/srv/irclogs.ubuntu.com/2012/05/09/#launchpad-dev.txt

gary_posterlifeless, fun idea.  I added it to the bug00:15
gary_posternight all00:16
wallyworld__StevenK: i can get a asus with i7-2670QM and 16GB RAM and 1920x1080 15.6" display for $1300. sounds ok i think00:27
wgrantThat's huge.00:32
wgrantBut nice res.00:32
wallyworld__wgrant: huge? you mean screen size?00:33
wgrantYeah00:33
wallyworld__i wouldn't go smaller00:33
wallyworld__15.6" is the smallest i've had ever00:34
wgrantI guess if you're using it as a desktop without an external monitor it might make sense.00:34
wallyworld__i have an external monitor too00:34
wgrantOK, so it doesn't make sense :)00:34
wallyworld__i like to move about a bit sometimes00:35
gary_posterlifeless, I was shutting down my juju instances and realized I forgot to give the new subunit/testtools/testrepository packages a whirl.  I just replaced everything on the master and slave and started a new run.  Everything seems great.  Once that packaging snafu is dealt with for subunit, everything looks good for a release from what I can tell.01:23
gary_posterthank you!01:23
* gary_poster disappears again01:24
StevenKBah01:24
StevenKDisappeared before I could tease him.01:24
wallyworld__StevenK: are you going to update factory methods to use info_type for branches?01:35
StevenKDown the line, yes.01:36
StevenKIt's a bit premature right now.01:36
wallyworld__StevenK: ok. i've got some tests with placeholders for when that happens01:36
StevenKwallyworld__: I'm happy to sprinkle in information_type into makeBranch() in the current branch I'm doing.01:37
StevenKI just won't remove private yet01:37
wallyworld__StevenK: that would be nice, thanks01:37
StevenKwallyworld__: Aye, I shall.01:37
wallyworld__you ocr?01:38
StevenKI should be.01:38
=== StevenK changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: StevenK | Firefighting: - | Critical bugs: 3.47*10^2
wallyworld__i will have a mp soon :-)01:38
StevenKI'll be sure to give it the appropiate amount of attention.01:39
StevenK(IE, none at all.)01:39
StevenKcjwatson: Re your db branch, don't worry about the +8, it's a DB patch, and I don't think it should count.01:43
lifelesswgrant: have you read the browserid spec?01:48
wgrantlifeless: Not in depth, but I know roughly how it works.01:53
wgrantlifeless: Assuming it becomes semi-standardised and gains some widespread support, I see no technical reason to forbid it.01:55
lifelessthe thing I wonder, quite apart from whether its a good idea01:55
lifelessis whether someone can create an openid thunk01:56
lifelessand we can avoid supporting it entirely01:56
lifelesswgrant: overhead is overhead01:56
wgrantSure, one could pretty easily implement an OpenID provider which effectively proxies BrowserID.01:57
lifelesswgrant: I skimmed it about 6 months ago01:57
lifelessbut the details are paged out01:57
lifelessso, if you can do that, one wonders my mozilla didn't just do that instead.01:58
wgrantlifeless: Well02:01
wgrantlifeless: Part of the idea of BrowserID is that browsers will implement it.02:01
wgrantSo there'll be no bouncing around.02:01
wgrantThe browser will authenticate directly with its keypair and a time-limited certificate signed by the provider.02:01
wgrantOne could implement an OpenID gateway, but it would still involving hideous redirects and ugliness.02:02
lifelesswgrant: yes, I'm aware of that02:09
lifelesswgrant: my HTTP nazi hate just loooves that02:09
wgrantIt does mean that you have to trust the browser, and you can't really seamlessly integrate server-side 2FA.02:12
wgrantBut it in all otherwise a tonne less hideous, sick and wrong.02:13
* wallyworld__ -> computer shop @#$%%$#!04:18
StevenKHahaahaha04:19
wallyworld__bastard :-(04:19
jtvStevenK, wgrant: just had another silly idea… one of the more expensive parts of publishing is now the “ls -lR.”  Could we maybe kick that off earlier, using the MF?05:34
wgrantWe might be able to, yeah.05:35
jtvIt's getting hard to single out culprits now, but IIRC this was a relatively big one.05:37
StevenKwallyworld__: Are you back yet?06:04
wallyworld__StevenK: back. bloody traffic accident. took 1 hour for a 15 minute trip :-(06:46
bigjoolshaha06:50
jtvwgrant: interested in reviewing the scrubber change?  https://code.launchpad.net/~jtv/launchpad/bug-994650-scrub-faster/+merge/10516906:50
wgrantjtv: It would be my pleasure.06:51
jtv\o/06:51
jtvI think I'll grab some food, meanwhile.06:54
wgrantjtv: Looks good.07:15
wgrantjtv: I'm not sure if avoiding loading the POFiles is a useful optimisation at this stage, but it can't hurt.07:15
wgrantStevenK: Have you run the garbo job over dev and test sampledata?08:24
StevenKwgrant: Nope, I was going to do that with the db patch08:26
wgrantStevenK: Sounds reasonable.08:26
wgrantJust means you have to practice a bit of necromancy.08:27
StevenKMeh, the garbo job is still in db-devel ... :-)08:27
wgrantTrue08:27
stubShould garbo tasks become celery tasks, or should we keep the garbo infrastructure?08:28
wgrantI think it'll all become celery once that's stable.08:34
jtvwgrant: are you well?09:01
jtvYou approved my branch without comments.  Naturally I'm concerned.09:02
wgrantHeh09:07
nigelbhahahahah09:12
nigelbjtv: remember to make some pep8 violations so you both can breathe easy :D09:12
jtvThanks.  I'll try to remember that.09:13
jtvAnd if I don't, then wgrant can point out that I forgot.09:13
nigelbHeh.09:13
StevenKOh, sure. "You forgot to make some silly mistakes in this branch so I can comment. Oh wait, I just did comment. Damn it!"09:14
adeuringgood morning09:40
rick_h_morning adeuring11:16
adeuringmorning rick_h_11:16
jtvAny reviewers about?  If so, have an MP!  https://code.launchpad.net/~jtv/launchpad/bug-994650-scrub-in-batches12:09
czajkowskimorning13:29
rick_h_morning czajkowski early out west isn't it?13:31
czajkowskiyes13:32
czajkowski6:3513:32
=== abentley changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: abentley | Firefighting: - | Critical bugs: 3.47*10^2
czajkowskiany advice folks https://answers.launchpad.net/launchpad/+question/19656814:00
abentleyadeuring: is "name.find('/eggs') >= 0" equivalent to "'/eggs' in name" ?14:17
adeuringabentley: ah, sure...14:17
abentleyadeuring: I find the latter more intuitive.14:17
adeuringright14:18
abentleyadeuring: So most of this code is about determining whether the src directory is sys.path, and adding it if it's not?14:20
adeuringabentley: yes14:21
abentleyadeuring: Have you tested this with lazr.jobrunner installed as an egg?  Seems like this kind of thing could easily break in that environment.14:22
adeuringabentley: I haven't tested it yet. But why should it break?14:23
abentleyadeuring: I find the whole eggs/buildout thing magical, and I never know what's going to break it, so I would want to test it.14:24
abentleyadeuring: In particular, I'd worry that /usr/bin/python -m celery.bin.celeryd_detach is not going to pick up all the other eggs that Launchpad installs.14:25
adeuringabentley: Have a look at the scripts in <lp-branch>/bin. Most of them are thin wrappers around a real script and just configure sys.path. Take celeryd as an example14:26
adeuringabentley: celeryd starts fine14:27
abentleyadeuring: Sure, but most of them specify a giant load of paths.  I thought yours was just about adding one path.14:27
adeuringabentley: no: extended_path = [name for name in sys.path if '/eggs' in name]14:27
adeuringthat adds the whole bunch14:27
adeuringabentley: the extra steps with this_path is about the development environment14:28
adeuringwhere lazr.jobrunner is not W"eggified"14:28
abentleyadeuring: Is there any need to configure CELERYD_PREFETCH_MULTIPLIER now that we're using CELERY_ACKS_LATE?14:29
adeuringabentley: I am not 100% sure but don't want to test it either. I must admit that I did not understand the related docs fully...14:30
abentleyadeuring: r=me14:35
adeuringabentley: cool, thanks14:35
cjwatsonhttps://code.launchpad.net/~cjwatson/launchpad/db-packageset-score/+merge/105113 looks like it has sufficient review now - could somebody mark the MP as Approved for me and then I can land it?14:47
adeuringabentley: what were the jobs again which we wanted to run via celery on stanging/qastaging? i.e., which classs names should appear in jobs.celery.enabled_classes ?14:51
rick_h_cjwatson: done14:51
abentleyadeuring: BranchScanJob.14:51
adeuringabentley: thanks14:51
cjwatsonrick_h_: ta14:53
rick_h_abentley: https://code.launchpad.net/~rharding/launchpad/ga_combo/+merge/105219 for you when you get a sec pls15:15
rick_h_abentley: note that there's a lot of qualifications and such with it15:16
rick_h_looks worse than it is because of the file move/copy of the ga.js around :/15:16
abentleyrick_h_: "a diff'd version"?15:16
rick_h_abentley: in the google-analytics directory is a diff to the raw google provided js file so that it strips the things against our 'outside js/css' policy15:17
abentleyrick_h_: I don't think that copyright statement is correct.  They don't nest like that.15:25
rick_h_abentley: yea, wasn't sure about that. I tried to state the "Code below" to help clarify15:25
abentleyrick_h_: I believe that including Google's code makes the whole file a derived work, so the header should be "Copyright 2012 Canonical Ltd, Copyright Google Inc." or similar.15:28
rick_h_abentley: ok, sounds reasonable to me. Is there someone I should run this by/check officially you know of?15:29
abentleyrick_h_: You could try Amanda Brock.  The dept is https://wiki.canonical.com/LegalServices15:31
rick_h_abentley: ty15:31
abentleyrick_h_: Where was this file before now?15:34
rick_h_abentley: icing/google-analytics/ga.js15:34
abentleyrick_h_: So with this change, do we have it in three places?15:35
rick_h_abentley: yes, for the moment15:35
rick_h_we have to support both combo loader/non-combo loader users15:36
rick_h_I guess we could get rid of the copy in the app/js/google-analytics directory, only the .diff file is *required*15:36
abentleyrick_h_: If we can get rid of a copy, that would be good.15:37
rick_h_abentley: yea, I don't think we'd lose anything. Will do.15:37
=== salgado is now known as salgado-lunch
=== matsubara is now known as matsubara-lunch
=== deryck is now known as deryck[lunch]
gmbUrsinha, czajkowski said you were looking for me.17:44
Ursinhagmb, yeah, I have a lot of questions about launchpad code... but will enter a session in a few minutes :/17:44
gmbUrsinha, Okay, let's try and catch up later on then. Some time after lunch if you're free.17:49
=== matsubara-lunch is now known as matsubara
=== salgado-lunch is now known as salgado
=== deryck[lunch] is now known as deryck
lifelessmorning18:30
lifelessallenap: hi18:30
lifelesssinzui: \o/ \o/18:33
sinzuilifeless, I am glad you are happy18:33
lifelesssinzui: contact-this-team18:38
sinzuiI was sure you were tracking that18:38
lifeless:)18:39
czajkowskisinzui: nice blog post have posted in all the places.18:40
lifelessI'm sure we'll still get admins of casual 'groups' making the same mistake, but hopefully at a much lower incident rate18:40
sinzuiI wish I could have solved the Cc sender issue too, but the the objects need restructuring. I reverted aft 3 hours so that I could get your bug fixed18:40
czajkowskinods I know tis fine18:40
lifelesssinzui: thats a shame18:40
sinzuiyes. barry and I over-specialised the recipient set object. I need to make it behave more like a normal object.18:41
=== cr3_ is now known as cr3
allenaplifeless: Hi there.19:33
lifelessallenap: hi, if you'd like some voice time, I'd be delighted to do so.19:34
allenaplifeless: Cool, that sounds good. I've just seen your reply, so I'll read that.19:34
lifelessallenap: skype ?19:37
allenaplifeless: Is G+ okay? I don't have Skype installed right now.19:41
lifelesstis fine19:41
lifelessit has nearly as good echo cancellation19:42
lifelesswhich is, for high latency, esssssential19:42
allenaplifeless: Okay, invited.19:43
rick_h_lifeless: ping, when you get a sec20:01
lifelessrick_h_: pong; afk for a sec, but write here and I'll reply in a few20:48
=== gmb` is now known as gmb
rick_h_lifeless: working on wrapping the ga.js code into a YUI module so I can combo load it and drop the request/icing/etc.20:48
rick_h_lifeless: in code review, dealing with the copyright notice stuff came up and trying to figure out how to handle documenting the "This is a LP YUI module, all this code inside here is Copyright Google" as it is now20:49
gmbczajkowski, about?20:49
rick_h_lifeless: so came up to ping legal on it and legal is asking me " How do we get permission from Google to modify and redistribute their code?" and can't find anything atm20:49
rick_h_lifeless: think you were involved and wondered if you 1) know how to handle it or 2) know the info I can give legal about how we got/get permission from google to mod/distribute the ga.js script20:50
rick_h_lifeless: see https://code.launchpad.net/~rharding/launchpad/ga_combo20:50
lifelessinteresting20:53
* lifeless runs screaming20:53
=== abentley changed the topic of #launchpad-dev to: http://dev.launchpad.net/ | On call reviewer: - | Firefighting: - | Critical bugs: 3.47*10^2
rick_h_lifeless: so in particular looking at lib/lp/app/javascript/google-analytics/google-analytics.js commenting20:54
* rick_h_ is waiting for bzr.lp.net to finish thinking20:54
rick_h_http://bazaar.launchpad.net/~rharding/launchpad/ga_combo/view/head:/lib/lp/app/javascript/google-analytics/google-analytics.js20:54
lifelessrick_h_:21:00
lifelesshttp://support.google.com/analytics/bin/answer.py?hl=en-GB&answer=103238921:00
lifelessI believe flacoste checked with Google when we first put it in the tree21:00
rick_h_lifeless: rgr, I understood there was some conversation but wasn't sure if it was you someone else and didn't find any reference in the docs/code21:01
lifelessrick_h_: so why do you need to convert it to yui ? is it a combo loader limitation ?21:01
czajkowskigmb: looking for me ?21:01
lifelessrick_h_: or is it because you're going to batch it with all the other modules needed? Could we not not do that and just reference that one file via a combo loader url ?21:02
rick_h_lifeless: well, it was blocking/hanging when I was doing combo loader testing. If I make it a YUI module I can combo load it in the same request with the other JS on every page21:02
lifelessrick_h_: keeping the unaltered google file and unaltered async-loading snippet from google21:02
lifelessthat would keep the delta lower too21:02
rick_h_so it's benifit of -1 request per page done, and servied via our combo loader removing one more thing in icing21:02
lifelesshttp://support.google.com/analytics/bin/answer.py?hl=en&answer=100808021:02
lifeless- goes in the end of <head>21:02
lifelessyou can serve via the combo loader just by using a combo loader url21:03
lifelessso its solely -1 request, per the very first page they land on21:03
lifelessrick_h_: or am I misunderstanding something about combo loader internals ?21:04
rick_h_lifeless: well in order for YUI to know how to request it from the combo loader we parse the YUI modules (make jsbuild) and it's listed21:04
=== _mup__ is now known as _mup_
rick_h_lifeless: but we could manually muck with/modify that I think21:04
lifelessrick_h_: nono, *not* yui21:04
rick_h_right, I gotcha21:04
lifelessif we use http://support.google.com/analytics/bin/answer.py?hl=en&answer=100808021:04
lifelesswhich we probably do today21:04
lifelessand replace the url in it with one pointing at the combo loader21:05
lifelesswe'll get back unaltered content.21:05
lifelessright ?21:05
rick_h_yes21:05
lifelessso just moving the url - trivial. No delta vs google's authoratitive copy.21:05
lifelessand still async, nonblocking on the page.21:05
rick_h_ok, so we can work with the async version, our old one just wasn't using that one?21:06
lifelessI don't know, you're the one doing the debugging :)21:06
rick_h_k, I'll pull down and compare tomorrow21:06
lifelessrick_h_: I gotta run, TL meeting; will pick this up after that.21:06
rick_h_rgr21:06
lifelessif we're blocking today, I'd say keep blocking, do the minimal work21:06
lifelessmove the url to the combo loader, add a bug saying we should be async, move on :)21:07
rick_h_jcsackett: hey, have you been testing out/using the auto reloader for the JS?21:23
* rick_h_ is curious21:24
=== Ursinha is now known as Guest64056
jcsackettrick_h_: i have not.21:28
jcsacketthow does one use that?21:28
rick_h_make jsbuild_watch21:29
rick_h_then just edit the js files and on save it'll auto jsbuild them21:29
jcsacketti will have to check that out.21:29
rick_h_that's the idea at least21:29
jcsackettthat would make somethings *way* easier.21:29
rick_h_yea, I got wondering, I've not done a ton of JS since adding it, but seems you are21:30
jcsackettand the more i do, the more it seems i need to do. :-P21:30
rick_h_ruh roh21:30
jcsackettah, the life of cleaning up old bad code.21:30
jcsackett:-P21:30
rick_h_that's the truth21:30
rick_h_give it a try next time and let me know if you hit issues21:30
jcsackettso, by "next time" you mean tomorrow? :-P21:30
rick_h_:)21:31
rick_h_I didn't want to presume21:31
rick_h_just deeper than you thuoght it was or issues you're hitting?21:31
jcsackettmore things making use of it than i thought.21:31
rick_h_going to be hacking at the coffee house tonight if you need a second set of eyes on anything21:31
* jcsackett nods.21:31
rick_h_ah, gotcha k21:31
jcsackettquick thought? know any reason that a BuiltClass from Y.Base.create might barf when it's constructor.NAME is grabbed?21:32
jcsackettin that, it's constructor has no .NAME.21:32
rick_h_yea, you didn't create the formal constructor. You'd have to inspect it and try to see if it even has a NAME property21:32
rick_h_I think you'd have to get it from your instance?21:33
jcsacketthm. wonder why Banner was fine but something derived from Banner dies ...21:33
rick_h_not sure, honestly, never used the NAME bit :/21:33
jcsacketti'm not either, YUI throws the error.21:33
jcsacketti'll keep poking.21:33
rick_h_well that NAME should change to the Base.create('ISNAME'... right?21:33
rick_h_make sure your requires is correct21:33
rick_h_the error I got today on that was waaay out of left field21:33
jcsackettalready checked requires. would that it was that easy.21:34
jcsacketti'll keep poking. :-)21:34
rick_h_k21:34
cjwatsonI forgot to add a COMMENT in db-devel r11582.  I guess since stub didn't comment on that during review it's not the end of the world; but is there anything I should do?21:38
cjwatsonAs in is it worth landing a follow-up commit?21:38
jcsackettrick_h_: just realized. no `new` call. :-P21:38
jcsackettit's always the little things.21:38
rick_h_jcsackett: ah, yea that one bites me all the time. I've learned that error hard core21:38
sinzuiwallyworld, how goes the morning? What is your preferred means of voice today?22:00
=== matsubara is now known as matsubara-afk
sinzuijcsackett, mumble?22:03
jcsackettsinzui: yup. signing on.22:03
sinzuiwallyworld, wgrant_ look at https://bugs.launchpad.net/launchpad/+bug/70242922:04
_mup_Bug #702429: Pillar owners and private non-security bug visibility <bugs> <disclosure> <qa-ok> <Launchpad itself:In Progress by wallyworld> < https://launchpad.net/bugs/702429 >22:04
wgrant_cjwatson: Not really, but you can if you want.22:05
lifelesscjwatson: it is worth it, jfdi it, and it can land on devel22:09
lifelessthe comments are only used for dev machines anyhow22:09
=== Guest64056 is now known as Ursinha
jcsackettdamn thunderstorms...22:18
=== elmo_ is now known as elmo
StevenKwallyworld_: wgrant_ is stealing your thing!23:53
* wallyworld_ wants it back23:54

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!