[12:27] <jcole> $ dmesg | grep drm
[12:27] <jcole> [4294686.182000]  [drm]  Initialized drm 1.0.0 20040925
[12:27] <jcole> isn't that pretty old?
[12:28] <Surak> jcole: more than ubuntu :-D
[12:28] <jcole> this guy has a newer version -->  http://ubuntuforums.org/showthread.php?t=134069&highlight=20060119
[12:29] <jcole> maybe that's why there are so many xgl problems with i810
[12:30] <jcole> on ubuntu... my colleague has suse running fine with xgl... ubuntu wouldn't work
[12:30] <jcole> i've got: [4294686.185000]  [drm]  Initialized i915 1.1.0 20040405 on minor 0:
[12:31] <mdz> mjg59: around?
[12:32] <mdz> mjg59: is it at all expected that sleep/hibernate via  the Fn key combinations is no longer working on my T42, and what can I do to remedy it?
[12:34] <theCore> why Gnome create an icon for NTFS drives, even when the user doesn't have the permission to access it?
[12:34] <mdke> theCore, it's a known bug, and fixed, I think
[12:35] <theCore> mdke: oh, well 
[12:35] <theCore> mdke: thanks for the info
[12:36] <mdke> I think it's https://launchpad.net/distros/ubuntu/+source/partman-basicfilesystems/+bug/25071
[12:36] <Ubugtu> Malone bug 25071 in partman-basicfilesystems "More reasonable defaults for ntfs mounts during installation" [Major,Fix released]  
[12:37] <mjg59> mdz: It ought to work
[12:37] <mjg59> mdz: What's the sleep button action in gnome-power-preferences?
[12:40] <mdz> mjg59: hibernate
[12:41] <mjg59> mdz: Hm.
[12:42] <mjg59> mdz: If you run xev and focus it, do fn+f4 and fn+f12 generate keycodes?
[12:43] <lifeless> mdz: hi
[12:46] <jcole> do you think drm being so old in ubuntu might be the cause of alot of these bugs? -> https://launchpad.net/distros/ubuntu/+bugs?field.searchtext=direct+rendering
[12:46] <jcole> my /lib/modules/2.6.15-20-686/kernel/drivers/char/drm/i915.ko kernel module is dated 20040405
[12:48] <jcole> 2.6.15-20 is the latest kernel i can find in synaptic
[12:53] <Surak> hum. is gnome's bugzilla down?
[12:54] <HiddenWolf> yes
[12:54] <HiddenWolf> too many connections error. :)
[12:56] <mdz> mjg59: need to get the laptop back in hand later today...should it generate keycodes or not?  I thought those were acpi events
[12:56] <mdz> lifeless: hi?
[12:59] <KaiL_> that was only 1, not 20 ;)
[01:00] <jcole> lol, that was a bad idea
[01:01] <KaiL_> "suse" is mostly something, which has to do with bad ideas ;)
[01:01] <Kamion> Surak: yes, I knew about the localechooser-data thing but wasn't terribly worried; after all for most people it's a live CD and they won't bother upgrading it
[01:02] <Surak> Kamion: so I shall close the bug.
[01:02] <Kamion> also I'm more likely to be unhappy at people for closing bugs than for opening them
[01:02] <Kamion> Surak: no, don't
[01:02] <Kamion> please leave closing espresso bugs to espresso developers :)
[01:02] <Surak> there are two or three regressions from 0.99.36 to 0.99.43.
[01:03] <Surak> Kamion: ok. Even fixed bugs? Sorry. I closed one of them. Just as triage, to make your time better used on coding.
[01:03] <Kamion> please don't close bugs on the basis that it now works for you if you aren't the original submitter
[01:04] <Kamion> that actually uses up more of my time, because I have to go back and check
[01:04] <Surak> sorry.
[01:05] <mjg59> mdz: Ot's all done via keycodes now
[01:05] <Surak> Kamion: did you remove the resize partition option in 0.99.43? It was there friday.
[01:05] <Kamion> Surak: no, I did not
[01:05] <Surak> Kamion: forget it. I zeroed the hard drive to test some GRUB bugs.
[01:05] <Kamion> Surak: you almost certainly no longer have a partition table that partman can resize and insert more partitions into; this often happens because you have too many primary partitioner
[01:05] <Kamion> partitions
[01:05] <Kamion> ... or that would do it too
[01:07] <lifeless> mdz: I heard there was a distro sprint coming up soon
[01:07] <Surak> specifically, bug 35614
[01:07] <lifeless> mdz: I'd like to come along to that to get a stronger sense of the QA challenges you guys face
[01:07] <Ubugtu> Malone bug 35614 in espresso "grub fails" [Normal,Confirmed]  http://launchpad.net/bugs/35614
[01:08] <Kamion> I wish people would open new bugs for that
[01:08] <Kamion> grub can fail for all sorts of reasons
[01:08] <Surak> it works here. I'm doing some testing to make it break so I can help you, but it insists in working!
[01:09] <Kamion> with Flight CD 6, it will only break if there's a candidate version of grub in the archive newer than what you have installed
[01:09] <Kamion> I believe the bug to be fixed in more recent versions of espresso, and so I've just closed the bug
[01:10] <Kamion> (thought I'd already done so, thanks for the reminder)
[01:11] <Surak> Oh, you just closed it. It was still as open here, as I probably opened the window about 2h ago :-)
[01:12] <Surak> Kamion: bug 37199 was fixed too. 
[01:12] <Ubugtu> Malone bug 37199 in espresso "Espresso does not check wheter the entered passwords are the same" [Normal,Needs info]  http://launchpad.net/bugs/37199
[01:12] <KaiL_> jcole, do you know, if there are known deadlocks with i915?
[01:12] <mjg59> mdz: Oh, BTW - I'll be travelling during tech board tomorrow
[01:13] <HiddenWolf> lmao, if you order packages by number of bugs, 81 is less than 9. :)
[01:13] <HiddenWolf> someone confirm I'm not crazy? 
[01:14] <Burgwork> HiddenWolf, linky?
[01:14] <Kamion> Surak: thing is, I could never reproduce that problem in the first place
[01:14] <HiddenWolf> https://launchpad.net/people/seb128/+packages is the page I was checking
[01:14] <KaiL_> jup, same here
[01:14] <Kamion> Surak: actually, no, ignore that, see my comment in the bug
[01:14] <Burgwork> HiddenWolf, the sorting it is doing is by alphabetical, rather than numbeical
[01:14] <Kamion> Surak: the bug should remain open
[01:14] <KaiL_> doesn't say, that you aren't crazy, but that bug really exists :p 
[01:15] <lifeless> I've put my name up to get universe upload rights in the next tb meeting. Its at 5am here - how much of a requirement is to be at the meeting ?
[01:15] <HiddenWolf> Burgwork: makes the sorting for that column pretty useless
[01:15] <Surak> Kamion: because of the last comment? About being empty instead of different?
[01:15] <Kamion> Surak: yes
[01:16] <Kamion> Surak: thanks for triage though, it helps remind me of things sometimes
[01:17] <Burgwork> HiddenWolf, indeed
[01:18] <Surak> Kamion: as I am not as good as coder as you, but I can still help anyway. Specially in this case, as I know python a little and I did that awful shell script which does the same thing as espresso... 
[01:19] <Kamion> generally speaking, I'd rather not have people looking around for bugs to close (as mentioned, that mostly just creates more work, because in my experience at best 50% of those closures are valid), but anyone who can dig into a bug a bit and clarify what's going on can make themselves very useful
[01:20] <Kamion> what my bug list needs is clarification more than anything else, particularly since many installer bug reports are pretty unclear
[01:20] <Surak> Kamion: the translation is being done through rosetta's debian-installer item, isn't it? I noticed that the "next" "previous" and "cancel" button are never translated, and there are not such items there.
[01:21] <jcole> KaiL_: the ubuntu drivers are just really outdated 
[01:21] <Kamion> Surak: yes; next/previous/cancel are untranslated at the moment, that's a bug, although hopefully fairly easy to fix
[01:21] <jcole> KaiL_: go here to get the latest videoa driver for your card -> http://dri.freedesktop.org/snapshots/
[01:22] <Kamion> "fix up various untranslatable strings" is on http://wiki.ubuntu.com/UbuntuExpress/ToDo
[01:22] <KaiL_> well, if you only go by the date of the file, something seams to be wrong on your system ;)
[01:22] <Surak> I would never look at it there. :-)
[01:22] <KaiL_> here the file is dated 2006-04-05
[01:22] <jcole> KaiL_: $ dmesg | grep drm
[01:22] <jcole> KaiL_: [4294686.182000]  [drm]  Initialized drm 1.0.0 20040925
[01:23] <Surak> Night Kamion: same thing here. Better go home and leave this for tomorrow. Thanks for your attention.
[01:23] <Surak> duh- the first line in wiki is a bug i just opened.
[01:24] <KaiL_> jcole, interesting - I wonder, from where this version comes..
[01:24] <jcole> KaiL_:  [4302209.541000]  [drm]  Initialized i915 1.4.0 20060119 on minor 0: <-- notice this line http://ubuntuforums.org/showthread.php?t=134069&highlight=20060119
[01:25] <KaiL_> really interesting
[01:25] <jcole> KaiL_: it's the date the debian/ubuntu kernel maintainter(s) downloaded the driver which was a few years ago... i suspect that all drm/dri drivers in ubuntu are outdated
[01:26] <KaiL_> maybe only the version number didn't get changed?
[01:26] <jcole> KaiL_: suse and fedora are up to date
[01:26] <jcole> KaiL_: no, i'm missing extensions that should be on my card
[01:26] <KaiL_> I can't belive, that all these drivers are 2 years old...
[01:27] <jcole> KaiL_: i suspect this update will make my 3d card work the same on these other distros
[01:28] <jcole> KaiL_: probably because they are not part of the main kernel source tree?
[01:28] <KaiL_> aren't they?
[01:28] <jcole> KaiL_: no
[01:29] <KaiL_> I remembered, that the kernel has DRM modules in it's main source tree....
[01:35] <jcole> KaiL_: where?
[01:36] <jcole> all right... here goes
[01:36] <jcole> module compile done...
[01:36] <KaiL_> let's see.. (didn't look into the source recently)
[01:37] <Surak> night all.
[01:48] <mdz> lifeless: where did you hear that?
[01:49] <lifeless> mdz: some comment during the london sprint. May have misheard/misinterpreted.
[01:49] <mdz> lifeless: happy for you to join us, but we won't get together again until after the release
[01:49] <lifeless> mdz: thats cool. Wanted to raise it as soon as possible in case it was near
[01:50] <lifeless> thanks
[01:52] <csmall> I have a question that possibly only developers can answer. Regarding performance, comparing RPM based distributions to debian based distributions. Why do RPM based dist seem to run smoother, faster, softer (I know this sounds funny) while debian based dists seem to ...I dunno...hit the disk harder I guess...not the soft smooth feeling of RPM based dists
[01:52] <csmall> I just could never figure it out
[01:53] <Amaranth> uh
[01:53] <Amaranth> how can writing new files to the disk not "hit the disk"?
[01:53] <Kamion> RPM vs. .deb is just a delivery mechanism for files, so that cannot have anything to do with how the system runs
[01:53] <csmall> thats all fine and I understand what you mean
[01:54] <csmall> I am just trying to explain it in the best way I can
[01:54] <Kamion> the difference must be elsewhere
[01:54] <csmall> yes
[01:54] <csmall> but there is a noticable difference
[01:54] <Kyral> in a way, RPM and DEB are just frontends to Tar.gz :D
[01:54] <csmall> yeah they are just compressed files
[01:54] <Kyral> OW!!!
[01:54] <Kyral> what?
[01:54] <Kamion> presumably the kernel shipped by e.g. Red Hat happens to be better for your hardware
[01:54] <Amaranth> you oversimplify things :)
[01:54] <Surak> csmall: try a "apt-get dist-upgrade" in fedora and ubuntu. Then tell me how many minutes more you have to wait in fedora for it just show you the list.
[01:54] <Kyral> I said "in a way"
[01:54] <Kyral> jeez :P
[01:55] <csmall> I am not really comparing the package managment
[01:55] <csmall> Surak: agreed
[01:55] <csmall> but I mean an overall feel
[01:55] <Amaranth> csmall: perhaps you have dma enabled on the other distro and not on ubuntu?
[01:55] <Surak> csmall: which means rpm can be, in fact, much slower on this particular case.
[01:55] <csmall> Amaranth: thats what I was wondering I just installed dapper Iwill chekc if its on
[01:55] <csmall> Surak: I mean an overall feel for the entire desktop experience
[01:56] <csmall> Amaranth: I will check. I have been using debian religiously for years and I have always wondered why there is this difference in the feel
[01:56] <lifeless> csmall: I dont think there is any relationship betwen the packaging and the feel
[01:56] <HiddenWolf> ubuntu and debian are the only ones building for 386?
[01:56] <csmall> Just thought maybe you guys could offer some insight
[01:56] <Amaranth> HiddenWolf: probably
[01:57] <csmall> lifeless: I agree, I only mentioned RPM to descibe the distributions I am refering to
[01:57] <HiddenWolf> Amaranth: would that _really_ be noticable?
[01:57] <lifeless> csmall: its misleading though, right.
[01:57] <Amaranth> HiddenWolf: doubtful
[01:58] <Amaranth> HiddenWolf: iirc it's march=386 but mtune=p4 so i think it's just a matter of not using a couple of newer instructions
[01:58] <jdub> 486
[01:58] <Amaranth> jdub: oh, it's 486 now?
[01:59] <Amaranth> can't really go any higher because of the via c3
[01:59] <jdub> it always was iirc
[01:59] <KaiL_> VIA should really repair their CPU crap..
[01:59] <Amaranth> i'm out of my league here
[02:00] <HiddenWolf> *chuckle* Well, whatever you do, current performance will always become the baseline. My dapper feels normal or slow, untill I boot up breezy. :)
[02:00] <Amaranth> KaiL_: They don't do anything wrong, they just don't implement one instruction (cmov?) that others do
[02:00] <Amaranth> HiddenWolf: hehe, yeah
[02:00] <HiddenWolf> Just now, eog took more than a blink to show me a photo, and I was 'damn, slow' 
[02:00] <KaiL_> Amaranth, bad enough - afaik except this single instruction, it's an i686, so it's only an i486 with many many additional instructions, which are mostly not used
[02:01] <HiddenWolf> Must just have been a microsecond or 2. :)
[02:01] <csmall> dma is on, so thats not it. maybeit will always be a mystery heh
[02:01] <Amaranth> csmall: it's harder on the system but from what i've heard it's also faster, maybe that has something to do with it
[02:02] <Amaranth> although that might be apt vs yum
[02:20] <Kamion> KaiL_: actually no, it really is an i686, it's just that every other i686 happens to implement i686+cmov
[02:21] <Kamion> oh god, changing gtk's language under its feet is incredibly tedious
[02:21] <Kamion> it caches the translations of all stock items
[02:56] <lifeless> Kamion: how can I tell if a motherboard has SATA without plugging it into a monitor and rebooting ?
[02:57] <lifeless> Kamion: is there some regular device/driver that hints this ?
[02:58] <Kamion> lifeless: lspci, look for SATA
[02:58] <Kamion> or possibly Serial ATA, or similar
[02:59] <lifeless> thanks
[02:59] <lifeless> pity though, looks like my server box is sataless
[02:59] <lifeless> :|
[03:02] <HiddenWolf> lifeless: just get the model of moderboard and google for a product/review page
[03:02] <lifeless> HiddenWolf: if I knew that, I'd just grab my manuals
[03:02] <Lathiat> dmidecode perhaps?
[03:03] <HiddenWolf> lspci grep ATA shows it for me.
[03:03] <lifeless> Lathiat: good call. thanks
[03:03] <Lathiat> dmidecode works for me on most things like that, if nothign else to get the motherboard model
[03:03] <lifeless> problem is I have lost touch with hardware-level skills
[03:03] <lifeless> been focused on software for long enough that I'm out of date
[03:04] <lifeless> Lathiat: yeah, hindsight - bleeding obvious
[04:24] <bddebian> dh_link creates hardlinks?
[04:26] <ajmitch> bddebian: no
[04:28] <bddebian> How do I symlink dirs?  The manpage isn't real clear.  Can I just do dh_link debian/foo/  debian/bar/ ?
[04:30] <bddebian> dh_link
[04:30] <bddebian> dh_link: link destination debian/qgis/usr/share/doc/qgis is a directory
[04:42] <bddebian> ajmitch: So, how can I do a dir?
[04:45] <desrt> is anyone else experiencing mime wonkiness?
[04:48] <bddebian> mime is wonky to begin with :-)
[06:33] <fabbione> morning
[06:34] <bddebian> Hello fabbione
[08:26] <dholbach> good morning
[08:53] <sivang> morning !
[09:00] <pitti> Good morning
[09:01] <ajmitch> morning pitti 
[09:03] <sivang> and dholbach  , ofcourse
[09:04] <tepsipakki> g'morning. what component is responsible for showing the volume when pressing volume buttons on laptops?
[09:04] <dholbach> hi sivang!
[09:05] <pitti> hey sivang
[09:05] <sivang> dholbach: :-)
[09:07] <mdke_> morning dholbach 
[09:07] <mdke_> dholbach, hey, we're thinking of branching the ubuntu-docs repo today sometime, ok with you?
[09:08] <dholbach> mdke_: that's ok with me, sure - sorry for not answering the mail earlier
[09:08] <mdke_> dholbach, no problem
[09:08] <mdke_> dholbach, it might permit us to eliminate some of those unused icons and reduce the source package size :D
[09:10] <dholbach> :-)))
[09:10] <jdub> dholbach: upload u-a crack!
[09:10] <dholbach> jdub: ok!
[09:11] <mdke_> jdub, i will never be able to read something you write in irc again without asking myself whether you are writing from the "love-sac" or not
[09:12] <Burgundavia> mdke: are you scarred for life?
[09:13] <sivang> mdke: what, when did this happen??
[09:14] <sivang> having such consequences, it must have been something very shocking..
[09:14] <mdke> sivang, I'm reluctant to consign you to the same fate, but it happened when i read http://perkypants.org/blog/2006/04/09/tin-roof-rusted/
[09:16] <sivang> oh dear
[10:14] <mvo> Riddell: https://launchpad.net/distros/ubuntu/+source/language-selector/+bug/37065/+index is about the kde interface guidlines. can you give me a hint please :) ?
[10:14] <Ubugtu> Malone bug 37065 in language-selector language-selector-qt "[OK]  not working." [Normal,Confirmed]  
[10:22] <pitti> dholbach: okay for me to upgrade clamav to new upstream release 0.88.1 (from 0.88)? it fixes three security bugs
[10:23] <pitti> dholbach: http://changelogs.debian.net/clamav -> top entry
[10:23] <dholbach> yeah
[10:49] <pitti> hi slomo
[10:49] <slomo> pitti: hi :)
[11:00] <hunger> Currently I have about 1GiB logfiles each day! And that on a laptop that is turned of most of the time.
[11:01] <mdke> you win
[11:01] <hunger> klogd is eating up more than 35% CPU time with dd taking the rest... what did you change that produces so much data?!
[11:02] <hunger> Looking into it I think I have problems with dm:-(
[11:03] <hunger> Any idea how I can find out what devicemapper device dm-13 is in normal terms?
[11:03] <hunger> Is that the minor number in /dev/mapper?
[11:06] <stockholm> i need to reach clair davis urgently
[11:06] <stockholm> is she online?
[11:07] <stockholm> does someone have a phone number?
[11:09] <fabbione> stockholm: what is the problem? she is in holidays
[11:11] <stockholm> fabbione: the problem is that she was (i thought) supposed to write the paper for mark's debconf6 talk and did not do so
[11:11] <stockholm> fabbione: and without a paper the talk wont take place
[11:12] <fabbione> stockholm: i am pretty sure she will sort it out after eastern as soon as she is back
[11:13] <stockholm> fabbione: i sent her reminderrs that the deadline was over last week
[11:13] <fabbione> also.. no need to panic.. it's not like Mark really needs paper to talk for a few hours
[11:13] <stockholm> fabbione: the stuff gets printed very soon
[11:13] <stockholm> fabbione: we require it for the proceedings.
[11:13] <fabbione> stockholm: nothing that can be done now.. she is not online and Mark isn't either.. 
[11:13] <Treenaks> fabbione: The organisers want a basic description of what the talk is about, etc., probably
[11:13] <Treenaks> stockholm: am I right?
[11:13] <stockholm> yes
[11:14] <Treenaks> I'd say: just wait for sabdfl to join, and ask :)
[11:14] <fabbione> stockholm: ok, you know what is the status, you will be able to sort it out with Claire and Mark when they will be back next week.
[11:14] <stockholm> hum, right, thanks
[11:19] <hunger> It is amasing how quiet the system gets as soon as the FS in fixed again:-| sorry for the noise.
[11:31] <mvo> jamesh: do you happen to know what magic I have to do to get something > 0 from gtk_get_current_event_time() ?
[11:43] <jamesh> mvo: have some events pending
[11:44] <infinity> dholbach: Are you doing UVF exceptions for universe?
[11:44] <jamesh> mvo: with a bit more context, I could probably give a better answer
[11:45] <dholbach> infinity: yes, with the motu-uvf team
[11:45] <infinity> dholbach: I'd like a UVF exception for mailutils (fixes a bunch of bugs, looks happy in sid, makes people stop filing bugs about /usr/bin/mail not having an MTA)
[11:46] <dholbach> infinity: go for it
[11:47] <Kamion> FYI: ubuntu-server seeds merged back into main Ubuntu seeds
[11:47] <Kamion> (with any luck)
[11:48] <infinity> \o/
[11:51] <fabbione> Kamion: cool
[11:51] <jsgotangco> nice
[11:54] <tseng> lamont: your beagle syscall change should go upstream, yeah?
[11:55] <infinity> tseng: Yes.
[11:55] <infinity> tseng: ia64 isn't the only arch without old-skool _syscall  magic numbers, and some day, they'll go away on all arches.  Eventually.
[11:59] <tseng> infinity: cheers.
[12:13] <mvo> jamesh: sorry, my isp hated me for a couple of minutes
[12:14] <jamesh> 17:43 <jamesh> mvo: have some events pending
[12:14] <jamesh> 17:44 <jamesh> mvo: with a bit more context, I could probably give a better answer
[12:15] <Kamion> oh, damnit, silbs is on holiday IIRC
[12:15] <mvo> jamesh: I need it for startup-notification support in gksu, I managed to get my events by sending a syntetic GDK_POPERTY_CHANGE event to a fake-window
[12:15] <Kamion> any ubuntu-art moderators here? silbs is the only one listed
[12:17] <jamesh> mvo: for startup notification, the time should be for the event corresponding to the user action that requested the window
[12:18] <mvo> jamesh: hm, gksu is called by the menu, so  I would have to extract and use this time to have the exact time information? (gksu itself starts some other app then)
[12:18] <jamesh> mvo: if the .desktop file says that the app supports startup notification, I think the time gets passed in the environment (iirc)
[12:20] <mvo> jamesh: I think the problem is that gksu itself is a gtk application so the startup notification stuff (from the panel) things all is done when the password dialog is up
[12:21] <jamesh> mvo: ah.
[12:21] <jamesh> mvo: I guess gksu should be starting the privileged process with startup notification then
[12:22] <jamesh> mvo: probably using the time the user pressed enter in the password dialog as the time
[12:23] <mvo> jamesh: yes, that is what I'm currently working on
[12:24] <jamesh> mvo: so a signal handler on the entry's "activate" signal and OK button's "clicked" signals would be appropriate
[12:25] <jamesh> use gtk_get_current_event_time() there (they are called while the event is being processed)
[12:25] <Riddell> mvo__: hmm, Ok is "make changes and quit" which Cancal is "quit without any changes being having been saved"
[12:25] <Riddell> mvo: I think they should be changed to "Install" and "Close"
[12:31] <mvo> Riddell: ok, fine with me
[12:31] <mvo> jamesh: thanks!
[12:31] <jamesh> mvo: it worked?
[12:34] <mvo> jamesh: my approach with the fake-window works (tested it now). the trouble with using the event time from the dialog is that sometimes no dialog is needed (when the sudo password is cached). this just occured to me, so I may have to use (the rather ugly) fake-window approach
[12:44] <sabdfl> mvo: ping, are you the right guy to forward some dapper Chinese issues to post l10n sprint?
[12:44] <sabdfl> what's abel's email?
[12:46] <dholbach> sabdfl: abelcheung@gmail.com
[12:46] <mvo> sabdfl: yes, please forward the mails to me
[12:46] <sabdfl> thanks guys
[12:46] <sabdfl> Riddell: ping
[12:46] <Kamion> sabdfl: on last night's fc-cache thing, yeah, as elmo/infinity said it's easy to take care of by a diversion - been meaning to do that for a while. I'm just waiting for a current install CD to rsync down so that I can test it now.
[12:46] <Riddell> sabdfl: hi
[12:47] <sabdfl> Riddell: mvo just answered my question in pvt msg, i was wondering if Kubuntu always uses the same fontconfigs as Ubuntu
[12:47] <sabdfl> i have questions from Intel China re Kubuntu fontconfigs for China
[12:47] <Riddell> sabdfl: yes it does
[12:48] <mvo> sabdfl:  http://people.ubuntu.com/~mvo/bzr/language-selector/language-selector--mvo/fontconfig/zh_CN is the fontconfig for mainland china
[12:49] <jamesh> mvo: if no dialog is needed, then the right time to use would be the start time for the application itself
[12:50] <jamesh> mvo: I wonder if you can get at that easily?
[12:50] <infinity> Kamion: divert update-initramfs while you're at it.
[12:50] <jamesh> (application == gksu here)
[12:50] <Riddell> mvo: slightly confusing how Language Selector is called Language Support in the menu
[12:51] <Kamion> infinity: ok, I'll have a look, thanks
[12:51] <Kamion> will do some stopwatch timing
[12:51] <mvo> jamesh: the start-time of the app started via sudo? I'm not sure that this is possible at all because we need a server timestamp for the startup notification stuff
[12:52] <mvo> Riddell: ups, I'm fixing this now
[12:52] <mvo> Riddell: hm, hold on
[12:52] <infinity> Kamion: We should compare lists sometime, so whatever speedups you do to d-i, I get in livefs, and vice versa.
[12:52] <mvo> Riddell: IIRC there was some discussion about this at some point and that was why it was renamed ...
[12:52] <infinity> Kamion: In the next week, pre-beta, might be nice for the installer case. :)
[12:53] <jamesh> mvo: looks like gdk_x11_display_get_user_time() will give you gksu's startup time if it hasn't popped a dialog
[12:53] <Kamion> infinity: at the moment I only divert scrollkeeper-{update,rebuilddb} and run scrollkeeper-update
[12:53] <jamesh> mvo: and if it has interacted with the user, it'll return the time of the last user interaction, which is what you want
[12:53] <jamesh> sounds perfect :)
[12:53] <infinity> Kamion: Ahh, kay.  Well, fc-cache is a killer, update-initramfs can easily be another few minutes on slow hardware (another 20 or 30 seconds on fast machines)...
[12:54] <infinity> Kamion: And I assume you already divert invoke-rc.d to avoid that hassle?
[12:55] <Kamion> infinity: debootstrap installs a fake start-stop-daemon, but pkgsel doesn't at present
[12:55] <Kamion> hasn't seemed to be a problem so far
[12:55] <Treenaks> infinity: my /boot doesn't understand DMA, so I know what you mean :)
[12:56] <jamesh> mvo: the only case where gdk_x11_display_get_user_time() will return 0 is if you haven't interacted with the user, and you weren't started with startup notification or the startup notification ID didn't include the timestamp
[12:56] <infinity> Kamion: It's not a problem, per se (especially since you reboot right afterward anyway, to make sure the system is in a sane state), but if you're going to be rebooting anyway, it's pointless to start daemons in the postinsts and waste time there.
[12:57] <jamesh> that last case shouldn't ever happen with recent libstartup-notification
[12:57] <mvo> jamesh: that seems to be working nicely, thanks
[01:03] <mdke> infinity, thanks for the server guide reviewing *hugs*
[01:03] <infinity> mdke: NP, sorry I was a bit cranky and harsh.
[01:05] <mdke> infinity, i don't think you were. Much appreciated, thanks
[01:05] <Kamion> infinity: yeah, not disagreeing, just seems to be higher-hanging fruit
[01:06] <infinity> Kamion: Well, most postinst use invoke-rc.d (and, IMO, they're buggy if they don't), so just diverting that to /bin/true is pretty effortless, and works well.
[01:17] <sabdfl> Kamion, infinity, Riddell: thanks muchly
[01:23] <sabdfl> what's the best way to reset the system fontconfig? if I have a machine which has been around a while, and want a virgin config there to test with as if it were a new install?
[01:24] <Treenaks> sabdfl: check if there's a .dpkg-new file, if so, replace config with that
[01:24] <Treenaks> sabdfl: and there might be some stuff left-over in ~/.font*
[01:32] <Treenaks> Ok, launchpad _really_ needs a qdb module
[01:56] <desrt> Ubuntu 5.10
[01:56] <desrt> Our Price:$3.95
[01:56] <desrt> You Save:$4.00 (50%)
[01:56] <desrt> uh.
[01:57] <_ion> Mkay. :-)
[01:57] <sivang> desrt: ?
[01:57] <desrt> http://www.osdisc.com/cgi-bin/view.cgi/products/linux/ubuntu/inst_5x.html?ad=overture
[01:57] <_ion> Quite a bargain.
[01:57] <desrt> ya.  50% is a lot to save
[01:59] <simira>  http://www.osdisc.com/cgi-bin/view.cgi/products/linux/ubuntu/inst_5x.html?ad=overture
[01:59] <simira> au
[01:59] <simira> sorry
[02:00] <simira> are they really allowed to do that?
[02:00] <desrt> of course they are
[02:03] <desrt> an interesting hack would be to burn a slightly-altered version of the gold master containing some 'mere aggregation' of a trivial proprietary product with restricted commercial redistribution terms
[02:03] <desrt> so that the CDs that ubuntu sends out for free can't be resold
[02:04] <desrt> not sure what right of first sale would say about that, though
[02:04] <Kamion> not to mention free software morality
[02:04] <tseng> wait, these are canonical cds?
[02:04] <desrt> tseng; maybe?
[02:05] <desrt> Kamion; pah.  you can always download the official iso :p
[02:05] <tseng> *that* would be immoral
[02:05] <tseng> i dont see a problem with them selling their own cdrs with ubuntu
[02:05] <desrt> Kamion; just prevents people from abusing the free-cd program for commercial purposes
[02:05] <desrt> tseng; ya.  of course that's totally fine
[02:06] <Kamion> desrt: over my dead body as cdimage guy, anyway
[02:06] <Kamion> *shrug*
[02:06] <desrt> Kamion; it's not a serious suggestion
[02:06] <Kamion> a lot of people watch #ubuntu-devel, I'd rather not have a random unserious suggestion turn up on the news sites as Ubuntu's plans
[02:07] <Kamion> so I figured an early refutation was a good plan :)
[02:07] <desrt> eugenia lurking behind the bushes?  heh.
[02:07] <ajmitch> desrt: given the quality of 'news' on there, I wouldn't be surprised :)
[02:07] <desrt> Kamion; i say let these people do it :)
[02:07] <Kamion> (not saying that folks reselling the Canonical-sent CDs aren't immoral)
[02:08] <desrt> Kamion; just makes their sites look more ridiculous
[02:08] <desrt> (as if osnews isn't already the laughingstock of our world)
[02:13] <ajmitch> desrt: that, and the sky is falling
[02:31] <Kamion> sabdfl: right, by my count in vmware, that fc-cache change saves 2 minutes out of 14 or so
[02:31] <Kamion> (that 14 being just the pkgsel step, not the whole installation, obviously)
[02:37] <Riddell> desrt: they published a retraction
[02:47] <Mithrandir> heno: your accessibility changes have been merged now, you should hopefully see them in tomorrow's daily-live
[02:47] <jdub> mjg59: http://fridge.ubuntu.com/node/332
[02:48] <heno> Mithrandir: wooooo! Thanks :)
[02:49] <heno> That gives me a chance to track down that keyboard issue I have too on a clean system
[02:49] <Mithrandir> I need to fix casper a tiny bit too, but I'm hoping to get that done now.
[02:50] <Mithrandir> gconftool is called as root, so the settings don't affect the ubuntu user. :-P
[03:00] <Riddell> heno: what's the plan about kubuntu winfoss?
[03:01] <heno> Riddell: we should get together and look at available disk space and any KDE-like apps you want
[03:01] <heno> Riddell: then I'll tweak the look a bit for kubuntu
[03:02] <heno> Riddell: what are you using now, the old one?
[03:02] <Riddell> heno: currently it's the same as breezy
[03:02] <Riddell> heno: disk space seems about right for now
[03:02] <Riddell> heno: I wouldn't mind having scribus on there if it's any good
[03:02] <heno> Riddell: ok, I'll use that as a starting point then
[03:03] <Riddell> maybe replace audacity or something
[03:03] <heno> Riddell: ah, yes that's a nice new addition
[03:03] <heno> yep, I've tested scribus briefly. seems ok
[03:03] <heno> I'll test it a bit more. It's a good candidate for theopencd as well
[03:04] <Riddell> heno: also speedcrunch would be cool http://speedcrunch.digitalfanatics.org/download.html
[03:06] <heno> Riddell: does it look as pretty as that icon now, or as ugly as it was in the past? ;)
[03:07] <heno> I guess it has a rather small user group (but you decide)
[03:07] <Riddell> heno: does which?
[03:07] <heno> speedcrunch is rather plain looking, no?
[03:08] <Riddell> heno: it's got a keypad now, which keeps the moaners quiet
[03:09] <_ion> Does it support RPN?
[03:09] <heno> Riddell: Right, I'll have a look. 
[03:10] <heno> 3.8 MB is rather large for a calculator. Can these things be made to share the Qt libs in some way?
[03:10] <heno> As gimp and gaim do with gtk
[03:10] <Riddell> speedcrunch is qt 4, scribus is qt 3 so no
[03:10] <heno> ah, ok
[03:11] <heno> how is kexi and kde-pim, and hot new developments there?
[03:11] <Riddell> oh yes, there might be a new version of kexi
[03:11] <heno> or should I go with the same packages as before?
[03:11] <heno> ok, cool. I'll look
[03:14] <heno> ... kexi 2006 edition, released in march
[03:23] <pitti> dholbach: tang{o,erine}-icon-theme approved, sorry for the delay
[03:26] <pitti> doko: openoffice.org-en-au approved; however, I hesitate adding it to l-support-en, since it would blow up the CD
[03:29] <pitti> mvo, jordi: ttf-lao approved for main; however, how should we handle that? will fontconfig etc. work well enough to include it into u-desktop? or should it become a l-support-lo dependency (ugly)?
[03:31] <doko> pitti: ok
[03:33] <mvo> pitti: it should work well enough for u-d, there are no common glyphs with other CJK fonts
[03:33] <pitti> cool
[03:34] <pitti> heno: xcursor-themes approved, so please seed it/depend on it/whatever is appropriate
[03:35] <heno> pitti: great, thanks :)
[03:36] <heno> Mithrandir: ^^ After testing the current CD I'll probably ask you to add xcursor-themes to the seed as well. I'll coordinate the config settings with Luke
[03:37] <Mithrandir> sure
[03:37] <Mithrandir> he will want to merge some changes I'm doing now first.
[03:38] <ogra> Kamion, this fifo stuff seems to be a metter of luck, sometimes it works and sometimes the backgrounded process i try to read from just dies silently ... 
[03:38] <ogra> *matter
[03:44] <Lathiat> hrm, quagga
[03:44] <Lathiat> has a question on prerm of medium priority which is "do you really want to stop quagga"
[03:44] <Lathiat> which on ubuntu gets skipped, and causes prerm to fail
[03:44] <Lathiat> would it be reasonable to bump that to high?
[03:45] <Lathiat> breaks upgrading, for example
[03:45] <Lathiat> (as it just did for me)
[03:45] <infinity> If the prerm fails when the question is answered incorrectly, why does the question exist at all?
[03:45] <Lathiat> infinity: the question is
[03:45] <Lathiat> infinity: "do you relaly want to stop quagga"
[03:45] <Lathiat> which is medium, and edefauls to false
[03:45] <Lathiat> and is ignored so it just abortsd
[03:45] <Lathiat> i assume in debian the default priority shows medium questions
[03:45] <infinity> Yes, my point is that if there's only one correct answer, the question shouldn't be asked.
[03:46] <Lathiat> i think the question is
[03:46] <Lathiat> "i'm about to upgrade quagga"
[03:46] <Lathiat> "if i stop it, you might lose connetivity"
[03:46] <Lathiat> "you sure?"
[03:46] <infinity> pcmcia-cs had a similar question, with equal irritation.
[03:47] <infinity> See what that does now and follow suit.
[03:47] <Lathiat> i mean, its a valid question, bu shoudlnt it be priority high?
[03:47] <Lathiat> hrm
[03:47] <infinity> My gut feeling would be to change the default answer.
[03:47] <Lathiat> i'd be inclined otherwise
[03:47] <Lathiat> to make it high?
[03:47] <infinity> Most users just want to upgrade (and at that point have already downloaded packages)
[03:47] <Lathiat> not tha ti have alot of experience in that area or on what others do
[03:48] <ajmitch> most users wouldn't want to run quagga without really knowing what it does :)
[03:48] <Lathiat> sure, but it might not be obvious why your dist-upgrade barfs halfway through
[03:48] <Lathiat> hell it took me a few minues to figur eit out, and i had to dig aroudn the prerm file.
[03:49] <Lathiat> infinity: it seems to lack the question entirely
[03:50] <ajmitch> Lathiat: I'm not saying it should be kept broken
[03:52] <Lathiat> i guess whether or not to ask at all is another question
[04:05] <maswan> heh, after a workalike to that xscreensaver vulnerability hit me in the late 90ies, I alwasy make sure to not have focus in any window when locking the screen. :)
[04:17] <Diziet> Yay!  I finally managed to get firefox to start with a Spanish home page.  Now all I have to do is automate my fiddlings.
[04:18] <sladen> Diziet: can you get it to replace the start-page URL with an empty URL too rather than file:///mucho-wankage
[04:25] <Diziet> Err ?  You want it to start with a blank page ?
[04:27] <lifeless> Diziet: I think he wants it to show the start page
[04:27] <lifeless> Diziet: but not the url. 
[04:28] <lifeless> so that users dont have to delete the url every-single-time-they-hit-ctrl-T
[04:28] <Treenaks> replace about:blank with the intro pagE?
[04:29] <Treenaks> scary :)t 
[04:31] <Diziet> I'm definitely unconvinced by that as an idea.
[04:33] <infinity> I'm fairly sure it's a simple XUL/JS one-liner to clear the address bar.
[04:34] <infinity> The idea does have merit.
[04:34] <Kamion> they don't have to put any effort into deleting the URL - just ctrl-L <type stuff>
[04:34] <Kamion> or highlight the URL and type
[04:34] <infinity> No one cares that they're looking at file:///usr/share/ubuntu-artwork/home/index.html, they care that they just started a browser and want to use it.
[04:34] <infinity> (It makes the page feel more "built-in" that way)
[04:35] <infinity> An empty address bar screams "type in my and see what happens!"
[04:35] <infinity> s/my/me/
[04:37] <Mithrandir> TheMuso: please merge my casper changes if you want to give me something to update from.
[04:37] <infinity> It also resolves the "should a single click insert or hilight?" debate once and for all, I think, since the number one time a user wants to single-click-to-retype-the-url is when they start a fresh browser.
[04:38] <infinity> If a fresh browser has nothing in the URL bar, problem solved. :)
[04:38] <infinity> (For the record, I'm firmly on the "single click should insert" side of that debate)
[04:39] <ploum> Is there a way to get the previous version of a package in Dapper ? I'm looking to the previous nvidia-glx one in order to track something. (if anyone has it in his /var/cache/apt/archives, I would appreciate)
[04:40] <infinity> ploum: The 8178 driver, from lrm 2.6.15.7-4?
[04:40] <ploum> infinity: yes, I guess it's that one
[04:41] <infinity> ploum: Hang on a second.
[04:41] <ploum> (I have a lot of freeze with nautilus and epiphany and I want to try if they are related to the driver)
[04:43] <infinity> ploum: i386 or amd64?
[04:43] <ploum> i386
[04:43] <sladen> Diziet: yes.  intro-page (unless they have set their own start-page), but with empty URL bar and focus in the URL bar
[04:44] <infinity> ploum: http://people.ubuntu.com/~adconrad/linux-restricted-modules-2.6.15_2.6.15.7-4/
[04:45] <infinity> ploum: Be sure to downgrade all related packages (LRM and LRM-common as well), not just nvidia-glx.
[04:45] <infinity> ploum: And reboot between tests.
[04:45] <ploum> infinity: many thanks :-)
[04:45] <ploum> I will do that
[04:45] <freeflying-ibook> jdub: ping
[04:46] <sladen> Diziet: I did look at it once.  It requires starting having that start page load with permission to talk to the chrome
[04:47] <ploum> so, let's go for reboot.. (I hate that)
[04:47] <ploum> cheers
[04:50] <sladen> infinity: /proc/bus/pnp seems to have disappeared and the sysfs equivalent seems underpopulated.  what's a good way to 'lspnp' these days?
[04:54] <bandini> mmh bug-buddy seems to segfault with today's update. anyone else seeing this? (crashes in md5_get_digest_from_file())
[04:56] <seb128> that's not from today
[04:57] <bandini> mmh ooutdated mirror of mine?
[04:59] <bandini> 2.14.0-0ubuntu1, here
[05:02] <janimo> infinity: any eta on xubuntu-live? thanks
[05:17] <ogra> wow, fabbione, you made someone really happy in bug 3527 :)
[05:17] <Ubugtu> Malone bug 3527 in Ubuntu Breezy "OpenGL screensaver freezes!!" [Normal,Rejected]  http://launchpad.net/bugs/3527
[05:32] <dholbach> ogra: btw the new dia is in debian - it might only be a sync - just saw the upload mail, dunno how well it'd work for us
[05:33] <ogra> i'll do a testbuild today and ask for the sync
[05:33] <ogra> (in fact that was what i was hoping for)
[05:33] <dholbach> rocknroll
[05:33] <ogra> :)
[05:34] <dholbach> seem it was the right day to sign up for debian-devel-changes :)
[05:34] <ogra> hehe :)
[05:42] <lifeless> ogra: gss is still bust for me
[05:42] <lifeless> ogra: and when it goes bad, so does my metacity mouse-focus behaviour
[05:43] <ogra> lifeless, with 2.14.1 ?
[05:44] <lifeless> ogra: ah, no.
[05:44] <lifeless> will try
[05:45] <Diziet> Anyone here use es_ES, fi_FI, lt, pl_PL, ru_RU, sk, sv_SE, or zh_CN ?
[05:45] <Diziet> And would you be willing to try an under-the-table m-f-l-* package for me ?
[05:45] <Diziet> I've tried it myself of course and it seems to work but IWBN to test it in a `real' system.
[05:46] <Diziet> OK, I'll try in #ubuntu.
[06:27] <bddebian> How can I symlink dirs?  dh_link seems to fail on directories?
[06:33] <Kinnison> according to the manpage it shouldn't
[06:33] <Kinnison> the second example indicates a symlink of directories
[06:33] <bddebian> I know but it does
[06:34] <bddebian> fail that is
[06:34] <seb128> maybe you don't use it correctly?
[06:34] <bddebian> Gah, will guys quit fixing things, my dist-upgrades take forever.. :-)
[06:34] <Kamion> bddebian: it won't work right if the target directory already exists
[06:34] <bddebian> seb128: Probably not, that's why I'm asking
[06:35] <bddebian> Kamion: Ahhh
[06:35] <Kamion> i.e. it won't remove the target directory for you if something else already created it
[06:35] <Kamion> such as make install or dh_install
[06:35] <bddebian> Shite.
[06:35] <Kamion> why so bad? it's usually easy to fix up paths in whatever's doing the installation
[06:35] <bddebian> Hmm, can I wildcard it?  dh_link /foo/bar/* /bar/foo/* ?
[06:35] <Kamion> no
[06:36] <Kamion> just fix whatever's installing to the directory that you want to be a symlink
[06:36] <Kamion> it should probably install to the target of the symlink instead
[06:36] <seb128> speaking about link, I'm not sure but is there a bug 
[06:36] <bddebian> Kamion: Well qgis's help path is hardcoded in the app for /usr/share/qgis/doc and Debian is installing in /usr/share/doc/qgis/
[06:36] <seb128> if package n-1 has dir_name and package n makes it a symlink on other_dir_name
[06:37] <seb128> does that work?
[06:37] <azeem> bddebian: then that hardcoded path should be patched, IMHO
[06:37] <Kamion> bddebian: change qgis' Makefile to install somewhere else, then
[06:37] <bddebian> azeem: Agreed
[06:37] <seb128> I think that's something known but I'm not sure
[06:37] <Kamion> or mv and rmdir, whatever's needed
[06:37] <bddebian> But it always comes bad to the question of how far to veer from Debian
[06:37] <bddebian> Err s/bad/back/
[06:38] <Chipzz> bddebian: given the rules of shell wildcard expansion, how *would* that work?
[06:38] <azeem> bddebian: well, it looks like the right fix
[06:38] <bddebian> Chipzz: It was just a question. :-)  I'm a little slow
[06:38] <azeem> so submit the patch to Debian as well, maybe?
[06:38] <ogra> mdz, the patch for fullscreen preview in g-s-s landed in g-s-s CVS today, if i could include it, that would make doko a happy man (we talked about it before and you said we'd have to wait if upstream accepts it)
[06:38] <Chipzz> bddebian: ;)
[06:39] <mdz> ogra: is the openGL visual problem fixed yet?  that is much more important
[06:39] <bddebian> azeem: You Debian types ignore my patches.. ;-P
[06:39] <azeem> whatever.
[06:39] <bddebian> :-)
[06:39] <ogra> mdz, nope
[06:39] <ogra> (i know about its importance)
[06:44] <bddebian> ack
[06:44] <sladen> regarding #22045, I've also noticed that the openGL hacks aren't getting a z-buffer, so it's possible they're not getting double-buffered either
[06:44] <ogra> mdz, i was hoping for the announced patch in that bug
[06:45] <bddebian> whee
[06:45] <ogra> else i could indeed need help with that one, i still have no machine where i could reproduce it
[06:45] <mdz> ogra: announced patch?
[06:46] <ogra> mdz, see the comment from Bart Hartgers in that bug
[06:46] <mdz> ogra: many of us here in the channel can reproduce it
[06:46] <ogra> "I might be able to help. I've a radeon 9200 in my desktop that has the same problem, and I am halfway done with a patch agains 2.14 to fix this on my system."
[06:47] <mdz> that was a week ago
[06:47] <mdz> ogra: I'm sure you can get an account on a machine you can use to reproduce the bug
[06:48] <mdz> I can give you an account on mine, but it would probably be better with someone in a closer time zone so they could answer questions about the display if needed
[06:49] <mdz> perhaps email Bart and ask if he can give you his incomplete work for you to use as a starting point
[06:49] <mdz> anything but wait
[06:49] <ogra> yep
[06:49] <ogra> that was my plan, i wanted to give him some time since he volunteers
[06:55] <mdz> there's no shame in asking for his code; you're offering to help him finish it
[06:56] <ogra> yep
[06:57] <mdz> ogra: also, was the xscreensaver patch for thin clients carried over to gnome-screensaver?
[06:57] <mdz> or does it provide some other mechanism?
[06:58] <ogra> its using gconf keys i was thinking about to just unset the key for the hacks ... 
[06:58] <ogra> (g-s-s calls that throttling)
[06:59] <mdz> how does it work?  does ldm set them on the ssh command line?
[06:59] <ogra> but adding it to g-s-s is trivial (its a two line patch)
[06:59] <mdz> oh, so that's currently broken?  is there a bug open?
[06:59] <ogra> i'd set them from ltsp's postinst ... so we have a system setting and if the user insists to use hacks, he's still able to
[07:00] <ogra> nope, i think you closed it ... it was the same bug i carried over ...
[07:00] <ogra> its still on my TODO, so dont worry, wont get lost
[07:01] <ogra> Mithrandir, did you set the RUNNING_UNDER_GDM env var from casper again (while we're at g-s-s)
[07:01] <mdz> ogra: if I close a bug because it was fixed, and it has been unfixed without my knowledge, it is appropriate to correct me :-)
[07:02] <ogra> mdz, ok, i'll reopen it (it were two bugs about the same issue in two packages and i have it on my list anyway)
[07:03] <mdz> ogra: I didn't close it; it's bug #18595
[07:03] <Ubugtu> Malone bug 18595 in xscreensaver "Needs to disable graphics-intensive modes for LTSP" [Normal,Unconfirmed]  http://launchpad.net/bugs/18595
[07:03] <ogra> meh, right, but it indeed doesnt show up in the g-s-s buglist :)
[07:04] <bddebian> Heya LaserJock
[07:04] <ogra> (added)
[07:11] <Riddell> Kamion: i18n working well in kde-espresso
[07:11] <Riddell> Kamion: is there a way on the Ready label to not include "ubuntu"?  so it can be changed to Kubuntu?
[07:11] <Riddell> and probably ogra will want his to say edubuntu somehow
[07:12] <sladen> Riddell: could it be de-branded completely ?
[07:13] <ogra> isnt that what espresso-ubuntu-artwork is for ? 
[07:14] <Riddell> sladen: the string says "Ubuntu is now ready to install", that could be de-branded but it's not a bad string to have like that
[07:14] <Riddell> of course I could just run s/Ub/Kub/ in the frontend easily enough
[07:15] <Burgwork> Riddell, why not just say "Ready to install"?
[07:15] <Burgwork> as branding in espresso is bad and should be avoided
[07:17] <sladen> ogra: grep -rci opengl gnome-screensaver-*  = 0  does it actually do any  gdk_gl_ specific calls, or leave that to the hacks?
[07:17] <Riddell> Burgwork: that could well be the best way, although it does seem quite nice to know what you're installing
[07:18] <Burgwork> Riddell, you already do - Ubuntu
[07:18] <Riddell> Burgwork: I'm not installing ubuntu, that's the point :)
[07:19] <Burgwork> if an Ubuntu live cd installs Fedora, I would concerned
[07:19] <sladen> Riddell: the image in the background, the desktop, the front of the CD, and the CD sleave should give you a good hint
[07:19] <Riddell> I'm not against the idea, I just follow Kamion, poke him
[07:20] <sladen> anything that is shipped with 'Ubuntu' in the name is unattractive for other derivers
[07:21] <Burgwork> sladen, I am fairly certain this is a simple mistake, as Kamion is pretty busy and probably just missed it
[07:21] <ogra> sladen, it sadly leaves it to the hacks, which is wrong, since you cant force the hacks visual in --root mode
[07:22] <ogra> xss just selects the visuall according to what xscreensaver-gl-helper gives it
[07:22] <ogra> and sets it in the daemon
[07:28] <Kamion> Riddell: the "ready" label was awkward, I couldn't come up with a good phrasing that excluded Ubuntu
[07:28] <Kamion> suggestions welcome
[07:28] <Kamion> you can't do s/Ub/Kub/ as not all translations will include "Ub"
[07:29] <Burgwork> "Finish Install"
[07:29] <Kamion> the sentence needs to be rephrased from the current "Ubuntu is ready to install on this computer."
[07:29] <Kamion> Burgwork: not a heading
[07:29] <Burgwork> ah
[07:29] <Kamion> the heading is "Ready to install"
[07:29] <Burgwork> "You are now ready to complete the installation"
[07:29] <Burgwork> ?
[07:29] <Kamion> sounds clunky, I'm afraid
[07:29] <sladen> or  Complete Installation.
[07:30] <Kamion> to me anyway
[07:30] <Burgwork> yes
[07:30] <ogra> s/Ubuntu/The System/ ?
[07:30] <Kamion> sladen: that's a heading, not appropriate
[07:30] <Kamion> ogra: ugh
[07:30] <Kamion> the system is ready to install onto your computer machine thing noun.
[07:30] <Kamion> "system" works elsewhere, but not there, IMHO
[07:30] <ogra> heh
[07:30] <Kamion> the welcome label needs to be rephrased too
[07:30] <Kamion>  Ready to install? Once you answer a few questions, Ubuntu can be installed
[07:30] <Kamion>  on this computer so you can run the system at full speed and without the
[07:31] <ogra> probably "I am" ?
[07:31] <Kamion>  CD.
[07:31] <ogra> has a personal note
[07:31] <Kamion> ogra: I'm avoiding anthropomorphisation
[07:31] <Kamion> doesn't sound right in espresso
[07:33] <sladen> ogra: xscreensaver-gl-helper is only a separate binary to avoid linking xss against libGL.   gdk_gl_query() && gtk_gl_area_new([GDK_GL_RGBA, GDK_GL_DOUBLEBUFFER, GDK_GL_DEPTH_SIZE, 1, GDK_GL_NONE] );   should do what's needed.  (or abuse glXChooseVisual() directly on the root window)
[07:33] <janimo> Kamion, for xubuntu art in the gfx boot will a .pcx file suffice?
[07:34] <janimo> if so how do you prefer it? bug against d-i, personal mail etc
[07:34] <ogra> sladen, thanks !
[07:34] <Kamion> janimo: just point me to a usplash image and I can do the conversion
[07:34] <Kamion> janimo: bug against debian-cd
[07:34] <janimo> Kamion, oh cool. xubuntu-artwork has the usplash
[07:34] <janimo> that's the source package
[07:34] <Kamion> ok, file me a bug and I'll try to look at it tomorrow
[07:35] <janimo> ok, thanks
[07:56] <sivang> re all
[07:57] <mdz> Kamion: is the lack of an initrd.list in the archive a d-i problem or a soyuz problem?
[07:58] <jordi> pitti: thanksf or ttf-lao
[07:58] <Kamion> mdz: it's called udeb.list now, in line with changes made in Debian to help the archive track which sources it needs to keep around for GPL requirements; might have slightly different contents
[07:58] <Kamion> it's a directory or two up, too
[08:04] <mdz> thanks
[08:10] <ogra> oh fun, libgtkgl2.0-dev is indeed in universe ...
[08:10] <ogra> ergh, but gtkglarea5-dev (gtk 1.2) is in main ????
[08:11] <wasabi_> http://www.amazon.com/exec/obidos/ASIN/032142722X/desktoplinuxa-20/104-8693816-3330301   <--- nice
[08:11] <Burgwork> wasabi_, http://www.amazon.com/gp/product/0132435942/ref=pd_sxp_elt_l1/002-9576091-6069611?n=283155 <-- better ;)
[08:11] <ogra> pitti, ^^^ didnt you clean up the gtk 1.2 mess already ? 
[08:12] <pitti> ogra: ~ 7 packages still b-dep on it
[08:12] <_mvo_> "foreword by Mark Shuttleworth"
[08:12] <ogra> meh
[08:12] <pitti> ogra: it's still on my wishlist to clean that up, but it's work
[08:12] <wasabi_> mvo, grab latest GAptI. 
[08:12] <wasabi_> When you get a chance. ;)
[08:12] <ogra> i might need libgtkgl2.0-dev for g-s-s :/
[08:13] <mdz> yay, awty unbroken now
[08:13] <Burgwork> _mvo_, I might be a little biased of course...
[08:13] <pitti> mdz: what does that tool do?
[08:13] <_mvo_> Burgwork: sure ;)
[08:13] <mdz> pitti: reports the versions of packages in many places, from the archive to the CDs
[08:13] <_mvo_> wasabi_: pulling
[08:13] <mdz> so we can see when a new version is propagated
[08:14] <pitti> ah
[08:14] <wasabi_> mvo, it's pretty much all implemented and works. Some parts are a bit hacky, because I need to find proper ways to do it.
[08:14] <wasabi_> I also need a little better tie between gapti and synaptic for installation.
[08:14] <_mvo_> wasabi_: anything you need in synaptic for this?
[08:14] <pitti> mdz: it sounds like a tool that collects build failure states, anastacia inconsistencies, etc. and print a summary report :)
[08:14] <mdz> Kinnison: is the drescher archive exported via http anywhere?
[08:14] <wasabi_> Yeah. I'd like a way to add a key to apt-key, with the proper confirmation dialog.
[08:14] <mdz> pitti: it can grow ;-)
[08:14] <wasabi_> Right now I'm using some of your code in gnome-software-properties.
[08:15] <mvo> wasabi_: you send a mail/bug-report for this, right?
[08:15] <wasabi_> the apt_key class in dialog_apt_key.py
[08:15] <wasabi_> Yeah.
[08:15] <sivang> Burgwork: I'm honnored to know you :)
[08:15] <mvo> wasabi_:  ok, I'll have a look. loads of stuff going on currently, sorry for the lag
[08:15] <wasabi_> I was also considering writing my own package installation dialog for it, like update-manager. I would like to talk with you 1on1 about it though.
[08:16] <wasabi_> Since it might actually fit into Synaptic.
[08:16] <wasabi_> I don't just want to initiate an install for whatever packages are listed. Would rather have a screen so the user can see what is going to be installed, and not just as "raw" as synaptic.
[08:17] <mvo> wasabi_: update-manager is using synaptic as backend, the dist-upgrader is independent 
[08:17] <mvo> wasabi_: if you want your own (based on apt.InstallProgress/FetchProgress) we should probably move it into python-apt or python-apt-gtk or something like this
[08:17] <wasabi_> Something akin to how OS X and windows installers are.   It shows a list of what will be installed, top leve items, like "Photoshop", and then under that has conditionals that you can choose like "Extra Brushes"... see wha tI mean? Those are recognized as Suggests and Recommends.
[08:17] <wasabi_> But the interface needs to be a bit better at conveying the idea.
[08:18] <wasabi_> And not inunduating the user with tons of things like package-common package-base, package-foo, libfoo, libbar, etc.
[08:18] <sivang> pitti: hehe
[08:18] <sivang> pitti: yes, paper is going to deprecated in not time anywyas
[08:19] <pitti> sivang: still, rc1 was a real mess, as were previous snapshots; so I'm hopeful again
[08:19] <mvo> wasabi_: right. if it's only about installing that shouldn't be a problem
[08:20] <wasabi_> It's something which we could benefit from in synaptic primem but which makes a real lot of sense when real users start clicking on .apt files.
[08:20] <wasabi_> I'm going to put together at least a little UI mockup for it.
[08:21] <mvo> wasabi_: cool, thanks!
[08:21] <wasabi_> Anyways, gapti actually works. Adds the key, etc.
[08:21] <wasabi_> Installs the packages.
[08:21] <wasabi_> Not sure ow to tell the browser it's allowed to open the file yet, though. ;0
[08:22] <Mithrandir> ogra: running_under_gdm has always been set.
[08:22] <ogra> uuh, really 
[08:22] <ogra> thats odd, then something in g-s-s is wrong
[08:22] <wasabi_> Obviously it's not set in stone yet. I think I'm going to change the .apt file format... and want some other input on that.
[08:23] <wasabi_> Instead of having PublicKey, require the entire file to be cleartext signed by the key.
[08:23] <wasabi_> That way the file must originate from the same source whose archives are being added.
[08:25] <mvo> right, that sounds sensible
[08:26] <Riddell> pitti: will you be able to review kmplayer and knetworkmanager for main inclusion before beta freeze?
[08:26] <wasabi_> Also, I want to talk about this in a better forum. For the entire idea to succeed, there has to be some marketing push behind it.
[08:26] <pitti> Riddell: yes, will process the list in the next days
[08:26] <wasabi_> We want ISVs to actually support it, and publish their software with it.
[08:27] <pitti> Riddell: kmplayer sounds suspicious :) is it *that* mplayer/
[08:27] <pitti> ?
[08:27] <Riddell> pitti: we use the xine plugin
[08:27] <pitti> ah
[08:27] <Riddell> and don't even build the mplayer one as I remember
[08:29] <jordi> Riddell: any news about that pot?
[08:29] <Riddell> jordi: which one?  katapult?
[08:30] <jordi> yah
[08:30] <Riddell> jordi: no response from mez yet, I'll send him another e-mail
[08:30] <bddebian> Uhm, interesting new background :-)
[08:30] <Riddell> bddebian: where?
[08:30] <jordi> ok
[08:30] <bddebian> Riddell: Gnome 
[08:30] <pitti> wow, cups prints!!!
[08:31] <netstar> It seems the r200_dri.so is not bundled with Dapper
[08:31] <Mithrandir> I'd be more surprised if cups was scanning.
[08:31] <pitti> Mithrandir: after all the mess and failures I have seen in the last weeks, this is a miracle, really!
[08:32] <Chipzz> arf!
[08:33] <Chipzz> 20:33 -NickServ(NickServ@services.)- You have already identified
[08:33] <Chipzz> ?
[08:34] <netstar> how can I search all packages in repositories for a file name?
[08:36] <mvo> Kamion: re launchpad bug #36022, I understand that we work-around the problem for now?
[08:36] <Ubugtu> Malone bug 36022 in launchpad-upload-and-queue "change-override can't handle pockets" [Major,Confirmed]  http://launchpad.net/bugs/36022
[08:43] <sivang> indeed interesting background
[09:28] <lucas> I've got a problem with librmagick-ruby. There are several bugs which would justify an update for breezy.
[09:30] <lucas> I initially asked mdz and Kamion about an upload through breezy-updates. I was asked to contact the breezy-backports team instead. I did, and after a long delay, I'm told that packages in breezy-backports must be backported without modifying the dapper source.
[09:30] <lucas> this is not possible with librmagick-ruby.
[09:30] <lucas> what should I do ?
[09:36] <mdz> lucas: why do you feel that it is impossible?
[09:37] <mdz> lucas: it is acceptable to modify the package in dapper to allow it to build correctly on breezy as well
[09:37] <lucas> because, for example, there's a build-dep on debhelper >= 5
[09:37] <lucas> I don't understand why breezy-backports packages should be built using dapper sources
[09:38] <tseng> because it would be a mess to have specific breezy-backports sources
[09:39] <lucas> ok. I'm tired of those librmagick-ruby issues. What should I do : mark the bugs as rejected, or leave them open, hoping that somebody will fix them someday ?
[09:40] <pitti> lucas: these bugs aren't fixed in dapper?
[09:40] <lucas> they are
[09:41] <lucas> they are Fix Released for Dapper, but still open in breezy
[09:54] <Burgwork> infinity, do you enjoy doing janitor work as other people play in the sandbox? :)
[09:55] <mdke> who knows what the "P" stands for in the ubuntu-server "Install a LAMP server" option?
[09:55] <tseng> mdke: thats a trick question
[09:55] <sivang> mdke: python probably :)
[09:56] <tseng> mdke: it stands for php/perl/python, take your pick.
[09:56] <mdke> tseng, I've written: to install a LAMP server (Linux, Apache, MySQL, PHP/Perl/Python), select blah
[09:56] <mdke> oh good
[09:56] <mdke> thanks
[09:56] <tseng> i am not sure that it installs all those, though
[09:56] <sivang> mdke: we hae something like that in ubuntu server?
[09:56] <tseng> is there a meta package?
[09:57] <mdke> tseng, I'm not sure... there is an option in the boot screen, it seems: http://www.flickr.com/photos/jsgotangco/126410755/
[09:57] <tseng> oh
[09:57] <tseng> is that something he has made custom?
[09:58] <tseng> i havent seen that.
[09:58] <tseng> the official server cd installs ubuntu-minimal and ubuntu-server, not lamp out of the box last I saw
[09:59] <mdz> pitti,Kamion: would it make sense to consider Malone for UbuntuMainInclusionQueue, or do you prefer the wiki workflow?
[10:00] <mdke> tseng, I think that's a screenshot from flight 6
[10:00] <mdke> (of the server cd)
[10:02] <janimo> mdz, hi. I can answer your mail in detail tomorrow
[10:05] <mdz> janimo: ok, thanks
[10:14] <luopio> hi all. I was hired to develop a live-cd for the IOI (International Olympiad in Informatics). The current live cd uses Morphix, but I'm leaning toward switching to Ubuntu. I'd be interested to hear how stable 6.06 live cd is. Any experiences?
[10:15] <bluefoxicy> "The kernel freeze is a deadline for kernel updates, since they require several lockstep actions which must be folded into the CD building process."
[10:15] <Tm_T> luopio: toimii
[10:15] <bluefoxicy> Someone explain to me why exactly this can't be automated
[10:15] <Tm_T> luopio: I mean. works, though this is wrong channel for that kind of questions
[10:15] <luopio> Tm_T: mutta kauanko? :)
[10:15] <tseng> bluefoxicy: because its a waste of time to solve something in software when you don't have to
[10:16] <Tm_T> luopio: see you in #ubuntu-fi, ok?
[10:16] <luopio> Tm_T: ok
[10:16] <Tm_T> thanks
[10:16] <bluefoxicy> tseng:  so as opposed to changing the kernel package, you have to...?   Change the kernel package, alter scripts to copy vmlinuz-{version}{thing}-01, change the location of modules hard-coded into a script...?
[10:16] <tseng> uh, what
[10:17] <bluefoxicy> tseng:  I just don't understand the problem at all, perhaps you could explain what it takes to build a CD besides picking packages and running a build script
[10:17] <bluefoxicy> is there an #ubuntu-mkcd or such?
[10:17] <tseng> exactly that, but you dont just change core packages like the kernel in the middle of the process
[10:18] <tseng> "lockstep actions" include having a matching version of linux-restricted-modules
[10:18] <bluefoxicy> I would assume you would do something like drop in a new kernel core package, rebuild the CD, burn/qemu, test, and then nod and smile when it works
[10:18] <bluefoxicy> ah
[10:18] <tseng> which can sometimes lag by days
[10:18] <tseng> and would give you a broken cd image
[10:18] <bluefoxicy> yes I've seen that
[10:18] <tseng> and the testing process for flight cds is a litle more involved
[10:19] <tseng> including installs and livecd tests on at least 3 archs
[10:19] <bluefoxicy> however I'm assuming that you don't release kernel-x without -restricted-modules-x
[10:19] <tseng> you assumed wrong
[10:19] <bluefoxicy> ah
[10:19] <tseng> rolling a flight release is a big production.
[10:19] <ogra> the kernel is also part of the installer and bootprocess of the CD ... that needs manual intervention
[10:19] <Mithrandir> tseng: for each of ubuntu, kubuntu, edubuntu and now xubuntu, yes.  So it's a large amount of testing.
[10:19] <tseng> Mithrandir: indeed.
[10:20] <bluefoxicy> this isn't fairly distributed between devs?
[10:20] <bluefoxicy> I would think for each project there's a dev sitting around who can test it
[10:20] <bluefoxicy> as opposed to one guy in his basement with 60 hot pockets rabidly changing CDs between 5 machines and 2 qemu installs
[10:20] <tseng> there are about 30 "core dev" folks atm
[10:20] <ogra> you still have to download the isos, burn them and install them before testing
[10:21] <Mithrandir> bluefoxicy: sure, it'd be ideal if we had an army of testers who jumped at every opportunity to test a new fresh image.  Unfortunately we don't.
[10:21] <tseng> only a handful are responsible for release management
[10:21] <bluefoxicy> ah
[10:21] <ogra> edubuntu are 6 CDs i have to sync ... if i discover something critical i can rebuild the whole set and start over 
[10:21] <Mithrandir> bluefoxicy: it's also that we don't want to have to stop development for a day or two every second week because we're releasing a flight.  Just freezing the archive slows us down enough.
[10:21] <ogra> testing flights is very time consuming ...
[10:22] <bluefoxicy> so one more question
[10:22] <bluefoxicy> kernel freeze is may 18, 2.6.16 is out, 2.6.17 is in RC1 so it probably won't be out by next month, unless tehy release before RC9 this time
[10:22] <bluefoxicy> (what was Linus' record, RC14?)
[10:23] <Mithrandir> our 2.6.15 is probably fairly close to 2.6.16.
[10:23] <bluefoxicy> is it likely we will see 2.6.16 in Dapper, or is it more probable that we're sticking with .15 + bug fixes?
[10:23] <bluefoxicy> probably
[10:23] <tseng> .15 is a sure bet
[10:23] <ogra> .15
[10:23] <bluefoxicy> alright
[10:24] <tseng> if you have a bug solved in a later release, file it now
[10:24] <tseng> get it backported
[10:24] <bluefoxicy> I don't, I am more trying to find a bug solved and don't know if it is even relavent.  The LKML totally ignored the message I posted asking.
[10:24] <tseng> well
[10:24] <tseng> I have some ideas why :)
[10:24] <bluefoxicy> that would, btw, be that the via DRM driver in the kernel causes hard freezes
[10:25] <bluefoxicy> when i boot X using driver 'via' I pretty much die on screen savers :/
[10:25] <bluefoxicy> if I look in time, dmesg starts piling up with oopses on various things, once in a while there's a NULL pointer deref in kernel space, etc.
[10:26] <bluefoxicy> I have no idea if it's fixed in a newer release, maybe I should build a 2.6.17 some time and see.
[10:26] <bluefoxicy> it's a mix between plain laziness, and not wanting to test something that's going to crash my production PC, which I hate shutting down anyway
[10:27] <crimsun> (you're already running dapper on your production...)
[10:27] <bluefoxicy> but either way, it otherwise works
[10:27] <ogra> crimsun, shhh
[10:27] <bluefoxicy> crimsun:  yes, witth visible effects
[10:27] <bluefoxicy> (rhythmbox:16578): GStreamer-CRITICAL **: gst_pad_push: assertion `GST_IS_PAD (pad)' failed
[10:27] <bluefoxicy> Segmentation fault
[10:27] <bluefoxicy> That being a recurrant one :)
[10:28] <bluefoxicy> crimsun:  my strategy is pretty much, "If it breaks, I find a work-around to keep the system alive until they fix it, and file a bug."
[10:28] <bluefoxicy> at least I'm contributing something ;)
[10:29] <dieman> wow
[10:29] <bluefoxicy> at any rate, no chance of 2.6.16+ in dapper, not a big deal.  I was just curious.
[10:29] <dieman> tetex-bin paper size bugs are 6 years old in debian
[10:33] <bluefoxicy> tseng:  what I'm hoping for is Dapper+1 having the network authentication stuff.  Fedora and RHEL already have a nice quick set-up interface for that on first boot, just check off LDAP and configure it.
[10:34] <bluefoxicy> of course they also have rpm rollbacks, which is important too.  deb rollbacks would be great, install a broken package, roll it back to a working one.  That's an important feature for enterprise data centers.
[10:34] <Mithrandir> it'd be really sweet to have a proper AD killer.
[10:35] <bluefoxicy> Mithrandir:  you can set up roaming profiles on LDAP, what else does AD do?
[10:35] <Mithrandir> central handling of user accounts, both the user info and the password part.
[10:36] <bluefoxicy> yes
[10:36] <bluefoxicy> ldap does that
[10:36] <Mithrandir> it also gives you nice things like two-way authentication of services.
[10:36] <Mithrandir> single-sign-on.
[10:36] <wasabi_> I've been having some probs with restricted-modules.
[10:36] <bluefoxicy> what we need is a UI to configure OpenLDAP and administrate it (Directory Administrator is nice... but something closer to our current utils would be good)
[10:36] <wasabi_> Not running depmod properly.
[10:36] <bluefoxicy> A configuration util for OLDAP is non-existent
[10:36] <Mithrandir> bluefoxicy: you shouldn't have your passwords in ldap.
[10:36] <wasabi_> Havne't been able to pinpoint the problem, beacuse running depmod fixes it. ;0
[10:37] <bluefoxicy> we'd need something to set up the server, initialize the things for storing passwords and accounts in MD5, import or generate an openssl X.509 cert, set up TLS...
[10:37] <wasabi_> bluefoxicy: I've written a large wiki article describing such a utility.
[10:38] <wasabi_> And how it would fit into the entire lifecycle.
[10:38] <wasabi_> what was it called. one sec.
[10:38] <wasabi_> https://wiki.ubuntu.com/DomainAuthenticationUtility
[10:39] <wasabi_> Basically it just becomes a meta model around a few pieces of knowledge: domain name, computer account, etc.
[10:39] <pitti> mdz: I don't have a problem with the wiki, it's fast to handle; you think about handling reports like specs?
[10:40] <mdz> pitti: just a thought; since you two are most directly involved, do what works best for you
[10:40] <mdz> the sync processing in malone seems to be working very smoothly
[10:42] <dieman> wasabi_: oo
[10:42] <dieman> wasabi_: very interesting
[10:42] <wasabi_> THe individual pieces should really be fixed before we move on though.
[10:43] <dieman> yeah
[10:43] <dieman> i just skimmed it
[10:43] <wasabi_> libnss_ldap should discover it's info from DNS properly, which it doesn't.
[10:43] <wasabi_> libpam-krb5 barely functions.
[10:43] <dieman> we're actually still a nis shop here
[10:43] <dieman> heh
[10:43] <wasabi_> Doesn't create keys right.
[10:43] <wasabi_> Need a daemon per-user and per-machine that keeps cache updated
[10:43] <wasabi_> needs a daemon per machine that changes password on a schedule.
[10:43] <wasabi_> Needs something to keep /etc/krb5.keytab uptodate with all of that.
[10:44] <wasabi_> libnss_ldap needs to fall back properly to a slave LDAP server.
[10:44] <wasabi_> it finds one from DNS< but it if fails it doesn't look again.
[10:45] <bluefoxicy> Mithrandir:  why sohuldn't passwords be in ldap
[10:45] <wasabi_> Nothing wrong with storing passwords in LDAP imo... Just shouldn't use LDAP for actual auth. ;0
[10:46] <wasabi_> krb5 is there. ;)
[10:46] <bluefoxicy> wasabi_:  ldap over TLS
[10:46] <wasabi_> Not much need for that with krb5.
[10:46] <wasabi_> Which provides channel encryption on it's own.
[10:46] <bluefoxicy> KRB5 does what
[10:46] <wasabi_> kerberos
[10:46] <bluefoxicy> I know
[10:47] <bluefoxicy> TLS does certificated authentication with X.509 certificates verifiable against a certificate authority
[10:47] <wasabi_> Ticket based auth. It's what AD uses.
[10:47] <bluefoxicy> ticket based auth?
[10:47] <wasabi_> A user logs in by asking kerberos to send them a ticket.
[10:47] <wasabi_> They use that ticket to access subsequent services, without being reasked for password.
[10:47] <bluefoxicy> ah
[10:47] <wasabi_> single sign-on, etc.
[10:47] <bluefoxicy> you could LDAP with KRB5 as a backend
[10:47] <wasabi_> For instance, reading the user list from LDAP should be a privledged operation.
[10:48] <bluefoxicy> err.
[10:48] <bluefoxicy> other way
[10:48] <bluefoxicy> KRB5 with LDAP as a backend
[10:48] <wasabi_> Sure, but mit krb5 doesn't do that well, and heimdal only does that with samba built extensions.
[10:48] <Mithrandir> wasabi_: uh, how can that be a priviledged operation?
[10:48] <bluefoxicy> so security wise they're equivalent, but functionality wise KRB5 allows for single sign-on?
[10:48] <wasabi_> Mithrandir: You log onto your system with a kerberos password and get a ticket. 
[10:48] <wasabi_> You use that ticket to list users later.
[10:49] <wasabi_> LDAP shouldn't allow non-ticketed connections, ideally.
[10:49] <wasabi_> in AD it doesn't, etc.
[10:49] <Mithrandir> wasabi_: stuff run from cron needs to be able to see who owns a file too.
[10:49] <wasabi_> Stuff run from cron should have a keytab.
[10:49] <wasabi_> as it does in windows.
[10:49] <bluefoxicy> Mithrandir:  easily, you remove NULL BASE dn listing, turn off anonymous LDAP, and set up each authenticating system to use a privileged cd,dn + password to access the data needed to authenticate?
[10:49] <wasabi_> Each machine in AD has a network user account. When the machine boots it logs in, just like a user.
[10:49] <wasabi_> And uses that account for network access.
[10:50] <bluefoxicy> Mithrandir:  Or just put krb5 in front of ldap, either or.
[10:50] <wasabi_> It also changes it's own password randomlly, etc.
[10:50] <bluefoxicy> wasabi_:  does krb5 allow other information to be stored, as ldap does?
[10:50] <wasabi_> cron jobs that run as root should be able to access LDAP using the computer account.
[10:50] <wasabi_> No.
[10:50] <Mithrandir> bluefoxicy: congratulations, you just broke ls -l if your network is down. :-P
[10:50] <wasabi_> Mithrandir: more underlying stuff that needs to be fixed.
[10:50] <bluefoxicy> wasabi_:  public keys, roaming profiles..?
[10:50] <wasabi_> Mithrandir: need a client cache in front of NSS.
[10:50] <wasabi_> Or behind it. ;)
[10:50] <Mithrandir> uh, what information is it that you store in ldap which is secret?
[10:50] <bluefoxicy> Mithrandir:  if you're already logged in ls should work o.o
[10:50] <wasabi_> nscd, if it didn't suck horribly, would be it.
[10:51] <bluefoxicy> Mithrandir:  user names are secret.
[10:51] <wasabi_> Mithrandir: the user's themselves.
[10:51] <wasabi_> A random computer sitting on my network should not be able to access that.
[10:51] <bluefoxicy> Mithrandir:  haven't you noticed using a bad user or password gives "bad username or password"?
[10:51] <Mithrandir> bluefoxicy: no, they're not.  User names are usually trivially discoverable.
[10:51] <bluefoxicy> not "That user doesn't exist" or "bad password"?
[10:51] <wasabi_> Mithrandir: Might also be a requirement of corporate/gov policy.
[10:51] <bluefoxicy> Mithrandir:  yeah, by hitting www.foo.com/~user with various users
[10:52] <wasabi_> Mithrandir: for example, AD doesn't allow non kerberos authetnicated connections, by default.
[10:52] <wasabi_> So somebody can't walk in with a laptop and access the user list.
[10:52] <wasabi_> They actually have to hack something.
[10:52] <Mithrandir> wasabi_: that doesn't say anything about why being able to list the users on a system which you have authenticated to is bad.
[10:52] <wasabi_> It's not. I didn't say that.
[10:52] <wasabi_> If you've authenticated, you can list the users.
[10:52] <wasabi_> But you must be authenticated.
[10:52] <bluefoxicy> Mithrandir:  being able to list users on a system you don't have access to.
[10:53] <wasabi_> being able to walk onto the company premisis and plugin a laptop and access the company user list is bad.
[10:53] <bluefoxicy> Mithrandir:  I want to hack Accounting, I'm pretty sure Larry's password is something about his dog spot, probably spotty1 or such
[10:53] <wasabi_> That's all.
[10:53] <wasabi_> Some privacy laws even protect against it.
[10:54] <bluefoxicy> Mithrandir:  Either A) I don't work for the firm; or B) I can get the users for my area, but not for the accounting network, which stores its data in a separate privilege area.
[10:54] <bluefoxicy> Mithrandir:  So I have to guess what Larry's accounti s first, blindly.
[10:54] <wasabi_> too many cooks.
[10:54] <bluefoxicy> Or, I just dump a list of users, notice lwall06, figure out this is Larry, and try 2 attempts and get in.
[10:55] <wasabi_> Blue, that doesn't matter. You need to be doing so from a company system.
[10:55] <lamont> pitti: ping
[10:55] <wasabi_> It needs to be auditable.
[10:55] <wasabi_> Mithrandir: they aren't, but they are secret.
[10:55] <wasabi_> Mithrandir: ie you shouldn't be able to get a big list of user names unless you're allowed to.
[10:55] <bluefoxicy> confidential, perhaps not secret.
[10:56] <wasabi_> Sure.
[10:56] <wasabi_> Also it needs to be auditable. The LDAP server needs to be able to record what human being requested the list, so they can be investigated.
[10:56] <bluefoxicy> wasabi_:  I do not understand how kerberos auth is supposed to work, though.  Where is the ticket stored?
[10:56] <wasabi_> bluefoxicy: local machine.
[10:57] <bluefoxicy> wasabi I  mean Icould set up krb5 here right? I have an ldap server behind me
[10:57] <wasabi_> Yup.
[10:57] <bluefoxicy> and I could rig pam to use krb5 auth right?
[10:57] <wasabi_> I have it set up at home, because I do this a lot.
[10:57] <bluefoxicy> but what would that get me
[10:57] <wasabi_> All my computers, laptops included, use pam_krb5 and libnss.
[10:57] <bluefoxicy> besides pam logging in with krb5 auth?
[10:57] <wasabi_> Get you?
[10:57] <bluefoxicy> I could try to modify stuff in the ldap directory
[10:57] <bluefoxicy> and it would probably not run to pam and say "You have a ticket right?  Lemme use it"
[10:57] <wasabi_> My needs are a) to test it, because I deploy it in companies. I need to know it works. b) single sign on.
[10:58] <bluefoxicy> explain single sign on
[10:58] <Mithrandir> hmm, I wonder if you could write a dhcp extension to tell the computer about its kerberos domain and such.
[10:58] <Mithrandir> or if somebody has already done that.
[10:58] <wasabi_> bluefoxicy: I log into my computer once, open my email client, and read my email. Then I browse to a corporate web page, and it says Hello Jerry.
[10:58] <wasabi_> Neither the email client nor the web page has a stored password.
[10:58] <bluefoxicy> wasabi I mean the underlying mechanics
[10:58] <bluefoxicy> I am sure if I log into my mail client it will be like "what?"
[10:59] <wasabi_> Evolution supports kerberos auth to IMAP.
[10:59] <wasabi_> Mozilla doesn't yet.
[10:59] <bluefoxicy> thunderbird doesn't though?
[10:59] <wasabi_> IE and Outlook both do.
[10:59] <wasabi_> I doubt it.
[10:59] <Chipzz> Mithrandir: not necessary
[10:59] <Chipzz> Mithrandir: kerberos can do realm discovery via dns
[11:00] <wasabi_> Mithrandir: In AD the machine has two pieces of info. It's domain name and it's machine name.
[11:00] <wasabi_> It uses the domain name to do service discovery in DNS.
[11:00] <wasabi_> And it's machine name to login.
[11:00] <wasabi_> Oh and it has a machine password. ;)
[11:00] <wasabi_> Which is basically random and very long.
[11:00] <bluefoxicy> wasabi_:  so the target service or PHP script has to have krb5 authentication built-in?
[11:01] <wasabi_> bluefoxicy: Yup. IIS supports it. Most internal windows apps are built that way.
[11:01] <Mithrandir> wasabi_: you don't need to try to teach me how kerberos works.
[11:01] <Ubugtu> Malone bug 38893 in gnome-screensaver "Screensaver Still NOT Working!! CRASH INSTANTLY!!" [Normal,Unconfirmed]  http://launchpad.net/bugs/38893
[11:01] <bluefoxicy> wasabi_:  mind, I don't quite get how to do krb5 over a web browser :)
[11:01] <wasabi_> Mithrandir: Well, you were asking for a DHCP extension thing... which is covered by those two.
[11:01] <torkel> bluefoxicy: google SPNEGO
[11:01] <wasabi_> bluefoxicy: SPENGO WWW-Authenticate header. MS invented it. Firefox implemented it.
[11:01] <wasabi_> SPNEGO I mean.
[11:02] <Mithrandir> bluefoxicy: you use another authentication mechanismn than the regular basic authentication.
[11:02] <bluefoxicy> sweet.
[11:02] <wasabi_> Firefox's implementation doesn't read from the ticket cache though.
[11:02] <wasabi_> At least last I checked.
[11:02] <Mithrandir> huh?  It sure does when I tested it a year ago.
[11:02] <Mithrandir> or rather, did.
[11:02] <bluefoxicy> wasabi_:  and krb5 can be used as an authenticaton server to IPSEC, TLS, and PPTP (fuck PPTP) authenticators?
[11:02] <wasabi_> Really? It asks me for my username and password.
[11:02] <wasabi_> It uses kerberos, but it still asks.
[11:03] <wasabi_> Thought it was a bit silly. ;0
[11:03] <wasabi_> bluefoxicy: yeah.
[11:03] <wasabi_> Err, wait, no.
[11:03] <wasabi_> PPTP sure. :0
[11:03] <bluefoxicy> oh
[11:03] <bluefoxicy> shit.
[11:03] <bluefoxicy> hmm.
[11:03] <wasabi_> AD also has services which auto generate certificates for each user.
[11:03] <LaserJock> Who decides the name of Dapper+1? sabdfl, TB, CC ?
[11:03] <bluefoxicy> LDAP will work for IPSEC and TLS?
[11:03] <wasabi_> For various purposes.
[11:03] <Mithrandir> bluefoxicy: pptp and ipsec both work on a host level, not on a user level, so it's slightly muddy when you try to do that.
[11:03] <wasabi_> LDAP doesn't work for it. LDAP stores the certs.
[11:04] <wasabi_> yeah and also that. MS figured it out though, mostly it's a hack.
[11:04] <sabdfl> LaserJock: me, with some help
[11:04] <mdke> LaserJock, some say that the name comes in a dream
[11:04] <bluefoxicy> wasabi_:  I am trying to find the best way to implement a network authentication system over here so I can help small businesses get their IT shops started for a small fee ;)
[11:05] <wasabi_> Cool.
[11:05] <wasabi_> I still use AD.
[11:05] <LaserJock> sabdfl: is there a name yet? I'm working on some shipped documentation (Ubuntu Packaging Guide) and since people should be packaging for Dapper+1 when they install Dapper (for the most part) it would be good if I could include the name
[11:05] <wasabi_> The pieces just aren't there to compare yet.
[11:05] <wasabi_> I mena, the pieces are, but nothing tying them together in any way usable by most small shop IT admins.
[11:05] <sabdfl> LaserJock: there's a proposal, not yet confirmed or announcd
[11:05] <bluefoxicy> wasabi_:  I'm thinking all open source with OSX and Windows able to use it for auth/roaming profiles
[11:05] <sabdfl> i expect it will be out tihs week or next
[11:05] <wasabi_> bluefoxicy: Windows cannot auth against non-Windows kerberos.
[11:06] <wasabi_> bluefoxicy: Samba guys are working hard on it, but not yet.
[11:06] <bluefoxicy> wasabi_:  red hat active directory server?
[11:06] <wasabi_> There's code where it works, but it's not released.
[11:06] <mdke> dholbach, y0
[11:06] <wasabi_> Red Hat Directory Server, not Active.
[11:06] <wasabi_> It's just LDAP.
[11:06] <dholbach> mdke: hm? :)
[11:06] <wasabi_> Windows requires the machine kerberos tickets to have a special piece of data in them called the PAC.
[11:06] <wasabi_> Which contains group membership and SID values.
[11:06] <LaserJock> sabdfl: ok thanks, will look forward to the announcement and try to make the necessary changes
[11:06] <mdke> dholbach, just branched ubuntu-docs to repos/branches/dapper
[11:07] <ogra> LaserJock, i think that justifies a string freeze exception :)
[11:07] <dholbach> mdke: ok, cool - good to know
[11:07] <wasabi_> They've sort of made it hard to implement with licensing restrictions.
[11:07] <Kamion> mdz: hmm, I think I'm happy with the wiki workflow for now for main inclusion
[11:07] <mdke> dholbach, I'm now busy rampaging through the source package to make it 1/10 of the size
[11:07] <pitti> lamont: pong
[11:07] <LaserJock> ogra: heck yeah! We love exceptions
[11:07] <mdke> dholbach, can you do an upload tomorrow?
[11:07] <dholbach> mdke: yeah, can do that
[11:07] <mdke> danke
[11:08] <dholbach> good night guys
[11:08] <mdke> night
[11:08] <mdz> night dholbach
[11:08] <bddebian> Gnight dholbach
[11:08] <Kamion> bluefoxicy: changing the kernel ABI is a lot easier now than it used to be, now that we've got it down to a fairly fine art; it's four uploads distributed across about three people, plus a seed change which I normally do
[11:08] <dholbach> night mdke, mdz, bddebian :)
[11:08] <Kamion> bluefoxicy: but that doesn't mean that we shouldn't be conservative about changing such a core part of our system just before a release
[11:09] <Kamion> bluefoxicy: the CD build process itself is entirely automated; what goes into it is not necessarily
[11:11] <bddebian> I've broken 11,000 just uploading .desktop files.  How lame :-)
[11:11] <bluefoxicy> Kamion:  of course, you want a wide testing buffer before changing the kernel out.
[11:11] <bddebian> Whoops, sorry, wrong window
[11:12] <bluefoxicy> Kamion:  Just I'd rather the wiki reflect reality, and not made-up excuses :)
[11:12] <tseng> < bluefoxicy> Kamion:  Just I'd rather the wiki reflect reality, and not  made-up excuses :)
[11:12] <tseng> erg stupid mouse
[11:12] <tseng>  /topic amature hour in #u-d
[11:12] <bluefoxicy> haha
[11:13] <bluefoxicy> tseng he just said it was fairly easy, the wiki says it's massively complicated
[11:13] <bluefoxicy> but whatever 
[11:14] <bluefoxicy> wasabi_:  so the next krb release from samba should auth windows?
[11:14] <tseng> fairly easy is relative
[11:14] <wasabi_> No. Samba doesn't release krb.
[11:14] <bluefoxicy> oh
[11:14] <wasabi_> They are working with Heimdal though.
[11:14] <bluefoxicy> you just said ... ah.
[11:14] <wasabi_> Either way, kerberos is just one part of authing with windows.
[11:15] <wasabi_> You really lose a lot of windows features when not using AD.
[11:15] <wasabi_> Enough to make AD worth it.
[11:15] <bluefoxicy> there isn't an AD server on linux though :(
[11:15] <wasabi_> Now, integrating AD into a Unix domain. That has promise.
[11:15] <wasabi_> ie so you can share kerberos keys and LDAP.
[11:15] <tseng> wasabi_: what is a "unix domain"
[11:15] <wasabi_> But windows systems still use AD abilities.
[11:15] <wasabi_> tseng: a mad up term for what we're talking about
[11:15] <wasabi_> made
[11:15] <seb128> Keybuk, Diziet: around?
[11:15] <Keybuk> I'm around
[11:15] <bluefoxicy> wasabi_:  I would rather be able to build a full open source solution but have it flexible enough to add Windows and OSX later.
[11:15] <Keybuk> what's up?
[11:16] <wasabi_> bluefoxicy: Then start coding for Samba.
[11:16] <seb128> Keybuk: 
[11:16] <wasabi_> There are massive amounts of work to be done still.
 epiphany-browser 2.14.0 ships a /usr/share/gnome/help/epiphany-browser folder
 epiphany-browser 2.14.1 ships a /usr/share/gnome/help/epiphany folder and a /usr/share/gnome/help/epiphany-browser symlink to
 to it
 after upgrade /usr/share/gnome/help/epiphany-browser is an empty folder, not a symlink :/
 that's a "classic" one, no?
[11:16] <Keybuk> yup
[11:16] <Keybuk> classic dpkg
[11:16] <seb128> k
[11:16] <bluefoxicy> wasabi_:  what else does AD do?  Application servering?
[11:16] <seb128> what is the usual way to workaround it? :)
[11:16] <Keybuk> can't replace a directory with a symlink
[11:16] <Keybuk> fix it in postinst
[11:16] <wasabi_> bluefoxicy: certificate authority, group policy, client side extensions to that.
[11:17] <Keybuk> if it's left as an empty directory, remove it and restore the symlink
[11:17] <seb128> Keybuk: like, if that's a dir, rm and redo the symlink?
[11:17] <Keybuk> yup
[11:17] <seb128> k, thank you
[11:17] <pitti> Keybuk: is that a bug or a feature?
[11:17] <bluefoxicy> wasabi_:  OpenLDAP can function as a CA AFAIK.  Group policy...?
[11:17] <wasabi_> openldap can store CA data, but not function as a CA.
[11:17] <bluefoxicy> oh
[11:17] <wasabi_> openldap is a ldap server.
[11:17] <Keybuk> pitti: feature
[11:17] <bluefoxicy> hm.
[11:18] <wasabi_> group policy are objects in AD which machines can apply to enable/disable features and change settings, network wide.
[11:18] <wasabi_> Such as automated software installs, hiding the run menu, changing hte home page.
[11:18] <wasabi_> To stuff such as distributing certificates.
[11:18] <bluefoxicy> oh, very windows specific then
[11:18] <mdke> seb128, did you see my mail ages ago about making epiphany take different default homepages, depending on locale? Any idea if it's possible somehow? np if not
[11:18] <wasabi_> Yup.
[11:18] <Burgwork> bluefoxicy, gpo are very powerful but also fairly arcane and a source of much headache
[11:18] <wasabi_> I think some PADL project actually made headway there.
[11:18] <seb128> mdke: yeah, sorry, it's on my pile of "to reply" mails, I'm catching up atm
[11:19] <wasabi_> http://www.padl.com/Articles/PADLReleasesXADIdentitySe.html
[11:19] <seb128> mdke: I've just replied to the "utch, no quality control from translations on rosetta" replies on the l10n french list
[11:19] <dieman> group policy is evil stuff
[11:19] <bluefoxicy> wasabi_:  okay so, is there a way to implement authentication through an open source server such as OLDAP or Heimdal KRB5, storing the roaming profiles on an open source samba server
[11:19] <mdke> seb128, no problem. I thought the answer might be a simple "no", so I thought I'd ping you. if the answer is "possibly yes", then awesome!
[11:19] <dieman> we've already got cfengine
[11:19] <dieman> really
[11:19] <wasabi_> bluefoxicy: Sure, Samba. ;)
[11:19] <bluefoxicy> wasabi_:  and then set up a Windows Active Directory Server to get its auth info from those sources?
[11:19] <wasabi_> Nope.
[11:19] <seb128> mdke: they forwarded your mail to discuss how to organize the team
[11:19] <mdke> seb128, in french, I guess?
[11:19] <dieman> cfengine can exact its policy from netgroups specified via LDAP
[11:20] <bluefoxicy> wasabi_:  so in other words I can't set up the network to be expanded later?
[11:20] <mdke> seb128, my french is rubbish
[11:20] <wasabi_> You can't do what you want to do right now.
[11:20] <wasabi_> And if you kill AD, you will lose windows features.
[11:20] <seb128> mdke: yeah, but nothing really useful. I pinged carlos today, my idea of an easy way would be "lock when somebody starts a translation, mail the diff when he unlocks it"
[11:20] <bluefoxicy> wasabi_:  what I want to do right now is set up small IT shops with full open source setups
[11:20] <wasabi_> bluefoxicy: then use Linux desktops.
[11:21] <dieman> you can do roming profiles on samba, at least in old school pdc mode.
[11:21] <mdke> seb128, sounds like a decent idea. 250 is a big group to handle :)
[11:21] <dieman> im not sure the extent of what samba doing ad can do
[11:21] <wasabi_> Yeah. You still do NT4 mode.
[11:21] <seb128> mdke: mail going to a list, so it's easy to read what the person did and to fix errors and notice if the guys does it wrongly
[11:21] <dieman> i thought the newer releases have reasonable ad capabilities
[11:21] <seb128> mdke: yeah :/
[11:21] <seb128> mdke: for epiphany I don't know but I'll ping upstream ;)
[11:22] <mdke> seb128, rocking. thanks dude
[11:22] <wasabi_> dieman: For joining to AD, not for hosting it.
[11:22] <dieman> wasabi_: ahh, ok
[11:22] <bluefoxicy> wasabi_:  I'm waiting for Ubuntu having the easy network authentication setup, like RHEL and Fedora do, because then I can use Ubuntu desktops ($250 for 1 year of support up to 10 incidents?) and RHEL servers ($2500 service contract???) in a mixed shop.
[11:22] <wasabi_> Linux can join to AD wonderfully.
[11:22] <wasabi_> CIFS support still sucks bad.
[11:22] <seb128> mdke: np
[11:22] <bluefoxicy> wasabi but AD costs too much money :)
[11:23] <bluefoxicy> if someone so desired they could go supportless fedora core/ubuntu desktop and CentOS/Ubuntu server.
[11:23] <pitti> Keybuk: what's the use case for that feature?
[11:23] <wasabi_> Well, I know of no open source/free alternative to AD which actually delivers on what AD does.
[11:23] <wasabi_> Which is massive ease of use, for a small shop.
[11:23] <wasabi_> And automatic configuration of each host.
[11:23] <bluefoxicy> wasabi a small shop may not have the $3000 for Win2k3 server
[11:23] <wasabi_> The pieces are there, but you set them up yourself.
[11:23] <bluefoxicy> they may have a good $100,000 start-up loan sitting around, but you don't want to blow a grand here and there.
[11:24] <wasabi_> I'd charge any shop more than $3000 for the man hours involved to set up a Linux LDAP/kerberos setup to serve windows.
[11:24] <wasabi_> Nuff said.
[11:24] <seb128> mdke: <chpe> seb128: the home page is stored in gconf, and gconf supports localised prefs, so yes :)
[11:24] <wasabi_> And win2k3 server is actually like $800.
[11:24] <mdke> seb128, wooo
[11:24] <wasabi_> Or get SBS. It's less. ;)
[11:24] <bluefoxicy> wasabi start-ups can start on full linux but I'd like to make an expansion path that allows for integrating OSX and Windows if desired.
[11:24] <wasabi_> Or help with the programming heh.
[11:25] <mdke> seb128, we've already started shipping translations (or links, where I haven't got translations yet), they are in /usr/share/ubuntu-artwork/home
[11:25] <bluefoxicy> nothing would be cooler though then setting up a small 5-10 PC linux shop for a small business and 4 years down the line they have 40,000 machines and no Windows in site.
[11:25] <dieman> omfg, the russians want to drive a gold plated golf ball of the ISS
[11:25] <dieman> off of
[11:25] <dieman> as a stunt
[11:25] <bluefoxicy> dieman so, they're russians, they're up there drinking wodka
[11:26] <dieman> oh, and im in the wrong chan for this :)
[11:26] <dieman> i thought i was in my idle chan
[11:26] <dieman> sorry about that :)
[11:26] <bluefoxicy> send them some female astronauts, they'll stop worrying about driving golf balls...
[11:27] <ogra> mdz, haha, see bug 22045  :-D
[11:27] <Ubugtu> Malone bug 22045 in gnome-screensaver "Poor performance / flickering with OpenGL hacks" [Normal,Confirmed]  http://launchpad.net/bugs/22045
[11:28] <bluefoxicy>  - turn 3D acceleration on
[11:28] <Chipzz> is it just me, or did X get a whole lot faster to start recently? :)
[11:28] <bluefoxicy> I heard X7 was going to have SE-X
[11:29] <bluefoxicy> wasabi_:  i am in general untrusting of microsoft's products.  I don't pretend to understand them, but I fear the defaults.  SQL server had default users and passwords that got owned by SqlSlammer.F ><
[11:30] <wasabi_> I'm not suggesting to use WIndows. I'm just telling you what it is and what it does, and what *I* have yet to find a replcement for.
[11:30] <wasabi_> And this has become offtopic. PM or something.
[11:30] <bluefoxicy> My idea of installing something is that anything that requires attention, such as administrative passwords, dangerous default settings (i.e. unencrypted connections?  OpenLDAP without tls!) are mandatory for you to set up
[11:30] <mdke> #ubuntu-offtopic?
[11:30] <bluefoxicy> that's a better idea.
[11:30] <bddebian> Please
[11:31] <seb128> mdke: k, noted. I'll think about the best way to do that and let you know
[11:32] <Surak> Kamion: espresso 0.99.44 is calling the systems $username+"-desktop" in the "who are you?" screen. however, it does not accept the dash. Should I open it? Did you already notice?
[11:32] <Surak> hello folks
[11:32] <mdke> seb128, thanks so much.
[11:32] <seb128> np, thank you for pinging me on the topic and for the translated pages ready to use :)
[11:33] <mdke> :)
[11:33] <mdke> Surak, it's very late evening Kamion's time, he might not be around. if not, maybe file a bug
[11:33] <ogra> grmbl ...
[11:35] <Surak> mdke: Yes. He tends to be awake working these days until late hours, for desperation of his wife :-)
[11:42] <pitti> so I wasted TWO HOURS with diff -Nur, tar, dpkg -P and sudo make install/uninstall just to discover that a simple dangling stupid symlink broke cups *grrrr*
[11:52] <Keybuk> pitti: :(
[11:52] <Keybuk> pitti: I'm having one of those "I *know* this worked once, because I tested it; so why is it flat-out not working now?" moments
[11:53] <pitti> Keybuk: I wish diff -Nur would consider symlinks :/
[11:53] <Keybuk> pitti: I tend to use comm on the file lists
[11:53] <pitti> Keybuk: what broke for you?
[11:53] <Keybuk> pitti: if you rename eth1 to "foo", it still calls "ifup eth1"
[11:54] <pitti> 'it' being?
[11:54] <Keybuk> udev
[11:54] <pitti> oh, udev ifup's stuff?
[11:54] <Keybuk> this is why it sometimes hangs on "Configuring network interfaces"
[11:54] <Keybuk> both devices call "ifup eth0", leaving the "ifup -a" to do eth1 later
[11:54] <Keybuk> I know it damned well worked when I wrote the patch
[11:54] <Keybuk> so something in udev changed since, and we didn't notice (other than the occasional "hangs for a minute at ifup -a" bugs)
[11:55] <pitti> hm, I never encountered that one
[11:55] <Keybuk> yeah, it's a hard one to spot ;)  most people have eth0 and eth1, and both on auto, and it looks like it works :p  it doesn't matter that the wrong event is ifuping the wrong interface <g>
[11:56] <Keybuk> but is a *bad* bug
[11:56] <Keybuk> I know what's causing it now, fortunately
[11:56] <Keybuk> the $env{INTERFACE} in the RUN rule is expanded *before* the device is renamed
[11:56] <Keybuk> so there must have been a change in the way that works
[11:57] <Keybuk> and then, after this, I have to work out why "blacklist pcspkr" doesn't