[01:13] <d34df00d> Hi!
[01:13] <d34df00d> I've just updated a new version of po, selecting a new import, for Rosetta.
[01:13] <d34df00d> I'm the owner of the translations, and Launchpad says I have full access.
[01:14] <d34df00d> Why do my changes are marked as suggestions and therefore the corresponding strings are considered to be untranslated?
[01:20] <d34df00d> And, how one can delete a duplicate translation template?
[04:19] <alpharesearch> hello, is there a guide that tells me how to use my PPA for lucid and maverick... right now I did an upload for lucid and than I change the debian/changelog but this didn't work...
[04:19] <alpharesearch> I got a rejection email that the debian.tar.gz already exists
[04:26] <micahg> alpharesearch: https://help.launchpad.net/Packaging/PPA/Uploading
[04:32] <alpharesearch> micahg: dput ppa:your-lp-id/ppa <source.changes> that is what I did... I tried -f but this didn't help
[04:32] <wgrant> alpharesearch: You can't upload the same version twie.
[04:32] <wgrant> +c
[04:32] <wgrant> You need to change the version string (eg. append ~lucid1, ~maverick1, etc.)
[04:35] <alpharesearch> in the changelog file I guess?
[04:38] <alpharesearch> wgrant: thank you...
[05:28] <MTecknology> So.. is it bad at all if I have a user account in Launchpad that's only there as basically an ssh keyring for logging into launchpad from many users on many servers?
[05:42] <thumper> MTecknology: heh, probably
[05:43] <MTecknology> thumper: Any suggestions? I like the fact that each system is using different keys, each development project in progress gets a key, then removed when not used
[05:44] <thumper> MTecknology: in reality it will work
[05:44] <thumper> MTecknology: but not considered best practice from our point of view
[05:45] <MTecknology> thumper: what would be better?
[05:45] <MTecknology> thumper: just have some developer keys that can get dropped in .ssh ?
[05:46] <thumper> MTecknology: I don't know how you have your users set up, or your projects working
[05:49] <spiv> It's pretty unusual to want anything other than "one LP account per actual human being".  (There are reasonable cases of not one-to-one correspondence, like automated branch gatekeeper accounts like ~bzr-pqm, but it's not the common case.)
[05:50] <spiv> If you want access control, then I'd recommend using teams rather than fake people with lots of SSH keys.
[05:51]  * spiv really goes to lunch this time
[05:51] <spm> MTecknology: the problem you create by many:1, is that any of the many can 'break' things, and you have no recourse to discovering who did it. I simplify to the dramatic case.
[05:52] <spm> it's not just a 'bad in launchpad' thing; it's pretty much a bad in everything thing.
[05:54] <MTecknology> spm: For projects, we make a user account on the system, drop project data in there, setup a shared key for that user in LP so it can grab som private branches, then we use authorized_keys for who can ssh in as that user
[05:57] <MTecknology> spm: I'll try to start thinking of better ways to manage it
[05:57] <spm> Ahh. That's a subtly different case from your earlier description. It's not ideal; but it's feasible.
[05:59] <MTecknology> spm: I'm bad at explaining things :P
[05:59] <spm> :-)
[06:00] <spm> so what you're describing is more in line with the gatekeeper that spiv mentions. it does weaken the trust model, but often the tradeoff is worth it. just beware of what you lose by doing this.
[06:01] <MTecknology> alrighty, thanks :)
[06:53] <mtaylor> spm: ola amigo!
[06:53] <spm> mtaylor: hola!
[06:53] <mtaylor> spm: how's the land of oz?
[06:53] <spm> hmm. only perfect today. bit of a let down tbh.
[06:54] <mtaylor> I hear that. sucks to be surrounded by such wonder all the time
[06:54] <mtaylor> spm: so ... I have come by to poke you in the eye about the ~swift mailing list of death
[06:55] <spm> truly. :-) (fwiw, the reference was an older slogan my home state had: Queensland: Beautiful one Day, Perfect the next)
[06:55] <spm> yah. that's giving me a real headache that one.
[06:55]  * wgrant doesn't remember that one. It must be old.
[06:55] <mtaylor> that's a good slogan
[06:55] <mtaylor> now quite as good as "what happens in vegas, stays in vegas" ... but good nonetheless
[06:56]  * spm pokes wgrant with the bluntish end of a stick, or walking frame...
[06:59] <spm> mtaylor: so, I believe! the last thing to try is bouncing mailman itself. for fairly obvious reasons, not something I want to just do for the heck of it.
[06:59] <mtaylor> indeed
[06:59] <spm> supposedly a rollout is scheduled for later this week - so was hoping to progress then; failing that, I know barry is somewhat available again from todayish; and was hoping to grab him at some point
[07:00] <mtaylor> cool. well, good luck! it certainly seems like a nice weird state we got that bad boy in
[07:01] <spm> seeing as the whole mailman/launchpad thing is largely his baby. I'm confidant he'll have some ideas on how to extricate us.
[07:01] <spm> yeah. bit of a bugger.
[07:01] <spm> So it looks like no progress, largly because there hasn't been :-/ but surely has not been forgotten
[07:03] <mtaylor> hehehe
[08:48] <fta2> what's going on with the builders??
[09:41] <geser> could someone please kick the build dispatcher? It looks like it doesn't dispatch anymore since around 12 hours
[09:48]  * bigjools looking
[09:50] <bigjools> geser: what makes you say that then?
[09:52] <geser> bigjools: https://edge.launchpad.net/builders/ shows many empty buildds (while there are jobs in the queue if the page doesn't lie) and the buildd history for the idle buildds show the last job for around 12 hours ago (https://edge.launchpad.net/builders/palmer/+history, https://edge.launchpad.net/builders/vernadsky/+history, https://edge.launchpad.net/builders/yellow/+history)
[09:52] <bigjools> ok
[09:53] <bigjools> they are not in the pool right now
[09:54] <geser> why does the builders page list them then? (Or am I misunderstanding something?)
[09:55] <bigjools> oops sorry they are, I was looking at a different builder, my bad
[10:15] <bigjools> geser: ok we're bringing them back up now.  It's probably bug 463046 if you're interested.
[10:16] <bigjools> wgrant: are you really working on that? :)
[12:14] <Daviey> Hi, using the API - is there a way of getting a list of user/team PPA's; without manually handling the JSON?
[12:14] <bigjools> Daviey: are you not using launchpadlib?
[12:15] <Daviey> bigjools: Oh yes!  But I can't see how to handle ppas_collection_link
[12:15] <wgrant> Just use .ppas
[12:15] <bigjools> don't use that
[12:15] <bigjools> wot wgrant said
[12:16] <wgrant> launchpadlib applies magic to turn it into a less useless attribute.
[12:16] <wgrant> blah_link -> blah
[12:16] <wgrant> blahs_collection_link -> blahs
[12:16] <Daviey> Ahh!  Didn't realise that.. or see it in the docs..   thanks!
[12:16] <wgrant> Yeah, +apidoc doesn't document launchpadlib's magic, since it doesn't know about launchpadlib.
[12:17] <wgrant> It probably should.
[12:19] <Daviey> Awesome, thanks..
[12:19] <Daviey> One more API question, is there a way of viewing ubuntu archive uploads by person?
[12:19] <bigjools> what do you mean by upload?
[12:20] <bigjools> as in - what API object are you expecting
[12:21] <Daviey> Essentially, access to https://edge.launchpad.net/~PERSON/+related-software
[12:22] <Daviey> Ie, for a given person, return the uploaded packages
[12:30] <Daviey> I could do getPublishedSources(created_since_date), then determine the uploader - but i hoped there would be a by_person parameter.
[13:27] <bigjools> Daviey: I don't think there's anything exported to do that yet
[13:27] <bigjools> feel free to contribute :)
[13:28] <Daviey> bigjools, Adding to my todo.. :)
[14:17]  * flacoste is back.
[15:14] <serg> hi here. I've a question about bug interface. say, a bug affects a project and is targeted to a release. Like in here: https://bugs.launchpad.net/zope3/+bug/98275 - how to make these " Status tracked in" to appear ? All I get is two rows with two independently changeable statuses
[19:07] <mtaylor> any losa around?
[19:09] <mtaylor> I have a public project (that has some other issues anyway) but which is requiring that people be logged in to launchpad to view ... launchpad.net/swift
[19:09] <mtaylor> if there's anyone awake who can look at that, it would be really stellar
[20:41] <bac> hi mtaylor
[20:42] <mtaylor> hi bac !
[20:42] <bac> mtaylor: on staging could you change driver and maintainer to be you, not the teams: https://staging.launchpad.net/swift
[20:42] <bac> i'm confused as to what is causing the login.  it is usually the attempt to display private stuff.
[20:43] <bac> but, given the state of swift-core it could be triggering it
[20:43] <mtaylor> bac: is swift-core private? I thought we'd finally killed all of the private things
[20:43] <bac> mtaylor: no it isn't but i thought it retained some oddness
[20:44] <bac> mtaylor: it's the only thing i know to try at the moment
[20:44] <mtaylor> well, the ~swift team retained oddness re: mailing list
[20:44] <mtaylor> bac: I'll try right now
[20:44] <bac> thx
[20:45] <mtaylor> got a timeout error :)
[20:45] <mtaylor> but now is done
[20:46] <bac> ok, that's not it.  :(
[20:48] <mtaylor> darn
[20:48] <mtaylor> bac: so we seem to have been helpful in pointing out some places where taking private things public are painful :)
[21:49] <tommie-lie> I hope this is the right place to ask for "correct behavior" before reporting bugs
[21:50] <tommie-lie> I reported a bug https://bugs.launchpad.net/bugs/610567, added a patch, marked it as fix released and assigned the bug to myself, however, the bug does not appear neither in the project's list of bugs, nor in my personal list of bugs
[21:50] <tommie-lie> and obviously even the bot can find it
[21:51] <tommie-lie> is this normal, did I do something wrong, or should I file a bug against launchpad?
[21:52] <geser> bugs in state "Fix released" aren't shown by default, you have to use the advanced search to find them
[21:53] <tommie-lie> ah
[21:53] <tommie-lie> thanks
[21:53] <hexmode> do daily builds not really work?
[21:53] <hexmode> also, estimated build times for receipes seem more like wild guesses
[21:55] <tommie-lie> geser: then I have to hope that the project owner reads his launchpad mails in order to see my report, right?
[21:58] <geser> why did you mark it as "Fix released"? "Fix released" is to be used when the fix is released e.g. part of the next upstream version
[21:58] <tommie-lie> err
[21:58] <tommie-lie> hm
[21:58] <tommie-lie> good question
[21:59] <tommie-lie> I meant something like "fix committed", but then again this is not quite right either
[22:00] <geser> let it at "New" and let the project owner update it
[22:01] <geser> e.g. when they included your patch
[22:03] <tommie-lie> I just set it to "In Progress" as I have "taken responsibility to fix the bug" and have "begun work"
[22:03] <geser> ok
[23:07] <michaelh1> Morning.  How can I find out the corresponding anonymous URL for a bzr lp:project URL?
[23:08] <michaelh1> I want to use buildbot to access a LP based project, but it insists on using bzr+ssh with no credentials
[23:08] <poolie> michaelh1, you can look at the web page, or if you have the ssh url just change the scheme
[23:08] <poolie> it's always the same path
[23:08] <lifeless> bzr info -v lp:project
[23:08] <poolie> hi lifeless, michaelh1
[23:09] <lifeless> hi poolie
[23:09] <michaelh1> Ah! there's the magic: I was trying http://bazaar.launchpad.net/linaro-gcc, but it's actually .../~linaro-toolchain-dev/gcc-linaro
[23:09] <michaelh1> Ta.
[23:10] <geser> can someone check why the PPA buildds are idling with jobs in the queue?
[23:10] <mtaylor> hey lifeless
[23:11] <wgrant> I was about to ask about that.
[23:11] <wgrant> I think there's another broken build happening.
[23:11] <lifeless> isn't this just the secheduler bug ?
[23:11] <wgrant> It's redispatching frequently.
[23:12] <wgrant> lifeless: Want to look at the buildd-manager log?
[23:12] <wgrant> Or is that LOSAy?
[23:13] <lifeless> losay AIUI
[23:13] <geser> and do we currently have that many private builds in the official queue? I see them half the day "building private source" only
[23:13] <lifeless> wgrant: also, just finished schipol -> shanghai -> akl -> chc
[23:13] <lifeless> wgrant: I'm in /no/ state to be touching anything production
[23:13] <wgrant> lifeless: Haha, forgot that bit.
[23:15] <wgrant> LOSA ping...
[23:16] <geser> and btw: is the computation of the queue length on i386 broken? I can't believe that 3 i386 builds need only 2 min for 32 jobs
[23:16] <wgrant> geser: Those are probably translations jobs.
[23:16] <wgrant> Which have a, er, bad duration calculation.
[23:18] <poolie> hi wgrant
[23:18] <wgrant> Morning poolie.
[23:19] <michaelh1> What's a good way to cache bzr data coming from launchpad?
[23:20] <michaelh1> I want to download the same 500 MB tree to many machines many times
[23:20] <michaelh1> (All read only at least)
[23:22] <micahg> michaelh1: branch once and scp?
[23:22] <michaelh1> micahg: I want to pull down the changes as they occur as well
[23:23] <lifeless> michaelh1: setup a bzr repository in a common root directory on each machine
[23:24] <lifeless> michaelh1: if you have the machines all local here, do that, and then also push the trunk to each machine once, to see the repos
[23:25] <michaelh1> OK.  Then pull changes from LP into the repository, and the slaves will pull that down?
[23:26] <lifeless> yes
[23:27] <michaelh1> What verbs does bzr use in read-only HTTP mode?  Can I put a HTTP proxy in the way?
[23:27] <lifeless> bzr cache busts
[23:27] <poolie> so yes, you can
[23:27] <poolie> only get and head
[23:27] <lifeless> and post
[23:28] <lifeless> you need to ensure post gets a sane error if you deny it
[23:28] <lifeless> but
[23:29] <lifeless> http non-smart mode is much much much slower than authenticated bzr+ssh mode - use that if you can. _really_.
[23:29] <lifeless> you can create an account for your machines and give them that accounts ssh key
[23:29] <michaelh1> I'm having trouble getting the SSH key to propagate into buildbot - not sure why
[23:30] <michaelh1> Does bzr fork out to SSH or use a Python implementation?
[23:31] <poolie> on unix by default forks
[23:32] <poolie> check the .ssh permissions?
[23:33] <lifeless> michaelh1: you might consider hudson instead - many teams at canonical are migrating/using it
[23:33] <michaelh1> I think that buildbot cleans the environment in some way and looses the connection to the ssh-agent
[23:34] <michaelh1> Right, will go have a play...