/srv/irclogs.ubuntu.com/2010/07/01/#ubuntu-devel.txt

=== micahg1 is now known as micahg
=== bjf[afk] is now known as bjf
=== bjf is now known as bjf[afk]
=== jjohansen is now known as jjohanse-afk
AndpheHi, anyone can help?, I'm upgrading PHP 5.2 to 5.2.13 for myself and as an excercise03:36
Andphelaunchpad build the amd64 but not the i386 package03:37
Andphehttp://launchpadlibrarian.net/51213960/buildlog_ubuntu-lucid-i386.php5_5.2.13.dfsg.1-0ubuntu0~andpheppa1_FAILEDTOBUILD.txt.gz03:37
Andpheconfigure: error: recode extension can not be configured together with: mysql03:38
Andphethis is the error message ↑↑03:38
Andphefor what I've research, it's because the config file gets rebuild and the patches are loose03:38
micahgAndphe: please try #ubuntu-packaging03:39
Andphemicahg:oh, ok, sorry03:39
micahgAndphe: np03:39
=== dendrobates is now known as dendro-afk
un214you have managed to annoy people [no not just me] pretty thoroughly with plymouth04:19
un214I'm trying to take some drastic action to get rid of it -- in my mind there can be no justification for it04:19
un214but you managed to wire mountall to plymouth pretty thoroughly so mountall must go as well04:20
un214and brings a good chunk of /etc/init with it04:20
RAOFSomething like plymouth is fundamentally necessary for parallel init to work properly.  You don't have to have the splash enabled.04:21
un214I'm about to prove that wrong04:21
RAOFFor the special-case of not having more than one thing ask for input during boot, I presume.04:22
un214what's the second thing that asks during early boot?04:22
RAOFfsck, 1 passphrase input per encrypted device, etc.04:23
un214easy fix for that one -- attach all encrypted devices before fsck of nonroot04:24
RAOFfsck of multiple drives running in parallel?04:24
un214my old slackware system handled that pretty darn good04:25
un214[in the rare case where fsck needed to ask something it reran it in serial mode]04:25
RAOFI guess it comes down to: are you going to enumerate all the things which could possibly ask for input before login, and work around each of them separately?04:26
un214I'm actually going to say there should not be any more for any reason04:26
un214by the time anything has to ask a boot question except for efs you have a sick system04:27
un214and if you want to ask at boot time then it cannot be fixed remotely anymore04:28
RAOF(Unless you extend something like plymouth ;))04:28
un214it is my judgement the better move is to try to bring up console login, networking, and sshd and wait unless you can't even go multiuser04:29
un214[can't mount /usr = fatal]04:30
RAOFYeah, but wouldn't it be cool if it was fatal in a remotely-resolvable fashion? :)04:31
un214well if you moved sshd off /usr it would be04:31
un214I retooled a system I was using for development once so it could start sshd even if it couldn't remount / rw.04:32
RAOFAnyway, you're welcome to work at removing Plymouth from the startup sequence.  I don't think it's a huge issue, though.04:32
un214I like your attitude.04:32
RAOFAnd it's solving real problems, so you'll need to re-implement solutions for them.04:32
un214it just managed to wrankle a lot of people because for most people it doesn't solve anything interesting04:33
RAOFI'm not sure about “a lot”, but for many people it won't be solving anything interesting, no.  From a distro point of view, though, we need to care about the more general cases.04:34
un214I actually worked this problem years ago, and cut about a third off of boot time by bolting a parallel init system on top of inittab04:35
un214everything after mount -a can be done in parallel no problem, everything before is not really parallelizable anyway04:35
RAOFExcept that's not true - it's entirely possible to parallelise things before mount -a (although some of this is in the kernel), and there's dependencies after mount -a.04:37
un214I'm really curious how plymouth would handle /dev/console -> /dev/ttyS104:40
un214You see, I'm really concerned as neither plymouth nor mountall have demonstrated the rock solidness I've come to expect from Linux systems.04:44
=== nixternal is now known as Guest69583
=== bjf[afk] is now known as bjf
=== bjf is now known as bjf[afk]
dholbachgood morning06:52
pittiGood morning07:04
dholbachhey pitti07:04
pitticcheney: oh, please don't worry, your wife is much more important :)07:04
pittihey dholbach07:04
=== Amto`OFF is now known as Amto_res
tkamppeterCan someone tell me which is the default scanning tool in Kubuntu?08:37
raphinktkamppeter: I think that would be skanlite08:38
pittitkamppeter: simple-scan08:43
tkamppeterpitti, is simple-scan not a GTK application?08:50
pittitkamppeter: argh, sorry, misread08:50
pittitkamppeter: trust raphink, not me :)08:50
pittitkamppeter: (unlucky line break just before "in Kubuntu" in my IRC client :) )08:50
tkamppeterpitti, raphink, thanks. I have put xsane | simple-scan | skanlite instead of xsane into the Recommends of hplip-gui now, should avoid unneeded GTK installation for Kubuntu users.08:52
pittitkamppeter: shouldn't that be a suggests rather?08:53
pittitkamppeter: oh, and as for bug 600504, should python-notify be a dependency of hplip-gui then, instead of hplip itself?08:53
ubottuLaunchpad bug 600504 in hplip (Ubuntu Maverick) "Dependency on python-notify makes hplip unsuitable for servers" [High,Confirmed] https://launchpad.net/bugs/60050408:53
tkamppeterpitti, xsane was already in Recommends:, so I added the other there.08:54
pittitkamppeter: hmkay08:55
tkamppeterpitti, one would need to check whether hplip-gui works without python-notify. Now I have python-notify also in the Recommends: of hplip-gui.08:56
pittitkamppeter: hplip-gui is fine; the problem was that hplip itself pulled in python-notify, which caused half the desktop to appear in the server isos08:57
tkamppeterpitti, this I understand. There were tons of small changes done by the Debian maintainer, he was probably not aware of hplip being used on servers or python-notify pulling so many GUI libraries. Therefore I have introduced hplip-gui earlier.09:00
IraqiWho know APT line for ubuntu please? thanks:)09:28
Iraqiand i invite you all to new channel #ubuntu-iq is support all things  :)09:29
=== tkamppeter_ is now known as tkamppeter
=== Daviey_ is now known as Daviey
=== almaisan-away is now known as al-maisan
=== smb is now known as smb-afk
=== al-maisan is now known as almaisan-away
=== almaisan-away is now known as al-maisan
=== MacSlow is now known as MacSlow|lunch
=== MacSlow is now known as MacSlow|lunch
seb128didrocks, let's move there12:38
seb128so current maverick CDs have notify-osd and notification-daemon installed12:38
seb128does anybody has a clue why notification-daemon get installed?12:38
seb128knowing that notify-osd provides notification-daemon12:39
didrocksnotify-osd is providing notification-daemon and only notify-osd is in the seeds12:39
seb128I've no found any versioned depends on it either that would explain it being pulled in12:39
=== smb-afk is now known as smb
smbpitti, Would you be able to rescue ddebs from the last update? ddebs.ubuntu.com is again stripped of any Lucid12:45
pittismb: I rescued everything yesterday, and fixed the problem of disappearing amd64 ddebs12:54
smbpitti, Hm, was I blind this morning then?...12:54
pittino, you aren't12:54
pittismb: the maverick ddebs are back12:55
pittibut the lucid ones not12:55
pittismb: but the lucid ones were built ealier than a week ago, so there's nothing to rescue12:55
pittibut something on them must be wrong still, since they keep disappearing12:55
smbpitti, Right. I feared that would be related to the build time of proposed12:56
smbpitti, So I should get together with you right after the next proposed upload to Lucid12:56
pittiwell, we need to find out why they disappear in the first place; currently RTFS12:57
pittismb: ah, I still have the "-image-debug" hack in place12:57
pittismb: which I should remove now, since the lucid/maverick ones are properly named now?12:57
smbRight they are12:58
pittidone12:58
pittiso the maverick ones should stay now12:58
pittiand the next lucid round as well12:58
smbHm, one would hope the hack would be a either or hack and not a one xor the other. :)12:58
pittihm, but now the karmic ones will disappear..12:58
seb128cjwatson, pitti: do you have any clue about notification-daemon?12:59
pittiseb128: what about it?12:59
seb128cjwatson, pitti: or rather about why it's on the current iso, cf backlog half an hour ago or so12:59
pittiseb128: will look in a bit, still messing with ddebs12:59
seb128pitti, thanks12:59
smbpitti, Depends on whether the rebuild of debian would get accepted. ;-) But then this would still leave hardy and jaunty out. Not to mention dapper, though I cannot remember anyone asking for _those_13:00
pittismb: the one-liner to rename the packages certainly would :)13:01
smbpitti, I guess I cannot win against that argument. :)13:01
pittiso as for the "or" hack, I could change the hack to never clean linux ddebs at all13:01
cjwatsonseb128: looking13:01
pittibut they'd pile up quickly13:01
seb128cjwatson, thanks13:02
smbYeah, I think that would soon make is the sworn enemy of IS13:02
* pitti ponders13:02
cjwatsonseb128: recommends from libnotify113:02
smbpitti, I would have guessed the hack would be something like if packagename modufied from old way has a matching package in archive or moified the new way has one, then it can stay13:03
cjwatsonwhich comes in from python-notify and gnome-settings-daemon (at least)13:03
seb128cjwatson, but it's not versioned?13:03
smbpitti, But I might think in a too simple way13:03
seb128cjwatson, and notify-osd provides notification-daemon13:03
seb128cjwatson, that worked in lucid, I don't see what changed...13:04
cjwatson"not versioned"?13:04
cjwatsonoh13:04
seb128cjwatson, well the provides should work if the depends is not versioned13:05
cjwatsonI'll look in more detail13:05
seb128thanks13:05
cjwatsonah yes, I see13:05
pittismb: ah, I'll just special-case the versions < 2.3213:05
cjwatsonthe thing is, libnotify1 is coming in through a package in the desktop-common seed (python-notify)13:05
smbpitti, That sounds simple and reasonable13:05
cjwatsonnotify-osd is seeded in desktop, not desktop-common13:05
cjwatsonso when germinate is processing desktop-common, it doesn't yet know that notify-osd is to be preferred13:06
cjwatson(or, put another way, what handles notification-daemon in flavours other than Ubuntu?)13:06
smbpitti, And if we ever get Karmic to the other side, we can lower that barrier13:06
cjwatsonchanging the ddeb names for karmic is a much smaller proposition than rewriting the whole packaging, and does not depend on the latter in any way13:07
cjwatsonI would have no problem with changing the ddeb names being part of an SRU13:07
cjwatsonthat change on its own would be effectively reviewable13:07
* smb covers13:07
smbI don't want to start over on that thread. I am awaiting whatever comes from above13:08
seb128cjwatson, hum ok, thanks, what would be the best way to solve that then? having notify-osd in desktop-common could be an issue for other desktops?13:09
seb128didrocks, ^13:09
cjwatsonseb128: maybe avoid using python-notify in stuff that's in desktop-common13:09
didrocksseb128: thanks, saw that, so it's an install order issue as we inferred. thanks cjwatson :)13:10
cjwatsonso either rip the python-notify dep out of hplip, or move hplip to the various desktop seeds rather than having it in desktop-common13:10
seb128I think I've read pitti talking about that earlier today13:10
* seb128 reads backlog13:10
cjwatsonbecause effectively, this is happening because desktop-common isn't sufficiently self-contained13:10
=== echidnaman is now known as JontheEchidna
pittiI'm on the phone, but I dealt with that this morning13:11
pittibrb13:11
seb128juil. 01 09:57:43 <pitti>tkamppeter: hplip-gui is fine; the problem was that hplip itself pulled in python-notify, which caused half the desktop to appear in the server isos13:11
seb128ok, so another case of pitti fixing the bug before having it reported ;-)13:11
seb128cjwatson, thanks a lot13:11
seb128"hplip (3.10.5-3ubuntu2) maverick; urgency=low13:12
seb128  * debian/control: Drop python-notify to suggests, it's pulling half of13:12
seb128    the desktop into server images."13:12
seb128didrocks, ^ that's fixing it for the record13:12
didrocksoh nice :-)13:12
tkamppeterpitti, why did you set bug 600504 to "Fix Committed" when you have uploaded the fix already hours ago?13:13
ubottuLaunchpad bug 600504 in hplip (Ubuntu Maverick) "Dependency on python-notify makes hplip unsuitable for servers" [High,Fix committed] https://launchpad.net/bugs/60050413:13
pittire13:13
pittitkamppeter: since it wasn't quite correct; adding the recommends to -gui is better13:13
pittitkamppeter: and you said that you committed that to debian/bzr13:13
seb128pitti, just for the record that breaks notify-osd on une13:13
seb128well broke13:14
pittiseb128: hplip? hardly..13:14
seb128but the bug is in alpha2 une isos13:14
pittiseb128: oh, I see what you mean13:14
pittiseb128: because hplip only got that dependency yesterday13:14
seb128pitti, it does, it brings notification-daemon in which is prefered to notify-osd13:14
pittiseb128: ok, understood; thanks for figuring out13:14
tkamppeterpitti, I have both currently both Recommends: of hplip-gui and Suggests: of hplip. I did not upload it to Ubuntu yet.13:14
pittiseb128: so this is essentially "fix committed", the next daily will have it13:15
didrocks(I have added une session to notify-osd dbus service startup so that even if it does reproduce, we won't have that later on)13:15
pittitkamppeter: sounds fine; thanks!13:15
seb128pitti, right, and didrocks fixed notify-osd to include "une" in the list of sessions which prefer notify-osd over notification-daemon13:15
pittisplendid13:15
tkamppeterpitti, should I upload the current version into Ubuntu, so that this gets the one of a2?13:15
pittitkamppeter: no hurry; it won't get into a2 any more anyway13:16
pittitkamppeter: the server ISOs got respun, and their size is fine again13:16
pittitkamppeter: and it's not a real issue for desktops, just a minor inconvenience for UNE; but it's only alpha-213:16
pittismb: ddeb-retriever h4ck adapted; should DTRT for jaunty/karmic/lucid/maverick now13:17
smbpitti, PHP (probably help paying) :)13:18
pittismb: :)13:18
smb(should have meant probably helps praying, but i seem too weak to hit the keys)13:19
pittismb: Freudian typo, eh?13:19
smbpitti, Heh, maybe. Who knows.13:20
smbDoesn't help to switch between two keyboards13:20
pittismb: you got a new shiny ergonomic one?13:21
smbpitti, Nah, just the other one and the Thinkpad on which irc is running13:21
=== fddfoo is now known as fdd
=== MacSlow|lunch is now known as MacSlow
=== dendro-afk is now known as dendrobates
=== sconklin-gone is now known as sconklin
=== sebner_ is now known as sebner
=== dendrobates is now known as dendro-afk
=== Amto_res is now known as Amto`OFF
=== dendro-afk is now known as dendrobates
=== bjf[afk] is now known as bjf
=== deryck is now known as deryck[lunch]
apwpitti, when do you do your magic for a milestone 'start' for a-3 ?16:30
pittiapw: I don't have to16:30
pittiapw: it's built into the code16:30
pittias soon as the previous milestone ended, it's considered the start of the current one16:30
pittiso tomorrow all charts will be "clean" and trend line will be correct16:31
apwpitti, oh so it will happen automatically then, the 'clear out' for the later milestones16:31
apwawsome ...16:31
pittithis works well as long as you do the planning before the milestone starts16:31
apwpitti, yep indeed16:31
pittiapw: yes, it rewards planning which happens on time :-P16:31
directhexso when does main unfreeze?16:37
directhexi want to finish one last test-build of mono 2.6.3-3 before sending it to experimental, and merging from that16:37
pittidirecthex: go ahead16:38
directhexpitti, few hours left of my test-build on agricola before it's an issue, I just wanted to check16:39
directhexARM test-builds are slow16:39
=== dendrobates is now known as dendro-afk
chrisccoulsondoko - do you know what creates /usr/lib/jvm/java-6-sun-1.6.0.20/jre/plugin/i386/ns7/libjavaplugin_oji.so as an alternative for mozilla-javaplugin.so on hardy? i don't see it created on my machine with all the sun-java6 packages installed17:01
chrisccoulsonbut some users have that selected as an alternative, which is causing some problems17:02
=== dendro-afk is now known as dendrobates
dokochrisccoulson: I think this is the old sun plugin, maybe we need to remove the old alternative on upgrade17:04
dokobut I'm afk now17:04
chrisccoulsondoko - yeah, that's what i'm thinking. do you think we should do a sun-java6 upload to remove the old alternative, and increase the priority of the libnpjp2.so plugin?17:06
chrisccoulsonthat's the one they should be using really17:07
chrisccoulsonanyway, thanks17:07
=== deryck[lunch] is now known as deryck
=== MacSlow is now known as MacSlow|break
=== amitk is now known as amitk-afk
=== pitti changed the topic of #ubuntu-devel to: Alpha-2 released | Development of Ubuntu (not support, not app development) | #ubuntu for support and general discussion for dapper-lucid | #ubuntu-app-devel for application development on Ubuntu | http://wiki.ubuntu.com/UbuntuDevelopment | See #ubuntu-bugs for http://wiki.ubuntu.com/HelpingWithBugs
* ogra grumbles about weird cups udev rules 18:52
ograit forcefully loads parport and lp on my ARM board18:53
pittiogra: that's in the init script now18:58
pittiit's not really how it's meant to work indeed18:58
pittiparport_pc and parport ought to be loaded automatically via modaliases18:58
ograpitti, well, it produces oopses on my beagleboard18:59
ograand its pretty unlikely you will ever find a parport on arm HW :(18:59
ograerr18:59
ogra:)18:59
=== dendrobates is now known as dendro-afk
=== dendro-afk is now known as dendrobates
=== al-maisan is now known as almaisan-away
=== MacSlow|break is now known as MacSlow
=== fta__ is now known as fta
Daenyth|WorkHi. I'm trying to split one of the packages I maintain and the docs are a little bit confusing because they expect a handrolled debian/rules file where I have one using cdbs. I have more details posted here: http://superuser.com/questions/158689/how-to-create-a-debian-split-package-when-using-debhelper19:54
Daenyth|WorkHow do I need to change my rules file to get both packages built?19:54
Daenyth|WorkAre there any examples of packages that are split and use cdbs like that?20:13
=== sebner_ is now known as sebner
cndI see that the xorg-server source package bzr branch hasn't been auto-updating right: https://launchpad.net/ubuntu/+source/xorg-server20:53
cnddoes anyone know who handles these issues?20:53
pitticnd: james_w mostly20:56
cndpitti: ok, thanks20:56
=== cnd is now known as cnd`
=== dendrobates is now known as dendro-afk
=== fta__ is now known as fta
ricotzpitti, hello21:17
pittihello ricotz21:17
ricotzpitti, i just want to kindly remind you of docky, perhaps you can let is pass despite the lack of verifications21:18
pittifor an update of this size we should get more than just one or two "works for me"21:18
ricotzi release the new upstream version today which will be in debian/maverick soon21:19
ricotzthere are a least a verfication which isnt noticed for the most serious bug https://bugs.edge.launchpad.net/docky/+bug/55556221:20
ubottuLaunchpad bug 555562 in docky (Ubuntu Lucid) "crash accessing Gnome keyring in Lucid" [Critical,Triaged]21:20
pittiricotz: do you know some folks who are using docky and can confirm that the proposed version still works for them?21:21
LaneyMe!21:22
LaneyDid I verify it already?21:22
Kanohi, when will the webfrontend give results for maverick21:22
Kanoi really need that info21:23
ricotzpitti, most people we are in contact with using the development or stable ppa21:23
pittiLaney: seems not21:23
Laneypitti: which is the master bug?21:23
pittiLaney: see http://people.canonical.com/~ubuntu-archive/pending-sru.html, pick one :)21:24
Laneyok then21:24
* Laney blinks, that's a lot21:24
ricotzpitti, if there are regressions i think they are overweighted by bugfixes multiple times21:25
pittismall regressions perhaps, but if it stops working for someone else, it's worse21:25
pittiknown bugs are better then breaking existing functionality21:25
ricotzyes, for now the keyring bug is stopping 2.0.2 from working very often21:26
Kanoalso could you update libdrm to 2.6.21, thats critical for libva21:26
pittimathiaz: are you usually sponsoring landscape-client? There are two versions in -proposed, and it would be good if you could upload the current one with -v to include both changelog records; or alternatively condense them into one record21:26
Kanono git version needed, just tested the debian experimental release21:27
ricotza21:28
pittigood night21:28
ricotzpitti, good night21:28
ricotzLaney, the next sru proposal will have 24 connected bugs21:29
=== dendro-afk is now known as dendrobates
Laneythis is a lot of work for SRU verification21:30
Laneyit'd be good if you could organise testing21:30
ricotzLaney, it is unlikely to verify all of them in a normal timeframe21:33
Laneyit's not really on to subjugate the process though21:34
LaneyI verified a few of the bugs21:35
shadeslayerLaney: hi21:36
Laneyhello21:36
shadeslayerLaney: can you do a SRU21:37
shadeslayerhttps://bugs.launchpad.net/ubuntu/+source/desktopcouch/+bug/56537621:37
ubottuLaunchpad bug 565376 in desktopcouch (Ubuntu) "bughugger does not work in kubuntu lucid" [Undecided,New]21:37
shadeslayerpatch is attached21:37
LaneyI can't upload that package :(21:37
shadeslayer:(21:37
Laneythat's not a properly formatted SRU anyway21:38
ricotzLaney, of course i could through these bugs myself but this wont be real testing this should be done by real users ;-)21:38
Laneyyour diff should be targetted to lucid-proposed and the version should be ...3.121:38
Laneyand you need a test case21:38
shadeslayerLaney: ahh21:38
ricotzLaney, thanks for testing some of these bugs21:38
Laneyricotz: You just need to get a couple of users to test the bugs for you21:39
LaneyIf you call for this then I'm sure you'd get it21:39
Laneyit's easier if the bugs have decent steps to reproduce in the description21:40
ricotzyeah, but i like to invest my spare time in really developing docky ;-)21:41
ricotzLaney, who is supposed to change the tag to verfication-done?21:42
LaneyI'm actually not sure21:43
sebnerricotz: I've seen normal users and/or archive admin doing this21:44
ricotzScottK, hello, can you answer this question about sru bugs? ^21:44
ricotzsebner, yeah, i have seen this, but i thought there might be a rule21:46
sebnerricotz: first come first serve maybe? ^_^21:46
ricotzLaney, sebner, i will change the tag myself for the verified ones21:48
LaneyI don't know if more than one verification should happen21:49
* sebner agrees with Laney 21:49
ricotzLaney, https://bugs.edge.launchpad.net/docky/+bug/58221221:52
ubottuLaunchpad bug 582212 in Docky "GMail docklet opens wrong "compose mail" page when using Google Apps" [Low,Fix released]21:52
Laneyok then21:53
=== fta_ is now known as fta
=== dendrobates is now known as dendro-afk
jdongHmph does Ubiquity ignore the http mirror hostname preseeded?22:33
jdongIt seems to remember the directory on the server but look on CC.archive.u.c22:33
jdonghmmmm *scratches head* so what do I have to preseed to manually set the mirror?22:57
dupondjehttps://launchpad.net/ubuntu/+archivemirrors => who maintains this list ? there seems to be an small error in it :)23:10
=== lifeless_ is now known as lifeless
=== asac_ is now known as asac
jdongcjwatson: maybe dumb question, but does ubiquity not support preseeding mirror/http/hostname? It seems to always set it back to CC.archive.ubuntu.com...23:37
cjwatsonjdong: you'll at least need mirror/country set to "manual" as well, but it should support it23:48
* cjwatson -> bed23:48
jdongcjwatson: hmm I've got d-i mirror/country string manual set23:48
jdongand d-i mirror/http/hostname string my_host...23:48
jdongit seems to ignore mirror/http/hostname, but accept mirror/http/directory23:49
jdong(10.04)23:49

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