[04:36] <ScottK> slangasek: Thank you for fixing pinentry.
[04:41] <mgunes> lool, could you look at bug #195257? why may it be that the fix (in Debian) hasn't made it into Ubuntu?
[04:41] <ubotu> Launchpad bug 195257 in acpi-support "missing man page for acpi_fakekey" [Low,New] https://launchpad.net/bugs/195257
[04:45] <lamont> mgunes: it could be because things don't automatically sync now (for the rest of the hardy development cycle...)
[04:46] <mgunes> lamont, it's been fixed a year ago.
[04:46] <lamont> heh
[06:06] <slangasek> ScottK: sure thing
[07:08] <warp10> Good morning
[07:09] <dholbach> good morning
[07:09] <ion_> morniŋ
[07:10] <warp10> hey dholbach
[07:10] <dholbach> hey ion_
[07:16] <pitti> Good morning
[07:17] <Mithrandir> hiya Martin
[07:18] <pitti> pochu: hm, indeed that's actually supposed to work
[07:18] <pitti> pochu: where did you put that file?
[07:19] <pitti> jdong: filed wrongly in what way? If it's wrong, then it should be reported to the hal upstream guys, since many distros use that patch now and it works for many people
[07:19] <pitti> jdong: it's ultimately a kernel problem, but if you have a specific case which could be worked around in hal, we can do that
[07:42] <bryce> pitti: if you have a moment, I put in my core-dev application a few days ago and would greatly appreciate having your feedback about it on the motu-council list
[07:47] <StevenK> Oooh.
[07:48] <Mithrandir> pitti: thoughts on updating bluez-gnome and obex-data-server to 0.23 and 0.3?  They seem to work fine for me in light testing here at least.
[07:48] <Mithrandir> pitti: it's blocked by getting o-d-s into main though
[07:50] <pitti> Mithrandir: can't really judge without looking at changelogs at least
[07:51] <pitti> Mithrandir: I can do the MIR today, though
[07:52] <Mithrandir> pitti: bluez-gnome changelog: http://pastebin.com/f322f84e2 ; http://pastebin.com/f49eade61 for o-d-s
[07:52] <Mithrandir> current bluez-gnome is 0.15
[07:55] <pitti> Mithrandir: o-d-s looks good to me, if the impact of the ABI change is confined
[07:55] <pitti> API
[07:56] <pitti> Mithrandir: but I don't see any reverse dependencies, so that should be good?
[07:56] <Mithrandir> pitti: new bluez-gnome depends on it, but nothing else.
[07:56] <Mithrandir> o-d-s got through NEW about a week ago
[07:57] <pitti> ah, then it can hardly break much already :)
[07:58] <Mithrandir> indeed, but I figured I'd at least pass it by another release team member rather than just sneaking it in. ;-)
[07:59] <pitti> Mithrandir: bluez-gnome looks fine, by and large bug fixes; did you test the two new features? (file sharing and file receiving)
[08:00] <Mithrandir> pitti: I tested file sending, I'll test the two others before uploading.
[08:00] <Mithrandir> (I need to dig out my laptop then, since my phone doesn't seem to have a usable obexftp client
[08:00] <Mithrandir> )
[08:00] <pitti> Mithrandir: cool; go ahead then
[08:00] <Mithrandir> cheers
[08:03] <Mithrandir> file receiving works just fine.  I'd like to have it forced to authenticate beforehand, but absolute worst case is somebody fills up your desktop with pr0n.
[08:14] <stgraber> morning
[08:23] <pochu> pitti: it's at /usr/share/apport/package-hooks/tracker.py
[08:23] <pitti> pochu: that sounds correct
[08:26] <pitti> pochu: ah, I bet I know
[08:26] <pitti> pochu: there's a missing \ after the 'and'
[08:26] <pitti> pochu: do you get this exception in /var/log/apport.log?
[08:26] <pitti> not sure whether I got the logging right
[08:27] <pitti> pochu: in the line-wrapped string, too
[08:38] <pochu> pitti: I thought python didn't need those "\"
[08:38] <pochu> pitti: thanks, I'll add it and test again :)
[08:38] <pitti> pochu: it does; it heavily relies on correct whitespace and indentation :)
[08:39] <seb128> hey there
[08:39] <pochu> morning seb128
[08:39] <pochu> pitti: oh, it's not needed for lists and the like, right?
[08:40] <pitti> pochu: right, not within parenthesis, brackets, and braces
[08:40] <pitti> hey seb128
[08:41] <seb128> hello pitti pochu
[08:42] <seb128> pitti: I didn't fix the retracer, it has a login error and I'm not sure if that's p-l-b error or if you need to reinstall a cookie or something
[08:42] <pitti> seb128: ok, I'll have a look; thanks
[08:43] <seb128> pitti: thanks
[08:44] <pochu> pitti: oh, and yes, the crashes are listed in apport.log
[08:48] <pitti> pochu: cool
[08:53] <ScarFreewill> please excuse me being off-topic but I've been searching over an hour and I don't know where else to look, I'm trying to send Virtual Key Codes to my system but I've got no idea how to do this, does anyone know an open source appliction that does this? I'd like to do this in java, though i don't think this is possable so if i can do it in c/c++ and run it either through jni or runtime.exec, thanks for you time
[08:53] <ScarFreewill> by the way, its for an alarm clock :-P
[09:01] <geser> good morning
[09:06] <pitti> hi geser
[09:09] <geser> Hi pitti
[09:28] <Mithrandir> mvo: any idea about a fix for https://bugs.launchpad.net/ubuntu/+source/update-manager/+bug/153980 ?
[09:28] <ubotu> Launchpad bug 153980 in update-manager "7.04->7.10 "Upgrade" fails: "Getting upgrade prerequisites failed"" [High,Confirmed]
[09:28] <mvo> Mithrandir: let me check
[09:29] <Mithrandir> mvo: a friend of mine has the problem right now, so if you want him to test something for you, I can surely get him to post more info or test stuff for you
[09:31] <mvo> Mithrandir: the error messsage is given for different error conditions, it may mean that the archive server/mirror refuses the connection, that for some reason there is a authentication error or no network is available, if your friend could give me the output of /var/log/dist-upgrade/main.log, that would be much appreciated
[09:31] <asac> pitti: i have an action to figure out if we can produce ffox/xul/midbrowser locales out of launchpad somehow - even though there won't be a xpi export for hardy. my idea would be to map the locale chrome entities during build to .po files and then produce chrome jars from the .po files exported from launchpad. ok, what do you think in general? possible, impossible?
[09:32] <Mithrandir> mvo: thanks; just asked him.  He might join here too.
[09:32] <mvo> Mithrandir: thanks
[09:33] <pitti> asac: there's a project which does that
[09:33] <pitti> asac: formerly called 'pootle'
[09:33] <pitti> asac: now it's something else, carlos knows
[09:33] <asac> pitti: really?
[09:33] <kolla> mvo: hello.. you want my main.log
[09:33] <Mithrandir> mvo, meet kolla, kolla, meet mvo
[09:33] <kolla> *nod* :)
[09:33] <asac> pitti: well ... he didn't really know
[09:33] <mvo> kolla: hello! yes please
[09:34] <asac> pitti: so in general that should be possible. good.
[09:41] <pochu_> pitti: what makes apport not send more than 3 reports (or whatever its limit is)? I'd like to test this tracker hook but apport seems to be ignoring tracker-extract crashes now, perhaps because I've crashed it enough :)
[09:41] <pochu_> I don't have ~/.apport-ignore.xml
[09:42] <pitti> pochu_: there's a counter in the /var/crash/tracker.crash file
[09:42] <pitti> pochu_: just delete the file
[09:42] <pochu_> thanks
[09:44] <pochu_> "not a genuine Ubuntu package" -- I shouldn't have bumped the version to test it ;)
[09:45] <tarzeau> hi geser
[09:45] <geser> Hi tarzeau
[09:45] <tarzeau> geser: can you make ubuntu have http://packages.ubuntu.com/bubbros ?
[09:49] <geser> tarzeau: unfortunately not, as we are already in FeatureFreeze which also means no new packages anymore
[09:49] <tarzeau> geser: pity, but maybe for after that?
[09:50] <geser> for intrepid (= hardy+1) it should get autosynced
[09:50] <tarzeau> geser: ok thanks
[09:51] <pitti> thekorn: hm, current hardy p-lp-bugs seems to wildly mix 'attachment' and 'attachments' method attributes; you recently fixed that in trunk, right?
[09:58] <pitti> seb128: ah, seems that p-lp-bugs trunk works
[09:58] <seb128> pitti: ah, good
[09:58]  * pitti installs that
[10:01] <pochu_> pitti: oh, I need to scape the single-quoted sentence
[10:01] <pochu_> SyntaxError: EOL while scanning single-quoted string
[10:01] <pitti> right, what I said; just append a \
[10:02] <pochu_> yay it works now :)
[10:08] <pitti> pochu_: rock
[10:08] <pochu_> pitti: thanks for the help
[10:09] <pitti> pochu_: you're welcome, no problem
[10:30] <thekorn> pitti, yes, the attachement(s) issues should be fixed in the trunk, hopefully
[10:31] <pitti> thekorn: seems like it, a first test looks good; thank you!
[10:39] <pitti> thekorn, seb128, bdmurray: phew, retracer is running again; I installed p-lp-bugs 'main' head into the hardy chroot, which seems to help
[10:39]  * seb128 hugs pitti
[10:40]  * thekorn hugs pitti 
[10:41]  * \sh hugs slangasek for fixing ncurses...
[10:44] <TheMuso> =/c
[10:45] <seb128> hi TheMuso
[10:45] <TheMuso> Hey seb128.
[10:46] <seb128> TheMuso: there is new atk and at-spi tarballs to package if you want to do those updates
[10:46] <TheMuso> seb128: Oh ok, thanks. I guess they'll appear on gnome-announce some time, but I'll get on them.
[10:47] <seb128> TheMuso: not sure if every maintainer is using the list, but there is the ftp list which has all the uploads
[10:47] <TheMuso> seb128: Ah ok, I'll get on that as well, thanks.
[10:49] <seb128> you are welcome
[10:53] <seb128> dholbach: there is a guy on #ubuntu-bugs which would like to discuss making 5-a-day easier, is there a chan or something where I could point him since you are not hanging there? ;-)
[10:58]  * ogra_cmpc wonders whats th reason for gnbd to exist, it semms not to offer more than nbd
[11:01] <ogra_cmpc> thegodfather, ^^ do you know ??? is there any extra feature gnbd has over nbd ?
[11:11] <pitti> mvo: is there an equivalent for apt.Cache()[pkg].candiateOrigin for the installed version?
[11:12] <Mithrandir> hm, why is it ps no longer sorts by start time?
[11:12] <Mithrandir> (yes, I know this changed a while back)
[11:13] <mvo> pitti: no and most of the time it would give anyhting interessting because apt does not store were a package came from. so if you have "foo 1.0" and "foo 1.1" and only 1.1 is in the archive, the information were 1.0 came from ist lost
[11:13] <pitti> mvo: ok
[11:13] <pitti> mvo: alright, I'll continue to use the candidateOrigin then; bdmurray pointed out that apport would currently file bugs for PPA packages
[11:14] <pitti> mvo: component: 'main' archive: 'gutsy' origin: 'Ubuntu' label: 'Ubuntu' site 'ppa.launchpad.net' isTrusted: 'False'
[11:14] <pitti> is what I get for a PPA package
[11:14] <pitti> i. e. origin is still 'Ubuntu'
[11:14] <pitti> mvo: should I rely on isTrusted==True?
[11:14] <mvo> pitti: sounds like a bug in the ppa archive Release file to me
[11:14] <pitti> mvo: well, it is a PPA for gutsy
[11:15] <mvo> pitti: for now, until they change either origin or label
[11:15] <pitti> I'm just not sure how to interpret these field
[11:15] <pitti> mvo: what does isTrusted mean? that the release file could be verified using the archive keyring?
[11:15] <mvo> pitti: yes
[11:15] <mvo> pitti: that together with origin or label sounds like a good way to me
[11:16] <pitti> hm, so using this might be a little too harsh?
[11:16] <mvo> pitti: I would also suggest to ask PPAs to have a different origin or label than the main archive
[11:16] <pitti> mvo: I'll file that bug on soyuz
[11:17] <pitti> mvo: but anyway, for now, do you recommend "isTrusted==True && origin == Ubuntu" or "origin==Ubuntu && not site.startswith('ppa')" ?
[11:17] <mvo> pitti: the later I would say
[11:18] <pitti> mvo: ok, thank you
[11:18] <mvo> thanks
[11:20] <pitti> mvo: ah, yay LP dup detection on filing; it's already there, bug 140412
[11:20] <ubotu> Launchpad bug 140412 in soyuz "Release file should show that it comes from a PPA" [Wishlist,Triaged] https://launchpad.net/bugs/140412
[11:30] <ogra> soren, did you ever think about creating a loopback blockdevice with nbd to get your grub-install stuff going ?
[11:41] <tkamppeter> pitti hi
[11:41] <pitti> hey tkamppeter
[11:42] <tkamppeter> pitti, WDYT about an UVF for CUPS 1.3.6? It contains many bug fixes, especially it seems that a crash auditing was done this time.
[11:43] <pitti> tkamppeter: we don't have UVF any more; if .6 doesn't have new features and only bug fixes, we should definitively get it
[11:43] <tkamppeter> Yes, Upstream Version Freeze was replaced by Feature Freeze.
[11:47] <\sh> tkamppeter: good to see you....
[11:48] <\sh> tkamppeter: do you know any way to use a dell CL3110cn printer with ubuntu? i didn't find any usable reference for it on cups or linuxprinting.org
[11:48] <asac> pitti: bug 193941 ... do i just ened to add need-XXX-retrace to tags to make retracers look at this?
[11:49] <ubotu> Launchpad bug 193941 in firefox-3.0 "firefox crashed with SIGSEGV in __kernel_vsyscall()" [Undecided,New] https://launchpad.net/bugs/193941
[11:49] <tkamppeter> pitti, I went to the announcement and all items are bug fixes, no new features. AFAIK the CUPS upstream policy is that minor releases are only bug fixes, so we should be able to use the highest CUPS 1.3.x coming out before release of Hardy.
[11:49] <\sh> tkamppeter: the printer is attached as IP printer...and only starts working when using it via port 9100
[11:49] <asac> pitti: hmm. how comes that this isn't private?
[11:49] <pitti> asac: right; usually it gets done automatically, but due to a load of bugs in p-lp-bugs we lost some tags
[11:49] <pitti> tkamppeter: agreed
[11:50] <tkamppeter> pitti, can you do the update to also cover Debian?
[11:50] <asac> pitti: should i make it private now? or is it too late?
[11:50] <pitti> tkamppeter: it should be done in trunk and merged to the ubuntu branch, yes
[11:50] <pitti> asac: you can do it
[11:50] <asac> pitti: ok, and subscribe ubuntu-crashes-main, right?
[11:51] <pitti> asac: but not sure why it was un-private-ed
[11:51] <asac> pitti: it happens from time to time
[11:51] <pitti> the bug reporter himself can choose to do so
[11:51] <asac> pitti: well, it looks like that those are never private
[11:51] <tkamppeter> \sh, you should try this driver: http://foo2hiperc.rkkda.com/
[11:51] <asac> (vs. unprivated)
[11:51] <pitti> asac: the retracer will automatically subscribe the right team
[11:52] <asac> ok
[11:52] <\sh> tkamppeter: does it mean, that dell uses oki hw and just relabled it?
[11:54] <tkamppeter> pitti, what about an FF exception of foo2zjs make Hardy support a lot more ugly winprinters, and it cannot break anything as other printers are supported by other drivers.
[11:54] <pitti> tkamppeter: -> slangasek
[11:54] <pitti> tkamppeter: please file an FF exception request and subscribe ~ubuntu-release
[11:54] <pitti> tkamppeter: it will have good chances
[11:57] <tkamppeter> \sh sorry, the model name looks so similar to the ones of the OKIs. In general Dell does not develop printers, they relabel printers of many brands, Samsung, Lexmark, ... So the chance is not bad that yours is an OKI, but it can also be a Lexmark.
[11:58] <\sh> tkamppeter: grmpf..I dislike dell serverwise .. now I know another reason printer-wise
[11:58] <tkamppeter> \sh, so try the mentioned OKI driver but also this Lexmark driver: http://foo2slx.rkkda.com/
[11:59] <tkamppeter> Perhaps you can even try the driver for the cheapo Samsung color lasers (SpliX).
[12:01] <\sh> tkamppeter: ok...let's see what the outcome is...:)
[12:04] <tkamppeter> \sh, if you get your printer to work, tell me which model/driver selection did the trick. Thanks.
[12:06] <\sh> tkamppeter: will do :)
[12:07] <ivan_> hi, i've just fixed latex209-bin package for ubuntu 7.10 in order to prevent bug https://bugs.launchpad.net/ubuntu/+source/latex209/+bug/149145
[12:07] <ubotu> Launchpad bug 149145 in latex209 "package latex209-bin 25.mar.1992-10 failed to install/upgrade: Unterprozess post-installation script gab den Fehlerwert 1 zurück" [Undecided,Confirmed]
[12:07] <ivan_> but i'm in doubt what to do nex
[12:07] <ivan_> t
[12:08] <ivan_> how should i contact a supervisor of package in order to publish it?
[12:17] <cjwatson> ivan_: start with https://wiki.ubuntu.com/MOTU/TODO/Bugs
[12:17] <persia> ivan_: https://wiki.ubuntu.com/SponsorshipProcess is the process description.  As latex209 is in universe, you might ask for help in #ubuntu-motu, although you may be redirected to https://wiki.ubuntu.com/MOTU/Contributing
[12:20] <ivan_> cjwatson, persia: thanks
[12:24] <ogra> geez, thast cool ... http://spritesmods.com/?art=picframe&amp;page=1
[12:30] <Riddell> evand, cjwatson: kubuntu kde 4 daily-live CD working surprisingly well, however ubiquity freezes when trying to access debconf, might something be missing from the CD?
[12:31] <cjwatson> Riddell: would be difficult to break it in that particular way by leaving stuff out without ubiquity being totally absent - anything interesting in the usual logs?
[12:31] <ogra> Riddell, probably something thats pulled in by using universe that wouldnt be available otherwise ? (a recommends etc) ?
[12:31] <Riddell> cjwatson: syslog just has the one entry "Feb 25 12:22:39 ubuntu ubiquity[10785]: Ubiquity 1.7.10"
[12:32] <cjwatson> Riddell: how do you know it's freezing when trying to access debconf?
[12:32] <Riddell> cjwatson: hmm, hang on, installer/debug has "QMutex::lock: Deadlock detected in thread"
[12:32] <cjwatson> ogra: I really don't think guessing is helpful ...
[12:33] <ogra> sorry
[12:33] <cjwatson> Riddell: is the startup process in ubiquity-dm still right for kde4?
[12:34] <Riddell> cjwatson: ok, I got it
[12:34] <Riddell> it's the kde 4 widget style
[12:34] <cjwatson> ogra: (don't get me wrong, thanks, it just seems like a bit of an extensive logical leap :-) )
[12:34] <Riddell> it loads QtDbus which breaks with python-qt-dbus is loaded
[12:34] <ogra> cjwatson, i dont, but i know i'm a big guesser and shouldnt be :)
[12:35] <cjwatson> ogra: the way I handle that is to figure out some evidence which might support or contradict the hypothesis, and ask about that
[12:35] <cjwatson> intuitive guesswork is important, but also the careful deduction that can support it :)
[12:36] <ogra> yep ... :)
[12:36] <cjwatson> Riddell: ah, and ubiquity loaded python-qt-dbus first and got in the way?
[12:36] <ogra> i'll try to not start off with a guess first place in the future :)
[12:36] <Riddell> yes, so as it happens ogra is right, since that kde 4 widget style is something in universe which has never been used with ubiquity before :)
[12:37] <cjwatson> :-)
[12:37] <cjwatson> probably easier to get there stepwise though
[12:51] <Riddell> ArneGoetje: qt 4 fonts aren't anti aliased by default, but I notice large text is (http://muse.19inch.net/~jr/tmp/qt4.png), any idea why?
[12:53] <ion_> I have the same problem, but i haven’t got around to investigating it.
[12:55] <ogra> ... phew ... what a changelog for edubuntu-meta ....
[12:59] <mvo> pitti: if you have a bit of time, it would be cool if you could review #195419 and let me know if that is ok to upload to gutsy-proposed
[13:02] <pitti> mvo: the diff appears twice because it's a symlink somewhere, right?
[13:03] <pitti> mvo: confirmed, bug updated
[13:13] <Mithrandir> pitti: I finally managed to test the file sharing thingy by using my N810.  It works fine then.
[13:13] <Mithrandir> pitti: so a new bluez-gnome is just waiting on the MIR. :-)
[13:15] <seb128> Mithrandir: speaking about new bluez-gnome, bug #190405
[13:15] <ubotu> Launchpad bug 190405 in bluez-gnome "[FFE] please upgrade bluez-gnome to 0.21" [Wishlist,Confirmed] https://launchpad.net/bugs/190405
[13:16] <Mithrandir> seb128: yes, it's including the bits from there
[13:16] <Mithrandir> it'll be 0.23
[13:16] <Riddell> evand, cjwatson: ok if I apply this patch and upload ubiquity? http://muse.19inch.net/~jr/tmp/ubiquity.diff
[13:16] <seb128> Mithrandir: ok, good ;-)
[13:16] <Mithrandir> seb128: are you following the OBEX bits of gvfs?
[13:17] <Mithrandir> (http://bugzilla.gnome.org/show_bug.cgi?id=509621)
[13:17] <ubotu> Gnome bug 509621 in general "Implement obex-ftp backend" [Enhancement,Assigned]
[13:17] <seb128> Mithrandir: not really no, but looks like bastien nocera is working on that
[13:17] <Mithrandir> seb128: yeah, though it's blocked on some changes in gvfs itself.
[13:18] <seb128> alex is responsive so that should be alright
[13:18] <Mithrandir> ok, coolie
[13:18] <cjwatson> Riddell: sure
[13:20] <seb128> mdke: hi
[13:23] <slytherin> Hi all. Can anyone comment on MIR for obex-data-server. I didn't file the bug, but I promised the bug report that I will handle it in his absence.
[13:26] <seb128> slytherin: is the mir team subscribed to the bug?
[13:27] <slytherin> seb128: yes it is.
[13:28] <seb128> slytherin: so they will likely comment on it when they have a look
[13:28] <slytherin> ok
[13:30] <slytherin> The reason I was asking is that recent versions of bluez-gnome are using obex-data-server as backend for file transfer and I am waiting for o-d-s to be in main for getting latest bluez-gnome in Ubuntu.
[13:43] <Chipzz> who has the "contentless ping" script here?
[13:44] <Mithrandir> http://err.no/src/contentless_ping.pl
[13:47] <seb128> slytherin: ah, you just joined, Mithrandir was speaking about that some minutes before you joined, he did the bluez-gnome 0.23 update
[13:48] <ktk> hi all
[13:48] <ktk> tkamppeter: here?
[13:48] <Mithrandir> seb128: we spoke about it in #ubuntu-mobile, but thanks. :-)
[13:48] <Chipzz> Mithrandir: thx! :)
[13:49] <pitti> Mithrandir: ack
[13:55] <cjwatson> mvo: bug 195121: gnome-app-install doesn't depend on software-properties-gtk any more. Should we add it to the desktop seed?
[13:55] <ubotu> Launchpad bug 195121 in ubiquity "software-properties-gtk is not installed on Hardy Alpha 5" [Undecided,New] https://launchpad.net/bugs/195121
[14:06] <asac> Riddell: back?
[14:08] <asac> Riddell: oh i see that you talked above ... do we have kde libs in archive yet that are GPLv3 compatible?
[14:08] <soren> ogra: No, but I doubt that'll work.
[14:09] <ogra> soren, works fine up to the point where i need to have partitions
[14:10] <Riddell> asac: yes
[14:11] <Riddell> asac: qt is now gpl 3 so gnash etc is fine
[14:11] <asac> Riddell: hmm i think they need some kde headers as well?
[14:11] <Riddell> asac: kdelibs is LGPL so no problem there for GPL 3
[14:12] <asac> ah ok, so it was just QT. great
[14:14] <mvo> pitti: yes, it appears twice because of the symlink
[14:15] <soren> ogra: Erm... Ok.
[14:16] <ogra> soren, grub sees it as real blockdev (which it actually is) if you adjust device.map it works fine ... the prob s that i want two partiotions and nbd or udev seems to not support them (not sure yet)
[14:16] <jdstrand> lamont: re bind9 autotool-- I was not aware I did.  I checked my debdiff against -3 and I only see my apparmor changes
[14:17] <mvo> cjwatson: I added it as a recommends for synaptic, but it should probably go to the seeds as well
[14:17] <soren> ogra: kpartx is your friend.
[14:17] <ogra> soren, yeah, i saww you use it
[14:17] <ogra> i didnt have time today yet, to much general work
[14:17] <ogra> i'll look into it
[14:18] <ogra> someone told me debian live uses grub to install to loop devices, but looking at the code didnt show any evidence for that
[14:18] <soren> ogra: To be honest, I forget what the exact problem was..
[14:19] <ogra> i need/want to create an image from scratch that has two patitions and uses grub ....
[14:19] <soren> ogra: Oh, that's not what I meant.
[14:19] <ogra> like: dd it, fdisk on it, make filesystems and install grub ...
[14:20] <soren> ogra: I mean I forgot why I couldn't get grub to do what I wanted it to do.
[14:20] <ogra> it doesnt support loop devices
[14:20] <cjwatson> mvo: yes, please do - I'll assign you the bug?
[14:20] <mvo> cjwatson: yes, thanks
[14:21] <ogra> looks like your code works around that by picking pieces out of stage1 and 2 to assemble an mbr
[14:22] <keescook> Keybuk: say, can you look at my comments on 45842?  I wanted to get your thoughts before we worked on fixing it.
[14:22] <ogra> i'll use yours if i dont find a better way, but nbd loopdevices (unlike gnbd you actually can have loopback blockdevs with nbd) seem to be a tad cleaner here
[14:22] <lamont> jdstrand: hrm.. it's possible that I did... :)
[14:22] <lamont> the autofiles get regenerated when I build the source package
[14:23] <Keybuk> keescook: opening...
[14:23] <lamont> OTOH, they get built by bad means - bind needs some auto* love to come into the current version and have autoreconf -f -i -v work
[14:23] <Keybuk> keescook: err, I can't remember why it only waits for certain things
[14:24] <Keybuk> keescook: I think it was to avoid waiting on things like cdrom drives
[14:24] <sistpoty|work> infinity: any chance, that you could increase the inactivity kill timeout on the sparc buildd for ghc6 (it just takes its time..., see bug #194912)?
[14:24] <ubotu> Launchpad bug 194912 in ghc6 "ghc6 6.8.2-1ubuntu1 FTBFS on sparc" [Undecided,Confirmed] https://launchpad.net/bugs/194912
[14:24] <keescook> Keybuk: I'm not sure I follow (this is for NFS)
[14:24] <Keybuk> keescook: was reading the patch at the bottom, it reverts the fact that waitnfs only waits for /usr and /var right?
[14:25] <kirkland> Keybuk: that's one part of it
[14:25] <keescook> Keybuk: right, there's that part too.  It should still only be paying attention to network fs's when building that list
[14:26] <keescook> what do you think of using watershed to solve the races?
[14:27] <Keybuk> I'm not sure I ever understood what the races were?
[14:27] <Keybuk> or what the problem was
[14:27] <keescook> Keybuk: basically, the ifup script can run before rcS finishes
[14:27] <Keybuk> right
[14:27] <keescook> so it will try to NFS mount before mtab is ready
[14:27] <Keybuk> because it's run from udev
[14:28] <Keybuk> mtab isn't really that critical though?
[14:28] <keescook> right, and it needs to be there due to that.
[14:28] <keescook> (i.e. it should be an ifupdown script)
[14:28] <Keybuk> mtab shouldn't be
[14:29] <Keybuk> I don't see the link with mtab here
[14:29] <keescook> let's ignore the mtab bit -- that's not the real problem
[14:30] <Keybuk> ok
[14:30] <keescook> there appears to be problems with NFS coming up before we get to waitnfs in rcS
[14:30] <Keybuk> NFS should come up before waitnfs
[14:30] <Keybuk> the idea is that NFS is mounted from an if-up.d script
[14:30] <Keybuk> if-up runs from udev
[14:31] <keescook> well, that relates to the problem where waitnfs only waits for things in /usr and /var.
[14:31] <keescook> it should wait for all "auto" network fs's
[14:32] <Keybuk> but that shouldn't stop them being mounted in the first place?
[14:36] <keescook> Keybuk: afaict, the ifup script can run too early.
[14:36] <Keybuk> "too early" ?
[14:37] <keescook> Keybuk: and waiting until waitnfs is running is late enough to DTRT
[14:37] <keescook> it fails to mount... I'll turn to kirkland now, who has reproduced this....
[14:39] <keescook> Keybuk: kirkland and I will get a better determination on the root cause...
[14:39] <Keybuk> if you wait until waitnfs - you may as well just put mountfs where waitnfs is ;)
[14:39] <Keybuk> (with some kind of spin to wait until network is up)
[14:40] <keescook> heh, I thought so too, but it seems to do additional nice things when interfaces change via udev
[14:40] <kirkland> Keybuk: that was actually how I was working around the problem...  doing the mountnfs inside of the waitnfs
[14:40] <ArneGoetje> Riddell: cannot say for now, need to investigate by myself. Is this onm Alpha 5?
[14:42] <Keybuk> kirkland: that doesn't solve the race with ifup though
[14:42] <Keybuk> if you're ifup takes a long time, your boot can take an extraordinary long time
[14:42] <Keybuk> to undo the general problem, you could put a spin at the front of mountnfs, and put it back in the ordinary boot sequence
[14:42] <Keybuk> the difficulty will be making sure you only spin if there are nfs mounts
[14:42] <Keybuk> and spin until a network is up, only while ifup is likely to run
[14:43] <Keybuk> you don't want to halt the boot because the network cable was unplugged, for example
[14:43] <keescook> but if an NFS is auto, it _should_ wait until that's up.
[14:45] <geser> zul: Hi, it seems we have currently three xen source packages in the archive: xen-3.2 (main), xen-source (universe), xen-unstable (universe). Are xen-source and xen-unstable of any use in hardy or can they be removed?
[14:45] <zul> xen-unstable should be removed and xen-source should be removed for hardy
[14:46] <Riddell> ArneGoetje: yes but it's not new
[14:47] <ArneGoetje> Riddell: I didn't play around with kde4 yet, I am currently fixing font issues on gnome through fontconfig.
[15:04] <kagou> hi
[15:04] <seb128> hey kagou
[15:06] <kagou> hey seb128
[15:32] <mvo> cjwatson: I added it to the hardy desktop seed now, should I do a upload with it now or is there more stuff pending ?
[15:33] <ogra_cmpc> woah, i wouldnt have guessed the langpacks on the addon make up so much space
[15:33] <ogra_cmpc> (it just dropped to 430M from over 600)
[15:45] <cjwatson> mvo: I don't mind
[15:54] <lool> I wonder whether we have wiki macros for LP; I think I saw some for blueprints for example; how can I list available macros on wuc?  It seems these are files on the actual wiki server
[15:58] <fbond> Hi, what is the difference between -dgb packages in the regular repostories and the -dbgsym packages in the ddeb repos?
[15:59] <cjwatson> lool: you mean https://wiki.ubuntu.com/InterWiki ?
[16:00] <cjwatson> lool: (if you need more, file an RT request)
[16:00] <lool> cjwatson: It looks like what I want; thanks!
[16:00] <cjwatson> fbond: -dbg in the regular archive is constructed deliberately by the source package and may (or may not) have extra debugging facilities compiled into it; for example python*-dbg has extra debugging facilities, more than just debugging symbols
[16:01] <cjwatson> fbond: (sometimes they're just leftovers, though)
[16:01] <cjwatson> fbond: the -dbgsym ddebs are automatically constructed and only contain debugging symbols
[16:01] <fbond> cjwatson: if I'm using qprof to do some profiling, is there any way to make use of the -dbgsym packages to make qprof's output more useful?
[16:02] <fbond> cjwatson: do the -dbgsym packages get used "automatically" everywhere, or do, e.g., gdb need special options to be told to use them?
[16:02] <fbond> cjwatson: as you can probably tell, I'm in just a bit over my head :)
[16:04] <ogra_cmpc> cjwatson, i wonder if we could teach the addon to include the corresponding kde langpacks to the ones included in ubuntu-desktop, for hardy i'll handle them manually but it seems to make sense to automate that in the future
[16:04] <ogra_cmpc> (now that edubuntu-desktop depends on kdeedu again)
[16:06] <cjwatson> fbond: you want pitti for that, not me :-)
[16:07] <cjwatson> ogra_cmpc: ubuntu-desktop doesn't contain any language packs; you mean the Ubuntu ship seed?
[16:07] <pitti> fbond: just add "deb ddebs.ubuntu.com gutsy main restricted universe multiverse" to your apt sources, then you can apt-get insall them (<packagename>-dbgsym)
[16:07] <ogra_cmpc> cjwatson, err, indeed
[16:07] <cjwatson> 'deb http://ddebs.ubuntu.com ...' surely?
[16:07] <pitti> fbond: erm, sorry: "deb http://ddebs.ubuntu.com"...
[16:07] <ogra_cmpc> i just cant effort to ship *all* langpacks as we did before
[16:07] <ogra_cmpc> err
[16:07] <ogra_cmpc> afford
[16:08] <cjwatson> ogra_cmpc: hmm. In theory we might be able to add a germinate feature to do that but it would take some pondering
[16:09] <cjwatson> ogra_cmpc: feel free to file a wishlist bug on germinate for it
[16:09] <ogra_cmpc> yeah, as i said, nothing for hardy
[16:09] <fbond> pitti: right, I know how to get them.  I'm wondering if I have to do anything special to make gdb, qprof use them?
[16:09] <pitti> fbond: gdb finds them automatically; I don't know about qprof
[16:10] <fbond> pitti: did gdb have to be patched to find them thusly?
[16:10] <pitti> fbond: ideally it does; our Ubuntu packages have a 'pointer' to the debug symbols (.gnu.debuglink ELF section)
[16:10] <fbond> Ah.
[16:10] <pitti> fbond: no, gdb has supported that since ages
[16:10] <fbond> Interesting.  I'll give it a shot.
[16:10] <cjwatson> pitti: BTW, I noticed that it didn't seem to be finding them for a GTK application I was debugging
[16:10] <cjwatson> although I was using the regular -dbg packages, not -dbgsym
[16:11] <pitti> cjwatson: hm, the original dh_strip should do the same
[16:12] <cjwatson> ii  libglib2.0-0                                            2.15.5-0ubuntu1                            The GLib library of C routines
[16:12] <cjwatson> ii  libgtk2.0-0                                             2.12.8-1                                   The GTK+ graphical user interface library
[16:12] <cjwatson> ii  libglib2.0-0-dbg                                        2.15.5-0ubuntu1                            The GLib libraries and debugging symbols
[16:12] <cjwatson> ii  libgtk2.0-0-dbg                                         2.12.8-1                                   The GTK+ libraries and debugging symbols
[16:12] <cjwatson> but gdb just pretends like those aren't there in a backtrace through libglib.so
[16:14] <fbond> pitti: thanks, btw
[16:19] <lool> I plan asking for the addition of [Bug ] to the InterWiki list cjwatson mentionned previously
[16:19] <lool> As to allow [Bug 1234] in wiki pages, reports etc.
[16:19] <ubotu> Bug 1234 on http://launchpad.net/bugs/1234 is private
[16:19] <ogra_cmpc> :P
[16:26] <geser> pitti: I'm just looking why ocamlrpcgen is still broken. I found the problem but I don't know whose fault it is. Do you know if a space is allowed between dh_strip -X and the argument?
[16:26] <pitti> geser: it's not documented to be allowed
[16:27] <pitti> geser: and pkg-create-dbgsym assumes that there is no space
[16:27] <pitti> geser: dh_strip might get along with it, though, I haven't checked the implementation
[16:27] <geser> that seems to be the problem, cdbs calls dh_strip with spaces between -X and the argument
[16:28] <geser> dh_strip.pkg-create-dbgsym -plibocamlnet-ocaml-bin -X usr/bin/netplex-admin -X usr/bin/ocamlrpcgen leaves the binary intact
[16:28] <geser> but it breaks if I use dh_strip
[16:28] <geser> but dh_strip works if I remove the space after -X
[16:29] <pitti> geser: ok, I need to fix that in pkg-create-dbgsym then
[16:31] <pitti> geser: I reopened the bug, thank you
[16:34] <cjwatson> lool: I'm not sure the syntax would work quite that way ("Bug:1234" surely?), but sounds good in principle
[16:34] <cjwatson> or possibly [wiki:Bug/1234] if it doesn't like the above syntax
[16:37] <tkamppeter> pitti, \sh, slangasek: foo2zjs FF exception posted, bug 192948
[16:37] <ubotu> Launchpad bug 192948 in foo2zjs "FF Exception for foo2zjs: NEW updated version, supports more printers (like OKI c3300n)" [Medium,Triaged] https://launchpad.net/bugs/192948
[16:38] <tkamppeter> ktk, I am here.
[16:53] <asac> calc: sorry if i missed it, but did you roll ooo xul1.9 update yet?
[17:01] <calc> asac: not yet, i plan to roll it into the OOo 2.4.0 upload i do this week
[17:01] <asac> calc: great
[17:01] <asac> thanks for the info
[17:02] <calc> asac: i was going to put it into 2.3.1 upload but found out it would be a good idea to get 2.4.0 in this week so stopped working on 2.3.1 and updated 2.4.0 as quickly as i could :)
[17:02] <calc> asac: its ready pending OOo 2.4.0 rc2 release
[17:06] <asac> good
[17:15] <cjwatson> calc: any reason to block it on that, out of interest? I meant to ask before ...
[17:15] <cjwatson> calc: after feature freeze it's generally better to get things in early
[17:17] <calc> cjwatson: ok, see privmsg :)
[17:31] <Riddell> siretart: why does xine depend on libxine1-plugins | libxine1-misc-plugins ?  seems to me libxine1-plugins beings in some optional extras that aren't necessary (and it breaks things since it tried to bring bad packages onto the CDs)
[17:37] <calc> i may just not be able to read but where do i set the bug contact for a project in lp?
[17:37] <calc> i see actions on the left but haven't found which one is the proper one
[17:37] <pitti> Mithrandir: does that obex server process run as root or dbusdaemon?
[17:37] <calc> also how do i add a group to a project?
[17:38] <cjwatson> https://bugs.launchpad.net/PROJECTNAME/+bugcontact
[17:39] <cjwatson> possibly only if you're the registrant
[17:39] <cjwatson> you can do https://bugs.launchpad.net/PROJECTNAME/+subscribe
[17:39] <calc> cjwatson: yea i am
[17:39] <cjwatson> which doesn't require magic privileges
[17:40] <cjwatson> what do you mean by a "group" in this context?
[17:42] <calc> cjwatson: er 'team
[17:42] <calc> ' actually
[17:42] <pitti> Mithrandir: ah, root, nevermind; hm
[17:42] <cjwatson> oh, teams behave just the same way as persons for nearly all purposes
[17:43] <cjwatson> except for things like special UI entries for teams you're a member of
[17:43] <calc> cjwatson: ok
[17:43] <cjwatson> you can only subscribe a team to a project if you're the team administrator
[17:43] <calc> cjwatson: yea i already did that for bugs, how do i have the team joined as members of the project?
[17:44] <cjwatson> projects don't have members as such ...
[17:44] <calc> cjwatson: oh ok
[17:44] <cjwatson> they have contacts for various specific purposes
[17:44] <cjwatson> like driver, bug contact, security contact
[17:45] <calc> cjwatson: i tried +security for security contact but it didn't seem to work
[17:45] <calc> ah its +securitycontact
[17:45] <cjwatson> calc: it's linked from the bugs tab
[17:45] <cjwatson> e.g. https://bugs.launchpad.net/openoffice, look at the bit on the left
[17:46] <calc> cjwatson: ah ok
[18:46] <Riddell> infinity: could you update livecd-rootfs on king.buildd, or work out why it isn't updating
[18:46] <infinity> Riddell: It upates once a day, in theory.  when was the last upload?
[18:47] <Riddell> infinity: days ago.  the i386 one has updated fine
[18:47] <infinity> Friday...
[18:47] <infinity> Hrmph.
[18:47] <infinity> Looking.
[18:48] <infinity> buildd@king:~$ sudo chroot build-hardy-live/chroot-hardy/ dpkg -l livecd-rootfs | grep ^i
[18:48] <infinity> ii  livecd-rootfs              0.53                        construction script for the livecd rootfs
[18:48] <infinity> Looks like it auto-updated fine...
[18:49] <slangasek> he may need the updated BuildLiveCD script
[18:49] <infinity> Ahh, yes, this is possible. ;)
[18:50] <infinity> -       *ubuntu|*ubuntu-dvd|kubuntu-kde4|ubuntu-lpia|base|tocd) ;;
[18:50] <infinity> +       *ubuntu|ubuntu-lpia|base|tocd)  ;;
[18:50] <infinity> Picky, picky.
[18:51] <infinity> Riddell: Should be happier now.
[18:52]  * infinity needs to move the valid project test out of BuildLiveCD...
[18:53] <infinity> Though, adding new project shouldn't happen often.
[18:53] <Riddell> thanks infinity
[18:53] <Chipzz> infinity: but it would have the advantage of being easier to configure if you want to make your own live cd ;)
[18:53] <Chipzz> infinity: so ++ for derivatives ;)
[18:54] <infinity> Chipzz: Well, people home-brewing don't need to use BuildLiveCD at all, unless they're really keen on publishing things the way I do, etc.
[18:54] <infinity> Chipzz: There's a reason it's packaged as an "example". :)
[18:56] <infinity> Still, in one of these mystical "free time" periods, livecd.sh and BuildLiveCD both could use some "No one's really done anything to them other than minor changes for years" love.
[18:56] <Chipzz> :)
[18:57] <infinity> And they're the worst kind of example of what happens when LaMont writes something, I take it over and half rewrite it, and then some other poor schmucks (yay, distro) get stuck trying to make sense of it.
[19:01] <Riddell> evand: can I upload another ubiquity with a kde4 fix?
[19:01] <xivulon> slangasek: did you discuss metalink urls with the webmaster?
[19:02] <evand> Riddell: sure
[19:02] <slangasek> xivulon: yes; the conclusion is that it's not feasible to host it on www.ubuntu.com, so we will need to have two different metalink urls - one for alphas, one for beta/rc/final
[19:04] <xivulon> do you mean, that it is possible to host metalinks on ubuntu.com, but without redirection?
[19:04] <slangasek> xivulon: I'll try to get some final URLs settled on cdimage.u.c and releases.u.c this week so you can start making use of them
[19:04] <xivulon> thanks
[19:04] <xivulon> couple of minor things
[19:05] <slangasek> xivulon: I mean, www.ubuntu.com is not provisioned to scale the way that releases.ubuntu.com is; anything that's going to be hammered at release time needs to go to the latter
[19:05] <xivulon> release.ubuntu.com is perfectly fine, provided the url does not change I do not care
[19:06] <xivulon> can I have daily-live + final?
[19:06] <slangasek> the daily-live will *not* be on releases.ubuntu.com, it'll be on cdimage.ubuntu.com
[19:06] <slangasek> so - two urls to check
[19:06] <infinity> slangasek: Oh, BTW, regarding things like #184210, if you thing it should never build on !x86, please poke me to P-a-s it and close the bug.  Might be more sane in this case, though I'm not sure if it's meant to be portable and fails to be.
[19:06] <xivulon> will be 3: alpha6/beta/final
[19:07] <xivulon> slangasek I assume that having MD5SUMS + MD5SUMS.gpg is ok
[19:07] <xivulon> the ones in use now are: http://www.wubi-installer.org/metalinks/8.04/
[19:07] <xivulon> no partial sum and only 1 mirror in there
[19:08] <xivulon> the signature there is mine for obvious reasons
[19:08] <slangasek> infinity: I think it's meant to be portable and fails to be; so AFAICS it's a bug, but not release-critical by any means
[19:09] <slangasek> xivulon: I'm assuming that we want to put the metalink file in the per-release root directory (http://cdimage.ubuntu.com/releases/hardy, not http://cdimage.ubuntu.com/releases/hardy/alpha-6), so that you don't have to have an index of all the possible milestone names to poll
[19:09] <infinity> slangasek: Yeah, I was pretty slack on the "check the arch, determine the priority" thing, as I had 3 failure logs for most builds, and would just pick one at random to read.
[19:09] <slangasek> so in that case, there would be only two urls; one on cdimage. and one on releases.
[19:10] <xivulon> slangasek that would be nice
[19:10] <slangasek> infinity: right :)
[19:11] <slangasek> xivulon: now, what do you mean about MD5SUMS + MD5SUMS.gpg?  I don't remember this being mentioned as a new requirement related to metalink
[19:11] <xivulon> slangasek: we could already put in there the metalinks for both, they will not have any checksum
[19:11] <xivulon> slangasek: I need the md5 of the metalink files, and a signature of that
[19:11] <slangasek> ok
[19:11] <xivulon> to avoid spoofing attacks
[19:11] <slangasek> please document that in the bug report, if it's not there already
[19:11] <xivulon> will do that
[19:13] <slangasek> given that this is a single file, is there a reason we wouldn't want to provide just a .gpg sig for the metalink files themselves?
[19:13] <xivulon> I do use 1 metalink for each distro
[19:13] <xivulon> see http://www.wubi-installer.org/metalinks/8.04/
[19:13] <ogra_cmpc> xivulon, googling for grup on loopdevices revealed that you seem to have done work on that area ... was there any outcome (i'm trying to do something similar here)
[19:13] <ogra_cmpc> err
[19:13] <ogra_cmpc> s/grup/grub/
[19:13] <ogra_cmpc> :0
[19:13] <slangasek> a single MD5SUMS for all the metalink files for a release would be nice, but impractical because of the way in which each flavor is generated separately
[19:14] <slangasek> oh, there'll be different metalink files for each arch, that's true
[19:14] <xivulon> ogra_cmpc: we do not install grub on loop device if that is what you mean
[19:14] <slangasek> in that case we can do MD5SUMS{,.gpg}, sure; there'll be a separate one for each flavor, all the same
[19:14] <xivulon> since the boot folder is a windows folder bindmounted into the loop device
[19:14] <ogra_cmpc> xivulon, well, i only saw you worked on fixing it
[19:15] <ogra_cmpc> ah, thats quite different from my setup, thanks
[19:15] <xivulon> ogra_cmpc: I have added 1) detection of host device when menu.lst is generated 2) making sure that device.map is there
[19:15] <xivulon> what is your requirement exactly?
[19:16] <ogra_cmpc> loopmounted device with loopmounted partitions
[19:16] <xivulon> in what context do you boot them?
[19:16] <ogra_cmpc> i.e. /dev/loop0 is the image, /dev/loop1 the first partition (mounted with offset)
[19:17] <xivulon> or better what is your boot setup?
[19:17] <ogra_cmpc> i dont boot them, i want a dd-able image you can just dump on a usb stick
[19:17] <ogra_cmpc> without forcing the user to manually install grub during that process
[19:17] <xivulon> ah I see
[19:18] <ogra_cmpc> i have something half way working with a loopback nbd device but not quite ...
[19:18] <xivulon> what about using grub4dos in mbr?
[19:18] <xivulon> that can use --find-root
[19:19] <xivulon> which allows you to get the right device without coding it into menu.lst
[19:19] <ogra_cmpc> i'll look into it, thanks ... i was just googling for grub and loopback on the weekend and the results have your name scattered all over
[19:19] <ogra_cmpc> so i thought i'd ask
[19:20] <xivulon> yes they kept me busy...
[19:20] <xivulon> with wubi stuff
[19:20] <xivulon> by the way, wubi images can also be useed on USB
[19:20] <ogra_cmpc> developer life was so much easier with dumb old lilo .... sigh ...
[19:21] <xivulon> ahah
[19:22] <xivulon> I could add wubi support to target USB if you wish, that is the same as a regular wubi installation + dropping wubildr.mbr in the mbr
[19:22] <ogra_cmpc> nope
[19:22] <xivulon> one less for me then :P
[19:23] <ogra_cmpc> what i do here is quite specific :) wubi wouldnt help here
[19:24] <ogra_cmpc> but thanks for the offer :)
[19:24] <xivulon> slangasek: if we have a signature for each metalink, I see no much point in having also the MD5
[19:25] <xivulon> I mean no point in having md5+signature for each metalink. metalink + signature is enough
[19:26] <slangasek> xivulon: there will be a separate metalink for each architecture, so it makes sense to only sign once for each flavor; in which case there's no reason for us to break your existing code
[19:26] <xivulon> sure
[19:57]  * slangasek stares at the python2.5 testsuite.  Did it just play a Monty Python clip at me?
[20:03] <Keybuk> slangasek: entirely possible
[20:04] <slangasek> test_normalization and test_ossaudiodev, apparently
[20:17] <\sh> guys, how do I debug an apache2 which is segfaulting? (i guess it's php5)
[20:18] <slangasek> run it under gdb in debug/foreground mode?
[20:19] <\sh> slangasek: good idea.../me is braindead at this time of the day
[20:21] <Keybuk> The following NEW packages will be installed
[20:21] <Keybuk>   libpam-doc
[20:21] <Keybuk> ...cover me, I'm going in!
[20:22] <\sh> hmmm
[20:22] <\sh> it's gutsy...and there was an pcre upgrade...
[20:22] <\sh> the segfault is coming and gdb tells me : 0xb7f517ac in ?? () from /usr/lib/libpcre.so.3
[20:23] <\sh> trying a bt full gives me
[20:23] <\sh> #0  0xb7f517ac in ?? () from /usr/lib/libpcre.so.3
[20:23] <\sh> Cannot access memory at address 0xbf121c20
[20:24]  * LaserJock throws Keybuk a few grenades ... just in case
[20:25] <Keybuk> LaserJock: you're not supposed to throw them *at* me
[20:25]  * Keybuk recovers his limbs
[20:25] <LaserJock> hmm
[20:25] <\sh> slangasek: any idea how to tell gdb to be more noisy? :)
[20:25] <Keybuk> \sh: install the debug sym packages?
[20:26] <Keybuk> deb http://ddebs.ubuntu.com gutsy main universe
[20:26] <Keybuk> deb http://ddebs.ubuntu.com gutsy-security main universe
[20:26] <Keybuk> should be an apache-dbgsym
[20:26] <Keybuk> and maybe a libpcre3-dbgsym there
[20:27] <\sh> Keybuk: thx
[20:27] <Keybuk> (there is, I just checked)
[20:28] <\sh> Keybuk: ah well...unmet deps ;)
[20:29] <\sh> Keybuk: the dbgsyms are not automatically rebuild when a -security or -update package is uploaded?
[20:29] <Keybuk> \sh: not sure, I think they are
[20:30] <Keybuk> that being said, gutsy-security is kinda empty
[20:30] <Keybuk> pitti: ^
[20:30] <\sh> Keybuk: and now there were again my problems ;)
[20:32] <\sh> and this damn system must be running tomorrow morning at 8am UTC
[20:33] <Keybuk> \sh: does it work if you unapply the security update?
[20:34] <\sh> Keybuk: well..I'll have to revert the all the apache2+module deps etc...
[20:34] <Keybuk> \sh: you will?
[20:35] <\sh> Keybuk: I have to actually...if I can't find the problem until tomorrow morning, I'm fcked
[20:35] <Keybuk> have you tried just reverting the security update though
[20:35] <Keybuk> to see whether that fixes the problem?
[20:35] <Keybuk> that would tell us whether we have an issue
[20:36] <\sh> Keybuk: you mean of the pcre fix?
[20:36] <Keybuk> yes
[20:36] <\sh> give me a min :)
[20:38] <\sh> Keybuk: it would be easy if libpcre3-0ubuntu0.7.10 would be still in the archives
[20:38] <\sh> Keybuk: which means, the last update was a -security fix, and the version before it, was also a security fix,
[20:39] <slangasek> hrm? where is that version number from?
[20:39] <Keybuk> \sh: https://edge.launchpad.net/ubuntu/gutsy/+source/pcre3/7.4-0ubuntu0.7.10.1
[20:39] <Keybuk> you can download from there
[20:39] <Keybuk> (minus the edge)
[20:40] <Keybuk> pick the appropriate architecture on the left to get to the resulting binary packages
[20:42] <\sh> Keybuk: the same...it's not pcre it seems
[20:43] <lamont> infinity: livecd.sh wasn't supposed to live this long...
[20:43] <evand> ogra_cmpc: Will there continue to be full edubuntu desktop CDs going forward, or will that eventually be disabled on cdimage?
[20:44] <infinity> lamont: Hey, it still works... More or less. :)
[20:45] <\sh> Keybuk: and actually the dbgsym packages of apache2 are out of date
[20:45] <slangasek> my understanding was that edubuntu desktop would remain a discrete image, only server was being rearranged into an ubuntu-server adjunct?
[20:47]  * \sh upgrades now to hardy...and if this is not working I'm reinstalling the gutsy image..I'm tired
[20:47] <evand> hrm, ok.  Thanks slangasek
[20:48]  * \sh 's wife will kill him ... since 3 weeks in the company and already a nightshift
[20:48] <ScottK2> Shouldn't have bought the company life insurance then.  You might be safer.
[20:49] <\sh> ScottK2: *g*
[20:49] <\sh> we had a similar crash under debian etch..I solved it with replaceing debian with ubuntu..and now we added some stuff to the php backend which uses ZF...and bang again a segfault
[20:50] <slangasek> evand: that's my understanding; plans may have changed when I wasn't looking
[20:50] <\sh> (debian <- old php release)
[20:50] <evand> slangasek: indeed, noted
[20:52] <\sh> ScottK2: but actually a good news: zend is working with me on making ZendFramework in ubuntu rock...direct upstream contact established
[20:54] <LaserJock> slangasek: I'm not sure that a Desktop CD will be possible for Edubuntu, because of size issues
[20:54] <LaserJock> unless we want to do some funky stuff like not putting edubuntu-desktop on the Edubuntu Desktop CD :-)
[20:54] <slangasek> LaserJock: it's currently oversized; but how is it better to not have a desktop CD than to trim the metapackage?
[20:55] <LaserJock> because the Desktop CD is stupid for Edubuntu anyway ;-)
[20:56] <LaserJock> it's kind of a bad thing, IMO, to have a Desktop CD that is radically different from what you get when you install Edubuntu the "recommended" way
[20:56] <stgraber> slangasek: having an edubuntu desktop CD which is basically an Ubuntu with a different artwork is useless from my point of view
[20:56] <stgraber> you would need to add the content of the add-on CD to the desktop CD to have a real educational desktop
[20:56] <LaserJock> exactly
[20:57] <LaserJock> it wouldn't have very many of the apps actually available
[20:57] <stgraber> you will need both edubuntu desktop and add-on CD, then why not have the ubuntu desktop and the edubuntu add-on CD ?
[20:57] <stgraber> LaserJock: what do we have more in Edubuntu desktop than Ubuntu desktop ? gobby and the artwork ?
[20:58] <LaserJock> ogra added most of the edu apps
[20:58] <slangasek> stgraber: does the add-on CD include all of the edubuntu desktop packages?  I don't know that it currently does
[20:58] <stgraber> slangasek: it does, it was the test I did for Alpha-5
[20:58] <LaserJock> it now contains most of the stuff that's on the add-on CD
[20:58] <stgraber> slangasek: I started from an Ubuntu Alternate and then installed the whole Edubuntu desktop from the Edubuntu add-on CD
[20:58] <LaserJock> i.e. the way it was before we split the CDs
[20:59] <slangasek> stgraber: oh, it includes the *desktop* add-on packages?  hrm, ok; that means the edubuntu server went untested, I guess :)
[20:59] <stgraber> slangasek: edubuntu server won't exist for Hardy
[20:59] <slangasek> well, not my decision anyway how edubuntu CDs should be structured
[20:59] <slangasek> stgraber: er, I thought that was also part of what was on the edubuntu add-on
[21:00] <\sh> bah....apache2 upgrade path is broken
[21:00] <stgraber> LTSP was moved into the Ubuntu Alternate
[21:00] <slangasek> hmm, ok
[21:00] <stgraber> so I can now install LTSP from Ubuntu Alternate (also one of the tests I did for Alpha-5)
[21:00] <stgraber> and then install the add-on CD to have the same as the previous Edubuntu server
[21:08] <\sh> keescook: ping pcre3 security fix
[21:09] <\sh> keescook: looks like that it breaks here (on hardy now)
[21:09] <\sh> keescook: same problem as on gutsy...
[21:09] <ScottK2> \sh: Sounds promising (the zend bits, not the pcre).
[21:11] <\sh> ScottK2: well, the pcre bit is giving me a headache..
[21:12] <\sh> but I wonder how can I resolve this: 	rrc = Cannot access memory at address 0xbf69ad18
[21:13] <slangasek> that typically points to stack corruption
[21:13] <slangasek> try valgrind?
[21:13] <slangasek> (which may or may not be useful, given that php does its own memory management)
[21:14] <\sh> slangasek: well...if you hint me on running apache2 + php5 inside valgrind? or just valgrind /usr/sbin/apache2 -X
[21:14] <\sh> ?
[21:14] <slangasek> exactly
[21:16] <\sh> slangasek: trying
[21:25] <\sh> slangasek: could it be that, while running in valgrind, that php5 is not reading it's ini?
[21:27] <slangasek> \sh: valgrind should be invisible to php.
[21:27] <\sh> slangasek: I got the error now
[21:28] <\sh> slangasek: http://paste.ubuntu.com/4983/
[21:28] <shaya> anyone have any idea how to get home button in navigation toolbar in firefox 3? Ubuntu's version doesn't drag for me in "customize" mode
[21:28] <\sh> slangasek: your guess was correct it seems
[21:29] <slangasek> \sh: heh, nice
[21:30] <slangasek> that's a very, very strange use of memcpy()...
[21:30] <\sh> slangasek: as it is the very same secfix in hardy and gutsy I think we have a problem
[21:31] <\sh> hoping that kees is still awake ;)
[21:31] <slangasek> well, the first error returned by valgrind is not caused by pcre
[21:32] <\sh> slangasek: the message of Access not within mapped region at address 0xBE094FFC
[21:32] <\sh> is raised in pcre_exec.c, or do I read something wrong?
[21:33] <slangasek> no, you've read that right; but I'm not sure what that signifies
[21:33] <slangasek> because before that, it's giving a stack overflow
[21:35] <el_cubano> Hello.  Does anyone know if (or where) the kernel team maintains their work in a VCS of some sort?
[21:35] <jdong> el_cubano: kernel.ubuntu.com
[21:35] <el_cubano> jdong: Wow!  That was easy?
[21:35] <jdong> :)
[21:36] <el_cubano> Next question, is anyone aware of how easy/hard it would be to remove the AppArmor patch from the kernel?  Any potential problems with that?
[21:36] <\sh> slangasek: hmm...what could give us the information we need? I installed apache2*, php5 and libpcre3 dbgsym packages
[21:36] <jdong> el_cubano: why do that?
[21:36] <jdong> would blacklisting the module not suffice?
[21:37] <slangasek> \sh: probably some google searching on that error message
[21:37] <el_cubano> jdong: Because AppArmor makes it impossible to build Lustre modules
[21:37] <el_cubano> jdong: Using kernel 2.6.22 from Gutsy
[21:37] <slangasek> \sh: have you confirmed that downgrading prce3 fixes the problem?
[21:37] <jdong> el_cubano: you should be able to just stop apparmor's init script and unload the apparmor module
[21:37] <jdong> and in that case, it'd be pretty much identical to an un-apparmored system
[21:37] <el_cubano> jdong: That does not help if the compilation of the lustre modules never succeeds
[21:38] <el_cubano> AppArmor changes the vfs_* calls around.
[21:38] <jdong> el_cubano: ah, well I've never tried reverting the apparmor patches, I wouldn't think it'd be that hard to do
[21:38] <\sh> slangasek: I'm downgrading now...
[21:38] <el_cubano> jdong: OK.  Thanks for the pointer.
[21:39] <\sh> slangasek: strangly using the zend-core binaries of apache2+php5 doesn't raise this...
[21:40] <siretart> Riddell: because it would break dapper->hardy upgrades. in dapper all plugins were included in the libxine1 package. if that dependency was dropped, those plugins would silently vanish from the user system
[21:40] <siretart> Riddell: vorlon explained me that was unacceptable for debian, and I don't see why it therefore should be for ubuntu
[21:41] <\sh> slangasek: downgraded to libpcre3 7.4-1ubuntu1 now
[21:43] <geser> shaya: wait for firefox 3 beta 4 (see bug 192505)
[21:43] <ubotu> Launchpad bug 192505 in firefox-3.0 "Where's my home button?" [Low,Fix committed] https://launchpad.net/bugs/192505
[21:44] <shaya> danke
[21:45] <\sh> slangasek: ok...the same happens with the old version...
[21:48] <slangasek> \sh: right.  you say you're using some Zend extension on this system?  Is that extension relevant to reproducing the bug? (i.e., can you put together a test case that doesn't depend on the extension?)
[21:49] <\sh> slangasek: nope...we are using ZendFramework it's a simple php source lib (so it's interpreted by php5)
[21:49] <\sh> slangasek: and nope I can't setup a testcase...normally a module shouldn't segfault when something is wrong with the source files (regarding php actually)
[21:50] <slangasek> er, of course it /shouldn't/, but the test case is still needed in order to let others try to reproduce the problem
[21:52] <\sh> slangasek: well, if I would knew what part this triggers I would cry out loud and fix it...but there are two possibilities: 1. something goes wrong in detected system locales (which was the cause when using ZendFramework and debian php5) or 2. I don't know...because on the other machine with the zend provided binaries, it works
[21:55] <fbond> pitti: Just FYI, I'm now using oprofile, and it makes use of debug symbols nicely.  It's a great package!
[22:20] <soren> Isn't gdb supposed to automagically pick up the -dbgsym stuff if I have those installed.
[22:24] <Riddell> seb128: your last change to libgpod broke libgpod3-nogtk, it can't be installed
[22:28] <\sh> slangasek: just FYI..with the zend-core binaries on this ubuntu... it segfaults to
[22:28] <\sh> too
[22:36] <seb128> Riddell: I didn't change this one in a while now
[22:37] <seb128> Riddell: and why installing the common would break the nogtk version?
[22:37] <Riddell> seb128: and it's been broken for a while
[22:37] <Riddell> seb128: because it only depends on the gtk version
[22:38] <seb128> ?
[22:38] <Riddell> really it shouldn't depend on any of course, they should depend on it
[22:38] <seb128> no
[22:39] <seb128> the depends is right
[22:39] <seb128> the hal helper which is used to make new ipods work is using libgpod
[22:39] <seb128> and they install it there
[22:39] <\sh> slangasek: I wonder if it's really this locale thing...that this ZendFramework doesn't know anything about utf-8 locales
[22:39] <seb128> but right it should have a | libgpod3-nogtk there
[22:40] <seb128> seems that nobody is using nogtk though
[22:40] <seb128> the change has been ages ago
[22:42] <Riddell> seb128: amarok does
[22:52] <slangasek> \sh: there's really not much more I can say about this without a reduced test case I can use to reproduce the problem
[22:54] <\sh> slangasek: well, I'm stopping now, because I think the bug is inside the application and/or in ZendFramework. I can reproduce the crash with two independent installations
[22:55] <\sh> slangasek: I'm writing an email to the ZendFramework guys with providing all necessary infos ... well, anyways I'm fcked tomorrow.
[22:56] <slangasek> a segfault in PHP is a bug in the engine, no matter how it's triggered; but it still needs a reduced test case to be debuggable
[22:58] <\sh> slangasek: both are using 5.2.4 so this is the common thing
[22:58] <\sh> slangasek: and I'll talk tomorrow to our php devs..they should reproduce something
[22:59] <komputes> whats the best way to install Firefox 2 on Hardy?
[23:00] <infinity> komputes: apt-get install firefox-2
[23:00] <komputes> infinity: from what source repository?
[23:00] <\sh> slangasek: and the bad thing is...the version which works is php 5.2.1
[23:01] <komputes> infinity: firefox-2 is not a package in any of the hardy repos from what I can see
[23:02] <infinity> komputes: Err, it's probably in binary NEW currently.
[23:02]  * infinity goes to see about that.
[23:02] <komputes> infinity: can you let me know which repository that is so I can add it to my sources.list
[23:03] <infinity> komputes: It's not a repo at all, it's a "waiting on archive admin intervention".
[23:04]  * komputes is confused
[23:04] <geser> komputes: it will be in universe in a few days
[23:05] <infinity> komputes: New binaries (new names) need manual intervention.  I've just intervened, so "firefox-2" will exist in an hour or two, after the publisher gets to them (in universe)
[23:05] <komputes> infinity: thx
[23:10] <komputes> infinity: is there a .deb of firefox 2 somewhere?
[23:10] <komputes> infinity: can I add a gutsy repo and download it from there?
[23:10] <infinity> komputes: You could just wait an hour or two, surely?
[23:11] <komputes> infinity: actually, i kind of wanna go home
[23:11] <infinity> komputes: Alternately, you can download the binaries before they hit the archive from: https://edge.launchpad.net/ubuntu/+source/firefox/2.0.0.12+2nobinonly+2-0ubuntu3/+build/524435
[23:11] <infinity> komputes: (Assuming you use i386, follow links up and to another arch, if you need another)
[23:12] <komputes> infinity: the .changes file?
[23:13] <infinity> komputes: The links to the binaries on the left...?
[23:13] <infinity> (Which are 404, go soyuz)
[23:14] <komputes> haha
[23:14] <infinity> https://edge.launchpad.net/ubuntu/hardy/+queue?queue_state=2&queue_text=&start=20
[23:14] <infinity> Expand "firefox (your arch)"
[23:15] <infinity> Those links work.
[23:15] <komputes> i386
[23:16] <komputes> infinity: URL please :o
[23:17] <komputes> my bad
[23:21] <komputes> infinity: thanks two fold