/srv/irclogs.ubuntu.com/2008/11/09/#ubuntu-devel.txt

=== asac_ is now known as asac
NCommanderScottK, I'm glad you agree on debhelper since it would make life a lot easier w.r.t. to backporting stuff00:56
directhexNCommander, DH7 for hardy!01:00
NCommanderyup01:00
directhex7.0.13ubuntu1~dhx1 worked fine for me, before upgrading to intrepid ;)01:01
NCommanderIt means making backports will be tons easier since now we don't need to mangle the rules01:01
directhex0 changes needed from 7.0.13ubuntu1 according to my changelog01:01
NCommanderright01:02
directhexi'd be very happy about an official backport, since i could erase it from my ppa01:02
NCommanderWe could also do DH7 for dapper01:03
NCommanderSo if we backport something, it will stop meaning having to fudge things down to DH401:03
directhexehm... my spidey sense is screaming "check the required version of dpkg-dev" at me01:04
directhexi wonder why. let me check...01:04
=== jrib is now known as Guest64524
NCommanderdirecthex, it built w.o issue01:05
* NCommander sees if it installs and works01:05
NCommanderdirecthex, note, you can't drop dh7 from your PPA, since PPAs don't have backports enabled :-/01:05
directhexoh anuses -_-01:05
NCommanderIts a long standing bug01:05
directhexso is lack of debian targets for debian/ubuntu synergy01:06
NCommanderScottK, when backporting something to hardy from jaunty, do we also need to backport it to intrepid?01:06
directhexbut important things like new launchpad logos take precedence!01:06
NCommander(since hardy-backports > intrepid base)01:06
Hobbseedirecthex: you'll never win that battle.01:07
=== Guest64524 is now known as jrib
NCommanderHobbsee, we've been promised it was going to happen "real soon now"01:07
HobbseeNCommander: yeah, yeah.01:07
directhexwhich battle? i fight a lot of losing battles01:07
directhexthen grumble & move on to other things01:07
* NCommander notes we should just dump everyting in Hardy backports into a PPA, then make it so other PPAs can depend on it 01:08
Hobbseedirecthex: the one about wanting (and getting) launchpad to implement more possibly useful stuff, like fixing regressions, or adding useful features01:08
HobbseeNCommander: that would be the sensible optoin01:08
directhexi probably shouldn't mention that opensuse's PPA-like service can build against pretty much any major distro, from RHEL to Debian01:08
directhexit might be seen as pushing a pro-novell agenda or some crap01:08
NCommanderdirecthex, they don't take debian source packages01:08
directhexNCommander, sure they do01:08
NCommanderI'm told you put a .orig.tar.gz, and then a debian folder01:09
NCommandernot the normal tar.gz/diff/dsc01:09
directhexit takes orig/diff/dsc01:09
directhexand some other random undocumented things01:10
directhexi ran a test build, let me find it01:10
NCommanderScottK, poke01:10
NCommanderScottK, if we backport something from jaunty to hardy, should we also backport it to Intrepid?01:10
directhexNCommander, here we go. an old mono backport build, targeting etch & hardy, from orig/diff/dsc. https://build.opensuse.org/package/show?package=Mono&project=home%3Aoerc01:11
NCommanderah01:11
NCommanderNeat01:11
StevenKNCommander:  debhelper | 7.0.13ubuntu1 |      intrepid | source, all01:11
NCommanderJaunty has a slightly newer one01:11
NCommanderwhich fixes a few bugs01:12
StevenK"Meh"01:12
NCommanderStevenK, https://bugs.edge.launchpad.net/intrepid-backports/+bug/29549501:13
NCommander:-)01:13
ubottuLaunchpad bug 295495 in intrepid-backports "Please backport debhelper 7 from intrepid" [Undecided,Won't fix]01:13
StevenKWon't fix01:13
NCommanderin intrepid01:13
NCommanderin Hardy its In Progress01:14
StevenKThat was actually my point. Backporting debhelper from Jaunty to intrepid is silly01:14
directhexbecause it doesn't fix the core issue of the bug01:14
directhexsince it's still 701:14
directhexand the purpose of the bug is to have 7 in 7less versions01:15
directhexmuch as bug fixes are nice01:15
NCommandermy concern is the upgrade path01:15
NCommanderIs it a bad thing is hardy-backports > intrepid?01:16
directhexknow what? time for beds.01:16
=== jelmer_ is now known as Guest29580
=== _jason is now known as jrib
ScottKNCommander: There isn't a strict requirement to backport to every release.  I think it's a good idea for upgraders though.04:39
NCommanderScottK, Well, I just requested to backport from intrepid. so no issues04:39
NCommanderBut we maybe should draft a policy on backporting across multiple releases04:39
ScottKNCommander: I think that's a good idea.  It needs to be a function of the upgrade path.   For example, if you backport to Dapper there's no need to hit Gutsy because that's not on the upgrade path for Dapper anymore.04:41
NCommanderBut ATM, due to our backports, the Hardy->Intrepid upgrade path may not work as expected04:41
NCommander(well, anything we've backported from jaunty, which isn't much)04:41
ScottKAgreed.04:41
NCommanderGenerally speaking, if you backport to Hardy, intrepid should just work04:42
NCommanderTHe only issue is do-release-upgrade disables backports on upgrade AFAIK04:42
ScottKNCommander: It doesn't.04:42
NCommanderOh04:42
NCommanderSo as long as backports remains enabled, no issues04:43
* ScottK just looked at a box I recenly upgraded and backports is still enabled.04:43
NCommanderWe should talk to jdong and make it an offical policy I guess04:43
OltreIrc`1517.:::] Ci40 @ Tutti [:::. »BuTT3rF|y sCr|pT«»rEvOLuTiOnZ»v3.1.5«06:32
OltreIrc`1517!list06:32
ubottuHi! I'm #ubuntu-devel's favorite infobot, you can search my brain yourself at http://tinyurl.com/5zfb6t - Usage info: http://wiki.ubuntu.com/UbuntuBots06:32
=== dfiloni is now known as devfil
=== RainCT_ is now known as RainCT
=== _nightwish is now known as nightwish
alex-weejis the console font supposed to be something other than the default, old IBM(?) PC one?14:34
alex-weeji just booted intrepid live CD and noticed that the console font was different14:35
alex-weejbut in my hardy->intrepid upgraded machine it's not14:35
=== bureflux is now known as afflux
=== chuck_ is now known as zul
=== johanbr_ is now known as johanbr
bdheemanhello17:43
bdheemancan i build x86_64 packages of an x86 machine, if yes - how?17:43
directhexno.17:44
bdheemans/of/on/17:44
bdheemandirecthex: but, i saw openwrt build system builds kind of some ARM and other packages even on an x86 machine17:46
directhexyou can cross-compile, but you can't use a cross-compiler for package building in the conventional sense17:47
directhexcross-compiling means using a different gcc (or different gcc flags)17:47
bdheemanok, thanks17:47
directhexand you need all the libs and so on...17:47
bdheemanso, better buy an amd64 machine17:48
directhexwell... it's certainly an easier prospect than trying to build packages you can't test, with a cross-compiler17:49
directhexconsider also: a PPA, since PPAs do amd6417:49
bdheemandirecthex: i have not heard about PPA's here in India17:55
bdheemanbut AMD machine are quite popular there days17:56
directhexbdheeman, PPA. personal package archive. a build service on launchpad.net18:05
bdheemandirecthex: thanks, i was not aware of that service18:07
=== thegodfather is now known as fabbione
superm1slangasek, I believe i've got all of the important pieces in order for mythbuntu live disks via livecd-rootfs and cdimage now.  i can reliably build livefs'es that are the right size at least.  could you set up the cron job(s) for building those now for whenever dailies start back up again?19:08
jdongdoes Ubuntu squashfs support the LZMA compression method?19:37
jdongi.e. in-kernel?19:37
pwnguini doubt it19:37
jdongI figured19:38
jdongthe userspace tools support it which made me curious19:38
jdongI'm re-doing my backup strategy... again... and I'm considering squashfs archiving19:38
pwnguinif squashfs makes it into the kernel19:38
pwnguinit might make sense to visit it19:38
pwnguinlzma19:38
pwnguinthe liveCD is perennially out of space19:39
pwnguin(or maybe perpetually)19:39
jdongmakes it into... the linus tree you mean?19:39
jdongwell bi-anually out of space ;-)19:39
pwnguinyea, when i looked into it last, squashfs decided to try to get into the linus tree again19:40
jdongright, I understand that's the reason why there's no LZMA support19:40
hyperairvlowther: ping19:40
jdongi.e. upstream doesn't want any more compression methods in the kernel19:40
pwnguinjdong: you should probably double check with the ubuntu kernel ML19:41
pwnguini donno if there's plans or if i missed something, etc19:42
RainCT(btw, does someone know how to fix/workaround some applications which are displaying solid black instead of input from the webcam? I believe this to be because the webcam uses JPEG compression and the applications don't support it.)19:54
mohbanahi, anyone familiar with fontconfig? what's wrong with this; http://pastebin.com/m53266bde19:56
jdonggrumble stupid pulse dying after suspend20:12
=== ajmitch_ is now known as ajmitch
hyperairjdong: there's a bug regarding that20:17
hyperairjdong: and a fix ready20:18
jdonghyperair: neat :)20:18
hyperairjdong: bug 20208920:18
ubottuLaunchpad bug 202089 in pulseaudio "Pulseaudio is blocking normal sound after resume" [Low,In progress] https://launchpad.net/bugs/20208920:18
hyperairit seems to be assigned to someone else, and in progress20:18
hyperairi realized that after i came up with a workaround and created a debdiff20:18
hyperairso now i'm actually wondering what i should do20:19
directhexcry!20:20
jdonghyperair: well it's been assigned to crimsun, so I'd say it's in the best hands :)20:21
hyperair=\20:22
hyperairdirecthex: hmph20:22
pwnguinwho assigned it to crimsun?20:23
* hyperair has no idea20:23
pwnguinif it wasn't crimsun himself, might as well at least attach the debdiff20:23
hyperairpwnguin: it's attached20:24
jdongpwnguin: I see a SRU-compatible debdiff attached20:24
ScottKAccording to the activity log he assigned it to himself.20:24
pwnguinhyperair: then i guess you just wait ;)20:25
hyperairi guess20:25
hyperairspeaking of sru and debdiffs... could someone help me out with bug 267922 /20:25
hyperair?20:25
ubottuError: Could not parse data returned by Launchpad: The read operation timed out (https://launchpad.net/bugs/267922/+text)20:25
pwnguineep20:26
hyperairlol20:26
hyperairi don't have any trouble accessing that20:26
hyperairbug 26792220:26
ubottuLaunchpad bug 267922 in banshee "Banshee hangs up/crashes when pluggin in MTP-USB-Player" [Undecided,Confirmed] https://launchpad.net/bugs/26792220:26
hyperairthere we go20:26
jdonghyperair: ok that one is within my umbrella of authority, lemme take a look at your debdiff20:29
hyperairjdong: thanks =)20:30
jdonghyperair: where did the patch come from? did you write it?20:31
jdongit's pretty big, just trying to understand it20:31
hyperairjdong: extracted from trunk20:31
jdonghyperair: gotcha20:31
jdonghyperair: ok I acked the SRU for you20:32
hyperairthanks20:33
jdongsubscribe ubuntu-universe-sponsors20:33
jdonglet me know if you don't get sponsorship in a reasonable timeframe20:33
hyperairalright20:33
hyperairwhat's a reasonable timeframe?20:33
jdongwithin a few days, I'd say.20:33
hyperairokay20:33
hyperairjdong: should i unsubscribe motu-sru?20:35
jdonghyperair: no need ,keep them subscribed20:36
hyperairokay20:36
Heller_Bardehai21:04
Heller_BardeI am an Archlinux user and would like to know how i can find out how ubuntu packages are built and packaged. Can someone help me with that?21:05
directhexHeller_Barde, specifically...?21:05
Heller_Bardedirecthex: actually the background is that i would like to build some packages (around xorg) with ubuntu patches because i have serious performance issues with opengl applications21:06
hyperairHeller_Barde: what do you want to know?21:07
sebner!packageing | Heller_Barde21:08
ubottuSorry, I don't know anything about packageing21:08
sebner!packaging | Heller_Barde21:08
ubottuHeller_Barde: The packaging guide is at http://wiki.ubuntu.com/PackagingGuide - See https://wiki.ubuntu.com/UbuntuDevelopment/NewPackages for information on getting a package integrated into Ubuntu - Other developer resources are at https://wiki.ubuntu.com/UbuntuDevelopment - See also !backports21:08
Heller_Bardeand i found my way to the patches and everything and i found out that its not only the patches that made the difference but obviously too how it is built.  i got as far as debian/rules but that's source-package specific, i now would like to know how the actual packages are built21:08
hyperairdebian/rules21:08
hyperairthat's the key to everything21:08
Heller_Bardeubottu: oh cool, ill look at that and get back to you with questions21:08
ubottuError: I am only a bot, please don't think I'm intelligent :)21:08
Heller_Bardeah21:08
Heller_Barde^^21:08
hyperairheh21:08
Heller_Bardesebner i meant21:08
Heller_Barde^^21:08
hyperairHeller_Barde: debian/rules is a Makefile. everything goes in there21:09
sebnerHeller_Barde: :) np21:09
Heller_Bardehyperair: yes, i saw that. but from one source packages several binary packages are made21:09
directhexokay, the whirlwind tour. source packages (with a few exceptions) come in three files - foo.orig.tar.gz contains the upstream software, converted to gzip where required. foo.diff.gz contains a gzip diff of all the changes between the upstream tarball, and what is needed for packaging - typically it should only contain the "debian" folder, which has the files in it about how to build. foo.dsc is a manifest, pointing to the other tw21:09
directhexo files21:09
directhexthe key files in the debian/ folder are "rules" and "control"21:10
hyperairHeller_Barde: patches in debian/patches, configure flags in debian/rules somewhere, post installation stuff also in debian/rules21:10
Treenaksdirecthex: hurricane warning ;)21:10
directhex"control" is the list of information about the package - dependencies, general package metadata, and contains the list of binary packages (and their deps) as well as info on the source package. the metadata is used by apt/dpkg21:10
directhexTreenaks, i prefer to think of it as a monsoon21:11
Heller_Bardemhhmhh but lets say i have the source packages, i patched everything and then built it (which is pretty tough without a debian system ^^) and then what? where comes the splitting stuff up in the several packages?21:11
directhex"rules" is a makefile, which does the configure.make stuff. in most cases, it'll use a "helper" app of some kind, such as debhelper, which will do assorted cleanup - generating automatic deps, putting files into actual packages, and so on21:11
hyperairHeller_Barde: then there's postinst, postrm, preinst, prerm which are hooks that are run when installing/upgrading/removing the package on the user's system. kinda like the one archlinux has.. what was it called? something.install?21:11
Heller_Bardehyperair: yep21:12
hyperairdamn my archlinux packaging is rusty.21:12
directhexHeller_Barde, the splitting is done by a debhelper command in debian/rules21:12
Heller_Bardehyperair: how so? there is soo little to know ^^21:12
Heller_Bardedirecthex: ahh, i seem to have missed that21:12
Heller_Bardei'll scour through mesa-src again21:13
hyperairHeller_Barde: i know. i forgot it heh. the names mostly. i still remember how to make a PKGBUILD build() function21:13
directhexHeller_Barde, specifically, look for lines about dh_install and dh_installdeb21:13
hyperairwait, you're poking around with mesa? i've got experience with that one.21:13
Heller_Bardeyes21:13
hyperairHeller_Barde: exactly what are you looking for?21:13
directhexHeller_Barde, those two will use a list of files matching the binary package name - e.g. libmoon0.install contains the list of files which dh_installdeb will put into the libmoon0 package21:14
Heller_Bardei use compiz as performance: 3-8 fps in archlinux (with testing) and 30-40 fps in ubuntu 8.10 on a intel x3100 (965gm chipset)21:14
Heller_Bardeand i tracked it down to some packages21:14
hyperairregarding the package splitting, you needn't really look into those... archlinux doesn't split packages21:15
hyperairso yo wouldn't be able to use it anyway21:15
Heller_Bardehyperair: actually it does21:15
hyperaireh? since when?!21:15
Heller_Barde*smiles sheepishly*21:15
Heller_Bardejust some21:15
hyperairO_o21:15
Heller_Bardeexactly the ones i need21:15
hyperairhow?21:16
Heller_Bardewell its just that intel-dri is built from mesa-src21:16
hyperair@_@21:16
Heller_Bardeno idea why its not integrated21:16
Heller_Bardeno fucking clue (sry for language)21:16
hyperairlol21:16
hyperairanyway21:16
hyperairhere's how the build process works.. mostly21:16
hyperairthe debian/tmp folder is ismilar to archlinux's pkg/ folder21:17
Heller_Bardeone moment i'll get my project folder, i have one or two questions21:17
hyperairthat's where everything goes just before splitting happens21:17
hyperairthen it'll go into debian/package_name21:17
Heller_Bardehyperair: okay21:17
hyperairwell sometimes you can skip debian/tmp21:17
hyperairand go straight into debian/package_name21:17
hyperairbut generally for stuff with many packages, we dump everything in debian/tmp first21:18
hyperairand stuff is split based on package.install files21:18
hyperairwhich contain lists of files21:18
hyperairwhat goes into where kinda thing21:18
Heller_Bardeholy moly, i get it now21:18
hyperairokay =)21:19
Heller_Bardei just looked at it again and went am i frickin stupid21:19
hyperairheh21:19
hyperairwell21:19
Heller_Bardesry, i had an epiphany21:19
Heller_Bardei guess i was too tired last time21:19
hyperairheh21:19
hyperairno worries i get stuff like that once in a while21:19
Heller_Bardeit was pretty frustrating... well, that's what you get going from warm, fuzzy, easy packaging in arch to the hard world of debian :P21:20
Heller_Bardebrb21:20
hyperairwell it depends on whether they're using raw debhelper or cdbs21:20
hyperaircdbs packaging is awesome21:20
hyperairi use it all the time21:20
RAOFhyperair: CDBS is magic harvested from the rich plumes of arcanum which spew from deep-sea volcanoes.  When you're doing exactly what it's designed for, it's great.  When you try to bend it a bit, you play "chase the make rule" through endless includes, all alike.21:36
hyperairRAOF: nice analogy21:37
sebnerhyperair: another try would be: "Black magic"21:38
hyperairxD21:39
hyperairi like cdbs. it does most of what i need, compared to debhelper and its 186940747134 different dh_* commands21:40
sebnerhyperair: already tried dh7? :)21:40
hyperairnop21:40
hyperairi haven't even bothered learning the names of all the dh_* commands21:41
sebnerhyperair: take a look at it and tell me again the difference between dh and cdbs ;)21:41
hyperair=.=21:41
RAOFIt's all the joy of cdbs, but with the ability to clearly see what the hell's happening.21:41
hyperairany docs i can read?21:41
RAOFman dh21:41
lifelessdh(3SSL)                                                                                            OpenSSL                                                                                           dh(3SSL)21:42
lifelessNAME21:43
lifeless:P21:43
lifeless       dh - Diffie-Hellman key agreement21:43
lifeless?21:43
hyperairlifeless: wrong dh21:46
RAOFlifeless: You'd be after "man 1 dh", then.  :P21:56
lifelessRAOF: I think I was hinting that you should be more precise :>21:58
RAOFlifeless: Pfft.  Precision is for mathematicians!22:00
lifelessRAOF: zigactly22:01
lifelessRAOF: unless you're a statistician22:01
RAOFThen you have precise things to say about how imprecise you are.22:01
lifelessprecisely22:02
=== geser_ is now known as geser
Heller_Bardeokay, i do have some questions now. wtf do all these dh_* commands do?22:14
peciskHeller_Barde: dh stands for debian helper scripts22:14
Heller_Bardeah -.-22:15
directhexeach script has a specific job22:15
Heller_Bardewhat is the $* variable in a makefile?22:16
radixHeller_Barde: each of the dh_ programs has a man page22:33
radixand there's a 'debhelper' man page22:33

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!