[00:00] wgrant, the C::B developers are quite frilled that C::B is going into the archive, and agreed to accept changes to make it properly LSB compliant :-) [00:00] wgrant, oh [00:00] NCommander: A sane upstream! Yay! [00:00] yeah [00:01] wgrant, I don't think they're going to backport the changes, but the next stable won't need a patch with a little luck [00:02] wgrant, thanks. now I upload correct file [00:02] And they're working to help resolve the other issue [00:02] finally I think that I have a consistent package. [00:02] do you know how I can a advocates from motus? [00:03] k0p, post it to revu, and then poke the MOTUs [00:04] NCommander, post on revu? revu.tauware.de? [00:04] yeah, upload it to revu after signing it with your GPG key [00:05] It's already done [00:05] Then just upload it to revu [00:05] I'm waiting [00:06] but it's already upload [00:06] but how poke MOTUs? [00:07] Just ask around here ;-) [00:07] cody-somerville, wake up [00:07] no : ( [00:08] k0p, I found you an MOTU ;-) [00:08] http://revu.tauware.de/details.py?package=umit it's the package :) [00:08] NCommander, who? [00:09] today I understand what was a native and non native package.. lol. finally :) [00:09] k0p, cody-somerville [00:10] cody-somerville, can you review/sponsor my package? [00:11] k0p, well, your package is lintian clean [00:12] (I'm not an MOTU, but as a porter, I know quite a bit about packaging) [00:12] yeah [00:12] You pass step one ;-) [00:12] NCommander, i'm working hard to mantain this package last days [00:12] * NCommander has an anonying habit of packaging things that cause false-postive checks in lintians -_-; [00:13] * NCommander checks the Python packaging guide just to confirm I'm right about this possible oops [00:13] ok, that's right [00:14] NCommander, what? [00:14] I was checking to see if there were any special rules with listing python modules [00:14] python, python (>= 2.5) | python-pysqlite2 [00:14] THAT threw me through a loop [00:15] hmm [00:15] delete python, python (>= 2.5)? [00:16] Well, if I read it right, you either depending on a python > 2.5, or pysqlite2 [00:17] opssssssssssss [00:17] lol [00:17] of course. thanks :) [00:17] NCommander: While you're reviewing uploads...any chance you could look at my first ever Ubuntu (Java) packaging attempt: http://revu.ubuntuwire.com/details.py?package=iriverter , please? [00:18] k0p, reviewed. [00:18] jmarsden, I'm not an MOTU, but I'd love to take a quick look [00:18] cody-somerville, very thanks ! [00:18] i'l read the comments :) [00:18] OK... cody-somerville can review it too if he likes :-) [00:19] jmarsden, loose the ${misc:Depends}; that's just an example put there by dh_make [00:19] NCommander: OK, I didn't add that it was there in the original package, but that makes sense. [00:19] jmarsden, you have a copyright symbol in your control file, it should be a (C), simply so the copyright file is pure ASCII [00:20] jmarsden, you have two patchs changing the same line of code O_O; [00:21] cody-somerville, i'll solve the problems. about RC it's because it's RC2.. We already have a RC1 [00:21] it's release candidate 2 [00:21] OK... same deal, that copyright char was there in the iriverter package I started with... will fix. Why doesn't lintian complain about that? [00:21] now we're working to release stable [00:21] NCommander: Two patches... one by the previous maintainer, one by me... should I merge them into a single patch file? [00:22] jmarsden, it edits the same line; delete the old patch [00:22] OK, can do. [00:22] jmarsden, In addition, edit each patch and fill out the header [00:23] NCommander: fill out with what?? Can you point me at an example of one that is filled out? [00:23] http://revu.ubuntuwire.com/revu1-incoming/iriverter-0807192230/iriverter-0.16/debian/patches/01_change_launcher_classpath.dpatch [00:24] Oh wait [00:24] nm on the patch licenses [00:25] er, patch headers [00:25] sorry, my brain is slightly frayed [00:25] OK :-) [00:25] One minor thing [00:25] Fix the bug close field [00:25] YOu have (LP#91237) [00:26] It should be (LP: #91237) [00:26] OK. Good catch. [00:26] jmarsden, did you build it in pbuilder/sbuild? [00:26] NCommander: No, it fails with an issue about a virtual dependency in pbuilder... any advice on how to proceed to get over that? [00:27] Oh. THAT issue [00:27] pbuilder only has the main repo by default [00:27] It does build OK in my PPA though... [00:27] You need to enable universe [00:27] THe PPA usually means it willl work, but fixing pbuilder is a good thing [00:27] jmarsden, https://wiki.ubuntu.com/PbuilderHowto [00:27] Enjoy ;-) [00:27] Thanks. [00:28] It looks good to me once you correct it, and if I were an MOTU, you'd get my rubber stamp. [00:29] * NCommander pokes the cody-somerville [00:32] NCommander, moo? [00:34] jmarsden, can yo look at jmarsden's package once he uploads an updated version; I looked it over and caught as many mistakes as I could find. [00:36] er, cody-somerville , not jmarsden [00:36] * cody-somerville grins. [00:36] I suppose so. [00:36] wgrant: ok. [00:39] cody-somerville: Thanks, I'm sorting out my pbuilder right now ready to try building it that way. [00:41] cody-somerville, out of curiosity, when did you become an MOTU? [00:43] NCommander, March 27th [00:47] wgrant: done. grab debdiff from bug #246408 [00:47] Launchpad bug 246408 in matplotlib "Please merge matplotlib 0.98.1-1 from Debian/unstable" [Undecided,New] https://launchpad.net/bugs/246408 [00:49] cody-somerville, "3. Why did you change the upstream version from 0.9.5RC2 to 0.9.5-RC? " [00:49] are you talking about tarball name? [00:50] k0p, it looks like you change everything from 0.9.5RC2 to 0.9.5-RC2 [00:50] yes. it's true. I change. it's there was a trouble with last release [00:50] but our secript generate files like 0.9.5-RC2 [00:51] but it's no trouble.. in few days we release the stable version :) [00:51] You must make the upstream version 0.9.5~RC2 [00:51] Or 0.9.5~rc2 [00:53] wgrant, name folder too? [00:54] k0p: Pardon? [00:54] 0.9.5~RC2 is it affect the tarball orig name? [00:56] You must rename the upstream tarball to that. [00:56] And the version in the changelog. [00:56] That's it. [00:56] sure [00:56] k0p, but you don't need to patch the source [00:57] cody-somerville, I need patch setup.py [00:57] I'll use dpatch as you recommend [00:57] k0p, okay :) [00:58] but I don't need to patch nothing more [00:58] www.medibuntu.org is down... this some legal thing to be worried over or just server maintenance or such? [00:58] cody-somerville, debhelper is my rules? do you suggest that I only use cdbs? [00:58] genii: We are not Medibunut. [00:59] *Medibuntu [01:00] k0p, no, I'm saying that if you don't need to call one of the dh_ scripts then don't :) [01:00] wgrant: I know. Just figured if any group, someone here would know something about it [01:00] Why? [01:00] Because both groups do packaging for Ubuntu, it's not unlikely one of the 178 nicks here are in Medibuntu too... is it? [01:01] does medibuntu have its own channel? [01:01] kaminix: medibuntu is not so big. [01:02] cody-somerville: http://revu.ubuntuwire.com/details.py?package=iriverter is ready for your review :-) Thanks! [01:02] cody-somerville, yeah.. i'm overwriting dh_scripts thats no need right? so what I need to do it's a cleanup on my rules right? [01:02] Still, I'd guess it's driven by the people who do ordinary packaging for Ubuntu as well. [01:02] cody-somerville, and yeah.. the desktop files it was wrong. my mistake too. :) [01:03] kaminix: We have no official ties. [01:03] Anyone in here on the mailing list who can say if my mail about edict packages reached the list? [01:03] wgrant: Not officially, I know. [01:03] kaminix: Anyhow that was along my line of thinking also. The #medibuntu channel seems to have 8 people in it [01:03] I also think it advisable that we be kept as separate as possible. [01:05] wgrant: Anyhow I'll look elsewhere for answers on this issue now. [01:07] jmarsden, revu isn't the appropriate place for package updates. [01:08] jmarsden, please file a bug on launchpad and attach a debdiff, subscribe ubuntu-universe-sponsors [01:08] Ah. OK. well, there is already a bug, LP #91237 [01:08] Launchpad bug 91237 in iriverter "java library not found" [High,Confirmed] https://launchpad.net/bugs/91237 [01:08] So just attach the debdiff to that one? [01:09] what command line entry would you use to search a folder for a specific string? [01:09] nickellery: grep -r "specific string" foldername [01:09] jmarsden, yup [01:10] jmarsden, thanks you! [01:10] good night guys [01:12] jmarsden, did this package do any patching in the current version? [01:15] cody-somerville: yes. In the current 0ubuntu2 package it had one patch. Chanegd the smae line I changed to fix this bug, so NCommander said to merge my second patch into that one. [01:17] cody-somerville: BTW I just uploaded a fresh debdiff to that bug. [01:18] Does this mean you can now "sponsor" that fix for me? [01:20] jmarsden, Can you make your changelog more like this?: http://pastebin.ubuntu.com/28576/ [01:22] cody-somerville: Sure. But to me that looks like too much low level detail in the changelog? I'll go ahead and redo it, just trying to understand the reasoning for the request -- can you explain? [01:23] jmarsden, because I like changelogs to be useful without having to actually do a debdiff [01:24] OK. [01:26] jmarsden: Particularly because we team maintain packages in Ubuntu, debian/changelog should be clear to someone who's never touched the package before what was done and why. [01:37] ScottK: OK. That's a very different approach from other projects I've coded for, where the changelog entries were for a summary of what a change was for, not for the details of which files were touched which can be determined from diffs. Is this general approach to changelogs documented somewhere in tutorials for Ubunteros? Should it be? [01:39] cody-somerville: So is http://pastebin.ubuntu.com/28578/ more what you are looking for? [01:42] jmarsden, thats more verbose than what I would have done but sure [01:43] OK. Any more things to change you've spotted, or should I attach a new debdiff to LP# 91237 with that changelog included? [01:46] jmarsden, I think you're good. [01:46] * cody-somerville needs to go make dinner. [01:47] cody-somerville: Cool, I'll upload it and you can sponsor it after you cook and eat :-) Thanks for your advice and help. [03:19] wow [03:20] I know [03:20] The Avatar series finale was amazing ... [03:20] I'm awesome. [03:20] oh :/ [03:20] * NCommander notices cody-somerville [03:20] oh [03:20] Its just wow [03:32] Hey guys im having an issue with dh_shlibdeps finding libraries that are in other binary packages of the same source package, im packging this for debian but the guys in #debian-mentors are unresponsive, so i hope you guys can help :), the build log can be found at http://pastebin.com/f45e4a0cb and the package is here: http://mentors.debian.net/debian/pool/main/g/googlegadgets/googlegadgets_0.10.0-1.dsc any help would be greatly ap [03:32] preciated. [03:45] jscinoz, let me take a look [03:45] thanks NCommander [03:45] jscinoz, pastebin never seems to work from here, please repost to paste.ubuntu.com [03:45] ok one moment [03:45] also [03:45] Just the section of the log where it begins running shlibs [03:45] Hi [03:46] NCommander, http://paste.ubuntu.com/28591/ [03:46] oh [03:46] thats the whole thing >_< ill paste the other bit, one sec [03:46] No this is fine [03:46] SOmetimes paste.u.c chokes with a massive log [03:47] But this is small enough I was able to load it [03:47] NCommander, ah alrighty :) [03:48] jscinoz, if you would like to upload your package to REVU (Ubuntu's version of m.d.n, other MOTUs can help review it, and it can be put into Ubuntu; and then the Utnubu team can work on pushing it back into Debian; I took this route with codeblocks) [03:48] ah alright, hang on though i need to rework the dependencies as some of the packages dont exist or have different names in Ubuntu [03:48] jscinoz, nope, it pretty much will work out of the box [03:48] libqt4-opengl-dev i think is a virtual package in ubuntu >_< [03:48] oh [03:48] We keep the numbers the same [03:49] * jscinoz tries build in hardy chroot [03:49] jscinoz, use intrepid [03:49] that's our current unstable branch [03:49] ah [03:50] gah [03:50] "xulrunner-dev which is a virtual package." [03:50] In hardy? [03:50] oh wait, its xulrunner1.9-dev in ubuntu? [03:50] in intrepid [03:50] Oh [03:50] * NCommander looks what its in Ubuntu [03:50] er, debian [03:51] lol [03:51] In Debian, its xulrunner1.9-dev [03:51] hrm [03:51] uhh... [03:52] hmm *starts intrepid VM* [03:52] * jscinoz realises it'd be faster just to check packages.ubuntu.com [03:52] You could just use debootstrap to grab intrepid ;-) [03:52] http://packages.ubuntu.com/intrepid/xulrunner-1.9 [03:53] ah [03:53] so it has the 1.9 [03:53] one sec, will change. [03:53] Ok, I *think* I see your problem [03:54] yeah because i had xulrunner-dev as dep [03:54] but in ubuntu you need xulrunner-1.9-dev [03:54] i think. [03:54] RIght, but it should work fine in Debian [03:54] SInce there is a proper xulrunner-1.9-dev package [03:54] hopefully :P [03:55] yay for cowbuilder :) [03:55] Ok, dh_shlibdeps is building the library your missing [03:55] * NCommander perfers pbuilder, but to each his own [03:55] Where's your control file? [03:56] or better [03:56] link to the .dsc [03:56] jscinoz, I don't see it on mentors, where is it? [03:57] one moment. [03:58] http://mentors.debian.net/debian/pool/main/g/googlegadgets/googlegadgets_0.10.0-1.dsc [03:58] Hello everyone :) I want to contribibute to open source communtity. . .How can i join MOTU mentoring [03:58] * NCommander grabs [03:58] also i tried using xulrunner-1.9-dev on unstable [03:58] but it failed as virtual package [03:58] jscinoz, ouch, there is no love here [03:58] jscinoz, I'll help you with that issue in a moment [03:58] Lets just get it to build first on SOMETHING [03:58] so basically, Ubuntu Intrepid needs xulrunner-1.9-dev, while Debian Unstable needs xulrunner-dev [03:58] alrighty [03:58] :P [03:59] you can probably specify xulrunner-dev | xulrunner-1.9-dev and get it to work on both [03:59] oh yes forgot about that [04:00] Ok, lets see here [04:02] blarg my ISP's mirror of intrepid archive is broken >_< [04:02] ouch [04:03] your lacking an shlibs file [04:03] shouldnt dh_makeshlibs be creating it? [04:04] It won't find libraries for things not in the archive [04:04] (call it a bootstrapping problem) [04:04] Hello everyone :) I want to contribibute to open source communtity. . .How can i join MOTU mentoring [04:04] any help please :( [04:05] rraj-be, well, what would you like to do specifically? [04:05] * NCommander hits his head [04:05] It's been awhile since I did anything with shared libraries, and I'm right now coming up with a blank [04:05] It will hit me on how you fix this in a moment [04:06] rraj-be:https://wiki.ubuntu.com/MOTU/Mentoring/Contributor [04:06] hmm *reads manpage of dh_shlibdeps and dh_makeshlibs* [04:07] Hrm [04:07] How'd you generate the .install files? [04:07] manually. [04:07] oh [04:07] i see the problem i think [04:08] yeah [04:08] YOu forgot to package the library ;-) [04:08] i didnt include libggadget-js-1.0.so.0 *facepalm* [04:08] Do yourself a favor [04:08] rraj-be: https://wiki.ubuntu.com/MOTU/Mentoring/NewModel?highlight=%28motu%29%7C%28mentoring%29 [04:08] Add --fail-missing [04:08] It's your best friend [04:08] NCommander, there's an easier way to make .install files? [04:08] :_ [04:08] ^to your dh_Install line [04:08] NCommander, ok thanks [04:09] what will that do? [04:09] jscinoz, not really, but --fail-missing will cause the build to fail if any file isn't moved from the debian/tmp folder [04:09] ah thanks [04:09] Since your dealing with a lot of individual libraries, you can't really wildcard things to make life easier [04:09] aye [04:10] It's one of those necessary evils of packaging [04:10] shouldnt be too hard to fix now though i dont know why i didnt realise i forgot that lib before [04:10] You probably had accidently included it in one larger package [04:10] Or something [04:10] dpkg_shlibsdep is kinda weird [04:10] I've seen it do extactly this, and I've seen it work when it should have failed [04:11] Usually the later happens because you installed an earlier version of the package, and its finding the required library from that [04:11] ah [04:12] but that shouldnt happen when building in a chroot though? [04:12] That's why it might build with dpkg-buildpackage, but oblierate itself in a chroot [04:12] Hrm [04:12] I've looked at dpkg's code [04:12] I'm not going to try and figure it out ;-) [04:12] I *really* wish the New Maintainers manual told you about things like --list/fail-misisng [04:13] hmm well i've fixed it so that lib is included lets hope it builds now [04:14] anyways i must go away for 30mins (BBQ next door, yay australia) thanks for the help, hopefully when i get back the thing will have built :) [04:14] jscinoz, if you wish to submit your package for direct inclusion into Ubuntu, I can help you with that; otherwise when it finally gets into the Debian archive, you can request a sync, and it will get pulled into intrepid from sid [04:18] jscinoz, I can see some things that need to be fixed when you return [04:21] NCommander, hey im back, actually got 15mins or so until i have to leave for BBQ, also i thought downstream sync closed on june 6th? [04:21] and what were the other things i need to fix? [04:21] jscinoz, your copyright file [04:21] NCommander, doh i hate those things :P [04:21] Apache 2.0 license isn't a "common" license [04:21] oh. [04:21] YOu need to put the full text in your copyright file [04:22] Essentially cat COPYING >> debian/copyright ;-) [04:22] ANd then move things around [04:22] thats fine then, thought itd be more annoying than that [04:22] The upstream offer and actual Copyright lines look good [04:22] SO aside from including the actual license text, it looks fine [04:23] ok :) also gotta update that URL since im packaging a newer version now [04:23] yeah [04:23] THat's important [04:23] YOur watch file looks good [04:23] whee it built :D [04:23] dputting again to mentors [04:24] ack, you use quilt [04:24] * NCommander feels the burn [04:24] * NCommander is a dpatch man ;-) [04:24] <3 quilt [04:24] bah [04:24] You should clarify your changelog [04:24] with what? [04:24] Since you added a patch, you should explain why, and what it does [04:25] oh [04:25] alright one moment. [04:25] Add a docs file, and list the NEWS and README files; if there are any other documentation files that aren't already installed, add them too [04:25] alrighty, what is the syntax of a docs file? [04:26] just the filename [04:26] So just README, newline, NEWS [04:26] Pretty straightforward ;-) [04:26] ok, and is it one for the whole sourcepackage, or one for each binary package? [04:26] ... [04:26] argh [04:26] * NCommander checks the manual [04:26] what man page should i look at for this thing? [04:26] dh_installdocs? [04:26] They don't have manpages? [04:27] oh [04:27] wait it mentions it hang on [04:27] looks like one for each sourc epackage [04:27] Pretty much [04:28] Put these in the main library source [04:28] should i do it for all source packages? or just the main [04:28] oh answer it :P [04:28] er, binary [04:28] sorry [04:28] I'm a little off it today ;-) [04:28] ah yeah same :P [04:29] but anyways, why is dpatch better than quilt? [04:29] I find with quilt I'm fighting the tool more then actually writing the patch [04:29] THere is no actual difference in terms of writing the patch [04:29] But I find quilt just to be a downright pain to use [04:29] er, I meant applying the patch [04:30] nice job on the get-orig-source target [04:30] :D [04:30] I usually have to nag people to add that [04:30] is upstream a .tar.gz then? [04:30] yes [04:30] ok, that will work fine thne [04:30] (I'm usually a little warry of using uscan because it will grab the latest source package vs. the current, but I'm not going to veto a package over it) [04:31] You don't call the patch target in your rules file [04:31] WHich means your patch isn't being applied [04:31] yeah i do [04:31] Oh [04:31] QUILT_STAMPFN [04:31] -_-; [04:31] * NCommander bites tounge about his "love" of quilt [04:31] :P [04:31] anyways one other question. [04:32] Do you build any of the software in third_party? [04:32] umm. [04:32] i dont know >_< [04:32] probably. [04:32] Its under the MPL [04:32] YOu'll have to note that in the copyright file [04:32] * jscinoz wonders if it needs to be built or if shared one can be used. [04:33] It should still be noted because your not repackaging the tarball to remove them ... [04:33] * NCommander is checking to make sure nothing non-free is sneaking in [04:33] how can i check if they are actually built or not? [04:34] grep the Makefile.am's [04:34] I'd just note them because they'r ein the actual upstrem package [04:34] Put something like this [04:35] Source files in the third_party/codesighs, and third_party/editline, are covered under the Mozilla Public License [04:35] *include license text* [04:35] I assume SUBDIRS = third_party means it is going in there. [04:36] if i remove that line, those things wont be built then right? [04:36] note that third_party/valgrind/valgrind.h is under a BSD license (this one is a common license so instead of including the full license text, just refer to the file in /usr/share/common-licenses) [04:36] ok thanks. [04:36] It's all open and free, just different licenses [04:36] It's easier to note it in the changelog then anythign else ;-) [04:37] * NCommander blows the dust off his FSF Savannah licensing skills [04:37] :P [04:37] one of the perks of being an admin on Savannah is that you get "really" good at checking package licensing [04:37] Although this package would still fail to meet Savannah standards [04:37] wuh? [04:38] don't worry about it ;-) === asac_ is now known as asac [04:38] :P [04:39] Lets just say that the FSF is a little *ahem* crazy when it comes to making sure packages properly are documented and licensed with header files, copyright notices, and no ambiguity [04:39] hmm [04:40] WHen I was an admin on Savannah, we actually checked those things in the source packages [04:40] it says the things in third_party/codesigh can be used under GPL or MPL (choice) [04:40] or LGPL even [04:40] This must be from Mozilla [04:40] So .... *insert foot in mouth* [04:40] same for those in editline [04:40] so what should i say for them in debian/copyright? [04:41] Say that its available under the MPL1.1/GPL 2.0, and LGPL 2.1 [04:41] Put the full text of the MPL 1.1 in [04:41] And then simply say that the text of the GPL 2.0/LGPL 2.1 are in the /usr/share/common-licenses [04:41] (use the template at the bottom of the copyright file that was used to describe the licensing of your packagin) [04:42] alright thanks [04:45] I know, the copyright file is painful [04:45] If you want to see pain, go look at glibc's [04:48] <3 quilt so easy to use :P [04:48] what did you find hard about it just wondering? [04:50] It wasn't that it was hard [04:50] It just felt very clunky compared to dpatch [04:50] quilt is more powerful because of the way it can selectively add and remove patches [04:50]