mozilla :: #ateam

21 Apr 2017
00:32gpsAutomatedTester, ted: if you want to see betamax in action: https://reviewboard.mozilla.org/r/132230/diff/1#index_header
00:33gpsi like how you can substitute out parts that don't matter, such as API tokens
00:34gpsand it captures all HTTP headers, so that mitigates bugs around varying headers
00:38tedpretty cool
00:38tedwonder if you could wire that up as a http proxy to use with non-python software
00:38ted(he thinks, not having tests for the s3 interaction in sccache)\
00:41gpsthere are enough "fake s3" servers out there that this particular problem has been solved
00:41gpsthere's a whole "fake aws" python package providing stub implementations of various services IIRC
00:41tedyeah
00:42gpspython-moto maybe?
00:42tedi have an issue open to use one, i just haven't gotten around to it
00:42tedbut faking is not quite the same as testing against the real API
00:42gpsyeah
00:43gpsted: https://github.com/ucarion/http_replayer
00:43tedah, neato
00:43gpsthat's the only "rust" from the list at https://github.com/vcr/vcr
00:44gpsalso, geez, how fragmented is the ruby http request library ecosystem?
00:45gpsthere are so many choices that i thought i was looking at a Node.js module at first
09:17whimbooato: hey. I would need an immediate r? from you for https://reviewboard.mozilla.org/r/132348/
09:18whimbooif we can get this landed on central and beta we can unblock SV from signing off on update tests for 54.0b1
09:18whimboothanks
09:20whimbooi wish we could test those cases easier :/
09:21atowhimboo: You should be passing -marionette instead.
09:22whimboono
09:22whimbooas you know arguments are lost after a restart
09:22whimbooits still crazy for me to understand why this only happens when changing the update.status file
09:23atoRight, you say further down in the commit message.
09:23whimboomaybe matt has an idea
09:23ator+
09:23whimboocould be that we have special restart behavior when the file exists
09:23whimboooh wait
09:23whimboofirefox checks this file during a restart!
09:24whimbooato: thanks
09:24whimbooTomcat|sheriffduty: hello. you have to do us a big favor
09:24whimboowe need the patch on bug 1358402 for central and beta
09:24bugbotBug https://bugzilla.mozilla.org/show_bug.cgi?id=1358402 Marionette, blocker, hskupin, ASSIGNED , Total bustage of firefox-ui fallback update tests for Firefox 54.0b1 because Marionette is not getting started after a restart of Firefox
09:25Tomcat|sheriffdutyok
09:25whimbooso if you can land you will be our hero :)
09:25whimbooSV is blocked on the 54.0b1 signoff due to that
09:30Tomcat|sheriffdutywhimboo: all done
09:31whimbooTomcat|sheriffduty: +++
09:32Tomcat|sheriffdutywhimboo: np
09:33* whimboo needs lunch now after such a stressful morning
10:39atoAutomatedTester: We need to do a release soon. Ive even had people email by privately asking us to.
10:40atoAutomatedTester: But it would be great if you could do another run of the Selenium test suite again with master (-:
11:27armenzg_afkrwood|afk: what is H2 from that email?
11:29AutomatedTester|cloudato: can you do a geckodriver release now please?
11:38armenzg_afkrwood|afk: got it; Http2
11:48Standard8smacleod: do you know if there was a specific reason that ahal choose .lint extensions rather than .lint.py in bug 1258341? It breaks editor & flake8 integration, so I was thinking about renaming them
11:48bugbotBug https://bugzilla.mozilla.org/show_bug.cgi?id=1258341 Lint, normal, ahalberstadt, RESOLVED FIXED, Refactor |mach eslint| to use mozlint library
13:34atoAutomatedTester: Sorry, was in a meeting.
13:34atoAutomatedTester: I can absolutely do a release. Are we confident about Selenium?
13:42smacleodStandard8: I do not remember a specific reason
13:42AutomatedTester|cloudato: yup, they have tagged and shipping too
13:43Standard8smacleod: ok, thanks
14:11atoAutomatedTester: Great!
14:11atoAutomatedTester: Do we need to put in a new required Selenium version in our README and changelog?
14:37vladikoffGetting a send keys error in Sel 3.3.1 and geckodriver 0.15.0: `Error: [POST http://localhost:4444/wd/hub/session/611d26b4-6217-4eaa-bbbe-98e5f0af8559/element/28/value / {"value":["s","i","g"]}] Expected [object Undefined] undefined to be a string` If anyone knows what that is about please let me know
14:41vladikoffah https://groups.google.com/forum/#!topic/selenium-users/yqYM0COYO1A
14:44atovladikoff: https://github.com/mozilla/geckodriver/issues/659#issuecomment-295727985
14:44vladikoffato, hero!
14:45atoExpect 0.16.0 to be out within the next hour or so.
14:45ato(Please dont tell the users.)
14:45vladikoffato, sorry I searched through issues but i think i missed it because it was closed :(
14:46atoThats fine (-:
14:46atoAt least you didnt open an issue and publicly berated us for being incompetent (-;
14:57AutomatedTester|cloudAto yup, 3.4+
14:58AutomatedTester|cloudToday has been a fucking "fun" day
15:01atoAutomatedTester|cloud: Just tested Windows with some exploratory testing. We seem good there!
15:01atoIll go ahead and push the button.
15:01AutomatedTester|cloudAto thank you
15:02jgrahamato: Thanks!
15:05atojgraham: r? https://github.com/mozilla/geckodriver/pull/664/files
15:06atoAutomatedTester|cloud: or you
15:06AutomatedTester|cloudato: r+
15:08atoHm, I wonder if I did this correctly. Pushed the tag separately from merging.
15:08atoWell see.
15:10jgrahamI think that's how you have to do it?
15:10jgrahamI'm not sure but I never managed to push a tag and a commit at the same time
15:10atoYes, I just thought pressing the merge button in the UI made it different.
15:10atoI guess its really the same thing.
15:14atomaja_zf: FWIW I see the test expectations I used in the maximize window patch are not true on Windows 7: https://treeherder.mozilla.org/#/jobs?repo=try&revision=4e11fc5b2a4f
15:14atomaja_zf: It just looks like I need to increase the delta a little bit.
15:18vladikoffato, is `Error: call to Function() blocked by CSP` new? We happily removed `unsafe-eval` from our app because geckodriver didn't need it
15:18atoI dont think Ive seen that before.
15:18vladikoffato, oh ok, that's via `execute_async`
15:18atoBut whatever it is, it shouldnt be a result of the geckodriver release.
15:19atogeckodriver is only a HTTP proxy.
15:31marcoekyle: I'm joining
15:31ekylemarco: ok!
15:43vladikoffAutomatedTester|cloud, were Linux 32-bit builds of geckodriver discontinued? (seems gone since ver 0.15
15:51jgrahamvladikoff: We have way more problems making that build than any other configuration
15:51jgrahamato is investgating, but do you really not have a 64bit machine to run this on? :)
15:51vladikoffjgraham, and usage is hopefully low on that one
15:51vladikoffjgraham, nah I maintain a lib that links to downloads, so I'll just remove 32-bit linux support ;)
15:53atoOK, so funny story.
15:54atoThe Linux 32-bit musl build works when compiled as a debug binary, but not optimised.
15:54vladikoffheh
15:54atoThis is why the build is marked as green on GitHub, but failed to upload to the release.
15:54atoI might do a debug compile myself or try with a different rustc version and upload the 32-bit binary manually.
15:55ato32-bit Linux should be burnt to the ground, but I promise Ill upload _something_ within a few hours.
15:55marcomaja_zf: you can file an issue here https://github.com/marco-c/grcov/issues for gcc < 4.9
15:55maja_zfmarco: will do, thanks
16:09atoOuch: http://www.somsubhra.com/github-release-stats/?username=mozilla&repository=geckodriver
16:10atoI guess I may have broken someones CI now.
16:10atoWell see soon! (-:
16:11maja_zfheh
16:11jgrahamHonestly, if you are downloading untested software into your CI occasional breakage is entirely your own fault
16:11vladikofflock them deps
16:12atoYes, its a recipe for a disaster waiting to happen.
16:12jgraham(FWIW I am actually doing that, but I&#39;m also testing the latest Firefox and Chrome nightly builds, so I&#39;m happy to accept responsibility when something goes wrong)
16:12atoThe best bug we had on geckodriver was when GitHub Downloads went down due to the Amazon S3 outage though.
16:13jgraham*bugs
16:13atohttps://github.com/mozilla/geckodriver/issues/501
16:13atoRELEASE DOWNLOAD IS UNAVAILABLE
16:13atoPlease fix ASAP!!! Thank you
16:33AutomatedTester|cloudato: did you see Twitter today (specifically stuff from me)
16:34AutomatedTester|cloudato: chap was complaining you were not professional because you said &quot;it will be out when it&#39;s out&quot; to the sendKeys geckodriver bug
16:53tedAutomatedTester|cloud: ridiculous
16:56AutomatedTester|cloudted: it got worse and he kept digging a bigger hole
16:57tedi saw a bit of your twitter back-and-forth
16:57AutomatedTester|cloudDay in the life of a lead in a successful OSS project
16:59AutomatedTester|cloudIt just frustrates me how people pay $0 for something but feel the developers are beholden to them
17:07wlachit&#39;s not a helpful attitude, no
17:27atoAutomatedTester|cloud: Oh no, I did not see that at all!
17:28atoIve been trying to stay away from Twitter for a while. Havent missed the animosity.
17:30AutomatedTester|cloudato: tldr he was a dick and continued to act like one
17:30atoWow, I can see that now.
17:30atoIve been in this game for long enough to know never to make a promise on a release time.
17:32jgrahamAutomatedTester|cloud: Pretty happy to offer a geckodriver support service for, say $10,000/monthly. I mean I wouldn&#39;t actually do anything for that money, but people would have the satisfaction of their Twitter complaints being justified.
17:33atoOver the past few weeks, Ive literally answered and triaged hundreds of issue reports on geckodriver, but apparently you can never be too careful about the tone you use when you reply.
17:33ato(Proof: https://sny.no/2017/04/diary)
17:34atohttps://irccloud.mozilla.com/file/qoSUfJtK/screenshot-2017-04-21-183119.png
17:34ato is pretty despicable.
17:45ekylechmanchester: i am trying to use the linked-files-map.json, but all sources from the gcno files can not be found in the keys of the dict. I am trying to characterize the problem before I report it; in the meantime...
17:45ekylewhat does &#39;MOZ_OBJDIR&#39;: &#39;obj-firefox&#39;, do?
17:58ekylenm, found some docs
18:59booterrorekyle: Hi, you up?
18:59ekylebooterror: yes, it is the middle of the day for me
19:00booterrorMiddle of Night, here
19:04booterrorekyle: 3pm EDT is 12 am IST
19:04booterrorekyle: I&#39;m living in tomorrow :D
19:06ekylebooterror: we must talk verbally, text is not fast enough
19:11booterrorekyle: How do want to do it? Hangouts call?
19:11ekylebooterror: sure, we can try that
19:13ekylebooterror: we have a codecoverage group, if you want https://hangouts.google.com/hangouts/_/mozilla.com/mozillacodecoverage
19:20ekylebooterror: muted?
19:20ekylebooterror: network issues?
19:23booterrorekyle: Can you hear me?
19:23ekylebooterror: no sound
19:23booterrorekyle: Ahh damn :?
19:23ekylebooterror: keep trying, I am not going anywhere
19:24booterrorekyle: Ok, thanks
19:27booterrorekyle: From your previous mail, you send me a mozilla vidyo link
19:27ekyleyou can try that, yes
19:27ekylebooterror: tell me where you are, so I can switch
19:32ekylebooterror: brb
19:34booterrorekyle: Sure, Let me figure out the sound problem
19:35ekylebooterror: back
19:38booterrorekyle: I&#39;m really sorry, I&#39;m not a day-to-day user of any video calling client, so having a little trouble setting up
19:40ekylebooterror: np, take your time.
19:43AutomatedTesterjgraham: we may offer that service in the future
19:44AutomatedTesterjgraham: with an extra cost of $1000 per hour to fix their &quot;high priority&quot; bugs over doing normal triaging
20:00armenzg_afkgps: thanks for your review the other day. It will be a bit before I can get back to get to it; I&#39;m recofused on a Quantum perf request
20:06ekylebooterror: how&#39;s it going?
20:10booterrorekyle: Trying skype
20:11ekylebooterror: that has the most success
20:20booterrorekyle: I&#39;m up on skype!
20:20ekyleok, just a sec
22 Apr 2017
No messages
   
Last message: 35 days and 17 hours ago