[00:03] why patch is not applied due to DEP3 tags? [00:04] ari-tczew: something must be wrong with the formatting [00:04] or..wait, what release? [00:04] micahg: maverick. package uses cdbs [00:04] micahg: sorry, karmic! [00:05] micahg: simple-patchsys.mk in debian/rules [00:05] * micahg doesn't know if karmic recognizes DEP-3, [00:05] * micahg doesn't know enough about this [00:05] ari-tczew: do you want to pastebin the patch and we can see if anything looks obviously wrong [00:08] micahg: this is OK patch: http://paste.ubuntu.com/483679/ [00:10] micahg: patch WRONG: http://paste.ubuntu.com/483680/ [00:10] Trying patch debian/patches/01-fix_XSS_IE.patch at level 1 ... 0 ... 2 ... failure. [00:12] ari-tczew: http://dep.debian.net/deps/dep3/ take a look at the structure section [00:13] you might need a # before each header line [00:14] micahg: I tried to use ## and doesn't work [00:14] ari-tczew: 1 # [00:16] micahg: still failed [00:17] ari-tczew: maybe try usind cdbs-edit-patch to make the patch and see what headers it adds [00:17] *using [00:17] * micahg isn't good with non-quilt systems [00:19] micahg: I'll solve this through another way [00:19] ari-tczew: k [05:28] nxvl: why not update in Debian and file a sync request? [05:28] re: terminator [05:28] micahg: my debian sponsor is sloooow [05:28] micahg: and i need a FE anyway [05:28] nxvl: you did specify why in the bug [05:28] oh [05:29] well, you mention it, but you should make it clear that's the FFe reason [05:29] nxvl: I had something similar, I got a package updated today in Ubuntu to make the freeze and will update Debian soon [05:30] well, i'm trying to get it into debian at the same time === freeflyi1g is now known as freeflying === ivoks_away is now known as ivoks === ivoks is now known as ivoks_away === hannesw_ is now known as hannesw [09:12] Hello [09:23] good morning [09:25] أثممخ [09:25] Hello [09:39] Wow. The ubuntu font handles that surprisingly well. [09:40] I see that you use CDBS. Nice! :-D [09:40] Then I suggest to use auto-resolving of build-dependencies (currently they are not quite optimal): [09:40] 1) copy debian/control to debian/control.in [09:40] 2) edit debian/control.in replacing cdbs build-dpendency with @cdbs@ [09:40] ^ ewwwwwww [09:41] O_o [09:41] Mmmmmm, autogenerated debian/control! Everyone loves it! [09:42] RAOF> Sarcasm? :) [09:42] Indeed! [09:44] Most perspicatious of you :) [09:45] lucidfox: where did you see that? [09:45] hyperair> Debian pkg-multimedia-maintainers mailing list [09:46] lulz [09:47] More dh, please! All the conciseness of cdbs with significantly less crazy arcana. [09:47] What RAOF said [09:47] Really, it's once again, a toolkit vs framework ideological dispute [09:48] but dh>=7 please. [09:48] With frameworks, all too often you end up fighting the system if you need to do something nontrivial [09:48] hyperair++ [09:48] * hyperair is utterly terrified of all these dh_blahblah debian/rules [09:48] in fact, i'm dreading having to stare at libgpod's debian/rules. [09:48] * maco used dh_make to avoid learning how to do write debian/rules until dh7 happened [09:49] hahahaha [09:49] i read through the cdbs documentation, and amazingly emerged with less than 50% understanding of CDBS. [09:50] lucidfox: lfaraone says his sponsor until he became a DD was the guy that wrote cdbs and so would refuse to sponsor anything not using it [09:50] for the other 50%, i had to dig through the files in /usr/share/cdbs/1 [09:50] (or possibly current maintainer of cdbs... meh, someone with bias) [09:50] heh === yofel_ is now known as yofel [10:50] does anyone fancy helping me with a bunch of haskell rebuilds? [10:54] :( [10:54] too much going on [10:55] at least i can move my neck today though! [10:56] necks are overrated [11:00] sharks don't have necks [11:00] so they don't look back [11:01] sharks would probably be dangerous on the M40 [11:13] lucidfox_: oh my, the person who received those tips is now in #debian-mentors [11:14] mmm, cdbs hackery [11:15] i can give a concrete example of why cdbs is a problem [11:17] you need to employ unpleasant hacks, or cli:Depends won't work on source packages with C and mono outputs [11:17] as cdbs will discard the dh_clideps output [11:17] how about: any package that requires multiple build passes [11:17] yes, or that [11:18] Laney, I've some spare cycles: I'll take 5. Just need test-rebuild and build1 upload? Which packages? [11:18] libubuntuone is an example of the issue, anyway. [11:18] hence [11:18] binary-predeb/libubuntuone1.0-cil:: binary-fixup/libubuntuone-1.0-1 [11:18] dh_clideps -plibubuntuone1.0-cil [11:18] libubuntuone is an issue in itself [11:19] * ajmitch would be rather tempted to rip out cdbs from that package [11:39] persia: Test build, check it's installable (but should be if it builds), upload [11:39] persia: let me know which ones you take so we don't collide [11:39] this is what makes me want binNMU-a-like in LP [11:40] List me 5 and I'll make my otherwise idle build-server sweat a bit. I'm not familiar enough to pick them wisely. [11:40] ok [11:40] haskell-configfile haskell-haxr haskell-unix-utils haskell-hsh haskell-vty [11:40] chrisccoulson: why didn't you sync nspr from Debian? :( [11:41] ari-tczew, because we package it completely differently [11:41] ari-tczew, all mozilla products are blacklisted [11:42] so, please don't go merging or sync'ing mozilla products from debian without speaking to me first ;) [11:47] Laney, E: Unable to find a source package for haskell-unix-utils [11:48] * persia strongly encourages folks to talk to all of mozillateam, rather than only chrisccoulson, as it saves chrisccoulson repeating the statements, thereby increasing his patience [11:48] heh :) [11:50] persia: that'll be haskell-unixutils, sorry [11:50] That one exists. Starting testbuilds. [11:51] this is where I really need to set up a gpg-agent [11:52] heh. [11:52] gpg-agent + debsign -r is bliss [12:05] Laney, All successful. Uploading. [12:06] persia: cool, thanks [12:07] What transition causes the rebuild (for the changelog entries)? [12:08] Various. I just put “No-change rebuild to pick up new library ABIs” [12:08] heh. OK. [12:08] It makes more sense to do them all at once rather than repeatedly rebuilding over the cycle [12:10] If you're paying attention, yes. I usually catch stuff in NBS. [12:12] * persia vaguely grumbles about environments like haskell and java that require installation of lots of stuff before one can build source === dholbach_ is now known as dholbach [13:52] hi all. i'm new. i need some help, i try to apply to developer community of ubuntu but is so comfusing can anyone can guide me step-by-step what to do. i created wiki page and i have launchpad account, what i should do ? in the site there is too much information that comfusing me [13:53] Status0, First step: don't worry about any of that. The application process isn't at all important to get started. [13:53] Next: what sort of stuff do you like to do? [13:54] persia: develop. use linux. contribute linux. [13:54] By "linux" do you mean the kernel, or the set of software often used with linux? [13:55] linux, means everthing i have no problems to develop to kernel and modules of selinux, and i have no problems to develop new software [13:57] basicly i want to be a part of community, and help where i can. and give ideas. [13:58] OK. So, most developers spend time working in kernelspace *OR* userspace. You are welcome to do both, but I can only usefully tell you how to get involved in doing stuff with Ubuntu userspace. For kernelspace, you want to talk to the #ubuntu-kernel folk. [13:58] You've become part of the community just by joining our channels, and being interested. Welcome. Next step: stay part of the community by doing stuff :) [13:59] As distribution developers, we tend not to develop much new software, rather focusing mostly on bugfixes and integration. You're certainly welcome to participate by creating new software that works well as part of Ubuntu, but that path doesn't usually lead to the title of "Ubuntu Developer". [14:00] Some of the folks who do new software specifically for Ubuntu hang out in #ubuntu-app-devel [14:00] ahh, i came in at just the right time... [14:01] In this channel we're mostly focused on QA stuff: bugfixing, integration coordination, trying to reduce distribution maintenance overhead through communication with upstream, etc. [14:01] If that interests you, my best advice is to pick a bug, and start tearing into it. [14:01] I haven't done much MOTU work since Maverick was released. I was going to ask what items are needing attention? [14:01] where and how i do it ? [14:02] the folks in #ubuntu-bugs can help track down the affected software, and collect a lot of information about the cause, and the folks here can help with finding a patch getting it applied, communicating it to all the right places, and getting it uploaded. [14:03] https://launchpad.net/ubuntu/+bugs has *lots* of bugs. [14:03] * persia checks to see if there are some bitesize or packaging ones open [14:04] Status0: i thought this was a pretty inspirational blog post on soomeone getting involved with fixing bugs. http://brunogirin.blogspot.com/2010/08/contributing-to-shotwell.html [14:05] https://launchpad.net/ubuntu/+bugs?field.tag=bitesize are supposed to be easy bugs, which might be a good place to start. [14:05] !development [14:05] Interested in becoming an Ubuntu Developer? Get started here: https://wiki.ubuntu.com/UbuntuDevelopment [14:06] That just leads to the wiki pages, and ends up confusing people about the application processes. [14:06] The key fact that we encourage everyone to just join in, and only worry about applications after they've been around a while and need special perks isn't made clear enough. [14:08] persia: oh, but isnt about developer membership , mostly links to various sources and how they can help. maybe the factoid needs a cleanup ? :) [14:08] *it isnt [14:08] vish, The factoid is probably fine. [14:09] good afternoon everyone [14:09] But lots of folks starting there end up getting worried about "How to become a developer". I'm not sure how (or I'd have changed it), but I think we need to make it clearer that there are absolutely no requirements before starting as an Ubuntu Prospective Developer, or working in bugsquad. [14:10] yeah.. [14:11] thanks about the information, but its still comfusing what in the site wrote that i should become a member and do a long process in order getting membership . why its not easy. [14:12] it's should be open community. and i see alot of Bureaucracy it's not fun :-( [14:12] It's only a communications issue. [14:12] q. is "1.2.3beta1" a valid version for a .deb filename [14:13] It's a completely open community, but we honor those who have made significant and sustained contributions by granting them an email address, IRC cloak, etc. [14:13] Perhaps calling that status "member" may feel a bit exclusionary. [14:13] jetienne: It's a technically valid upstream part of a Debian version, yes. But you probably want it to be considered "lower" than 1.2.3, right? So I'd suggest 1.2.3~beta1. [14:13] jetienne: no i thing you should provide also a name like "demo-software.1.2.3-beta" [14:14] not being a member does not exclude you from participating [14:14] Indeed. membership is granted *after* participating for a while, rather than being a first step. [14:14] * duanedesign thinks he found a bitesize bug he can work on \o/ [14:14] Status0: as persia was explaining , you dont have to be a member to help. there is no Bureaucracy here :) [14:15] Laney: Status0: ok so a - or a ~ is needed. i was looking at http://semver.org/ this is the version seen from github guys [14:15] Status0, The issue is that 1.2.3~beta1 sorts earlier than 1.2.3, but 1.2.3-beta1 sorts after than (and confuses things, as 1.2.3-xxx is typically used when xxx represents the distribution revision) [14:15] * persia prefers + or ~, depending on semantic intent [14:15] persia, we used to call it "Ubuntero" initially :) [14:15] the word Member came up later [14:16] I know. I'm becoming increasingly unsure "Member" is better. [14:16] ++ [14:17] I thought that Ubuntero was just someone who had signed the code of conduct [14:17] i think that espatialy ubuntu www site should explain this instad buffer-overflow of information [14:17] Laney, That was later, but is the current meaning. [14:18] Laney: i singed this also. [14:18] I see [14:18] * Laney is too new-school, clearly [14:18] Status0, I'd agree, but we're a lot better at maintaining software than we are at maintaining the website. There's plenty on the website that is not only confusing, but actually completely wrong. [14:28] how much time do you think is takes a newbie to create his first package? [14:29] Korbit: not a lot, depends a lot on the software [14:29] For a nice clean upstream, an hour or two at most. [14:29] Korbit: http://io.debian.net/~tar/irc/debian-packaging/ [14:30] ok, i'm trying to package my first package of a basic java program that is not currently in the repositories [14:30] i'm reading debian's packaging guide [14:30] good luck :) [14:30] persia: we are talking newbies, though. I had some experience when I started, and I still probably spent a few days getting my first package into sponsor-approval-ready state [14:30] (my sponsor was very pedantic) [14:30] tumbleweed, Then you either didn't start from good docs, or didn't have good people telling you what to do. [14:31] persia: I had a good person, but he tried hard to make me work out things myself [14:31] * persia is exceedingly pedantic, but has led people through packaging simple stuff in less than an hour: it all depends on the upstream. [14:31] i'm reading this now: http://www.debian.org/doc/maint-guide/ch-dreq.en.html [14:32] about sponsor, my neoip-webpeer needs some love [14:32] :) [14:32] persia: I'd say *now* it takes me around an hour or two to package an arbitrary thing. [14:32] (assuming a fair degree of weirdness, because all upsteams are weird) [14:32] Really? Even a hello world shell script? [14:33] well, obviously more than that :) [14:33] Ah, yeah :) [14:33] how should i know in which section i should place my program in the control file? [14:33] Korbit: there's a list of sections linked to from the debian policy [14:33] Korbit: which software is it? [14:33] sablecc, a parser generator for compilers === simar__mohaar is now known as simar [14:34] Korbit: heh, good luck with sablecc (btw I've got a package of it) [14:34] Korbit: devel [14:34] tumbleweed, you do? but it's not in the repositories [14:34] https://launchpad.net/~jerome-etienne/+archive/neoip/+packages <- i got neoip-webpeer deb in my ppa... how can i include it in normal ubuntu repo ? [14:34] tarzeau, thanks [14:36] jetienne, There's quite a bit of discussion about that currently (see the ubuntu-devel@lists.ubuntu.com archives for this month). You might get lucky uploading to REVU. Debian is frozen for squeeze, so NEW is backed up. You may do best to wait, although starting the process of getting into Debian may be ultimately easiest to continue to maintain. [14:37] Korbit: it was removed from debian because nobody was maintaining it: http://bugs.debian.org/508361 [14:37] persia: wild guess on how long it may be ? rough estimate is ok [14:37] Korbit: my package looks workable, I can't remember why I didn't try to get it back into debian (probably because I don't use it personally) [14:38] tumbleweed, should i continue packaging it then? [14:38] I can't. Some packages have been in REVU over a year, but I've seen others go through in a week or so. I hope Debian will release real soon, but that won't happen until the bugs are closed. [14:39] Korbit: if you are keen to maintain it in debian yourself, you are welcome to my packaging. https://edge.launchpad.net/~stefanor/+archive/tsl/+sourcepub/1218761/+listing-archive-extra (packaging is in bzr: lp:~stefanor/+junk/sablecc ) [14:39] shadeslayer: shadeslayer: the new version of xserver-xorg-input-synaptics is released upstream here http://xorg.freedesktop.org/releases/individual/driver/xf86-input-synaptics-1.2.99.901.tar.gz [14:39] persia: hmm ok thanks [14:40] ppa will stay then :) [14:40] Korbit: sorry, my "good luck with sablecc" was incorrect, it's a perfectly sane package. It was other stuff that depended on it that was insane. [14:40] I'd recommend following the discussion on ubuntu-devel@ : I suspect the conclusions from that will be better guidance than anything anyone can tell you now. [14:41] simar: i have exams for the next 10 days ;) [14:41] tumbleweed, i'll try to follow the debian instructions to create my first package with sablecc =) [14:41] simar: if its a bug release update, package it [14:41] tumbleweed, i guess i can do it [14:42] if its a absolutely new release with new features, dont bother [14:42] tumbleweed, if i manage to create it, i'll create a package for the new beta release also [14:42] i doubt itll go through FFe [14:43] Korbit: cool. You learn more doing it yourself :) [14:43] tumbleweed, before i know it, i'll be fixing bugs and submitting patches for other programs ;) === ivoks_away is now known as ivoks [15:01] tumbleweed, there is so much to read to create a single package! [15:02] tumbleweed, it's a miracle we have contributors to debian [15:02] and i'm a technically oriented person, i can't imagine non technical people trying to do this [15:03] Korbit, http://ubuntulinuxtipstricks.blogspot.com/2010/08/is-packaging-new-software-hard.html outlines a fairly simple procedure to package fairly quickly. [15:04] persia, thanks, i'll read that and i'll also blog about this once i get my first package accepted [15:05] yeah! dpkg-buildpackage exited successfully!! [15:05] As I mentioned to jetienne, getting a package accepted is a bit fuzzy just now... [15:05] echo $? -> 0 (i'm getting somewhere!) [15:06] with a now of undefined length :) [15:07] i got a sablecc_3.2-1_amd64.deb i'm happy [15:08] Korbit: yes, packaging is a big world :) [15:09] * tumbleweed must find some nice easy example packages for global jam this weekend [15:14] following debian packaging instructions is really discouraging [15:15] Korbit: which ones? (I'd call that a bug) [15:16] http://www.debian.org/doc/maint-guide/ch-build.en.html (6.5 pbuilder package) [15:16] i'll try to use the ubuntu instructions to run pbuilder [15:17] sudo apt-get install ubuntu-dev-tools && pbuilder-dist sid create && pbuilder-dist sid build mypackage.dsc [15:17] this makes sense to you? [15:17] sudo pbuilder build sablecc_3.2-1.dsc [15:18] * persia praises sbuild for not needing to be root to build stuff [15:18] looks good (but Laney is right about pbuilder-dist being easier to use for newbies) [15:18] * Laney cuddles sbuild [15:18] but the chroots mk-sbuild creates seem to be far from minimal [15:19] Same chroot as you get from pbuilder-dist. [15:19] there should be a GUI to help building packages [15:20] http://people.ubuntu.com/~persia/pull-soyuz-chroot gets ideal test chroots for Ubuntu, but isn't so useful for Debian. [15:20] (someone should add pbuilder support) [15:20] didn't build in pbuilder, i'll work on it [15:20] root@chicken:/# apt-cache policy openssh-client [15:20] openssh-client: Installed: (none) [15:20] (sid-amd64)root@chicken:/home/laney/temp/haskell-unixutils-1.22# apt-cache policy openssh-client [15:20] openssh-client: Installed: 1:5.5p1-5 [15:20] pbuilder, sbuild [15:22] OK. Check the soyuz chroot. I think either pbuilder isn't adding enough, or sbuild is adding too much, but I don't know which is true. [15:26] pbuilder failed again :( [15:27] Build-Depends: debhelper (>= 7), ant, default-jdk-builddep [15:27] dpkg-buildpackage is breaking on that last dependency [15:28] dpkg-checkbuilddeps: Unmet build dependencies: default-jdk-builddep [15:28] i'm stuck here :( [15:31] in http://ftp.de.debian.org/debian/pool/main/a/ant/ant_1.8.0-4.dsc [15:32] it's also using that dependency [15:32] so mine should work [15:33] IIRC that got dropped [15:33] Korbit, Try apt-get install default-jdk-builddep [15:33] (and yes, it was dropped for maverick) [15:34] and gcj-native-helper provides it [15:34] Oof. That's annoying. [15:35] persia, thanks, it worked [15:35] i don't know why i need gcj for this... completely stupid since i already have openjdk [15:36] They do different things. [15:36] trying pbuilder again [15:37] starting creating my first package with a java programa was definitely not a good idea :D [15:44] that's what i thought :) [15:52] Korbit: yes, java packages tend to be tricky to build (although the one you've got is quite straight-forward, as they go) [16:10] tumbleweed, pbuilder ran successfully =) [16:10] echo $? -> 0 === jcfp is now known as Guest95978 [16:16] ScottK: btw, did you see my post on -devel about rainbow? OLPC's using it quite successfully in the field. (as in, it doesn't break things, and there are a bunch of applications that various developers have created for the platform under its restrictions) [16:16] lfaraone: I did. [16:16] It's interesting. [16:17] * lfaraone likes it, but then again, I'm the maintainer in Debian, so I'm biased. :) [16:17] ScottK: recently they were thinking about using Xephyr or VNC to sandbox X apps. [16:18] lfaraone: At this point I'm more interested in getting the requirements right than implementation. [16:18] We tried using Xephyr for ubuntu-mobile work during feisty-hardy: at that time it was too painful for words. Maybe it's better now. [16:19] persia: "sudo apt-get install sugar-emulator-0.88; sugar-emulator" will spawn an Xephyr window with a psedu-resolution of an OLPC XO display. [16:20] persia: so it can be used out of the box in Maverick. Not sure how the work on using Rainbow for it has gone. [16:20] I believe that. Does it handle segregation of X properties cleanly yet? [16:20] persia: hm? [16:21] At least then we had issues if the X configurations were significantly different. I'll have to play: maybe it works better now. === echidnaman is now known as JontheEchidna === apachelogger is now known as apache2logger === Guest95978 is now known as jcfp === fta_ is now known as fta [20:04] \sh: I assume you saw the 1.10.8 release of zf === ivoks is now known as ivoks_away [21:13] Anyone's brain up to fixing autoreconf in a CDBS package? We need to rebuild osgal for NBS, but it FTBFS due to it's autoreconf not working. If I autoreconf -f -i -Wall,no-obsolete and then build the binary without running clean, it builds fine. [21:13] If so, please grab osgal and have a look. If you aren't a MOTU, I'll be glad to sponsor it. === fta_ is now known as fta === fta_ is now known as fta === fta_ is now known as fta === fta_ is now known as fta === fta_ is now known as fta === fta_ is now known as fta === JanC_ is now known as JanC