[12:36] <dudanogueira> hello there! does anyone knows about the ekiga sound quality bug on edgy?
[12:42] <Burgwork> dudanogueira: this is not a user support channel, nor a bug tracker. Have you looked in malone?
[12:43] <dudanogueira> Burgwork, yes, and saw a related post on ekiga mailing list, a confirmed bug on malone...
[12:43] <dudanogueira> http://www.mail-archive.com/ekiga-list@gnome.org/msg00761.html
[12:43] <dudanogueira> https://launchpad.net/distros/ubuntu/+source/ekiga/+bug/30083
[12:43] <Ubugtu> Malone bug 30083 in ekiga "Sound quality horrible" [Medium,Confirmed]  
[12:43] <dudanogueira> cool =)
[12:44] <Burgwork> dudanogueira: add that information to the bug report
[12:44] <dudanogueira> on the mailing list it says that ekiga 2.0.4 will fix this bug...
[12:45] <dudanogueira> "I am waiting that the Ubuntu guys sort #359655 before releasing 2.0.4,
[12:45] <dudanogueira> but I think I won't wait a long time anymore because they seem unable to
[12:45] <dudanogueira> test if the proposed patch fixes the problem. Ah well, Ubuntu Edgy...
[12:45] <dudanogueira> Our daily nightmare."
[02:16] <jdong> cjwatson: can you estimate when the next archive day will be?
[03:06] <jdong> what's ascii code 09
[03:06] <jdong> and what's it doing in xchat-gnome's debian/control
[03:06] <HrdwrBoB> man ascii
[03:06] <jdong> 000003c0  2d 64 65 76 20 28 3e 3d  09 32 2e 31 36 2e 30 29  |-dev (>=.2.16.0)|
[03:06] <HrdwrBoB>        011   9     09    HT  \t (horizontal tab)   111   73    49    I
[03:06] <jdong> HrdwrBoB: that was sarcastic
[03:06] <jdong> there is a tab instead of a space in debian/control
[03:06] <jdong> which causes pbuilder-satisfydepends to bork out
[04:55] <fabbione> morning
[05:05] <Hobbsee> hey fabbione!
[05:05] <fabbione> hi Hobbsee 
[05:30] <LaserJock> fabbione: do you have any time for a spec review?
[05:31] <fabbione> LaserJock: i guess i can find the time...
[05:32] <LaserJock> fabbione: if you get the chance: https://blueprints.launchpad.net/distros/ubuntu/+spec/edubuntu-menus-completion
[05:32] <LaserJock> fabbione: it should be pretty simple
[05:33] <fabbione> LaserJock: ok.. looking at it
[05:38] <fabbione> LaserJock: pending approval
[05:39] <LaserJock> fabbione: thank you very much
[06:48] <towsonu2003> I usually don't post the same message to all the chat rooms I'm in, but you've gotta see this: http://en.wikinews.org/wiki/Kurdish_operating_system_under_investigation_by_Turkish_attorney_general
[06:48] <towsonu2003> Kurdish operating system in that link is... Ubuntu.
[06:50] <fabbione> towsonu2003: i see no problem in that but thanks for the info
[06:51] <towsonu2003> you're welcome
[06:51] <fabbione> Ubuntu can be "Multilingual" etc. etc. etc.
[06:51] <fabbione> it's just a matter of default install
[06:51] <towsonu2003> crimsun didn't like me posting that here, but it's kinda related to development I figured
[06:52] <fabbione> no it's not related to development and this isn't the right forum, but well you did it 
[06:52] <fabbione> it can't be undone
[06:52] <towsonu2003> fabbione, the fact that it could be installed in Kurdish by default was a big step for Turkey
[06:52] <towsonu2003> fabbione, sorry
[06:53] <fabbione> towsonu2003: i don't agree or disagree with whatever they did.. it's their choise of freedom :)
[06:53] <fabbione> i am just saying that English can always be there anyway
[06:53] <towsonu2003> I'm not sure I understand what you're saying
[06:53] <towsonu2003> :)
[06:54] <fabbione> i mean that i am not taking a position on their decision.
[06:54] <fabbione> and that's their choise (freedom) to do whatever they think is right.. like for example using a different language by default
[06:55] <towsonu2003> I feel like there is some misunderstanding between us :)
[06:55] <fabbione> please explain
[06:55] <towsonu2003> my point in posting this was to let ppl know that Ubuntu is under investigation ;)
[06:56] <towsonu2003> the "Kurdish operating system" title is to appeal to wikinews readers not familiar with ubuntu
[06:56] <fabbione> towsonu2003: and i am telling you that there is not much to investigate. their looking at a derivate
[06:56] <towsonu2003> ah okay
[06:57] <fabbione> and it's a decision of the derivate to include extra languages like English
[06:57] <fabbione> as Ubuntu we do provide all of the above
[06:57] <fabbione> so it makes the case somehow mutt
[06:57] <fabbione> if you understand what i mean
[06:58] <towsonu2003> yes
[06:58] <towsonu2003> though
[06:59] <towsonu2003> the translation team might get into trouble...
[06:59] <towsonu2003> and for no real reason
[06:59] <towsonu2003> as you said. 
[07:04] <fabbione> sfllaw: ping?
[07:44] <Burgundavia> fabbione: is now 2am there
[07:46] <fabbione> Burgundavia: sleeping is for weak
[07:48] <dholbach> good morning
[07:48] <Burgundavia> fabbione: quite possibly
[07:48] <Burgundavia> hey dholbach
[07:48] <dholbach> hi Burgundavia
[08:34] <siretart> keescook: glad I could help. nice setup script! :)
[08:36] <ajmitch> hey siretart 
[08:36] <Burgundavia> hey siretart
[08:37] <siretart> morning, ajmitch. hi Burgundavia!
[08:49] <mvo> Mithrandir: could you please have a look at my https://blueprints.launchpad.net/distros/ubuntu/+spec/dynamic-mirror-decisions spec? you are the approver
[08:52] <Burgundavia> Mithrandir: quick marketing question: when is the final decision on specs that are going in Feisty, because I want to finish preparing that LookingForwardAtFeisty doc
[08:54] <sivang> morning!
[08:55] <Mithrandir> Burgundavia: is the questions when we'll stop accepting new specs or stop dropping unimplemented spec?
[08:55] <Mithrandir> +s
[08:55] <Burgundavia> Mithrandir: the former
[08:55] <Burgundavia> this is a speculative doc, after all
[08:55] <Burgundavia> planned to be finished by this weekend at the latest
[08:55] <Mithrandir> iirc, the deadline for having specs approved is tomorrow.
[08:55] <Burgundavia> ok
[08:59] <Mithrandir> mvo: you don't have any information on handling upgrades?
[09:00] <mvo> Mithrandir: in the spec? no, that is true
[09:00] <Mithrandir> mvo: as in, upgrading sources.list.
[09:01] <ajmitch> hi Mithrandir, pitti 
[09:01] <mvo> Mithrandir: right. do you want me to add it?
[09:01] <ajmitch> & mvo :)
[09:01] <mvo> hey ajmitch
[09:01] <mvo> good morning pitti
[09:02] <pitti> Good morning
[09:04] <sivang> morning mvo , ajmitch , pitti 
[09:04] <sivang> is Keybuk around today?
[09:05] <mnepton> a bit early for him
[09:06] <mdke> he's often around at this time, I've seen
[09:06] <mdke> but, he's clearly not in the channel, so, no
[09:12] <Mithrandir> mvo: it'd be useful to have at least a mention of it, even if you think it's out of scope.
[09:12] <mvo> Mithrandir: ok
[09:16] <Mithrandir> pitti: could you process the cmake main inclusion request?  It's needed for cdrkit to build which is needed to get *-desktop installable which is needed for herd 1. :-)
[09:17] <pitti> Mithrandir: yup, just saw it in my mbox, will do soonish (after finishing debugging avahi with slomo)
[09:17] <Mithrandir> pitti: ook, fine.  Thanks.
[09:22] <mvo> fabbione: could you please have a look at #68888? is this really a duplicate? 
[09:23] <mvo> Mithrandir: I added the sources.list conversition bit
[09:23] <fabbione> mvo: i dunno.. sombody did mark it as duplicate
[09:24] <fabbione> mvo: yes it is yet another
[09:24] <fabbione> mvo: i am working on an SRU
[09:24] <mvo> fabbione: for this particular bug? great! 
[09:24] <sivang> Mithrandir: cdrkit is part of u-d now?
[09:24] <Lathiat> pitti: hrm avahi bug?
[09:25] <fabbione> mvo: well given that there are a few tons of duplicates i have no other solution other than fixing it or run away screaming
[09:25] <mvo> :)
[09:25] <pitti> Lathiat: bug in our detection of unicast .local domains
[09:25] <Lathiat> ah right
[09:25] <pitti> Lathiat: boiling down to 'host' not always returning a non-0 exit code on errors *headdesk*
[09:26] <Lathiat> pitti: heh
[09:26] <pitti> yeah, checking exit codes would be too easy, let's have people parse i18n'ed output
[09:27] <Lathiat> try lookup $HOSTNAME.local or?
[09:27] <pitti> host -t soa local.
[09:27] <HrdwrBoB> er
[09:27] <Lathiat> ah right
[09:27] <Lathiat> i guess that makes sense
[09:27] <Lathiat> hrm
[09:27] <Lathiat> that said
[09:27] <Lathiat> we use 'office.local' here
[09:27] <pitti> Mithrandir: ^ is that halfway accurate?
[09:28] <Lathiat> that would fail that test
[09:28] <pitti> Lathiat: what does the command show for you?
[09:29] <Lathiat> http://bur.st/~lathiat/local.txt
[09:29] <Lathiat> im asking 192.168.10.2 directly cus i actually dont use our internal dns servers on this machine, but the rest of our network uses that
[09:29] <pitti> Lathiat: can you please add the respective exit codes to that?
[09:29] <Lathiat> sure
[09:30] <pitti> Lathiat: (and without specifying the DNS server)
[09:30] <pitti> Lathiat: many thanks
[09:30] <Keybuk> pitti: I didn't realise we were frozen yet/
[09:30] <Lathiat> http://bur.st/~lathiat/local2.txt
[09:30] <Keybuk> there's been no announcement to the ML
[09:31] <pitti> Keybuk: don't shoot the messenger
[09:31] <Lathiat> pitti: i guess its hard to detect $blah.local where .local itself isnt setup properly
[09:31] <pitti> Lathiat: darn; this is way more complicated than I thought
[09:31] <Lathiat> not sure how common that case is
[09:32] <Lathiat> i guess detecting .local is a good start
[09:32] <Lathiat> perhaps check the search domain and see if that ends in .local and if so -t soa that
[09:32] <Lathiat> i mean i guess its impossible to be completely fool proof about this
[09:32] <Lathiat> (thats a windows DNS server, if that maters)
[09:32] <pitti> Lathiat: did you happen to have nss-mdns running with that?
[09:32] <Lathiat> hrm actually
[09:32] <Lathiat> i do
[09:32] <pitti> Lathiat: I wonder whether avahi DTRT and fails for $foo.office.local
[09:33] <Lathiat> host doesnt use NSS tho right?
[09:33] <pitti> Lathiat: I'm 90% sure
[09:34] <Lathiat> my nss line is hosts:          files dns mdns
[09:34] <pitti> Lathiat: ah, that would shadow it
[09:34] <Lathiat> i can ping both dargo.local (mdns) and central.office.local (regular dns) now
[09:34] <pitti> good
[09:35] <Lathiat> so perhaps nss-mdns doesnt interfere with .local sub-domains?
[09:35] <Lathiat> only <x>.local
[09:35] <Lathiat> in which case it wouldnt matter
[09:35] <Lathiat> also i need to bat lennart aroudn the head for turning off ipv6 by default everywhere
[09:35] <pitti> Lathiat: slomo_ just told me that avahi-resolve-host-name would give a timeout for foo.office.local
[09:36] <Lathiat> correct
[09:36] <Lathiat> however getent hosts works
[09:36] <pitti> Lathiat: did you notice any long delay for ping central.office.local?
[09:36] <Lathiat> nope instant
[09:36] <Lathiat> ditto for getent
[09:36] <Lathiat> altho is that cached somewhere?
[09:36] <pitti> great
[09:36] <_ion> I've actually been thinking i'll report a wishlist bug about use-ipv6 being off by default (at least the last time i installed avahi).
[09:36] <Lathiat> _ion: nod
[09:36] <Lathiat> _ion: i was thinking of changing it
[09:37] <Keybuk> pitti: one thing I noticed, btw; my nsswitch.conf wasn't converted to the "correct" format
[09:38] <Lathiat> _ion: main annoyance is it causes 'duplicate' local services, but thats partly good because people will implement proper duplicate service detection
[09:38] <pitti> Keybuk: what did you have before? just 'files dns'?
[09:38] <_ion> lathiat: Yeah, turning it off is the wrong "fix" for that problem.
[09:39] <Keybuk> pitti: files dns mdns
[09:39] <Keybuk> (the previous default)
[09:39] <Lathiat> the other problem is that most apps wont correctly handle link-ocal addresses
[09:39] <pitti> Keybuk: ah, current code doesn't touch nsswitch.conf if it already configures mdns
[09:39] <pitti> Keybuk: 'previous default' -> from where?
[09:40] <pitti> Keybuk: edgy's libnss-mdns didn't configure nsswitch at all for me
[09:40] <Keybuk> pitti: base-files, I imagine
[09:40] <pitti> i. e. even after installing everything I still had 'files dns'
[09:40] <pitti> Keybuk: ok, if you are 100% sure that you didn't touch that manually, we should fix that automatically
[09:41] <Keybuk> pitti: definitly sure
[09:41] <Keybuk> my server has that same line too
[09:41] <Keybuk> which I would definitly not have configured avahi on :p
[09:41] <Lathiat> WHY NOT? :)
[09:41] <pitti> ok, since that's actively wrong, we should fix that
[09:41] <Keybuk> Lathiat: no reason to
[09:42] <Lathiat> Keybuk: blasphemy :)
[09:42] <jdub> cf. lathait's screenshot of rhel
[09:42] <hunger> With avahi turned on now: Are you planning to have services like SSH announced automatically in feisty?
[09:42] <Keybuk> pitti: ah, that's how it was shipped in dapper
[09:42] <Keybuk> nsswitch.conf was a conf-file from base-files, with that line in it
[09:42] <pitti> Keybuk: ouch
[09:43] <pitti> 'k, will fix
[09:43] <Mithrandir> pitti: (topic) looks accurate, yes.  Thanks.
[09:44] <Keybuk> Mithrandir: could you mail devel-announce to make sure everybody's aware
[09:44] <Mithrandir> Keybuk: yes, I'm going to.
[09:44] <pitti> Mithrandir: is that a hard freeze already? or just a soft one to control what's going in?
[09:44] <jordi> pitti: so the gtetrinet version in edgy crashes the first time you launch it.
[09:45] <Mithrandir> pitti: if people have stuff they want in herd 1 and which isn't too intrusive, I won't object very loudly.
[09:45] <jordi> pitti: that, mixed with the new bug buddy which makes it very easy to send dupe bugs, its frying me with reports.
[09:45] <jordi> pitti: what can I do to get gtetrinet patched in edgy?
[09:45] <pitti> jordi: first, we should add a bug pattern for that to avoid more duplicates
[09:46] <pitti> jordi: second, https://wiki.ubuntu.com/StableReleaseUpdates
[09:46] <jdub> Keybuk: will publish-addresses be on by default?
[09:46] <pitti> jordi: can you please point me to two different bugs about that, and I'll write a pattern for you?
[09:46] <Keybuk> jdub: ?
[09:47] <jordi> pitti: I'm talking about gnome bugzilla
[09:47] <jdub> Keybuk: avahi setting
[09:47] <Keybuk> jdub: what's that setting do?
[09:47] <jdub> publishes <hostname>.local
[09:47] <Keybuk> oh
[09:47] <Keybuk> yes
[09:47] <Keybuk> that's on by default
[09:47] <jordi> pitti: http://bugzilla.gnome.org/show_bug.cgi?id=377958
[09:47] <jdub> it was commented in edgy
[09:47] <Ubugtu> Gnome bug 377958 in general "crash in GTetrinet: tried to start gtetrinet..." [Critical,Resolved: duplicate]  
[09:47] <jdub> afaict
[09:47] <pitti> jordi: oh, right; I'm not sure whether SRU applies fully to universe as well
[09:48] <jordi> pitti: http://bugzilla.gnome.org/show_bug.cgi?id=361961
[09:48] <jordi> ah, right.
[09:48] <Ubugtu> Gnome bug 361961 in general "crash in GTetrinet: je voulais ouvrir ce jeu" [Critical,Resolved: duplicate]  
[09:48] <jordi> It's a double free
[09:48] <jordi> the patch is totally trivial
[09:48] <pitti> jordi: sorry, can't do an apport bug pattern for bug-buddy
[09:48] <jordi> it only crashes the first time you run it, subsequent runs are good
[09:48] <jordi> pitti: I guess.
[09:48] <jdub> Keybuk: apparently yes is the default (thus the comment)
[09:48] <jdub> Keybuk: perhaps my avahi-daemon was wedged.
[09:49] <Keybuk> jdub: we decided that one's mac address, IP address and hostname are so tediously trivial, and sufficiently difficult to keep secret and interact with other services, that it wasn't "private data"
[09:49] <jdub> yeah
[09:49] <jordi> pitti: where do I find out if I can update universe?
[09:50] <pitti> dholbach: does SRU apply to universe as well, or are the rules less strict?
[09:50] <jordi> I found it
[09:50] <niktaris> cjwatson, hi, I get reports that ubiquity freezes if you try to change the default timezone in 6.10. Is that a known issue ?
[09:51] <mvo> Mithrandir: I'm looking over upgrade issues right now and some people report problems with the edgy mount-by-UUID mechanism. is it ok if I subscribe you to these bugs?
[09:51] <jordi> ha, dholbach!
[09:51] <jordi> dholbach: dude, help me out on this :)
[09:52] <Mithrandir> mvo: not a problem, but why would you do that?  I don't do stable.  :-)
[09:52] <mvo> Mithrandir: right. who does stable? 
[09:52] <mvo> ...
[09:52] <mnepton> mount by UUID is going to be a giant PITA for this office.
[09:52] <Mithrandir> mvo: as in SRUs?  mdz, cjwatson
[09:52] <Mithrandir> mnepton: why?
[09:53] <mvo> Mithrandir: well, someone has to check and fix those problems first before we can do a SRU. 
[09:53] <mnepton> walking new users through fstab changes or modifications will be horrifying
[09:53] <Lathiat> pitti: http://wiki.ubuntu.com/MOTU/SRU ?
[09:53] <Lathiat> hrm 
[09:54] <Lathiat> my firefox dns broke with that nsswitch line
[09:54] <Lathiat> the recommended one
[09:54] <Mithrandir> mvo: mount-by-UUID is Scott's baby, so maybe he'd be a better fit?
[09:54] <mvo> Mithrandir: I'm sorry. you are listed as the assignee that I why I assumed you are the right person to talk to
[09:54] <Lathiat> hrm broken my dns completely
[09:55] <Mithrandir> mvo: assignee on which spec?
[09:55] <mvo> Mithrandir: probe-for-root-filesystem
[09:56] <pitti> jordi: ok, just follow SRU, MOTU/SRU is only proposed and looks very similar
[09:56] <mvo> Keybuk: hello! I'm looking over upgrade issues right now and some people report problems with the edgy mount-by-UUID mechanism. is it ok if I subscribe you to these bugs?
[09:56] <pitti> Lathiat: ugh, for .local or for everything?
[09:56] <Keybuk> mnepton: walking people through "my hard drive no longer mounts" will be more horrifying
[09:56] <Keybuk> mvo: sure
[09:56] <Lathiat> pitti: for everything but local
[09:56] <Keybuk> I should be subscribed to them already, no?
[09:56] <Lathiat> pitti: and i have no search local
[09:57] <pitti> Lathiat: and after stopping avahi it starts working again, I assume?
[09:57] <Lathiat> hrm it seems to work as root
[09:57] <Keybuk> mvo: SRU of what?
[09:57] <mnepton> Keybuk: yeah. thanks. (bitch.)
[09:57] <mnepton> :P
[09:57] <jdub> seb128!
[09:57] <Keybuk> mnepton: "why is my hard drive called /dev/sdc1 now?  it used to be /dev/hda1"
[09:57] <mvo> Keybuk: it seems like some of the problems are not filed to the right package/person. I'm currently looking over that
[09:58] <Lathiat> pitti: hrm, nope, but it works as root
[09:58] <seb128> hey jdub
[09:58] <Lathiat> pitti: wonder if i have some permission weirdness somewhere
[09:58] <pitti> Lathiat: whoa, it doesn't work with avahi stopped?
[09:58] <mvo> Keybuk: no SRU (yet :), that was me confusing the probe-for-root-fs with the mount-by-uuid spec 
[09:58] <Lathiat> pitti: ya
[09:58] <Lathiat> i have a compiled install of avahi on thsi machine
[09:58] <Lathiat> i wonder if its getting in the way
[09:58] <Lathiat> (as well as a package)
[09:59] <pitti> Lathiat: hm, did you happen to leave resolv.conf/nssswitch behind with 0600 permissions or so?
[09:59] <Keybuk> mvo: there isn't a formal mount-by-uuid spec, iirc
[09:59] <Lathiat> nope
[09:59] <jdub> it was a drive-by deep-system "enhancement" ;-)
[09:59] <Keybuk> mvo: probe-for-root-filesystem was the original spec that never really got implemented other than the initramfs changes
[10:00] <Keybuk> which tollef did for dapper
[10:00] <mvo> Keybuk: ok, that explains my confusion then I guess
[10:00] <Keybuk> libata-for-all-ata-disks kinda dragged it back out of the depths to be "finished"
[10:00] <Lathiat> whats a simple command line tool that will do an nss-lookup and isnt setuid?
[10:00] <Keybuk> I think it's the libata one that specifies the migration
[10:00] <Lathiat> hrm getent i guess
[10:01] <Lathiat> except that works
[10:01] <Lathiat> but firefox and gaim do not
[10:02] <Lathiat> pitti: ohhhh
[10:02] <Lathiat> pitti: i have 0.7 on this machine
[10:02] <Lathiat> let me install 0.8
[10:03] <Keybuk> mvo: what are the bugs?
[10:03] <Keybuk> I know of one, that complex smb mounts aren't migrated properly
[10:03] <Lathiat> pitti: that line completely breaks if nss-mdns isnt installed, so its a bad default i guess
[10:03] <Keybuk> but then I think the same guy previously filed a bug saying those same mounts aren't even mounted in the frist place
[10:03] <Keybuk> so I suspect that's a bogus bug :)
[10:03] <Lathiat> pitti: so NOTFOUND=RETURN works beacause it cant find the mdns4_minimal library, i think
[10:03] <pitti> Lathiat: simple tool> finger?
[10:03] <Keybuk> (or, at least, a bug affecting anywhere we parse fstab)
[10:03] <Lathiat> pitti: yeh i ended up using telnet
[10:04] <jordi> pitti: righto
[10:04] <Lathiat> strace shows it looks everywhere for mdns4_minimal and fails (which doesnt exist in this version)
[10:04] <Keybuk> there's also the "every time I run mkswap, the uuid changes, and I have to edit fstab/resume.conf/etc." one
[10:04] <pitti> Lathiat: libnss-mdns' postinst changes that line on installation and reverts it on removal
[10:04] <pitti> Lathiat: so I guess that should be fine
[10:04] <Lathiat> i guess
[10:04] <Lathiat> would also break dns if /usr is not yet mounted ?
[10:05] <Lathiat> if /usr is on nfs could be bad?
[10:05] <pitti> Lathiat: I'll try (by uninstalling libnss-mdns and setting the nsswitch line manually)
[10:06] <pitti> Lathiat: if so, we need libnss-mdns to install its libs to /lib, I guess
[10:06] <Lathiat> pitti: well i have no nss-mdns installed now and with that line all dns lookups fail
[10:06] <Lathiat> pitti: and yes i guess that would make sense then
[10:06] <Lathiat> arguably NSS could diferentiate between 'plugin said not found' and 'i couldnt find the plugin'
[10:06] <sivang> Keybuk: do you normally stick to dhcdbd version in debian or do you sometime merge and update to a new upstream (from redhat, i.e.) ?
[10:06] <Lathiat> i wonder if theres a better state for that?
[10:06] <Keybuk> sivang: Debian
[10:07] <pitti> Lathiat: I agree
[10:07] <pitti> Lathiat: maybe that's 'unavail'?
[10:07] <pitti>        unavail
[10:07] <pitti>               The  service  is  permanently  unavailable.   This  can  either mean the needed file is not available, or, for DNS, the server is not available or does not allow
[10:07] <pitti>               queries.  The default action is continue.
[10:07] <pitti> Lathiat: ^ according to that, the library missing should work
[10:08] <mvo> Keybuk: I subscribed you to them, bug #71217, bug #72016 and bug #72738
[10:08] <Ubugtu> Malone bug 71217 in Ubuntu "bad uuid reference in /etc/fstab" [Undecided,Unconfirmed]  http://launchpad.net/bugs/71217
[10:08] <Ubugtu> Malone bug 72016 in Ubuntu "Upgrading from 6.06 to 6.10 --> misconfigured swap partition entry in fstab" [Undecided,Unconfirmed]  http://launchpad.net/bugs/72016
[10:08] <Ubugtu> Malone bug 72738 in e2fsprogs "findfs reports /dev/sda1 instead of /dev/md0" [Undecided,Unconfirmed]  http://launchpad.net/bugs/72738
[10:08] <mvo> Keybuk: maybe more to come I'm not even close to catchup with my outstanding bugmail
[10:09] <Lathiat> pitti: is that like UNAVAIL=return or?
[10:10] <pitti> Lathiat: the doc says the default would be unavail=continue
[10:10] <Lathiat> (sorry I don't know much about nss..)
[10:10] <Lathiat> as opposed to notfound?
[10:10] <pitti> Lathiat: I'll check this out
[10:10] <Keybuk> oh, yeah, the NTFS thing is known
[10:10] <Lathiat> pitti: ok thanks
[10:11] <Keybuk> swap one is almost certainly that they've run mkswap
[10:11] <dholbach> jordi: ha!
[10:11] <dholbach> pitti: yes, they do
[10:12] <Keybuk> no idea what findfs is
[10:12] <dholbach> pitti: https://wiki.ubuntu.com/MOTU/SRU - I'll integrate it with the StableReleaseUpdates page
[10:12] <Keybuk> mvo: what do you want to do about these bugs though?
[10:12] <mvo> Keybuk: feel free to reassign/ignore. I just wanted to make sure that someone with more knowledge about this sees them :)
[10:12] <Keybuk> it's already too late to fix upgrade problems
[10:13] <jordi> dholbach: I just subscribed motu-sru to https://bugs.launchpad.net/distros/ubuntu/+source/gtetrinet/+bug/73720
[10:13] <Ubugtu> Malone bug 73720 in gtetrinet "Crashes on first startup" [Undecided,Unconfirmed]  
[10:14] <dholbach> jordi: super
[10:14] <jordi> dholbach: I can't seem to link it to GNOME bugzilla, the component is missing
[10:14] <dholbach> jordi: http://launchpad.net/products/+new
[10:16] <jordi> dholbach: no, the product in lp is there
[10:16] <jordi> I can't seem to add a gnome bugzilla link though
[10:16] <seb128> jordi: your bug has no patch :p
[10:16] <seb128> no debdiff rather :)
[10:16] <sivang> Keybuk: do you mind if I do dhcdbd ? (e.g. merge)
[10:17] <jordi> ah. :)
[10:17] <jordi> I could post the diff for debian's -2. Would that be enough?
[10:17] <dholbach> jordi: https://launchpad.net/products/gtetrinet - it does NOT exist
[10:18] <jordi> dholbach: oh I see. :)
[10:18] <jordi> so I need to create it?
[10:18] <dholbach> jordi: http://launchpad.net/products/+new
[10:18] <pitti> Mithrandir: I'd like to get http://paste.ubuntu-nl.org/34723/ into herd-1; I tested it properly, and we cannot fix it in an update
[10:20] <jordi> done :)
[10:21] <Mithrandir> pitti: you're aware that host(1) is useless for dns debugging and that you should use dig(1) instead?
[10:21] <Mithrandir> pitti: apart from that, looks good to me.
[10:22] <pitti> no, I'm not; it's just what the spec says
[10:22] <Keybuk> sivang: sure, go for it
[10:23] <pitti> Mithrandir: I can switch it to dig, if that's better (but maybe after herd-1
[10:24] <Mithrandir> pitti: sure
[10:24] <pitti> Mithrandir: dig always exits with 0, too, but I guess I could check for non-empty lines that don't start with ';'
[10:25] <pitti> Mithrandir: what's the principal difference to host?
[10:27] <sivang> Keybuk: thanks :)
[10:28] <Mithrandir> pitti: dig works, host doesn't.  At least in some cases, it's that way; I don't remember the exact problem.
[10:29] <Mithrandir> pitti: you probably want dig +sh -t soa local. and check for an empty answer.
[10:33] <Keybuk> dig is on /usr, isn't it?
[10:33] <Keybuk> actually, so is host
[10:33] <Mithrandir> we should totally just get rid of /usr and merge everything into /
[10:33] <Mithrandir> :-)
[10:34] <Mithrandir> mvo: spec approved
[10:34] <Keybuk> Mithrandir: we could rename the directories while we're at it
[10:34] <Keybuk> make it easier for everyone
[10:35] <Keybuk> /Programs, /Libraries, /System, /Users
[10:35] <mvo> Mithrandir: thanks!
[10:35] <_ion> Let's replace mountpoints with C:\, D:\ etc.
[10:36] <highvoltage> eek
[10:36] <Mithrandir> Keybuk: /usr served a useful function once upon a time.  I don't believe it does.  As for the directories having arcane names, well, changing those is a lot more work.
[10:37] <Keybuk> Mithrandir: I don't disagree
[10:37] <Keybuk> the main problem is you'd need to have /games, /include, /local and /share
[10:37] <_ion> I've heard MacOSX shows the /Users etc. directories transparently translated by locale settings, which is quite neat.
[10:38] <Mithrandir> _ion: it does.  It also transparently translates / in file names to : on the file system and vice versa.
[10:38] <Keybuk> http://forge.novell.com/pipermail/hula-general/2006-November/002084.html
[10:38] <Keybuk> ouch
[10:38] <Mithrandir> Keybuk: so we just need a /cruft directory too. :-)
[10:38] <Keybuk> /cruft/openoffice/...
[10:38] <Keybuk> I'm in!
[10:39] <Mithrandir> Keybuk: actually, I think games should be merged into bin.  A bit less sure what to do about include and share.
[10:39] <_ion> keybuk: Is that the announcement about Novell not more providing full-time developers for Hula? Microsoft has nothing to do with that, of course. ;-)
[10:39] <dade`> why gnome-power-manager does not work with 2.6.19 kernels ?
[10:39] <Keybuk> I've often wanted to split /usr/lib into "libraries" and "development pieces"
[10:39] <pitti> Lathiat: hmm, uninstalling libnss-mdns, restoring nsswitch.conf -> DNS still works for me; was that the scenario you tested?
[10:41] <cjwatson> one problem with merging games into bin is that there are filename clashes
[10:41] <Spads> heh
[10:42] <Mithrandir> cjwatson: I'm sure we could fix that.
[10:45] <siretart> pitti: I notice there is a spec called 'firewall' in blueprint, where you are assignee. it links to SoC-Firewall. do you know what the current status about that is?
[10:45] <pitti> siretart: I didn't hear back from the student for a year now, so I call it officially dead
[10:46] <siretart> pitti: has any work been started on this?
[10:47] <pitti> siretart: yes, there is some code from him; by the end of SoC it was very buggy still, though
[10:54] <pitti> Mithrandir: cmake approved
[10:55] <Mithrandir> pitti: yay you. :-)
[10:55] <Mithrandir> .. and there it got promoted
[11:01] <lifeless> hmm, no reason for a squid3-client
[11:01] <lifeless> the squidclient is no different
[11:01] <Mithrandir> no reason for subdirectories in /usr/bin either
[11:02] <lifeless> wtf
[11:02] <lifeless> very wrong
[11:02] <Mithrandir> that was more or my response too.
[11:02] <Mithrandir> more or less, even
[11:05] <ogra> pitti, yippie
[11:09] <cjwatson> Keybuk: btw, did you notice that Debian's pulsing twice a day now?
[11:10] <Keybuk> cjwatson: I had noticed, yes
[11:10] <Keybuk> haven't quite divined the times
[11:16] <cjwatson> Keybuk: seems to start at 0752 and 1952 UTC
[11:17] <cjwatson> I don't know exactly how long it takes, but apparently ftp-master is ridiculously overpowered now so not long
[11:17] <boitono> https://launchpad.net/distros/ubuntu/+source/initramfs-tools/+bug/47768  bug number 47768 has yet to be resolved and I am be afflicted, and news/help would be appreciated.
[11:17] <Ubugtu> Malone bug 47768 in initramfs-tools "Mount Root Files System Failed" [Critical,Confirmed]  
[11:17] <cjwatson> I imagine the mirror pulse dominates the time
[11:19] <boitono> And what really sucks is that the bug has been open since may of this year!
[11:34] <lifeless> boitono: so help fix it, put a patch forward. This channel is for development activity.
[11:39] <boitono> lifeless, ok, I think I just might, where can I go to get more information about what has already been done?
[11:40] <boitono> and how would I submit such a patch?
[11:42] <mnepton> boitono: https://wiki.ubuntu.com/NewDeveloperProcess
[11:43] <boitono> is the bug report the only source of information on this matter?
[11:44] <Keybuk> boitono: it would appear so
[11:45] <Hobbsee> mnepton: you dont deflate unless i poke you with a pin!
[11:46] <mnepton> yes, dear. whatever you think is best, dear.
[11:46] <Hobbsee> haha
[11:48] <Lathiat> pitti: i had the [NOTROUND=return]  etc
[11:49] <fabbione> regexp question :)
[11:49] <Lathiat> pitti: with no nss-mdns installed (or the wrong plugin specific - mdns4_minimal [which isnt in 0.7 even when installed] ] ) and telnet/firefox/gaim etc dont work but when i was root it worked
[11:49] <fabbione> cat mdadm.conf | sed -e :a -re '$!N;s/\n[[:space:] ] +/ /;ta' -ne '/^(ARRAY|DEVICE)/P;D' 
[11:49] <fabbione> DEVICE partitions
[11:49] <fabbione> ARRAY /dev/md0 level=raid5 UUID=e4ac4774:c730d28b:0978c89e:2173da3d num-devices=4
[11:49] <fabbione> how can i make sure to extract only the value for UUID ?
[11:49] <fabbione> sed today is not my friend
[11:51] <Mithrandir> sed -e s/.*UUID=([0-9a-f:] *)/\1/ , maybe?
[11:51] <Spads> foo=${line%% num-devices*}
[11:51] <Spads> bar=${foo##*UUID=}
[11:51] <Spads> in bash
[11:52] <fabbione> Spads: what's after the uuid is random
[11:52] <Spads> aha
[11:52] <Spads> pity.
[11:52] <fabbione> or can be random
[11:52] <Spads> in that case
[11:52] <fabbione> and it needs to be posix .. so no bashism
[11:52] <Spads> oh.
[11:53] <Mithrandir> what Spads posted is pure posix sh.
[11:53] <Lathiat>  cat /etc/mdadm/mdadm.conf |grep /dev/md1| sed -e 's/.*UUID=\([0-9a-f:] *\)/\1/' works for me
[11:53] <fabbione> Lathiat: please note that your mdadm.conf is not custom 
[11:54] <fabbione> Lathiat: just swap order of things and it will break
[11:54] <Spads> I'd have done + instead of *, but that regex looks pretty robust
[11:54] <Spads> I'd also gobble everything after the uuid value
[11:54] <fabbione> cat mdadm.conf | sed -e :a -re '$!N;s/\n[[:space:] ] +/ /;ta' -ne '/^(ARRAY)/P;D' | sed -e 's/.*UUID=\([0-9a-f:] *\)/\1/'
[11:54] <fabbione> e4ac4774:c730d28b:0978c89e:2173da3d num-devices=4
[11:54] <fabbione> num-device still needs to go
[11:54] <Spads> yeah, add a .*$ at the outside of the capture parens
[11:54] <cjwatson> uuid random> so foo=${line##*UUID=}; bar=${foo%% *}
[11:54] <Spads> s/^.*UUID=\([0-9a-f:] *\).*$/\1/
[11:54] <Mithrandir> you could do something like foo=${line##*UUID=} ; foo=${line%% *}
[11:55] <fabbione> Spads: thanks that works
[11:55] <cjwatson> Mithrandir: snap
[11:55] <Mithrandir> cjwatson: indeed.
[11:55] <fabbione> Mithrandir, cjwatson: spads wins a beer at next AllHands :)
[11:55] <Mithrandir> cjwatson: apart from mine not working since I used line in the second one too.
[11:55] <Spads> those ## and %% cuts are what i did second, too...
[11:56] <Spads> oh heh
[11:56] <Spads> I said "in that case" and never wrote them
[11:56] <Spads> ok
[01:05] <iwj> cjwatson: AYT?  Would you care to take another look at that login spec of mine ?
[01:11] <sivang> Keybuk: Why is 20-remove-bind-user.patch not mentioned in any of the changelog entries for the base->ubuntu delta btw? (is it related to the create pid with useful mode?)
[01:12] <Keybuk> sivang: probably because it comes from our own original packaging of dhcdbd
[01:12] <Keybuk> and the changelogs for that are missing
[01:12] <sivang> Keybuk: okay, I take it this is needed for this to work nicely with upstart then?
[01:12] <Keybuk> no
[01:12] <Keybuk> nothing to do with upstart
[01:13] <Keybuk> we simply don't need to poke the bind user from dhcdbd, or vice-versa
[01:13] <Keybuk> so the permission is not required
[01:13] <sivang> ah, okay
[01:13] <Keybuk> (and leaving it there would be a security hole)
[01:13] <sivang> fare enough
[01:13] <sivang> indeed
[01:13] <Keybuk> I think it's used in RH to reload bind when the dhcp interface changes
[01:13] <Keybuk> we don't use bind
[01:13] <cjwatson> iwj: will do
[01:14] <sivang> Keybuk: right, thanks for the info lots
[01:15] <iwj> cjwatson: Thanks.
[01:25] <cjwatson> iwj: if the gdm login from a screensaver fails or is cancelled, the VC should probably be switched back to the screensaver it came from and the screensaver made to run again
[01:25] <iwj> Good point.
[01:26] <cjwatson> iwj: "If the user requests to switch to a different user, the screensaver will be started immediately on that VC" -> meaning the VC of the switched-from user rather than the switched-to user?
[01:26] <iwj> Exactly.
[01:27] <Mithrandir> we generally put docs for packages in main in main?
[01:27] <cjwatson> Mithrandir: yes
[01:28] <cjwatson> iwj: it looks fine to me now; however I don't know any of the code in question. Has anyone in the GNOME team looked at it to do a feasibility check?
[01:28] <Mithrandir> as of next publisher run, *-desktop should be installable.
[01:28] <iwj> cjwatson: Note that I know of, no.  I should talk to seb.
[01:28] <Mithrandir> (yay!)
[01:29] <cjwatson> iwj: if seb doesn't scream, I'm happy to approve it
[01:29] <cjwatson> Mithrandir: rah
[01:29] <iwj> cjwatson: OK.
[01:29] <cjwatson> d-i nearly merged
[01:29] <Mithrandir> cjwatson: yay you! :-)
[01:30] <iwj> seb128: Are you around at the moment ?  Would you care to cast your eye over https://wiki.ubuntu.com/UnifiedLoginUnlock and tell me whether I'm mad or not ?
[01:49] <cjwatson> heno: your shell in braille-setup is a bit wonky - you want [bB]  rather than [b,B]  etc.
[01:49] <cjwatson> heno: also, you must use /bin/sh not /bin/bash or it won't work in the installer
[01:50] <cjwatson> there's nothing bash-specific about your script
[01:52] <cjwatson> heno: (it's otherwise fine, though - approvd)
[01:52] <cjwatson> +e
[01:57] <pitti> Keybuk: FYI, new nss-mdns uploaded which should automatically fix your nsswitch.conf as well
[01:57] <pitti> Keybuk: I tested it locally, but feedback appreciated, of course
[02:11] <seb128> iwj: I was away for lunch, will have a look to it now
[02:21] <seb128> iwj: any reason for "The X server used by gdm will remain around forever" instead of using gdmflexiserver?
[02:26] <seb128> iwj: the spec makes sense to me
[02:27] <Keybuk> iwj: sweet
[02:27] <mvo> cjwatson: could you please have a look at https://blueprints.launchpad.net/distros/ubuntu/+spec/dist-upgrader-fixes ?
[02:28] <Keybuk> seb128: flexi would create a new gdm on a new VT, for each switch or unlock?
[02:29] <seb128> Keybuk: right, the time to log-in
[02:29] <cjwatson> mvo: yep, it's in my queue
[02:30] <Keybuk> so you'd get (gdm on VT 7) -> login -> (X session on VT 7) -> lock -> (screensaver on VT 7) -> unlock -> (screensaver on VT 7; gdm on VT 8) -> enter password -> (gdm on VT 8 exits; X session on VT 7)
[02:30] <mvo> cjwatson: thanks
[02:30] <Keybuk> in the spec, it's always gdm on VT 7; X sessions or screensavers on VT 8 plug
[02:30] <Keybuk> s/g$/s/
[02:30] <Keybuk> the advantage there I guess is that it's quick to get back to gdm at any point, and you only create a new VT for a new login
[02:30] <seb128> Keybuk: is that scheme an issue?
[02:31] <seb128> right
[02:31] <Keybuk> the disadvantage is that you'd always have a gdm on VT 7
[02:31] <Keybuk> the advantage of flexi is that you don't have the spare gdm at all
[02:31] <seb128> the advantage of using gdmflexiserver is that it is already working
[02:31] <Keybuk> the disadvantage is that you need a new VT for any password operation
[02:31] <seb128> and it's fast enough I think
[02:31] <cjwatson> hmm, what happens if you're trying to log in but have run out of VTs?
[02:31] <Keybuk> does flexi know that if you enter a username and password, it's not to login that user, but to exit and switch to a different VT?
[02:31] <seb128> right, but opening gdm on VT is cheap enough no?
[02:31] <Keybuk> if so, where does it get that state from
[02:31] <cjwatson> sorry, trying to unlock
[02:32] <cjwatson> failing to log in (at first) because you've run out of VTs is one thing; failing to unlock because you've run out of VTs would be much worse
[02:32] <cjwatson> (imo)
[02:32] <seb128> we could fallback to gnome-screensaver unlock in that case
[02:32] <cjwatson> yeah, true
[02:32] <seb128> cjwatson: but right, you have a point
[02:32] <Keybuk> iwj's does seem more robust
[02:32] <seb128> yeah
[02:33] <seb128> it's just extra work
[02:33] <Keybuk> I don't think it is extra work; I think flexi would need the same amount of work
[02:33] <seb128> if he's happy to do the changes that's all good :)
[02:33] <cjwatson> iwj's requires more in the way of gdm<->screensaver communication framework
[02:33] <Keybuk> I just played with it, it doesn't seem to actually do what the spec requires
[02:33] <Keybuk> firing up flexi started a login on VT 8, entering my username there started a new X session on VT 8; so I had two
[02:33] <Keybuk> it didn't unlock VT 7
[02:33] <seb128> it should open a dialog "do you want to user the running session or open a new one"
[02:34] <seb128> s/user/use
[02:34] <seb128> rather "use current session, open new one, go back to gdm"
[02:34] <cjwatson> "would you like GNOME to work or not work?"
[02:34] <cjwatson> (do two concurrent sessions actually work properly now?) :-)
[02:34] <seb128> (no they don't)
[02:35] <Keybuk> seb128: also flexi took several seconds to start the greeter
[02:35] <Keybuk> didn't time it exactly, but it was around 5s
[02:35] <giskard> dade`, ?
[02:35] <dade`> you got me
[02:35] <cjwatson> I don't really see a particular non-desktop-developer need for two concurrent sessions though. Does that really need to be an explicit UI option?
[02:35] <seb128> Keybuk: right, I've just tried too
[02:35] <seb128> it's around 5 seconds for me too
[02:35] <seb128> I'm convinced, I like the spec way better too ;)
[02:35] <heno> cjwatson: thanks! I have no doubts there is lots of room for improvement in my shell scripting :)
[02:36] <seb128> cjwatson: no, I think we have a bug about that, it should just switch back to the previous one
[02:36] <Keybuk> if we then have hibernate lock the screen and always come up to VT 7
[02:36] <Keybuk> that'd be sweet
[02:36] <cjwatson> right, I was just foollowing up on your 'it should open a dialog "do you want to user the running session or open a new one"' - or were you describing the current intended behaviour?
[02:37] <seb128> cjwatson: yeah, I was describing the current behaviour
[02:37] <seb128> because Scott said
 firing up flexi started a login on VT 8, entering my username there started a new X session on VT 8; so I had two
[02:37] <Keybuk> seb128: ah, weird, my .desktop had "-s" -- without that, I do get the popup
[02:37] <Keybuk> though the mouse doesn't work in that popup
[02:37] <Keybuk> in fact, the mouse doesn't work in the flexi server at all
[02:37] <cjwatson> seb128: ah, right, I misunderstood your "should" then
[02:38] <Keybuk> amusingly, when it flips back, it flips back to a locked screen
[02:38] <seb128> Keybuk: I just noticed the same issue on my laptop one hour ago
[02:39] <iwj> seb128: Thanks for reading that.
[02:39] <Keybuk> tbh, I quite like the idea that "Ctrl-Alt-F7" is a "switch user" command <g>
[02:39] <Keybuk> "Press Ctrl-Alt-F7 to login" <g>
[02:39] <Keybuk> if only we could make it Del
[02:39] <iwj> Keybuk: Write a spec for feisty+1 :-).
[02:39] <Keybuk> we probably could
[02:40] <seb128> iwj: np, thank you for writing that spec, user switching is something that needs love for a long time now
[02:40] <Keybuk> iwj: I imagine it'd just be an X patch of some kind
[02:40] <iwj> I watched my parter try to use it and neither of us had any idea what was going on.
[02:41] <iwj> I particularly liked the `would you like to make a horrid mess' option.
[02:43] <iwj> If we're patching X we could make the X VC switching not hideously racey.
[02:50] <Keybuk> VT switching in general is racey
[02:52] <ogra> but i honestly think user switching has nothing to do in X  ... put it properly into gdm rather ...
[02:59] <sfllaw> fabbione: Pong.
[02:59] <fabbione> sfllaw: any progress on bug #56496?
[02:59] <Ubugtu> Malone bug 56496 in lvm2 "Cannot install clvm" [Undecided,Confirmed]  http://launchpad.net/bugs/56496
[03:00] <sfllaw> It has an undecided priority, so I haven't gotten to it yet.
[03:05] <sfllaw> fabbione: I work in order of priority, from most important to least.
[03:06] <fabbione> sfllaw: ok.. 
[03:06] <fabbione> ok let's try again
[03:06] <fabbione> sfllaw: any progress on bug #56496?
[03:06] <Ubugtu> Malone bug 56496 in lvm2 "Cannot install clvm" [High,Confirmed]  http://launchpad.net/bugs/56496
[03:08] <sfllaw> fabbione: Not yet.  Soon.
[03:10] <sfllaw> fabbione: Your SRU e-mail was quite charming.
[03:10] <Slant_Laptop> Can anyone take a look at https://wiki.ubuntu.com/FeistySuspendOverview and give any pointers to improving it?
[03:11] <fabbione> sfllaw: hmm i don't recall what i wrote in it
[03:12] <sfllaw> Something about holy roosters.
[03:12] <Hobbsee> hehe
[03:13] <Hobbsee> holy roosters/
[03:13] <fabbione> sfllaw: ahaha yeah i remember now
[03:15] <sfllaw> I had no idea why poultry was involved with LVM2, but decided I didn't want to know.
[03:15] <ogra> grilled roosters ... mmm 
[03:22] <iwj> cjwatson: So did you actually want to set that spec to `Approved' ? :-)
[03:22] <iwj> Or shall I ?
[03:26] <Keybuk> iwj: he's keeping you shivering with anticp...
[03:29] <cjwatson> iwj: I was waiting for the results of the conversation above to be encoded into the spec, if necessary
[03:29] <cjwatson> iwj: or for somebody to say no changes were necessary :) I wasn't entirely sure
[03:30] <Keybuk> cjwatson: I think we agreed that iwj's spec was right?
[03:30] <cjwatson> yeah, on reflection it looks like it
[03:31] <iwj> That's what I thought but then I would say that :-).
[03:31] <cjwatson> indeed :)
[03:31] <cjwatson> approved. thanks for pestering ;-)
[03:32] <iwj> NP.
[03:32] <iwj> Thanks.
[03:57] <seb128> rodarvus: hi
[03:57] <seb128> rodarvus: around?
[04:23] <cjwatson> heno: are the configuration changes to orca mentioned in https://wiki.ubuntu.com/Accessibility/Specs/ColorFilters going to be trivial?
[04:24] <heno> cjwatson: yes, it would be making gui elements to set gconf keys
[04:25] <heno> cjwatson: likely as an extra tab in Orca
[04:25] <heno> cjwatson: Upstream are also traditionally very helpful with this stuff
[04:26] <cjwatson> ok, if it's basically straightforward extensions to similar things that are done already, that's fine
[04:27] <heno> it is yes, more of the same
[04:28] <cjwatson> "heno - 2006-27-11: ..."
[04:28] <cjwatson> heno: your date format is mutant. :)
[04:29] <heno> ops, typo :)
[04:29] <cjwatson> heno: I think that's everything you sent me approved
[04:29] <cjwatson> thanks, all seemed pretty reasonable
[04:29] <heno> cjwatson: cool!
[04:30] <heno> cjwatson: I also still thinking about the multilingual speech synth one, but I'm unsure about the community help I will get there
[04:31] <heno> It needs to be defined quite clearly and not be unrealistic
[04:31] <heno> need to do some expectation management there :)
[04:42] <jorgp> I was trying to package a backport, I gave the package a name of  1.2.9-1~edgy1 per https://wiki.ubuntu.com/BackportsHowto
[04:42] <jorgp> however lintian says bad-version-number 1.2.9-1~edgy1
[04:43] <Keybuk> lintian probably needs updating ;)
[04:43] <cjwatson> jorgp: you have an old version of lintian; upgrade it to 1.23.23 or newer
[04:44] <cjwatson> the one in feisty will do
[04:44] <jorgp> ok
[04:44] <jorgp> yes, I am using edgy's lintian
[04:45] <jorgp> 1.23.22ubuntu1
[04:45] <cjwatson> (lintian messages are meant to be read and decided upon by humans)
[04:47] <jorgp> thanks cjwatson, I install 1.23.25 from feisty
[04:54] <pitti> Keybuk: can you please reject ifupdown_0.6.8ubuntu2_source.changes from the incoming feisty queue?
[04:55] <pitti> Keybuk: ajt and I just discussed the ifupdown/avahi-autoipd integration and we found something better
[04:55] <Keybuk> oh?
[04:55] <Keybuk> what did you reject, and what did you decide on?
[04:55] <pitti> so I'd rather discuss it with some more guys again
[04:56] <pitti> Keybuk: instead of adding a new ifupdown method, ajt would prefer autoipd installing a pre-up.d script
[04:56] <pitti> Keybuk: that would start avahi-autoipd on each interface, so that it's always available for completely unconfigured interfaces
[04:56] <pitti> that works, because autoipd has the nice property of immediately releaseing itself when the interface gets a 'real' IP assigned
[04:57] <Keybuk> ah
[04:57] <Keybuk> you wouldn't need the dhclient hook then?
[04:57] <pitti> this is a really interesting approach IMHO, and we should play with that and discuss it again
[04:57] <pitti> Keybuk: right
[04:57] <pitti> any manual ifconfig, dhclient, etc. would assign a real address and remove the zeroconf one
[04:58] <pitti> Keybuk: as I said, I'd like to play with the idea before committing to it, but at least I'd like to defer my ifupdown upload a bit
[04:58] <pitti> if it gets into the archive, no harm is done either, but might confuse people
[05:01] <cjwatson> rodarvus: what's happening with bullet-proof-x drafting?
[05:02] <cjwatson> there is not much time left
[05:02] <cjwatson> iwj: do you have time to review ubiquity-driver-updates for me?
[05:02] <rodarvus> seb128, I'm here now
[05:03] <seb128> rodarvus: somebody asked on #ubuntu-bugs if bug #68607 should be marking as wishlist
[05:03] <Ubugtu> Malone bug 68607 in xorg "No direct rendering by default with i810/i815 video" [Undecided,Unconfirmed]  http://launchpad.net/bugs/68607
[05:03] <rodarvus> cjwatson, indeed, I'm sorry. I had problems with my main laptop, but will finish it today
[05:03] <cjwatson> Keybuk: intel-mac-support is in Review, and you're the approver; do you fancy shortcutting the review step? it's fairly short
[05:04] <rodarvus> seb128, I'm not sure if we should configure X to run at 16 bpp, or instead disable dri on these machines
[05:05] <seb128> rodarvus: ok, I told him to let the bug untriaged if he was not sure since you were not around
[05:05] <Keybuk> cjwatson: I usually do if I'm the approver
[05:05] <rodarvus> imho, a machine with a video board older than a i830 is not very fast, and will have mixed results using dri
[05:05] <seb128> rodarvus: and he left the chan since so don't bother
[05:06] <rodarvus> *nods*
[05:07] <iwj> cjwatson: Sure.
[05:12] <cjwatson> iwj: thanks
[05:12] <cjwatson> Keybuk: (likewise)
[05:14] <iwj> cjwatson: I see there are no plans for vendor udebs, so only `optional' drivers make sense.
[05:14] <iwj> You might want to make that clear.
[05:14] <iwj> Also, isn't there a timing issue wrt udev events for the devices and installation of the kernel modules ?
[05:14] <cjwatson> iwj: that's not true - this is only for the live CD, which doesn't use udebs.
[05:14] <Keybuk> iwj: elaborate (timing issue)
[05:15] <iwj> Keybuk: Err, I may be confused - I'm only the reviewer :-) - but AIUI the idea is that something will install these debs into the livecd so that the modules are available and relevant devices can be found.
[05:15] <iwj> (I'm reading https://wiki.ubuntu.com/Ubiquity/DriverUpdates btw, to save you url hunting)
[05:16] <cjwatson> it'll be done in the initramfs. the thing that does it might have to poke uevents again, I suppose ...
[05:16] <Keybuk> will the modules be just available, or loaded?
[05:16] <iwj> I don't know if you start a new udev in the live system.
[05:16] <iwj> Keybuk: The idea is that the vendor supplies a .deb full of modules.  I take it they won't be loaded explicitly except if udev decidese to.
[05:16] <cjwatson> there's only one udev, but you can poke it to process events again
[05:17] <Keybuk> will these modules override any in the existing system?
[05:17] <cjwatson> Keybuk: they'll typically be installed in /updates
[05:17] <cjwatson> so yes, they might
[05:17] <iwj> I don't really understand how the livecd works, apparently, but I seem to have found something that needs further explanation ...
[05:17] <Keybuk> note that the referral to new-pci-ids is kinda broken, given that spec (as summarised) is impossible
[05:17] <iwj> The initramfs just builds the whole root fs straight away and there are no udebs ?
[05:17] <Keybuk> how will you cope with
[05:18] <Keybuk> *drum roll*
[05:18] <Keybuk> DEPMOD
[05:18] <cjwatson> new-pci-ids can sort itself out in time, and if necessary be replaced by something that works :)
[05:18] <Keybuk> the new udeb won't include a new modules.dep will it?
[05:18] <Keybuk> and if it does, how will you run two of them?
[05:18] <cjwatson> iwj: the live CD contains a root filesystem which is mounted and pivoted into
[05:18] <iwj> cjwatson: Right, I mean, the stack of union mounts and that kind of thing.
[05:18] <cjwatson> it also has an initramfs with all the modules it needs
[05:18] <cjwatson> there are, indeed, no udebs
[05:19] <iwj> cjwatson: all the modules it needs> So if a module it `needs' is on the vendor CD it won't get it.
[05:19] <cjwatson> Keybuk: surely just run depmod in the vendor .debs' postinst?
[05:19] <cjwatson> iwj: that's, er, the point of the specification ...
[05:19] <iwj> cjwatson: depmod> Maybe that should be documented.
[05:19] <Keybuk> cjwatson: which could break if volatile hasn't been unpacked then
[05:19] <Keybuk> you can only run depmod if you know that all the modules you'll ever need are under you right now
[05:19] <cjwatson> iwj: yeah, take it as read that I'll document the results of this IRC discussion, but I want to have the discussion first
[05:19] <Keybuk> and lrm isn't unpacked or mounted until the boot sequence
[05:19] <iwj> cjwatson: So I'm confused.  You're installing these debs into the real root not the initramfs.
[05:19] <iwj> cjwatson: Right.
[05:19] <cjwatson> Keybuk: you can run depmod multiple times ...
[05:20] <Keybuk> cjwatson: lrm doesn't cause depmod to be run
[05:20] <iwj> real root> So the initramfs won't have those modules available until it pivots and runs the real init (I assume there's something like a real init).
[05:20] <cjwatson> it does, in the postinst. it seems to make sense that a .deb containing modules would do the same
[05:20] <Keybuk> cjwatson: I'm clearly misunderstanding where these debs are handled
[05:20] <cjwatson> iwj: ok, so we could copy the installed modules into the postinst
[05:20] <iwj> And if the real root has different kernel modules to the initramfs you have to arrange to poke udev.
[05:20] <cjwatson> er
[05:21] <cjwatson> sorry, the above is garbage :)
[05:21] <Keybuk> cjwatson: these are unpacked and installed at livefs build time?
[05:21] <Keybuk> rather than boot?
[05:21] <cjwatson> no
[05:21] <cjwatson> we have to handle this .deb before the real root is mounted
[05:21] <Keybuk> ok, so when are they unpacked into the system?
[05:21] <Keybuk> ok
[05:21] <Keybuk> so you can't run depmod
[05:21] <cjwatson> why not?
[05:21] <Keybuk> because you'll overwrite the correct modules.dep created at livefs build time
[05:21] <Keybuk> with one that's missing the volatile modules
[05:21] <iwj> cjwatson: You're going to install the .deb into the initramfs and treat it like a udeb ?
[05:21] <cjwatson> if using this scheme, we could run depmod again later
[05:22] <cjwatson> iwj: forget about udebs.
[05:22] <cjwatson> please. it's horribly confusing and irrelevant.] 
[05:22] <iwj> Are you going to install this .deb into the initramfs or the real root ?
[05:22] <Keybuk> cjwatson: that would have to wait until the root filesystem is writable, which is significantly after modules have been loaded
[05:22] <Keybuk> unless you move lrm into the livefs initramfs
[05:22] <Keybuk> which isn't unreasonable
[05:23] <cjwatson> iwj: the modules need to be available in both
[05:23] <iwj> cjwatson: So you're going to unpack it twice ?
[05:23] <cjwatson> they may, for example, include modules needed for the kernel to understand your CD drive
[05:23] <iwj> And you're going to run the vendor's postinst in the initramfs ?
[05:23] <cjwatson> (the bootloader clearly understands it, and loaded the initramfs, but the kernel might not)
[05:23] <iwj> cjwatson: CD drive> That's not going to work - that's where the modules are coming from.
[05:23] <cjwatson> iwj: see above.
[05:24] <cjwatson> it's entirely possible for the bootloader to be able to read the initramfs without the kernel being able to do so
[05:24] <cjwatson> that's why this all goes in the initramfs
[05:24] <cjwatson> the bootloader is typically using BIOS calls, which vendors are rather more careful to implement :)
[05:24] <iwj> ???  But you say the initramfs is prompting for this CD.
[05:25] <Keybuk> the initramfs has to find the CD to mount it, yes
[05:25] <Keybuk> initramfs is loaded by the boot loader
[05:25] <cjwatson> oh, sorry, yeah, you're right. mm
[05:25] <Keybuk> oh, I see, heh
[05:25] <iwj> This conversation seems to be making more confused, not less.
[05:25] <iwj> s/more/me more/
[05:26] <cjwatson> ok, so let me back up and actually read the spec I wrote
[05:26] <cjwatson> "Any module packages (in .deb format) found on the driver CD will be stored in memory and installed into the live session by casper at some point after filesystems have been mounted."
[05:26] <cjwatson> so in fact that does not require them to be installed in the initramfs - which on reflection makes more sense.
[05:26] <iwj> Right.  So they're `optional' - not needed for boot.
[05:26] <iwj> And you can install them later whenever is convenient.
[05:27] <iwj> But you still have to retrigger udev.
[05:27] <Keybuk> which invalidates the first use case
[05:27] <cjwatson> err, in an exceedingly loose sense of optional
[05:27] <cjwatson> they might still include disk drivers
[05:27] <Keybuk> they have to be installed into the initramfs if you want to find the disk
[05:27] <cjwatson> they're not optional for booting a normal (non-live-CD) system
[05:27] <cjwatson> but they are optional for booting a live CD
[05:27] <iwj> cjwatson: Right, `optional' for the livecd, quite possibly non-optional for the installed system.
[05:27] <cjwatson> Keybuk: the initramfs only needs to find the CD
[05:27] <iwj> Keybuk: No, no, that's a different initramfs.
[05:27] <iwj> There are two initramfs's.
[05:27] <Keybuk> ah true
[05:28] <iwj> Something has to make sure the installed initramfs gets regenerated with the right modules.
[05:28] <cjwatson> but yes, you're quite right, we have to run depmod and trigger udev somehow.
[05:28] <cjwatson> iwj: that bit's straightforward, but needs to be expanded in the spec
[05:28] <iwj> Right.
[05:28] <iwj> cjwatson: Can I suggest we have the casper thing do that, rather than the vendor's postinst ?
[05:29] <iwj> That way the vendor has only to know `drop these files here' and of course it only has to run once.
[05:29] <cjwatson> That would work. I had been intending to deliver an SDK to vendors anyway, mind you; I don't expect them even to know how to construct .debs
[05:29] <Keybuk> which udev are you triggering?
[05:29] <cjwatson> so I don't think it makes too much difference, but it would probably be easier to QA in casper
[05:29] <Keybuk> there are two udevs in our boot sequence
[05:29] <iwj> cjwatson: Right.  But the simpler you make it the less they'll put `crash the computer', `annoy the user', and `fuck everything up' in their maint scripts.
[05:30] <cjwatson> heh
[05:30] <cjwatson> Keybuk: I thought the whole point of the fiddly /dev tmpfs stuff was to have only one udev ...
[05:30] <Solarion> this probably isn't the place to talk about getting stap to work is it?
[05:30] <Keybuk> cjwatson: one udev database, two udev processes with differing rules and module sets
[05:31] <Keybuk> the udev in the initramfs has a minimal rule set and module set, and just gets the root filesystem mounted
[05:31] <Keybuk> we kill that one, exec init
[05:31] <Keybuk> and some seconds later, start a new udev up with the complete rule and module set
[05:31] <Keybuk> that udev's job is to do everything else, including getting the other filesystems mounted
[05:31] <Keybuk> they share the same /dev filesystem, and udev db
[05:31] <Keybuk> but that's it
[05:32] <cjwatson> Keybuk: so, it perhaps helps here that the live session's root filesystem is writable from before the real init starts
[05:33] <Keybuk> the "root filesystem" the first udev mounts, in this case, is the casper image, squashfs, etc.
[05:33] <Keybuk> and from above, these drivers don't "fix" that
[05:33] <Keybuk> so it's the second udev that you want to influence
[05:33] <cjwatson> right
[05:33] <Keybuk> which means you want the drivers unpacked into the livefs
[05:33] <cjwatson> correct
[05:33] <cjwatson> as the spec says
[05:33] <cjwatson> I was temporarily confused at the start of this IRC discussion; sorry about that
[05:34] <cjwatson> so casper can unpack the drivers into the livefs after unpacking and making it mountable, and run depmod on the livefs
[05:34] <iwj> Yay!  gs-gpl is just a sync!
[05:34] <cjwatson> s/mountable/writable/
[05:34] <iwj> Shame I had to do so much archaeology to find that out.
[05:34] <Keybuk> ok
[05:34] <Keybuk> so your issue is making sure udev (really, modprobe) knows about these new drivers
[05:34] <Keybuk> which means you need to run depmod
[05:34] <cjwatson> and it can tweak the init scripts to run depmod immediately after mounting volatile?
[05:35] <Keybuk> right
[05:35] <Keybuk> that'll take about 30s to 1m though
[05:35] <Keybuk> depending on CD drive speed
[05:35] <Keybuk> which is a bit of a bitch, but hey ho
[05:35] <Keybuk> I was thinking of being more clever <g>
[05:35] <cjwatson> I think that's possibly acceptable if it's only in the driver disk case
[05:35] <cjwatson> but if you have a better idea, I'm happy to hear it :)
[05:35] <Keybuk> you could have a modules.dep and modules.alias snippet included in the packages
[05:35] <Keybuk> and concat that onto the end of the ones in the livefs <g>
[05:36] <cjwatson> is there a sane way to generate that, given that it might need to override modules we ship?
[05:36] <Keybuk> that's the tricky bit
[05:36] <cjwatson> I'd have expected modprobe to honour the first thing it finds, not the last
[05:37] <Keybuk> that doesn't help the installed system though
[05:37] <Keybuk> you have another problem there
[05:37] <iwj> What does `A copy of the entries from debian/changelog corresponding to the changes relative to the current version in Ubuntu' mean ?
[05:37] <Keybuk> as you need to, after u6y installs them, also run depmod there
[05:37] <Keybuk> but to run depmod in the installed target, you need to have volatile mounted there
[05:37] <Keybuk> does u6y do that at the moment?  (lrm postinst does it)
[05:39] <cjwatson> Keybuk: ubiquity reconfigures linux-image and linux-restricted-modules, which has the effect of running depmod - I'm pretty sure it does that with volatile mounted
[05:39] <cjwatson> but if it doesn't, that can be fixed
[05:41] <Keybuk> right
[05:41] <Keybuk> the lrm thing mounts it
[05:41] <Keybuk> (in postinst)
[05:41] <Keybuk> so I guess u6y ends up with a mounted volatile in the target for a while
[05:41] <Keybuk> but that's ok
[05:41] <cjwatson> iwj: it means the debian/changelog entries from Debian that are later than the last version in Ubuntu
[05:42] <iwj> cjwatson: Cripes.
[05:42] <cjwatson> Keybuk: why did you add that requirement anyway? I always ignore it
[05:42] <cjwatson> it seems pointless to me
[05:42] <Keybuk> which requirement?
[05:42] <iwj> I haven't included those in my sync request bug 73783.
[05:42] <Ubugtu> Malone bug 73783 in gs-gpl "please sync gs-gpl from Debian sid main" [Undecided,Unconfirmed]  http://launchpad.net/bugs/73783
[05:42] <cjwatson> Keybuk: for sync requestors to include new debian/changelog entries from Debian
[05:43] <iwj> I assume the purpose of `if there are Ubuntu changes' is to convince ubuntu-archive that I actually checked.
[05:43] <iwj> So I decided to provide the evidence I had rather than conform to the letter of the requirement ...
[05:43] <cjwatson> iwj: Indeed. That part makes sense to me, but not the other ...
[05:43] <Keybuk> cjwatson: that was mostly documenting existing practice -- makes it easier to bounce the bug for approval; notice " * included patch from ubuntu to do FOO" if they asked for overriding changes, etc.
[05:44] <cjwatson> in my experience it's a load of text I skip over; if I decide I have to do additional verification then I'm just going to look at the actual source packages anyway
[05:44] <Keybuk> I'd reject iwj's request on two counts -- no changelog, and no summary of the changes being overriden
[05:44] <cjwatson> and I think the first of those is excess bureaucracy
[05:44] <Keybuk> I tend to check the changelog against the changes they claim debian takes
[05:45] <Keybuk> and it's something they should have to copy and paste anyway
[05:45] <iwj> Keybuk: That's duplicating my work.
[05:45] <iwj> I haven't looked at the Debian changelog entries at all.  I looked at the actual code.
[05:45] <cjwatson> iwj: (the original problem was that people were writing scripts to lie for them about having checked)
[05:45] <iwj> (Since I started out assuming it was going to be a tricky merge.)
[05:46] <Keybuk> the problem there is that do I trust iwj when he says "I have verified that a sync is the right answer" more than I trust random joe bloggs motu
[05:47] <Keybuk> the answer is obviously yes, I'm damned sure iwj checked it more thoroughly than even I would
[05:47] <Keybuk> but will joe bloggs complain that I subject his requests to more checks than I do iwj's?
[05:47] <Keybuk> and that I require more information from joe bloggs than iwj, because I don't believe him automatically
[05:47] <cjwatson> actually, I'd ignore such a complaint on the basis that "er, yes. you didn't write dpkg" or something approximately similar
[05:48] <cjwatson> yes, that's "unfair"
[05:48] <Keybuk> actually, in reality, I rarely bother checking anyone in core-dev
[05:48] <cjwatson> but we aren't robots; a belief about another developer's working practices is a legitimate thing to take into consideration
[05:48] <Keybuk> I just read the name of the source package, and paste it into the terminal
[05:49] <Keybuk> and to a large extent, do the same with prominent motu as well
[05:49] <iwj> Maybe the requirement should say something like "information demonstrating that; typically XYZ will be sufficient".
[05:50] <Keybuk> iwj: I think that's too vague, people will spend more time arguing over what's sufficient than it would have taken them to paste the common information in the first place
[05:51] <Keybuk> the changelog is certainly a lot more relevant after UVF
[05:51] <Keybuk> but even before, I've caught things in there the requestor didn't notice
[05:52] <Keybuk> though to some extent, the point is to prove you've looked at it, so we don't need to
[05:54] <Keybuk> cjwatson: actually, I think I agree with you
[05:54] <Keybuk> the changelog is probably a bit irrelevant
[05:54] <Keybuk> the point was to have something they couldn't just have as a template
[05:54] <Keybuk> but the changelog is extracted by most people's scripts
[05:54] <Keybuk> it's available to us in the changes file, if we want to check
[05:54] <Keybuk> so I'd be happy to just have a list of the ubuntu changes being overridden
[05:55] <Keybuk> though that makes the bugs less useful for UVF approvals?
[06:02] <cjwatson> iwj,Keybuk: Thanks for all the comments. I've updated https://wiki.ubuntu.com/Ubiquity/DriverUpdates following this discussion; does it make more sense now?
[06:05] <iwj> That's much clearer.
[06:06] <cjwatson> thanks!
[06:06] <Keybuk> cjwatson: could you mention that the lrm init script modification will work because the livefs is always writable
[06:06] <Keybuk> normally that runs "before the filesystem is writable"
[06:07] <cjwatson> Keybuk: sure
[06:07] <Keybuk> in the long-term, we won't be using modprobe/depmod for module loading anyway, so there should be a better solution for this
[06:07] <Keybuk> maybe not feisty though
[06:07] <Keybuk> depends how eager kay gets
[06:08] <cjwatson> Keybuk: done
[06:09] <mdz> Keybuk: oh?
[06:09] <Ubugtu> Malone bug 69094 in apt-setup "Please refer to upgrade instructions in sources.list" [Wishlist,Confirmed]  
[06:10] <Keybuk> mdz: in a nutshell, we'll have files that map device information strings to driver names
[06:10] <Keybuk> and a table of driver to module mappings
[06:10] <mdz> Keybuk: we already do have those
[06:10] <Keybuk> no, we have device information strings to MODULE names
 difference
[06:11] <seb128> can Breaks be used now?
[06:11] <Keybuk> on seeing a new device, udev checks the tables to see which drivers it needs, and if necessary will directly load any modules and their dependencies (without calling to modprobe, so the information can be cached in udev)
[06:12] <Keybuk> it will then bind the device and driver together
[06:12] <mdz> seb128: I don't think so; confirm with mvo
[06:12] <seb128> mdz: ok, thank you
[06:12] <seb128> mvo?
[06:12] <ogra> Keybuk, thats what i was suggesting since ages, you can split it in small categories and speed up everything 
[06:12] <Keybuk> this means, for example, if we want to switch from nvidia to nvidia-legacy, we just change the mapping text files
[06:12] <Keybuk> the modules and drivers within them don't change
[06:13] <ogra> but when i suggested it there was no such thing as udev :)
[06:13] <Keybuk> or if we know of a new pci id, we just add it to the text file
[06:14] <Keybuk> plus we can have fun
[06:15] <ogra> :)
[06:16] <Keybuk> the decision about whether to bind the device to bt878 or bttv can be made by udev based on other factors, such as configuration file
[06:16] <Keybuk> or even delayed until you fire up an application
[06:16] <Keybuk> and have one driver unbound and the other bind on the fly
[06:17] <Keybuk> the principal blocker atm is that it's pretty hard to map module names to the names and buses of the drivers within them
[06:18] <Keybuk> e.g. the PCI nForce2_smbus driver comes from the i2c_nforce2 module
[06:19] <Keybuk> there's a patch to do that, but it's broken for pnp for some reason
[06:27] <mvo> seb128: yes, for edgy->feisty we have a apt that can use them so there shouldn't be a problem
[06:27] <seb128> mvo: so I can use it now?
[06:27] <Keybuk> won't that break dapper->NEXT-LTS
[06:27] <Keybuk> as dapper apt can't parse them?
[06:29] <mvo> Keybuk: yes, but we can't put them on hold until then :)
[06:29] <mvo> seb128: yes
[06:29] <Keybuk> (we actually have another interesting problem with making dapper upgradable to the next LTS; dapper is "sarge era", the next LTS must occur within "etch era".  If Debian get another release out before we get the next LTS out, we'll lose migration support (transitional packages, postinst, etc.) from Debian)
[06:30] <Keybuk> in fact, it's entirely possible that as soon as etch comes out, we'll lose the ability to update from dapper because a lot of it will have been considered "upgrade from sarge only"
[06:32] <mvo> maybe we should sync our lts with debian releases ,)
[06:39] <Surak> Does anyone know this guy which is claiming autorithy for ubuntu on turkey?
[06:43] <mjg59> Which guy?
[06:46] <Surak> which is producing noise on ubuntu-devel-list
[08:04] <comm[A|n] der> ello, is there a list of patches, which ubuntu applies to vanilla kernel for the distribution kernel?
[08:04] <Burgwork> comm[A|n] der: that is in git
[08:05] <Burgwork> https://wiki.ubuntu.com/KernelGitGuide
[08:05] <comm[A|n] der> thanks
[08:06] <highvoltage> Burgwork: did you read #ubuntu-classroom tonight?
[08:07] <highvoltage> Nat from opensuse spoke.
[08:11] <BenC> damnit
[08:11] <BenC> Mithrandir: ping
[08:14] <Mithrandir> BenC: hello
[08:14] <BenC> Mithrandir: Any way I could convince you to let my lrm and linux-meta uploads build for Herd-1 so that the new kernel is used instead of -6?
[08:14] <mdke> highvoltage: what time?
[08:14] <BenC> kernel is already built and published, and lrm/linux-meta are quick builds
[08:15] <Mithrandir> BenC: sure, sounds fine.
[08:15] <Mithrandir> BenC: when can you have them uploaded?
[08:15] <mdke> highvoltage: oh, I see it, nm
[08:15] <BenC> Mithrandir: lrm is already done and accepted, linux-meta got uploaded 5 minutes ago, so should be hitting lp soon
[08:16] <BenC> Mithrandir: linux-restricted-modules-2.6.19 2.6.19.3-2, linux-meta 2.6.19-2
[08:16] <Mithrandir> BenC: I hope you got a mail saying "the upload awaits approval by a distro manager" and not a real accept for lrm?
[08:17] <BenC> This upload awaits approval by a distro manager
[08:17] <BenC> linux-meta just got the same pending-accepted
[08:17] <BenC> so they only need to be pushed through by a distro manager :)
[08:18] <Mithrandir> I guess that means me. :-)
[08:18] <Mithrandir> I'll go find my laptop, then.
[08:26] <sfllaw> fabbione: The lvm2 bug is verified.
[08:34] <Lin> where can I setup the VGA parameter  to be set on grub after a dpkg-reconfigure linux-image-$(uname -r) (I know about edit menu.lst.. but everytime that I run dpkg-reconfigure it is reseted).
[08:35] <Lin> im build a artwork package.
[08:35] <Lin> building.
[08:35] <mdke> Lin: try #ubuntu for support
[08:36] <Lin> mdke: ok. I have tried..
[08:36] <Lin> thank you anyway :-)
[08:36] <dade`> Lin: and read carefully /boot/grub/menu.lsg around "defauloptions"
[08:37] <Lin> dade`: weird way to display default options (commented) :-)
[08:39] <Lin> Before I have searched on /etc and /var/cache/debconf .. and nothing.. thank you both!
[09:01] <keescook> Keybuk: can you (or someone?) change the MoM tarball's default changelog entry from "Merge from Debian unstable" to something like "Merge from Debian unstable.  Remaining Ubuntu changes:" to remind folks to list changes?  :)
[09:02] <joejaxx> fabbione: Herd1 still scheduled for tomorrow?
[09:02] <Keybuk> keescook: done
[09:02] <keescook> Keybuk: sweet.  :)
[09:02] <Mithrandir> joejaxx: no, I just sent an email to ubuntu-devel-announce saying that about eight hours ago.
[09:03] <joejaxx> Mithrandir: oh ok because the wiki (http://wiki.ubuntu.com/FeistyReleaseSchedule) needs to be updated
[09:03] <Mithrandir> joejaxx: huh?
[09:04] <joejaxx> Mithrandir: it says Herd 1 is scheduled for November 30th 2006
[09:05] <Mithrandir> joejaxx: I was being sarcastic.  Yes, I wouldn't have sent an email saying "we're still aiming for thursday" eight hours ago if it still wasn't true.
[09:05] <joejaxx> Mithrandir: oh ok
[09:05] <cjwatson> in general, please try not to do the "are we there yet" thing here ...
[09:06] <Mithrandir> I would recommend subscribing to ubuntu-devel-announce for, well, announcements about ubuntu development.
[09:06] <cjwatson> it doesn't help, and the release manager is generally working on it as hard as possible anyway
[09:06] <joejaxx> cjwatson: i did not mean my statement to be nagging
[09:06] <Keybuk> keescook: it'll only change new merges
[09:07] <Keybuk> ie. those already generated will stay as they are
[09:07] <joejaxx> i know how hard you all work to put out releases
[09:07] <joejaxx> Mithrandir: yeah i will i will have to setup a filter as i am subscribed to ubuntu-bugs
[09:07] <Keybuk> cjwatson: ...and prone to sarcasm when asked
[09:07] <cjwatson> bingo
[09:08] <Mithrandir> Keybuk: even more so when there has been a recent announcement of the status.
[09:08] <cjwatson> hmm, wonder if I have time to update ubiquity to the new partman-auto or if I have to hold that bit back somehow
[09:09] <Mithrandir> cjwatson: before herd 1?
[09:09] <Mithrandir> cjwatson: working late tonight? :-)
[09:09] <cjwatson> well, you need *a* ubiquity upload surely
[09:10] <Mithrandir> yeah, I do.
[09:10] <cjwatson> if nothing else to make it point sources.list at feisty
[09:11] <Mithrandir> it'd be nice if it didn't blow up if you squint at it, but that might be a bit early in the cycle to ask for
[09:14] <keescook> Keybuk: yeah, totally.  no need to regenerate everything.  :)
[09:15] <cjwatson> haha, ubiquity actually kind of works with the partman-auto questions the other way round from what it was expecting
[09:15] <cjwatson> you end up with two pages in reverse order and then it gets confused
[09:15] <cjwatson> but the half-workingness is amusing
[09:19] <jdong> cjwatson: when will ubuntu-archive get processed next?
[09:19] <cjwatson> Mithrandir: ok, I think I'm going to defer the partman merge and update everything else
[09:19] <cjwatson> jdong: Friday
[09:19] <jdong> cjwatson: ok, thanks
[09:19] <cjwatson> jdong: (or if I feel like it before then)
[09:19] <jdong> :)
[09:19] <cjwatson> or if Mithrandir feels like it. whatever
[09:21] <Mithrandir> cjwatson: ok
[09:29] <jdong> would cdbs be something sane to backport if it builds correctly?
[09:29] <jdong> (i.e. are new version backwards-compatible with older ones?)
[09:33] <cjwatson> I'd want to be testing whether the entire rest of the archive built correctly, not so much whether it did itself
[09:40] <jdong> cjwatson: yeah, point taken, hence I've been trying to dodge it for a while now.. :D... was just wondering if you experts knew of any off-the-bat issues :)
[10:10] <ere> I have a problem with VGA out and Feisty on Dell Latitude D52X laptops. I get a lot of flicker and distortion in the image when both the display of the laptop and vga out are enabled. I wonder if anyone know if intel's efforts (http://intellinuxgraphics.org/) will be included in Feisty? 
[10:10] <Burgwork> ere: they already are
[10:11] <ere> Burgwork: in the xorg i810 driver?
[10:12] <Burgwork> yes
[10:12] <Burgwork> intel has been working on that for awhile, that is more of a reannounce and reaffirmation
[10:12] <crimsun> cf. xserver-xorg-video-intel in edgy/universe
[10:14] <jdong> yeah, i810 tends to do that when VGA and LCD are both enabled....
[10:14] <jcole> is ubuntu available for arm?
[10:14] <jdong> unfortunately my 835 chipset can't run the new intel drivers without hardlocking... :(
[10:14] <ere> jdong: any suggestions for how to fix? I have 100 laptops with that problem :(
[10:15] <jdong> ere: does xserver-xorg-video-intel work for you?
[10:15] <crimsun> jdong: (fwiw Debian Sid's xserver-xorg-video-i810-modesetting works better)
[10:15] <jdong> ere: I'm still working on this one myself
[10:16] <ere> jdong: I have not tried it yet
[10:16] <jdong> ere: anyway, I'll keep on playing with it when I get some spare time...
[10:19] <ere> well, thanks for the input. I will try more tomorrow.
[10:37] <BenC> Mithrandir: thanks for the lrm/linux-meta pass
[10:54] <cjwatson> jcole: not at present
[11:24] <mdz> seb128: feisty-login is pending approval with you as approver; can you look at it?
[11:27] <seb128> mdz: looks fine to me, should we pass it by Mark to know if he likes the mockups before approval though?
[11:27] <mdz> seb128: sure
[11:28] <mdz> seb128: make sure you are OK with the implementation though
[11:30] <seb128> mdz: I've discussed with MacSlow about the implementation during UDS and I'm fine with it
[11:31] <mdz> ok, thanks
[11:33] <seb128> np
[11:44] <niktaris_> cjwatson, ntfs partitons are not added to fstab after install ?
[11:46] <cjwatson> niktaris_: they're meant to be (assuming you didn't deselect the mountpoint that should have been preselected for those partitions), and there's code for that
[11:47] <cjwatson> if they're not, and you didn't manually deselect them, something went wrong
[11:47] <niktaris_> thanks
[11:54] <decko_home> Guys, there some way to change the "notification-theme" on Ubuntu???