[01:19] <IiLuminated> some project got open sourced http://code.google.com/p/macaw/ and http://drpetter.proboards.com/index.cgi?board=musagi&action=display&thread=82
[03:06]  * mwhudson lunches
[03:29] <lifeless> is there a standard environment variable for selecting EDGE/STAGING for API clients ?
[04:10] <thumper> today seems to be full of fail
[04:17] <thumper> mwhudson: do you know of a list like object that is fed by a generator but reusable?
[04:18]  * thumper tries something
[04:19] <mwhudson> thumper: there might be something in itertools
[04:29] <thumper> mwhudson: look sane enough? http://pastebin.ubuntu.com/412900/
[04:30] <thumper> mwhudson: if I were to put that into the launchpad tree, where should it go?
[04:31] <lifeless> thumper: so, queued status doesn't seem to have any major side effects
[04:31] <thumper> cool
[04:31] <lifeless> thumper: are you happy with my using that on production ?
[04:31] <thumper> if you break it, you fix it, how does that sound?
[04:32] <lifeless> I'm not in the DBA group anymore
[04:32] <lifeless> or I'd say yes,
[04:34] <lifeless> thumper: https://bugs.edge.launchpad.net/launchpad-code/+bug/561160 would seem to be a blocker
[04:34] <mup> Bug #561160: API: 'Code failed to merge' setting doesn't work <Launchpad Bazaar Integration:New> <https://launchpad.net/bugs/561160>
[04:35] <thumper> lifeless: heh
[04:35] <thumper> lifeless: convince jml that queues are a priority and I'll get it all working
[04:37] <lifeless> thumper: slack time :)
[04:37] <thumper> heh
[04:37] <lifeless> thumper: do you have any thoughts about 561160 ?
[04:37] <thumper> yes
[04:40] <mwhudson> thumper: looks neat enough, probably lp.services.utils
[04:40] <thumper> mwhudson: I've added an extra fix for when it hits completion
[04:40] <thumper> mwhudson: shorter and simpler than the active state example
[04:41] <lifeless> thumper: and your thoughts were ?
[04:41] <thumper> lifeless: it is a trivial fix
[04:42] <lifeless> cool
[04:45] <lifeless> thumper: does that imply you'll do the trivial fix ?
[04:45] <thumper> not right now, no
[04:45] <thumper> I'm in the middle of three other things
[04:45] <lifeless> Oh, I wasn't meaning instantly
[04:47] <thumper> in which case, yes probably
[05:05] <lifeless> is https://dev.launchpad.net/Getting still accurate about the supported versions?
[05:07] <wgrant> It's Hardy, Jaunty, Karmic, and Lucid most of the time, though it sometimes breaks for a few hours, and two tests fail.
[05:08] <lifeless> so the page is stale ?
[05:09] <wgrant> In that Lucid should probably be added? Yes.
[05:12] <lifeless> ugh wow
[05:12] <lifeless> 650MB of pacakes rocketfuel-setup wants to isntall
[05:13] <wgrant> lifeless: Disable Recommends installation.
[05:14] <wgrant> (add '-o Apt::Install-Recommends=false' to the apt-get command-line)
[05:14] <lifeless> does that work for aptitude command line too ?
[05:15] <wgrant> Hopefully.
[05:15] <jtv> hi folks
[05:15] <wgrant> Morning jtv.
[05:16] <lifeless> much better
[05:16] <lifeless> perhaps that should be in rocketfuel-setup ?
[05:18] <wgrant> Patches welcome :P
[05:22] <spm> wgrant: I have to ask. how long have you been waiting to throw that line back at lifeless? ;-) months? years...?
[05:22] <wgrant> spm: Heh heh.
[05:23] <spm> must have worked. he's speechless. /me records the moment for posterity. something to tell the grandkids. "Yes, I was there when lifeless was rendered speechless."
[05:24] <wgrant> But yes, seriously, I can't think of a reason not to have it in there.
[05:25] <spm> sounds good to me as well; if anyone cares...
[05:27] <lifeless> wgrant: rs=lifeless
[05:27] <lifeless> spm: what were you saying about speechless? :P
[05:28] <spm> lifeless: it depends on the person. so eg wgrant speechless == > 2-3 hours; you, 10-15 mins. me 2-3 seconds. it depends.
[05:28] <jtv> lifeless: stop pretending... the "rs" means you weren't really reviewing.  You were speechless.
[05:28] <wgrant> rs='really speechless'?
[05:28] <lifeless> lol
[06:52] <mwhudson> thumper: http://staging.launchpadlibrarian.net/42124669/mwhudson-linux-trunk.log <- log with my latest branch from the kernel import
[06:53] <thumper> mwhudson: it still took over an hour to get the branch?
[06:54] <mwhudson> "Getting exising bzr branch from central store." step took an hour :(
[06:54] <mwhudson> thumper: looks like it
[06:54] <thumper> 2010-04-12 03:04:41 INFO    29134781 bytes transferred | Fetching revisions:Inserting stream
[06:54] <thumper> 2010-04-12 03:27:01 INFO    Fetching revisions:Finishing stream
[06:54] <mwhudson> there's 20 minutes at the end of that step without output, dunno what's happening there
[06:54] <thumper> what happens there?
[06:54] <mwhudson> thumper: jinx!
[06:55]  * thumper EODs
[06:55] <thumper> good work though
[06:56] <mwhudson> thumper: the import should complete overnight, i think
[06:57] <mwhudson> then we can see how long the scanner takes to process it, that'll be good for a laugh
[07:00] <wgrant> Why does the 'N bytes transferred' not increase? Is that only the volume transferred between messages?
[07:00] <mwhudson> wgrant: something like that yes
[07:00] <mwhudson> i actually forget the details
[07:07] <lifeless> mwhudson: repack on the server
[07:07] <lifeless> mwhudson: perhaps
[07:08] <mwhudson> lifeless: it's fetching over sftp
[07:08] <lifeless> mwhudson: ok, not that then.
[07:08] <wgrant> Is the slowness also perhaps attributable to strawberry being slow?
[07:09] <mwhudson> wgrant: some of it, yeah
[07:35] <lifeless> spm: is there a standard for LP API credential files for services you maintain ?
[07:36] <spm> lifeless: I'm not sure I appreciate the question? we deal with SSL certs - so I suspect you mean something else. Client side by implication? if Y, not that I'm aware of. we/losas have very little interaction with the API's.
[07:36] <lifeless> spm: oauth cred files
[07:37] <lifeless> spm: e.g. PQM, tarmac, other scripts that run and speak to LP API's
[07:37] <lifeless> do you like to have the cached credentials files that they need in some specific place ?
[07:38] <spm> short answer is they don't talk to the LP api's - again aiui; they're using ssh specific keys where we need to split user access. ???
[07:38] <lifeless> spm: ok, weirdd ;)
[07:38] <spm> legacy? :-)
[07:38] <spm> keep in mind we were having to do this stuff before the api's existed.
[07:39] <spm> actually i can only think of 1 method where we do access the api's - killing spam in bugs - and that's a pita to use.
[07:41] <spm> PIA, more from a work flow perspective. request: "pls delete spam <url>" so we click on the url; verify it's spam. ssh to server X; sudo; setup api env; lookup docco on how to do this... ; run funky api command/login; disabled spam. Yukness.
[07:41] <lifeless> spm: yes, I hate that too.
[07:41] <spm> i personally find it easier to switch to DB and do it in sql... less faffing around.
[07:42] <lifeless> spm: however, it is the blessed be official way to do stuff
[07:42] <lifeless> spm: I figure documenting the path the creds file at should make it marginally easier
[07:42] <spm> tho - from a pqm/etc perspective where we are scripting large amounts of "stuff" - there may be a case to do stuff in the API.... hrm....
[07:42] <lifeless> spm: but perhaps you should also have a LP bug
[07:43] <lifeless> 'create credentials for remote server with no browser hurts my bwain'
[07:43] <spm> I am a sysadmni of very little brain, so the hurt is correspondingly smaller
[07:43] <lifeless> spm: will it break you if I have a new PQM to deploy tomorrow that used LP API's ?
[07:44] <lifeless> spm: we'd like to attach the error messages so that everyone can see the failure, not just the person that hit the big green button.
[07:44] <spm> for bzr? probably - or rather *possibly*. we'd need to split the other users on the same box. I don't anticipate major drams.
[07:44] <lifeless> spm: shouldn't need to split any users
[07:44] <lifeless> spm: I can make the creds file a PQM config option
[07:45] <spm> ahh, k.
[07:45] <lifeless> spm: this is why I'm discussing it with you
[07:45] <lifeless> spm: to know what you need
[07:45] <spm> I'd want to ok with the 2nd group - we can't break them. Tho I suspect they'd be sympathetic.
[07:45] <lifeless> they are ?
[07:45] <spm> U1 :-)
[07:46] <lifeless> I'll mail statik now
[07:46] <spm> LS are configured, but not activly using
[07:46] <spm> lifeless: cc losas@ pls too. if only so the other guys get a heads up.
[07:46] <lifeless> c.c ?
[07:46] <spm> aye
[07:46] <lifeless> with or without s ?
[07:47] <spm> um....
[07:47] <spm> with :-)
[07:47] <lifeless> I have both in my mail client :P
[07:47] <spm> ha. wonder if without works....
[07:48] <spm> Ahh that may go to just tom. as that's his title: losa of the losa's (apostrophe for clarity)
[07:48] <lifeless> oh man
[07:48] <spm> the 'l' is severely overloaded these days
[07:48] <lifeless> don't say that aloud
[07:48] <spm> heh, esp in a bad kiwi accent?
[08:00] <lifeless> running make schema I get
[08:00] <lifeless> ImportError: No module named tickcount
[08:00] <lifeless> in a new setup following the current dev howto
[08:00] <lifeless> on lucid
[08:00] <lifeless> any advice ?
[08:15] <noodles775> lifeless: yep, I think you're missing python-support from the launchpad ppa?
[08:21] <adeuring> good morning
[08:22] <al-maisan> Good morning noodles775 and adeuring
[08:22] <adeuring> hi al-maisan, hi noodles775!
[08:22] <noodles775> Hi al-maisan and adeuring
[08:40] <mwhudson> holy ****: the kernel import completed!
[08:41] <mwhudson> spm: is the puller running on tellurium, or is it broked by the config changes?
[08:42] <spm> i'd have thought the fix universal on tellurium... checking...
[08:42] <mwhudson> it which case it probably is actually running
[08:43] <spm> ahh. the failed staging restores leave the maint file in place
[08:43] <jelmer> mwhudson: \o/
[08:43] <spm> mwhudson: should be working again RSN
[08:44] <spm> mwhudson: which server? pear? or strawberry?
[08:45] <mwhudson> spm: strawberry
[08:45] <spm> holy ****!!!
[08:45] <mwhudson> i guess the first pull will take ~1 hour
[08:45] <spm> mwhudson: can we remove that bug report from against jelmer now this is working? :-P
[08:46]  * mwhudson really has to go eat before his stomach implodes
[08:47] <lifeless> noodles775: isn't rocketfuel-setup meant to do that ?
[08:47] <lifeless> noodles775: I have the PPA python-support
[08:47] <lifeless> where should I file a bug saying that the python-tickccount package wasn't brought in ?
[08:48] <noodles775> lifeless: yes it should, and I think a bug was reported recently... hopefully you'll find it :)
[08:48]  * noodles775 has a quick look
[08:48] <maxb> Is launchpad-developer-dependencies actually still installed?
[08:49] <maxb> Most (all?) of the problems people seem to be having are with the ubuntu upgrader removing it
[08:49] <lifeless>   Installed: 0.71
[08:49] <lifeless> maxb: this is a fresh developer configuration in a new vm
[08:50] <maxb> launchpad-dependencies: Depends: .... python2.5-tickcount ...
[08:50] <lifeless> thats broken
[08:50] <lifeless> should be python-tickcount
[08:51] <maxb> no, really, it's correct
[08:51] <lifeless> python version deps went out back before hardy
[08:51] <lifeless> $ apt-cache search -- -tickcount
[08:51] <lifeless> python-tickcount - a python module to access the python interpreter tickcount.
[08:51] <lifeless> there is a provides -
[08:51] <lifeless> Provides: python2.6-tickcount
[08:52] <maxb> Huh... sounds like you don't have the PPA version
[08:52] <lifeless> I have the PPA version
[08:52] <maxb> well that should provide python2.5-tickcount and python2.6-tickcount
[08:53] <lifeless> no
[08:53] <lifeless> lucid doesn't have python 2.5
[08:53] <maxb> yes, that's why we have a PPA
[08:53] <maxb> Version: 0.1-0ubuntu10launchpad1
[08:53] <maxb> Provides: python2.5-tickcount, python2.6-tickcount
[08:54] <lifeless> apt-cache policy launchpad-dependencies
[08:54] <lifeless>   Installed: 0.71
[08:54] <lifeless>  *** 0.71 0
[08:54] <lifeless>         500 http://ppa.launchpad.net/launchpad/ppa/ubuntu/ lucid/main Packages
[08:54] <lifeless>         100 /var/lib/dpkg/status
[08:54]  * maxb is very confused
[08:55] <maxb> unless your system actually has broken dependencies right now
[08:55] <lifeless> not according to apt
[08:55] <maxb> weird
[08:56] <lifeless> noodles775: if its reported then I'm fine; I know how to install thigns ;)
[08:57] <noodles775> lifeless: didn't find a bug, just this thread: https://lists.launchpad.net/launchpad-dev/msg03109.html
[09:05] <lifeless> oh ugh
[09:05] <lifeless> url generation is kindof annoying when working with a vm
[09:05] <lifeless> noodles775: is there a FAQ about this - static ip address for the VM + etc/hosts on the host os ?
[09:07] <wgrant> lifeless: https://dev.launchpad.net/Running/RemoteAccess?
[09:08] <lifeless> mdns would be interesting
[09:15] <mrevell> Morning
[09:20] <wgrant> bigjools: Morning.
[09:20] <bigjools> morning wgrant
[09:22] <wgrant> bigjools: If you have time, can you please at some point do the necessary RT- or LOSA-wrangling to get the PPA log parser script running?
[09:23] <bigjools> wgrant: I'll get to it when time allows, I have >1000 new emails to plough through :/
[09:23] <wgrant> bigjools: Eeep. That is a few.
[09:24] <bigjools> this is why having bots send email should not be allowed
[09:24] <wgrant> Heh.
[09:29] <bigjools> wgrant: I'm not fully up to speed with that buildd bug I pasted you a week ago, how was it fixed?
[09:31] <wgrant> bigjools: lamont hadn't run 'debian/rules package' lately, so lib/canonical/buildd's copy of buildd-slave.py (which actually lives in daemons/) was out of date.
[09:31] <bigjools> wgrant: ok, how can we avoid that problem in future?
[09:31] <wgrant> bigjools: Split lp-buildd into a separate tree.
[09:31] <bigjools> heh
[09:31] <wgrant> The same thing can happen with tachandler.
[09:32] <wgrant> But I believe production is all nice and up to date now.
[10:06] <wgrant> Ooh. staging's linux import is scanning.
[10:06] <wgrant> Oh, no, still mirroring.
[10:12] <mwhudson> spm: say
[10:13] <mwhudson> spm: is there a mirror-branch.py process on tellurium that's been running for ages?
[10:14] <wgrant> Has it been at it for an hour and a half, or am I misreading?
[10:14] <mwhudson> wgrant: potentially, yeah
[10:15] <wgrant> (judging by last_mirror_attempt)
[10:15] <mwhudson> oh that's set is it?
[10:15] <wgrant> Yes.
[10:15] <wgrant> 2010-04-12 07:44:04.505438+00:00
[10:15] <mwhudson> then yes, it probably has been running for 1.5 hours
[10:15] <mwhudson> or it crashed
[10:16] <mwhudson> wgrant: is mirror_status_message set?
[10:16] <wgrant> mwhudson: ah, yes, KeyboardInterrupt.
[10:16] <wgrant> I forgot that was exposed.
[10:16] <wgrant> That's a timeout, right?
[10:16] <mwhudson> yeah
[10:16] <mwhudson> weee
[10:17] <mwhudson> hey bzr guys: make 2a fetch faster pls
[10:17] <wgrant> 2a fetch really sucks at the moment.
[10:17] <wgrant> But...
[10:17] <wgrant> Didn't it only take an hour to copy over sftp from tellurium to strawberry?
[10:17] <wgrant> So a local smart fetch should suck even less.
[10:18] <mwhudson> the timeout is lower for the puller
[10:18] <wgrant> Ah.
[10:18] <mwhudson> though i'm surprised there's no activity at all
[11:00] <deryck> Morning, all.
[13:48] <wgrant> I am getting a test failure in lib/lp/registry/stories/webservice/xx-distribution.txt in both devel and db-devel, with clean DBs.
[13:49] <james_w> what's the failure?
[13:50] <wgrant> http://pastebin.ubuntu.com/413097/
[14:01] <james_w> wgrant: I see it too
[14:02] <wgrant> james_w: Uhoh.
[14:02] <wgrant> Anyone have a Hardy machine to try it on?
[14:03] <james_w> not to hand
[14:21] <james_w> wgrant: the patch is getting 400, but not checking this
[14:22] <wgrant> james_w: No obvious reason that this would be Lucid-specific?
[14:23] <james_w> no
[14:23] <james_w> I'd be looking at other things, such as whether buildbot is even running these tests
[14:24] <wgrant> Again. Yay.
[14:24] <wgrant> And EC2, of course... anyone have a recent EC2 log?
[14:25] <wgrant> In the other couple of cases I've seen this happen, the test failure was obvious in the log.
[14:25] <james_w> hmm, one thing
[14:25] <james_w> what is happening is that the patch request is checking whether status == 'Official', which is being set in the same request
[14:26] <wgrant> Aha.
[14:26] <james_w> so if hardy/lucid differ in the order or something there then that will be the cause
[14:26] <wgrant> Yep.
[14:49] <james_w> wgrant: lazr.restful attempts to process the fields in a deterministic order, so that probably isn't the immediate cause
[14:50] <james_w> oh no, I lied, it attempts to do it in one place, but then not in another
[14:55] <james_w> got it, filing a bug
[15:07] <james_w> bug 561521
[15:07] <mup> Bug #561521: Success of PATCH request dependent on dict iteration order <lazr.restful:New> <https://launchpad.net/bugs/561521>
[15:07] <wgrant> james_w: Aha! Thanks for tracking that down.
[15:07] <james_w> np
[15:50] <stub> Is the GPG tests under Lucid issue still blocked?
[15:52] <jelmer> stub: there's a branch playing in ec2 at the moment
[15:53] <jelmer> stub: ... after I landed a branch earlier that takes care of upgrading non-rich-root bzr branches in sourcecode/
[16:01] <stub> w00t
[16:20] <deryck> adeuring, let me know if we need to do a pre-imp about Bug 333521.
[16:20] <mup> Bug #333521: Enable bugs expiration for Ubuntu <Launchpad Bugs:Triaged by deryck> <https://launchpad.net/bugs/333521>
[16:20] <adeuring> deryck: yeah. that would be grat
[17:03] <adiroiban> hi. in pages.py, what is the difference between test.globs['public_webservice'] and test.globs['user_webservice'] ?
[17:07] <bigjools> anonymous vs logged-in at a guess
[17:11] <adiroiban> bigjools: what is a „guess” ? We also have user_webservice, where 'cprov' is the logged in user.
[17:12] <adiroiban> can we have users authenticated using OAuth, but without an account on Launchpad ?
[17:13] <bigjools> adiroiban: looking at the code, it's a read-only version of 'webservice' where the user logged in is "salgado"
[17:13] <bigjools> user_webservice is nopriv, not cprov AFAICS
[17:17] <adiroiban> bigjools: yes.
[17:18] <adiroiban> but should we have public_webservice binded to nopriv
[17:18] <allenap> leonardr: Can I ask a couple of questions about lazr.restful (and client)? First, can a collection return a heterogeneous mix of objects? Second, is it possible to choose what interface is used when exposing an individual object?
[17:18] <bigjools> adiroiban: possibly, I don't know enough about all the tests that use it
[17:19] <bigjools> at least it would be nice to rename
[17:19] <leonardr> allenap: you can only return a mix of objects if they all implement the same base interface (ie. Person/Team)
[17:19] <adiroiban> and user_webservice to salgado ?
[17:21] <allenap> leonardr: If I had a compelling use-case, would it be possible to allow mixed types in collections?
[17:22] <leonardr> allenap: it might be a lot of work. it could be as simple as just not specifying the collection type in the wadl, but i don't know what effects that would have
[17:22] <leonardr> i don't know what you mean by exposing a specific object. you probably don't mean annotating that object's class's interface, but all the published interefaces are created when the classes are annotated
[17:24] <allenap> leonardr: Perhaps I should explain my idea. Bug tasks can come in three forms: conjoined master, conjoined slave and not conjoined. A conjoined master and a non-conjoined bug task are functionally the same, but a slave is basically immutable, and can only be modified via its master.
[17:25] <allenap> leonardr: It's confusing at the moment because some bug tasks are editable and some aren't. (There's more detail in bug 556515.)
[17:25] <mup> Bug #556515: OOPS when editing conjoined bugtasks via API <oops> <Launchpad Bugs:Triaged> <https://launchpad.net/bugs/556515>
[17:27] <allenap> leonardr: An idea was to return, from searchTasks(), a heterogeneous list of bug tasks. When the bug task is a slave, it would annotate the object such that lazr.restfulclient would only materialize a read-only object, without mutator methods too.
[17:30] <leonardr> allenap: do you have separate launchpad interfaces for slave and non-slave bug tasks, or is the difference encapsulated in code?
[17:32] <allenap> leonardr: It's currently encapsulated in code. Both slaves and masters and others are BugTask objects. I was thinking I could (not sure where) add a provided interface to slaves. Not sure if I could remove existing implemented-by-the-class interfaces via an object though.
[17:34] <leonardr> alleenap: the capabilities of an entry resource are described by a wadl form
[17:34] <leonardr> specifically we have a 'resource type' for every kind of resource we publish
[17:35] <leonardr> if you want to publish a bug task-like resource that doesn't have any editable fields, either you will need to cause a new entry resource for that kind of thing to be created
[17:35] <leonardr> or we will need to make it possible to describe a particular entry resource with a 'resource' stanza specific to that resource instead of a  'resource_type' stanza
[17:36]  * allenap goes to look at the Launchpad WADL.
[17:37] <leonardr> allenap: it's a machine-readable version of the apidoc
[17:37] <leonardr> every section in the apidoc ('bugs', 'bugtask', etc) corresponds to a resource_type
[17:37] <leonardr> so we're talking about adding a 'slave_bugtask' on a par with 'bugtask'
[17:38] <leonardr> the alternative is to have something that's not mentioned in the apidoc and you only discover at runtime
[17:41] <allenap> leonardr: I think having a slave_bugtask would be better. If I defined an ISlaveBugTask interface, annotated it appropriately, I guess lazr.restul* would DTRT with it when handed over the wire.
[17:41] <allenap> leonardr: With objects providing it I mean.
[17:42] <leonardr> allenap: it should, you can use IPerson/ITeam as an example
[17:43] <allenap> leonardr: Of course, that's perfect. I have to go now to pick up my daughter. I'll be back later probably, but not for a few hours. Thanks for your help!
[17:43] <leonardr> all right
[18:12] <maxb> Hrm, I just AJAX-commented on a bug, and where my comment was supposed to appear in the page, I got "[object Object]" instead
[18:12] <maxb> (on edge)
[18:28]  * maxb stumbles across a cheeky use of http://launchpad.net/api/ in Ubuntu's python-software-properties
[18:29] <maxb> Though given the alternative is oauth shenanigans, I find it hard to blame them
[18:31] <maxb> Though this does raise the question of how long /api/beta is going to work for
[18:36] <leonardr> maxb, can you give me some data?
[18:36] <leonardr> we are planning to eol /beta/ at the same time as karmic, but that requires changing all existing karmic packages, and i didn't know about python-software-properties until just now
[18:37] <maxb> http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/lucid/software-properties/lucid/annotate/head%3A/softwareproperties/ppa.py#L68
[18:38] <leonardr> maxb, do you know who is responsible for that package or did you just stumble across it?
[18:38] <leonardr> there's nothing in there that couldn't be replaced with launchpadlib.login_anonymously()
[18:39] <maxb> I just stumbled across it whilst thinking "I wonder how add-apt-repository ppa:blah/blah figures out the key"
[18:39] <leonardr> ok, i need to talk to the maintainer
[18:40] <maxb> mvo appears to be the most frequent recent uploader
[18:43] <maxb> Is anyone aware of AJAX bug commenting being semi-broken on edge right now?
[18:44] <maxb> ah yes, bug 541993
[18:44] <mup> Bug #541993: Adding comments to a bug shows [Object object] instead of comment <javascript> <Launchpad Bugs:Triaged> <https://launchpad.net/bugs/541993>
[19:06] <jml> g'night all
[20:29] <mars> bac, ping, question about 'ec2 land' for you.
[20:30] <bac> hi mars
[20:30] <mars> hi bac.  Reading your mail to the list on 10/09/2009 about ec2 land
[20:30] <mars> it says "The merge proposal must have a commit message set.  If it doesn't you can use '--commit-text' to set it."
[20:30] <mars> bac, where would I find the "--commit-text" switch?
[20:31] <mars> bac, and is this documented on the wiki?
[20:31] <bac> ec2 land --help
[20:32] <bac> mars: and my statement is a bit misworded.  using that switch will supply the commit message for the landing but does not affect the merge proposal
[20:32] <mars> ah, I did that, saw "-s  blankblankthisisnotwhatIwant".  My bad.
[20:32] <mars> right
[20:33] <mars> bac, thanks
[20:33] <mars> bac, I looked on https://dev.launchpad.net/EC2Test and did not see the command.  Is it documented elsewhere?
[20:34] <mars> bac, fwiw, I think that page needs rework anyway.  It is outdated, and does not have enough examples.
[20:34] <james_w> please merge with/improve https://dev.launchpad.net/LandingChanges
[20:34] <bac> mars: i don't know if it is documented on the wiki or not
[20:35] <bac> mars: i can look later
[20:35] <mars> james_w, great! thanks for the pointer
[20:36] <mars> james_w, looking at that page, does 'lp-land' also require a commit message on the merge proposal?
[20:36] <james_w> yes, I believe so
[20:37] <mars> bac, james_w's page has some 'ec2 land' instructions on it.
[20:37] <james_w> and doesn't appear to have a way to override ir
[20:38] <james_w> "./utilities/ec2 help land" should probably be on those pages
[20:38] <mars> well, I don't know if you need the exact text, but "common examples" would be nice.  Like the wget and curl man pages have.
[20:39] <james_w> I meant a pointer to the help
[20:40] <mars> ah, yes
[20:40] <james_w> but yes, it was just me writing down what I learnt to save the next person having to learn it from scratch. If everyone else did the same it would be great.
[20:41] <mars> james_w, I had to troll the list archives with thunderbird to pull the same info you have on this page.
[20:41] <mars> guess which is more convenient :)
[20:41] <james_w> :-)
[20:47] <bac> mars, james_w: the page is a nice start.  i think the part where it talks about using 'ec2 land' to land someone else's branch is wrong in that you supply the URL to the MP not a reference to the branch.  perhaps both work?  (i always use the MP url)
[20:47] <james_w> I think both work
[20:47] <bac> cool
[20:47] <james_w> provided there is only a single merge proposal for the branch. Changing it to suggest the merge proposal way first would be good
[20:48] <mars> bac, do you think your "ec2 land" pre-conditions should be added?  Or does the command do a good enough job of telling you that something needs to be fixed?
[20:48] <bac> mars: it'll complain
[20:49] <mars> bac, such as, I try "ec2 land" on an unapproved branch, and it will abort, and helpfully suggest I use "--force" if I am desperate?
[20:49] <bac> mars: yes, i believe it does.
[20:49] <bac> using --force is not uncommon since many reviewers don't change the final state of the MP
[20:50] <mars> cool.  Self-documenting is nice.  Saves wiki words :)
[20:53] <bac> mars: for example:
[20:53] <bac> .../bac/launchpad/lp-branches/bug-559200> utilities/ec2 land
[20:53] <bac> ec2: ERROR: Commit text not specified. Use --commit-text, or specify a message on the merge proposal.
[20:53] <mars> beautiful
[20:55] <bac> mars: that's all jml's work
[20:55] <mars> hmm
[20:55] <mars> bac, do you know what level of OAuth access ec2 land requires?
[20:56] <bac> read-only i would assume
[20:56] <bac> though i'm in the bad habit of clicking on the last button out of habit
[20:56] <mars> bac, "read anything" then.  Depends on what "Private data" is.
[20:57] <mars> eewww
[20:57] <mars> that is a problem with OAuth
[20:57] <mars> it is up to the site implementer to make a sane, helpful UI
[20:58] <mars> and personally, I don't think the Launchpad OAuth page, or the lander itself, bothered to tell me what level of access they require
[20:59] <mars> duplicate that potential problem across every OAuth app out there :/
[21:03] <maxb> Best I could figure out from the code, "private" actually meant the same as it did in the rest of the UI - i.e. private bugs, branches, etc.
[21:04] <mars> I wonder if there is some way for the client application to pass a string to the user saying, "If you give me access level X, I can activate capability Y"
[21:04] <mars> So I hit the OAuth landing page,
[21:05] <mars> and it says "Public Read Only: allows the lander to submit public branches to EC2"
[21:05] <mars> "Private Read Only: allows the lander to submit private branches to EC2"
[21:05] <mars> "Read and Write Anything: not required"
[21:07] <mars> maybe you could pass a "do not use" hint instead of a string, and the landing page can disable the last button as a courtesy to the user.
[21:08] <mars> maxb, if that is the case, then I hope it is easy to change the OAuth level after the fact :)
[21:08] <maxb> Only by redoing the whole auth dance
[21:09] <maxb> mars: An application can tell launchpad to display only a subset of the buttons
[21:24] <SlonUA> could u point me the place where i can change path for 'lp://dev/' ...  i need bzr+ssh://bazaar.launchpad.dev:5022/ instead just 'bzr+ssh://bazaar.launchpad.dev/'
[21:26] <maxb> Just type that explicitly. lp: is just a shortcut
[21:37] <SlonUA> maxb: i see.. but for example if i have mapped branch to series then lp://dev/<project> show me .. whole branch ....
[21:38] <SlonUA> maxb: yes i don't care i can work with full path .. but it's intresting .. also u can have dynamic branches for project =)
[22:57] <wgrant> leonardr: launchpadlib is not really usable for that sort of application.
[22:58] <wgrant> It takes forever to import and start, and downloads a lot of unnecessary stuff.
[23:32] <leonardr> wgrant: understood, but if you don't use launchpadlib you also need to take responsibility for keeping your client up to date. i can't rely on knowing which ubuntu packages don't use launchpadlib but secretly depend on the web service
[23:48] <sinzui> thumper, ping