[00:46] Hello everyone. aMSN 0.98 sources uploaded to sourceforge a few minutes ago, available from http://sf.net/projects/amsn/files . I think the deadline for making it into karmic is today or something? [00:47] you have 23 hours [00:47] ok, anything i can do as a non-motu amsn developer? === micahg1 is now known as micahg [00:48] i think the timing is awkward. how significant is the jump from 0.98 from whatever was previously in the archive? [00:49] uhh, new protocol in use, audio and a/v conferences supported, tons of bugfixes, [00:49] content roaming (share your display picture and personal message across various computers) supported [00:50] so "pretty significant" then [00:50] yup [00:50] that's just what i can think about OTOH [00:50] package it or convince someone else do [00:50] to* [00:50] hope you or they don't mess it up ;) [00:51] from a packager perspective, any changes that would affect packagers? changes to the build system, build dependencies, or binary files installed by "make install"? [00:51] hmm.. [00:51] pochu: congrats on becoming DD =) [00:52] if you want a/v support there's a dependency chaos with farsight [00:52] binary files - i think farsight and tclISF ? [00:52] wait i think a guy is already doing proper packages (of earlier svn revisions) [00:52] the he's got 23 hours! [00:53] they're non-official packages though :S [00:53] directhex, speaking of 23 hours, any chance you'd be willing to give the second advocation for http://revu.ubuntuwire.com/p/xajdeo ? [00:53] ah they're for jaunty [00:53] hmm. i'll try poking Adri2000 for now :) [00:55] stochastic, apparently marked as already advocated [00:56] directhex, whoops, typo http://revu.ubuntuwire.com/p/xjadeo [00:58] advocated. you owe me one piece of gushing praise about mono. [00:59] * stochastic loves the soft elegant structure of mono. He can't think of a nicer piece of engineering. [00:59] thx ppl, see you~ [00:59] thta was a quick review! [01:01] i analyse debdiffs at double speed after four pints [01:01] 'analyse' [01:02] after four pints I suspect that word becomes a pejorative [01:03] i would typically reject anything with an upstream debian/ still in orig, but if james-w is happy with it, then it'd be wrong for my personal distaste to be a deciding factor [01:03] and other than that, it seems to be in good shape [01:03] as much as cdbs nonsense is capable of such things [01:03] hah [01:04] stochastic: you could use .install files for some of the stuff in rules [01:04] see, i actually looked at it! and here you thought i merely clicked the shiny button in a drunked stupor [01:05] and you should reword the comment of your desktop file to be in line with http://developer.gnome.org/projects/gup/hig/2.0/desktop-integration.html [01:05] but yeah I'd still upload it [01:05] Laney knows his shizzle, listen to Laney [01:05] Thanks Laney, it's good to learn these things. So this does mean that it's made feature freeze now, right? (provided the archive admin doesn't reject it) [01:05] also, Laney for DD! [01:05] provided directhex uploads it [01:06] and it will be ok to fix even after a reject [01:06] at least that's my interpretation [01:30] directhex, should I be worried that xjadeo hasn't moved to "archived packages" and is still on the "new packages" listing of revu? [01:30] stochastic, no idea. RainCT knows about REVU, i'm a mere peon with it [01:30] stochastic: please ask upstream to remove the debian directory in their next release. [01:30] now, bedtime! [01:30] bdrung, will do. [01:31] stochastic: and please contact the person who wants to package xjadeo for debian (see mail response) [01:31] bdrung, will do. [01:32] stochastic: thanks. [01:32] bdrung, no thank you for your help. [01:32] yw [01:38] any chance of fontconfig 2.7.1 getting into karmic? [01:39] will there be another debian sync before Karmic is released? [01:40] I believe syncs are by request only at this point [01:40] ah okay, thanks micahg [01:40] https://wiki.ubuntu.com/KarmicReleaseSchedule [02:14] hrm, "./pbuilder-dist karmic create" on my sid box is complaining that "karmic" is an "Unknown suite" === zooko` is now known as zooko [02:15] Howdy, warner. [02:15] maybe sid's /usr/bin/cdebootstrap doesn't know about karmic.. [02:15] heya [02:17] indeed, /usr/share/cdebootstrap/suites doesn't mention "karmic". maybe if I just edit it.. === Cuddles is now known as Richie [02:41] hrm. cdebootstrap is hanging on "P: Retrieving Release" === keylocker is now known as leleobhz [03:34] what's the correct chan for asking packaging questions? [03:35] this one [03:35] prefrontal: this channel, if you're packaging for universe (most people are) [03:36] binarymutant: ohai, i saw you on #debian-ruby :) [03:36] :D [03:38] can you peek at my rules file please, i am only getting stuff installed to /usr/share/doc. http://pastebin.ca/raw/1542672 [03:39] basically, i only get the stuff i listed on the dh_installdocs line [03:40] debuild builds my package.. [03:40] use dpkg-buildpackage -rfakeroot to build your package. [03:41] thanks. i also get many of these messages: Compatibility levels before 4 are deprecated. [03:41] yeah, put '7' in compat. [03:41] debian/compat [03:41] * jtimberman isn't a motu btw, just a user that recently packaged some thigns [03:42] thats cool, thanks. [03:42] here's a live one. [03:42] https://bugs.launchpad.net/ubuntu/+source/renpy/+bug/418992 [03:42] Launchpad bug 418992 in renpy "Ren'py 6.9.3 released" [Undecided,New] [03:53] is it necessary to find the package that provides every single library listed on ldd $(whatever lib) and put that package on the Depends line? [03:54] is it safer to do so in terms of getting pkg approved? [04:00] prefrontal, I think a lot of those will be taken care of with sh:Libs [04:01] ok === ApOgEE__ is now known as ApOgEE [04:46] when i name this package libQuarter0 i get `dpkg-source: error: source package name `libQuarter0' contains illegal character `Q'', and when I name it libquarter0 i get package-name-doesnt-match-sonames [04:51] prefrontal: You cannot have uppercase letters in package names. [04:52] If I recall correctly. [04:52] Would some MOTU please be so kind to review (and hopefully advocate) the package ffgtk I just uploaded? http://revu.ubuntuwire.com/p/ffgtk I'm shooting for upload before the Feature Freeze. === Laibsch1 is now known as Laibsch [05:11] is lintian insane, or am i insane? either way, i'm going insane! === johe__ is now known as johe [05:31] can someone please help me figure out what this means: native-package-with-dash-version http://lintian.debian.org/tags/native-package-with-dash-version.html [05:31] prefrontal, is you package up on REVU? [05:31] no i am workin on it [05:32] trying to nail these lintian errors [05:33] i download Quarter-1.0.0.tar.gz, open it up to Quarter-1.0.0, mv it to libquarter0-1.0.0. the package is called libquarter0 [05:33] prefrontal, it would be easier for people to give you feed back if it was up on REVU [05:37] warner: what's the status of Tahoe-LAFS-in-Karmic? === Guest14205 is now known as santiago-ve [05:38] prefrontal, https://wiki.ubuntu.com/MOTU/Packages/REVU/ [05:39] no change since this afternoon. I haven't been able to get a karmic pbuilder environment built on my home sid box, to conveniently fix the minor things which I already know about [05:39] it's gotten no feedback yet [05:40] porthose, i'm all REVU-ready, i was just hoping to fix all this crap myself before pestering you guys 2 much. but i am about to upload this thing right now [05:40] iulian: you mentioned that you might be able to help with Tahoe-LAFS. Brian Warner here has gotten it uploaded to REVU. [05:44] prefrontal, having it up on REVU is not pestering people :) [05:47] porthose, i just uploaded it, but i don't see it on revu yet [05:48] porthose, here it is: http://revu.ubuntuwire.com/p/libquarter0 [05:49] * porthose looks [05:51] hyperair: :D [05:52] pochu: :) [05:52] pochu: weren't you interested about gtk2-engines-aurora the other day? ;-) [05:52] hehe, maybe :) [05:53] poke me about it tomorrow, ok? [05:54] porthose, fyi, this is just a dependency of the package i realy want to get into karmic, which we just released yesterday. so any advice you can provide will help me tons. its a neural network simulator called emergent: http://grey.colorado.edu/emergent [05:54] pochu: sure :-) [05:54] ive been providing my own apt repo for years, but figured its time to do it right [05:56] prefrontal: are you at CU Boulder? Emergent looks cool. [05:56] yep i am [05:57] Cool! I did my undergrad studies there and now years later I've returned to live in Boulder. [05:57] hello fellow boulderite ;) [06:10] Okay, I have asked all the Ubuntu and/or Debian developers that I can think of if they will please review Tahoe-LAFS, and now I must go to bed, as I should have done two hours ago. :-) [06:12] good morning [06:15] Good morning, dholbach. [06:16] hey zooko [06:26] dholbach: did you ever have a chance to check out those videos? [06:27] which one? [06:28] I was looking at them a while ago and their seemed to have changes that made them hard to follow - I could be that thick too [06:29] oh, now I see what you mean [06:29] no, sorry - I didn't have the time to dive into them yet [06:29] ok [06:30] can you make vmware finish installing for me? [06:30] I probably can't :) [06:30] installing this thing is a painful process :( - so painful - I was using vbox but apparently that's not good for developing machines you want ot give to others [06:59] I noticed a package of mine (xjadeo) that got it's second advocation on revu about five hours ago is still listed on the 'new' page. Is this correct? Has it been uploaded? Do I need to click the "archive" link at the top of it's page? I don't want it to get messed up and miss feature freeze. [07:07] prefrontal, I left a comment :) [07:12] good mornin dholbach [07:14] hiya porthose [07:30] stochastic: You could check the new queue. [07:30] TO be sure. [07:31] TheMuso, where is that? [07:32] stochastic: afaicr https://launchpad.net/ubuntu/karmic/+queue [07:35] TheMuso, eek looks like directhex didn't upload it. Do you have a second? [07:38] stochastic: Sure. [07:39] * TheMuso logs into revu. [07:41] * TheMuso gives it a cursary build/lintian run just to be safe. [07:41] Not that I mistrust the MOTU who gave it the ok. [07:43] hi folks [07:43] hey DktrKranz, thanks for the upload! :) [07:44] sistpoty|work: you're welcome, I didn't commit to svn though (I forgot :/) [07:44] stochastic: Uploading now. [07:44] DktrKranz: no problem, will do that after work [07:45] sistpoty|work: ... and thanks for let me do some practice ;) [07:45] heh [07:47] porthose, thank you [07:48] prefrontal, yw :) [08:37] hi guys [08:37] memaker has just had a new release [08:37] and I have built a new package for it [08:37] it has LOADS of bug fixes and some new features [08:37] what is the best step to take to get it into the universe before feature freeze? [08:37] do I need to make a bug? [08:38] rugby471: Yes. [08:38] Attach diff.gz. [08:38] only the diff.gz [08:38] what about orig.tar.gz [08:38] rugby471: We will get it. We only need the diff.gz file. [08:38] ok [08:38] rugby471: And subscribe ubuntu-universe-sponsors. [08:38] what do I file the bug in, memaker? [08:39] rugby471: Yes, file it against memaker. [08:39] and does it needs any specific tags? [08:39] rugby471: No. [08:39] kl [08:39] Tell me when you're done. [08:39] thanks iulian! [08:39] kl [08:39] I will [08:40] is there any recommended title of the ubg I should use? [08:40] ubg > bug [08:40] hey rugby471! [08:40] long time no see [08:40] sorry one last question do I file it against the memaker project or the memaker package? [08:40] hey [08:40] I thought you ere walking the dog... [08:40] :-) [08:40] joking [08:40] back now :) [08:40] rugby471: Against memkaer package in Ubuntu. [08:40] thanks [08:41] Hello dholbach. [08:42] hey iulian [08:42] How goes it? [08:42] very good very good [08:42] how 'bout you? [08:43] I'm happy. I've just passed my driver's license test. [08:44] congrats [08:44] Thanks! [08:44] woohoo [08:46] iulian: ok filed the bug here https://bugs.launchpad.net/ubuntu/+source/memaker/+bug/419077 [08:46] Launchpad bug 419077 in memaker "memaker package needs updating to 1.5" [Undecided,New] [08:46] hehe thanks bot :-) [08:47] dholbach: sorry forgot to ask how are you? [08:47] thanks, I'm good, just very busy :) [08:48] but that's nothing new :) [08:48] hehe [08:48] * warner finally gets a karmic pbuilder environment set up on sid [08:51] * juanje hugs dholbach :-) [08:51] * dholbach hugs juanje back [08:52] now, will it actually build the zfec package that I want to get into karmic.. [08:53] do people usually install lintian into their pbuilder environments? mine doesn't seem to have it.. [08:55] cool, it did. But, of course, without lintian in the karmic chroot, I can only run it against sid's lintian, which will probably complain about different things [09:09] Is a maintainer for the CouchDB package in Karmic around? === e-jat is now known as ejat [09:18] jtimberman, I think that the maintainer is motu as whole [09:18] binarymutant: indeed.. [09:19] jtimberman, it's synced from debian [09:19] I'll try to find somebody couchdb [09:19] just a sec [09:19] binarymutant: not the version in karmic [09:20] ah :D [09:20] the version in karmic is 0.10, which has some backwards incompatible changes per one of the guys on #couchdb. [09:20] which also breaks chef-server. [09:20] they said 0.9.1 is stable on #couchdb. [09:27] pbuilder just blew my mind. what a nice tool === warner is now known as warner_sleep === WelshDragon is now known as Richie [09:43] prefrontal: seen cowbuilder yet? :-) [09:44] or sbuild? [09:45] sbuild, as i understand, is used on the ubuntu infrastructure to build out packages across the board. [09:46] yes, but it's more complex to get going, and requires LVM for its copy-on-write functionality [09:46] So pbuilder is better when you're getting started [09:46] maxb: I'm a sysadmin with years of experience with LVM, so I'm not afraid :-D [09:48] sbuild is hard to get setup. [09:51] <\sh> slytherin: hmm..nope...you should just setup local package mirrors [09:52] <\sh> mk-sbuild-lv does the rest [09:52] That is why I said it is hard to setup. With pbuilder it is just 'sudo pbuilder --create'. [09:53] However sbuild + schroot gives you much more flexibility with having chroots on hand to test things as a normal user. [09:53] The same chroot can be used to test package functinality/build by hand, as well as a build run with sbuild. [09:53] maxb, pbuilder is actually not working for me bcz the packages im testing require opengl [09:53] And there is no waiting for tarballs to pack/unpack. [09:54] does cups-ppdc replace cupsddk? [09:54] prefrontal: so? [09:54] cowbuilder? [09:55] alas, i resorted to vmware [10:03] oh man.. so far i have tried to test my package in pbuilder (no opengl), by installing alpha 4 to a vm (never booted up), and by upgrading jaunty to karmic (rebooted to busybox) [10:03] i am feeling screwed === dholbach_ is now known as dholbach [10:29] when a motu gets a chance can you review my upload on revu, gnome-video-arcade. i've addressed rainct's comments, thanks. http://revu.ubuntuwire.com/p/gnome-video-arcade [10:32] hi, I want to remove some code from an upstream package, not because it is not free, but just because it has another license as the rest of the package, it contains files missing a license header and because there is already a ubuntu package with this part of code as a standalone lib. [10:33] My question is if I have to add the dfsg suffix to tha orig.tar.gz [10:33] logari81: yes. [10:36] slytherin: thank you, I thought dfsg is only for the case that non-free code has to be removed [10:37] logari81: as far as I know, files missing license header have ambiguity regarding dfsg freeness. [10:37] ok actually you are right [10:39] logari81: if the files in question are free (and this seems to be the case if they are in another package already in the archive), there's no need to remove them [10:39] logari81: just make sure that the library from the package, not the internal copy is used [10:40] logari81: and of course document the different license in copyright [10:44] sistpoty|work: but so I will still have a problem with the missing license headers [10:45] but I see in the standalone package (superlu_3.0) the problem has not been corrected [10:46] there are many files with no license [10:46] logari81: there's not too much of a (legal) problem with missing license headers (unless e.g. it's a gpl'd lib, and the license headers have been *removed*, which the gpl doesn't allow) [10:46] it is BSD code [10:47] logari81: a license is not bound to license headers. if there's a file that says "this is lib blabla, with the license foobar", that works as well [10:47] logari81: of course license headers are way preferred (exactly to find out about it, if used in a different project), but not mandatory [10:53] sistpoty|work: then what you suggest is to let the upstream package as is, to copy the same copyrights from the stand alone superlu to my debian/copyright even if during the building I use the ubuntu standalone superlu lib instead of the superlu copy provided with my upstream package. If that is right I ll prepare a new version and upload it for further proofing. [10:54] logari81: yes. Of course you should also doublecheck superlu lib files/the copied copyright to be in order [10:57] Can anyone recommend a relatively simply Python application package I can use as a guide to creating a package for a new Python application? [11:02] IntuitiveNipple: I recommend you to try Quickly (https://edge.launchpad.net/quickly) that makes all the stuff you need for a new Python app and package. After that you can see inside how the package has been done [11:03] Thank-you. I'm working on getting the video editor OpenShot packaged and went down the python-support route and then realised its mainly for supporting modules rather than applications [11:03] IntuitiveNipple: also you can check packages like gtg or usb-creator [11:05] IntuitiveNipple: ok, anyway, you can create som dummy app with quickly just to see how the package is and copy the stuff you need [11:06] I'll have a play around with those :0 === jweyrich_ is now known as jweyrich [13:23] ttx: ping [13:24] slytherin: quick-pong [13:24] ttx: is it ok to merge ivy from Debian unstable, if we skip maven related changes? [13:25] The version in Debian is 2.1.0~rc2 (ubuntu has rc1). [13:25] any missing feature / bugfix as a reason why we should ? [13:25] I haven't checked changelog. I will and let yuo know. [13:26] slytherin: I'm ok with it if tere is a compelling reason to do it. If it's just to have the latest, then I prefer we pass [13:26] ok [13:44] I'm trying to understand the reasons for splitting a single upstream application into several logical binary packages, e.g. application, application-data, application-themes application-profiles, where 'data' 'themes' and 'profiles' are distinct data file sets. I thought it was so these packages could be updated without the application being updated, but it seems like from the same source package, the buildd will always build all the bin [13:44] ary-packages in the source, so what reason would there be to logically split like this? === rmcbride_ is now known as rmcbride [13:54] ttx: ivy rc2 changes - http://paste.ubuntu.com/259826/ [13:55] geser: Ping. I'm working on updating libwiimote. There's this patch from you http://paste.ubuntu.com/259827/ but just adding "-fPIC" to the CFLAGS:= line seems to be enough for it to build, so I'm wondering whether the other stuff is needed? [13:56] slytherin: hm. So nothing outstanding... if it were to move to the final release, I'd agree. But to move to rc2 ? I just don't need any risk introduced right now. [13:58] Ok. I will check when final release is made. [13:58] That will probably need FFE. [14:08] geser: ah nevermind, upstream fixed the file so -fPIC can be set from debian/rules now [14:20] RainCT: hi! are you back from holidays? [14:20] pochu: Hey. Not yet, but I thought I'd do some stuff before FF :) [14:21] RainCT: ah ok :) webboard can wait some more then ;) [14:27] I have a question : I am maintaining a package in debian and I am dealing with it on Ubuntu too.. [14:27] I would like to have a ne release in ubuntu [14:28] but since we are in sync currently [14:28] pochu: Maybe I can get to it later today. You can sponsor the upload then :) [14:28] I would avoid to create diversion... [14:28] RainCT: alright! [14:28] hi. I get an error when logging in @ http://revu.ubuntuwire.com/ -> "MOD_PYTHON ERROR [...]" [14:28] do I need a FFe to sync toward 1.1.10-1 if it's e.g. 1.1.10-0ubuntu1 in Ubuntu, and 1.1.10-1 in Debian ? [14:29] does this happen for other people as well? [14:30] RainCT: ^^ [14:30] huats: I don't think so, but if both versions are equal there isn't much point in requesting a sync for that. Just remember to sync once Debian has something interesting. [14:30] huats: No. [14:30] * RainCT looks [14:31] RainCT: The idea was to avoid diversion... [14:31] RainCT: and ScottK thanks ! [14:31] than I will upload it right now to ubuntu... [14:32] huats: (Well, being diverted isn't necessarily bad. You've to consider what is more important, the version number or giving the builds work and forcing people to download the package again) [14:32] is requestsync broken for everyone? bug #418802 [14:32] Launchpad bug 418802 in ubuntu-dev-tools "requestsync crashed with ValueError in validate_param_values()" [Undecided,New] https://launchpad.net/bugs/418802 [14:32] bakkdoor: Works here. Can you paste the error on http://paste.ubuntu.com, and check if Ctrl+F5 fixes it or the error is persistant? [14:32] RainCT: sure [14:33] I'll keep that in mind [14:33] thanks RainCT [14:38] "every package needs a "needs-packaging" bug in Launchpad" when did that become a requirement and what's its purpose? [14:38] RainCT: http://paste.ubuntu.com/259838/ <- its persistent [14:38] I don't think it is a requirement. [14:39] ITP isn't even a strict requirement in Debian. [14:39] Riddell: this spreads on getdeb too [14:39] anyone started it [14:39] -_- [14:39] Riddell: where'd you see that? [14:40] REVU tells you off if you don't have one too [14:40] yeah it's on https://wiki.ubuntu.com/MOTU/Packages/REVU/CheckList [14:40] and I've seen commenters moan about it too [14:40] on revu [14:40] Riddell: well, it *should*. [14:41] it's a good idea to prevent duplication of work, but nothing more imo [14:41] lfaraone: Why? [14:41] Laney: Definitely. [14:41] * RainCT agrees [14:42] Riddell: iirc, that was discussed some time ago on the -motu list, let me try to find the thread again [14:43] bakkdoor: Sorry, no idea why that happens. Maybe ask in #launchpad? [14:44] RainCT: alright [14:49] Riddell: https://lists.ubuntu.com/archives/ubuntu-motu/2007-November/002694.html (iirc the idea came up to avoid duplication and to be able to link to a itp, but I'm no longer 100% sure) [14:51] Hi! Can someone please give me a hint how to change the ownership of files to install when creating a package? [14:52] RainCT: when you get a chance, can you review my upload on revu, gnome-video-arcade. i've addressed your comments, thanks. http://revu.ubuntuwire.com/p/gnome-video-arcade [14:52] flohack: chmod? :) [14:52] RainCT: I tried a 'chown -R www-data:www-data $(CURDIR)/debian/mypackage/www-data-dir' within the binary-install/mypackage target in rules, but that does not seem to work [14:53] (or if you use 'install' instead of 'dh_install', the former has an option for ownership) [14:53] RainCT: I use a .install file to copy files from the source tree, which command would then be best to change the ownership? [14:53] yeah I'd go with chmod then [14:53] *chown === milli` is now known as milli [14:54] RainCT: Hmm...any ideas why my commands from above don't work? [14:55] I see the command being executed when running 'debuild binary', but neither the ownership is right after installing with dpkg, nor is the ownership correct when looking at the debian/mypackage directory [14:58] Which seems reasonable to me, as changing the ownership to www-data is 'permission denied' when attempting to do that as a normal user [14:58] Any ideas? [14:58] debian/rules runs as root (or fakeroot) [15:03] flohack: try to see how it is done in apache2 package. [15:21] Hello... aMSN 0.98 sources is released... any kind soul available to package it before FeatureFreeze? :) [15:22] Is CDBS capable of calling autoconf? Else, what's the best way to handle a tarball where autoconf wasn't run (and which requires patching of .in files to work for Ubuntu) [15:23] vivia: Don't you think it is a bit high expectation. [15:26] RainCT: without cdbs: "autreconf -i" should do the right thing [15:26] slytherin: yeah I know it's a last-moment [15:26] but there's a huge load of new features of bugfixes [15:27] vivia: if the features and bug fixes are important I am sure the packagers will apply for FFE. [15:27] RainCT: Do you need to run it at the build time? How about create a patch. [15:27] slytherin: It fails to unapply [15:28] hmm, that is bad. [15:28] sistpoty|work: Well I'm trying with that (well I tried with 'autoconf -f' but same with autoreconf) in makebuilddir:: which generated the files correctly (./configure; make works) but makefile.mk doesnt' seem to be running ./configure [15:28] slytherin: sorry, what is FFE? [15:28] IIRC, cdbs has provision to do certain tasks before build. Btu I have never used it. cdbs manual may help. [15:28] vivia: Feature Freeze Exception [15:29] vivia: feature freeze exception [15:29] thanks :) [15:29] RainCT: hm... don't use cdbs then :P [15:29] (not a clue about cdbs, sorry) [15:31] has anyone ever seen "python setup.py install --root=blah" install scripts in /usr/local/bin? [15:31] I thought --root would be enough and e.g. --prefix wouldn't be needed? [15:33] ok so I'm off, if FFE is possible for a big difference in features and bugfixes. Thanks all! :) [15:40] How can I make sure that pbuilder keeps the build result, so I can examine what it did? [15:42] Or is there a way to tell pbuilder to drop me to a shell within the build environment if the build fails? [15:44] flohack: --logfile will write the stdout to a file. you can also drop into the build env. by using hooks [15:45] stefanlsd: That would be great...by any chance have you got a working example how to drop into a shell. [15:46] flohack: man pbuilder and look for hookdir. I use the following http://pastebin.ubuntu.com/259866/ . Its called C10shell [15:47] stefanlsd: Thanks a lot! [15:50] OK got that working. How do I fix ldconfig-symlink-missing-for-shlib? [15:50] Hi, I'm trying to upload a package to my PPA but I keep getting FAILEDTOBUILD errors due to incorrect build-dependencies. Is there any way to check this beforehand? [15:50] flohack: np. good luck. oh yeah. dont forget to to tell pbuilder your hookdir, else it doesnt pick up the hook [15:51] stefanlsd: Thanks for the hint, I think I got it right. [15:51] sluimers: pbuilder [15:53] Something different: I'm trying to backport alfresco for hardy, but the build fails with 'The import sun.security cannot be resolved', which suggests that the rt.jar (which contains the proprietory java extensions) is not added to the build. I simply took the jaunty package and changed the openoffice dependency to 2.4 ( from 3.0 ). Any ideas what the problem could be? The control file Build-Depends on sun-java6-jdk, so the sun compilier should [15:53] in use... [15:55] flohack: What is correct name of package? [15:57] slytherin: alfresco-community [15:59] flohack: I can not find any such package on packages.ubuntu.com [15:59] wait, fund it [15:59] it's in the partner repository [16:01] flohack: sun-java6-jdk can not be installed interactively. On build servers this is solved by preseeding of debconf answer (to the question that requires acceptance of agreement). I believe this is not present on PPA. So the other JRE which is automatically pulled by ant is getting used. And my guess is that JRE is GCJ. [16:01] I did that, parts of it compile just fine [16:01] I executed: echo -e "sun-java6-bin shared/accepted-sun-dlj-v1-1 boolean true\nsun-java6-jdk shared/accepted-sun-dlj-v1-1 boolean true\nsun-java6-jre shared/accepted-sun-dlj-v1-1 boolean true" | debconf-set-selections [16:02] flohack: Where did you do that? [16:02] in the pbuild chroot (sudo pbuilder login --save-after-login) [16:03] flohack: What is the value of JAVA_HOME in debian/rules? [16:04] "alfresco-community-3.2/debian$ grep -R JAVA_HOME *" Does not return anythin [16:05] I'm currently in the pbuilder shell after the build failed (hook) [16:05] JAVA_HOME is not set there [16:05] "javac -version" in the chroot shows that it's ecj which is the default java compiler [16:06] flohack: you can try setting JAVA_HOME to the sun java installation directory in debian/rules. See if that works. Other option is to use openjdk-6-jdk as build-dep instead of sun-java6-jdk. [16:07] POX_: are you there? [16:07] zooko`: You sent me a contentless ping. This is a contentless pong. Please provide a bit of information about what you want and I will respond when I am around. [16:07] flohack: and ask the maintainer to use openjdk for builds henceforth. [16:07] POX_: Wow! Good script! [16:07] slytherin: Where exactly would I set that? which target? is setting it with 'export JAVA_HOME=/usr/lib/jvm/java-6-sun' correct? [16:08] POX_: I gotta go to the bus to work now. I hope you look at Tahoe-LAFS on http://revu.ubuntuwire.com/p/tahoe-lafs today. Bye! :-) [16:09] slytherin: Would depending on openjdk-6-jdk make that the default? Does the openjdk include the sun.security package? [16:19] slytherin: I wrote the email proposing the dependency change to the maintainer's address and tried setting JAVA_HOME. I'll report back! [16:22] Anyone actually know the proper way to make a deb file -- not using checkinstall [16:23] kevdog: read the topic [16:24] flohack: using openjdk should fix it. But I don't remember exactly if it will. Setting correct JAVA_HOME is the best way. [16:24] kevdog, https://wiki.ubuntu.com/PackagingGuide/Complete [16:24] Oh -- sorry -- do you know where I can ask this question then? #ubuntu told me to check here [16:24] has the vuze question been asked in here a million times? [16:29] kevdog: Have you gone through the packaging guide? If you have specific questions then ask here. [16:29] I just can't get the chown www-data:www-data within the rules to work. I looked at the apache2 package and can't find any differences there. I can see that the command is executed when running 'debuild binary'. After installing the dpkg, the directory is still owned by root:root. [16:30]