• jgriffin
    fish+tonic, wow that sounds bad
  • philor
    when mochitest-robocop reports a totally different number of tests passed each run, is that because it really does silently fail to run a bunch often as not?
  • glob
    face, tell me a story about your past
  • » face runs
  • glob
    :(
  • Ms2ger
    Eheheheh
  • Ms2ger
    Morning, edmorley
  • edmorley
    good morning! :-)
  • jmaher
    bah, the other timezones are talking
  • » Ms2ger kicks jmaher out
  • AutomatedTester
    edmorley: have you done your rypple feed back
  • AutomatedTester
    or do you have to skip it?
  • edmorley
    AutomatedTester: i was planning on doing it today
  • edmorley
    although yeah i guess it's going to be pretty much ignored due to me falling under the too new category
  • AutomatedTester
    this stuff feels like "Look how amazabubble I am"
  • Ms2ger
    That's the point, no?
  • AutomatedTester
    Ms2ger: yes but doesnt mean I like it
  • Ms2ger
    Of course not, you're not a 'murrican :)
  • » AutomatedTester is glad he isnt japanese
  • AutomatedTester
    "Even with all my immense failures I have managed to these minor things"
  • AutomatedTester
    where minor is actually quite big
  • Ms2ger
    Maybe you should send that
  • AutomatedTester
    reverse psychology on the HR people...
  • Ms2ger
    Who cares about HR people anyway
  • edmorley
    raises...
  • Ms2ger
    Who cares about money anyway
  • AutomatedTester
    Ms2ger: did you read Nicholas Zakas blog post from yesterday
  • Ms2ger
    No
  • » AutomatedTester finds it
  • » Ms2ger googles
  • AutomatedTester
  • AutomatedTester
    Ms2ger: ^
  • AutomatedTester
    the bit about money rings true, at least for my european friends
  • AutomatedTester
    cant talk for the 'murricans
  • Ms2ger
    pedantic details, not pendantic, dammit
  • » Ms2ger has a feeling he got nerd-sniped
  • AutomatedTester
    Ms2ger: going to MozCamp Europe?
  • Ms2ger
    When/where/why?
  • AutomatedTester
    Warsaw, Poland in September
  • AutomatedTester
    because you can hang out with nerds
  • AutomatedTester
    and you dont have to put your IRC nick down on docs I am sure to keep your psuedo-anonyminity
  • AutomatedTester
    well I am sure they would at least make a special case for you
  • Ms2ger
    Mm, September is unlikely
  • AutomatedTester
    fair enough
  • Ms2ger
    (Or, for that matter, June through August and October through January)
  • AutomatedTester
    school?
  • Ms2ger
    Yeah
  • AutomatedTester
    Ms2ger: when do you finish?
  • AutomatedTester
    or are you lecturing now?
  • » Ms2ger wonders how old AutomatedTester thinks he is
  • Ms2ger
    "If I need a queue or a linked list, I use the implementation that�s native to the language I�m using. These are all solved problems."
  • Ms2ger
    (Finishing in a year or two, probably)
  • AutomatedTester
    Ms2ger: I think you're ~22
  • AutomatedTester
    well from the conversations we have had
  • AutomatedTester
    comments in bugs etc
  • Ms2ger
    Not bad, turning 21
  • sawrubh
    Ms2ger: if I may ;), do you plan to reveal someday :P
  • mcote
    *gasp* details about Ms2ger
  • mcote
    well unverifiable ones anyway
  • Ms2ger
    mcote, for you I'm 57 and retired
  • mcote
    nice, early retirement
  • AutomatedTester
    sawrubh: Ms2ger has left his camera on for a second before in a meeting
  • sawrubh
    AutomatedTester: so you guys have seen him ;)
  • AutomatedTester
    I havent, wasnt in that meeting
  • Ms2ger
    Well, the camera happens to be covered physically :)
  • sawrubh
    AutomatedTester: you can be a nice detective(given Ms2ger was correct in saying 21)
  • mcote
    omg why does adb keep going crazy
  • Ms2ger
    mcote, what does the A stand for?
  • mcote
    ah-can't-keep-going
  • mcote
    said with a bit of a drawl
  • sawrubh
    Ms2ger: so you started working even before entering school ?
  • Ms2ger
    Damn
  • Ms2ger
    Gmail claims I started getting involved with this stuff in October 07
  • » Ms2ger feels old
  • sawrubh
    heh
  • AutomatedTester
    Ms2ger: at least you arent like one the current contributors in the UK who is 14
  • AutomatedTester
    think about when he gets to your age
  • » Ms2ger tries to remember how old pavlov was when he got involved
  • » sawrubh wishes he had gotten into this thing earlier
  • » sawrubh they keep getting younger and making me feel bad
  • » mcote thinks all these "linked-in is useless anyway; I'm closing my account" emails are kind of funny, considering that's how I got into Mozilla :)
  • » sawrubh everybody now start thinking why they got here ;)
  • sawrubh
    mcote: how did that happen :)
  • » sawrubh thinks the story might be interesting
  • Ms2ger
    15 or so
  • mcote
    heh not extremely interesting :)
  • » sawrubh still wants to know :P
  • mcote
    recruiters were having a hard time finding people with the a-team's skill set, and they decided to look to "the Great White North"
  • sawrubh
    Ms2ger: what was that 15 or so for ?
  • sawrubh
    Ms2ger: was that when you started contributing ?
  • mcote
    so I guess they started looking around linked in and found me somehow
  • Ms2ger
    Nah, pavlov
  • mcote
    shot me an email asking if I was interested in an interview
  • mcote
    day before my birthday :)
  • Ms2ger
    (He's interviewed in the Code Rush movie about the Netscape code release)
  • Ms2ger
    Along with his parents
  • sawrubh
    mcote: seriously do people get job requests on Linked-in (I still think it's useless)
  • mcote
    I do, most of them not terribly interesting
  • mcote
    except Mozilla's :)
  • mcote
    I get occasional emails from Amazon as well
  • mcote
    via LinkedIn I mean
  • sawrubh
    Ms2ger: I can't find pavlov on Mozillian(maybe he prefers to remain anonymous too :P)
  • mcote
    Google too, of course, though they may have found me elsewhere
  • Ms2ger
    Stuart Parmenter, I don't think he's still around much
  • sawrubh
    Ms2ger: yes I've seen that movie :)
  • » sawrubh thinks he won't close down his Linked-in account after all
  • mcote
    heh
  • mcote
    I mean it hasn't been *that* useful, but without it I may have never been hired here at the bestest job evar
  • sawrubh
    Ms2ger: are you under some legal obligations ?
  • Ms2ger
    I'm not
  • sawrubh
    Ms2ger: just stop me when I'm too annoying, your anonymity just has that allure :P
  • Ms2ger
    mcote, sounds like useful!
  • Ms2ger
    :)
  • mcote
    exactly :)
  • Ms2ger
    Otoh, I get job offers thrown my way without LinkedIn, so I guess I'll manage :)
  • sawrubh
    Ms2ger: which country are you from ? I thought you are still studying(do people get job offers while they are studying ?)
  • sawrubh
    Ms2ger: how do those recruiters find you :P
  • Ms2ger
    I hang around on IRC ;)
  • » sawrubh ehsan must have gotten tired of this :P
  • sawrubh
    heh
  • AutomatedTester
    Ms2ger: do you get job offers outside on Mozilla?
  • sawrubh
    Ms2ger: from within the Mozilla irc :P
  • Ms2ger
    AutomatedTester, not explicitly
  • AutomatedTester
    just curious since there isnt much of a trail on you so thought recruiters would be scared off by that
  • » sawrubh just thinks about how much of my information is on the interwebs
  • sawrubh
    Ms2ger: have you tried mirror.me ? Bet they can't make a map of you
  • » AutomatedTester does love how Ms2ger has put a little bit of fear into W3C because of his psuedo-anonyminity
  • sawrubh
    I mean a cloud
  • sawrubh
    AutomatedTester: why has he proposed some spec too ?
  • Ms2ger
    AutomatedTester, silly legal people...
  • AutomatedTester
    Ms2ger: I know
  • mcote
    for most people anonymity is virtually impossible; the best most of us can do is attempt to present ourselves the way we want to be seen
  • AutomatedTester
    sawrubh: Ms2ger has made some really good changes to DOM4
  • Ms2ger
    I care about Boston, MA, apparently
  • Ms2ger
    And kickboxing?!
  • sawrubh
    Ms2ger: did they use your twitter account for creating your cloud, coz certainly you don't have a footprint anywhere else.
  • Ms2ger
    Yeah
  • AutomatedTester
    Ms2ger: is canadian?!?!?!
  • » sawrubh reads about DOM4
  • Ms2ger
    Also Swedish
  • Ms2ger
    "birth"
  • sawrubh
    Ms2ger: dude you've got like 0 tweets how can they create your cloud. What heuristics do they use- your followers :P
  • AutomatedTester
    Ms2ger: dont worry, mirror.me got me wrong too
  • Ms2ger
    sawrubh, I guess that's how "India" ended up in there :)
  • AutomatedTester
    wtf is Ruby there...
  • sawrubh
    heh
  • Ms2ger
    I'm fascinated by the huge "java"
  • » sawrubh seriously doubts if that India has been cause by him ;)
  • AutomatedTester
    *sigh* I hate it when I find Firefox bugs when I dont really want to
  • Ms2ger
    Vikash too, I guess
  • sawrubh
    we are causing turbulence in your cloud :P
  • Ms2ger
    That looks about it
  • edmorley
    boo twitter
  • edmorley
    I was expecting mirror.me to be something more interesting
  • sawrubh
    Ms2ger: why do you have a dutch tag to our mozillians account(is that for distracting people :P)
  • » sawrubh thinks it's time to stop interrogating so much and will quite after this last question :P
  • » sawrubh *quit
  • Ms2ger
    AutomatedTester, thanks for the link, enjoyed the article
  • AutomatedTester
    Ms2ger: I enjoyed it and found it quite true for a lot of Mozillians
  • AutomatedTester
    this rypple stuff makes me a sad panda
  • AutomatedTester
    Ms2ger: I think I might have to put "Even with all my immense failures I have managed to these minor things" in
  • edmorley
    AutomatedTester: what aspect specifically?
  • Ms2ger
    AutomatedTester, do it! :)
  • AutomatedTester
    I have done my accomplishments and now doing challeneges
  • AutomatedTester
    and feeling that "Everything is a challenge" might be the most valuable thing
  • edmorley
    I don't even have anything to talk about in the gaols bit, since my gaols weren't added for the last time
  • mcote
    heh gaol
  • mcote
    that is an old word for "jail" :)
  • mcote
    brb
  • Ms2ger
    Goal: get a better job ☑
  • » face showed that to wlach last month :)
  • AutomatedTester
    mcote: do we have to submit our rypple stuff this week
  • mcote
    yeah by Friday I believe
  • AutomatedTester
    I was hoping I could speak to ctalbert|afk before submitting
  • AutomatedTester
    oh well
  • mcote
    you'll be speaking with him afterward :)
  • AutomatedTester
    true
  • mcote
    the process is you write up a self-eval, he writes up an eval of you, then you compare and chat about it
  • mcote
    probably best if the former is done entirely by yourself... more honest that way
  • AutomatedTester
    mcote: it was more "should I put X in" type questions that I had
  • mcote
    ah, well I could perhaps help with those questions :)
  • AutomatedTester
    mcote: thanks for the offer
  • AutomatedTester
    I might take you up on that
  • mcote
    sweet finally getting some reasonable predictions out of orange seed
  • AutomatedTester
    nice mcote
  • whimboo
    hm is jeff not around today>
  • whimboo
    ?
  • mcote
    jhammel?
  • mcote
    well it's only 7:46 am PT
  • whimboo
    heh, i always forget about the timezones :)
  • whimboo
    especially if its raining outside and it's dark
  • whimboo
    i feel it's already way later
  • jmaher
    yeah, jhammel usually pops on in 30-60 minutes from now
  • mcote
    I mean he's an earlybird but not normally *quite* this early :)
  • whimboo
    i have a massive patch for him today to review :)
  • whimboo
    he will be happy
  • jmaher
    ha
  • mcote
    as long as you don't have "opinions"
  • mcote
    :)
  • whimboo
    no strong opinions, no
  • glob
    heh
  • wlach_
    jmaher: your patch looks good for bug 754873-- I take it you resolved your issues?
  • bugbot
    Bug bugzilla.mozilla.org/754873 normal, --, ---, jmaher, NEW , fennec automated tests should capture logcat output while test is running
  • marshall_law
    do we have any async helpers in marionette for queueing up tests in a given order?
  • jmaher
    wlach_: yes, it logs great now
  • jmaher
    logbot: you can see logs here: tbpl.mozilla.org/?tree=Try&rev=1102b444f235
  • logbot
    jmaher, found 0 results
  • wlach_
    jmaher: what was the problem before?
  • jmaher
    oh, wlach_ ^
  • mdas
    marshall_law: no, the python marionette test runner uses unittest, and I don't think they guarantee order
  • marshall_law
    mdas: I guess I just mean on the JS side
  • jmaher
    wlach_: I needed to split on \r, not \n
  • jmaher
    that was one problem
  • marshall_law
    mdas: np, just curious :)
  • mdas
    marshall_law: can you give me an example usage? Marionette itself doesn't really understand the concept of test suites
  • mdas
    it just drives user actions
  • jmaher
    and another was getting everything correct between devicemanager and the respective subclasses
  • wlach_
    jmaher: cool. well I'm glad we figured that out-- was worried that shell() was buggy or something
  • mdas
    marshall_law: it has a mechanism for keeping track of asserts and stuff (is, isnot methods), but that's all it really knows about
  • marshall_law
    kk
  • marshall_law
    mdas: I'm really just looking for some simple helpers on the JS side I guess, nothing from the driver specifically
  • mdas
    marshall_law: well, if all your tests are js tests, you can set them in order in a unit-test.ini file and they should get run in that order. But for JS tests, people are using the test framekwork that lightsofapollo set up
  • mdas
    and I'm not sure how/if it has ordering
  • marshall_law
    oooh
  • marshall_law
    mdas: I didn't realize there was a higher level framework :)
  • marshall_law
    mdas: you happen to have a link?
  • jmaher
    wlach_: thanks!
  • mdas
  • marshall_law
    mdas: sweet, thanks!
  • marshall_law
    ahh this is for general unit tests I guess
  • marshall_law
    mdas: FWIW, the suite I've been working on is an emulator-specific test suite for the mobilenetwork APIs
  • marshall_law
    probably needs to be device/emulator specific
  • mdas
    marshall_law: what makes it emulator specific?
  • mdas
    are you setting network parameters and stuff on the emulator?
  • marshall_law
    mdas: it depends on the mobile networks present in the emulator modem
  • marshall_law
    yup
  • mdas
    ah okay
  • marshall_law
    it's for switching networks, etc
  • mdas
    ohh
  • mdas
    then you want one of philikon's tests
  • mdas
    I'll show you
  • » mdas needs to update docs *soon*
  • marshall_law
    mdas: FWIW it's already partially landed, but I have some updates that are coming in another patch this week
  • mdas
    can I see it?
  • marshall_law
    sure!
  • marshall_law
  • marshall_law
    mdas: lots of async madness
  • marshall_law
    :)
  • mdas
    marshall_law: ah okay, do you plan on running emulator commands during the test?
  • mdas
    for example, to shut off network access etc
  • marshall_law
    mdas: nope shouldn't need to. the test is pretty encapsulated
  • marshall_law
    mdas: at least for now. I guess it might make sense to do that, now that you mention it though ;)
  • mdas
    heh
  • mdas
    yeah that's a good test if there are qemu/android console commands to control network, which I'm fairly certain there is
  • marshall_law
    mdas: the networks in the emulator aren't really configurable right now, they're built in to the emulator's modem
  • mdas
    but yeah from the looks of it, this can run either through marionette, or through lightsofapollo's test framework
  • marshall_law
    I've already had to fix several emulator modem bugs to get this working
  • mdas
    marshall_law: well acutally, it should probably be checked in as a marionette test, since it will fail in a desktop environment.
  • mdas
    marshall_law: ah, that sounds unpleasant.
  • marshall_law
    fair enough. with marionette I can specify that the test runs on the emulator only, though..
  • marshall_law
    yeah
  • mdas
    yup you can do that in the unit-test.ini
  • marshall_law
    yup I did :)
  • mdas
    awesome!
  • marshall_law
    mdas: how hard is it to get marionette running on device? ;)
  • marshall_law
    it would be nice to be able to run venv_test.sh directly without having to go through the interactive console :)
  • mdas
    marshall_law: it runs by default on the device :_
  • mdas
    :)
  • marshall_law
    hrm
  • marshall_law
    mdas: is there a script / args I should be aware of?
  • marshall_law
    right now I'm mostly using the b2g test.sh frontend to venv_test.sh
  • mdas
    marshall_law: you'll need to forward the port from the device to your computer so you can talk with it
  • mdas
    u
  • Ms2ger
    A new guy?
  • marshall_law
    mdas: yeah, I was hoping there was something that automated that as well ;)
  • marshall_law
    Ms2ger: I'm pretty new.. started on b2g platform ~3 weeks ago
  • Ms2ger
    Oh, pff, b2g ;)
  • marshall_law
    :)
  • mdas
    marshall_law: no but that's the only thing you need to run to get it going
  • Ms2ger
    You'll get to know me, I troll this channel
  • Ms2ger
    (Well, not just *this* channel)
  • mdas
    Ms2ger is mozilla's very own trollbot
  • mdas
    or troll something. no one quite knows
  • marshall_law
    mdas: kk fair enough. I guess if --arch (whatever) is passed that will still try to launch / use the emulator?
  • marshall_law
    Ms2ger: hah nice ;)
  • mdas
    marshall_law: --arch? what is the parameter and what are oyu sending it to?
  • mdas
    ohhhhh, you mean --type=<arch> for the test runner?
  • marshall_law
    ah eyah
  • marshall_law
    my bad
  • marshall_law
    sorry I was pulling from memory
  • mdas
    it will only launch the emulator if you don't specify an address
  • marshall_law
    let me check the test.sh frontend in b2g
  • mdas
    marshall_law: if you specify and address, then it assumes you already have it running
  • marshall_law
    venv_test.sh --emulator=$ARCH --homedir=$B2G_HOME --type=b2g
  • marshall_law
    I guess I'll need to update our frontend to support both then
  • marshall_law
    mdas: so --address=localhost:<whatever adb forward port>
  • mdas
    marshall_law: yeah and you won't nee the --emulator command either, that's for launching the emulator.
  • mdas
    yeah so all you'll need is venv_test.sh --address=localhost:<port> --type=b2g
  • marshall_law
    k
  • marshall_law
    nice! k
  • marshall_law
    I'll update the test script
  • mdas
    cool
  • marshall_law
    mdas: would you accept a bug/patch to do the port forwarding part for a device or specific serial in marionette directly?
  • marshall_law
    you guys already have the code for finding an open port and using it, it would be nice to reuse that :)
  • jmaher
    we have code in devicemanager to find an open port
  • jmaher
    ^ (sans windows)
  • marshall_law
    I don't think B2G has a prayer of building in Windows atm anyway :)
  • jmaher
  • marshall_law
    hrm
  • mdas
    marshall_law: hm, yeah if the you specify the address and some option like --forward, then sure you can add it and I'lll accept it!
  • mdas
    jmaher: marshall_law: we aren't using devicemanager in the test runner for the marionette tests
  • mdas
    we use it for the mochitest stuff
  • jmaher
    mdas: but you could copy that code if you wanted to as a starting piont
  • marshall_law
    jmaher: that seems a little overkill? can't you just open a socket on port 0 then get the port and close? it's not 100% fullproof but should work in most situations?
  • jmaher
    err point
  • mdas
    ah true
  • jmaher
    marshall_law: we write automation to kick ass :)
  • marshall_law
    hehe
  • marshall_law
    fair enough :)
  • marshall_law
    looks like that's what marionette is doing right now anyway
  • jmaher
    well, this runs on a machine with hundreds of processes and each one could be trying to open a port in the same general range
  • marshall_law
  • marshall_law
    jmaher: ahh yeah that makes sense
  • marshall_law
    mdas: I was thinking something even nicer like --device=MY_DEVICE_ID
  • wlach_
    gbrown: any particular reason you set the time for sleep to 500ms in bug 764147? I'm wondering if it might not be better to set it to something a bit smaller
  • bugbot
    Bug bugzilla.mozilla.org/764147 normal, --, ---, gbrown, NEW , SUT agent: near 100% cpu during StartPrg2
  • marshall_law
    (from adb devices)
  • wlach_
    gbrown: I think even a 25ms sleep should fix the CPU problems
  • Ms2ger
    Oh, jmaher is here
  • gbrown
    wlach_: FindProcThread uses 500 ms. but you are right -- that is a long time. I'll reduce it.
  • jmaher
    I am going afk to run an error, back in a while
  • wlach_
    gbrown: maybe we should lower it there too. ;) I'm a bit nervous about long timeouts partially because I've been having problems with latency with eideticker lately. not sure if that applies here but I don't see any good reason to have really long timeouts
  • gbrown
    agreed
  • mdas
    marshall_law: yes, even better!
  • mdas
    forgot about that. I'm so used to just running one emulator
  • marshall_law
    :)
  • marshall_law
    I usually have emulator + a device
  • mdas
    yeah same
  • mdas
    but I use -e
  • marshall_law
    soon probably multiple devices
  • mdas
    but that doesn't work with multiple emus ahah
  • marshall_law
    yup
  • marshall_law
    in adb you can just use adb -s <device id>
  • mdas
    yup
  • whimboo
    jhammel: heyas
  • jhammel
    whimboo: hi; how's it going?
  • whimboo
    jhammel: i would need your help. do you see anything like that on windows?
  • whimboo
  • jhammel
    on windows? ;)
  • whimboo
    i constantly see it on os x and windows
  • whimboo
    haven't tested linux yet
  • jhammel
    i have seen that though i can't recall when exactly (and not just with mozmill, with other things using sockets)
  • jhammel
    any python sockets experts?
  • whimboo
    jhammel: would you mind to run
  • whimboo
    mozmill -b /Applications/Firefox/Nightly.app/ -m mutt/mutt/tests/js/frame/tests.ini
  • whimboo
    just with an updates app path
  • whimboo
    after testThree for persisted tests it always fails for me
  • whimboo
    it has been started today
  • whimboo
    not sure why
  • jhammel
    updates app path?
  • whimboo
    updated
  • jhammel
    sure, just a minute or two
  • whimboo
    i will file the bug
  • whimboo
    it blocks me from finishing the jsbridge patch
  • whimboo
  • bugbot
    Bug bugzilla.mozilla.org/764442 critical, --, ---, nobody, NEW , Constant 'Disconnect Error: Application unexpectedly closed' errors
  • jhammel
    thanks
  • whimboo
    i will try with a fresh git clone
  • whimboo
    i will also try on linux now
  • mdas
    ted: ping
  • jhammel
    i'm trying too, now that i'm done posting to bugzilla
  • jhammel
    whimboo: i do get INFO | Timeout: bridge.execFunction("50bcbbfc-b574-11e1-9d11-0024d7065b60", bridge.registry["{4601040d-b523-4dec-9477-8bcc006e9220}"]["runTestFile"], ["/home/jhammel/mozmill/src/mozmill/mutt/mutt/tests/js/frame/persisted.js", false, null])
  • jhammel
    between each restart :(
  • ted
    mdas: pong
  • whimboo
    jhammel: do you run only this test or all from frame/tests.ini?
  • jgriffin
    ahal: dd you get that error figured out you saw when running mochitests?
  • jhammel
    whimboo: i gave it the whole manifest
  • jhammel
    whimboo: its still running but i'm having all sorts of timeouts :(
  • jhammel
    (hopefully related)
  • whimboo
    jhammel: i assume you have't seen anything like that in the last couple of days
  • whimboo
    how about /tests/js/tests.ini
  • jhammel
    e.g. INFO | Timeout: bridge.execFunction("0bc34cd6-b575-11e1-9d11-0024d7065b60", bridge.registry["{d5026cf6-5956-4420-b3b5-f897305055a9}"]["runTestFile"], ["/home/jhammel/mozmill/src/mozmill/mutt/mutt/tests/js/frame/stop_runner.js", false, null])
  • whimboo
    do you see that for all tests?
  • jhammel
    i'll run that when this run is done
  • whimboo
    k
  • jhammel
    the timeouts make the runs very long indeed :(
  • whimboo
    i also tried with aurora down to release but all the same
  • mdas
    hey ted, I'm having trouble getting marionette built into the package. I've added the needed files in the package-manifest file under the [browser] section and updated the Makefile.in file with the ENABLE_MARIONETTE changes... but i'm still not seeing it
  • mdas
    do you mind taking a look at my changes to see if you see something wrong?
  • whimboo
    jhammel: the whole network stack is playing wild here
  • whimboo
    and i have ohter crazy behavior
  • mdas
    ted: pastebin.mozilla.org/1661893 I have this, plus ENABLE_MARIONETTE=1 in my mozconfig
  • jhammel
    whimboo: okay, running the whole js test suite
  • jhammel
    whimboo: fwiw i haven't run the tests in awhile
  • mdas
    oh and AC_SUBST(ENABLE_MARIONETTE) is in configure.in
  • jhammel
    i haven't had good results from the js tests this year and i don't think i've had good luck with the mutt python tests since returning from london :(
  • whimboo
    jhammel: all tests work now!
  • whimboo
    that was my work last week
  • ted
    mdas: okay
  • ted
    mdas: oh, crap that makefile hunk needs to be in browser/installer/Makefile.in
  • ted
    sorry
  • jhammel
    whimboo: they haven't for me yet :(
  • whimboo
    jhammel: are you on an old revision?
  • jhammel
    whimboo: nope
  • jhammel
    just pulled today
  • mdas
    ted: oh okay thanks I'll give that a shot
  • ted
    mdas: yeah, that's where package-manifest gets preprocessed
  • mdas
    ted: gotcha
  • jgriffin
    mcote|lunch: so I looked at the OrangeSeed code
  • jgriffin
    but you're at lunch, we can discuss when you get back
  • jhammel
    whimboo: do you want my log file?
  • mcote
    mid way through lunch :)
  • mcote
    actually i fixed it up more
  • jgriffin
    I didn't find any bugs, I think it's just more a matter of how we're using the algorithm and what we expect from it
  • mcote
    i think i'm getting not-too-bad results now
  • jgriffin
    cool
  • mcote
    actually one big problem was that I was using the dates of the overall first and last occurrences for extrapolation
  • jgriffin
    instead of OS-specific dates?
  • mcote
    I should have been using the first & last according to each combination
  • mcote
    yeah OS/type/test/tree
  • jgriffin
    yeah
  • mcote
    once I fixed that I got much better results, at least from the one or two bugs I've tried so far
  • mcote
    I just finished refactoring the hell out of it as well :)
  • jgriffin
    I also wonder how much value there is to trying to find this per-OS, rather than in aggregate
  • jgriffin
    since persumably a given bug was actually introduced just once, and not multiple times, once per os
  • mcote
    well I think it is worthwhile to split it up into combinations; however, we don't have to display all the results
  • mcote
    just maybe the ones that make the most sense--say, with the highest number of hits and/or frequency
  • mcote
    and prefer m-i to all other trees
  • whimboo
    jhammel: no. i see the same on linux
  • jgriffin
    yeah
  • mcote
    anyway I have to run out for an hour; I'll show you some results when I'm back and get your opinions
  • jgriffin
    cool
  • mcote
    also wouldn't mind chatting about how to do this in the UI
  • jgriffin
    have fun at the SAQ
  • mcote|lunch
    ha CostCo actually :)
  • jgriffin
    heh
  • mcote|lunch
    (though I will be buying wine there...)
  • jmaher
    jgriffin: I gave you a patch to review this morning, should be about final for the android mochitests
  • jgriffin
    jmaher: I just pulled it up!
  • jmaher
    awesome
  • jgriffin
    nice solution
  • jmaher
    jgriffin: thanks; it feels a bit hacky, but not too insane
  • jhammel
    edmorley: i'm guessing this is a can of worms, but it would be nice to know on TBPL if all jobs were finished
  • jmaher
    jhammel: maybe a progress bar?
  • jhammel
    jmaher: that would work
  • jhammel
    i'd even settle for <h3>You're done!</h3> ;)
  • jhammel
    or <h3>This try job is over!</h3>
  • ted
    not sure if TBPL has enough info to know that
  • ted
    there's probably not anything that defines "the set of all jobs that would run on this push"
  • jhammel
    yeah :( that's why i mentioned the canned annelids
  • ted
    heh
  • edmorley
  • jhammel
    edmorley: itym bz new 'Tinderboxpushlog'
  • jhammel
  • bugbot
    Bug bugzilla.mozilla.org/764464 normal, --, ---, nobody, NEW , Indicate if all jobs are done
  • whimboo
    jhammel: anything you can do on this bug today?
  • whimboo
    jhammel: i mean the one i have filed
  • edmorley
    jhammel: :-)
  • jhammel
    whimboo: not at the moment....looking at other bugs right now, sorry
  • edmorley
    jhammel: btw, by all done, would you be happy with "no jobs in buildapi listed as pending/running" or does it have to be "no jobs pending/running plus none somehow omitted by the scheduler by cross referencing against a non-existent master list"?
  • jhammel
    edmorley: unfortunately, the latter :( I can see currently if there are any greys and know its not done but there have been several occassions where i had thought all jobs were done but i guess(?) weren't scheduled yet or lord knows what and did not show up as grey
  • edmorley
    ok
  • carljm
    samliu: are you available at 3pm (PT) for a meeting about that mobile evangelism tool project?
  • carljm
    samliu: and what's your email?
  • ted
    wlach_: ping
  • wlach
    ted: pong
  • ted
    wlach: hey, if we were to get you a crappy armv6 android tablet with HDMI output, how hard would it be for you to get eideticker results for stock browser/fennec?
  • wlach
    ted: shouldn't be terribly hard, except for the fact that I'll be away for the next two weeks starting friday
  • ted
    okay
  • ted
    probably not critical to get in that timeframe
  • ted
    thanks
  • wlach
    ted: I guess the main thing is that it be HDMI out that mirrors the screen. some android devices only do hdmi for specific apps
  • ted
    ah
  • ted
    yeah, i have no idea
  • ted
    it's probably cheap enough that we can just buy one and find out
  • ted
  • mdas
    ted: yeah i saw that! super cool. who designed the rack?
  • dividehex
    i did
  • ted
    mdas: came up again in the meeting :)
  • mdas
    dividehex: nice job:)
  • dividehex
    thanks
  • ted
    yeah, it's super awesome
  • dividehex
    it will look really cool when we have 60 of them in SCL3
  • jhammel
    how big is it? 6u?
  • jhammel
    i assume these will go in actual rack racks
  • jhammel
    oh dromaeo...oh dromaeo....why dost thou take so long, dromaeo?
  • dividehex
    its 4U and yes it will go into the server racks in SCL3
  • dividehex
    it will be able to handle both tegras and pandas
  • jhammel
    *actual* pandas? ;)
  • ted
    not the kung-fu variety
  • ted
    wlach: don't suppose you can tell from the crappy alibaba page whether it's likely to support display mirroring or not? ananda.en.alibaba.com/product/53635…PC_processor_ARMv6_compatiable.html
  • wlach
    ted: looking
  • ted
    the thing is literally like $130, so i'll just have someone order one for you unless you think it's really unlikely to work
  • wlach
    ted: yeah my guess is that it'll work. I bet display mirroring is the easy way of doing hdmi
  • ted
    okay
  • ted
    great
  • dividehex
  • bugbot
    Bug 731670: normal, --, ---, ctalbert, NEW , Figure out changing MAC address for panda boards
  • dividehex
    joduinn-mtg: also ^^
  • ted
    dividehex: oh ho
  • ted
    thanks
  • ted
    wlach: filed bug 764491 FYI
  • bugbot
    Bug bugzilla.mozilla.org/764491 normal, --, ---, desktop-support, NEW , Hardware Request: armv6 tablet for William Lachance
  • jhammel
    beh, still no action on Bug 123456 : Hardware Request: cray XK6 for Jeff Hammel
  • bugbot
    Bug bugzilla.mozilla.org/123456 is not accessible.
  • wlach
    ted: awesome. ty
  • wlach
    I'm sure Diane will be thrilled to have yet another computer in our house
  • mdas
    ahal: the first cappuccino thing I made was the best. I keep making fail ones now:(
  • jhammel
    wlach: get another cat to offset it :)
  • mdas
    wlach: how many do you have at this point?
  • ahal
    mdas: you'll get better with practice
  • ahal
    maybe
  • mdas
    heheh
  • wlach
    mdas: I think between all my mobile phones and computers, something like 8-9
  • wlach
    it's a bit appalling
  • jhammel
    wlach: i think she meant cats
  • jhammel
    where 8-9 is a completely acceptable number
  • mdas
    wlach: yeah i have about 10 computers and phones, but also 2 xboxes and 1 ps3...
  • jhammel
    wow
  • mdas
    then one laptop died and i gave back the tablet! so i'm down to 8 or so
  • jmaher
    I have 9 devices, but 3 usually are offline
  • mdas
    it's kind of insane .
  • jhammel
    between home and work i have 3 working laptops, 1 phone, one tablet, and two items in the "dead hardware i should probably get rid of" drawer
  • jhammel
    though, speaking of, if anyone wants a first gen iphone...
  • mdas
    hahaha
  • mdas
    I do need a new doorstop
  • jhammel
    mdas: it will only work with iDoors
  • mdas
    lol!
  • ahal
    haha
  • mdas
    wlach: do you have more than one desktop computer?
  • mdas
    those take up so much space
  • wlach
    mdas: just one fortunately
  • mdas
    and generate so much heat. and are loud'
  • mdas
    good!
  • jhammel
    i guess
  • jgriffin
    not to mention making one very hot in the summer
  • mdas
    I dislike them generally
  • jhammel
    i've seen desktops that are pretty quiet
  • jhammel
    mdas: you can give me yours ;)
  • mdas
    this one here is very quiet
  • jhammel
    i actually really need a desktop
  • mdas
    the one I have at home is like, the loudest
  • jhammel
    laptops are kinda annoying due to the cord factor
  • mdas
    I could buy a quieter fan for it, but i don't actually use it often heh
  • jhammel
    vs a desktop where you can get a wireless keyboard and mouse
  • jhammel
    and store e.g. on a bookshelf
  • ted
    most of my desktops are fairly quiet nowadays
  • ted
    my MBP is louder
  • ted
    when the fans spin up
  • ted
    WHIIIIIIRRRRRRRR
  • ted
    wlach: having more devices is awesome
  • jhammel
    my netbook is actually the loudest
  • ahal
    jhammel: ew wireless mouse
  • jhammel
    ahal: ?
  • ted
    especially when they're more crappy chinese devices
  • mdas
    yeah it sounds like it's taking off on a jetpack
  • jhammel
    ahal: oh, i see what you mean...why have a mouse? good point ;)
  • ahal
    jhammel: the battery always dies and it sucks for gaming
  • jhammel
    ahal: i don't game ;)
  • » jmaher either
  • wlach
    ted: do crappy chinese armv6 devices make you weep for humankind?
  • jmaher
    I haven't figured out how to use a mouse for 'head games'
  • jhammel
    actually, i play this text based adventure called POSIX
  • wlach
    just want to know what to expect
  • jmaher
    jhammel: sounds like a fun game
  • ted
    wlach: no, i just would avoid putting your hands in your mouth after handling it
  • ahal|afk
    have a doctors appointment should be back in an hour or two
  • jhammel
    jmaher: ahal is offended at our non-gaming stance
  • jmaher
    well, we are bleeding edge americans
  • wlach
    ted: do they use melamine as an additive with these things?
  • ted
    i'm not sure, but it wouldn't surprise me
  • jhammel
    sounds nutritious
  • jhammel
    jgriffin: btw, the meeting info wiki.mozilla.org/ReleaseEngineering/Mozharness
  • jgriffin
    jhammel: cool, thanks
  • » AutomatedTester wonders if he should play LDAP roulette...
  • jhammel|lunch
    AutomatedTester: sounds "fun"
  • jhammel|lunch
    i know Russian roulette and Polish roulette
  • AutomatedTester
    jhammel|lunch: on the news it looks like there was a riot about that last night
  • AutomatedTester
    all because of stupid football
  • jhammel|lunch
    AutomatedTester: i thought being English you *had* to like football...or does your SA past excuse you?
  • AutomatedTester
    jhammel|lunch: I enjoy the other English game of rugby
  • AutomatedTester
    which is well known in SA
  • AutomatedTester
    its like American Football but without all the protective gear
  • dividehex
    ted: what product and component should i file a bug under if its for you?
  • jhammel|lunch
    AutomatedTester: no cricket?
  • AutomatedTester
    and you cant throw the ball forward
  • AutomatedTester
    not cricket
  • AutomatedTester
    jhammel|lunch: I am thinking of staying in SF for my entire visit in just over a week and commuting down to Santa Clara for my conf
  • jhammel|lunch
    AutomatedTester: :)
  • jhammel|lunch
    otoh, commuting :(
  • AutomatedTester
    do you think BART to Freemont and then VTA to Santa Clara or CalTrain to MV and then VTA to Santa Clara?
  • AutomatedTester
    the BART option also means I can use AmTrak if need be
  • jhammel|lunch
    oh, the Caltrain still requires a transfer? :(
  • AutomatedTester
    yea, its not near the station :/
  • AutomatedTester
    Google maps wasnt helpful
  • AutomatedTester
    said it was the same either way
  • jhammel|lunch
    hmmmm.....i'd do whichever is shorter. i haven't done Freemont -> VTA myself
  • AutomatedTester
    ok
  • xabolcs
    whimboo: can i have some question about the ntt issues?
  • jhammel|lunch
    unless you're staying near the caltain, though, there is the getting to the caltrain step
  • » AutomatedTester is staying near the SF office
  • AutomatedTester
    well, I should book that
  • AutomatedTester
    2 blocks away and nearer the Market street BART
  • jhammel|lunch
    yeah; i'd try bart -> freemont....see how awful it is
  • jhammel|lunch
    AutomatedTester: looks like bart -> fremont is 46 min
  • whimboo
    xabolcs: sure
  • AutomatedTester
    jhammel|lunch: total commute is up to 2hrs
  • jhammel|lunch
    :( :(
  • jhammel|lunch
    awful
  • AutomatedTester
    jhammel|lunch: 4hrs of commuting vs having multiple hotels and dragging luggage all over the bay area
  • AutomatedTester
    rock > automatedtester > hard place
  • jhammel|lunch
    AutomatedTester: tell them to move the conference to somewhere more convenient to you ;)
  • whimboo
    jhammel|lunch: i assume you will not have any time to check the jsbridge breakage?
  • jhammel|lunch
    whimboo: probably not today
  • xabolcs
    whimboo: thanks! in issue #83 the recommended 'if (PopupNotifications)' check drops 'PopupNotifications is undefined' :(
  • » AutomatedTester phones Steve Souders and Tim O'Reilly quicky
  • whimboo
    jhammel|lunch: :( so i will have to look into now. because it completely blocks my work
  • xabolcs
    so i would like no to address that nit
  • whimboo
    xabolcs: it's ok
  • mcote
    samliu: ping
  • xabolcs
    whimboo: ok. another one. in #75 (push to tip menuitem) you wrote to remove the extra currChangeset variable in openPushlogSinceCurrentBuild()
  • samliu
    mcote: pong
  • mcote
    samliu: I need your hands again
  • mcote
    adb blew up again for some reason
  • mcote
    I think this time we should just leave 3 devices connected for now at least
  • mcote
    see if that's related
  • samliu
    mcote: cool, can we do this at 1 (in an hour)
  • mcote
    yup yup
  • xabolcs
    whimboo: should i remove in openPushlogToCurrentBuild() too?
  • samliu
    aiite I'll let you know when I get into haxxor :D
  • whimboo
    xabolcs: please give me links so I can recheck
  • whimboo
    i do not have all in my head. sorry :)
  • xabolcs
  • xabolcs
  • whimboo
    xabolcs: no, in the other method it is fine
  • xabolcs
    whimboo: ok
  • xabolcs
    whimoo: in resize columns issue - mozilla/nightlytt #74/files#r966123
  • xabolcs
    whimboo: what about to file a follow-up bug for the 'hidecolumnpicker' stuff?
  • xabolcs
    whimboo: to be a preference controlled stuff?
  • xabolcs
    whimboo: 'cause i think it not really related to resizing :)
  • whimboo
    xabolcs: why should it be configurable via a preference?
  • whimboo
    let users show/hide columns they don't want to have
  • whimboo
    you are introducing the this attribute there with the tree feature
  • xabolcs
    whimboo: yes, but i don't want to be showen that columnpicker
  • whimboo
    so let it be
  • xabolcs
    whimboo: and there are another stuffs, like column order :)
  • xabolcs
    which could be also introduced with that follow-up issue
  • whimboo
    it's fine if you leave both alone for now
  • xabolcs
    whimboo: by the way, a preference window seems to be come with the Suite support of tiltebar customization
  • xabolcs
    whimboo: another question, in #74 too - mozilla/nightlytt #74/files#r966111
  • xabolcs
    whimboo: what will call setupTree()? (to acces the tree with origialTarget)
  • xabolcs
    whimboo: because it's is called by init() actually
  • xabolcs
    whimboo: ahh sorry
  • xabolcs
    missunderstood that comment
  • mdas
    no brain, it's 'make package', not 'make pancakes'...
  • » mdas sighs
  • jhammel
    mdas: make pancakes only works on a few devices, sadly
  • mdas
    heh
  • jhammel
    make: *** No batter to make target `pancakes'. Stop.
  • jhammel
    i always wondered why the first quote was a grave quote and the second quote was normal o_O
  • jhammel
    i guess its a grave situation but not *that* grave
  • mdas
    whoa
  • mdas
    really? i never noticed
  • mdas
    hah
  • AutomatedTester
    hmmm pancakes *drool*
  • jhammel
    ew, i never liked drool pancakes
  • xabolcs
    whimboo: one question to #74 - resize columns - mozilla/nightlytt #74#discussion_r966085
  • xabolcs
    whimboo: you spotted to move that style definition to css file
  • mdas
    clear
  • xabolcs
    whimboo: but in #55 - mozilla/nightlytt 32a6bc6#L1R44
  • mdas
    lol whoops
  • jhammel
    mdas: clear: removing all files from filesystem....done.
  • jhammel
    i actually never use clear
  • xabolcs
    whimboo: was left there. if i commit that css fix on top of my #55 branch would you mind to cherrypick that?
  • mdas
    jhammel: when my brain wants to make pancakes, i use clear
  • mdas
    cluttered screens don't help the situation haha
  • jhammel
    mdas: i have a script that outputs, say, 1000 lines of nothing
  • jhammel
    since clear + shift+pageup are questionably useful
  • ahal
    mdas: you need a latte!
  • jhammel
    you know what they say....better latte than never?
  • ahal
    I hate myself for smiling
  • mdas
    Hah!
  • mdas
    and yeah, why not anohter latte.
  • jhammel
    don't you know that lattes are a gateway drug?
  • mdas
    jhammel: gateway drug to... huffing kombucha?
  • whimboo
    xabolcs: just comment on the pull what you want to do. sounds fine to me
  • mdas
    god that sounds disgusting.
  • whimboo
    xabolcs: sorry, but I'm deep in a problem right now
  • jhammel
    mdas: pretty much, on both counts
  • jhammel
    or yerba mate
  • whimboo
    jhammel: this issue kills me!!
  • jhammel
    :(
  • » AutomatedTester is getting excited about bugzilla.mozilla.org/753022
  • bugbot
    Bug 753022: normal, P3, ---, mdas, NEW , debug builds should have marionette enabled and tested
  • xabolcs
    whimboo: ok, no problem
  • xabolcs
    whimboo: thanks for the answers, all my question was answered :)
  • whimboo
    xabolcs: sweet. :)
  • whimboo
    jhammel: lemme try to at least find a simple testcase
  • jhammel
    we should probably have a testcase for this anyway
  • jhammel
    and...CI... :(
  • jhammel
    a man can dream
  • » Ms2ger snorts
  • whimboo
    jhammel: CI?
  • jhammel
    whimboo: continuous integration
  • whimboo
    ah
  • whimboo
    you mean another system for tests
  • Ms2ger
    Confidence interval
  • ahal
    mdas: ok, I'm building again. Hopefully I can at least make some progress today
  • AutomatedTester
    for Mozmill?
  • jhammel
    whimboo: well, having the mutt tests run somewhere, yes
  • whimboo
    jhammel: probably we could reuse mozmill-ci for it
  • jhammel
    whimboo: wfm, as long as it yells loud enough when something breaks
  • whimboo
    jhammel: sure. you will get an email
  • whimboo
    :)
  • mdas
    ahal: yay!
  • mdas
    that's a good sign
  • mdas
    also, this cup is much better than my last attempt
  • jhammel
    aki: you in SF?
  • ahal
    mdas: so maybe cream is the answer
  • aki
    jhammel: i am
  • jhammel
    aki: i think jgriffin may be joining today, perhaps
  • aki
    ok
  • jhammel
    aki: same room?
  • jgriffin
    yep, I am!
  • mdas
    yes cream + 1%
  • aki
    jhammel: yeah. there's someone parked there, checking zimbra
  • jhammel
    mdas: 99% !
  • Ms2ger
    glob|away, oh, Referenced Bugs says [Bug 123456] (Secure bug) now, nice
  • bugbot
    Bug bugzilla.mozilla.org/123456 is not accessible.
  • aki
    kicked him out, we've got slim's
  • jhammel
    Ms2ger: that's where we talk about you behind your back
  • jhammel
    Ms2ger: as opposed to #ateam where we talk about you to your face
  • jhammel
    aki: cool, i'll be right over
  • Ms2ger
    Actually, I know the story behind bug 123456 ;)
  • jhammel
    Ms2ger: i was *also* reading #developers yesterday :P
  • Ms2ger
    Excellent ;)
  • aki
    jmaher: jgriffin can you guys hear me?
  • jgriffin
    aki: no can hear!
  • aki
    hm, i'm dialed into 354. maybe 9354
  • jgriffin
    yeah you need 9354
  • Callek
    aki: ooooo meeting
  • » Callek joings
  • Callek
    (sorry)
  • mcote
    jgriffin: I can query on starttime in the same way as date, right?
  • whimboo
    jhammel: i think it's somewhat related to persisted
  • jhammel
    jgriffin: your voiping badly
  • mcote
    e.g. { 'starttime': [1234567890, 1234567891] }, right?
  • jgriffin
    jhammel: sorry about that, let me plug in ethernet
  • jhammel
    jgriffin: better now :)
  • jgriffin
    mcote: yeah, I believe that should work
  • jgriffin
    but I have not tried it
  • mcote
    hm didn't seem to
  • whimboo
    jhammel: so it's controller.restartApplication() which is broken
  • jgriffin
    hmm, it may be that ranges don't work with strings, and that isn't being analyzed as a number
  • mcote
    hm well the test run record indicates that it's a number
  • jgriffin
    yes you're right, it is
  • jgriffin
    let me see if mozauteslib is doing something strange
  • mcote
    could be that I'm doing something strange too :)
  • ted
    dividehex: god, i don't know
  • jgriffin
    hmm, vidyo just died on me
  • jgriffin
    mcote: so it doesn't look like mozautoeslib is doing anything weird
  • samliu
    mcote: I'm in
  • samliu
    :D
  • dividehex
    ted: no worries, its in relops queue for now but I assigned it to you
  • mcote
    samliu: ah cool, okay, hm, now to decide which phones to leave connected :)
  • ted
    cool
  • samliu
    hm one is on a browser
  • samliu
    the other is black
  • samliu
    (nexus one and gs2)
  • samliu
    are those the ones acting up?
  • mcote
    the gs2 *was* working well up until last week some time
  • mcote
    okay so the best one appears to be the nexus s with address 78:d6:f0:cf:8d:11
  • mcote
    so we'll keep that one
  • samliu
    kk
  • mcote
    maybe just unplug the droid and the nexus one for now
  • samliu
    aiite
  • samliu
    done
  • mcote
    k let's see what adb says...
  • mcote
    huh so 3 are connected right now?
  • samliu
    mcote: yeah
  • mcote
    oof
  • mcote
    I see only two
  • samliu
    aw
  • samliu
    which two?
  • samliu
    (last 4 mac address numbers are fine)
  • mcote
    wow the nexus one and one nexus s
  • mcote
    but you unplugged the nexus one right?
  • samliu
    ohhhh I unplugged the nexus s
  • samliu
    x.x
  • samliu
    oops
  • mdas
    jgriffin: ping
  • samliu
    one sec
  • mcote
    heh
  • jgriffin
    mdas: pong
  • samliu
    haha I dont know phones x.x
  • samliu
    ah yeah okay I plugged the other nexus s back in
  • samliu
    so now it's 2 nexus s and the gs2
  • samliu
    the gs2 is on a browser screen
  • mcote
    so keep them all plugged in except the ones with MAC addresses HT9CPP809750 and 0A3AA1FD1402E01D :)
  • mdas
    jgriffin: have a marionette test that measures the startup times of each b2g app. I was going to ask the gaia team if I can put it in their repo
  • mcote
    okay can you restart the g2?
  • samliu
    haha thanks
  • mcote
    gs2
  • samliu
    k
  • mdas
    and then set up the CI to run it
  • jgriffin
    mdas: sounds great!
  • mcote
    yeah it sees the two nexus s but not the gs2
  • mdas
    if they're like 'no!' for some reason, then i'll just push it to testing/marionette/tests/perf
  • jgriffin
    haha, yeah
  • jgriffin
    is this a mocha test?
  • mdas
    jgriffin: nope
  • mdas
    it needs marionette stuff
  • jgriffin
    cool
  • mdas
    i can turn it into a mocha test once we have support for it
  • samliu
    gs2 is plugged in
  • samliu
    says "not registered with testserver"
  • mcote
    yeah that's okay; autophone isn't running
  • mcote
    okay I see 3 phones, good good
  • mcote
    okay let's try this yet again
  • jhammel
    mdas: mocha test? you live in a coffee world :)
  • mdas
    it's a delicious, delicious living
  • jgriffin
    lol
  • mcote
    samliu: okay hopefully the three are rebooting
  • mcote
    hm I guess shut the other two off for now
  • mdas
    hah!
  • mdas
    i just git clinted something
  • » mdas gives up on life
  • jgriffin
    git clint? lol
  • mdas
    yes.
  • mdas
    i meant to do git clone
  • mdas
    instead i just did git clint <enter>
  • jhammel
    mdas: itym git rdone
  • jgriffin
    heh
  • jhammel
    mdas: don't disturb clint on his vacation
  • mdas
    heh
  • mcote
    samliu: okay thanks, I think we're good for now. one was disabled immediately, but whaddyagonna do
  • mcote
    hopefully the other two keep running
  • mcote
    or at least one
  • mcote
    sigh
  • samliu
    aiite
  • samliu
    lol
  • samliu
    was it the gs2
  • samliu
    ?
  • mcote
    no, one of the nexus s's
  • samliu
    weird
  • samliu
    the gs2 is on a browser screen
  • mcote
    it's got sd card problems or something
  • mcote
    yeah that's good
  • mcote
    hopefully the other nexus s too?
  • samliu
    yep
  • mcote
    awesome
  • samliu
    should I unplug the nexus s that stopped / isnt on a browser?
  • mcote
    meh let's leave it for now
  • samliu
    aiite
  • mcote
    I'm more curious if 5 phones was just getting to be too much for the mini
  • samliu
    haha yeah or the cheap usb hub that has 6 things attached...
  • mcote
    it gave up on the one nexus s, so it shouldn't be trying to use it, and hopefully just having 3 phones attached will work
  • jhammel
  • mcote
    we'll see
  • samliu
    wow these phones are being powered from it too
  • samliu
    it could totally be a power issue
  • mcote
    yup
  • mcote
    yeah we already knew that having 6+ phones plugged in was a problem
  • mcote
    5 wasn't before, though
  • samliu
    x.x
  • mcote
    adb just seems to get confused easily
  • samliu
    yeah...
  • samliu
    is there an alternative?
  • mcote
    I need to change autophone to use the sut agent where possible
  • mcote
    yeah the sut agent
  • Callek
    aki: hrm could you hear me allright when I was talking, or was I extra quiet. Just went to unplug my headset and noticed that it wasn't plugged in -- I MIGHT have done so right after the call, but just want to be sure
  • Callek
    :-)
  • mcote
    which is pretty stable
  • samliu
    aiite
  • samliu
    well, ima leave haxxor
  • mcote
    cool cool
  • mcote
    thanks again
  • samliu
    :D np
  • aki
    Callek: i could hear you
  • samliu
    anytime
  • samliu
    hehehe
  • mcote
    good to hear you say that ;)
  • Callek
    ok, then it probably was plugged in
  • Callek
    (my built-in mic is horrid, I have found)
  • samliu
    doing missions in here makes me feel like batman
  • mcote
    ha good good :)
  • mcote
    k gonna walk my dog & let this run for a while
  • jmaher
    jhammel: ping
  • jhammel
    pong
  • jmaher
    jhammel: I will get to your review tonight
  • jmaher
    I was about 75% done last night, then you uploaded new versions
  • jhammel
    heh; well the new versions don't change much
  • jmaher
    yeah, just need to start over
  • jhammel
    except that test i removed....stupid me for running tests
  • jmaher
    heh
  • jhammel
    its too bad you can't post a patch update
  • jmaher
    just wanted to give you a heads up on that
  • jmaher
    you can post a patch update, just create a new patch
  • jhammel
    jmaher: yes, but then you need to start your review from scratch ;)
  • jhammel
    (sorta)
  • jmaher
    only on the updates
  • jhammel
    jmaher: still running on try, fwiw
  • jmaher
    err secondary patch
  • jhammel
    green so far
  • jmaher
    ok, I will go through it later tonight
  • jhammel
    coolz
  • jhammel
    thanks
  • jmaher
    jhammel: if you have a few minutes can you look at: people.mozilla.org/~jmaher/sxs/tp5n/sxs.html
  • jhammel
    lemme know if you have any questions
  • jhammel
    jmaher: yikes!
  • jhammel
    jmaher: what is going on there/
  • jmaher
    it has been that way for a long time
  • jmaher
    the svg stuff I think we should leave the original way
  • jmaher
    but the rest is better
  • jhammel
    hmmm, i just don't recall it being so bad :(
  • jhammel
    yeah, the svg specifically
  • jmaher
    tscroll has some hiccups though
  • jmaher
    but I am not too concerned about tscroll
  • jmaher
    anyway, if you have thoughts it would be nice to get second opinions
  • jmaher
    I would like to turn the side by side stuff off this week if possible
  • jhammel
    jmaher: iirc we only run tscroll on android these days
  • jmaher
    jhammel: naw, it runs on all desktop as well
  • jhammel
    jmaher: yeah, other than tsvg it looks good
  • jmaher
    under chrome
  • jhammel
    jmaher: ah, it is hidden
  • jmaher
    tomorrow Iwill file bugs to convert these
  • jhammel
    Y U NO CONVURGE EKSPEE?!?
  • jmaher
    it sort of sucks that we can't have tsvg run the same as the others
  • jhammel
    yeah....i wonder what's going on there
  • jmaher
    I suspect once we go to page centric views it will change, but we are really not there yet
  • jhammel
    its not just worse...it is a lot worse
  • jhammel
    jmaher: we're outputting to datazilla though....i wonder if we can/should mine that data
  • jmaher
    anyway, I need to afk
  • jmaher
    jhammel: we could mine the data
  • jhammel
    jmaher: cool, talk to you tomorrow
  • jmaher
    have a good night jhammel
  • jhammel
    jmaher: i'll ticket talos+versioning+talos.json+mozharness will wor
  • jhammel
    you too :)
  • jhammel
    and enjoy your pi :P
  • jhammel
    i made all the digits of pi for jmaher yesterday k0s.org/pi.html
  • jhammel
    need some css animations to scroll it or what not
  • » AutomatedTester lols at jmaher|afk's blog post title
  • AutomatedTester
    wow... marionette(its probably Firefox) is very unforgiving for syntax errors
  • mdas
    AutomatedTester: for execute_script stuff?
  • mdas
    t
  • mdas
    that would be firefox heh
  • AutomatedTester
    mdas: no I forgot to add + to concat some strings in a log message and Firefox launches and just hangs
  • mdas
    ah
  • AutomatedTester
    totally my fault but still...
  • jhammel
    wow ;)
  • jhammel
    AutomatedTester: sadly, he didn't include a lolcat with his post
  • » AutomatedTester hasnt read it
  • AutomatedTester
    I was checking twitter while a test ran
  • » AutomatedTester stars things to read them later
  • jhammel
    stars things?
  • AutomatedTester
    Favourite
  • AutomatedTester
    what ever it is called
  • jhammel
    ah....bookmar?
  • jhammel
    k
  • » AutomatedTester uses ifttt so when I favourite a post in twitter it then pushes the first link into Pocket ( was known as ReadItLater)
  • jhammel
    i have too many bookmarks for that...id never find anything
  • jhammel
    ah
  • jhammel
    i need a queue of stuff or i just lose it
  • AutomatedTester
    jhammel: thats what pocket does for me
  • » AutomatedTester just reads from top to bottom in that
  • jhammel
    my goal for the next decade is to get down to a single queue that isn't my inbox
  • jhammel
    i also want to sync my bookmarks with some service, preferably toolbox
  • jhammel
    i also want to pay off my student loans ;)
  • » face weeps
  • jhammel
    face: my student loans make me weep too
  • » face changes nick to seleniumbot
  • AutomatedTester
    face: its selbot... get it right *sheesh&
  • » face wonders if ezri is a ds9 reference
  • jhammel
    god i hope not...that was a mostly awful show
  • AutomatedTester
    face: what other DS9 factoids do you have?
  • » face wonders if jeads has his clippy script on ;)
  • AutomatedTester
    interesting...
  • AutomatedTester
    face: anything else?
  • » face patpats jmaher
  • jhammel
    yes, he's had a rough day
  • AutomatedTester
    jhammel: do we need rough in that sentence ^
  • jhammel
    AutomatedTester: can probably get it down to "he's...day"
  • jhammel
    i think it depends on which side of the pond you're on
  • jhammel
    face: which side of the pond are you on?
  • » face will work on converting wlach to irssi on a remote server
  • ted
    good job face
  • whimboo
    jhammel: this really fails since your implementation of restartApplication
  • jhammel
    i'm sorry
  • whimboo
    no worries
  • whimboo
    jhammel: mind at least having a quick look at the test?
  • jhammel
    it worked once upon a time...or at least appeared to work
  • jhammel
    sure, one second
  • whimboo
    i think we are trying to send something when the application is down
  • whimboo
    jhammel: i have attached a patch to the bug
  • whimboo
    really simplified
  • jhammel
    whimboo: thanks
  • whimboo
    jhammel: darn!
  • whimboo
    it's related to the manifest
  • AutomatedTester
    lol, my son has a whistly snore atm
  • jhammel
    whimboo: orly?
  • whimboo
    jhammel: give me some minutes
  • whimboo
    i think i know it
  • jhammel
    whimboo: sure, hopefully i'll be done filing this bug by then
  • » jhammel needs brain -> bugzilla
  • whimboo
    jhammel: i was wrong. running it with -t against the folder sadly runs test_post before test_pre
  • whimboo
    so it's not related
  • jhammel
    ah, yes...since its just (usually) alphabetical order
  • whimboo
    no
  • jhammel
    hmmm?
  • whimboo
    oh
  • whimboo
    sorry, heh
  • whimboo
    yeah. post comes before pre :)
  • jhammel
    the order isn't even guaranteed, iirc
  • jhammel
    maybe we alphabetize it, but os.listdir can certainly be in random order
  • jhammel
    well, probably hash order or what not
  • whimboo
    especially on linux
  • whimboo
    jhammel: you know what? with mozmill --restart it works
  • whimboo
    jhammel: happens when we call 'frame.runTestFile(path, False, name)'
  • jhammel
    whimboo: lemme look...just got done with the bugs i needed to write
  • jhammel
    well, the critical ones anyway
  • whimboo
    sure. i dive iin more
  • xabolcs
    whimboo: can i have one more question after your diving? it is about coding style ... (won't like to pause your debugging)
  • whimboo
    xabolcs: just go a head
  • whimboo
    jhammel: haha
  • xabolcs
    whimboo: just noted that i introduced some new functions which brake consistent style of the affected file
  • xabolcs
    whimboo: in ntt the function's opening "{" is on new line
  • xabolcs
    whimboo: but in my actual pulls i put in on the end of the function defining line
  • xabolcs
    whimboo: and also they doesn't have a space between "function" and "("
  • mdas
    jgriffin: cool, my perf test stuff got landed in gaia. I'm going to add a new job to the builder1 machine so it'll run the gaia perf tests once a day
  • whimboo
    xabolcs: we will have to change the old code to be consistent
  • mdas
    perhaps around 12pm, after the build work
  • jgriffin
    mdas: nice
  • xabolcs
    whimboo: ok for me. then it would be a broad change :), like the move to MPL2
  • xabolcs
    whimboo: will contain some indentation change too, like you noted in mozilla/nightlytt #74/files#r966106
  • whimboo
    jhammel: so yeah, the frame object gets destroyed after a restart
  • whimboo
    jhammel: or better is referenced to an object which doesn't exist anymore
  • whimboo
    when Imove the instantiation into run_test_file it works
  • xabolcs
    whimboo: i will file an issue about this coding style consistency
  • whimboo
    k
  • jhammel
    whimboo: is this for user restarts or non-user restarts or both?
  • whimboo
    jhammel: no user restart
  • whimboo
    therefore we have another method
  • jhammel
    so something is being set screwy....or not being set
  • whimboo
  • bugbot
    Bug 764442: critical, --, ---, hskupin, ASSIGNED , Constant 'Disconnect Error: Application unexpectedly closed' errors by using controller.restartApplication()
  • whimboo
    jhammel: ^
  • whimboo
    that should be the reason
  • whimboo
    the last frame object should be at 0x104213c10
  • whimboo
    but not 0x104213e50
  • whimboo
    and i probably know what that happens
  • whimboo
    jhammel: got it fixed!!! ]o/
  • whimboo
    jhammel: what about having the frame object as property in the mozmill class?
  • whimboo
    and not passing it around as parameter
  • whimboo
    jhammel: i will update the bug
  • jhammel
    whimboo: that is probably a good idea if we make sure to unset it in the right place
  • jhammel
    e.g. stop_runner
  • jhammel
    and maybe also in the shutdown event listener
  • whimboo
    jhammel: sure. mind commenting?
  • jhammel
    whimboo: sure, lemme make sure i understand what's going on here
  • whimboo
    the last comment explains it
  • whimboo
    jhammel: pastebin.mozilla.org/1662043 simplest fix
  • mcote
    jhammel: DS9 wasn't terrible!
  • mcote
    once you get into it, at least
  • whimboo
    but again I would favor the class member
  • whimboo
    jhammel: so probably you will get my largish jsbridge patch today :)
  • jhammel
    mcote--
  • jhammel
    mcote: oh i tried ;)
  • mcote
    well Sisko is terrible, don't get me wrong
  • whimboo
    i have to second jhammel. it wasn't my favorite too
  • mcote
    can't believe he teaches acting
  • mcote
    I wasn't into it for the first few seasons, but I kept trucking, and I was sad when it ended
  • jhammel
    mcote: he was actually one of the few people i could tolerate on the show ;)
  • whimboo
    jhammel: i would propose for this bug that we use my simple fix and I file a new one to move frame to a class member?
  • mcote
    it's all about Quark and Worf
  • mcote
    jgriffin: ping
  • jgriffin
    mcote: pong
  • jhammel
    whimboo: i'm not even entirely sure why your fix works ;)
  • mcote
    jgriffin: I am so happy you still answer my pings :)
  • jhammel
    whimboo: but sure, sounds good
  • whimboo
    jhammel: we can hop on skype or vidyo and i can explain
  • mcote
    jgriffin: so this returns nothing: eslib.query({'starttime': [1336437030, 1339029030]})
  • jgriffin
    mcote: haha
  • jhammel
    whimboo: sure lemme find a room
  • whimboo
    jhammel: skype?
  • jhammel
    whimboo: what makes me nervous about an instance variable is that if we don't unset it properly we'll be in a bad state
  • jgriffin
    mcote: ugh, let me try
  • jhammel
    whimboo: yeah, one sec
  • mcote
    weirdly, aggregates work
  • mcote
    e.g. this returns a reasonable count:
  • mcote
    eslib.aggregates({'starttime': [1338969030, 1339486343]}, aggregate_by={'buildtype': [u'opt'], 'testgroup': [u'mochitest-other'], 'os': ['fedora'], 'tree': [u'"mozilla-central"']})
  • jgriffin
    that is very weird
  • mcote
    different timestamps, I realize, but aggregating over the starttimes in my query statement above also works
  • jhammel
    whimboo: beh, skype got borked with my ubuntu upgrade; vidyo?
  • jgriffin
    hmm, I have an idea what might be going on
  • whimboo
    sure
  • whimboo
    jhammel: we can use my room
  • mcote
    jgriffin: cool. I decided that it's more accurate to look at starttime instead of date
  • mcote
    frankly I'm not even sure how 'date' is calculated
  • jgriffin
    date is just the date of the starttime, PST
  • jhammel
    whimboo: connected
  • mcote
    oh it is, okay
  • mcote
    thought I saw one that didn't match, but maybe I was looking at the wrong thing
  • whimboo
    jhammel: one sec
  • mcote
    s/maybe/probably/
  • whimboo
    jhammel: have i killed you?
  • jhammel
    beh, disconnected
  • jhammel
    yeah
  • whimboo
    ups
  • jhammel
    reconnecting ;)
  • whimboo
  • AutomatedTester
    laters
  • mcote
    toodles
  • jgriffin
    mcote: this is fallout from bug 733418
  • bugbot
    Bug bugzilla.mozilla.org/733418 normal, --, ---, nobody, NEW , ES not returning consistent results for some queries
  • whimboo
    jhammel: i hate vidyo
  • jhammel
    whimboo: heh, yeah need to reinstall skype
  • whimboo
    i never get those lags with skype
  • mcote
    oh
  • mcote
    blah
  • jhammel
    for some reason just about every ubuntu upgrade kills it
  • mcote
    wah
  • jgriffin
    yeah :(
  • mcote
    okay not a huge huge deal
  • mcote
    I'll just query by date, find the one with the timestamp I want, and count back from there
  • mcote
    I would also like for us all to never use the comment "will be easy to do" ever again (bugzilla.mozilla.org/669316#c1) ;)
  • bugbot
    Bug 669316: normal, --, ---, mcote, ASSIGNED , Add tool for estimating when an orange was introduced
  • jhammel
    mcote++
  • jhammel
    mcote: how about "it will be a cakewalk"?
  • mcote
    walking on cake is probably pretty hard
  • mcote
    at least if you want it to look like a cake afterward
  • whimboo
    jhammel: also this is one reason why i don't like that mutt restarts all the time in between tests :(
  • mcote
    and be edible
  • jhammel
    mcote: though i might have to add on a rider that we should avoid "We'll hack this for now, but we'll eventually fix it" ;)
  • mcote
    well yes, that's been documented many many times :)
  • mcote
    the issue here is that easy programming tasks do exist, but they are never in the automation domain
  • jhammel
    we should at least be honest and just "we'll hack this for now" and be done with it ;)
  • jhammel
    mcote: heh, nah, we have several
  • mcote
    oh? haven't met 'em
  • jhammel
  • bugbot
    Bug 755930: normal, --, ---, nobody, NEW , No option for cycles; there should be
  • jhammel
    i know because i just added it for a different bug ;)
  • jhammel
    since i got tired of ts tests taking 20 minutes
  • mcote
    okay there are easy bugs, fine, just no easy *projects* :)
  • jhammel
    ah, well that's different :)
  • mcote
    haha "The Art of Tooting your own Horn without Blowing It"
  • jhammel
    heh
  • mcote
    "I'd like to step up and toot Mr. Birdman's horn"
  • whimboo
    jhammel: patch is up on bug 764442
  • bugbot
    Bug bugzilla.mozilla.org/764442 critical, --, ---, hskupin, ASSIGNED , Application disconnect errors because of invalid frame objects if a test file restarts the browser
  • whimboo
    including a test
  • mcote
    man hard to believe I used to do all this stuff in C++
  • mcote
    talk about making difficult tasks more difficult
  • jhammel
    mcote: heh, i feel the same way
  • jhammel
    i remember writing parsers in C++
  • mcote
    heh yeah that was some time ago for me
  • jhammel
    and, sadly, feeling how much easier it was than writing them in C
  • mcote
    university
  • whimboo
    jhammel: here a quick preview for you: github.com/whimboo/mozmill/compare/jsbridge :)
  • jhammel
    whimboo: thanks;
  • jhammel
  • jhammel
    whimboo: i guess you don't feel the extension should apologize :P
  • whimboo
    heh
  • whimboo
    thanks
  • jhammel
    whimboo: r+ :)
  • jhammel
    whimboo: thanks for looking into this
  • jhammel
    beh, and now i have other fires :(
  • whimboo
    jhammel: where is the above line from the pastebin?
  • whimboo
    i can't find it
  • whimboo
    oh wait
  • whimboo
    hah now I understand
  • jhammel
  • whimboo
    yeah. the dev is to blame here!
  • jhammel
    heh
  • whimboo
    i can change it
  • jhammel
    whimboo: i actually don't care one way or the other ;)
  • whimboo
    "You fool missed to set the jsbridge port preference"
  • jhammel
    "Where de hell your jsbridge port at?"
  • jhammel
    mcote: as long as we're on automation aphorisms, "Work is what happens in between putting out fires"
  • mcote
    oh so true, so unfortunately true
  • whimboo
    jhammel: mind marking the patch as r+
  • jhammel
    whimboo: i did?
  • mcote
    I have a feeling that no matter how many comments I put into this code or how much I refactor it, it is going to be total nonsense to me in 6 months
  • jhammel
    at least that is what bugzilla tells me
  • whimboo
    jhammel: no
  • jhammel
    whimboo: bugzilla.mozilla.org/764442 ? refresh?
  • bugbot
    Bug 764442: critical, --, ---, hskupin, ASSIGNED , Application disconnect errors because of invalid frame objects if a test file restarts the browser
  • whimboo
    hah... lags
  • whimboo
    now I see it
  • jhammel
    :)
  • whimboo
    might be the rain here
  • jhammel
    whimboo: the bits are slow to get all the way to Deutschland
  • mcote
    schnell schnell!
  • whimboo
    everyone is partying here
  • whimboo
    we have won the football match
  • jhammel
    awesome :)
  • jhammel
    vs nederlands?
  • whimboo
    yeah
  • mcote
    heh lookit jhammel's dutch
  • whimboo
    lol
  • mcote
    i'm actually watching a recording of saturday's roller derby match here
  • jhammel
    well i do like Amsterdam....haven't been too any other Dutch city
  • mcote
    Rotterdam is boring :)
  • mcote
    looks like a North American city
  • jhammel
    though i also liked Frankfurt....almost as far from Dresden as possible in Germany though
  • mcote
    ew really?
  • jhammel
    Frankfurt?
  • mcote
    I found Frankfurt pretty... uninspiring
  • jhammel
    heh
  • mcote
    the towns nearby were nice
  • mcote
    we went to check out the Romanesque cathedrals in Mainz, Speyer, and Worms
  • jhammel
    i meant to go to Hiedelberg, but woke up way too late that day :(
  • jhammel
    need to visit the Berlin office
  • mcote
    indeed
  • mcote
    Berlin is great
  • mcote
    opinions of German nonBerliners notwithstanding :)
  • jhammel
    i've had several people tell me i should move to Berlin...which feels strange since i've never been there
  • jhammel
    maybe they just wanted me far away ;)
  • mcote
    i know two people who moved to Berlin
  • mcote
    and are still there
  • jhammel
    or so they would have you think!
  • jhammel
    okay, let's run these mozmill tests again :)
  • whimboo
    jhammel: they should work now
  • mcote
    jgriffin: cool I got something working by querying for dates but searching for starttimes
  • jgriffin
    ha, nice
  • mcote
    so for bug 761488, using opt/linux/mochitest-other/mozilla-inbound, I calculated the orange seed as being test run starting 2012-06-03 05:36:22 PT, revision 04682cf84deb
  • bugbot
    Bug bugzilla.mozilla.org/761488 normal, --, ---, general, NEW , Intermittent crash in browser_pdfjs_main.js [@ js::detail::HashTable]
  • mcote
    hm that seems a bit funny though
  • mcote
    you'd think it would be the 5th
  • mcote
    jgriffin: okay here's an example. the tbpl record for the first opt occurrence of that bug says date 2012-06-06
  • mcote
    but the starttime is 1338957539
  • mcote
    which, if I am correct, translates to 2012-06-05 21:38:59-07:00
  • mcote
    oh
  • jgriffin
    it translates into 2012-06-06
  • mcote
    maybe starttime is in PT
  • jgriffin
    yes, it is
  • mcote
    yeah sorry figured it was UTC
  • mcote
    silly me
  • jgriffin
    you would think
  • jgriffin
    but no
  • mcote
    thinking we'd use a sane timezone
  • mcote
    ok
  • jhammel
    sane? us?
  • mcote
    anyway hm that still doesn't change my prediction
  • mcote
    hm
  • mcote
    jgriffin: hm, what's the difference between 'starttime' and 'timestamp'?
  • jgriffin
    starttime comes from buildbot
  • jgriffin
    timestamp is generated by us when we parse the log
  • mcote
    okay so is date based on timestamp or starttime?
  • mcote
    or lemme put it this way, seems that it's based on timestamp :)
  • jgriffin
    you could be right, I'd have to look at the code
  • mcote
    not sure if that affects anything though
  • jgriffin
  • jgriffin
    it comes from starttime
  • mcote
    er
  • mcote
    but I see an orange like this:
  • mcote
  • mcote
    date 2012-06-08
  • mcote
    wait ugh
  • mcote
    nm
  • jhammel
    are the timestamps real (utc) timestamps?
  • jhammel
    one would hope
  • jhammel
    one of the reasons i wish timezones would just go away
  • mcote
    I want datetime.datetime.utcfromtimestamp()
  • mcote
    because .fromtimestamp() translates into my local time
  • mcote
    guh
  • jhammel
    yes, yes it does
  • jhammel
    although python OOTB has very little timezone support
  • mcote
    why on earth did I think I was getting away from date/time math when I left that calendaring/meeting-scheduling start-up...
  • jhammel
    which friggin confuses me constantly
  • jhammel
    mcote: no programmer can ever get away from date/time math :(
  • mcote
    I am going to call myself a temporal mathematician
  • mcote
    at least it will sound cool
  • jhammel
    and i for one doubt i will ever get it correct
  • mcote
    even if the work is awful
  • jhammel
    mcote: i prefer the title 'temporal dynamicist' ;)
  • jgriffin
    mcote: I was totally wrong about timestamp
  • mcote
    heh
  • jgriffin
    it's actually the time the orange was starrred in TBPL, as reported there
  • mcote
    ahh
  • mcote
    okay unimportant to me
  • mcote
    hm so I guess my prediction is right according to the math in the comment
  • mcote
    which I don't understand heh
  • mcote
    he calculates that the orange happens 5 times in 128 pushes
  • whimboo
    jhammel: something is wrong on windows
  • jhammel
    whimboo: what is that?
  • mcote
    so that's about once every 26 pushes
  • whimboo
    jhammel: ERROR | ProcessManager UNABLE to use job objects to manage child processes
  • whimboo
    it's an error
  • whimboo
    and we get pass/skip/fail == 0
  • mcote
    but he concludes that there is a 95% chance that the "offending changeset" occurred 75 pushes before
  • mcote
    that seems... weird.
  • mcote
    oh
  • jgriffin
    yeah and btw I think 95% is pretty far out
  • jhammel
  • whimboo
    jhammel: and it takes a really long time to start firefox
  • jhammel
    whimboo: this should probably be WARNING
  • jgriffin
    we probably want to display like 80% to 95% or something
  • mcote
    jgriffin: ah, I think he means that there's a 95% chance that it happened between the original occurrence and 75 pushes before that
  • jgriffin
    right
  • jhammel
    whimboo: i'm also still getting weird hangs on linux
  • mcote
    jgriffin: I guess that makes sense
  • jhammel
    (nor do all of my tests pass :(
  • whimboo
    jhammel: python or js tests?
  • jhammel
    whimboo: just js for now
  • jhammel
    though iirc there were python errors too
  • whimboo
    where does it hang?
  • jgriffin
    but most of the time, it will occur close in from that 95% window
  • jgriffin
    that's kind of an extreme case
  • mcote
    yeah well you can adjust the probability I guess
  • jhammel
    whimboo: so i'm running like this: mozmill -b `which firefox` -m mutt/mutt/tests/js/tests.ini
  • mcote
    if we wanted less certainty but a smaller range
  • jhammel
    whimboo: that should be fine, right?
  • jhammel
    i wish mozmill had a pause button :(
  • mcote
    e.g. with 75% probability, it gets reduced to 58 pushes
  • whimboo
    jhammel: sounds fine
  • mcote
    er 90% I mean
  • jhammel
    whimboo: so it will hang on browser restarts
  • whimboo
    where does it hang for example?
  • whimboo
    which ones
  • mcote
    75% is down to 34 pushes
  • mcote
    interesting
  • jgriffin
    mcote: right
  • jhammel
    lemme see if i can find where...its hanging now...
  • mcote
    okay yeah, I think his calculations are going in the wrong direction
  • mcote
    I don't think we care about 95%, 99%, and 99.9%
  • jgriffin
    right, we don't
  • mcote
    rather, it's probably more interesting to see, say, 50%, 75%, and 95%
  • jgriffin
    yep
  • mcote
    hokay
  • jhammel
  • jhammel
    that is one case
  • mcote
    this is coalescing now
  • jgriffin
    woot!
  • mcote
    maybe I can actually get this done this week :)
  • jgriffin
    and bonus points for using the world coalescing
  • mcote
    need a glass of wine first though
  • jhammel
    whimboo: it actually passes the test, fwiw
  • jhammel
    grrr
  • jgriffin
    way ahead of you!
  • mcote
    hehe
  • whimboo
    jhammel: i dont have it here
  • jhammel
    whimboo: i think mozmill hates this computer :(
  • jhammel
    ever since last fall sometime
  • whimboo
    jhammel: or wait
  • whimboo
    you are right
  • jhammel
  • whimboo
    same here
  • jhammel
    whimboo: okay, well at least it is consistent
  • whimboo
    please file a bug for
  • jhammel
    will do
  • jhammel
    i wonder if this is just expected...though i recall it not doing this at some point :(
  • jhammel
    :sigh: coming back to software after 6+ months is never fun
  • whimboo
    jhammel: i do my best to make it better
  • whimboo
    jhammel: keep in mind that we did no QA before earlier this quarter
  • jhammel
    yeah
  • jhammel
    its just painful trying to remember what worked before, what didn't, and what i was intending when i wrote the code
  • whimboo
    jhammel: what would you say if we enalble jsbridge debugging if you have set --console-level=DEBUG?
  • whimboo
    we can do it with a preference too
  • jhammel
    whimboo: sounds good to me
  • jhammel
    whimboo: there is actually the --debug flag
  • whimboo
    well. even better, right
  • jhammel
    :) exactly
  • xabolcs
    whimboo: one question again? :)
  • whimboo
    jhammel: works like a charm! :)
  • whimboo
    xabolcs: sure
  • whimboo
    jhammel: do you remember why we set strict js optiopns? we really shouldnt do that
  • xabolcs
    whimboo: where could i talk about my refactoring thoughts of ntt? would like to "slice" it along features...
  • whimboo
    jhammel: nm. we only have it in debug mode
  • whimboo
    xabolcs: which type of refactoring?
  • xabolcs
    whimboo: global namespace pollution, automatic filling of features on about:nightly, moving features to code modules
  • whimboo
    hm, what's the second? the others sound great!
  • jhammel
    xabolcs: automatic?
  • xabolcs
    whimboo: "self" xul overlaying - to achive unified overlays
  • whimboo
    jhammel: the patch is not ready yet, i would have to create some tests. but would you mind to give a review already?
  • xabolcs
    the automatic fill? :)
  • jhammel
    whimboo: the jsbridge refactor?
  • whimboo
    jhammel: yes, all actual tests pass
  • xabolcs
    as you could see in mozilla/nightlytt #58 the contributors are filled automatic i.imgur.com/vNE96.png
  • whimboo
    jhammel: or wait. I don't need new tests. my disconnect test from today is enough. i will just make sure we also create a new profile
  • xabolcs
    so i thought if there would be nightly.features.pushlog, nightly.features.screenshot, nightly.features.titlebar, ... they could be tested against and included in about:nightly
  • xabolcs
    based on what features is implemented / enabled on running application
  • xabolcs
    for example titlebar customization was introduced for tb in the near past, and totally "unimplemented" for sb and sm
  • xabolcs
    "self" xul overlaying - i mean generalizing xul overlaying, to minimalize the duplicated content
  • whimboo
    jhammel: patch is up and waiting for your review :)
  • whimboo
  • bugbot
    Bug bugzilla.mozilla.org/686320 enhancement, --, ---, hskupin, ASSIGNED , Allow Mozmill to probe the OS for a free port instead of hardcoded value in JSBridge
  • jhammel
    whimboo: coolz, i will look now
  • whimboo
    xabolcs: i woldn't spend too much time in about:nightly for now. IMO it's more important to get the bugs fixed and the new feature requests implemented
  • xabolcs
    whimboo: yeah, i know. i thought i happily implement these changes if somebody review my proposal(s) and mentor me. :)
  • whimboo
    xabolcs: the most important thiing would really be refactorng the modules
  • whimboo
    xabolcs: i know it's only me at the moemnt who is working with you on NTT but I will do my best for mentoring
  • jhammel
    i'm theoretically a maintainer but haven't done any work on it since November
  • » jhammel blames talos
  • jhammel
    face: and you
  • » face wishes people would use try
  • jhammel
    face: sure...just shift the blame
  • » face blames Abraham Lincoln
  • jhammel
    exactly!
  • » jhammel widhes people would use Abraham Lincoln
  • xabolcs
    whimboo: in the respect of the sum size of my proposes, i'd like to write a "specification" for it ...
  • xabolcs
    whimboo: ... would you mind to review some draft docs before reviewing some pulls?
  • whimboo
    absolutely, i would even propose that method. it's always better to have a clear view before starting the implementation