/srv/irclogs.ubuntu.com/2011/01/26/#ubuntu-motu.txt

=== Amaranth_ is now known as Amaranth
=== Amaranth_ is now known as Amaranth
=== Amaranth_ is now known as Amaranth
=== Amaranth_ is now known as Amaranth
pabelangerWhat an exercise, finally have reprepro and rebuildd working together.  Yay, for local package repository04:13
MTecknologyok... so... override_dh_fixperms:  if I place that in debian/rules it'll automatically be used, nothing special, just all magic?04:39
micahgMTecknology: I thought with overrides you have to do something04:48
MTecknologymicahg: oh.. I'm only adding it to change the permissions of one directory, would it maybe make more sense to just put it right after dh_fixperms?04:50
micahgMTecknology: hmm, I thought that you can the dh_foo, if appropriate, in the override along with whatever using the override for04:51
micahg*call04:51
MTecknologyoh04:52
MTecknologymicahg: If a user installed the package, then upgraded the package but never changed that directory, could I make the permissions get updated on it?05:00
micahgMTecknology: which directory?05:00
MTecknologymicahg: /var/log/nginx05:00
MTecknology750 instead of 75505:01
micahgMTecknology: Yeah, that should be fine as long as you have a good reason :), you might want to add a news file so people aren't shocked05:02
MTecknologymicahg: I'll add it to the news file; i'd add the change to postinst but then every single update the permission will be reapplied; i'd rather just do it once for old updates and if the user changes it then leave it alone05:04
micahgMTecknology: yes, that's best, cleaning up after oneself is fine, but sysadmins should have the final say05:05
micahgMTecknology: check the version in the postinst05:05
MTecknologyoh... I was under the assumption that that postinst happened after the new package was installed05:06
micahgMTecknology: hmm, you might have access to the old version, I'm not so clear, maybe someone else knows05:07
micahg*version number05:07
micahgMTecknology: yeah, you're right, http://www.debian.org/doc/debian-policy/ch-maintainerscripts.html05:10
micahgMTecknology: you could check it in the preinst05:11
MTecknologymicahg: I had to run for a minute.. "new-preinst upgrade old-version05:35
MTecknologymicahg: I had to run for a minute.. "new-preinst upgrade old-version" does that mean there should be an old-version variable in there?05:36
micahgMTecknology: I believe so :)05:36
MTecknology$2 i assume?05:36
micahgI think so05:36
MTecknology:D05:36
MTecknologyheh.. so I'll just need to know how to compare the versions05:37
micahgMTecknology: dpkg --compare-versions :)05:37
MTecknologymicahg: and to grab the new version?05:41
micahgMTecknology: you shouldn't need to worry about the new version, right?05:42
MTecknologywhat else would i be comparing against?05:42
micahgMTecknology: the last bad version05:42
micahgor the first good version depending on how you compare05:42
MTecknologyoh!05:43
MTecknologyI created a preisnt and didn't even need it :P05:43
MTecknologyexcept postinst doesn't seem to have that version for an upgrade05:44
micahgMTecknology: whcih is why you should do it in preinst :)05:44
MTecknologyif dpkg --compare-versions "$most_recently_configured_version" lt "0.8.54-4"; then chmod 750 /var/log/nginx; fi05:48
MTecknologymicahg: thanks a bunch :D05:52
micahgMTecknology: you're welcome05:53
MTecknologymicahg: now to test it out and see if things still build; I made a lot of changes (should close every open debian and launchpad bug against the package.) :D05:53
micahgMTecknology: awesome05:54
MTecknologyeven added a feature through a patch that it doesn't look like igor will accept05:54
micahgMTecknology: is Igor upstream?, we'd want Debian to take that then05:55
MTecknology:S ... debuild failed05:56
MTecknologymicahg: because of all the work i did with the package, the guys that manage it in debian asked me to help maintain; so what i do winds up right there05:56
micahgMTecknology: that's great, work in one place and lots benefit05:57
MTecknologyby igor, I meant the guy that controls what winds up in nginx05:57
micahgMTecknology: right, so if upstream rejects and Debian takes it that's fine for Ubuntu05:57
MTecknologyapparently i messed up debian/rules bad enough that debuild won't even complete05:57
MTecknologyoh.. it was only complaining about a little itty bitty thing....05:58
MTecknologymicahg: ya know.... I put a LOT of work into this thing; now I'm looking back and wondering where I'd be if this channel didn't exist.. I'd probably have given up entirely and got nowhere06:00
micahgMTecknology: I'm sure that'll make a lot of people here smile :)06:01
MTecknologymicahg: I hope so; I wouldn't even want to /try/ to get a list of everyone that's helped me...06:03
jonathanhi everyone06:27
jonathanI'm interested in joining the prospective developers but I'm confused as to where to go to start06:30
jonathanuh oh, guess my text got cut off06:30
jonathanwould anyone happen to know where one would start off in joining the perspective developers?06:30
MTecknologyjonathan: which developers?06:32
jonathanI was looking at the ubuntu website and it said that prospective developers was a good place to start in joining the development team06:32
MTecknologyoh.. there's a lot of places developers can look to contribute; that's part of why it's such an easy place to jump into06:33
jonathanI'm just confused on what I need to do in order to get started06:33
MTecknologyyou could look into some bugs for a package you particularly like and try to fix them06:34
MTecknologyfix them, build a patch, and then add to the bug report06:34
jonathanthat's it?06:34
MTecknologythat's one place that's somewhat easy to jump into06:35
MTecknologyotherwise you could look through and try to triage all bugs in ubuntu :)06:35
jonathanhehe06:35
jonathanI'm sort of new to coding06:35
jonathanI'll have to handle easy things06:36
MTecknologythere's no one language used; a lot packages produced for ubuntu are python driven though06:36
jmarsdenjonathan: You may find https://wiki.ubuntu.com/BeginnersTeam/FocusGroups/Development/Devbeginnings#Ubuntu%20Development%20Beginnings has a few things you can try out as you get started.06:37
evilvishjonathan: bug triage is relatively easy and gives you a good idea about the packages.. and where to get involved..06:37
jonathanand I can just come here if I happen to have questions?06:37
MTecknologytriage is where i started out at; i thought that if i just sat there and hammered away, i could be 'that guy' that got through everything :P06:37
MTecknologyjonathan: nope, questions are hated in here.... actually, some of the most amazing devs i know hang out in here06:38
evilvishjonathan: depending on what you are doing, we have channels different channels..06:38
MTecknologyalways eager to help new guys06:38
jonathanwhere do I go if I have questions or happen to need to talk to someone?06:38
MTecknologythere's #ubuntu-packaging if you want to get into packaging06:39
MTecknologyjonathan: you can ask where the best place is to ask too :)06:39
jmarsdenjonathan: For questions that are definitely about triage, you may find #ubuntu-bugs is a good place to ask them.06:39
jonathanby ubuntu packaging do you mean maintaining packages?06:41
MTecknologyya06:41
jonathansomeone like me would be capable of doing that?06:41
MTecknologythat's what I'm working on right now actually06:41
MTecknologyand I'm very far from an expert06:42
jonathanhow much coding is involved in maintaining packages?06:42
MTecknologydepends.. ideally, not much06:43
jonathanI guess that would be a good place to start then06:43
MTecknologyif you start though.. my best suggestion is 1) setup a dev system using debootstrap 2) create a user inside of that 3) do your dev in there06:43
MTecknologyif/when you start breaking things or installing php5-dev (i'm doing so now) you're not messing with your base system and it makes new environments incredibly easy to play with06:44
MTecknologyjonathan: https://wiki.ubuntu.com/PackagingGuide/Complete06:44
jonathanI've never used debootstrap before06:46
jonathando you have a guide for that?06:46
MTecknologyaptitude install debootstrap; debootstrap natty natty; chroot natty06:47
MTecknologydebootstrap is so easy it's actually scary :P06:47
jonathanok, I think the first thing I'll work on is mastering all those packaging commands06:47
MTecknologynah, then you'll never get anything done :P06:48
jonathanwhich ones should I focus on?06:48
MTecknologyjonathan: up until a month ago I was always editing the timestamp line in debian/changelog by hand and screwed up a lot; then i found dch and life got easier06:49
jonathanif I may ask, how long have you been a linux user?06:50
MTecknologyjonathan: there's a billion (or so) tools to make your life easier; best to just take a walk through that complete guide, and then start playing and figure out what you don't understand06:50
jonathanI've used linux for a few months so far06:50
MTecknologyi started playing with linux when 5.04 came out (my senior year in HS)06:50
jonathanthat was last year for me06:51
jonathanok, I'll look through the guide then06:51
jonathanthanks for the help06:51
MTecknologyjonathan: always feel free to ask questions; sounds like #ubuntu-packaging may be a good channel for you06:52
MTecknologyonly 7min wait time estimated for amd64 ppa builders.... I'm about to skew that estimate...06:55
macoupoading OOo?06:56
MTecknologymaco: not that bad, 13596k06:56
MTecknologymaco: providing a backport of php5 for lucid and maverick because it was such a huge request with nginx06:57
MTecknologymaco: at least i didn't trigger a complete rebuild of python :D07:00
MTecknologydangit.....07:37
MTecknologyI have package nginx-full that depends on package nginx being already installed; package nginx depends on nginx-full | nginx-light.. so nginx needs to be installed before nginx-full is finished up; but installing nginx means you need to install one of the others07:40
MTecknologyany tips on making this work?07:41
jmarsdenMTecknology: install nginx-light first?  That will satisfy the dependency.  Then install nginx, then install nginx-full.  Sounds like it will do what you want??07:43
MTecknologyjmarsden: nginx-full and nginx-light are conflicting packages; they both require nginx to be installed first though; it's like apache-common07:45
jmarsdenThen you have a circular dependency, if nginx depends on -light or -full and they each depend on nginx.07:46
MTecknologyyup07:46
MTecknologyI'm not sure what the right way is to deal with it07:46
MTecknologyjmarsden: I'm sure it's my control file - lemme grab it07:47
jmarsdenDon't do that :)  Avoid recursive dependencies :)  apache-common doesn't depend on apache-mpm etc...07:47
MTecknologyjmarsden: oh......... that's making more sense now :)07:48
MTecknologyjmarsden: so instead of Depends: nginx-full; I'm looking at Reccomends: nginx-full ?07:49
MTecknologyor would I want Suggests?07:50
MTecknologyRecommends: nginx-full | nginx-light ?07:51
jmarsdenSuggests is less "forceful" to package managers, recommends will mean if you apt-get install nginx you *will* pull in nginx-full...  Eww, not sure you want the | stuf in Recommends.07:51
MTecknologyok- that makes sense :)07:52
jmarsdenAre the names set in stone?  Maybe you should rename what i snow nginx to nginx-common and then rename nginx-full to nginx ???07:55
jmarsdenThat way apt-get install nginx will do what most people expect.07:56
MTecknologyI considered different names like that; but the only thing held in nginx-full or nginx-light is a binary and one or two other tiny things; nginx has most everything; but to deal with the way things look now it made the most sense to have nginx instead of nginx-common07:57
jmarsdenOK.  If there is an existing "tradition" for that naming scheme, then it makes sense to keep it.07:58
MTecknologyjmarsden: uploaded again... we'll see if things get fixed :)08:03
jmarsdenOK.  You can sometimes avoid the upload/wait-for-builders/test/oops/edit... cycle delay by setting up your own local repository, incidentally.  But I need to go to bed, you'll need to google for info on how to do that yourself :)08:06
MTecknologyjmarsden: alrighty, i should do that too... I wanted to be sleeping 4hr ago :P    thanks :)08:07
jmarsdenGoodnight.08:07
dholbachgood morning08:18
MTecknology!info php5-fpm08:52
ubottuphp5-fpm (source: php5): server-side, HTML-embedded scripting language (FPM-CGI binary). In component universe, is optional. Version 5.3.3-1ubuntu9.3 (maverick), package size 2875 kB, installed size 7628 kB08:52
MTecknology!info php5-fpm natty08:52
ubottuPackage php5-fpm does not exist in natty08:52
MTecknologyzul: What happened? :(08:52
MTecknologyzul: I just bumped up to 11.04 and it seems to have gone buh-bye08:54
=== JanC_ is now known as JanC
=== bilalakhtar_ is now known as cdbs
paultagMTecknology: http://lists.debian.org/debian-devel/2009/09/msg00348.html  <-- related?13:22
paultagphp5 (5.3.3-2) unstable; urgency=low13:24
paultag  * Don't build FPM SAPI now13:24
paultag -- Chuck Short <zulcss@ubuntu.com>   Fri, 07 Jan 2011 22:44:56 +000013:24
paultagMTecknology: might want to get in touch with that feller if it's a big deal13:24
ari-tczewbdrung: bug 702765 is ready.13:57
ubottuLaunchpad bug 702765 in libgcrypt11 (Ubuntu) "Please sync libgcrypt11 1.4.6-4 (main) from Debian experimental (main)" [Wishlist,New] https://launchpad.net/bugs/70276513:57
udienzari-tczew, about bash merging. You can take it i will merging firestarter13:59
ari-tczewudienz: I doubt that there is anyone willing to sponsor.14:00
ari-tczewcjwatson wouldn't touch it. doko is maintainer, but he delayed merge to next upstream release.14:01
ari-tczewand I'm not suprised. bash is very impotant package.14:01
udienzhm.. ok ok14:02
RhondaI still don't really understand why bash has the need for such a lengthy diff in the first place.14:03
cjwatsonthe merge in question isn't even all that important; most of the important changes are already in Ubuntu14:04
cjwatsonso I don't understand the lengthy debate about it14:04
* Rhonda . o O ( especially since the maintainer on both sides is the same person, actually … )14:05
BlackZI'm agreed with cjwatson; I said that in the bug report, though :)14:05
cjwatsonRhonda: I can understand plenty of reasons for that from my own experience, THB14:06
cjwatsonTBH14:06
cjwatsonnot all my packages are synced14:06
cjwatsonimportant> I mean, I don't mean to diss udienz, but it does seem reasonable to spend effort on merges with important changes first14:07
RhondaLike pgadmin3, thanks for ACKing it, cjwatson ;)14:09
RhondaThough no merge but sync, actually.14:09
cjwatsonI think I just processed it14:10
cjwatsonthat's mostly scripted, no need to thank me :)14:10
RhondaWhatever, thanks anyway. :)14:11
ari-tczewI hope that in next Debian cycle most packages will support Ubuntu changes in modern, parallel system14:11
cjwatsonI don't find it that simple14:12
ari-tczewso maybe we will gain more syncs14:12
cjwatsonin particular I'm not convinced that that interacts brilliantly with revision control14:12
Rhondaari-tczew: Actually dpkg is ready for that and buxy did blog on how to do it.14:12
Rhonda… in source v314:12
Laneythe implementation there isn't as good as it could have been, IMHO14:13
Laneyyou need to maintain entire separate series files for each distro14:13
cjwatsonI know dpkg can do it, but think of what the unpacked trees look like and how you deal with that in revision control; you end up having diffs in the unpacked tree even though it's the same version, which is very confusing when you then come to modify it14:13
cjwatsonso I haven't implemented that for any of my packages14:13
RhondaThere is a lot of things that aren't as good as they could have been and annoy people indeed, in source v3 in general, but it's definitely a step in the proper direction14:13
ari-tczew+1 ^^14:14
LaneyI (and the cli teams) find that keeping VCS branches for Ubuntu is a good way to manage the delta14:14
Laneygit merging is excellent for this workflow14:14
cjwatsonright, I prefer that to the approach of multiple series files14:14
cjwatsonit's much easier to see what's happening14:15
LaneyI would probably use unapply-patches in the multiple-series workflow14:15
RhondaLaney: I also consider VCS branches proper.14:16
RhondaI just haven't anyone approach me to do an ubuntu branch for irssi yet.  %-/14:16
RhondaHmm, actually I could do that myself me thinks.14:17
LaneyIt's in main, so you will have to get upload rights for it14:17
RhondaAnd step on the toes of those who have worked without contacting me so far. :P14:17
Laney(if you want to upload yourself)14:17
Rhondaoh14:17
Laneythat should be a formality though14:17
RhondaI guess I will have to apply because of logcheck anyway.14:17
Laneyyou can get per-package rights, and the Debian maintainer should be almost guaranteed to receive them14:18
RhondaSo what else from me is in main that I'm not aware of? Any cheap way to find out?  :D14:18
LaneyUDD? :P14:18
Rhondanah, grep-dctrl!14:18
Laneyor g.. yes14:18
Rhondahuhm14:19
cjwatsonthe problem with unapply-patches (and traditional patch systems) is that you can't use $vcs blame coherently between upstream and patched source files14:19
RhondaLaney: I don't have the packages files at my hand, so I might need to ressort to udd.  %-/14:19
cjwatsonwell, one of the problems14:19
Rhonda… or start adding deb-src for ubuntu onto my system. :)14:19
LaneyI want to figure out if topgit makes sense14:19
cjwatsonRhonda: or chdist14:20
Rhondacjwatson: What's chdist?14:20
cjwatsonit's in devscripts14:20
Laneyand I also want a way to indicate to Ubuntu developers that the Debian maintainer is maintaining a branch for the package14:21
Laneyprobably Vcs-foo14:21
cjwatsonhttp://manpages.ubuntu.com/manpages/maverick/en/man1/chdist.1.html14:21
cjwatsonI use it on an Ubuntu system so that I can do 'chdist apt-get unstable -d source <package>'14:21
* Rhonda goes to http://deb.at/Mchdist instead :P14:22
Rhondahuhm14:24
Rhonda$ grep-dctrl -FMaintainer,Original-Maintainer,Uploaders 'Gerfried Fuchs' -sPackage,Section /var/lib/apt/lists/at.archive.ubuntu.com_ubuntu_dists_natty_* | grep-dctrl -v -FSection universe14:25
RhondaSo it's beep, irssi and logcheck  \o/14:26
RhondaLaney: Thanks for pushing me to dig it up14:27
Laney\o14:28
Laneynow get applying!14:28
Rhondahah14:31
RhondaIf time permits I would. The irc meetings aren't quite working for me these days.  %-/14:31
geser19:00 UTC doesn't suite you either?14:31
RhondaThat's around the time my young one goes to bed, and we have time to talk with each other, my SO and me14:36
=== zul_ is now known as zul
bdrungari-tczew: done15:26
ari-tczewbdrung: thanks15:27
RhondaOh. Still had an intrepid chroot and wondered why it gave me errors.15:36
bdrungall sync and merge requests processed15:38
al-maisanhello, I am having trouble installing a package (python-scipy) on natty; the errors are as follows: http://pastebin.ubuntu.com/558620/17:03
al-maisanwhat is puzzling is that the quoted dependency "Depends: python-numpy (< 1:1.5)" is not in the python-scipy package's debian/control file17:03
al-maisanany ideas why this breaks?17:04
geseral-maisan: http://packages.ubuntu.com/natty/python-scipy shows this versioned dependency too17:24
* al-maisan looks17:24
al-maisanoh, I see.17:25
al-maisanthanks geser !17:25
=== Kmos is now known as Guest50434
ScottKHopefully whoever uploaded the new numpy is also taking care of the transition.18:04
macoScottK: does your brain pronounce that package's name with the proper -py ending or do you keep rhyming with an adjective for mashed potatoes?18:05
ScottKDepends.  That one rhymes with pie.  Scipy however sounds like the peanut butter.18:06
gesernot http://en.wikipedia.org/wiki/Skippy_the_Bush_Kangaroo ? :)18:09
al-maisanre. the new numpy .. /me hopes the same .. I went back to rev. 1:1.4.1-5ubuntu4 for the time being18:11
tumbleweedScottK: he's supposed to (he's a mentoree of mine). I'll prod him again18:17
ScottKtumbleweed: Thanks.18:18
ScottKgeser: No.  http://en.wikipedia.org/wiki/Skippy_(peanut_butter) - Also not http://skippyslist.com/list/18:20
=== al-maisan is now known as almaisan-away
ari-tczewScottK: around?18:47
ScottKari-tczew: Somewhat.18:48
ari-tczewScottK: could you try build gnome-power-manager on natty (armel) without that change: http://bazaar.launchpad.net/~ubuntu-branches/ubuntu/natty/gnome-power-manager/natty/revision/123#debian/rules18:49
ScottKari-tczew: Link to a .dsc I can dget and build without change and I'll try it.18:49
ScottKLink/Link me18:49
ari-tczewScottK: give me 5 minutes18:50
m4n1sh i had the trunk build of gaj in my PPA19:08
m4n1shnow there is an official release19:08
m4n1shso for adding the entry in changelog19:08
m4n1shshould I remove the changelog entry of the trunk build19:08
m4n1shor keep it there?19:08
ari-tczewm4n1sh: remove19:10
ari-tczewkeep information only about ubuntu package19:10
m4n1shari-tczew: yeah. Since I would ask a DD to upload it19:11
m4n1shso I think it needs to go19:11
m4n1shari-tczew: thanks for help19:11
m4n1shjust wanted to learn the best practices19:11
ari-tczewm4n1sh: You're welcome.19:12
ari-tczewScottK: armel = ARM ?19:25
ScottKari-tczew: For our purposes yes.19:25
ScottKarm is the processor and also the name of a now defunct architecture in Debian that was replaced by armel (IIRC little endian versus big endian is the difference)19:26
ari-tczewScottK: http://people.ubuntu.com/~ari-tczew/ScottK/ let me know when you will start build :)19:38
ScottKari-tczew: I have to update the natty chroot on that box first, so it'll be a few minutes.19:41
ari-tczewScottK: How it's going?20:41
ScottKari-tczew: Started the build and then had to leave.  Just got back.  I'll have a look and see.22:10
ScottKari-tczew: Successful build.  Do you need the build log?22:11
ari-tczewScottK: if you could upload somewhere?22:11
ScottKSure22:12
ScottKari-tczew: http://pastebin.com/QRrF9W0E22:14
ari-tczewScottK: thanks :)22:15
ScottKari-tczew: You're welcome.22:16

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