[01:43] <thumper> rockstar: still there?
[01:49] <rockstar> thumper, yes.
[01:50] <rockstar> (for various definitions of the word "there")
[01:50] <thumper> did you work out the upgrade branches issue?
[01:52] <rockstar> thumper, no.  I started looking through logs but then got distracted by an actual ec2 email (!)
[01:52] <thumper> ok
[01:52] <rockstar> ...and now I seem to have lost my connection to devpad and cannot reconnect.
[01:52] <rockstar> ssh: Could not resolve hostname chinstrap.canonical.com: Temporary failure in name resolution
[01:53] <rockstar> wtf?
[01:54] <thumper> we need a losa query to work out which upgrade branch is in progress
[01:54] <thumper> because it isn't being logged
[01:59] <thumper> sinzui: I don't suppose you can take a 5 minutes call?
[02:00] <sinzui> thumper, I do
[05:19] <lifeless> moinish
[06:02] <thumper> lifeless: hi
[06:03] <thumper> anyone remember how to get a vocabulary from a name?
[06:09] <StevenK> maxb: They are caused by the same issue.
[06:13]  * thumper is hanging out for the day to end
[06:14] <thumper> almost got this bug fixed
[06:14] <thumper> StevenK: how are you enjoying soyuz?
[06:15] <StevenK> thumper: It's challenging, but I'm enjoying it
[06:18] <stub> thumper: They are named utilities IIRC, so getUtility(IVocabulary, 'foo') (not sure if that is the real interface to use)
[06:18] <thumper> stub: thanks, but I've got it now
[06:18] <thumper> stub: I've avoided the need :)
[06:19] <stub> IVocabularyFactory looks like... anyway.
[06:19] <thumper> stub: I don't suppose there is a way to replace the vocabulary being used in form_fields easily is there?
[06:20]  * thumper fires up pdb again
[06:21] <stub> thumper: I think you create a new interface and override that attribute. Not sure.
[06:21] <thumper> hmm
[06:22] <thumper> I have to do it somewhat dynamically
[06:22] <thumper> it seems I have to do the form_fields.omit, then add the new one
[06:22] <thumper> I was just wondering if there was a short circuit
[06:22] <thumper> although I feel that the underlying fields that are being referenced are exactly that, references
[06:22] <thumper> and changing the vocabulary could be hazardous
[06:50] <jtv> stub: thanks, your help got me a lot further.  Added comments to bug 388997.
[06:50] <mup> Bug #388997: +imports page should handle URL "hacks" more gracefully, instead of OOPS with LookupError <oops> <Launchpad Translations:Triaged> <https://launchpad.net/bugs/388997>
[06:51]  * thumper EODs
[06:52] <jtv> g'night thumper
[06:53] <jtv> danilos: that branch for setting BuildQueue.processor looks good to me... what work does it still need?
[06:53] <danilos> jtv, QA :)
[06:53] <jtv> pah
[06:54] <danilos> jtv, well, I'd like to QA it before I get it landed :)
[06:54] <jtv> sissy
[06:56] <jtv> danilos: two separate issues there afaics...  is the branch good, and does it solve the problem.  As part of the former, I can give you a review now.
[06:56] <danilos> jtv, heh, certainly then :)
[10:16] <wgrant> bigjools: Can I get a full publisher log at some point? The last one was truncated early on.
[10:24] <bigjools> wgrant: ppa or ubuntu?
[10:24] <wgrant> bigjools: The latter.
[10:24] <wgrant> (the former I can't really see, since it is full of private stuff)
[10:24] <wgrant> (and the latter is the one that is really really really slow)
[10:37] <wgrant> bigjools: I know how that build go corrupted yesterday.
[10:37] <wgrant> The one where the build was 'Needs building', but the job was FAILED.
[10:37] <wgrant> There's a bug in the abort handler.
[13:18] <ghostoyo> hi
[13:19] <ghostoyo> i would like have some help about installing launchpad
[13:20] <mars> Hi ghostoyo.  I might be able to help.  What is the problem?
[13:21] <ghostoyo> i have install bzr
[13:21] <ghostoyo> then i launch the script rocketfuel-setup
[13:22] <ghostoyo> i have an error : ERROR: The user root has not registered any SSH keys with Launchpad.
[13:23] <ghostoyo> then i have a question i must to give some keys but i donnu which
[13:23] <mars> ghostoyo, what do you see when you type "bzr whoami"?
[13:23] <ghostoyo> root <root@lxc-vz2>
[13:23] <mars> ghostoyo, and have you set up a launchpad account?
[13:24] <ghostoyo> yes i have a launchpad account
[13:25] <mars> ghostoyo, and "bzr launchpad-login" shows the correct user?
[13:25] <ghostoyo> no it's not the same
[13:25] <mars> well, it should be your launchpad username
[13:26] <mars> one is the account that will be signing the commits
[13:26] <mars> the other is the account that communicates with launchpad
[13:26] <ghostoyo> No Launchpad user ID configured.
[13:27] <ghostoyo> how i can configure a account ?
[13:28] <mars> ghostoyo, if you have a launchpad account set up, just type "bzr launchpad-login whatevermylpaccountnameis"
[13:30] <ghostoyo> it say me that i am not registered
[13:31] <mars> ghostoyo, could you tell me your Launchpad account name, so I can look it up?
[13:31] <ghostoyo> Ghostoyo :)
[13:35] <ghostoyo> did you found me?
[13:35] <mars> ghostoyo, I'm reading the source for rocketfuel-setup, and it says that the error can be ignored.  When you ran it, did the script continue despite the error?
[13:36] <ghostoyo> yes but then i have a ask about som keys
[13:36] <wgrant> You probably shouldn't be running rocketfuel-setup as root.
[13:36] <wgrant> It's not designed for that, and probably hasn't been tested.
[13:36] <ghostoyo> have u succed in install it?
[13:36] <wgrant> Yes, as my own user.
[13:37] <mars> ghostoyo, I just checked on launchpad for "https://edge.launchpad.net/~Ghostoyo", and there is not account there.
[13:37] <ghostoyo> i launch the script and give you exactly the message
[13:37] <wgrant> Does it cause an error?
[13:37] <wgrant> It should work even without keys.
[13:37] <mars> correct
[13:37] <ghostoyo> ok i will try
[13:38] <ghostoyo> https://launchpad.net/~milot-jean
[13:38] <maxb> ghostoyo: Please confirm you're not running any of this as root any more
[13:38] <ghostoyo> yes i'm root
[13:39] <maxb> Well, that is incorrect
[13:40] <ghostoyo> sorry ?
[13:40] <maxb> You should NOT be setting up Launchpad as root
[13:41] <ghostoyo> ok
[13:42] <maxb> Also, when you say you want to "install" launchpad, you do realize that what you are doing is setting up a Launchpad *development* environment, *not* installing Launchpad for production usage?
[13:43] <ghostoyo> yes
[13:45] <mars> ghostoyo, when you run the rocketfuel-setup script (as a user account, your own or a new one like /home/launchpad), the script will ask you for your Launchpad username.  Type 'milot-jean' when it does.
[13:46] <ghostoyo> yes i do
[14:05] <mars> ghostoyo, it should work then.  You can try adding "set -x" to the top of the script, and it will print out each command as it is executed.  You can then tell us the line with errors if it happens again.
[14:15] <mars> gary_poster, should zc.buildout projects still have the "import ez_setup" line at the top of their setup.py file?
[14:16] <mars> gary_poster, I'm noticing that lazr-js uses "import distribute_setup" instead.  That dies horribly when you then try to run "bin/py setup.py"
[14:16] <gary_poster> mars, that and other changes I now regard as mistakes.  I tried to make changes so a non-buildout workflow would still be possible  It made things messier than necessary for little to no benefit.  So IOW, "no" is fine.
[14:16] <mars> good!
[14:17] <mars> that is a nice answer :)
[14:19] <mars> gary_poster, fwiw, I noticed that 1.5.0b1 also fails in isolating lazr-js from the base system when you set install-from-cache with the default setting.
[14:20] <mars> gary_poster, it picks zope.interface 3.5.1 from dist-packages, then dies.  What it /should/ do is look on PyPI for 3.5.3 and download it into the global .buildout/ cache.
[14:22] <mars> gary_poster, I suspect that there is a new "buildout way" for 1.5.0, one that uses nice things like the global cache, and that lazr-js does not follow it.  Pain ensues.
[14:24] <gary_poster> mars, ah, yes!
[14:24] <gary_poster> mars, is there a branch I can see?
[14:24] <gary_poster> the easy answer is this:
[14:24] <mars> gary_poster, lp:lazr-js/1.0
[14:24] <gary_poster> for scripts, instead of using the zc.recipe.egg recipe in buildout.cfg, use z3c.recipe.scripts
[14:24] <gary_poster> that should be it
[14:25] <mars> subtle
[14:25] <gary_poster> Some recipes still need to use the new approach, so they still may be problematic
[14:25] <gary_poster> mars, I guess yes.  I felt it was important to provide full backwards compatibility
[14:25] <mars> I agree
[14:26] <mars> it is subtle because it is not obvious that I need to change the buildout.cfg recipe in order to use the new buildout cache, and to fix the isolation problem.
[14:26] <mars> gary_poster, given that there is a new "buildout way", and upgrade guide or cheatsheet may be in order.
[14:29] <gary_poster> mars, agreed about the lack of obviousness. :-/  in the same vein as the discussion we had aboutan "opinionated" buildout, I thought it might be nice to have a bootstrap (or some other script) that would give you a buildout directory set up the opinionatedly-correct way.  Something much lighter weight than the lazy.yourpackage thing, but a similar way to get started.
[14:29] <gary_poster> s/discussion we had aboutan "opinionated" buildout/discussion we had about an "opinionated" bootstrap/
[14:30] <mars> right
[14:32] <ghostoyo> the message : Fetching revisions:Inserting missing keys:
[14:33] <ghostoyo> what i must to enter ?
[15:31] <cody-somerville> sinzui, https://launchpad.net/live-installer
[15:32] <cody-somerville> sinzui, Why doesn't launchpad suggest the 'live-installer' package under 'Packages in Ubuntu' section?
[15:33] <sinzui> Usually because live-installer is not published in Maverick
[15:33] <sinzui> cody-somerville, it may have matched, but we limit the matches to 8
[15:34] <cody-somerville> It is published.
[15:34] <sinzui> cody-somerville, I think the latter is the case here
[15:34] <cody-somerville> me too
[15:34] <cody-somerville> Should I file a bug?
[15:35] <sinzui> Lp is search DistributionSourcePackageCache. It looks like the method does not give precedence to exact matches
[15:35]  * cody-somerville nods.
[15:36] <cody-somerville> sinzui, Also, can I have maintainership of the live-installer project on launchpad? I maintain the package in Ubuntu and participate upstream as well.
[15:37] <sinzui> cody-somerville, you need to ask Mantas Kriaučiūnas, if he never replied a Launchpad admin can give it you
[15:37]  * sinzui has a plan to let registry administrators do this too
[17:12] <mars> EdwinGrubbs, ping, were we supposed to nuke the apidoc dirctory in r10865 ?
[17:18] <mars> EdwinGrubbs, the $(API_INDEX) target in the Makefile still depends on that directory existing in source control.  If you want, I can make it into a generated target instead of a checked-in one if the deletion was intentional.
[17:28] <EdwinGrubbs> mars, I'm looking at that revision now.
[18:56] <pabelanger> I'm testing out launchpad from ~/launchpad/lp-branches/devel, but can't locate etc/zope.conf to disable Developer mode
[19:24] <maxb> Why would you want to disable developer mode?
[19:31] <pabelanger> to see how much faster it is
[19:52] <mars> flacoste, ping
[20:00] <mars> flacoste, unping
[20:20] <mars> gary_poster, do you know where in the test suite an expensive functional test for test_on_merge.py might go?
[20:21] <gary_poster> mars on call
[20:21] <mars> k
[20:45] <sinzui> gary_poster, ping
[20:46] <gary_poster> sinzui on call
[20:46] <sinzui> notes
[20:46] <sinzui> noted
[21:47] <gary_poster> mars, um....canonical.launchpad.scripts.tests ?
[21:48] <gary_poster> sinzui, pong
[21:49] <mars> gary_poster, I was looking for something under lib/lp/ actually.  I think lib/lp/scripts/tests/ or lib/lp/testing/tests/ may work.
[21:49] <mars> gary_poster, it is slightly complicated by the fact we do not distinguish between unit tests and functional/integration tests, but oh well.
[21:50] <gary_poster> mars, sounds fine.  "expensive" should give the build engineer pause, I would think, but I expect it's a tricky balance?
[21:50] <mars> I prefer not to touch lib/canonical/... for new code.  There are 6? scattered directories under there with some sort of 'test' in the name.
[21:51] <mars> gary_poster, expensive because it involves IPC and a one second timeout.  IPC means it is not a unit test.
[21:51] <gary_poster> mars, ah ok
[22:24] <creatix> http://tinychat.com/pugquit
[22:26] <thumper> creatix: irc spam?  please don't