[02:22] <emilio> hey guys
[02:41] <emilio> can someone help me?
[02:41] <wgrant> emilio: Probably -- ask your question, and someone will help if they can.
[02:42] <emilio> i'm installed the Launchpad "itself" on my ubuntu
[02:42] <emilio> and want to know some things...
[02:43] <emilio> what is the real purpose of this?
[02:43] <wgrant> Why have you done that, rather than using Launchpad.net? While you *can* run it yourself, it's very difficult, so nobody does, and that wasn't the intention behind releasing the code.
[02:43] <emilio> can i have a my own launchpad with my own projects running on a server without any connection with launchpad.net?
[02:44] <maxh> You CAN, but you (usually) SHOULDN'T.
[02:44] <wgrant> You could, but you'd have to replace all of the images and branding used by the application, and obtain all the necessary resources and knowledge required to run it successfully.
[02:45] <wgrant> It's a very rare case when it's a good idea to do it.
[02:45] <emilio> humm
[02:45] <maxh> You probably want to put your project on launchpad.net.
[02:45] <wgrant> If you are at all unsure, you want to put it on Launchpad.net.
[02:46] <emilio> on my local build of launchpad i don't have a "create account", and only can loggin with the default "admin@canonical.."
[02:47] <wgrant> That's right. It's set up as a development instance, not a production one.
[02:47] <wgrant> Unless you know the application well and are really sure that you want to run your own, you *do not* want to run your own.
[02:47] <wgrant> It's no small undertaking.
[02:48] <maxh> If you ARE really sure you want to run your own you still probably don't want to.
[02:48] <wgrant> That's also true.
[02:50] <emilio> yeah but you said "it's set up as a development instance, not a production one.", so i need do make another build to set up for production?
[02:51] <wgrant> Yes.
[02:51] <wgrant> But there are no instructions for that, and you don't want to do it.
[02:51] <emilio> oh ok.
[02:52] <emilio> ty for your attention!
[02:56] <nigelb> wgrant: you mean its impossible to set up a production instance?
[02:57] <wgrant> nigelb: Not impossible. But you need to know what you're doing, or to have access to wiki.canonical.com.
[02:57] <wgrant> And then you still need to know what you're doing.
[02:57] <nigelb> :(
[02:57] <nigelb> Its supposed to be oss :(
[02:57] <wgrant> Not impossible: just difficult and in most cases foolish.
[02:57] <maxh> (Now watch as 124 people attempt to go to said wiki.)
[02:57] <nigelb> heh
[02:58] <wgrant> Heh.
[03:00] <wgrant> Vostok may change everything, however.
[03:01] <nigelb> Vostok?
[03:01] <wgrant> The derivative distro management tool that Linaro is developing.
[03:01] <nigelb> oh, its based on LP?
[03:01] <wgrant> At the moment the plan is for it to be an alternative UI onto a subset of Launchpad, which OEMs can run locally.
[03:01] <wgrant> And it will probably interact in some ways with Launchpad.net.
[03:02] <maxh> How exactly does Launchpad determine how to link to PGP key pages? I'm getting results for '0xdc471ac2d963ea539409322b6926ebb2c16905f0', which shows my key, but not all the addresses I've added in the past... however long since I uploaded the key, whereas results for '0xc16905f0' (which is obviously the same) shows all addresses.
[03:07] <wgrant> maxh: The former is the fingerprint. They both show the same list of addresses for me -- I suspect the URLs you checked differ another way.
[03:07] <wgrant> (note that the lst eight hexadecimal degits of the fingerprint are your key ID)
[03:08] <maxh> I just manually replaced the relevant portion of the URL to make sure it's otherwise the same. The fingerprint gets two addresses listed, and the key ID gets four.
[03:08] <wgrant> Aha. Looks like the build queue might finally be empty in a day or two!
[03:09] <wgrant> maxh: Hm. In both cases for me the four addresses are listed.
[03:10] <maxh> Well, I suppose if it works elsewhere, it doesn't matter if I see it correctly, since I already have it.
[04:24]  * achiang wonders if anyone knows how to get to a private PPA using the API
[04:25] <wgrant> achiang: Same as a public PPA.
[04:26] <wgrant> getPPAByName should work.
[04:27] <achiang> wgrant: ah! thanks, i didn't see that method. i was poking around the people.archive attribute
[04:27] <wgrant> achiang: That's from the days when people had only one PPA.
[04:27] <wgrant> It just grabs the first one.
[04:27] <achiang> nod, got it, thanks
[06:08] <achiang> anyone want to take pity on a git user who's struggling with bzr?
[06:08] <achiang> in git, my working model is: many logical branches in the same physical source directory
[06:08] <achiang> and then i just move between branches with git checkout
[06:09] <achiang> what's the equivalent in bzr?
[06:11] <wgrant> achiang: #bzr may be more help, but for most projects I have a project directory, with each branch as a separate directory inside that. You can get a system similar to git by using something like bzr-colo, but for non-gigantic projects I see little point.
[06:11] <wgrant> So I'll do something like this to set up a project:
[06:12] <wgrant> mkdir project
[06:12] <wgrant> cd project
[06:12] <wgrant> bzr get lp:project trunk # Will get the default branch from LP, and put it in the 'trunk' directory locally
[06:12] <wgrant> bzr branch trunk my-feature-branch
[06:13] <wgrant> That will create a new directory, my-feature-branch, as a branch from trunk.
[06:13] <wgrant> You can then hack around in there.
[06:14] <achiang> wgrant: and then to push my-feature-branch, i cd into that directory and... what: bzr push lp:project/my-feature-branch ?
[06:16] <wgrant> achiang: bzr push lp:~username/project/my-feature-branch
[06:16] <wgrant> It will then remember that URL for subsequent pushes.
[06:16] <wgrant> You can also later configure bzr to automatically calculate that URL, so you can just 'bzr push' from a new branch.
[06:20] <achiang> wgrant: thanks
[07:17] <MTecknology> So this package has been hovering between 4 second and 4 minutes until it's time to build - for over an hour
[07:17] <MTecknology> I think the math is a little mucked up for some reason.. :P
[07:17] <wgrant> MTecknology: The maths is fine... the buildd manager is not :(
[07:18] <MTecknology> wgrant: what's up with it?
[07:18] <wgrant> MTecknology: It is very much not a fan of having large build queues and lots of builders building lots of small builds.
[07:19] <MTecknology> wgrant: oh- mine is probably about as small as there is :P
[07:19] <wgrant> It sort of ends up taking 180 times longer than expected -- 15 minutes to complete one cycle, instead of 5 seconds.
[07:20] <MTecknology> heh.. that's a bit of a difference
[07:20] <MTecknology> I see that the queue is about half now though
[07:21] <wgrant> Right, the queue is shrinking.
[07:21] <MTecknology> yay for positives
[07:21] <wgrant> Assuming nobody decides that they need to throw another 2700-build Python rebuild onto the queue and then delete it as soon as it's finished, we should have a reasonable queue in 24-48 hours.
[07:22] <MTecknology> wgrant: and then delete it??
[07:22] <wgrant> MTecknology: In the last week we have had two or three rebuilds of Python 2.7 and everything that depends on it.
[07:22] <wgrant> This takes a couple of days of build farm time.
[07:22] <wgrant> The PPAs were then mysteriously deleted.
[07:22] <wgrant> And a fourth one has been created, but is not yet building, and if it does start building then people are probably going to start strangling other people :)
[07:23] <MTecknology> wgrant: ..yucky
[07:23] <MTecknology> only word I could think of
[07:24] <wgrant> I would have chosen something more violent.
[07:24] <wgrant> But yes.
[07:24] <MTecknology> why four rebuilds?
[07:24] <wgrant> I do not know.
[07:24] <wgrant> There was probably some mistake in the first three.
[07:24] <wgrant> Which they didn't discover until they had built EVERYTHING.
[07:25] <MTecknology> shoulda build it locally first :)
[07:26] <MTecknology> pbuilder ftw (on a hefty hefty machine)
[07:27] <MTecknology> wgrant: so if I went along and decided to rebuild all of python and perl and firefox/xul - would you come over here and kill me?
[07:33] <bilalakhtar> noodles775: my daily build is pending from the last 4 days
[07:34] <bilalakhtar> https://code.edge.launchpad.net/~bilalakhtar/+recipe/daily-builds-test
[09:12] <bilalakhtar> noodles775: Back. Power cut over here
[09:13] <wgrant> bilalakhtar: The build queue has been horrific for the last weekish. It should return to normal in the next day or so.
[09:13] <wgrant> .... if the buildds ever come back.
[09:13] <bilalakhtar> hanks wgrant
[09:13] <bilalakhtar> thanks
[13:06] <rioch> I've made a release out of one of my milestones, and that milestone has some bugs set to "Fix committed". Will they automatically be set to "Fix released"?
[13:15] <geser> IIRC no
[13:16] <wgrant> No. We had that for a while, but it proved to be too slow, so it was switched off.
[13:17] <lifeless> wgrant: ?
[13:18] <wgrant> lifeless: Milestoned bugs used to be closed when the milestone was released. But that had to be disabled, as it timed out for large milestones.
[13:19] <lifeless> bah
[13:19] <lifeless> I've requested rabbit be installed
[13:19] <lifeless> if you want to write up a out-of-transaction task for it, please do.
[13:20] <wgrant> lifeless: Ha ha ha.
[13:20] <wgrant> We won't see a message queue for at least two years yet...
[13:20] <lifeless> why ?
[13:20] <wgrant> Because it's been proposed a few times in the last 18 months, and never happened.
[13:21] <lifeless> check lp-developer-dependencies
[13:21] <wgrant> Oh, yes, I know it's in there.
[13:21] <wgrant> It doesn't mean it's going to be on production or in the code any time in the next decade.
[13:21] <lifeless> right
[13:21] <lifeless> I'll be surprised if we don't have the staging instance up next week
[13:22] <wgrant> Ooh.
[13:22] <wgrant> This is is good news indeed.
[13:23] <wgrant> You've not left Prague yet?
[13:23] <lifeless> platform sprint next week
[13:23] <lifeless> GNU hackers meeting after that.
[13:23] <lifeless> Then home.
[13:23] <lifeless> sorry, 'platform rally'
[13:23] <wgrant> Oh, right.
[13:24] <lifeless> 200 ppl in one room is not a sprint
[13:24] <lifeless> anyhow
[13:24] <wgrant> Heh.
[13:24] <lifeless> amqplib
[13:24] <lifeless> play with it.
[13:24] <lifeless> doit.
[13:29] <wgrant> I hate bad tests.
[13:31] <lifeless> jml: http://pypi.python.org/pypi/contextdecorator/0.10.0
[13:37] <jml> wgrant, hello
[13:39] <wgrant> jml: Hi.
[13:42] <lifeless> wgrant: so do I.
[13:43] <lifeless> wgrant: lets fix them
[13:43] <wgrant> lifeless: Oh, I'm rewriting this batch, don't worry.
[13:43] <wgrant> It's just infuriating how bad the Soyuz tests are.
[13:44] <lifeless> jml: http://www.mypressi.com/
[14:10] <geser> is there some ready vm-image for LP development? I planned to setup LP for development (in a chroot or vm as I want to have it seperate from my normal system) but didn't find time to do it
[14:11] <wgrant> There's no pre-built image, but there's https://dev.launchpad.net/Running/VirtualMachine
[14:23] <infinity0> how do i delete a contact email from my launchpad login service account? (as opposed to launchpad itself)
[14:24] <wgrant> infinity0: Despite the name, that's not actually part of Launchpad -- it's developed and managed by another team entirely.
[14:24] <infinity0> oh
[14:24] <wgrant> At the moment it's not possible to remove email addresses. However, I saw some bugmail this morning which suggests that it's fixed in the next release.
[14:24] <infinity0> oh, ok, thanks
[14:24] <infinity0> any estimate of a timeframe on that?
[14:24] <wgrant> No idea, sorry.
[14:25] <wgrant> Bug 507214
[14:25] <infinity0> cool, thanks
[14:26] <wgrant> It's targetted to 2.7.0, which is scheduled for the 22nd.
[14:26] <wgrant> Not sure how reliable that is, though.
[15:19] <MysteryX> Can I host my own, closed-source, software on launchpad for a fee? (or even without one?)
[15:22] <nigelb> for a fee, yes
[15:23] <MysteryX> good
[15:25] <MysteryX> I'll host my websites source there
[18:49] <d1b> morning - is there a way to view all source repos for a project
[18:49] <d1b> like i can do with say a gitweb representation (or hg)
[18:50] <thumper> d1
[18:50] <thumper> d1b: hi
[18:50] <d1b> MORNING
[18:50] <thumper> d1b: things are somewhat different on Launchpad than git or hg
[18:51] <d1b> eg http://bazaar.launchpad.net/~zeitgeist/
[18:51] <thumper> d1b: for a project to look at the branches for the project, go to https://code.launchpad.net/<project>
[18:51] <d1b> nothing is there
[18:51] <d1b> thumper: :/
[18:51] <thumper> that is a person or team
[18:51] <thumper> not a repository
[18:51] <d1b> http://bazaar.launchpad.net/~zeitgeist/zeitgeist/trunk/files
[18:51] <d1b> ...
[18:51] <d1b> but i can't view http://bazaar.launchpad.net/~zeitgeist/zeitgeist/
[18:51] <thumper> http://code.launchpad.net/~zeitgeist
[18:51] <d1b> why not?
[18:52] <thumper> d1b: because we haven't set it up that way :-|
[18:52] <thumper> the filesystem isn't what you see
[18:52]  * d1b grumbles about usability
[18:52] <d1b> ok i never was linked at any point to
[18:52] <d1b> http://code.launchpad.net/~zeitgeist
[18:52] <d1b> or code.launchpad.net
[18:52] <d1b> so without coming on irc
[18:52] <d1b> how would i have found that page?
[18:52] <d1b> neva
[18:52] <thumper> d1b: the branch name is a hyperlink on the bazaar.launchpad.net site
[18:53] <thumper> d1b: which takes to the LP page
[18:53] <d1b> thumper: from the launchpad site
[18:53] <d1b> oh god
[18:53] <thumper> we could do better here
[18:53] <d1b> sorry yes i did
[18:53] <d1b> weird - maybe it was because i came through edge?
[18:54] <d1b> ok
[18:54] <d1b> so what happened was i hit
[18:54] <d1b> https://code.launchpad.net/zeitgeist-project
[18:55] <d1b> thumper: sorry you are correct
[18:55] <d1b> i haven't got much sleep
[18:55] <d1b> my applogies
[18:55] <thumper> d1b: the usability still sucks
[18:55] <thumper> and I accept that
[18:55] <d1b> i just had problems finding the correct zeitgiest page i wanted
[18:55] <thumper> we do need to work on it
[18:55] <d1b> like i was on one of them
[18:55] <d1b> and i wanted to go up a dir / thing and view for al lof them
[18:55] <thumper> d1b: can you file a bug for me?
[18:56] <thumper> d1b: file it on launchpad-code
[18:56] <d1b> let me nail down the behaviour i want
[18:56] <thumper> ack
[18:56] <d1b> so i ened up https://launchpad.net/zeitgeist there
[18:57] <d1b> sorry wrong entry point
[18:57] <d1b>  https://launchpad.net/gnome-activity-journal --> to view the code --> wanted to see where the zeitgeist stuff was
[18:58] <d1b> mm this is not a problem that is fixable - but the solution would be neat
[18:58] <d1b> if launchpad was able to do ctags but like for python - where the dev could give it hints that would have been a solution
[18:59] <d1b> sort of like - the registered branches thing but extending on that
[19:00] <d1b> thumper: anyways - when i click on the link from the code page i as a user expect to see the files and not the project page for a branch
[19:00] <d1b> that may be 'wrong'
[19:01] <d1b> in any case pages like
[19:01] <d1b> https://code.launchpad.net/~zeitgeist/zeitgeist/magic
[19:01] <d1b> should have the  View the branch content --> larger
[19:01] <thumper> d1b: very soon (hopefully) it'll be more obvious
[19:01] <d1b> awesome
[19:01] <thumper> I want to move to make the code browsing and history more a part of Launchpad
[19:01] <d1b> oh neat
[19:02] <thumper> but right now it has too much in the way of memory leaks, excessive memory usage, and bad caching
[19:02] <d1b> what about integrating bugs to code?
[19:02] <thumper> we're working on that
[19:02] <d1b> awesome
[19:02] <thumper> what do you mean by "integrating bugs to code"?
[19:02] <d1b> when i fix a bug, and in the commit i say "this fixes foo"
[19:02] <d1b> and then i close a ticket
[19:03] <thumper> ah, the related bug status setting
[19:03] <d1b> or saying "this is fixed in commit number xyz"
[19:03] <thumper> it's on my hit list
[19:03] <d1b> awesome
[19:03] <thumper> howevery my hit list is kinda long
[19:03] <thumper> a big feature for us right now is building directly from a branch into a ppa
[19:03] <d1b> also
[19:03] <d1b> why is it that http://bazaar.launchpad.net/~zeitgeist/zeitgeist/magic/annotate/head:/zeitgeist-daemon.py
[19:04] <d1b> for example shows me the revision it was changed - but i can't view who the author was?
[19:04] <d1b> nor the date
[19:04] <thumper> really?
[19:04]  * thumper clicks
[19:04] <d1b> no i mean i have to click on it
[19:04] <thumper> d1b: just mouse over and pause
[19:04] <d1b> yes but i would prefer to be able to see that without doing that
[19:05] <thumper> that content would be really repetitive
[19:05] <d1b> oh i meant as an optional thing
[19:08] <d1b> oh wait what?
[19:08] <d1b> there is no search
[19:08] <d1b> hehehe
[19:08] <d1b> but bzr-search no?
[19:08] <thumper> when loggerhead stops being such a memory hog, we may integrate it
[19:08] <thumper> but yes, it is feasible
[19:08] <d1b> well that's where gitweb /hg thingy are at :/
[19:09] <d1b> thumper: out of interest could you link me to the code that does this that "leaks"
[19:09]  * thumper shrugs
[19:09] <thumper> d1b: lp:loggerhead
[19:09] <d1b> ... im not so good with lp:loggerhead stuff
[19:10] <d1b> i mean i need a url ;P
[19:10] <thumper> if it was any more specific, we'd have fixed it by now
[19:10] <thumper> bzr branch lp:loggerhead
[19:10] <d1b> sorry i think i need bzr 2 or something no?
[19:10] <thumper> yes
[19:11] <d1b> not going to happen ^^
[19:11] <thumper> why?
[19:11] <d1b> debian lenny
[19:11] <d1b> and virtualbox is currently foobared on me
[19:11] <d1b> http://bazaar.launchpad.net/~loggerhead-team/loggerhead/trunk-rich/files/head:/loggerhead/
[19:11] <d1b> right location?
[19:11] <thumper> should be
[19:12]  * thumper goes to grab more fud
[19:12] <d1b> i take it i can't get a 'snapshot'
[19:12] <d1b> from the interface like i can in hg and gitweb
[19:14] <d1b> http://bazaar.launchpad.net/~loggerhead-team/loggerhead/trunk-rich/annotate/head:/loggerhead/config.py ... :/ def _level_to_int_level(value):
[19:14] <d1b> ..... omg
[19:15]  * d1b ok sleep time
[19:15] <thumper> night
[19:23] <d1b> thumper: question does the memory leaking occur with threads only - or does it also occur without
[19:23] <thumper> d1b: I've not looked into it much
[19:24] <d1b> thumper: well i would be interested
[19:24] <d1b> is it even possible to run it without threading?
[19:24] <d1b> sure i could 'mock' threading but
[19:25] <thumper> d1b: it is just a wsgi app
[19:25] <thumper> so sure you could using the python reference impl
[19:27] <d1b> also haaha - why only so so few commits per a page
[19:27] <thumper> hey, I didn't write it
[19:27] <thumper> I'm just kinda responsible for it in LP
[19:28] <d1b> heeh
[19:28] <thumper> there is a loggerhead mailing list with other devs hosted through LP I think
[19:28] <thumper> if you are seriously interested in getting involved with loggerhead, start there
[19:30] <d1b> not really
[19:30] <d1b> i have already seen too many things that hurt my eyes
[19:32] <thumper> heh
[19:32]  * d1b crawls back to github
[19:34] <d1b> don't you love it when people go and close stdin, then mark stderror ...... oh nevermind
[19:34] <rockstar> thumper, where are you?
[19:34] <thumper> rockstar: LHR
[19:35] <rockstar> thumper, ah.  I thought you were headed straight to Singapore from here and thought "That was quick."
[19:35] <thumper> heh
[19:35] <thumper> no
[19:35] <thumper> and it is hong kong