=== asac_ is now known as asac === tuantub_ is now known as tuatnub === tuatnub is now known as tuantub [02:09] would anyone be interested in packaging this personal spanish tutor app? http://linux.com/archive/feature/140574 [02:17] .py app uses some wx stuff. what should the Depends be for "latest stable python-wxgtkX.Y that works with installed version of python" ? [02:18] and how does this work: python-wxversion - wxWidgets Cross-platform C++ GUI toolkit (wxPython version selector) [02:56] with pbuilder on an up-to-date jaunty box, I'm getting complaints about being unable to get debhelper >= 7. Is there an easy way to update the debhelper in base.tgz? [02:56] actual complaint is: pbuilder-satisfydepends-dummy: Depends: debhelper (>= 7) but it is not installable === ryanakca is now known as Guest66699 === Guest66699 is now known as ryanakca [03:17] is it possible for a package to build-depends on itself? [03:21] sorry if that got answered i missed it lost connection [03:51] Does anyone know if you can create a package that depends on itself to build? [03:52] say, once you had the first package in the repos? [03:54] that seems odd [03:54] jedc: It is possible, but extremely frowned on. [03:54] how would it ever get built the first time? [03:54] jedc: How do you boostrap the first one? [03:55] CarlFK1: Generally you get a Canonical sysadmin to hand build it (ain't going to happen for Universe) or your first upload includes prebuild binaries. [03:56] 'neat' [03:56] Painful. [03:57] I can imagine [03:57] ScottK: you can't upload binaries though, can you? [03:57] ScottK: or do you mean, in the source package? [03:57] I guess gcc is like that. [03:57] lifeless: I mean in the source packages. [03:57] I am having gentoo flashbacks [03:58] Pain all around. [03:58] CarlFK1: gcc can boostrap and cross compile with any C compiler [03:58] We really try to avoid this. [03:58] generally too much trouble [03:59] fpc is an example of a package that had to be bootstrapped manually. [03:59] It was broken for several releases before it got done. [04:00] well, i am trying to package a scheme implementation written mostly in scheme, it can be build using mzscheme (for the first one), but is much easier to build using itself [04:01] ok, so better idea is just do it using mzscheme? [04:01] That's doable [04:02] Do the first one with mzscheme [04:02] Painful, but not impossible. [04:02] Build-Depends: itself | mzscheme and a rules that can deal either way could work, too [04:05] hmm, thanks guys for the info and advice === qiyong_ is now known as qiyong [05:49] good morning [06:04] good morning dholbach ! [06:04] hiya fabrice_sp [06:10] got a basic package going today in the style of Ubuntu... http://schwehr.org/blog/archives/2009-08.html#e2009-08-11T19_28_03.txt [06:17] I have a lot packages that needs a sync from Debian, but build depends as a chain (package c depends on b to build and b depends on a to build, and a and b are not in Ubuntu, but are in Debian). Should I open 3 sync requests for each package? Or only one with the 3 packages? [06:22] 3 sync requests. [06:23] ScottK, and I put a reference to each other, saying "you have to sync this bug first, then this, ..."? [06:24] I'm trying to get built maven-plugin-tools, and apart from the self dependency, it miss at least 6 packages ... [06:26] requestsync is broken in karmic? [06:29] not anymore after an apt-get upgrade... [06:37] fabrice_sp: Yes. [06:41] ok. Thanks ScottK [06:43] !backports [06:43] If new updated Ubuntu packages are built for an application, then they may go into Ubuntu Backports. See https://help.ubuntu.com/community/UbuntuBackports - See also !packaging === nellery_ is now known as nellery === mrooney1 is now known as mrooney_ === mrooney_ is now known as mrooney1 [08:06] cjwatson: ia32-libs is now effectively without the pulseaudio alsa plugins, and lib32asound2-plugins doesn't have them either. [08:37] what could be the reasons for dpkg-source to build native package? [08:37] package versioning is correct [08:38] dpkg-source: info: using source format `1.0' [08:38] format 1.0 is native, right? [08:39] nah, it can be both, silly me :) [08:43] hyperair: right, it's on my list, after alpha 4 ... [08:43] okay [08:43] =) [08:48] * ivoks <- idiot :) [09:06] ivoks: format 1.0 is the .dsc/.diff.gz/.orig.tar.gz or .dsc/.tar.gz format [09:07] hyperair: i know [09:07] hyperair: i've solved the issue [09:07] format 3.0 is the one which uses quilt [09:07] it was a typo :/ [09:07] ah [09:07] - instead of _ [09:07] :) [09:07] * hyperair wonders what format 2.0 is [09:08] man dpkg-source [09:09] ah [09:09] right [09:09] * hyperair hopes that dak would hurry up and get format 3.0 support [09:46] hyperair: soyuz love for format 3.0 would be even better :-) [09:46] siretart`: well it'll have to come before dak implements it or autosyncs will be painful. [09:47] indeed [10:51] nhandler: can you get http://ubuntupackaging.wordpress.com/2009/08/12/on-call-review/ on the fridge? :) [10:51] I just discovered dpkg.org, it's quite cool: http://www.dpkg.org/dpkg/DPKG?action=fullsearch&context=180&value=dpkg&fullsearch=Text [10:52] well some of the pages at least like http://www.dpkg.org/dpkg/ConffileHandling [10:53] persia`: did the packaging training session about upgrading packages happen? [10:53] there's no log at https://wiki.ubuntu.com/Packaging/Training/Logs yet [10:55] dholbach: how do you keep tabs on all these things? how do you keep yourself organised? you seem to be involved in so many things and yet you seem to be on top of everything so effortlessly [10:56] highvoltage: he has superpowers :-) [10:56] highvoltage: I struggle to keep on top of ONE thing :-) [10:56] This dholbach has Super Cow Powers. [10:57] mok0: heh, I can relate :) [10:57] highvoltage: I wrote a bit about it in http://daniel.holba.ch/blog/?p=455 [10:58] (the last part) [10:58] I guess I'll have to give GTG a shot myself then [11:10] StevenK: Do you plan to update gnome-do-plugins too? [11:11] jpds: Yes [11:11] Awesome. [11:34] dholbach, There weren't enough attendees. We talked a bit about other things, but didn't do a proper upgrade training. [11:34] persia`: ok [11:36] * slytherin is glad to see persia` back in action. :-) [11:38] huhu persia` \o/ [11:40] Um. "in action" isn't the best way to describe it, but I am about. [11:40] in action = online and writing in the channel ^^ [11:41] Ah :) [11:43] In action = sharing the wisdom with community. :-) [11:43] uploading stuff and breaking the archive \o/ [11:45] RIght. That's more how I usually think of "in action". === kiko is now known as kiko-phone [13:11] thanks nhandler [13:12] No problem dholbach [13:35] dholbach: hi [13:37] hey bdrung_! [13:37] how are you doind? [13:37] doing [13:38] dholbach: good. the exams are behind me. i have time for packaging. ;) [13:38] yeeeeehaw! :-) [13:38] dholbach: when i manage to setup the network for kvm i will try out harvest [13:39] that sounds great [13:39] thanks a lot for jumping in to help out [13:39] james_w: ^ seems we found one Harvest Hero already! [13:39] :-) [13:39] dholbach: ;) [13:40] dholbach: seams that i am hacking on basic things. harvest, mozilla-devscripts, ... [13:40] dholbach: i plan to apply for motu. what do you think? [13:41] bdrung_: I think that's a fantastic idea [13:42] dholbach: ok, then i start writing it. [13:42] :-) [13:43] ttx: Do you plan to ask for sync of jetty from Debian unstable, since your packages only provide libraries and we have outdated jetty in karmic. [13:43] I'm trying to package up my python program, and have read the debian python policy, and I still don't quite understand the handling of "private modules" for my python package [13:43] dholbach: do you come to the stammtisch today? [13:44] bdrung_: no, unfortunately not - we have Mimi's cousin visiting us tonight [13:44] Can anyone point me to an example of a python program that ships a private module, so I can see how it modifies the sys.path to enable the import? [13:44] slytherin: not for karmic... jetty6 must reach main and at this point the jetty6 debian packages are not good enough. But for karmic+1, we'll definitely sync [13:44] ok. [13:44] I'll spend a few cycles qa it in debian next month [13:45] I was just asking from point of view of solr merge. [13:45] we have a solr-jetty package which is not present in Debian because jetty was previously removed from Debian. [13:49] pi-meson2: see PAPT repository - you'll find losts of examples there (patching sys.path is avoided in most of them) [13:49] see also http://lists.debian.org/debian-python/2009/03/msg00091.html (search for "current") [13:50] PAPT repository = http://svn.debian.org/viewsvn/python-apps/packages/ [13:50] POX: fantastic, thanks === cprov1 is now known as cprov [13:52] "--install-lib=/usr/share/foo --install-scripts=/usr/share/foo" added to setup.py's arguments should be enough (if script name != module name) [13:52] (and then dh_link /usr/bin/bar /usr/share/foo/bar) [13:53] okay, so the crazy part here is that, at the moment, I'm not actually using setup.py, but rather cmake [13:53] I might have to change that [13:53] doesn't matter much [13:53] the idea is to have script and module in the same directory [13:54] /usr/share/foo/script and /usr/share/foo/module - this way you don't have to patch anything [13:54] and then /usr/bin just symlinks to the script? [13:54] yes [13:54] *whew* okay, that makes a ton of sense [13:54] and prevents a lot of nastyness, too [13:54] thanks! [13:54] join PAPT and you'll get more comments later :) [14:06] Does anyone exactly how the buil-depends-indep work from point of view of Debian buildd? [14:08] according to policy they are only needed to build arch:all packages. In Debian they're a build by the uploader, so not installed on the buildd at all while on a Ubuntu buildd only on i386 (as it builds arch:all) [14:09] practise may differ (as usual) [14:12] hmm, looks like this is the problem then. One of the package FTBFS on non-i386 Debian buildd because there is no JDK. The reason is default-jdk is in build-depends-indep. [14:13] but it's really required to build a non-arch:all package? [14:15] Laney: Actually the build process is continuous. It builds arch:all as well as arch:any component in one go. [14:15] is that policy compliant? [14:17] in what way is it non-compliant? [14:18] any depending on all I thought was a problem [14:18] Laney: or the reverse? [14:18] maybe [14:19] Consider this. There is a jar file which will go in arch:all package. And there is a .so file which will go in arch:any package. Both these files are built in a continuous process. [14:20] the arch:all package depends on arch:any in this case. [14:24] slytherin, the problem with any depending on all is NMUable version numbers [14:25] right, but here all depends on any. [14:25] slytherin, if a binNMU of javafoo-native is uploaded with version 1.0-1+b1, but javafoo-java 1.0-1 expects javafoo-native=1.0-1, then you have a problem [14:26] so you can make that NMU-safe by breaking your build process to depend on >=1.0-1, << 1.0-2~ [14:26] Hmm, will have to think about this. There are at least 2 other such packages. [14:27] however, the NMU is the key problem here, and why breaking the build system to allow building the binary-only part should be allowed [14:27] i.e. it's okay for the arch:all build to do the arch:any build first, but NOT okay for the arch:any build to also build the arch:all [14:27] i don't know if it actually breaks policy, but it can be... ungood [14:28] as an example, ikvm in Sid is on 9 arches, and only 3 in lenny - and the extra arches were enabled by splitting the arch:all build [14:29] (it needs about a gig of ram to build, and crashed on most buildds mid compile) - but i386 buildd had the ram [14:30] directhex: But in case of Ubuntu non-i386 buildd actually build arch:all packages. [14:31] mmm, nope [14:31] So should I then use some technic in rules file to specify which target to run for arch:all and which to run for arch:any? [14:35] slytherin, yeah - the buildd will run "debian/rules build" then "debian/rules binary-arch" on non-i386 arches. so make sure your build rule doesn't do the indep stuff [14:35] I am using cdbs [14:35] on i386 it does "debian/rules build" then "debian/rules binary" [14:35] cdbs i can't help i'm afraid [14:37] If we had cdbs 0.4.59 then you could set DEB_BUILD_DEPENDENCIES [14:37] I even did the merge but there's a build failure I don't know how to fix :< [14:38] or there is easier option to change arch:all package to arch:any. [14:38] Heya gang [14:38] hi bddebian [14:39] Hello Laney [14:43] Hey bddebian. [14:44] Hi iulian === jMyles_ is now known as jMyles [15:25] Hello, can a package in universe have a depends on a package in multiverse? [15:25] nope [15:25] ping: nixternal [15:28] james_w, thanks === bdrung_ is now known as bdrung [16:07] how do I request NEW packages be synced from debian to ubuntu? [16:08] Ryan52: command 'requestsync' [16:08] same way as normal? I'd assume that there's no package to assign the bug to, tho... [16:08] slytherin: it doesn't work on my system because I don't want to install random .debs from ubuntu. [16:08] I just need to know, if there's no package to assign the bug to, where do I assign it? [16:09] Ryan52, NEW as in ftp-master.debian.org/new.html ? [16:09] yes. [16:09] well no. [16:09] there's a package that just got ACCEPTED into Debian [16:09] but it's not in Ubuntu [16:09] I want it to be. [16:09] just use requestsync [16:09] Laney: 08:08 < Ryan52> slytherin: it doesn't work on my system because I don't want to install random .debs from ubuntu. === kiko-phone is now known as kiko [16:10] oh ok [16:10] then file a bug manually against Ubuntu [16:10] Ryan52: Ok. Then the package should be 'Ubuntu'. [16:10] ok, thanks. [16:10] but how are you testing that your sync works without an ubuntu environment? [16:11] I'm not. [16:11] :D [16:11] ¬_¬ [16:14] Ryan52: make sure you subscribe the ubuntu-universe-sponsors team to the bug. [16:15] nah, I just bother nhandler :) === sbasuita_ is now known as sbasuita === zul_ is now known as zul [16:43] dholbach: my application text is ready (hopefully). [16:44] bdrung: could you send me an email about it - I'm happy to comment on it tomorrow [16:44] I'm just about to head out [16:45] dholbach: i will. [16:45] thanks muchly! [16:46] dholbach: would the 27.8. to early? [16:47] bdrung: not at all - just ask a few folks to weigh in on the application and apply [16:47] dholbach: ok [16:47] ok... I need to head out now [16:47] see you around! [16:47] dholbach: cu [16:47] bye [17:02] stochastic: I tried your patch for ardour (bug #148585) but it FTBFS [17:02] Launchpad bug 148585 in ardour "".ardour" project files are not recognized by GNOME (no file type association)" [Low,Confirmed] https://launchpad.net/bugs/148585 [17:03] geser, what does FTBFS mean? [17:03] Fails To Build From Source [17:04] uh oh [17:04] * stochastic runs off to check why [17:05] see my comment in the bug [17:05] it's probably because of the inclusion of gnome.mk [17:05] but I don't know if it's needed for sharedmimeinfo and if yes how to fix it [17:06] geser, from what I've read on cdbs that's the line to include for sharedmimeinfo [17:11] but apparently it doesn't work as intended for package without a configure script (ardour uses scons as far as I can tell) [17:11] geser: chrisccoulson was saying that running dh_desktop is depreciated in Karmic, therefore probably including gnome.mk isn't needed [17:11] ^^ the therefore is my interpretation, not what he was saying [17:16] try it out and check if the sharedmimeinfo is correctly included in the package [17:43] I'm thinking of starting a new pygtk-based project, which may be ready in 6 months or so. Should I use python 3 or python 2? I'm asking because I think that python 3 won't be shipped by default in even Ubuntu 10.04... [17:45] Ohai, looks like there's a bug in the runit package on karmic. I added a patch to this bug: #406621, but not sure the process of getting the package updated.. should i repackage per the MOTU Contributing page? [17:45] (i'm not the maintainer of runit, just a user) [17:47] alkisg: does pygtk work already with python3? if you don't plan to get your app into main, you can use python3 if you want as it's already in universe (but I don't know if the packaging tools do the right thing for python3) [17:47] geser, my pbuilder doesn't want to successfully build the original ardour-2.8-0ubuntu1 package, does yours? I can't test the mime info otherwise [17:48] I guess I'd better stick with 2.x... :) [17:48] stochastic: I just checked if the same error happens in the 2.8-1ubuntu1 and aborted when configure got running (didn't let it build to the end) [17:50] geser, half way through it errors out on some invalid char conversion [17:50] is this a bug in the package that needs to be fixed before release? [17:50] or a bug in my pbuilder? [17:50] jtimberman: https://wiki.ubuntu.com/SponsorshipProcess [17:51] geser: the runit package is already in universe. [17:51] stochastic: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=505960 perhaps? [17:51] Debian bug 505960 in ardour "FTBFS with GCC 4.4: missing #include; overloaded 'bitset(int)' is" [Unknown,Open] [17:51] it doesn't need a sponsor for a bug fix does it?? [17:52] jtimberman: sponsor == some dev who uploads a bug fix for you (after review) [17:53] geser: okay, hopefully some attention will be given to this then? its a pretty serious bug for anyone using runit, as it won't start properly when the package is installed. [17:55] geser that debian bug doesn't have the same error that I'm getting. I just uploaded to my PPA to see if the same thing happens there [17:56] hi, I am looking for an example multi-binary source package which uses debhelper and builds both arch-dependent and arch-independent binary-packages [17:58] logari81: look for packages building a -doc package, it should be easy to find one using debhelper [18:00] logari81: e.g. mcpp (but I didn't check if it's a good example) [18:02] geser: mcpp seems just fine..., it is also debian native [18:08] Hi. Can some MOTU have a look at bug #412347 and Bug #412352? This are the first step for more than 7 sync/merge after. Thanks :-) [18:08] Launchpad bug 412347 in ubuntu "Sync libstax-java 1.2.0-1 (universe) from Debian unstable (main)." [Wishlist,New] https://launchpad.net/bugs/412347 [18:08] Launchpad bug 412352 in ubuntu "Sync maven-repo-helper 0.7 (universe) from Debian unstable (main) (to sync after bug #412347)." [Undecided,New] https://launchpad.net/bugs/412352 [18:18] fabrice_sp: done [18:21] geser, thanks :-) [18:21] I'm trying to fix all the maven packages, and it's a real mess! [18:26] good luck with that [18:50] if there's a needs-packaging request for a package that's currently in debian NEW, is it proper to wait for it to enter unstable and request a sync, or upload a candidate to REVU and allow the debian version to override it later? [18:52] jbernard__: if it's not high in the queue, I'd upload it to Ubuntu now as FeatureFreeze will start soon and after that it's hard to get a new package in [18:52] you can request a sync later when it's accepted into Debian [18:54] it usually takes a week, which it's nearly at now, but I never know the current ftp-master load or how fast they'll get to it [18:54] a week to be processed? I'd rather say a month :) [18:55] well then i should definately get my package uploaded to REVU soon ;) [18:55] thanks for the input, i appreciate it [18:57] yw === ripps_ is now known as ripps [19:29] <__theIdiotBox> hello everyone. i want to give back to the opensource community... and i think i can do that by contributing to the repo maintenace/building community...thats why i'm here to learn and get started...can someone help me get started. [19:30] !contributing [19:30] Sorry, I don't know anything about contributing [19:30] __theIdiotBox, look at the subject (https://wiki.ubuntu.com/MOTU/Contributing ) [19:30] <__theIdiotBox> fabrice_sp_, any useful link [19:31] <__theIdiotBox> thank you : fabrice_sp_ [19:31] yw :-) [20:50] fabrice_sp_, hello again, you told me yesterday to repack eclipse-common-nls because it has been kicked out ouf debian [20:51] fabrice_sp_, now there is a problem, it seems to be no upstream source greater than this [21:04] fabrice_sp_, should I maybe build the old version for ubuntu? At least I need it === sbasuita_ is now known as sbasuita === Pricey is now known as PriceChild === PriceChild is now known as Pricey [22:07] Does anybody know how to use gpg-agent and pinentry over ssh. I'm trying to pull bzr packges, build, sign and upload them from a remote machine === Pricey is now known as PriceChlid === PriceChlid is now known as PriceChild [22:52] hi [22:54] I've read http://davidsiegel.org/100papercuts-round6/ and I'd like to help. Anyway I don't know how to set up my local devel environment. Are there guidelines about it? [22:57] hmm, debuild -S -s -k$GPGKEY starts with clean and it fails because I didn't build it yet, how do I ommit clean? [22:59] https://wiki.ubuntu.com/PackagingGuide/PatchSystems seems a good source of info === JanC_ is now known as JanC