2011/06/09/#launchpad-yellow.txt

[13:26] <gary_poster> bac benji danilos: call RSN
[13:27] <gary_poster> bac benji, should I bother trying to run lp2kanban?
[13:27] <bac> no
[13:27] <danilos> gary_poster, ack
[13:27] <gary_poster> ok
[13:36] <danilos> gary_poster, btw, can I take gmb's slot for the call perhaps?
[13:36] <gary_poster> danilos, of course
[13:36] <danilos> gary_poster, cool, and when is that? :)
[13:36] <danilos> in 24mins?
[13:36] <gary_poster> :-) in 54 min
[13:36] <danilos> ah, ok, sounds good, thanks again
[13:37] <gary_poster> np
[13:39] <danilos> gary_poster, btw, I wonder if you remember this: how does one create a view with a logged in user (meaning self.user is set) for testing? all the stuff I tried ("with person_logged_in()", "LaunchpadTestRequest().setPrincipal()" and even "login(email)" have failed for me)
[13:42] <bac> danilos: a combo of login() and create_initialized_view(..., principal=) should work
[13:44] <gary_poster> that sounds good to me, danilos.
[13:45] <danilos> bac, thanks, I was hoping to use the view class instead of the view name, but I am fine with this (I don't want to reinvent create_view)
[13:50]  * gary_poster needs to take kids to school today.  will step away RSN, back about 10 or 15 min later
[14:14]  * gary_poster had to make funny faces at baby also
[14:16] <benji> heh
[14:17] <danilos> :)
[14:18] <danilos> still no luck getting self.user to be set in my view :/ not sure what am I doing wrong, but I can't see what am I doing different from some other view tests :/
=== Ursula is now known as Ursinha
[14:27] <gary_poster> danilos, can your force it somehow, by manipulating the request or the bag or something?
[14:28] <gary_poster> I know that's ugly, and it ought to work, but if you can't figure it out soon, maybe that's reaonable (or maybe doingthat would help you iunderstand what's going wrong?)
[14:28] <gary_poster> (it ought to work == what you are doing now ought to work)
[14:28] <danilos> gary_poster, I tried that as well, didn't get far, so I am probably missing something obvious
[14:28] <gary_poster> gotcha
[14:29] <gary_poster> dunno :-/  I could look at it with you if you wanted, but I don't have super high confidence that I'd see something you didn't
[14:30] <gary_poster> danilos, I'm ready when you are.  call me; no rush.
[14:32] <danilos> gary_poster, calling
[14:41] <danilos> bac, hi, I am still kind of stuck, do you think you could spare a few minutes to look and maybe figure out something obvious that I am doing wrong?
[14:41] <bac> danilos: sure, paste it?
[14:42] <danilos> bac, https://pastebin.canonical.com/48350/
[14:42] <danilos> bac, that's a test, and when it hits view.subscriber_data_js, self.user is not defined
[14:43] <gary_poster> benji, so you know, flacoste has given you a full 45 min for your coffeescript presentation,  No pressure. ;-)  https://wiki.canonical.com/Launchpad/Sprints/Thunderdome2011/Agenda#Call%20for%20presentations
[14:43] <danilos> bac, full branch at lp:~danilo/launchpad/bug-772754-other-subscribers-loading
[14:43] <gary_poster> I also said we might want to talk about lp2kanban
[14:43] <benji> heh
[14:43] <bac> danilos: did you try passing the principal to c_i_v?
[14:43] <danilos> bac, yeah, nothing changes
[14:46] <danilos> bac, well, not entirely nothing, self.request.principal is then defined, but self.user is still None
[14:48] <bac> gary_poster: how much of the thunderdome will be spent with the platform team?  will we go to their plenaries?  or will we just be in the same hotel?
[14:49] <gary_poster> bac, that's a great question.  My assumption was that we simply were sharing hotels, but that seems unlikely now that you've asked.  Care to ask flacoste and report back to us?
[14:50] <bac> i would if he were around...
[14:50] <bac> i will next time i see him
[14:51] <gary_poster> ok
[14:59] <danilos> bac, any ideas of what else I could try?
[14:59] <bac> danilos: i'm looking
[15:00] <gary_poster> is self.user one of those launchbag things?
[15:06] <danilos> gary_poster, I don't know, haven't really dug that much
[15:06] <gary_poster> k
[15:07] <danilos> and I just created a minimal view and test using exactly the same logic, and it has self.user set :(
[15:09] <bac> danilos: yeah, look at test_data_person_subscription_user_excluded ... it works and is pretty darned minimal
[15:10] <danilos> bac, that's a test I am creating, but it doesn't work for me
[15:10] <danilos> bac, does it work for you?
[15:11] <bac> danilos: sorry, paste error. i meant test_unsubscribe_link_appears_for_user
[15:16] <bac> danilos: i think the problem is the assert is outside of the 'with person_logged_in' context
[15:17] <bac> so at that point no one is logged in.  even though the view was created with a logged in users there isn't one when you evaluate the property
[15:17] <danilos> bac, yeah, so it seems to me as well, though I am pretty sure I've tried that as well
[15:17] <bac> your test still fails but for a different reason i think
[15:19] <danilos> bac, yeah, because I haven't implemented the action yet, thanks, that is it, with all the confusion I wasn't even sure what I tried and when
[15:19] <bac> np
[15:22] <danilos> bac, very, very much thankful! really appreciate the time you took to look into it
[15:23] <bac> danilos: sorry i didn't spot it right away.  those subtle things are hard
[15:37] <danilos> gary_poster, fwiw, my "full" branch is up at https://code.launchpad.net/~danilo/launchpad/bug-772754-other-subscribers-remove-cruft; still not ready
[15:37] <danilos> and I've got to run now, talk to you tomorrow all
[15:37] <gary_poster> danilos, cool.  I'll keep on with other bugs then
[15:37] <gary_poster> :-)
[15:37] <gary_poster> bye
[15:48]  * benji watches his laptop go down in flames as Unity segfaults.
[15:48] <gary_poster> :-/
[15:52] <bac> gary_poster: just so you'll know, i'm spending considerable time this morning trying to QA bug 789383
[15:52] <_mup_> Bug #789383: Oops calling findSimilarBugs api anonymously <api> <bugs> <oops> <qa-needstesting> <Launchpad itself:Fix Committed by bac> < https://launchpad.net/bugs/789383 >
[15:52] <gary_poster> bac, ok, thanks for heads up
[15:53] <bac> it works locally but on qastaging it looks like it is redirectly to +login and responding with 503s
[15:53] <bac> s/redirectly/redirecting
[15:53] <gary_poster> :-/
[16:02] <gary_poster> chr...
[16:06] <benji> ok, I have to reboot again
[16:07]  * benji hopes Unity feels better now.
=== Ursinha is now known as Ursinha-lunch
[18:43] <bac> gary_poster: i'm seeing bunches of timeouts on qastaging for bug queries, e.g. https://lp-oops.canonical.com/oops.py/?oopsid=OOPS-1986QASTAGING120
[18:44] <gary_poster> bac, could it be the critical bug summary thing lifeless talked about on the list?
[18:45] <bac> gary_poster: i've just read the incident report and it is unclear
[18:45] <gary_poster> bac, bug 794802
[18:45] <_mup_> Bug #794802: many bug activities timing out due to contention on bugsummary <canonical-losa-lp> <regression> <timeout> <Launchpad itself:Triaged> < https://launchpad.net/bugs/794802 >
[18:45]  * gary_poster looks an oops
[18:45] <gary_poster> at
[18:47] <gary_poster> bac, agreed that it is unclear, but that 9 second query in the OOPS you gave looks pretty innocuous.  contention makes a lot of sense.
[18:47] <bac> gary_poster: so what's one to do?
[18:48] <bac> gary_poster: qa-ok even though i can't prove it?  qa-bad even though i don't think it is my branch?
[18:48] <gary_poster> oh
[18:48] <bac> lie and mark it qa-untestable?
[18:48] <bac> qa-punt
[18:48] <gary_poster> heh
[18:51] <gary_poster> bac, if you think it will block release of stub's fix for that bug, I say mark it qa-ok and add a comment about the status (you believe it to be fine, and are clearing it so that the critical fix can be deployed).  Alternatively in that situation, wait and ask lifeless what he wants.  If you don't think it will block, then just comment about the status and leave the qa tag as is.
[18:51] <bac> gary_poster: by definition it is going to block unless i do a rollback, no?
[18:52] <gary_poster> bac, if it is on devel then yes
[18:52] <gary_poster> (db-devel no)
[18:52] <bac> gary_poster: i will mark it qa-ok and be ready for the consequences
[18:52] <gary_poster> ok
[18:52] <bac> yes, we're on devel
[18:52] <gary_poster> k
[19:02]  * bac returns to kanban
[19:06] <gary_poster> bac, I am writing a reusable test fixture to help me temporarily replace one view with another, within the component registry.  I'm doing this so I can override a particular method on the view, in this case so that I can force an exception to be raised.  An alternative approach would have been to make a monkey patch, but I was trying not to do that.  Am I reinventing a wheel--do we have a pattern for this that you 
[19:07] <bac> i've never seen such a thing
[19:07] <bac> gary_poster: sounds cool
[19:07] <gary_poster> ok cool thanks
[19:07] <bac> but you know sinzui has the encyclopedic recall of all things
[19:30] <gary_poster> bac, in five minutes OK?
[19:30] <bac> perfect
[19:30]  * bac tea
[19:31] <gary_poster> cool
[19:35]  * gary_poster calls without warning
[19:36] <gary_poster> bac, oops?
[19:37] <bac> gary_poster: dunno, skype was reporting my network speed is poor
[19:37] <bac> it lies
[19:37] <gary_poster> :-)
=== Ursinha-lunch is now known as Ursinha
[20:43] <benji> apparently unit any chromium don't mix; I'll be right back
[20:45]  * benji gives up on Unity and tries to figure out how to turn it off.
=== benji is now known as Guest88900
[20:49] <gary_poster> :-/
[20:49] <benji___> aaand we're back
[20:49] <gary_poster> this is not giving me inspiration to run and try installing Ubuntu on metal again. :-/
=== benji___ is now known as benji
[20:51] <benji> It's a shame too, because I kind of like Unity; I definately like the type-an-app-name launcher thing, not so much the list of icons on the right, I never use that
[20:51] <benji> I also like the merging of the title bar and menu.
[20:51] <gary_poster> yeah, I like app-name launchers too
[20:52]  * benji tries to remember what the heck he was doing.
[20:52] <gary_poster> getting ready for call in 8 min! :-)
[20:52] <benji> heh
[20:56]  * gary_poster goes for glass of water
[20:59] <bac> hi benji -- very quick question of urllib2 vs httplib2 for lp2kanban
[21:00] <bac> i don't see an easy way to turn on debugging for urllib2 like httplib2.debuglevel=1
[21:01] <benji> bac: I /think/ there is a way, but I don't think it's documented; I wouldn't object to using httplib2 though, urllib2 is evil incarnate
[21:01] <gary_poster> heh
[21:01] <bac> benji: ok.
[21:02]  * benji runs to get a Dr Pepper 
[21:25] <benji> gary_poster: http://blog.ssokolow.com/archives/2011/05/07/a-python-programmers-first-impression-of-coffeescript/
[21:26] <gary_poster> eldest = 24 > 21 ? "Liz" : "Ike";
[21:26] <benji> http://jashkenas.github.com/coffee-script/#operators
[21:28] <benji> gary_poster: zip = lottery.drawWinner?().address?.zipcode
[21:40] <benji> gary_poster: https://gist.github.com/870838
[21:54] <gary_poster> if (typeof elvis !== "undefined" && elvis !== null) {
[21:54] <gary_poster>   alert("I knew it!");
[21:54] <gary_poster> }
[21:55] <gary_poster> alert "I knew it!" if elvis?
[21:57] <gary_poster> kids =
[21:57] <gary_poster>   brother:
[21:57] <gary_poster>     name: "Max"
[21:57] <gary_poster>     age:  11
[21:57] <gary_poster>   sister:
[21:57] <gary_poster>     name: "Ida"
[21:57] <gary_poster>     age:  9
[22:00] <benji> futurists =
[22:00] <benji>   sculptor: "Umberto Boccioni"
[22:00] <benji>   painter:  "Vladimir Burliuk"
[22:00] <benji>   poet:
[22:00] <benji>     name:   "F.T. Marinetti"
[22:00] <benji>     address: [
[22:00] <benji>       "Via Roma 42R"
[22:00] <benji>       "Bellagio, Italy 22021"
[22:00] <benji>     ]
[22:07] <gary_poster>         fixture = ZopeViewReplacementFixture('+login', ILaunchpadApplication)
[22:07] <gary_poster>         class OpenIDLoginThatFailsDiscovery(fixture.original):
[22:07] <gary_poster>             def _getConsumer(self):
[22:07] <gary_poster>                 return OpenIDConsumerThatFailsDiscovery()
[22:07] <gary_poster>         fixture.replacement = OpenIDLoginThatFailsDiscovery
[22:07] <gary_poster>         self.useFixture(fixture)
[22:22] <gary_poster> http://cappuccino.org/learn/demos/
[22:22] <gary_poster> http://280slides.com/Editor/
[22:36] <gary_poster> benji, reply from leankitkanban:
[22:36] <gary_poster> We currently do not have any throttling setup but we are developing a policy and will be implementing some safeguards against abuse of our API.  We are currently working with a customer that, most likely inadvertently, was causing us a lot of issues the past two days with a huge number of expensive API calls per second.
[22:36] <gary_poster> As for your issue, we haven't heard about this problem from the customers that are using our API but that doesn't necessarily mean they didn't experience the problem.  If you like, you can let us know when you are testing something and we can monitor our logs to see if we see anything suspicious.
[22:36] <gary_poster> bac ^^
[22:37] <bac> benji, gary_poster: i have converted to use httplib2 and i am not getting errors from leankit any more
[22:37] <bac> i am not claiming that is causal
[22:37] <bac> but it is good
[22:38] <gary_poster> bac, cool either way though :-)
[22:38] <bac> i'm putting a little bow on it and about to push
[22:38] <gary_poster> awesome
[22:47] <bac> gary_poster: it may be overkill but was easy so i did it -- output from running the script:  https://pastebin.canonical.com/48377/
[22:47] <bac> and usage
[22:49] <bac> it shows we need to be more discriminating in the bugs we pull from LP, perhaps.
=== Ursinha is now known as Ursula
[23:09] <gary_poster> bac, heh
[23:09] <gary_poster> yeah
[23:09] <gary_poster> night
[23:09] <bac> bye
=== Ursula is now known as Ursinha