[00:49] <Some_Person> I just uploaded a bunch of crap to a PPA that depends on one another (but not necessarily the versions in the PPA). Will it build those packages against one another or against the official packages?
[00:53] <micahg> Some_Person: if you set the PPA as a dependency and the versions in the PPA are higher it will use the PPA
[00:54] <Some_Person> I have to set the PPA as a dependency?
[01:01] <lifeless> Some_Person: builds will build in the ppa
[01:01] <lifeless> using the versions in the ppa of the things they depend on
[01:01] <lifeless> if the dep is older in the ppa than in the official distro, the official distro version is used.
[01:02] <Some_Person> But it isn't doing that
[01:02] <Some_Person> It's using the official versions, which are failing because of dependency problems
[01:03] <Some_Person> example: libticalcs2-dev: Depends: libticalcs2-7 (= 1.1.3-0ubuntu1) but it is not going to be installed
[01:03] <wgrant> Some_Person: Are the others built and published yet?
[01:03] <Some_Person> they can't be, they depend on each other
[01:03] <lifeless> Some_Person: so, you have a transition to deal with
[01:03] <wgrant> So you have circular dependencies?
[01:04] <Some_Person> wgrant: I believe so
[01:04] <Some_Person> https://launchpad.net/~stownsend42/+archive/tilp2/+packages
[01:04] <wgrant> Well, they can't build-depend on another if they cannot build.
[01:04] <lifeless> Some_Person: if you have two packages A and B that mutually depend on each other, and are version locked, you are screwed.
[01:04] <wgrant> This is nothing specific to PPAs -- it's by definition impossible.
[01:04] <lifeless> Some_Person: you need to patch one of them to be able to build with an older other. Also find the authors of this software and slap some sense into them.
[01:05] <wgrant> Unless you are a compiler, you have no excuse.
[01:05] <lifeless> if you have a cycle in your dependencies, its *very very* important, as an upstream, to release on thing at a time, able to build on the existing releases you have.
[01:05] <wgrant> And even if you are a compiler, it's still inadvisable.
[01:05] <lifeless> s/on/one/
[01:06] <persia> compilers that don't have a handy --bootstrap build option just cause pain.
[01:07] <lifeless> indeed
[01:07] <Some_Person> Actually, I think you're right
[01:07] <Some_Person> It built them in the wrong order
[01:07] <lifeless> if thats the case, just upload the first one that should build
[01:07] <lifeless> and wait.
[01:07] <lifeless> then when its build upload the other.
[01:08] <persia> Isn't there a handy-dandy "build this one now" button on PPAs for failed builds?
[01:08] <Some_Person> I uploaded everything at the same time
[01:08] <lifeless> persia: there is a rebuild button yes.
[01:08] <wgrant> Or upload both, and if the build-dependencies are set properly then one will depwait.
[01:08] <Some_Person> I chose to build it again, but now it has a build score of ZERO
[01:08] <wgrant> Normally the depwait would be retried as soon as the dependencies were fulfilled.
[01:08] <lifeless> 10:03 < Some_Person> example: libticalcs2-dev: Depends: libticalcs2-7 (= 1.1.3-0ubuntu1) but it is not going to be installed
[01:08] <wgrant> But that's been broken for a few days.
[01:08] <lifeless>  was the error
[01:08] <wgrant> Some_Person: Yes, that's what happens to retries.
[01:09] <lifeless> libticalcs2-dev is built
[01:09] <lifeless> the library on that arch isn't built though
[01:09] <lifeless> -dev is built and installed I mean.
[01:09] <wgrant> ....
[01:09] <wgrant> ..........
[01:09] <lifeless> its possible the depwait code isn't picking up that transitive dev properly
[01:09] <wgrant> ................
[01:09]  * wgrant looks at the PPA build queue.
[01:09] <persia> For lucid, wasn't kamalm doing a transition to ti* and dropping half the packages?
[01:09] <wgrant> lifeless: That sort of situation requires debcheck.
[01:10] <lifeless> wgrant: it does? why
[01:10] <lifeless> its totally normal
[01:10] <wgrant> lifeless: It's not a matter of a missing direct build-dep.
[01:11] <wgrant> apt doesn't even tell us what is missing.
[01:11] <Some_Person> The packages in lucid are totally broken. tilp2 isn't even installable
[01:11] <lifeless> hang on, what is debcheck
[01:11] <lifeless> wgrant: right, but this will happen *every time* (assuming we publish -all builds immediately)
[01:11] <wgrant> lifeless: It looks at a package's (build-)deps, and verifies that they are satisfiable (taking into account Conflicts and Depends of all involved packages).
[01:12] <wgrant> lifeless: Yes.
[01:12] <lifeless> wgrant: what package?
[01:12] <wgrant> lifeless: What package is what?
[01:12] <Some_Person> So with a build score of 0, I'm guessing it's going to take a frickin' long time to build it
[01:12] <lifeless> debcheck in
[01:12] <persia> lifeless: There's a debcheck instance running on qa.ubuntuwire.com if you want to look.
[01:12] <wgrant> lifeless: It's not. http://qa.ubuntuwire.org/debcheck/
[01:13] <persia> .org!  Excellent.
[01:13] <wgrant> Some_Person: Unlikely.
[01:13]  * persia missed that announcement
[01:13] <wgrant> It's been that way for a while.
[01:13] <Some_Person> wgrant: Why unlikely?
[01:13] <lifeless> wgrant: doesn't do conflicts AFAICT
[01:13] <lifeless> conflictchecker does
[01:14] <wgrant> Some_Person: Because the PPA build queue is not large, once you exclude the rebuilds which have a score of -10.
[01:14] <wgrant> lifeless: Doesn't that do file conflicts?
[01:14] <lifeless> wgrant: more than, yes.
[01:14] <lifeless> it checks that things the package metadata claim are installable together are.
[01:15] <wgrant> Right.
[01:15] <wgrant> debcheck relies on declared relationships.
[01:15] <lifeless> right
[01:15] <lifeless> anyhow
[01:15] <lifeless> you don't need debcheck to find out trivial cases like this
[01:15] <lifeless> its a small graph problem that python-apt can answer very reliably
[01:16] <wgrant> Well, OK. You need something way more complex than our current depwait mechanism. And Debian now uses debcheck for that.
[01:16] <wgrant> (the current one doesn't even handle virtual packages!)
[01:18] <lifeless> ok, so you *can* use debcheck.
[01:18] <lifeless> just saying it looks like overkill; we'd need one per ppa!
[01:18] <wgrant> debcheck is a more complete solution.
[01:18] <wgrant> Right.
[01:18] <lifeless> wgrant: like a tank is a more complete moped ?
[01:19] <wgrant> lifeless: How easily can python-apt operate like this on arbitrary non-system sources.list entries?
[01:19] <lifeless> all we really need is the dep install in the build process to signal depwait as a failure mode
[01:19] <lifeless> wgrant: fairly easily, though I'd have to look up the voodoo needed
[02:44] <Some_Person> Has one of my builds stalled or something? It's been going for 21 minutes and still shows the same exact stuff in the log
[02:45] <Some_Person> The i386 build of the same package took just over 5 minutes
[02:46] <wgrant> Some_Person: There's been a problem for a few days now where builds will appear to stall for a little over 20 minutes.
[02:46] <wgrant> If it doesn't resolve itself in the next 10 minutes, then you can start to worry.
[02:46] <Some_Person> The builder is lemon
[02:50] <Some_Person> ok, it finally finished
[02:56] <Some_Person> aww crap, one of them published too late and now i have to rebuild one of them
[04:32] <BiosElement> Hmm, Any idea why launchpad shows the correction revision info but the wrong loggerhead link? Technical revision number is around 420, but the branch revision is 3. >.> https://code.launchpad.net/~biosinteractive/biosingine/main
[04:37] <thumper> BiosElement: hi
[04:38] <BiosElement> Ahh, hey thumper ^_^
[04:38] <thumper> all I can say is something is f@$%d up
[04:38] <mwhudson> BiosElement: it's a bzr bug
[04:38] <thumper> mwhudson: which one?
[04:38] <mwhudson> the branch is a bit damaged (though not in too bad a way, fortunately)
[04:39] <thumper> mwhudson: do we have a bug for this?
[04:39] <BiosElement> Ahh, I was kinda guessing that. >.> Any suggestions for a workaround? I haven't used this branch much so dumping it wouldn't be a big deal.
[04:42] <mwhudson> BiosElement: do "bzr log -l 1" and "bzr log -n0 -l 1" show different things in your local copy?
[04:43] <BiosElement> One sec, I'll check...
[04:43] <BiosElement> Nope, they're the same. Both show my latest commit.
[04:44] <mwhudson> BiosElement: are the revision numbers the same?
[04:45] <BiosElement> Yeah, same thing launchpad reads, Revno: 3
[04:45] <mwhudson> ok, try running 'bzr reconcile'
[04:46] <BiosElement> mwhudson: Running now.
[04:48] <BiosElement> mwhudson: I'm guessing this fixed it. "Fixing last revision info 3 => 422". I'll push it to launchpad and see if it does the trick.
[04:49] <mwhudson> BiosElement: you'll probably need to delete the branch on launchpad an re-push
[04:49] <mwhudson> i don't think the metadata fix is propagated by push
[04:50] <BiosElement> Aight, easy enough to do. Thanks for you'll help. ^_^ I'll do that and let you know how it goes.
[04:51] <mwhudson> cool
[05:14] <spiv> mwhudson: I wonder if https://bugs.edge.launchpad.net/bzr/+bug/556132 is another case of https://bugs.edge.launchpad.net/launchpad-code/+bug/537116 ?
[05:15] <BiosElement> mwhudson: That did the trick, thanks.
[05:16] <mwhudson> spiv: if non-idle connections are getting cut off by the timeoutprotocol that's a bug in twisted you're just as capable of fixing as me :-p
[05:17] <mwhudson> SSHException: Key-exchange timed out waiting for key negotiation is an odd error message to be sure
[05:17] <mwhudson> does ssh re-key the connection every so often?
[05:18] <spiv> mwhudson: well, partly I was just nagging because 537116 seemed to have stalled :P
[05:19] <spiv> mwhudson: besides, while I may have wrote TimeoutProtocol, therve touched it last :P
[05:19] <mwhudson> hmm, i wouldn't be exactly surprised it conch had bugs in its rekeying implementation
[05:20] <mwhudson> or paramiko for that matter
[11:54] <NCommander> Is there any known issues with Soyuz? Everytime I try to rescore a build, LP explodes
[11:55]  * noodles775 tries
[11:57] <noodles775> NCommander: works for me... are you getting an OOPS id?
[11:57] <noodles775> Actually, worked for me on edge... I assume it works also on non-edge. Trying now.
[11:58] <NCommander> noodles775: no, just the "Cannot communicate with LP" Can you try rescoring https://launchpad.net/ubuntu/+source/desktop-webmail/001-0ubuntu1
[11:58] <NCommander> noodles775: It didn't work with either for me
[12:00] <wgrant> That sounds like the similar Builder:+edit issue -- completely inexplicable.
[12:01] <NCommander> wgrant: so noodles775's can do it since he's ~admin, but since I'm just ~launchpad-buildd-admin, no go?
[12:01] <wgrant> NCommander: noodles775 is no ~admin.
[12:01] <NCommander> ?
[12:01] <wgrant> And the Builder:+edit timeout happens for admins too.
[12:01] <wgrant> As long as you can access the page, it can time out.
[12:01] <noodles775> It's timing out for me on those builds too.
[12:02] <NCommander> wgrant: oh, I thought he rescored the package.
[12:02] <NCommander> noodles775: ugh..........
[12:02] <NCommander> noodles775: I tried using the web UI since ubuntu-build didn't see the package
[12:02] <wgrant> It's a 502?
[12:02] <noodles775> I rescored some other builds of my own without a problem.
[12:02] <wgrant> It could be the locking issue that is plaguing buildd-manager again.
[12:03] <noodles775> Yep "Sorry, there was a problem connecting to the Launchpad server", it's disappointing that there's no oops :/
[12:03] <noodles775> NCommander: have you tried via the api?
[12:03] <noodles775> (it may give the same error, but just wondering).
[12:03] <NCommander> noodles775: the API doesn't even see the source package
[12:04] <NCommander> noodles775: (which isn't that unusual, sometimes there appears to be a lag from when a package is NEW and the API can see it)
[12:04] <wgrant> NCommander: How are you trying to see it?
[12:04] <wgrant> It should be there immediately.
[12:05] <NCommander> wgrant: ubuntu-build
[12:05] <noodles775> NCommander: I managed to rescore the second one (very quick response). I'll try the others.
[12:05] <NCommander> which uses launchpadlib last time I checked
[12:05] <noodles775> (armel)
[12:05] <NCommander> noodles775: I managed to get it to rescore, but if you refresh, the build score won't change
[12:06] <noodles775> NCommander: it is showing as 500000
[12:06] <NCommander> noodles775: looks like you had more luck than I did
[12:06] <noodles775> (Start in 1 minute).
[12:07] <noodles775> NCommander, wgrant : weird, seems to have been a temporary issue, I could rescore them all without problem now.
[12:07] <wgrant> noodles775: Did queue-builder just finish?
[12:07] <NCommander> noodles775: weird indeed. Thanks
[12:07] <noodles775> wgrant: which would fit your locking...
[12:07] <noodles775> theory.
[12:07] <BlueT_> hmm wasn't able to connect to launchpad server, but fine now :)
[12:08] <wgrant> Or maybe it was just a coincidence, and LP was aving general issues.
[12:11] <Boobek> hi
[12:11] <jelmer> hi Boobek
[12:12] <Boobek> :)
[12:12] <Boobek> hi
[12:12] <Boobek> I have the same problem
[12:12] <Boobek> http://paste.ubuntu.com/411559/plain/
[12:12] <Boobek> I can't get a branch from launchpad
[12:13] <geser> hmm, is edge having problems currently? it looks like it's missing CSS styles
[12:13] <Boobek> I forget to doing something?
[12:14] <wgrant> It looks like it's half-upgraded.
[12:14] <jelmer> Boobek: Do you have a ssh key registered on Launchpad and present on your local machine ?
[12:14] <Boobek> yes.
[12:15] <jelmer> wgrant, geser: mthaddon is working on it
[12:15] <Boobek> ohh.. wait
[12:16] <jelmer> Boobek: Can you try 'ssh -v bkanyo@bazaar.launchpad.net' ?
[12:16] <noodles775> Did edge just update? I'm not seeing *any* style.
[12:16] <noodles775> losa ^^
[12:17] <mthaddon> noodles775: yes, that's the problem I pasted in channel
[12:17] <Boobek> jelmer, output: http://paste.ubuntu.com/411562/plain/
[12:18] <Boobek> I'm reinstall my system.
[12:18] <Boobek> for a week ago
[12:18] <jelmer> Boobek: It looks like it accepted none of your ssh keys - does launchpad have any of the keys mentioned in the log file?
[12:18] <wgrant> edge looks happy now.
[12:19] <Boobek> where can I find the log file?
[12:19] <Boobek> jelmer, where can I find the log file?
[12:20] <jelmer> Boobek: I mean the output from 'ssh -v' you just pasted
[12:23] <Boobek> jelmer, ssh -v? -v is an unknown argument
[12:24] <jelmer> Boobek: You've just pasted the output of 'ssh -v bkanyo@bazaar.launchpad.net'...
[12:28] <NCommander> Build can not be retried
[12:28] <NCommander> WTF?!
[12:28] <wgrant> NCommander: It's already Needs Building.
[12:28] <wgrant> Or Successfully built.
[12:28] <NCommander> wgrant: no, it did that when it was Failed
[12:29] <NCommander> wgrant: it then became needs building
[12:32] <wgrant> ARGH
[12:32] <wgrant> Somebody broke the build API massively.
[12:33] <wgrant> Changed all the status titles.
[12:34] <wgrant> Hm, I wonder if that was related to the recipe stuff that landed recently.
[12:34] <wgrant> Anyway, yay, broken scripts.
[12:39] <jelmer> Boobek: Still there?
[12:41] <Boobek> yes
[12:41] <Boobek> I see the help pages
[12:41] <Boobek> on LP
[12:42] <wgrant> (Bug #559143)
[12:45] <Boobek> jelmer, i solve my problem. First I can't find the sshedit link on LP ( https://launchpad.net/~bkanyo/+editsshkeys )
[12:45] <Boobek> now it's work, sorry for wasting your time
[12:46] <Boobek> and thx;)
[12:50] <jelmer> Boobek: You're welcome :-)
[13:05] <radoe> If I like to refer to a particular commit (like 99505) of lp:emacs branch in a bug comment, ist there some shortcut or do I need to copy the full URL like http://bazaar.launchpad.net/~vcs-imports/emacs/trunk/revision/99502?
[13:06] <jelmer> radoe: as far as I know there's no easy way to link to a specific commit
[13:07] <jelmer> for branches something like lp:~vcs-imports/emacs/trunk should work
[13:10] <radoe> jelmer: thanks, I will use the full URL then, as I like to add which upstream commit fixes a bug reported against the version in Ubuntu.
[13:30] <humphreybc> Heya guys, I'm trying to create a session blueprint for UDS
[13:30] <humphreybc> I'm not sure what I project I should assign it to - "ubuntu" or the project(s) it affects
[13:31] <humphreybc> i've tried to assign it to ubuntu and create it, it takes me to this page, but it's not listed
[13:31] <humphreybc> https://blueprints.edge.launchpad.net/sprints/uds-m
[13:49] <jelmer> humphreybc: hi
[13:49] <rye> Hello, is there known bug about e.g.  https://launchpad.net/bugs/556687 returning "Not allowed here" but https://bugs.edge.launchpad.net/ubuntu/+source/ubuntuone-client/+bug/556877 is completely available ?
[13:49] <humphreybc> hey jelmer
[13:49] <wgrant> rye: Different number.
[13:50] <rye> oops
[13:50] <humphreybc> i *think* i figured it out
[13:50] <humphreybc> https://blueprints.edge.launchpad.net/ubuntu/+spec/doc-collaboration
[13:50] <humphreybc> could you have a quick look just to check?
[13:51] <jelmer> humphreybc: That appears to be associated with UDS-M ok.
[13:52] <humphreybc> jelmer: fantastic
[13:52] <humphreybc> :)
[14:18] <NCommander> can anyone tell me what x-lazr-oopsid: OOPS-1560EB1428 is?
[14:18] <NCommander> oh
[14:18] <NCommander> wow
[14:19] <NCommander> handy
[14:19] <NCommander> except I can't access that page :-/
[14:19] <NCommander> the armel apport retracer keeps getting that oops
[14:20]  * maxb wonders why there's no help contact today
[14:33] <jelmer> maxb: I am the help contact today
[14:33] <jelmer> Oh, I hadn't noticed I should put myself in the topic
[14:33] <jelmer> sorry
[14:34]  * maxb points jelmer and NCommander at each other
[14:39] <NCommander> jelmer: can you make heads and/or tails of my oops?
[14:39] <jelmer> NCommander: I'm looking at the moment, trying to find back my credentials first
[14:57] <james_w> NCommander: you are trying to edit e.g. a package task when there is a lucid targeted task for that package
[14:57] <james_w> NCommander: are you getting 500 or 4XX response code?
[14:58] <wgrant> Bug 556515
[14:58] <wgrant> It'd be a 500.
[14:58] <jelmer> james_w: that's what we just found out, based on the backtrace
[14:58] <jelmer> wgrant: thanks
[15:09] <NCommander> james_w: 500
[17:02] <ahasenack> hey guys, just got a "server internal error" with login.launchpad.net, thought I should let you know: http://pastebin.ubuntu.com/411672/
[17:02] <ahasenack> it's actually the 2nd or 3rd time in the past few days
[17:02] <ahasenack> right when I'm trying to login, give my credentials and hit login
[17:14] <ahasenack> I opened a bug about it, have fun
[21:18] <mtaylor> internal server error: http://bazaar.launchpad.net/~eraser029/boots/cursesui/files
[21:19] <maxb> LOSAs: Loggerhead in need of a kicking ^^^
[21:20] <Chex> maxb: on it, thanks
[21:24] <mtaylor> thanks Chex !
[21:38] <dickelbeck> Hi, I have a few simple requests
[21:39] <dickelbeck> There are 3 branches as part of project Kicad.  They are being updated every 6 hours from sourceforge.net.   I need this to stop for all 3 branches.
[21:40] <dickelbeck> We are about to complete our switchover, I just did the final sync from there manually by clicking the nice button.
[21:41] <dickelbeck> Is this better asked on the launchpad "questions" tracker, everyone out drinking already?
[21:42] <zyga> dickelbeck: I think it depends on the timezone you are in, no
[21:43] <dickelbeck> Not so much, there are bloody Marys.
[21:45] <dickelbeck> I know its still too early in Australia even for that, lightweights.
[23:18] <Zarathoustra> hi chan!
[23:20] <Zarathoustra> Is there somebody that can access to the launchpad registry database?
[23:22] <Zarathoustra> I'm still waiting for somebody fix the problem I get to create my account to join a project...