[00:26:09] <maxbinder>	 dr0ptp4kt: are you still in office?
[08:21:53] <phuedx>	 morning o/
[08:34:21] <joakino>	 morning
[08:34:59] <phuedx>	 hey joakino o
[08:35:00] <phuedx>	 o/
[08:36:19] <joakino>	 hii
[12:09:01] <jhobs>	 mornin' folks o/
[12:43:51] <phuedx>	 hey jhobs
[12:43:58] <jhobs>	 howdy
[12:44:01] <phuedx>	 jhobs: what time is it for you?
[12:44:09] <jhobs>	 8:45am
[12:44:18] <jhobs>	 started early today
[12:59:20] <phuedx>	 jhobs: what's on your radar for today?
[12:59:36] <phuedx>	 i ask because i'm mostly happy with the technical side of https://gerrit.wikimedia.org/r/#/c/309409/
[12:59:52] <phuedx>	 but i'd like to deploy it to our staging server and test it before merging it
[13:00:04] * phuedx is just updating staging now
[13:01:06] <jhobs>	 phuedx: I need review on https://gerrit.wikimedia.org/r/#/c/310483/ as I'm uncertain I did the EL sampling right, then picking up one of the Todos
[13:01:16] <jhobs>	 309409 LGTM
[13:01:37] <jhobs>	 I can help you test once it's on staging though if you ping me
[13:17:07] <phuedx>	 jhobs: it's on staging
[13:17:10] <phuedx>	 i've found a bug
[13:17:15] <phuedx>	 am fixing it now
[13:17:58] <jhobs>	 phuedx: ok, I'll re-review after the new PS
[13:19:50] <dr0ptp4kt>	 sorry maxbinder just saw your msg
[15:03:25] <niedzielski-afk>	 nzr: o/ we are here. livin la vida loca, etc
[15:03:39] <joakino>	 hashar: mwext-mw-selenium job runs on chrome, right? (https://integration.wikimedia.org/ci/job/mwext-mw-selenium/10072/consoleFull)
[15:03:48] <joakino>	 trying to troubleshoot a failure i can't reproduce locally
[15:04:54] <hashar>	 joakino: I am not sure
[15:05:14] <hashar>	 I think it depends on environments.yaml
[15:05:22] <joakino>	 i'm seeing the chromium path in the full log
[15:05:31] <joakino>	 aha
[15:06:25] <hashar>	 yeah cause we have to vary the bin path to chromium depending on the os
[15:06:31] <hashar>	 i would head to https://integration.wikimedia.org/ci/job/mwext-mw-selenium/10072/
[15:06:37] <hashar>	 there is a video of the browser running
[15:06:41] <hashar>	 the .mp4 file
[15:06:49] <hashar>	 download it / open with vlc or whatever is your favorite video player
[15:06:54] <hashar>	 and it shows the browser window
[15:07:10] <hashar>	 that is captured whenever a test fail
[15:07:19] <hashar>	 it looks like chromium to me
[15:07:26] <joakino>	 woot
[15:07:42] <joakino>	 http://media0.giphy.com/media/msriR5ybSpQgo/giphy.gif
[15:07:49] <hashar>	 there is a screenshot as well https://integration.wikimedia.org/ci/job/mwext-mw-selenium/10072/artifact/log/Popups%20settings%3A%20Popups%20can%20be%20enabled%20via%20the%20%22Enable%20previews%22%20footer%20link.png
[15:07:57] <hashar>	 ahah
[15:09:38] <joakino>	 hashar: definitely chrome, thanks for the info! very useful
[15:09:49] <joakino>	 now to see why that doesn't happen to me locally
[15:10:11] <hashar>	 browser tests are hard to repro :(
[15:11:21] <joakino>	 they get really annoying sometimes
[15:11:43] <joakino>	 hashar: is there any way to see js errors? (i'm guessing not, but i don't loose anything by asking)
[15:12:24] <hashar>	 such as capturing the browser console ?
[15:13:07] <hashar>	 I dont think we have that
[15:13:17] <hashar>	 you would want to ask zeljkof in #wikimedia-releng
[15:18:54] <joakino>	 thanks for the intro
[15:45:57] <mdholloway|afk>	 bmansurov: hey, would you mind forwarding or adding me to that email thread when you get a chance?
[15:46:08] <bmansurov>	 mdholloway|afk: ok
[15:46:29] <mdholloway|afk>	 bmansurov: sweet, thx
[15:46:29] <bmansurov>	 mdholloway|afk: done
[16:01:02] <joakino>	 gerrit's down...
[16:02:10] <niedzielski>	 joakino: seems to be working for me?
[16:02:57] <joakino>	 niedzielski: it was just for a minute! (i'mnotcrazypleasebelieveme)
[16:03:17] <niedzielski>	 joakino: i can't. i can't imagine a world in which gerrit goes down even for a second
[16:03:52] <joakino>	 ;)
[16:04:08] <joakino>	 niedzielski: have you tried putting emoji in gerrit replies?
[16:04:19] <niedzielski>	 joakino: i did on the old version of gerrit
[16:04:21] <niedzielski>	 caused a 500 or something
[16:04:24] <niedzielski>	 i don't remember
[16:04:24] <joakino>	 i think it's one of the few places on the internet that won't accept unicode
[16:04:27] <niedzielski>	 yeah!
[16:04:30] <joakino>	 yep, same with this new one
[16:05:19] <niedzielski>	 that's too bad. really changes culture and the way we think when i can't send a picture of a finger going into an ok hand in a code review like i need to
[16:05:53] <niedzielski>	 fortunately, it doesn't seem to care about unicode patches. that'd be a bummer
[16:31:47] <hashar>	 joakino: did you get a task to retrieve the browser console log in the browse tests ? just wondering
[16:32:41] <hashar>	 joakino: we have an old one to retrieve the selenium internal logs.  And it seems I found a solution to capture both the selenium and console logs 
[16:32:43] <hashar>	 https://phabricator.wikimedia.org/T94577 :D
[16:34:13] <joakino>	 i'll comment there or open a new one later, i'm in grooming meeting now hashar 
[16:36:13] <hashar>	 joakino: ok no worries :}
[16:52:45] <jhobs>	 olliv: D:
[16:52:55] <jhobs>	 internet issues, or hangouts?
[16:52:57] <olliv>	 got kicked out
[16:53:01] <olliv>	 hangouts
[16:56:55] <niedzielski>	 dbrant: o/
[16:57:10] <dbrant>	 \o
[16:57:26] <niedzielski>	 dbrant: would it be possible to get a merger deal on this? https://gerrit.wikimedia.org/r/#/c/310356/
[16:57:57] <niedzielski>	 dbrant: i'll be needing it soon for some of the ci work
[16:58:34] <dbrant>	 niedzielski: sure, will look
[16:59:30] <niedzielski>	 dbrant: thanks. the issue that this solves is it really allows me to identify tests that require a webview and those that don't. there's an issue where the current emulator support by the android jenkins plugin requires a physical gpu which a vmware instance can't provide
[16:59:50] <niedzielski>	 dbrant: this might change in the future, but for now i need to be able to run just non-webview tests
[17:00:00] <dbrant>	 i see!
[17:02:29] <joakino>	 maxbinder: you coming?
[17:04:29] <Pchelolo>	 bearND: question: the random portion of the feed endpoint isn't used anywhere and we can safely change the format, right? https://github.com/wikimedia/restbase/pull/669#discussion_r78761709
[17:04:53] <bearND>	 Pchelolo: right
[17:05:39] <Pchelolo>	 bearND: so what's your opinion - have 'items' or make 'random' just an array (if in future you'd want to have many random articles at once) ?
[17:06:25] <Pchelolo>	 also FYI, we're going to deploy RESTBase soon with some changes to the feeds endpoint (populating summaries in RESTBase)
[17:06:48] <bearND>	 Pchelolo: I'd be fine with just an array if that looks reasonable to you. Thanks for letting me know abou the deployment.
[17:07:15] <bearND>	 Pchelolo: Do you want to test this first on beta cluster?
[17:07:48] <Pchelolo>	 bearND: given that we've had an outage recently that might be a good idea :)
[17:08:28] <Pchelolo>	 I'll ping you when something would be actually happening on our side
[17:08:56] <bearND>	 Pchelolo: mobrovac : I wrote up some documentation on how to change the app to a different server: https://www.mediawiki.org/wiki/Wikimedia_Apps/Team/RESTBase_services_for_apps#Testing_with_the_Android_app     Will need to add a line to the URI hints for beta cluster, though. Until then you can manually type it in.
[17:09:15] <maxbinder>	 joakino: I'm out today and tomorrow for Onsite
[17:09:51] <Pchelolo>	 bearND: great, but for me that's not very useful - I don't have any android device..
[17:10:33] <mobrovac>	 Pchelolo: ditch the iphone, do it! :)
[17:12:02] <bearND>	 Pchelolo: ^ plus you could download an Android emulator, like Genymotion
[17:13:26] <Pchelolo>	 bearND: Hm, let me actually do that, because although I'm working on backend that Android app uses, I've never actually seen it :)
[17:14:15] <Pchelolo>	 bearND: oh, it's paid, nonono
[17:16:22] <joakino>	 sorry i didn't know maxbinder 
[17:16:40] <maxbinder>	 joakino: no worries, I could have been clearer before today :)
[17:36:40] <bearND>	 Pchelolo: Sorry, hadn't looked at that in a while. There are many free alternatives. The best would be to also setup a dev environment using https://www.mediawiki.org/wiki/Wikimedia_Apps/Team/Wikipedia_Android_app_hacking#Install_JDK_8 and following. Once you have Android Studio installed you can create an emulator, and also launch builds from there
[17:43:15] <joakino>	 alright i'm out
[18:45:33] <Pchelolo>	 bearND: MCS fails to return feed in beta: https://en.wikipedia.beta.wmflabs.org/api/rest_v1/feed/featured/2016/09/11
[18:45:45] <Pchelolo>	 so we can't test RB deployment in beta..
[18:47:13] <bearND>	 Pchelolo: ohoh. That's not good. Any ideas why that is?
[18:47:43] <Pchelolo>	 bearND: there's no 'in the news' section on beta?
[18:47:58] <bearND>	 Pchelolo: oh, I see
[18:48:44] <Pchelolo>	 bearND: ye, here's it: http://appservice.wmflabs.org/en.wikipedia.beta.wmflabs.org/v1/page/news
[18:49:16] <Pchelolo>	 bearND: So I think I will proceed with RB deploy, but let's test the app right after I finish?
[18:50:15] <bearND>	 Pchelolo: ok. Ping me when it is deployed, so I can test it.
[18:50:20] <Pchelolo>	 kk
[18:50:55] <Pchelolo>	 but it would be nice to fix labs too eventually
[18:52:18] <bearND>	 Pchelolo: yes, definitely.
[18:53:51] <niedzielski>	 dbrant bearND mdholloway|afk: i'm going to bump the minor versionCode
[18:54:15] <dbrant>	 bump it
[18:59:28] <niedzielski>	 dbrant bearND mdholloway|afk: https://gerrit.wikimedia.org/r/#/c/310611/1
[19:02:07] <dbrant>	 looks like alpha builds isn't updating :(
[19:03:32] <niedzielski>	 dbrant: bummer. if it's not critical, make a card i can check it out after ci shtuff (i'll likely be in the same parts)
[19:04:36] <Pchelolo>	 bearND: RB deployed to eqiad, could you please verify the app is fine?
[19:04:44] <bearND>	 👍 
[19:04:54] <bearND>	 Pchelolo: will do
[19:06:27] <jhobs>	 gonna take a much needed break after working late & early. Feel free to ping if there's anything urgent o/
[19:07:03] <niedzielski>	 dbrant bearND mdholloway|afk: just a reminder that we were considering api >= 16 after the final overhaul build
[19:07:36] <bearND>	 Pchelolo: oh no. I think you should revert the deploy. The array is causing issues. The app expects an object but it's an array now :(
[19:08:12] <bearND>	 Pchelolo: com.google.gson.JsonSyntaxException: java.lang.IllegalStateException: Expected BEGIN_OBJECT but was BEGIN_ARRAY at line 1 column 30221 path $.random
[19:08:31] <Pchelolo>	 bearND: ok revert
[19:09:30] <bearND>	 thanks
[19:10:17] <Pchelolo>	 bearND: so the app was using the random portion after all?
[19:12:40] <bearND>	 Pchelolo: regrettably yes. Only for reading in the response for the service but then it was not used. I'm going to remove that from the app to allow for more flexibility in the future until the time we actually use it. //cc: niedzielski dbrant 
[19:13:09] <Pchelolo>	 bearND: anyway the app is out, so we have to go back to 'items' property..
[19:13:36] <niedzielski>	 bearND: should i wait on the build?
[19:13:44] <bearND>	 niedzielski: yes
[19:16:48] <bearND>	 Pchelolo: the app expects summary data directly under "random". As far as I can see, the properties in that object are all optional and effectively not used. I think we could introduce a new items property that could be an array without problems. (remains to be tested)
[19:18:28] <bearND>	 Pchelolo: correction: title and normalizedtitle are marked as NonNull. But I doubt that matters since the random object is not used.
[19:26:34] <bearND>	 Pchelolo: Yeah. That's the thing. I'm not even sure if we ever need an array for the "random" property. We could also just leave it as a simple summary object. Not sure if the iOS guys and our PM are around to comment on whether they just need one random entry for sure, and not multiple. (coreyfloyd JoshM_ dbrant)
[19:28:56] <dbrant>	 bearND: just one is fine for our needs
[20:39:55] <niedzielski>	 dbrant bearND mdholloway|afk: apk is getting hefty at 17M :)
[21:00:10] <bearND>	 is that phat?
[21:04:17] <Pchelolo>	 bearND: ok, we've fixed that problem in RB feeds and I'm going to try to deploy again. I'll ping you when it's time to test
[21:04:34] <Pchelolo>	 are you gonna be here?
[21:05:13] <Pchelolo>	 during the next 15 minutes or so
[21:10:42] <bearND>	 Pchelolo: what does the random property look like now? 
[21:10:51] <Pchelolo>	 we've returned the items
[21:11:05] <bearND>	 random.items[] ?
[21:11:07] <Pchelolo>	 we thought it's safe to return it back as it was before
[21:11:10] <Pchelolo>	 ye
[21:11:58] <bearND>	 and in there we'll get the summary data now.
[21:12:06] <Pchelolo>	 yep
[21:12:10] <bearND>	 Pchelolo: yes, I'll be around for the next 15 minutes
[21:12:45] <Pchelolo>	 kk, I'm going to do the canary deploy on 1 node, wait ~10 minutes/look at the logs etc, then deploy on all servers and ping you
[21:24:10] <Pchelolo>	 bearND: time to test the app
[21:24:18] <bearND>	 k
[21:25:42] <bearND>	 Pchelolo: lgtm
[21:25:58] <Pchelolo>	 bearND: awesome :) finally... that was a painful deploy :(
[21:26:57] <Pchelolo>	 bearND: now we can transfer https://phabricator.wikimedia.org/T143933 to you - you can remove summary data from the MCS response
[21:29:28] <niedzielski>	 i kind of wonder if folks will be confused by the difference between up and the hardware back button in the app. other apps use this paradigm but i don't personally find it as intuitive in the current implementation
[21:34:27] <bearND>	 Pchelolo: ok, sounds good. Thank you
[21:37:22] <bearND>	 Pchelolo: one thing I noticed is that now the entries have a few more properties: dir, lang. That's not an issue but for consistency sake we should probably change the (featured) image portion of the feed response to also have these. For this, can we have summary data + and extra property for the (full-res) image?
[21:39:34] <mobrovac>	 hm, but dir doesn't have a defined meaning for pix, does it?
[21:40:35] <bearND>	 mobrovac: there's still a description associated with the featured image
[21:40:36] <Pchelolo>	 mobrovac: bearND: we could, but right now we don't apply summary to the image, because the extract returned by summary endpoint for File pages is supert-maningless
[21:40:59] <mobrovac>	 right
[21:41:46] <Pchelolo>	 bearND: but it seems like MCS takes description from some other source, not like summary
[21:42:04] <mobrovac>	 we can always add them later
[21:42:21] <mobrovac>	 adding fields is easy, but getting rid of them is almost impossible
[21:45:35] <bearND>	 Pchelolo: yes, there can be multiple descriptions associated with a featured image for the various languages, which I just realized that the lang field inside description makes the structure different than the other ones. So, let's keep the image portion as is for now
[21:45:49] <niedzielski>	 dbrant bearND mdholloway|afk: there are a few bugs that came up so far that i think we should hold off on the beta. i'm adding tickets as i go so just check out the latest
[21:45:52] <Pchelolo>	 kk