and unassigned [15:50:21] i'm trying to use mobilefrontend with varnish, but all the requests to load.php have a &version=2015.... parameter which is messing with the caching, since it changes on every request, even though nothing is really changing [15:50:22] i see a version key on en.m.wikipedia.org requests, but that one is some kind of hash and doesn't change. How do i get that on my wiki? [15:51:01] Hi. Could anyone from mobile team to answer question? [15:51:57] What hosts mobile application get information from? do those hosts have the same IP addresses as web? [16:09:29] niedzielski: hmm, something has changed in that patch... looking now. [16:09:39] it was behaving correctly before. [16:10:08] dbrant: for material ripples? maybe it's just my setup [16:10:24] dbrant: i'm using the API 23 AOSP emulator in a nexus 5 configuration [16:11:15] dbrant: they're definitely more materialish but i don't see ripples on things until the touch up event. ideally, we'd see them bounce out on touch down [16:13:23] working on it... [16:42:39] niedzielski: dbrant|food: bearND: since the week is dwindling, anyone object to my focusing on getting my two patches in before i head out, and leaving code review to the rest of you guys until that's accomplished? [16:43:03] mdholloway: no objections and thanks for the reminder mail earlier! [16:43:13] niedzielski: thanks and no prob! [16:43:26] mdholloway: i'm currently digging through bearND's perf patch [16:44:04] mdholloway: sounds good [16:53:35] doh we missed vlsergey [17:02:12] joakino: bmansurov [17:02:20] ^ stand or be damned [17:02:23] almost there [17:02:35] google is loading ;) [17:06:12] phuedx: your recommendation seems like it will work assuming the schemas which are registered via EventLogging hook runs before the RL registration hook. [17:06:22] idk we shall see [17:08:33] rmoen: have a stab and see what happens -- it /shouldn't/ matter (because we just need both hooks to run before we output the page) [17:08:37] but yeah, you're right [17:08:46] because rl might shout about dependencies [17:09:23] phuedx: I [17:09:30] lol, i'll try it right after the standup [17:15:05] @bearND: hey! were you interested in feedback on your retrofit wip right now or is this review just for curious devs? [17:15:49] niedzielski: yes, I'm interested in feedback for it sooner rather than later [17:15:58] bearND: ok, will review. thanks! [17:27:13] phuedx: there? [17:54:13] phuedx: good news your suggestion worked ;) [17:54:16] bearND|afk mdholloway|afk dbrant|food: this is interesting. i've been using the new api 23 emulator the past couple days and just noticed that there is a "data usage warning" notification. when i tap on it, it says the os has used a couple GB bandwidth but doesn't tell me if any app contributed to that [17:55:29] dbrant|food: mdholloway|afk bearND|afk: and from the graph it makes it look like half of that was in the past hour or so [18:21:53] codezee: around? [18:22:13] codezee: if we want that config change to go out today i need to deploy a config change later [18:22:30] jdlrobson: yes, I've saw the discussion on traveller's pub [18:22:44] jdlrobson: I've also replied to the doubts that popped up [18:22:58] *I've seen [18:25:14] jdlrobson, codezee I can swat today [18:25:23] Brb [18:33:24] codezee: are you fixing the toc issue on the site with help from Nicholas? [18:33:27] or do you need help [18:36:01] jdlrobson: I've already identified the issue and added a patch [18:36:07] yup and i merged it :) [18:36:14] but i'm asking about the common.css fix [18:36:33] sumit can we safely add the wikidata support today [18:36:34] ? [18:36:48] jdlrobson: yes, I'll wrtie a mail to him since I do not have access to it [18:37:07] codezee: i can do it [18:37:08] jdlrobson: yes, wikidata support should not create issues [18:39:15] https://en.wikivoyage.org/wiki/London?debug=true < codezee live [18:39:15] that would be great! [18:39:48] \o/ works temporarily :) [18:40:54] jdlrobson: I'll update it on traveller's pub that toc dropdowns now works [18:41:15] codezee: i'll put https://gerrit.wikimedia.org/r/232782 in the swat window [18:41:21] do you know any pages i can test on in production? [18:43:11] jdlrobson: most pages of wikivoyae that I've seen are having banner property, our initial test page https://en.wikivoyage.org/wiki/Red_Wharf_Bay also has one [18:59:03] kaity K-HammZ, do you have anything for design review this afternoon? I saw that joakino declined, but we may have JonKatz available. [18:59:27] kristenlans: nothing from me [18:59:59] kristenlans I have something until 1:30. [19:00:52] jdlrobson: I'll be looking out for the deployment for as long as I can today, if any problem arises(though I hope not), I'll look into it at the earliest. [19:19:05] kristenlans: should we cancel? [19:20:27] kaity I think so. [19:26:19] kristenlans: i like K-Hammz :) [19:26:53] kristenlans: K-Lanz! [20:08:32] jdlrobson: Can you explain in a nutshell why Gather aliases .mustache templates to a .hogan name? [21:23:07] niedzielski: so, i've been trying the new API 23 emulator, and I'm not sure I "trust" it in terms of a faithful visual representation. For example, if you just go to the "Settings" app in the emulator and press-and-hold the items, there's no ripple effect, just as you observed in our app... [21:23:46] dbrant: hm, let me take a look on my end too [21:23:49] niedzielski: either they changed this behavior at the system level, or the emulator is flawed. [21:27:15] dbrant: i see what you mean :( [21:27:57] dbrant: my api 21 physical device performs as expected in settings [21:32:51] niedzielski: dbrant: I don't think this should be a blocker for merging this and the release. Do you agree? [21:32:52] bearND (/cc dbrant mdholloway|bbl ) do you know how to mark a non-string response field optional in swagger? it looks spec.js expects all the fields to be present [21:33:20] bearND: i agree [21:33:42] bearND: yeah, no qualms. i'd like to test it on my api 21 device though [21:35:12] niedzielski: i would have to either research this or ask mobrovac or gwicke about the Swagger question. So far, I've kept theses things very simple. [21:35:51] niedzielski: which field would be optional? [21:36:31] you mean the whole infobox structure? [21:36:34] bearND: i was thinking infobox would not always be present [21:36:38] that's right [21:36:56] yeah, good point [21:41:45] bearND: what is the question? [21:42:12] how to mark a response property as optional [21:42:16] gwicke: ^ [21:42:24] in swagger spec [21:42:37] * gwicke pulls up https://github.com/swagger-api/swagger-spec/blob/master/versions/2.0.md [21:43:24] generally, JSON responses are described by JSON schema [21:43:30] so we'll probably have to look there [21:44:41] "required" seems to be the keyword for JSON schema v3 [21:44:45] gwicke: hey! so the place this seems to be an issue is when running the tests. the test checks that all fields in the response body are present [21:45:47] okay; I'm not 100% sure if the tests actually use the JSON schema [21:46:01] where is the code again? [21:47:16] nm, found it [21:47:54] which test is this about? [21:47:59] looking at https://github.com/wikimedia/mediawiki-services-mobileapps/tree/master/test/features [21:48:50] gwicke: http://git.wikimedia.org/blob/mediawiki%2Fservices%2Fmobileapps/05fb5e3f31440df7f5f5fc9090be47824c386ed8/spec.yaml [21:49:31] basically, those x-amples sections are really just specifying a test [21:49:47] http://git.wikimedia.org/blob/mediawiki%2Fservices%2Fmobileapps/05fb5e3f31440df7f5f5fc9090be47824c386ed8/test%2Ffeatures%2Fapp%2Fspec.js [21:49:59] so, if a property is listed there & tested for being present, then that means that it is required implicitly [21:50:07] to make it optional, you can remove the test for the property [21:50:09] gwicke: http://git.wikimedia.org/blob/mediawiki%2Fservices%2Fmobileapps/05fb5e3f31440df7f5f5fc9090be47824c386ed8/test%2Ffeatures%2Fapp%2Fspec.js#L191 [21:50:30] oops looks like bearND beat me [21:51:09] which property should be optional? [21:51:29] is http://git.wikimedia.org/blob/mediawiki%2Fservices%2Fmobileapps/05fb5e3f31440df7f5f5fc9090be47824c386ed8/test%2Ffeatures%2Fapp%2Fspec.js#L191 the line that fails? [21:51:41] niedzielski: usually when you run the test it tells you expected and actual. [21:52:31] gwicke: er, sorry it's L139 [21:52:42] niedzielski: if you push your patch to Gerrit then we can have a look and try it out, too [21:53:07] okay, in that case you should perhaps remove the offending property [21:53:09] from the spec [21:53:46] for example, if it's the 'revision' property that should be optional in http://git.wikimedia.org/blob/mediawiki%2Fservices%2Fmobileapps/05fb5e3f31440df7f5f5fc9090be47824c386ed8/spec.yaml#L96, then just remove it from the x-amples section [21:54:08] bearND: gwicke https://gerrit.wikimedia.org/r/232853 [21:54:53] gwicke: is there way to still document, other than in comments, that the endpoint will often emit this field? [21:55:31] gwicke: i don't see how i would remove the test for one property of the response body [21:56:03] niedzielski: hmm, your gerrit link doesn't work or me [21:56:15] niedzielski: absolutely; the x-amples actually don't show up in the docs [21:56:58] bearND: it's a draft so maybe i have to add you? i just did so hopefully works now [21:57:06] for documentation purposes, you can specify a set of responses and corresponding schemas [21:57:16] example: https://github.com/wikimedia/restbase/blob/master/specs/mediawiki/v1/content.yaml#L87-L99 [21:58:18] the schema for a 'problem' response is at https://github.com/wikimedia/restbase/blob/master/specs/mediawiki/v1/content.yaml#L1038-L1049 [21:58:51] those schemas are then showing up in doc sections like https://en.wikipedia.org/api/rest_v1/?doc#!/Page_content/page_graph_png__title___revision___graph_id__get [21:59:25] gwicke: can schemas inherit from each other? [22:02:15] it sounds like: https://github.com/swagger-api/swagger-spec/blob/master/versions/2.0.md#composition-and-inheritance-polymorphism [22:02:20] haven't used it myself [22:02:48] gwicke: ok this will take me a while to digest. thanks for all the pointers! [22:03:44] glad you found the bits you were looking for [22:05:53] gwicke: Thank you for the good info. Does spec.js or any other test (from the template) also test the actual spec (not just the x-amples)? [22:08:42] it only tests the x-amples [22:09:02] there are also some libraries that generate random values based on the spec, but I haven't tried them [22:09:13] a somewhat smart fuzzer, basically [22:12:09] gwicke: ok. Thanks. [22:14:17] http://swagger.io/open-source-integrations/ [22:15:33] niedzielski: I guess for now you can add another example that requests a page that actually has an infobox. (And remove the infobox line on the current one, which goes to the enwiki Main Page [22:17:43] bearND: ok cool, thanks [22:41:27] jdlrobson: updates bug :) [22:52:42] thx FlorianSW [22:52:52] np, thanks for the quick review :)