[12:08] <LaserJock> I believe *-live are replaced with tasks in Edgy
[12:09] <Burgwork> what interests me the most is on-the-fly task switching
[12:13] <LaserJock> Burgwork: yeah, I don't know how feasible that is right now
[12:13] <LaserJock> I don't know a whole lot about it, but it looks like most of the current tasksel stuff works on tags in the .debs
[12:14] <LaserJock> and I wouldn't think that would lend itself well to "on-the-fly"
[12:16] <mjr> a,21
[12:17] <Riddell> sfllaw: did the kubuntu winfoss get fixed?
[12:20] <Kamion> ogra: your CDs probably shrunk because linuxprinting.org-ppds was finally pulled back out of desktop post-RC, IIRC
[12:22] <tkamppeter> ogra, so it was finally decided to ship without manufacturer-supplied PostScript PPDs?
[12:22] <tkamppeter> People have complained that the PostScript printers were not fully functional.
[12:23] <tkamppeter> Kamion, or  should we split the linuxprinting-ppds package, as I suggested earlier.
[12:24] <Kamion> tkamppeter: we're not making any further changes there for edgy now
[12:25] <Kamion> the linuxprinting.org-ppds seed change was agreed and done weeks ago; it's just that we forgot to upload the metapackages for a while
[12:25] <Kamion> and it's nothing to do with ogra, so don't bug him about it
[12:27] <Quash> I'm having trouble loading the Live Desktop on 6.10 RC.  I get a black screen at the desktop stage.  No one on #ubuntu channel can figure it out.  Can anyone help?  I've also filed a bug report with more details.
[12:28] <Quash> I want to hash out the problem to provide further details for the bug report, as I think it may become a problem for a number of people when 6.10 is release Oct 26.
[12:29] <Kamion> try the safe mode at the boot screen?
[12:29] <Kamion> (which forces X to use the vesa driver)
[12:30] <Kamion> anyway, can't stay to debug, as I need to go to bed in order to get down to Canonical HQ at a sane time tomorrow morning
[12:30] <Fujitsu> Goodnight, Kamion.
[12:30] <robitaille> Quash:  what is the bug number of the report you have filed?
[12:30] <sfllaw> Riddell: Haven't gotten that CD yet.
[12:31] <sfllaw> Riddell: I think so, but only when mkisofs from heno's tarball.
[12:31] <Quash> robitaille: Bug #67487
[12:31] <Ubugtu> Malone bug 67487 in Ubuntu "Install 6:10 RC Failure: Black screen at Live Desktop Stage" [Undecided,Unconfirmed]  http://launchpad.net/bugs/67487
[12:31] <tkamppeter> Kamion, sorry, I wanted to write "Kamion" instead of "o gra".
[12:31] <Quash> Yes, in launchpad.  That's it.
[12:33] <tkamppeter> Kamion, and in the next version (7.04), we can perhaps reorganize the printer drivers on the CD completely, as there we will probably have automatic driver download from FSG OpenPrinting.
[12:42] <Kaleo> BenC: bug 56090 about the pwc drivers is stagnant
[12:42] <Ubugtu> Malone bug 56090 in linux-source-2.6.17 "Regression: Webcams using pwc driver produce all-grey images" [Undecided,Confirmed]  http://launchpad.net/bugs/56090
[12:43] <Kaleo> The bottom line is an important regression for webcams support
[12:43] <BenC> Kaleo: That's because we cannot include the "working" pwc driver due to questionable code
[12:43] <BenC> there's lots about it on the l-k mailing list
[12:43] <Kaleo> I see
[12:43] <Kaleo> I did not read the thread
[12:44] <Kaleo> Do you have any good pointer ?
[12:44] <BenC> not off hand
[12:44] <BenC> search for "pwc patent linux"
[12:44] <BenC> something along there
[12:45] <Kaleo> there was indeed a problem of NDA a while ago
[12:46] <Kaleo> but the driver was replaced by Luc Saillard's
[01:07] <sivang> jdub: you have an idea how to make headers, and how to get rid of the annoying underline although in original post only the linkified word is underlined problem in planet? (my feed is hosted on advogato)
[01:08] <jdub> i have no idea what that sentence means.
[01:10] <sivang> jdub: actually I have no idea why I wrote it that way ;)
[01:10] <sivang> jdub: so, if you take a look at my aggregated post on planet, you'll see that the header is the date which is redundent and useless, and that some text is underlined uneccessarily
[01:11] <sivang> jdub: the orig post in advogato doesn't have this, and I wonder :)
[01:11] <jdub> the date thing is due to advogato not having titles, but providing the date in the <title> element
[01:11] <jdub> yo ucan't fix that
[01:11] <sivang> ah, crap okay
[01:12] <jdub> the underline is due to <a> instead of </a> in your markup
[01:15] <sivang> jdub: this is the markup http://paste.ubuntu-nl.org/28023/ , AFAICT I haven't missed the closing anchor tags
[01:15] <sivang> oh crap, I did
[01:15] <sivang> jdub: thanks ;)
[01:28] <AlexFicelle> Hi!
[01:44] <zMott> is gaim going to get fix's
[01:48] <robitaille> looking at gaim's bug reports, 71 are currently open, 11 new filed in the last 7 days still open; 7 of those seem to be related to a crash in one form or another
[01:50] <sivang> zMott: what's your crasher?
[01:58] <zMott> sent off the bug debug dump to launchpad
[01:59] <zMott> gaim, can handle ctcp 
[02:02] <lexual> Can anyone help me out. There is now a patch available for an important epiphany bug: https://launchpad.net/products/epiphany/+bug/56610
[02:02] <Ubugtu> Malone bug 56610 in epiphany "Automatic search from address entry doesn't work anymore" [Unknown,Confirmed]  
[02:04] <neuralis> sivang: around?
[02:06] <sivang> neuralis: here, yes
[02:06] <sivang> neuralis: what's up?
[02:07] <ajmitch> neuralis: going to be at MV?
[02:07] <neuralis> ajmitch: for a few days, aye
[02:07] <neuralis> ajmitch: why?
[02:07] <ajmitch> just wondering, there's a bit of server-realted stuff to talk about
[02:07] <neuralis> certainly, yeah
[02:07] <ajmitch> including updates on network-auth
[02:08] <neuralis> how did your soc end up? i never quite figured out what the outcome was
[02:08] <ajmitch> mostly there
[02:08] <ajmitch> we'll be using it for feisty, it seems
[02:08] <neuralis> sivang: are you considering supporting things like incremental backup to an external HDD within hubackup, or do you see that as outside of the application's scope?
[02:09] <neuralis> sivang: i don't at all get excited about backing up to cd/dvd, but if i could do clicky incremental backups over ssh and/or to an external drive, that'd be rocking.
[02:11] <sivang> neuralis: I don't consider it completely out of scope, however as people noted to me there are a couple of CLI solutions that enable that, I want to look at GUI wrapping them at some point, but first I want to concentrate on delivering something for the common "non tehcnical" user use case, e.g., slices and CD/DVDs
[02:11] <sivang> neuralis: there seems to be a tool I've investigated today, written in python, called cedar-backup2
[02:11] <sivang> neuralis: might be a good candidate for a sich
[02:11] <sivang> *such
[02:12] <sivang> neuralis: it's currently in universe, though
[02:12] <ajmitch> sivang: if it's worthy, write up a MIR for feisty
[02:12] <sivang> ajmitch: yes, right after I write a MIR for hubacukp and dar ;)
[02:13] <ajmitch> there are a few things I need in main, too :)
[02:13] <sivang> hehe
[02:15] <sivang> neuralis: cedar-backup2 also supports automatic periodical backups, so that functionality could go in as well
[02:17] <neuralis> well, i think remote systems are a bit of a stretch, but backing up to an external hdd would fit nicely within hubackup, simply as a different backend.
[02:17] <neuralis> if you're open to patches, i might even provide it (with activation only if rsync is found on the system).
[02:18] <sivang> neuralis: that might already be supported using the current infrastrucutre, and if so, could be just a day of work. Would you be as so kind to see if we've got a bug about that and/or open if not?
[02:18] <sivang> neuralis: I
[02:18] <sivang> neuralis: I'm more then open for patches :)
[02:18] <neuralis> okay, i'll look into it
[02:19] <neuralis> i would imagine you don't support incremental backups, which are pretty key to making backups to an external hdd really useful.
[02:19] <sivang> neuralis: you see, even now you can choose an external hd as the backup target (using stock GNOME file selector sensitive to pluggable media)
[02:19] <sivang> neuralis: the code for making incremental backups is already there
[02:19] <sivang> neuralis: it's just needs adjustment in ordre to work using the new UI workflow
[02:19] <neuralis> ah, interesting.
[02:20] <sivang> neuralis: this already worked to some extent in the prvious version (dapper's) as I thought ot support X local to Y pluggable device backups.
[02:21] <sivang> neuralis: but during UDS Paris the team of UI and usability people and I realized we need to break this funcitonality in order to create less clutter and confusion in users.
[02:21] <sivang> neuralis: compare the dapper and edgy spec and you'll see the "log" ;)
[02:21] <sivang> woops, I've killed a kitten
[02:23] <sivang> neuralis: so now the restore functionality and diff backup is going to be integrated into the mime type association handler
[02:24] <neuralis> sivang: interesting
[02:24] <neuralis> sivang: are you going to be in mv?
[02:24] <sivang> neuralis: I wish, but no :-/
[02:24] <neuralis> ah, okay. i'll look at the code if i have a bit of time, and then we can talk about some of this again.
[02:25] <sivang> neuralis: cool, thanks.
[02:25] <neuralis> sure.
[02:29] <sivang> neuralis: if you don't find my on IRC please email me, I'd hate to miss your input on this.
[02:30] <neuralis> sivang: not a problem.
[02:46] <jjesse> has anyone been successffull booting ubuntu or kubuntu on the new intel q965 chipset? cause i am unable to
[03:07] <lastnode> imbrandon, ping please
[04:52] <zMott> quit  
[04:57] <lexual> There is a bug that I reported for hoary. The bug no longer appears in edgy. I want to close the bug, but I'm not sure what status to give it.
[04:57] <jjesse> fixed released
[04:57] <jjesse> is what i would give it
[04:58] <lexual> jjesse: even though I haven't released a fix, and am not aware of where/when a "fix" was released.
[04:58] <jjesse> welll i guess you could reject the bug, but if it was "fixed" somewhere between hoary and edgy i would asume a fix was made
[04:59] <lexual> ok, worksforme would seem to make a lot more sense if it was available.
[04:59] <infinity> lexual: Someone released a fix,even if it wasn't you.
[04:59] <infinity> lexual: What's the bug?
[05:00] <lexual> infinity: I guess it's somewhere in xorg 7.0 or 7.1
[05:00] <lexual> https://launchpad.net/distros/ubuntu/+source/xorg/+bug/12301
[05:00] <Ubugtu> Malone bug 12301 in xorg "hsync/vrefresh not remembered; monitor off at boot is bad" [Wishlist,Confirmed]  
[05:00] <lexual> now works with edgy
[05:01] <infinity> Oh, that's a messy bug with all sorts of tangents.
[05:01] <infinity> Don't clos eit, please, but just add a comment saying that things work for YOU as you expect them to.
[05:03] <lexual> cool
[05:35] <whiprush> stupid question but was the EOL for Hoary announced?
[05:36] <irvin> whiprush, yup
[05:37] <whiprush> ok, I either totally missed it, or I can't do simple arithmetic
[05:38] <ajmitch> hey whiprush 
[05:38] <whiprush> hi aj
[05:38] <ajmitch> what's up?
[05:39] <whiprush> nothing, just wondering when exactly support for hoary ended.
[05:39] <whiprush> can't find anything on u-d-a
[05:39] <ajmitch> yeah, I didn't see it announced
[05:39] <ajmitch> lp still lists it as supported
[05:39] <ajmitch> but it's past the 18 month mark
[05:40] <ajmitch> should have gone EOL on the 8th
[05:40] <whiprush> yeah just wondering
[05:40] <whiprush> since the mirror is getting some 404's, so I thought I might have been mistaken
[05:40] <irvin> it was announced on ubuntu-security-announce
[05:41] <whiprush> there it is, thanks irvin 
[05:41] <ajmitch> so someone just needs to flick the big switch on launchpad as well
[05:42] <whiprush> in the past there was an announcement and LWN story, etc. etc.
[05:42] <whiprush> probably why I missed it
[05:42] <wasabi> So, I'm fighting the unwinnable battle. Anybody wanna help? :)
[05:42] <ajmitch> whiprush: http://old-releases.ubuntu.com/releases/
[05:42] <ajmitch> wasabi: which one?
[05:42] <wasabi> New NSS table: realm
[05:42] <ajmitch> ah
[05:42] <wasabi> I'm patching up libc to support it now, to make sure my heads on straight.
[05:43] <whiprush> ajmitch: I know where it is. :D I was just serving out 404's to some people from my mirror looking for hoary.
[05:43] <ajmitch> heh ok :)
[05:43] <whiprush> which reminded me of "hey, I don't remember a mail ..."
[05:43] <ajmitch> sorry, I haven't read the u-s-a mail
[05:44] <ajmitch> and the mail states it goes EOL on the 31st, nice
[05:45] <ajmitch> wasabi: briefly explain why you must have this new table :)
[05:45] <wasabi> MOstly because it's the most preferable way to solve the problem. The long discertation is on libc-alpha
[05:45] <ajmitch> yeah, that's why I said briefly
[05:45] <wasabi> http://sourceware.org/ml/libc-alpha/2006-10/msg00076.html
[05:45] <wasabi> =)
[05:45] <ajmitch> it's a long email for me to read :)
[05:46] <wasabi> For cross realm auth.
[05:46] <wasabi> The first one is "how do I display a drop down of realms in GDM?"
[05:46] <wasabi> The second one is "how do I not make every nss lookup contain every remote LDAP server in a large enterprise".
[05:46] <wasabi> s/contain/contact/
[05:47] <wasabi> Drop down could be hacked into GDM hard coded. It would just be fugly.
[05:47] <wasabi> The second would require circumventing nss for every app.
[05:47] <wasabi> Which obviously is not ideal.
[05:47] <ajmitch> definitely not a good solution to just bypass nss
[05:48] <ajmitch> or hack it into gdm
[05:48] <wasabi> It's not a solution at all.
[05:48] <ajmitch> since you have to hack it into everything that needs it
[05:48] <ajmitch> ok, I managed to wade through that email
[05:48] <wasabi> You should read the post. :)
[05:48] <ajmitch> no replies to it yet?
[05:48] <wasabi> Nope. Just posted it today though.
[05:49] <wasabi> It really is the proper solution. This probably is at the core of nss...
[05:49] <wasabi> how do you seperate lists of users and groups by some named division of management.
[05:49] <ajmitch> yeah, your suggestion sounds sane at a glance
[05:50] <wasabi> Also, we still must hack user@REALM names.
[05:50] <ajmitch> though how much opposition there'll be to extending NSS is yet to be seen
[05:50] <wasabi> Part of this will also be "fixing" every app that stores any thing long term not using the uid.
[05:50] <wasabi> Yeah.
[05:50] <jelmer> nss is somewhat standard so it'll be a lot harder to modify it
[05:50] <jelmer> then to extend it
[05:50] <ajmitch> hey jelmer 
[05:51] <jelmer> s/then/than/
[05:51] <wasabi> Yeah. It looks like people have been okay with adding tables over the years.
[05:51] <wasabi> Just not with modifying any of the structures.
[05:51] <wasabi> So, passwd is not changing, period.
[05:51] <jelmer> hey ajmitch, wasabi 
[05:51] <ajmitch> how would you lookup users in a realm though?
[05:51] <wasabi> jelmer: Samba guy?
[05:51] <wasabi> ajmitch: New functions like getpwent
[05:51] <wasabi> getpwent_rlm or somethign
[05:51] <jelmer> wasabi: jup
[05:52] <ajmitch> wasabi: he's our friendly samba contact :)
[05:52] <wasabi> Probably make up a "realm id"
[05:52] <jelmer> is there any reason why workstations would generally have more than one realm?
[05:52] <wasabi> Yes. cross realm auth.
[05:52] <jelmer> or are you going the Apple way and installing a KDC on every workstation ? ;-)
[05:52] <wasabi> jelmer: You've seen Windows on an AD in a large forest?
[05:52] <wasabi> On the login screen you have a choice of this computer and all of the realms.
[05:53] <jelmer> wasabi: right, but the main difference is
[05:53] <wasabi> But you can also access resourses on those remote realms.
[05:53] <wasabi> And programs need to be able to resolve SIDs 
[05:53] <jelmer> they have SID's to identify users, we only have uint32s
[05:53] <wasabi> Yup.
[05:53] <wasabi> That sucks.
[05:53] <wasabi> The best and only solution I can come up with is name spacing the uid space.
[05:53] <wasabi> https://wiki.ubuntu.com/NetworkAuthentication/ScratchPad  <--- there
[05:54] <jelmer> that would have to happen in the file system as well though :-/
[05:54] <wasabi> Nope.
[05:54] <wasabi> Same 32 bits.
[05:54] <wasabi> We have limited space compared to windows though, but I've taken a look at the AD setups I've seen, and don't think we're hitting a wall.
[05:54] <wasabi> Who will have more than 4096 realms?
[05:54] <wasabi> heh
[05:54] <fabbione> morning
[05:55] <wasabi> jelmer: Of course, when we're joining to AD, the new 2003 creates unique UIDs for us apparently.
[05:55] <wasabi> Not sure if that is cross forest though.
[05:55] <ajmitch> morning fabbione 
[05:55] <jelmer> wasabi: Ugh
[05:55] <wasabi> I suspect it will be, and stored in teh GAC.
[05:55] <ajmitch> jelmer: I see you got tp3 of samba out :)
[05:55] <wasabi> jelmer: Yeah. Well. Other suggestions? :)
[05:56] <wasabi> I think this is hte best I can do with 32 bits.
[05:56] <wasabi> jelmer: 1048576 users, which isn't that big.
[05:57] <wasabi> But I've never seen a single domain with that many, yet.
[05:57] <wasabi> You start having to forest for other reasons way before then.
[05:57] <Seq> does anybody know the proper way to kill "maintainer-script-calls-init-script-directly"?
[05:57] <ajmitch> Seq: invoke-rc.d in the maintainer scripts?
[05:58] <wasabi> And the likely hood of a collision in a merge of two companies is "low", with random 11 bit domain ids
[05:58] <wasabi> 2048 I mean, actually.
[05:58] <jelmer> wasabi: I still think it's plain ugly though.. 
[05:58] <wasabi> I agree.
[05:58] <wasabi> But it needs to start happening.
[05:58] <wasabi> We need to get our systems into this space before people will fully understand the problem.
[05:58] <jelmer> Well, do you really want ot be able to select realms that way?
[05:58] <Seq> ajmitch: thank you very much
[05:58] <wasabi> jelmer: I want to be able to access cross realm resources.
[05:58] <wasabi> Including file shares, such as NFS.
[05:58] <jelmer> before hacving proper kerberos integration?
[05:58] <wasabi> And SSH.
[05:58] <wasabi> Especially SSH.
[05:59] <zMott> using irc, and was wondering what is the best irc client for ubuntu?
[05:59] <wasabi> jelmer: This is part of that.
[05:59] <wasabi> They all go together.
[05:59] <jelmer> wasabi: I can see the use for that, I really do. 
[05:59] <jelmer> I'm just wonder whether this might be a bridge too far for initial kerberos/ldap support
[05:59] <wasabi> Also, I do not honestly believe we will be backing outselves into a corner.
[05:59] <wasabi> Since we are already in the same corner.
[06:00] <wasabi> At some point, somebody will be unsatisfied with static length UIDs, and they'll have to fix it.
[06:00] <wasabi> But that is not a change that is possible to push at this point.
[06:00] <wasabi> Too much politics.
[06:01] <jelmer> Yeah, that's /a lot/ of politics
[06:01] <jelmer> as it requires richer (non-POSIX) interfaces
[06:01] <wasabi> I think the addition of a simple realm table, some functions to qualify searches of passwd and group based on realm, and a more robust libnss-ldap, that take that into consideration, can get us a long way.
[06:01] <wasabi> And I think it's possible.
[06:01] <wasabi> Where libnss-ldap can be configured to locate remote ldap servers of remote realms.
[06:01] <wasabi> And query those for remote users.
[06:02] <jelmer> and so you're going to add a bitprefix to remote users?
[06:02] <jelmer> to prevent clashes with local uids?
[06:02] <wasabi> To any network user, yes.
[06:02] <wasabi> This will be defined in libnss though, so you can override it just like you can now.
[06:02] <wasabi> libnss-ldap that is
[06:03] <jelmer> what about matching uids on NFSv4? 
[06:03] <jelmer> or winbind-created uids?
[06:03] <wasabi> Doesn't really solve the problem except with NFSv4.
[06:03] <wasabi> tars need to work.
[06:03] <wasabi> ya know?
[06:03] <jelmer> tars?
[06:03] <wasabi> backups of user data.
[06:03] <wasabi> You need to be able to extract it and retain perms.
[06:04] <jelmer> NFSv4 has its own way of sharing uids across the network and so has Samba
[06:04] <wasabi> Yeah. I know.
[06:05] <wasabi> Well, do YOU have a domain with more than a million user accounts?
[06:05] <wasabi> Also are you going to be at umv? :)
[06:05] <Fujitsu> wasabi, I would presume a number of enterprises do.
[06:05] <wasabi> Me too, and I *suspect* they are cross realm.
[06:05] <wasabi> Which is only be being unable to imagine a building that can fit a million people.
[06:05] <wasabi> s/be/me
[06:06] <jelmer> wasabi: no, but even if I did - this breaks winbind and NFSv4 mounts across hosts
[06:06] <wasabi> How so?
[06:06] <Fujitsu> Maybe, but not necessarily. It's very possible to have a cross-campus domain.
[06:06] <wasabi> winbind doesn't need to assign uids.
[06:06] <wasabi> NFSv4 can map, on string name, and still works.
[06:06] <jelmer> no, but it needs globally shared uids
[06:07] <wasabi> globally?
[06:07] <jelmer> well, it needs them to be shared globally if you want matching uids on differnet workstations
[06:07] <wasabi> Yes. How does this break that?
[06:07] <wasabi> It actually circumvents it.
[06:08] <jelmer> because if somebody copies a file foo from host a to host b, host b will mask the uid for that file
[06:08] <wasabi> libnss-ldap is used to deliver the "enoughily" unique UID to the workstation.
[06:08] <jelmer> wasabi: I won't be in mv unfortunately, too busy with uni :-/
[06:08] <wasabi> The tools for "our directory server" will generate that based on the namespacing of the uid.
[06:08] <wasabi> Everybody elses still has to do something.
[06:09] <ajmitch> jelmer: unfortunate, would have been good to have you there
[06:09] <wasabi> I actually do this at work now, just by hand.
[06:09] <wasabi> Yeah, i'd love a samba person there.
[06:09] <wasabi> Do any work/live in the area?
[06:09] <wasabi> This obviously is core to what they're doing.
[06:09] <jelmer> Yeah, Jeremy (who is with Novell) lives in San Jose
[06:09] <whiprush> anyone in google?
[06:10] <whiprush> jelmer: you think he'd be interested in dropping by?
[06:10] <wasabi> idra already gave me a mouthful about how I should just use winbind. I do understand where he's coming from, but there is a lot I can't ignore.
[06:10] <wasabi> Like, the whole, we'd like this to work on fully free LDAP/KDC configs, and keep about the same.
[06:10] <wasabi> Where winbind is pretty irrelivent.
[06:11] <jelmer> wasabi: I don't think you sould be using winbind, mainly just that
[06:11] <jelmer> using uids this way is going to come back to haunt you
[06:12] <wasabi> I know. But not using them at all doesn't solve a problem.
[06:12] <wasabi> Forced into a corner.
[06:12] <wasabi> I don't take this lightly. =)
[06:12] <wasabi> I know, somebody somewhere is going to hit a duplicate rid, if this takes off.
[06:12] <wasabi> Less likely a duplicate uid.
[06:12] <wasabi> But it's going to happen.
[06:13] <jelmer> Well, people will complain if two users with the same name turn out to have diffent uids
[06:13] <wasabi> Well, not likley to happen...
[06:13] <wasabi> username@REALM
[06:13] <wasabi> Forced into that corner too. ;)
[06:13] <jelmer> Well, there's no way to merge realms, right?
[06:14] <wasabi> Define merge?
[06:14] <jelmer> combine them to be one
[06:14] <wasabi> Technically it would operate the same as MS's does.
[06:14] <wasabi> They have a tool.
[06:14] <wasabi> Regenerates SIDs, I believe.
[06:15] <wasabi> Basically just scripts making new users in the new realm.
[06:15] <jelmer> It'll probably define an alias in the new domain pointing at the old sid
[06:15] <wasabi> Hmm.
[06:15] <wasabi> reading.
[06:15] <jelmer> so that you can still access your old files
[06:16] <wasabi> Yup, you're right.
[06:16] <wasabi> Well, this is why I say that creating our own directory server is going to be a multi year task.
[06:16] <wasabi> And that we can tackle the client side a bit easier.
[06:16] <wasabi> There is SO MUCH to consider in a fully directory service.
[06:16] <jelmer> whiprush: jeremy might be interesting in attending, though I'm not sure what Novell's policy is on attending Ubuntu conferences :-)
[06:16] <jelmer> whiprush: And that's why I think cross-realm authentication should be step (2)
[06:17] <jelmer> single-realm authentication and integration is going to be useful for a lot of people
[06:17] <whiprush> jelmer: I don't mind sending a mail. I was more interested in Jeremy the Samba Hacker more than Jeremy the Novell Employee. :D
[06:17] <wasabi> I just work in too many places who would be really disappointed by not being able to access remote resources properly.
[06:18] <jelmer> wasabi: Is there at least going to be an option to turn off this behaviour?
[06:18] <wasabi> And, I guess, I'd rather get forests using unique UIDs now, vs planning to create conflicting ones now.
[06:18] <wasabi> jelmer: All I'm doing is standardizing a bunch of methodologies and fixing a bunch of existing tools, which will set this up. libnss-ldap needs love, libpam-krb5 needs love.
[06:19] <wasabi> We need a realm table for the thing to be useful.
[06:19] <wasabi> libnss-ldap will still have as many configuration options as it does now.
[06:19] <jelmer> As long as I can turn off the uid trick, I'm happy :-)
[06:19] <wasabi> Dunno what you mean by that?
[06:19] <wasabi> This is something that happens on a server that nobody has created yet. :)
[06:19] <wasabi> The uid trick has nothing to do with the client.
[06:20] <wasabi> Hell, you can put whatever you want into your LDAP.,
[06:20] <jelmer> sorry, s/can/will be able to/
[06:22] <jelmer> wasabi: what exactly would this be useful for for remote SSh though?
[06:22] <wasabi> Tying the thing together, so admins don't have to go through mental gymnastics to copy files.
[06:22] <wasabi> You put a file on a share, you expect to be able to ssh someplace and get it.
[06:23] <jelmer> sure, but that's kerberos cross-realm authentication - why does that need uids?
[06:23] <wasabi> Because the system you log into needs to map your login name to a uid, in order to create said files.
[06:23] <jelmer> I can copy stuff to foo@someothercompany as long as I've got jelmer@VERNSTOK.NL in .k5login on my account there
[06:24] <jelmer> but you'll be using the uid of your account on the remote server anyway, that's the nature of ssh
[06:24] <jelmer> or am I misunderstanding you?
[06:25] <wasabi> Well, I don't know if you're seeing the entire picture. I don't really mean "ssh works!" as much as "assumptions made by the admin using ssh hold true as they should."
[06:25] <wasabi> tar czvpf somestuff; scp somestuff me@remotebox.remotecompany.com:/tmp
[06:26] <jelmer> if you do an scp that way, then the remote file will be owned by 'me' anyway
[06:26] <wasabi> Sure, but the stuff in it won't.
[06:26] <jelmer> not by whoever you are on the local host
[06:26] <jelmer> right
[06:26] <jelmer> so if you locally have files that are owned by root (0) that you tar
[06:26] <jelmer> and then upload
[06:27] <jelmer> then the remote root user would own the files - how's uid mapping going to solve that?
[06:27] <jelmer> or is root's uid going to be != 0
[06:28] <wasabi> Solve what? That worked as intended.
[06:28] <wasabi> There's no uid mapping...
[06:28] <jelmer> ok, so if that worked as intended, what are you trying to solve?
[06:28] <jelmer> sorry for being ignorant :-)
[06:28] <wasabi> Well, I don't know why you insist on tarring files owned by root.
[06:28] <wasabi> When I make files I tar files owned by users.
[06:29] <Fujitsu> jelmer: The other UIDs on the system need to be consistent throughout the network.
[06:29] <wasabi> For backups that is.
[06:29] <jelmer> Fujitsu: Yes, but adding a prefix to the uid isn't going to solve that
[06:29] <wasabi> Of course not. But it eases the job of keeping them unique for any given 2 domains that want to interact.
[06:29] <wasabi> You have a 1/2048 chance of getting screwed.
[06:29] <Fujitsu> If we don't have the prefix, then massive conflicts will occur, and the world will explode.
[06:30] <wasabi> Which is better than a "everybody start numbering at 10000" like they do now.
[06:30] <fabbione> hey jdub 
[06:30] <jelmer> 1/2048 is pretty high imho
[06:31] <wasabi> I agree. =)
[06:31] <wasabi> But it's better than 1/1
[06:31] <Fujitsu> jelmer: it is very high, but there's not much that can be done with 32 bits :(
[06:31] <jelmer> I'd rather be screwed constantly than occasionally, that way you know what you're up against
[06:31] <wasabi> It is, honestly, the best we can do.
[06:31] <jdub> morning fabbione!
[06:31] <wasabi> Haha.
[06:31] <wasabi> That's an interesting position to take.
[06:32] <Fujitsu> Until we want to take the non-trivial step of moving to 64-bit or so UIDs, 1/2048 is really the best that can be done.
[06:32] <wasabi> Fujitsu: MS has... "unlimited" length SIDs
[06:32] <jelmer> wasabi: also, and this is the samba dev in me talking, what if both realms talk to a 3rd realm ADS realm and both run winbind?
[06:32] <Fujitsu> wasabi, how do they manage that!?
[06:32] <wasabi> Fujitsu: The structure contains a length.
[06:33] <wasabi> Fujitsu: Each workstation does have a local ... 32 bit uid it uses, keeps a kernel map of SID->UIDs that is basically cleared on reboot.
[06:33] <wasabi> UIDs never leave a process.
[06:33] <wasabi> SIDs are used for all storage and communication.
[06:33] <wasabi> We're not in that situation.
[06:34] <Fujitsu> True...
[06:34] <wasabi> Fujitsu: the first domain you create gets basically a "domain component" of the sid, users created in it append to it.
[06:34] <wasabi> sub domains extend the firsts sid
[06:34] <jelmer> the limit is 16 rids in one sid if I'm not mistaken
[06:34] <wasabi> Oh. Well, that's ungodly huge.
[06:35] <jelmer> yep, and they're globally unique...
[06:35] <jelmer> a 
[06:35] <wasabi> As much as possible. =)
[06:35] <wasabi> ie, every RID is very large and very random.
[06:35] <jelmer> well, some of them are hardcoded
[06:36] <jelmer> such as the BUILTIN domain, administrator user
[06:36] <wasabi> jelmer: I guess what I hope to solve with this, is to get a basis for where we're heading towards... which is either going to be a massively huge UID, or a better schem.
[06:36] <Fujitsu> It wouldn't be that difficult to use the MS idea, would it? It's non-trivial to adjust the UID size throughout every application that uses UIDs, but it wouldn't be difficult to make FSes take larger ones; ACLs are already abitrary in length...
[06:36] <wasabi> I want to solve the client side, because I believe it's possible. ;)
[06:36] <wasabi> The "uid namespacing" is something that needs to be considered when we start our own LDAP server infrastructure.
[06:36] <jelmer> Fujitsu: fixing the applications is really really a lot of work
[06:36] <wasabi> Until then, the admin of the LDAP server is responsible for creating unique UIDs.
[06:37] <wasabi> WIndows 2003 R2 now does it for us automatically.
[06:37] <jelmer> Fujitsu: + upgrading the existing fs installs
[06:37] <jelmer> Fujitsu: and breaking posix compliancy
[06:37] <Fujitsu> jelmer: I said it's non-trivial to do the applications...
[06:37] <Fujitsu> But it has to be done at some point.
[06:37] <Fujitsu> (the FS stuff)
[06:37] <wasabi> I want to be wise to the policitcal situation.
[06:37] <wasabi> Current Linux people simply do not understand.
[06:37] <jelmer> Fujitsu: I think nontrivial even is an understatement :-)
[06:37] <wasabi> Seriously.
[06:37] <Fujitsu> We can't live with 32 bits for ever, I don't think.
[06:38] <wasabi> I want to get our systems into those situations, where we're pushed, and people are forced to think.
[06:38] <wasabi> When Canonical gets over a million users in a companies domain on their LDAP server, they'll have the clout to do such a change.
[06:38] <wasabi> They do not now. Nobody does.
[06:38] <jdong> when mvo wakes up, tell him I said thanks for dist-upgrader... that thing is just amazing :)
[06:39] <jelmer> wasabi: I agree it's a bad scheme at the moment
[06:39] <wasabi> I think it solves more problems than the local winbind idea.
[06:39] <wasabi> Additionally, the winbind idea simply doesn't work on non MS domains.
[06:39] <Fujitsu> Local winbind is really hackish, IMO.
[06:39] <jelmer> winbind isn't supposed to fix this sort of thing
[06:39] <wasabi> Yeah, I know.
[06:40] <wasabi> But it sort of makes the problem go away.
[06:40] <wasabi> You join, and it works, until you try to do something that requires matching UIDs.
[06:40] <wasabi> Most admins of course have NFI what that even means.
[06:40] <jelmer> you can share sid->uid mappings with winbind
[06:40] <jelmer> they can be in LDAP
[06:41] <wasabi> Fujitsu: Take NTFS. It has a table of SID->NTFSID mappings. Basically each FS has it's own concept of a unique user.
[06:41] <wasabi> That's a static length, because allocating blocks of an huge SID is not efficient.
[06:42] <wasabi> But userspace never sees that, ever.
[06:43] <jelmer> another solution, one which we're pursuing in Samba4
[06:43] <jelmer> is optionally storing a SID in an xattr when possible
[06:43] <wasabi> That's interesting.
[06:43] <wasabi> But I really don't want SIDs on the Unix box.
[06:43] <Fujitsu> jelmer: I was thinking of using them, because they have arbitrary length.
[06:44] <jelmer> wasabi: well, you can use whatever you want - for example, I think using the UUID of the machine they come from (mac address?) or something would make sense
[06:44] <jelmer> not sure how well tar et al cope with xattrs though
[06:44] <ajmitch> some tar variants have support for them
[06:45] <Hobbsee> hey ajmitch 
[06:45] <ajmitch> it's needed for things like selinux labelling
[06:45] <ajmitch> hello Hobbsee 
[06:45] <Fujitsu> Afternoon Hobbsee.
[06:45] <jelmer> ah, right
[06:45] <Hobbsee> hey Fujitsu 
[06:46] <wasabi> whiprush: Does that gossip/telepathy thing work?
[06:46] <jelmer> anyway, time to get to uni - bye!
[06:46] <wasabi> bye!
[06:46] <wasabi> thanks for the convo!
[06:46] <ajmitch> jelmer: see you later :)
[06:46] <Fujitsu> Bye, jelmer.
[06:48] <wasabi> libc compiles slowely. =(
[06:48] <Fujitsu> wasabi, what are you doing to it?
[06:48] <wasabi> adding a new nss table
[06:48] <Fujitsu> Oh, you're actually doing that?
[06:48] <wasabi> yeah
[06:49] <wasabi> seems to not be the hardest thing in the world to do
[06:49] <Fujitsu> (that's quite an impressive email you sent to libc-alpha)
[06:49] <wasabi> people read that list?
[06:49] <ajmitch> wasabi: sure
[06:49] <ajmitch> isn't it fun reading?
[06:50] <wasabi> I'd rather clean my ears.
[06:51] <Fujitsu> Wouldn't it be nice if we could migrate easily to, say, 64-bit UIDs in everything?
[06:52] <ajmitch> Fujitsu: you'd find it'd be quite painful
[06:52] <Fujitsu> ajmitch: there was an `if' there...
[06:52] <wasabi> Well, it took Linux hitting the 16 bit mark before anybody did the migration to 32 bit.
[06:52] <wasabi> Lets learn from that, nobody is going to do shit until we hit the 32 bit mark.
[06:52] <wasabi> ANd that's "ok"
[06:53] <ajmitch> and you can see that 32 bits just isn't enough for everyone
[06:53] <Fujitsu> wasabi, we've hit a barrier here because of that limit.
[06:53] <wasabi> Because of 32 bit limit?
[06:53] <Fujitsu> Yes. You can't easily split it into two reasonably-sized pieces.
[06:53] <wasabi> We're imagining we will hit it eventually, in an IRC channel.
[06:53] <wasabi> Which is quite different from a customer dropping our product because of it.
[06:54] <Fujitsu> When did Linux migrate to 32 bit?
[06:55] <wasabi> Looks like 2.3
[06:55] <Fujitsu> Oh, so not too long ago...
[06:55] <Fujitsu> That's only 6 or 7 years...
[06:55] <wasabi> Well, wish somebody had spoken up about realms back then. It's a very simple concept, and has been done for ages.
[06:55] <wasabi> Look at NT4 if you want. =/
[06:55] <Fujitsu> Yeah :(
[06:57] <Fujitsu> What to do, what to do...
[06:57] <wasabi> Well, I wish I could get realm put into passwd cleanly, but that's obviously not happening.
[06:57] <wasabi> Hey wonder if gecos has a field for that.
[06:57] <Fujitsu> Why not?
[06:58] <wasabi> passwd is posix.
[06:58] <wasabi> Extending it would break "everything"
[06:59] <Fujitsu> Hm...
[06:59] <Fujitsu> We need a POSIX-ng!
[06:59] <wasabi> Well, no reason to change the semantics of passwd... just add another table with new semantics.
[07:00] <wasabi> If there are APIs to query a list of users within a given realm, and a nss module can translate that into a DNS SRV lookup on a remote DNS server, and a LDAP query on that remote server, all's good.
[07:00] <Fujitsu> What would the format of that table be?
[07:00] <wasabi> name:rid:display name
[07:01] <wasabi> maybe just name:rid
[07:01] <Fujitsu> Yeah, I see no point for display name.
[07:01] <Fujitsu> That'd be really nice to put into passwd :(
[07:01] <wasabi> Could learn from past lessons and add like, 8 string fields to it.
[07:01] <wasabi> JUST IN CASE
[07:03] <wasabi> So, I expect some new method like... setpwent_rlm(rid)
[07:03] <wasabi> Or similar.
[07:06] <wasabi> me->bed
[07:06] <wasabi> night
[07:07] <ajmitch> night wasabi 
[07:07] <Fujitsu> 'night, wasabi.
[07:28] <BHSPitLappy> GRR.
[07:28] <fifteenrabbits> yeah?
[07:30] <BHSPitLappy> dapper went windows on me.
[08:08] <Sp4rKy> hi there
[08:08] <Sp4rKy> i need some information about the ubuntu user creation process in the live CD
[08:09] <Kamion> apt-get source casper and look in scripts/casper-bottom/10adduser
[08:09] <Sp4rKy> Kamion: i'll do :)
[08:10] <Sp4rKy> in fact, i want know if the process uses /etc/skel when it creates the ubuntu user
[08:10] <Kamion> it uses user-setup to do the hard work (apt-get source user-setup; it's an installer component)
[08:10] <Kamion> it uses adduser, so yes
[08:10] <Sp4rKy> ok
[08:10] <Kamion> (user-setup uses adduser under the hood)
[08:10] <Sp4rKy> because i'm "forking" the UBuntu live CD to an Ebuntu version
[08:10] <Sp4rKy> and i need the ubiquity icon in the bar
[08:11] <Kamion> the same script handles that for other versions, actually
[08:11] <Kamion> by explicit copying of desktop files
[08:12] <Kamion> copying to some other enlightenment-specific location, or prodding enlightenment in some way or whatever would be ok
[08:12] <Sp4rKy> yep
[08:12] <Kamion> I recommend against putting stuff specific to ubiquity in /etc/skel, because it will then be hard to rip it out for the installed system
[08:12] <Sp4rKy> i've just need to create all the .e rep in /etc/skel
[08:13] <Kamion> I recommend against that. easier to just shove it into the bar in 10adduser
[08:13] <Sp4rKy> i don't understand ?
[08:13] <Kamion> as in, after the user has been created, create the files in its home directory, appropriately owned
[08:13] <Sp4rKy> (i'm french, and my english isn't good) :/
[08:13] <Sp4rKy> ok
[08:13] <ajmitch> Sp4rKy: are you using proper packages for ebuntu this time?
[08:13] <ajmitch> hey pitti 
[08:13] <Kamion> you have to create those files dynamically anyway, as they should not end up on the installed system so they shouldn't be in the squashfs
[08:14] <Sp4rKy> ajmitch: i use our package
[08:14] <ajmitch> Sp4rKy: I mean ones that aren't made with checkinstall :)
[08:14] <Kamion> so you might as well just create them directly in the user's home directory rather than messing about with /etc/skel
[08:14] <jsgotangco> ajmitch: i dont think he's the guy who did the checkinstall
[08:14] <Sp4rKy> ajmitch: :)
[08:14] <ajmitch> jsgotangco: I know he's not :)
[08:14] <Sp4rKy> ajmitch: we try
[08:14] <Sp4rKy> Kamion: yes ..., so create it in the 10adduser script ?
[08:15] <pitti> Good morning
[08:16] <Sp4rKy> hi pitti 
[08:17] <Sp4rKy> another question :
[08:17] <Sp4rKy> the bin/ and program/ rep only contains windows applications, right ?
[08:18] <pitti> hi Sp4rKy, moin ajmitch 
[08:20] <Sp4rKy> Kamion: where are the casper script in the live ?
[08:20] <Sp4rKy> i don't find them :/
[08:21] <infinity> Sp4rKy: In the initramfs.
[08:21] <Sp4rKy> :)
[08:22] <Sp4rKy> infinity: initrd maybe, no ?
[08:22] <infinity> Well, the filename has "initrd" in it, for backward compatibility with bootloaders, but I assure you it's an initramfs. :)
[08:24] <Sp4rKy> ok :D
[08:26] <Sp4rKy> i have to go
[08:26] <Sp4rKy> thx for your help :)
[08:33] <tfheen> pitti: I have accepted the ubuntu-docs upload.  Once that is in, please do fill up the CDs with langpacks.
[08:33] <tfheen> pitti: but make sure to leave a little space, I'd hate having to reroll because you were too optimistic. :-)
[08:33] <pitti> tfheen: is 5 MB okay?
[08:33] <tfheen> that's plenty
[08:33] <pitti> or do you expect larger changes still?
[08:33] <tfheen> I was thinking more like 1-2MB
[08:33] <infinity> 5MB is a lot.
[08:34] <pitti> tfheen: what about oo.o-l10n? will there still be an upload?
[08:34] <infinity> doko_: ?
[08:35] <tfheen> pitti: what about oo.o-l10n?  I don't have any email about that and can't find it in my scrollback
[08:35] <pitti> tfheen: ISTR that there was another upload planned, but I didn't see one
[08:36] <tfheen> pitti: maybe I'm just ignorant or something, but I can't remember about it.
[08:36] <tfheen> s/about//
[08:36] <tfheen> I can't write English either.
[08:37] <pitti> tfheen: ok, I'll do the langpacks now, and *if* there should really be another oo.o upload, I'll do immediate adjustments
[08:37] <tfheen> there won't be a full oo.o upload, that's for sure.
[08:37] <tfheen> waaay to late.
[08:37] <pitti> (-l10n, I meant, sorry)
[08:37] <pitti> even that is likely too late anyway
[08:38] <tfheen> I was hoping to have the images done last night be the final ones, code-wise.
[08:44] <pitti> wow, there's really plenty of space on the desktop CDs
[08:45] <highvoltage> :)
[08:46] <Burgundavia> have we removed gthumb yet?
[08:46] <pitti> ogra: shall I fill up edubuntu CDs for you with langpacks as well, or do you have other plans for that?
[08:47] <pitti> ogra: and do you aim for 650 MB or 700 MB images?
[08:48] <tfheen> pitti: he's been doing 700MB images forever, afiak
[08:48] <tfheen> afaik, even
[08:49] <pitti> funny, ogra has always had the tightest images, and now there are 50/86/66 MB free on the lives
[08:50] <jsgotangco> yes i was a bit confused when i saw the build today
[08:50] <tfheen> I'm not exactly sure what's happened there.
[08:50] <jsgotangco> "we got space"
[08:50] <jsgotangco> heh
[08:50] <tfheen> I just hope we haven't run into a bug in germinate
[08:51] <pitti> tfheen: well, the images have been that small for quite some time now
[08:58] <tfheen> pitti: you're telling me when you're done?
[08:58] <pitti> tfheen: yes, I will
[08:58] <tfheen> thanks
[09:01] <Kagou> hi
[09:15] <pitti> tfheen: ok, I adjusted ship and live accordingly; hm, AFAICS this doesn't even require a -meta upload?
[09:16] <tfheen> it doesn't, afaik.
[09:16] <tfheen> it requires the publisher to run so it picks up the new Task headers.
[09:16] <infinity> publisher's doing it's thing in a sec.
[09:17] <infinity> There, publisher going.
[09:18] <pitti> tfheen: right, ./update shows 'no changes found' after updating live, and ship has never required a -meta upload
[09:18] <pitti> tfheen: Kubuntu CDs are full, and I'm not sure about edubuntu; I'd rather wait for ogra
[09:20] <tfheen> I can do ubuntu first.
[09:20] <tfheen> once the publisher is done
[09:27] <doko_> infinity: ?
[09:32] <infinity> doko_: Were there plans for a final OOo-l10n upload, or is what's in the archive what we're shipping with?
[09:33] <infinity> pitti: We removed the *-live packages, so we could mangle livecds with nothing but seed changes.
[09:33] <pitti> infinity: sweet
[09:33] <Burgundavia> infinity: you planning on nuking the meta packages altogether?
[09:34] <infinity> Burgundavia: No, just *-live, since it's not an end-user package.
[09:34] <Burgundavia> right
[09:35] <infinity> Running germinate... .........
[09:36] <doko_> infinity: I do have a new export from rosetta, and updates for eo and ku. but ... I do have to make OOo packages for -proposed/-updates anyway. 
[09:37] <infinity> So we'll call what's in the archive final, then?
[09:38] <infinity> tfheen: publisher done and back in full-auto.
[09:38] <infinity> tfheen: And no new OOo-l10n, apparently.
[09:38] <infinity> tfheen: So, let the CDs fly.
[09:39] <tfheen> infinity: hooray, thanks.
[09:40] <tfheen> asking because my brain just stopped working: I need to rebuild the livefs-es, right?
[09:44] <infinity> tfheen: Yes, pitti updated live.
[09:44] <infinity> tfheen: And ubuntu-docs is in desktop (which is on the livefs)
[09:45] <pitti> tfheen: ^ if not, then I need to adjust the live seed again
[09:45] <infinity> Vaguely similar.
[09:45] <pitti> I only have hard data for .deb sizes, so alternates should get fine
[09:45] <infinity> squashfs should be a bit better in some cases, since it's got largers streams of data to work with.
[09:45] <pitti> I'm only concerned about the 'squashfs gets it worse' case, then I ruined a livefs build
[09:46] <infinity> And locale data is highly repetitive, so compressing it all together instead of per-language should yeild better results, one would think.
[09:46] <tfheen> pitti: you'll get some numbers in a little bit
[09:46] <pitti> right
[10:05] <dholbach> good morning
[10:05] <janimo> hi dholbach
[10:06] <BHSPitLappy> hooooooo
[10:07] <dholbach> heya janimo
[10:13] <tfheen> ubuntu, kubuntu alternate up
[10:28] <tfheen> edubuntu install images up
[10:29] <fabbione> tfheen: is ubuntu server up?
[10:31] <tfheen> fabbione: it was up last night.  Langpacks are slightly less interesting to you, I guess.
[10:31] <fabbione> tfheen: yeah just making sure
[10:31] <dholbach> hey seb128
[10:31] <seb128> hi dholbach
[10:32] <dholbach> I find bug 66769, bug 66779 and bug 67784 slightly worrying
[10:32] <Ubugtu> Malone bug 66769 in ubiquity "Ubuntu 6.10 Desktop for i386 installation crashes on squashfs_read_data (Oops: 0000 [#1] )" [Undecided,Unconfirmed]  http://launchpad.net/bugs/66769
[10:32] <HeXiOn> hello
[10:32] <Ubugtu> Malone bug 66779 in ubiquity "Ubuntu 6.10 Desktop for i386 installation crashes on zlib_inflate_codes" [Undecided,Rejected]  http://launchpad.net/bugs/66779
[10:32] <Ubugtu> Malone bug 67784 in linux-source-2.6.17 "Edgy RC1 is crashing in the installation at 47%" [Undecided,Unconfirmed]  http://launchpad.net/bugs/67784
[10:32] <HeXiOn> there's an application installed by default in edgy (it wasn't in dapper) that manages remote control keys. Anyone knows what application it is and how to deactive it????
[10:32] <dholbach> I'll do some installs on i386 now too and see if it still happens
[10:33] <tfheen> dholbach: aiui, Colin described those as "mostly bad CDs"
[10:33] <dholbach> tfheen: I downloaded a new .iso, burned on a new CD, got a new cd drive and did a cd test
[10:33] <dholbach> and the people in the bug reports did those tests too
[10:34] <dholbach> anyway - I'll let you know how the fresh installs go
[10:34] <tfheen> dholbach: if you manage to reproduce them, please put a dmesg into the bug reports.
[10:35] <dholbach> tfheen: alrighty
[10:43] <fabbione> ajmitch: why do you need xen-utils to build libvirt?
[10:43] <ajmitch> fabbione: I need libxen3.0-dev
[10:43] <fabbione> afaict libxen-dev is enough
[10:43] <ajmitch> yes
[10:44] <fabbione> so that B-D can be dropped
[10:44] <ajmitch> some of that packaging is probably before I split libxen* from xen-utils-3.0
[10:44] <fabbione> ok
[10:44] <ajmitch> that's why I did the xen split in the first place :)
[10:45] <fabbione> yeah make sense.. just asking if i missed something when i did build it yesterday
[10:45] <mdz> re-morning
[10:45] <ajmitch> I'll check it over, I haven't used it on my laptop with xen yet
[10:45] <ajmitch> hi mdz 
[10:46] <ajmitch> xen 2.6.17 has some issues on amd64 currently
[10:46] <fabbione> ajmitch: ok.. i just need the pure library as B-D really
[10:46] <fabbione> i am not using xen itself
[10:46] <ajmitch> ok
[10:46] <tfheen> xubuntu alternate up
[10:46] <fabbione> but it's required for the redhat-cluster-suite support for xen fencing
[10:46] <ajmitch> interesting
[10:47] <ajmitch> just rebuilding it without the xen-utils b-d 
[10:47] <fabbione> the cluster suite has always been able to run inside xen.. but it was lacking a method to "reboot" a xen virtual machine
[10:47] <fabbione> and this xen support is pulling in a lot of crap
[10:48] <fabbione> i guess i am going to split xen support in another package
[10:49] <ajmitch> some of their other xen stuff is a bit too RH specific
[10:49] <fabbione> ajmitch: please fix the errors i /msg to you
[10:53] <ajmitch> great, rpath warnings
[10:55] <mdz> pitti: what happened with langpacks?
[10:55] <mdz> pitti: the current desktop CD is only 661M
[10:55] <pitti> mdz: I updated the seeds this morning and new images are currently building
[10:56] <mdz> pitti: oh, I saw a new build for 20061024 and assumed it was that. what was the earlier build for then?
[10:56] <mdz> oh, it just now went up
[10:56] <pitti> hm, strange
[10:56] <pitti> the alternate CDs are filled up nicely now
[10:56] <mdz> so we are talking about the same build
[10:57] <pitti> but http://cdimage.ubuntu.com/daily-live/20061024/ is still to small
[10:57] <pitti> tfheen: do the live scripts pick up changes directly from the launchpad seed branch?
[10:57] <cjwatson> they should now, but there's still a small delay
[10:57] <cjwatson> 20 minutes should cover it
[10:58] <cjwatson> sorry, I'm being unclear, they actually get it from http://people.ubuntu.com/~cjwatson/seeds/
[10:58] <tfheen> the desktop CDs just finished
[10:58] <cjwatson> since they aren't in ubuntu-core-dev they can't fetch from sftp directly, and the delay on http://bazaar.launchpad.net is roughly the same as that on my seed mirror
[10:58] <tfheen> but those images look too small, agreed.
[10:58] <pitti> tfheen: ah, so the desktop images you just built are *not 1024?
[10:59] <cjwatson> check the livefs build log ...
[10:59] <tfheen> pitti: "not 1024"?
[11:00] <cjwatson> tfheen: I just uploaded a new gfxboot-theme-ubuntu that fixes a crasher when you select any keymap that doesn't have an associated locale (there are 20 such keymaps)
[11:00] <pitti> tfheen: timestamp 20061024
[11:00] <cjwatson> not saying we have to use it, but if we're going to rebuild anyway I think we should
[11:00] <tfheen> cjwatson: arg. :-/
[11:00] <tfheen> cjwatson: what needs to be rebuilt for that to take effect?  just the live images?
[11:02] <tfheen> pitti: which langpacks did you add?
[11:02] <tfheen> (for ubuntu live)
[11:02] <cjwatson> tfheen: all images, no livefses
[11:02] <tfheen> cjwatson: ok.  I think I'm fine with that.
[11:02] <tfheen> sfllaw: are you coordinating testing and such?  I haven't seen any mails from you on the subject?
[11:02] <cjwatson> I can live with it being only on desktop images if you've already got final alternates
[11:02] <pitti> tfheen: xh pt ru ja de fr aa af am an as ast az be ber bg bo br bs bua ca ceb co cs csb cy da dv dz el eo et eu fa fi fil fo fur fy ga gd gl grc gu he hr hu hy ia id for all arches
[11:03] <pitti> tfheen: and a couple more for i386
[11:03] <cjwatson> wow, that's a lot
[11:03] <pitti> well, we had a lot of space to burn :)
[11:03] <tfheen> cjwatson: we're almost down to 650MB CDs..
[11:03] <tfheen> (without langpacks)
[11:03] <cjwatson> tfheen: http://people.ubuntu.com/~cjwatson/tmp/gfxboot-theme-ubuntu.diff FWIW
[11:03] <mdz> pitti: which do we have language-support for?
[11:03] <pitti> mdz: just -en, as usual
[11:03] <sfllaw> tfheen: I did send out an announcement of the new RC.
[11:04] <tfheen> sfllaw: where to?
[11:04] <mdz> pitti: but we have input methods for some of them?
[11:04] <mdz> pitti: which do we have input support for, I should say?  rather than oo.o translations
[11:04] <sfllaw> tfheen: Mmm, standard testing addresses.
[11:04] <pitti> mdz: yes, the general scim stuff is in -desktop
[11:04] <sfllaw> Not to warthogs though.
[11:04] <pitti> mdz: we do not have the particular Chinese/Taiwanese scim maps
[11:04] <tfheen> sfllaw: last email I have from you is from Thursday at least.
[11:05] <tfheen> (in my inbox)
[11:05] <mdz> pitti: we've said that we should not ship translations without input method support
[11:05] <sfllaw> tfheen: Well, it's been a day, so I will send out e-mails to remind people to do another pass at Testing/Current.
[11:05] <janimo> tfheen, Kamion: since you do rebuild anyway culd you consider the new casper too?
[11:06] <pitti> mdz: hm, we have always shipped Chinese, I just added some less common languages today
[11:06] <tfheen> janimo: that'd require a livefs rebuild, so no.
[11:06] <tfheen> kubuntu desktop cds up.
[11:06] <pitti> so we should revisit the scim stuff soon for feisty
[11:06] <mdz> pitti: I know, and sabdfl specifically said not to do it without the input packages
[11:06] <tfheen> those seems good, space-wise.
[11:07] <fabbione> sfllaw: want to update testing/current with the new iso info?
[11:07] <cjwatson> we're going to have to rebuild livefses to take zh *out* - it's not clear that that is worth it ATM
[11:07] <cjwatson> we're going to have to => we'd have to
[11:07] <mdz> cjwatson: I wasn't suggesting it, but now that you mention it, we're rebuilding anyway
[11:08] <sfllaw> tfheen: Did you spin another set of CDs?
[11:08] <cjwatson> mdz: not livefses yet
[11:08] <mdz> edgy-desktop-i386 is 661M
[11:08] <tfheen> sfllaw: yes, with more langpacks.
[11:08] <sfllaw> tfheen: OK.  Rsyncing again.
[11:08] <sfllaw> tfheen: Is this 20061024?
[11:08] <sfllaw> And which ones?
[11:09] <tfheen> sfllaw: sounds right.  All.
[11:09] <tfheen> sfllaw: those aren't final-final, we want Colin's fix for gfxboot-theme-ubuntu too, but I'd like people to start testing now.
[11:09] <cjwatson> mdz: hmm, yes, I thought the livefses had been rebuilt to actually pick up pitti's language pack changes but apparently not yet
[11:09] <tfheen> mdz: that'll be fixed as soon as the edubuntu livefs build finishes and I start an ubuntu one.
[11:09] <sfllaw> tfheen: Please send me a note over IRC when you spin new ones?
[11:09] <sfllaw> Thanks.
[11:10] <tfheen> sfllaw: I can tell you explicitly, sure.
[11:10] <cjwatson> hmm, drat, missed the publisher with gfxboot-theme-ubuntu
[11:10] <cjwatson> I'll put it onto cron.infinity^Wcjwatson once this run finishes
[11:10] <pitti> mdz: ok, all scim tables together are about 10 MB
[11:10] <mdz> pitti: we can't make a change like that now
[11:11] <pitti> the problem is, we have little to no feedback about the scim stuff ATM
[11:11] <mdz> we'll leave it alone
[11:11] <cjwatson> tfheen: please hold off buildlive until the seeds are finalised and I push a seed update
[11:11] <mdz> just please don't let this happen again for feisty
[11:11] <pitti> mdz: alright, I'll keep it in mind. We desperately need more testing in that area
[11:11] <pitti> or testing at all
[11:11] <tfheen> pitti: file a bug, milestone it later and we'll revisit it.
[11:11] <pitti> yes
[11:11] <tfheen> cjwatson: uh, seed update for what?
[11:12] <mdz> pitti: we do get some testing, there's even a team in launchpad
[11:12] <cjwatson> tfheen: as in until I run update-seeds on rookery to make sure it's *really* up to date this time
[11:12] <mdz> it's generally after release though
[11:12] <mdz> mvo: what happened to bug 57081?
[11:12] <Ubugtu> Malone bug 57081 in scim-chewing "scim-chewing: user database file can be corrupted, which render the whole package useless" [High,Confirmed]  http://launchpad.net/bugs/57081
[11:12] <cjwatson> since it doesn't seem to have been for the last livefs build
[11:12] <tfheen> cjwatson: oh, sure.  I just noticed that the kubuntu live ones I built were ~700 MB as they should.
[11:13] <tfheen> cjwatson: that was just the ubuntu seed, I think.  But, please go ahead.
[11:13] <cjwatson> the cron job probably kicked in before you built kubuntu
[11:13] <dexem> carlos, ping?
[11:13] <tfheen> I guess so, yes.
[11:13] <carlos> dexem: pong
[11:13] <dexem> carlos, do you have 5 minutes to talk? :P
[11:13] <cjwatson> tfheen: right, seeds are definitely up to date now
[11:13] <carlos> sure
[11:13] <tfheen> cjwatson: good.
[11:13] <sfllaw> Testing/Current updated.
[11:14] <sfllaw> Everyone's off to the races again.
[11:19] <mdz> tfheen: server from yesterday is still current, right?
[11:21] <ogra> pitti, feel free to do what you want :)
[11:21] <pitti> ogra: ok, so I add both kde and gnome langpacks and fill alternate and desktop to 698 MB?
[11:21] <ogra> pitti, 686M is the biggest install iso i have and 648M the biggest live one ...
[11:21] <sivang> re
[11:22] <ogra> pitti, yep
[11:22] <pitti> E/Desktop:      50      85      66
[11:22] <pitti> E/Alternate:    12      40      29
[11:22] <pitti> ^ free MBs for i386/amd64/powerpc 
[11:22] <ogra> yup
[11:23] <cjwatson> mdz: lrwxrwxrwx  1 tfheen cdimage 8 Oct 23 19:26 cdimage/www/full/ubuntu-server/daily/current -> 20061023
[11:24] <dholbach> then i could have just kept on using 20061023
[11:24] <dholbach> grmbl
[11:25] <mdz> cjwatson: current as in "will not be rebuilt", not "has not been rebuilt" :-)
[11:26] <fabbione> sfllaw: netboot line on the wiki is borked.. missing | somewhere
[11:26] <sfllaw> fabbione: Overzealous deleting on my part.
[11:27] <fabbione> sfllaw: i just need to make sure what i need to test
[11:27] <sfllaw> I have to admit that Wiki syntax is not very good for tables.
[11:27] <sfllaw> fabbione: Thanks!
[11:27] <fabbione> sfllaw: broken again..
[11:27] <fabbione> there is a preview button.. you know :)
[11:28] <cjwatson> I wish squashfs weren't quite such an oops-o-rama
[11:29] <fabbione> sfllaw: server image is still 23. it won't be regenerated with extra langpacks
[11:29] <tfheen> mdz: -server is still current, yes.
[11:29] <sfllaw> Ah.
[11:29] <sfllaw> tfheen: That's not "All".
[11:30] <sfllaw> fabbione: Fixed.
[11:30] <sfllaw> tfheen: Reverted Server CD testing.
[11:30] <tfheen> sfllaw: ubuntu-server keeps slipping my mind since it's published as ubuntu, not -server
[11:31] <sfllaw> Ah.
[11:31] <sfllaw> I'll ask next time.
[11:31] <tfheen> pitti: hmm, so edubuntu still needs a livefs rebuild?
[11:31] <pitti> tfheen: I'm changing the seeds ATM
[11:31] <tfheen> pitti: oh well, ok.
[11:31] <tfheen> cjwatson: anything against that?
[11:31] <mvo> mdz: the user-database corruption problem is fixed with the new version of scim-chewing/libchewing. as for the rest (discussion about the pro/cons of {GTK,QT}_IM_MODULES) I didn't want to switch that 3 weeks before the release. my gut feeling is that we should switch to scim-bridge for edgy+1, but I would like to discuss this with some people who use scim on a daily basis
[11:32] <pitti> tfheen: committed; will do edubuntu/alternate seed changes now
[11:32] <mvo> mdz: oh, this was of course about #57081
[11:32] <tfheen> pitti: thanks.
[11:32] <cjwatson> tfheen: nope
[11:32] <mdz> mvo: the best place for that information is in the bug report; as it is, the last comment is me asking you to look into it and then nothing further
[11:34] <tfheen> ubuntu livefs-es building
[11:36] <tfheen> cjwatson: tell me when the gfxboot fix is in the archive and I'll spin alternates and desktops then.
[11:36] <mvo> mdz: right, sorry for not following up there :/ I will do this now 
[11:37] <mdz> tfheen: if so, there's no point in testing the current build
[11:39] <pitti> tfheen, ogra: edubuntu/alternate seed changes committed as well now
[11:40] <tfheen> mdz: it's better to have people test it for other regressions than having them sitting twiddling thumbs.
[11:40] <cjwatson> tfheen: I'm publishing the source now
[11:40] <ogra> pitti, thanks :)
[11:40] <tfheen> cjwatson: thanks.
[11:43] <seb128> tfheen: I don't think people just "twiddling thumbs" when they are not testing CDs :)
[11:44] <ubuntu_demon> hi
[11:44] <mdz> tfheen: seb128 has a point; we are borrowing developers as testers :-)
[11:45] <tfheen> mdz: do you have a hidden armada of testers you haven't told me about so we can stop borrowing developers? :-)
[11:46] <tfheen> kubuntu desktop ready for testing
[11:46] <tfheen> sfllaw: ^^
[11:46] <mneptok>  /m tfheen please don't discuss Matt's multiple personality disorder on public channels
[11:46] <mneptok> oops.
[11:46] <sivang> mneptok: heh
[11:46] <tfheen> mdz: but well, you have a point.
[11:47] <tfheen> better to not wear people out.
[11:47] <mdz> tfheen: no, but that means we should spend their time as usefully as possible
[11:49] <sfllaw> mdz: Should we wait until the gfxboot fix comes out?
[11:49] <mdz> sfllaw: yes
[11:49] <sfllaw> OK.
[11:49] <sfllaw> Man, that's actually difficult.
[11:49] <seb128> sfllaw: desktop bugs are waiting for some hand 
[11:50] <mdz> sfllaw: how about updating the wiki page to show which images are real candidates in need of testing
[11:50] <sfllaw> I'm just kidding guys.
[11:53] <sfllaw> tfheen: I'm confused by what I'm blocked on.  I think I can let people test the server CD and EdgyUpgrades.
[11:53] <sfllaw> Does gfxboot affect Kubuntu and Edubuntu as well?
[11:53] <fabbione> sfllaw: gfxboot yes but only i386/amd64
[11:54] <fabbione> sfllaw: but server CD boots with gfxboot too iirc
[11:55] <cjwatson> correct
[11:55] <cjwatson> but we don't necessarily *have* to rebuild it
[11:55] <fabbione> well nobody did test them yet.. might as well doit
[11:55] <cjwatson> though dunno, the server CD might have more people with weird keymaps? hard to tell
[11:56] <cjwatson> (dvorak is one of the affected keymaps)
[11:56] <mdz> I use dvorak but didn't know until today that it could be selected in gfxboot
[11:56] <cjwatson> publisher is messing around with death row at the moment
[11:57] <cjwatson> mdz: I did it just for you
[12:00] <fabbione> mdz: you are sitting close to cjwatson .. can you please tickle an answer out of him? :P
[12:01] <mdz> fabbione: wait for the final build
[12:01] <fabbione> ok
[12:02] <mdz> tfheen: I recommend not delaying further for anything but showstoppers to leave enough time for a full test cycle
[12:03] <sfllaw> tfheen: Please let me know when certain classes of images come out.
[12:03] <sfllaw> I'll enable them on Testing/Current.
[12:05] <iwj> root@samual8:/media/hda4/ian# rsync cdimage.ubuntu.com::
[12:05] <iwj> rsync: connection unexpectedly closed (0 bytes received so far) [receiver] 
[12:05] <tfheen> sfllaw: server cd is testable, upgrades are testable.
[12:07] <sfllaw> tfheen: Merci.
[12:07] <cjwatson> hmm, yes, rsync from cdimage is unhappy here too
[12:07] <cjwatson> Spads: ^-- ?
[12:09] <Spads> okay, I see a lot of rsync clients at the moment
[12:10] <tfheen> fabbione: unless you complain loudly, I won't redo the server CDs for the gfxboot problem.
[12:10] <fabbione> tfheen: well i would prefer to have stuff in sync.. let headacke if we get bug reports
[12:10] <fabbione> s/let/les
[12:10] <fabbione> EKKKEKEKE
[12:10] <fabbione> less
[12:10] <Spads> iwj: which IP are you rsyncing from?
[12:10] <tfheen> fabbione: that invalidates all your testing, though.
[12:11] <iwj> 193.201.200.170
[12:11] <fabbione> tfheen: the server CD are all old except mvo and mine
[12:11] <fabbione> 2 tests against 12
[12:11] <cjwatson> Spads: I was rsyncing from my laptop here
[12:12] <fabbione> oh actually
[12:12] <fabbione> 4
[12:12] <fabbione> there is also dholbach 
[12:12] <cjwatson> i.e. 10.90.90.75 but I don't suppose that matters ;-)
[12:12] <tfheen> cjwatson: publisher done?
[12:12] <cjwatson> tfheen: yeah, but buildds don't seem to be picking up g-t-u; I've pinged malcc
[12:12] <tfheen> cjwatson: thanks.
[12:13] <fabbione> tfheen: well up to you.. as i said i would prefer to have them in sync
[12:13] <tfheen> fabbione: ok, let's wait for g-t-u to crawl through the buildds + publisher, then.
[12:13] <fabbione> tfheen: ok perfect
[12:13] <Spads> what happens if you specify the cdimage target explicitly?
[12:14] <cjwatson> ah yes, ::cdimage/ works
[12:14] <cjwatson> iwj: ^--
[12:14] <iwj> Oh, it's started working.
[12:15] <dholbach> fabbione: what about me?
[12:15] <fabbione> dholbach: -server images testing.. we are going to roll out new images for g-t-u fixes
[12:16] <dholbach> fabbione: alright
[12:16] <iwj> Urr, except that there is no 20061024 dvd.
[12:16] <tfheen> cjwatson: seems to have succfully built now?
[12:16] <tfheen> https://launchpad.net/distros/ubuntu/+source/gfxboot-theme-ubuntu/0.2.10
[12:16] <cjwatson> ah yes, thanks
[12:16] <tfheen> iwj: not yet.  There will be one soonish.
[12:17] <tfheen> iwj: just rsync the one in current for now.
[12:17] <iwj> I got that one yesterday (overnight, in fact ...)
[12:17] <iwj> So I'll go and buy a working dvd writer now.
[12:18] <tfheen> ah, ok
[12:18] <cjwatson> publisher running
[12:19] <mdz> gfxboot doesn't affect powerpc or sparc, right?
[12:19] <cjwatson> correct
[12:19] <fabbione> mdz: right
[12:19] <iwj> Unless I should wait a minute or two to start my rsync ...
[12:19] <mdz> sfllaw: so those two arches should be clear to test
[12:19] <cjwatson> iwj: publisher will take at least 20 minutes
[12:20] <cjwatson> then there's CD image build time, and DVDs won't be first in the queue there
[12:20] <tfheen> I'll do desktop CDs first, since those are the quickest.
[12:22] <iwj> cjwatson: Ah, OK.
[12:40] <thom> whiprush: i want to talk to you about ubuntu-update-server when you're around
[12:44] <dholbach> tfheen: http://daniel.holba.ch/temp/syslog http://daniel.holba.ch/temp/dmesg http://daniel.holba.ch/temp/debug
[12:45] <tfheen> syslog is 403
[12:45] <dholbach> just a sec
[12:46] <dholbach> should work now
[12:47] <tfheen> cjwatson: ^^ can you take a look at that?  It looks.. worrisome.
[12:50] <ogra> xfs ?
[12:50] <mdz> dholbach: reproducible?
[12:50] <tkamppeter> doko, tfheen, are you there?
[12:50] <mdz> dholbach: and does it pass an integrity test?
[12:50] <dholbach> mdz: yes to both questions
[12:51] <mdz> dholbach: which iso?
[12:51] <fabbione> hmmmmm
[12:51] <dholbach> mdz: ubuntu i386 dvd 20061023
[12:51] <cjwatson> dholbach: it's one of the billion "kernel oopses on bad CD" bugs
[12:51] <mdz> hmm, the previous DVD worked fine for me
[12:51] <mdz> haven't tested that one
[12:51] <mdz> will give it a try in vmware now
[12:51] <cjwatson> you have a bad disk or a dirty drive lens or something along those lines
[12:52] <dholbach> cjwatson: i doubt it's a bad cd
[12:52] <cjwatson> did you integrity-check it?
[12:52] <dholbach> cjwatson: i downloaded a new iso, burned it on a new cd, bought a new cd drive and integrity-checked it
[12:53] <dholbach> bug 66769 bug 66779 bug 67784  are similar bugs
[12:53] <Ubugtu> Malone bug 66769 in ubiquity "Ubuntu 6.10 Desktop for i386 installation crashes on squashfs_read_data (Oops: 0000 [#1] )" [Undecided,Unconfirmed]  http://launchpad.net/bugs/66769
[12:53] <Ubugtu> Malone bug 66779 in ubiquity "Ubuntu 6.10 Desktop for i386 installation crashes on zlib_inflate_codes" [Undecided,Rejected]  http://launchpad.net/bugs/66779
[12:53] <Ubugtu> Malone bug 67784 in linux-source-2.6.17 "Edgy RC1 is crashing in the installation at 47%" [Undecided,Unconfirmed]  http://launchpad.net/bugs/67784
[12:53] <cjwatson> yes, and they're all kernel bugs
[12:53] <cjwatson> I just haven't reassigned them yet
[12:53] <cjwatson> 0xfffffffb is Z_BUF_ERROR
[12:53] <tfheen> cjwatson: I wonder if we might be running into some kernel weirdness because we mount the squashfs twice or something
[12:54] <cjwatson> so I'm prepared to believe it's a kernel bug caused by something other than a bad CD, such as a borderline CD and not dealing with re-reading on errors correctly or something
[12:54] <tfheen> it'd still be a kernel bug.
[12:54] <cjwatson> but I doubt it's anything VFS-level
[12:54] <tfheen> publisher done?
[12:54] <cjwatson> and it's *not* an installer bug
[12:54] <cjwatson> oh, yes, it is
[12:54] <tfheen> ubuntu desktop build running
[12:54] <tfheen> argh
[12:55] <tfheen> I forgot to exclude powerpc from the build. :-(
[12:55] <dholbach> cjwatson: I was not trying to apportion blame. Sorry, if it sounded that way.
[12:55] <cjwatson> tfheen: you can ctrl-c
[12:55] <tfheen> cjwatson: done, re-running
[12:55] <cjwatson> dholbach: I know, I just have a billion of those bugs and there's absolutely nothing I can do
[12:55] <cjwatson> and most of them I remain convinced that it's a hardware problem
[12:56] <cjwatson> tfheen: (ideally you tweak etc/.next-build-* after ctrl-c'ing so that it reuses the build number ...
[12:56] <tfheen> dholbach: what would be very interesting is if you can reproduce that problem in vmware.
[12:56] <cjwatson> )
[12:56] <tfheen> cjwatson: I didn't.  Oh well
[12:56] <tkamppeter> tfheen, ping
[12:56] <cjwatson> I would love the kernel to react less violently to that particular problem
[12:56] <fabbione> crimsun: ping?
[12:56] <tfheen> tkamppeter: You sent me a contentless ping.  This is a contentless pong.  Please provide a bit of information about what you want and I'll respond when I am around.
[12:57] <cjwatson> returning EINTR and letting userspace retry would be good
[12:57] <cjwatson> or just retrying in squashfs
[12:58] <cjwatson> I don't know if retrying will help, though; it could be that zlib_inflate doesn't cope with short reads in particular places
[12:58] <dholbach> tfheen, cjwatson, mdz: anything else you'd like me to investigate apart from running it in vmware (which I'd need to set up)?
[12:59] <tfheen> dholbach: not that I can think of offhand, no.
[12:59] <dholbach> ok
[12:59] <tfheen> ubuntu desktop CDs up.
[12:59] <tfheen> sfllaw: ^^
[12:59] <dholbach> cjwatson: I think it happened to mvo too with his new t60
[12:59] <tfheen> kubuntu desktop CDs spinning.
[12:59] <tfheen> dholbach: is this an SMP or UP system?
[01:00] <dholbach> up
[01:01] <cjwatson> tfheen: should I leave the publisher on manual?
[01:01] <tfheen> cjwatson: yes, I think so.
[01:03] <tkamppeter> tfheen, what is the state about foo2zjs?
[01:04] <tfheen> tkamppeter: did you ever ask anybody to upload it?
[01:04] <tfheen> kubuntu desktop cds up.
[01:04] <tfheen> sfllaw: ^^
[01:04] <tkamppeter> Yes, I have sent several e-mails to you and doko_.
[01:05] <tfheen> cjwatson: edubuntu seed mirror is current, right?
[01:05] <ogra> pitti just updated it for th elangpacks afaik ...
[01:05] <tkamppeter> tfheen, doko_, bug 65618
[01:05] <Ubugtu> Malone bug 65618 in foo2zjs "Package broken/incomplete, missing firmware files" [Medium,Fix committed]  http://launchpad.net/bugs/65618
[01:06] <tfheen> tkamppeter: no, you didn't.  Matt sent an email on Sunday evening saying "this can be approved if it hits the archive before monday morning".  You sent an email on Monday evening asking about progress, which is far too late.
[01:06] <tfheen> tkamppeter: the gates for edgy have been closed, so please follow the procedure for uploading to -proposed.
[01:06] <mdz> tkamppeter: (http://wiki.ubuntu.com/StableReleaseUpdates)
[01:07] <sladen> shawarma: make your case to Kamion/theen---if you get the nod, I'll upload it
[01:07] <tfheen> Ubuntu Alternate CDs spinning.
[01:08] <tfheen> sladen: edgy main upload?  No, too late.
[01:08] <cjwatson> tfheen: dunno if it was, but it is now
[01:08] <sladen> shawarma: ^^note
[01:08] <tfheen> cjwatson: thanks.  Can you verify that pitti's langpack fixes is there?
[01:09] <tfheen> (sorry about pestering)
[01:09] <tfheen> pitti: did you update the xubuntu seeds too?
[01:09] <cjwatson> tfheen: timestamp: Tue 2006-10-24 11:39:18 +0200
[01:09] <cjwatson> ^-- most recent change
[01:09] <ogra> tfheen, looking at my current checkout it looks fine
[01:09] <cjwatson> dholbach: gnade upload from doko in unapproved, new upstream version?
[01:10] <tfheen> ogra: good, thanks
[01:10] <tfheen> edubuntu livefs-es building
[01:10] <dholbach> cjwatson: that's ok
[01:10] <tkamppeter> tfheen, soory, I am in Lexington, I will propose it to be an official update on the site which mdz mentioned during or after the Ubuntu Summit in Mountain View. Perhaps I ask the author to fix it upstream and then I take his new upstream version.
[01:11] <kervel> hello, is there anything one can do about bugs with a known fix already on launchpad (eg sync with debian) ... noteedit in edgy is unusable now, and the bugfix is on launchpad ..
[01:11] <cjwatson> dholbach: ok, thanks
[01:11] <tkamppeter> So don't worry, I will update the bug report telling that it will com e as update after the Mountain view week and in Edgy+1
[01:12] <cjwatson> kervel: uh, I synced noteedit a while back?
[01:12] <kervel> ah, i just saw it ... sorry
[01:12] <ajmitch> kervel: I see only 1 open bug against noteedit
[01:12] <kervel> cjwatson i put it on hold .. sorry for the fals report 
[01:12] <tfheen> tkamppeter: ok, thanks.
[01:15] <tfheen> kubuntu alternate up.
[01:15] <tfheen> sfllaw: ^^
[01:15] <Riddell> woo
[01:15] <highvoltage> :)
[01:16] <tfheen> Riddell: uh, sorry.
[01:16] <tfheen> I'm not sure how I managed to not see that my prompt wasn't there, but it's not done yet.
[01:16] <tfheen> sfllaw: ^^
[01:17] <tfheen> Riddell: ppc is the final one already, though.
[01:17] <tkamppeter> tfheen, mdz, bug 65618 updated.
[01:17] <Ubugtu> Malone bug 65618 in foo2zjs "Package broken/incomplete, missing firmware files" [Medium,In progress]  http://launchpad.net/bugs/65618
[01:17] <cjwatson> kervel: so nothing for us to do?
[01:18] <Riddell> tfheen, cjwatson: current kubuntu alternate amd64 is 701M, was there an issue about >700MB causing problems even if it's below the standard CD limit?
[01:18] <tfheen> Riddell: it should be ok with those images, I think.
[01:19] <cjwatson> yeah, it's best to keep it < 700 * 1024 * 1024 if you can
[01:19] <cjwatson> it's .5MB over
[01:19] <Keybuk> that should so be in /usr/bin
[01:19] <tfheen> Keybuk: killev?
[01:20] <Riddell> cjwatson: what about that powerpc image?
[01:20] <Keybuk> tfheen: kill evolution's little bonobo hinderances
[01:21] <tfheen> Keybuk: pkill evolution doesn't do that for you?
[01:21] <Keybuk> tfheen: no, that just kills the main binary
[01:21] <Keybuk> not e-d-s, e-a-n, e-i-e-i-o, et. al.
[01:21] <tfheen> Keybuk: you know the difference between pkill and killall?
[01:22] <Keybuk> tfheen: ?
[01:22] <tfheen> : tfheen@thosu ~ > pgrep -l evolution
[01:22] <tfheen> 5303 evolution-alarm
[01:22] <tfheen> 5336 evolution-data-
[01:22] <tfheen> 5355 evolution-excha
[01:22] <tfheen> pkill would have killed those
[01:22] <tfheen> it does substring matching.
[01:22] <Keybuk> tfheen: doesn't make them disconnect from the panel, etc.
[01:22] <Keybuk> which usually causes the clock to crash
[01:22] <tfheen> oh, maybe not
[01:22] <Keybuk> taking the panel with it
[01:23] <Keybuk> I'd file bugs, but seb would just ignore them at me :p
[01:23] <mdz> dholbach: DVD installs fine in vmware
[01:23] <dholbach> mdz: I try a dvd text install now
[01:23] <ogra> Keybuk, evolition --force-shutdown is the current right way, killev is dead
[01:24] <ogra> (since ages)
[01:24] <Keybuk> what's the difference?
[01:24] <ogra> dunno
[01:24] <tfheen> Riddell: kubuntu is done now.
[01:24] <tfheen> sfllaw: ^^
[01:24] <ogra> but seb told me once i should use it this way
[01:24] <ogra> i dont think evo even ships killev anymore
[01:25] <Keybuk> it must do, I just ran it
[01:27] <Keybuk> dholbach: do you have evolution installed? :)
[01:27] <dholbach> pfffft, of course
[01:27] <dholbach> but not the    keybuks-small-helpers    package ;)
[01:28] <mdz> we still have oaf-slay though
[01:28] <cjwatson> Riddell: what powerpc image?
[01:28] <dholbach> bonobo-slay? :)
[01:29] <Keybuk> syndicate scott% dpkg -S /usr/lib/evolution/2.8/killev
[01:29] <Keybuk> evolution: /usr/lib/evolution/2.8/killev
[01:29] <Keybuk> dholbach: ?
[01:29] <mdz> sfllaw: wiki needs updating for the current builds
[01:29] <mdz> sfllaw: we're at 24.1/2
[01:29] <dholbach> Keybuk: ok, looked in the wrong place
[01:30] <Keybuk> dholbach: where did you look?
[01:30] <dholbach> in /usr/bin
[01:34] <Riddell> cjwatson: http://cdimage.ubuntu.com/kubuntu/daily/20061024.1/ 700M  is that too big?
[01:34] <fabbione> hmmmm
[01:35] <fabbione> cjwatson: did something change in the seeds that might pull in oowriter on netboot/netinstall for LAMP server??
[01:35] <cjwatson> Riddell: no
[01:35] <cjwatson> cjwatson@lithium:~/cdimage/www/full/kubuntu/daily/current$ echo $((700*1024*1024))
[01:35] <cjwatson> 734003200
[01:35] <cjwatson> cjwatson@lithium:~/cdimage/www/full/kubuntu/daily/current$ ls -l edgy-alternate-powerpc.iso
[01:35] <ogra> Riddell, 734003200 is the limit :)
[01:35] <cjwatson> -rw-rw-r--  2 tfheen cdimage 733802496 Oct 24 09:12 edgy-alternate-powerpc.iso
[01:35] <mdz> Riddell: there will be an .OVERSIZED file if so
[01:36] <tfheen> edubuntu alternate images done.
[01:36] <cjwatson> fabbione: http://people.ubuntu.com/~cjwatson/germinate-output/edgy/rdepends/ALL/openoffice.org-writer is --> that way
[01:36] <tfheen> sfllaw: ^^
[01:36] <ogra> grmbl
[01:36] <Riddell> mdz: that's not the case for the amd64 CD there
[01:36] <ogra> pitti... you miscounted ...
[01:37] <cjwatson> .OVERSIZED doesn't trigger on 700*1024*1024 at present
[01:37] <cjwatson> it probably ought to, in theory
[01:37] <fabbione> cjwatson: ok.... it's the usual langpack's stuff
[01:38] <tfheen> xubuntu alternate spinning.
[01:38] <fabbione> pitti: we need to review these langpack situations... we can't pull in oowriter on server /netinstall stuff
[01:39] <cjwatson> hmm, there's an argument that it's a d-i bug
[01:39] <cjwatson> shouldn't be installing language-support-* on servers
[01:39] <cjwatson> I might be able to fix that in preseeds, not sure
[01:39] <cjwatson> which would be quick
[01:40] <cjwatson> what the ... it's *supposed* to be preseeded already
[01:41] <fabbione> cjwatson: this sparc.. i am not sure we preseeding anything from the a.out image
[01:41] <cjwatson> fabbione: you're not using silo?
[01:41] <cjwatson> we do configure silo.conf to use the standard preseeds
[01:41] <fabbione> cjwatson: on netboot? no.. you boot directly the image
[01:41] <fabbione> yes that's from CD..
[01:41] <cjwatson> fabbione: I blame your dodgy architecture then :)
[01:42] <fabbione> cjwatson: i blame langpacks :)
[01:42] <fabbione> cjwatson: why do we need to blame eachother when we can blame pitti?
[01:42] <Riddell> tfheen: I've altered the kubuntu ship seed, could you rebuild amd64 alternate CD?
[01:42] <fabbione> cjwatson: there is tilo.. perhaps we can use that one.. -> feisty anyway
[01:42] <tfheen> Riddell: yes, in a little bit.
[01:42] <Riddell> thanks
[01:43] <cjwatson> fabbione: no, your installation will be wrong in other ways too. just document that you should boot with preseed/url=http://people.ubuntu.com/~cjwatson/bzr/debian-cd/ubuntu/data/edgy/preseed/ubuntu-server/ubuntu-server.seed
[01:43] <cjwatson> ("just")
[01:43] <tfheen> xubuntu alternate done
[01:43] <tfheen> sfllaw: ^^
[01:43] <cjwatson> or "pkgsel/language-pack-patterns= pkgsel/install-language-support=false", which I guess is actually shorter in this case
[01:43] <fabbione> cjwatson: eeeekkkk
[01:43] <mdz> sfllaw: ping
[01:44] <fabbione> cjwatson: ok.. testing that one
[01:45] <Treenaks> http://news.bbc.co.uk/2/hi/technology/6080048.stm
[01:48] <sfllaw> mdz: Pong.
[01:49] <tfheen> Riddell: kubuntu amd64 rebuilt
[01:49] <tfheen> sfllaw: ^^
[01:49] <mdz> sfllaw: <mdz> sfllaw: wiki needs updating for the current builds
[01:49] <mdz>  sfllaw: we're at 24.1/2
[01:49] <mdz> tfheen can give you a complete list of what's ready
[01:49] <tfheen> ubuntu, kubuntu daily-live ready
[01:49] <tfheen> ubuntu, kubuntu, xubuntu, edubuntu daily ready.
[01:50] <pitti> arrgh, oversized
[01:50] <tfheen> still not there: edubuntu live and ubuntu-server
[01:50] <tfheen> pitti: which?
[01:50] <pitti> apparently squashfs compression is much worse than .deb compression
[01:50] <pitti> tfheen: daily-live/20061024.2/ (Ubuntu)
[01:51] <tfheen> pitti: *sigh*, ok.  Can you strip the live seed a bit, then?
[01:51] <pitti> I'll do, but I can only guesstimate
[01:51] <cjwatson> cjwatson@lithium:~/cdimage/www/full$ find -path \*/current/\*.OVERSIZED
[01:51] <cjwatson> cjwatson@lithium:~/cdimage/www/full$
[01:51] <pitti> I'll be generous
[01:51] <cjwatson> hmm, no, that's obviously wrong
[01:52] <pitti> and for some reason the powerpc CD didn't grow at all
[01:52] <pitti> ogra: see above, hard to guesstimate squashfs compression
[01:52] <cjwatson> ah, better
[01:52] <cjwatson> 1~./daily-live/current/edgy-desktop-amd64.OVERSIZED
[01:52] <cjwatson> ./daily-live/current/edgy-desktop-i386.OVERSIZED
[01:52] <cjwatson> ./edubuntu/daily-live/current/edgy-live-amd64.OVERSIZED
[01:52] <cjwatson> ./edubuntu/daily-live/current/edgy-live-i386.OVERSIZED
[01:52] <cjwatson> ./edubuntu/daily/current/edgy-install-i386.OVERSIZED
[01:53] <pitti> fabbione: why do you want to pull in oowriter on server?
[01:53] <ogra> pitti, i'm talking about install, not live
[01:53] <cjwatson> pitti: it's d-i installing language-support-en
[01:53] <tfheen> pitti: oh, I just rebuilt amd64 and i386 because they should have been enough, except that we wanted the new livefs there too.
[01:53] <fabbione> pitti: it's the other way around.. i do NOT want it
[01:53] <dholbach> pitti: because vim and emacs suck - oowriter is the one-true-editor
[01:53] <fabbione> pitti: but langpacks pull it in
[01:53] <tfheen> sfllaw: ok, ubuntu desktop not ready after all.
[01:53] <pitti> fabbione: I see, but what's pulling it in?
[01:53] <pitti> fabbione: no, language-pack-* does not pull in any oo.o stuff
[01:54] <fabbione> langpack-something...
[01:54] <cjwatson> pitti: language-support-en does
[01:54] <pitti> fabbione: only language-support-* does
[01:54] <pitti> right, but I didn't touch that
[01:54] <fabbione> ywah
[01:54] <cjwatson> pitti: it's not your fault
[01:54] <fabbione> cjwatson: yes it is :P
[01:54] <pitti> cjwatson: ah, ok; sorry, still catching up with reading backlog
[01:54] <cjwatson> in feisty, remind me to make pkgsel not install language-support-en if the desktop isn't installed or queued for installation
[01:54] <fabbione> pitti: well in one way or another we get there.. i am testing Colin's fix
[01:55] <tfheen> cjwatson: file a bug, mark it later.
[01:55] <pitti> tfheen: it seems that the ubuntu/powerpc/desktop image did not pick up the seed changes?
[01:55] <cjwatson> tfheen: will do
[01:55] <tfheen> pitti: no, it's a different bug.  I just didn't rebuild the image.
[01:56] <tfheen> (because I'm a moron)
[01:56] <pitti> ah, ok; I'll chop off 12 MB as well
[01:56] <tfheen> dholbach: thanks dude
[01:57] <ogra> ok, the edubuntu install iso should be fine again 
[01:58] <pitti> ogra: thanks, I was just about to fix it
[01:58] <pitti> tfheen: ubuntu live stripped down
[01:58] <ogra> pitti, i dropped zh and pt
[01:58] <ogra> what did you drop from ubuntu live ? 
[01:58] <tfheen> cjwatson: seed mirror update, please?
[01:58] <pitti> ogra: about a dozen langpacks
[01:58] <ogra> which ones :)
[01:58] <tfheen> ogra: does that mean you need a rebuild?  If so, of what?
[01:59] <pitti> ogra: (et eu fa fi fil fo fur fy ga gd gl grc gu he hr hu hy ia id)
[01:59] <ogra> pitti, thzanks
[01:59] <pitti> ogra: oh, no, sorry, those were just moved; I dropped is it jbo jv ka kab kk km kn ko ku ky la lb lg li lo lt lv mg mi mk
[01:59] <ogra> tfheen, install i386  ...
[01:59] <ogra> tfheen, working on a live seed fix ...
[01:59] <pitti> ogra: shall I fix your lives, too?
[01:59] <pitti> ogra: sorry for the trouble, btw!
[01:59] <ogra> pitti, that'd be nice so we have similar langpack sets
[02:00] <cjwatson> tfheen: done
[02:00] <ogra> pitti, i expected it, we never had a release where it fitted on the first attempt ;)
[02:00] <tfheen> anything else we need to hold on for livefs-es?
[02:01] <Riddell> tfheen: kubuntu alternate amd64 hasn't changed size, is there something that needs to happen before it picks up the language pack list from the seeds?
[02:01] <tfheen> Riddell: yes, Colin needs to update the seed mirror.  Done now.
[02:01] <tfheen> Riddell: spinning new amd64 update now
[02:01] <tfheen> ubuntu livefs-es building.
[02:04] <pitti> tfheen, ogra: edubuntu/live seed downsized; this *should* make all isos fit again
[02:04] <tfheen> Riddell: amd64 isos done
[02:05] <tfheen> sfllaw: ^^
[02:05] <tfheen> sfllaw: that gives us a .3 for all isos, but the images are the same as for .2 and .1
[02:07] <tfheen> ogra: do you want new i386 and powerpc ISOs?
[02:07] <tfheen> cjwatson: edubuntu seed mirror up to date?
[02:09] <ogra> tfheen, since amd64 is oversized as well, i think i need a full set of live 
[02:09] <ogra> for install i386 would suffice ...
[02:09] <tfheen> ogra: you want a new powerpc one too, since I forgot to rebuild that.
[02:10] <cjwatson> tfheen: yes
[02:10] <ogra> oki
[02:10] <tfheen> ARCHES="powerpc i386" for-project edubuntu cron.daily spinning
[02:13] <sfllaw> tfheen: Oy!
[02:13] <sfllaw> So confusing.
[02:13] <tfheen> sfllaw: what's confusing?  The copying of ISOs?
[02:14] <sfllaw> No, the point releases of images.
[02:14] <ogra> the numbering ?
[02:14] <tfheen> sfllaw: how would you rather number them?
[02:14] <ogra> why dont you point to /current and just veryify they point to the right iso each update ?
[02:14] <sfllaw> tfheen: Oh, I don't think there's a better way.
[02:15] <sfllaw> The iterations are just difficult to track.
[02:15] <cjwatson> ogra: because then people aren't explicitly conscious of what they've tested
[02:15] <ogra> hmm, right
[02:16] <sfllaw> OK, I've been up for too long.  I must have a nap.  Will be back in a couple of hours.
[02:18] <tfheen> ogra: ppc and i386 still oversized.
[02:18] <ogra> meh
[02:18] <tfheen> curiously, apache and ls disagree about how big a megabyte is.  Or rounding rules.
[02:19] <ogra> tfheen, hmm, looks like there was no change at all in these isos ? 
[02:19] <Keybuk> tfheen: so do your disk and ls, probably
[02:19] <ogra> didnt it pick up the seed change ?
[02:20] <tfheen> Keybuk: not by 1MB out of ~700.
[02:20] <tfheen> ogra: apparently not.
[02:20] <tfheen> cjwatson: any idea about that?
[02:20] <ogra> we're at rev 542 fwiw
[02:21] <ogra> probably just a race condition
[02:22] <tfheen> Colin's mirror is at 540
[02:23] <ogra> ah
[02:23] <cjwatson> cjwatson@rookery:~/public_html/seeds/edubuntu-edgy$ bzr revno
[02:23] <cjwatson> 542
[02:23] <tfheen> that changed in the last two minutes
[02:23] <cjwatson> may've just updated; the cron job was a minute ago
[02:23] <ogra> seems we were to fast on the trigger
[02:23] <tfheen> oh well
[02:24] <tfheen> ARCHES="powerpc i386" for-project edubuntu cron.daily
[02:24] <tfheen> running
[02:24] <tfheen> again
[02:24] <ogra> thanks :)
[02:24] <cjwatson> oh, gar, my mirror is from http://bazaar.launchpad.net
[02:24] <cjwatson> so the delay is that much greater
[02:24] <cjwatson> will look at that early feisty
[02:25] <tfheen> cjwatson: thanks.  We should be good now, though
[02:25] <cjwatson> mjg59: please can we have fewer things that need changing for new usplash command-line calling conventions? (uswsusp hook in initramfs-tools)
[02:25] <cjwatson> tfheen: yep
[02:25] <tfheen> cjwatson: can you file a later bug about it?
[02:31] <cjwatson> tfheen: done
[02:31] <tfheen> ogra: alternate build done.
[02:31] <tfheen> sfllaw: ^^
[02:32] <mjg59> cjwatson: I'm not quite clear on what you mean
[02:32] <ogra> tfheen, now that looks nice :), thanks a lot !
[02:35] <fabbione> cjwatson: that cmdline works..
[02:36] <fabbione> cjwatson:  i also filed a bug on pkgsel
[02:37] <tfheen> so, once this livefs build is done, I'll rebuild all of the ubuntu desktop cds and then the edubuntu ones.
[02:37] <tfheen> nothing else needed, right?
[02:37] <cjwatson> mjg59: the hook in initramfs-tools that calls /sbin/resume is the third thing that needs to be updated for the amd64 reversion
[02:37] <cjwatson> fabbione: I'd already done that
[02:37] <cjwatson> I'll dup
[02:37] <fabbione> oh ok
[02:38] <Riddell> tfheen: hmm, no change in size for amd64 alternate kubuntu
[02:39] <tfheen> Riddell: probably got bit by the same bug as the others, then.  Retrying..
[02:39] <fabbione> cjwatson: should we add something to the ReleaseNotes perhaps?
[02:39] <cjwatson> it's not a new problem
[02:39] <fabbione> yeah i know, but it got worst in edgy :)
[02:39] <fabbione> at least on sparc.. but hey
[02:40] <tfheen> ogra: building edubuntu livefs-es now.
[02:40] <ogra> thanks 
[02:43] <tfheen> Riddell: done
[02:47] <Riddell> tfheen: perfect, thanks
[02:47] <Keybuk> ogra: bug #67833
[02:47] <Ubugtu> Malone bug 67833 in metacity "wrong URIpath drug'n'droped with mouse" [Undecided,Unconfirmed]  http://launchpad.net/bugs/67833
[02:47] <Keybuk> no
[02:47] <Keybuk> ogra: bug #67383
[02:47] <Ubugtu> Malone bug 67383 in ltsp "Udev seems to fail on LTSP thin clients, local device support affected" [Medium,Unconfirmed]  http://launchpad.net/bugs/67383
[02:47] <Keybuk> :p
[02:48] <ogra> Keybuk, ah, yes i wanted to ask about that one ... i suspect its a kernel bug but the syslog doesnt show anything
[02:48] <Keybuk> ogra: why would it be a kernel bug?
[02:48] <ogra> the udev error is the only indicator there
[02:48] <Keybuk> you have several extra programs run by udev rules compared to the default?
[02:48] <ogra> dunno, but the system seems to get extremely slow
[02:49] <Keybuk> system gets slow => show me a bootchart
[02:49] <ogra> we have one rule that runs two different scripts (one that triggers mounts and one that unmounts)
[02:49] <Keybuk> do you use /dev/null in that script?
[02:49] <ogra> not that i'm aware of, i'll look
[02:50] <ogra> ogra@edubuntu:~/packages/ltspfs-0.4.3/scripts$ grep dev/null *
[02:50] <ogra> remove_fstab_entry:    umount -l ${DEV} 2>&1 >/dev/null
[02:50] <ogra> Keybuk, ^^^
[02:51] <ogra> but the remove part isnt triggered on boot
[02:51] <Keybuk> ok
[02:51] <ogra> highvoltage, are you around ? 
[02:51] <amachu> hi
[02:51] <mjg59> cjwatson: Ah, right. I'm not really sure how this can be done better.
[02:52] <Keybuk> it should create /dev/null first though
[02:52] <ogra> highvoltage, we're talking about bug 67383
[02:52] <Ubugtu> Malone bug 67383 in ltsp "Udev seems to fail on LTSP thin clients, local device support affected" [Medium,Unconfirmed]  http://launchpad.net/bugs/67383
[02:52] <tfheen> ogra: you're aware that 2>&1 >/dev/null != >/dev/null 2>&1 ?
[02:52] <Keybuk> static char *first_list[]  = {
[02:52] <Keybuk>         "/class/mem",
[02:52] <Keybuk>         "/class/tty",
[02:52] <Keybuk>         NULL
[02:52] <Keybuk> };
[02:52] <cjwatson> mjg59: something in usplash to get usplash arguments, maybe
[02:52] <ogra> tfheen, as long as it supresses output :)
[02:52] <Keybuk> tfheen: it won't
[02:52] <Keybuk> uh, ogra it won't
[02:53] <Keybuk> that'll make sure that errors are placed on standard output
[02:53] <mjg59> cjwatson: Sorry, I'm being horribly slow today. You mean have usplash parse the config file?
[02:53] <cjwatson> 2>&1 => dup *current* fd 1 (i.e. stdout) onto fd 2
[02:53] <cjwatson> >/dev/null => open /dev/null, dup onto fd 1
[02:54] <tfheen> : tfheen@thosu ~ > echo blah >&2  | cat 2>&1 > /dev/null
[02:54] <tfheen> blah
[02:54] <cjwatson> think about it in terms of syscalls, remember that they're done in command line order, and it makes more sense
[02:56] <ogra> in any case its not relevant for the bug ...
[02:56] <ogra> Keybuk, /dev/null will exist at some point and i dont see how local device access should be slowed down by it
[02:57] <cjwatson> mjg59: not sure. not a "right now" thing anyway
[02:57] <Keybuk> ogra: it won't be ... I don't think udev is to blame at all
[03:00] <Keybuk> cjwatson: except for pipelines, but let's not confuse the poor boy <g>
[03:00] <plb> any chance of getting gaim beta 4 and flash 9 in edgy before release? 
[03:00] <Keybuk> plb: none.
[03:00] <cjwatson>  /topic
[03:00] <ogra> flash 9 ?
[03:00] <ogra> if its released probably :)
[03:00] <plb> bah why not? people have been waiting years for flash 9
[03:01] <plb> ogra: flash 9 beta is out
[03:01] <cjwatson> edgy is FROZEN
[03:01] <plb> works better than 7
[03:01] <cjwatson> very, very solid
[03:01] <gnomefreak> plb: not on everyones system
[03:01] <tfheen> edgy is at roughly 0K.
[03:01] <gnomefreak> i have seen alot of people complaining
[03:01] <cjwatson> universe and multiverse are still a little bit slushy, but I'd be pretty cautious about big new things that require lots of user testing
[03:01] <thom> plb: if they've been waiting years they can wait another 6 montsh
[03:01] <plb> I've seen a few complain but not many
[03:01] <plb> heh
[03:01] <cjwatson> main and restricted are entirely locked down
[03:02] <plb> toss that crap in the nonfree section =P
[03:02] <Keybuk> cjwatson: so my scary util-linux patch is out? :p
[03:02] <highvoltage> ogra: sorry, been in a meeting
[03:03] <tfheen> Keybuk: any scary patches from you are out until feisty opens, yes.  :-P
[03:03] <Hobbsee> tfheen: you mean they'll be welcome even then?  :P
[03:03] <plb> how about an early release like firefox and fedora 6 ;] 
[03:04] <highvoltage> ogra: so if I understand correctly, the error message isn't a problem, just ugly?
[03:04] <tfheen> plb: no.  Please stop disturbing the developers working on preparing the release.
[03:04] <ogra> highvoltage, do you have a syslog from a local device access attempt ?
[03:05] <ogra> highvoltage, i'd like to see whats actually causing the slowness ... the bootlog isnt very informative
[03:05] <mdz> Riddell: I just did a fresh kubuntu OEM install, and got the KDE crash handler on login
[03:06] <mdz> Riddell: kdesktop crashed
[03:06] <Riddell> mdz: not good, I'll try out OEM
[03:06] <mdz> Riddell: sorry, ubiquity install
[03:07] <mdz> wrong vm
[03:07] <Riddell> mm, even worse
[03:07] <highvoltage> ogra: I don't have access to an edubuntu machine at the moment, when I get home I'll send that syslog to you first thing
[03:08] <mdz> Riddell: only oddity was that it failed to open the sound device (due to another vm using it)
[03:10] <mdz> Riddell: emailed details
[03:12] <Keybuk> tfheen: aww, I just wanted to rewrite mkswap to check the partition for an existing swap or hibernate partition and reuse the same UUID :)
[03:13] <tfheen> Keybuk: you're like three months late for that to go in. :-P
[03:13] <Keybuk> tfheen: how about udev just swapon'ing any block device with a swap signature? :)
[03:13] <ogra> Keybuk, ++
[03:14] <wasabi> Especially hot pluggable ones that you can yank our unexpectidly.
[03:14] <Keybuk> wasabi: the kernel copes just fine with that ;)
[03:14] <ogra> i have that in ltsp for feisty ...
[03:14] <wasabi> heh.
[03:14] <ogra> Keybuk, not true
[03:14] <Keybuk> it does
[03:14] <Keybuk> it's entirely happy to run the OOM killer if needed
[03:14] <wasabi> sounds insane.
[03:14] <wasabi> Yeah, but a running app was on the device.
[03:14] <ogra> if i kill an ndb device where i already use swap space in ltsp the client will die
[03:15] <ogra> where should it put the swapped data ?
[03:15] <Keybuk> wasabi: heard the one which goes "Doctor, it hurts when I do this" ?
[03:15] <cjwatson> wasabi: don't yank out block devices that might have swap on, then
[03:15] <wasabi> Eh. Sorry. It seems silly. What if I was putting in a drive just to see what was on it?
[03:16] <HrdwrBoB> then you get free extra swap
[03:16] <wasabi> ANd it happened to be an old Linux drive... Nobody is going to check if something auto enabled swap
[03:16] <HrdwrBoB> and you get to wonder what's using the partition table when you reload it
[03:16] <cjwatson> this would only be at boot
[03:16] <HrdwrBoB> and why it doesn't work
[03:16] <tepsipakki> hi.. I'm trying to reinstall initscripts on edgy, but /etc/network/if-up.d/mountnfs is not created for some reason
[03:16] <tepsipakki> I'm trying to reproduce some weird problems with nfs
[03:16] <ogra> cjwatson, the way i understood Keybuk we'll have a udev rule that generally matches all swap devices
[03:17] <Keybuk> tepsipakki: "reinstall initscripts" ?
[03:17] <ogra> that wouldnt restirct it to boot
[03:17] <wasabi> What's the point anyways?
[03:17] <tepsipakki> Keybuk: well, unpacking it would be enough :)
[03:17] <cjwatson> at boot> perhaps not
[03:18] <Keybuk> wasabi: of swap?  none, just buy more memory
[03:18] <wasabi> I mean, the point of auto enabling.
[03:18] <Keybuk> 2GB is a reasonable minimum :)
[03:18] <cjwatson> Riddell: what package do I need to install to get Qt Designer to know about KListView widgets?
[03:18] <Keybuk> wasabi: solves the "I ran mkswap, and now my fstab UUID is wrong" problem
[03:19] <cjwatson> tepsipakki: it's probably a conffile; use dpkg --force-confmiss
[03:19] <wasabi> Oh.
[03:19] <tfheen> Riddell: kubuntu dvd up
[03:19] <tfheen> ubuntu desktop running
[03:19] <Riddell> cjwatson: should be kdelibs4-dev
[03:19] <tepsipakki> Keybuk: that did the trick, thanks
[03:19] <Riddell> cjwatson: else kde-devel should bring in the rest
[03:20] <wasabi> I'm only against it because I know I'd probably hit it. I mean, I've done weird stuff like put swap on a usb disk temporarily.
[03:20] <mdz> Riddell: it didn't happen on the second login
[03:20] <cjwatson> Riddell: ok, thanks
[03:20] <mdz> Riddell: I've saved a vmware snapshot where it crashed if you need more info
[03:20] <wasabi> I don't erase that disk immediatly after... I set it aside and use it again later.
[03:21] <wasabi> I'd hate to plug it into a system later to get a file off and have some app swap to it, then pull it.
[03:21] <Keybuk> you shouldn't just pull it anyway ;)
[03:21] <mdz> true, and the eject process could deactivate the swap
[03:21] <wasabi> True. Guess you've got me then. What about SATA?
[03:21] <wasabi> Same deal?
[03:22] <Keybuk> what about SATA?
[03:22] <wasabi> I have a stack of 4 hard drives sitting here I have to go through soon... so I'll be plugging them into my SATA ports and seeing what's on them. Guess I just have to be careful to swapoff them.
[03:22] <wasabi> I suspect that'll bite people.
[03:22] <tfheen> wasabi: just like you should umount them, yes.
[03:23] <wasabi> Except I had to actively mount them to begin with.
[03:23] <wasabi> It didn't do it behind my back.
[03:23] <Keybuk> maybe it will in feisty ;)
[03:23] <wasabi> Heh.
[03:23] <wasabi> Actually I'd be fine with that.
[03:24] <wasabi> As long as the expectations remain the same for any device you plug in.
[03:24] <mdz> ogra,tfheen: is edubuntu up to date?
[03:24] <wasabi> I do sort of view swap as a bit more dangerous than some vfat partition though.
[03:25] <wasabi> I pull vfat all the time. If I know I didn't alter it, it's cool.
[03:25] <wasabi> The kernel handles it fine, nothing is killed.
[03:25] <ogra> mdz, live is building
[03:25] <ogra> mdz, install should be fine
[03:25] <tfheen> mdz: livefs-es are done, I'm building ubuntu desktop cds now, will do edubuntu live cds once that's done
[03:25] <tfheen> (then I'll do DVDs, then ports)
[03:25] <wasabi> Some of my running applications could have been swapped to the disk though, pulling that will trash them.
[03:26] <mdz> tfheen: oh, ubuntu desktop CDs aren't up to date yet?
[03:26] <mdz> I swear you said they were
[03:26] <ogra> mdz, only language pack changes ...
 ubuntu, kubuntu daily-live ready
[03:26] <wasabi> Oh also... no. I found a downside.
[03:27] <tfheen> mdz: they were oversized.
[03:27] <wasabi> Does a swap partition have a partition length check?
[03:27] <mdz> argh
[03:27] <tfheen> mdz: sorry. :-(
[03:27] <wasabi> So it will refuse to mount if the size of the disk it's on != the size the fs expects.
[03:27] <tfheen> uh, wtf?
[03:28] <pitti> tfheen: it seems the recent seed change wasn't picked up?
[03:28] <Keybuk> wasabi: in the name of all that is holy and good in this world, WHY WOULD YOU DO THAT?!
[03:28] <wasabi> Keybuk: md/raid1
[03:28] <iwj> cdimage.ubuntu.com is broken again
[03:28] <wasabi> My swap partitions on my servers are on md devices.
[03:28] <iwj> Spads: rsync: connection unexpectedly closed (0 bytes received so far) [receiver] 
[03:28] <Keybuk> wasabi: that wouldn't show up as a swap partition, it'd show up as a partial md block
[03:28] <ogra> wasabi, wow, you have md/raid on usbsticks ? 
[03:28] <ogra> intresting :)
[03:28] <Keybuk> RAID-1 MIRROR SWAP!
[03:29] <wasabi> Keybuk: raid1, it's show up as the underlying device.
[03:29] <tfheen> pitti: *sigh*, yes.
[03:29] <wasabi> /dev/sdNY would still be a valid partition.
[03:29] <Keybuk> wasabi: we don't mount devices, we mount block devices (aka partitions)
[03:29] <ogra> iwj, works fine here
[03:29] <Keybuk> yes, but it wouldn't be a swap partition
[03:29] <Keybuk> it'd be an md partition
[03:29] <wasabi> md puts metadata at the end.
[03:29] <Keybuk> they still show up as md
[03:29] <wasabi> It's perfectly mountable manually.
[03:29] <wasabi> I get into this all the time with lvm/md probe orders.
[03:30] <wasabi> where lvm will actually detect the underlying md device as the lvm pv
[03:30] <ogra> iwj, there is a connection restriction ... only two from the same IP are allowed iirc
[03:30] <wasabi> and screw everything up
[03:30] <Keybuk> heh
[03:30] <iwj> ogra: I'm testing from my colo and no-one else there is rsyncing.
[03:30] <ogra> hmm
[03:30] <iwj> And it doesn't work from here via NAT and DSL either.
[03:31] <iwj> cdimage.ubuntu.com A INET 195.248.90.25
[03:31] <StevenK> wasabi: md_component_detection = 1
[03:31] <Spads> iwj: okay, lemme check
[03:31] <ogra> i'm geeting my recent edubuntu isos just fine atm 
[03:31] <iwj> ogra: Say   rsync cdimage.ubuntu.com::  and what does it say ?
[03:31] <wasabi> StevenK: device_size_prompt = no actually makes the problem go away.
[03:31] <mdz> working fine here
[03:31] <wasabi> evms. =)
[03:31] <wasabi> StevenK: But the point remains, Keybuks auto-swap-mounter will need an equiv to md_component_detection
[03:31] <Spads> iwj: load is heavy
[03:32] <ogra> iwj, what kind of syntax is that ?
[03:32] <iwj> ogra: Standard syntax for an anon rsync.
[03:32] <mdz> one of them
[03:32] <iwj> Spads: So it just hates me ?
[03:32] <iwj> mdz: Right.  It's less typing than the same thing in URL format :-).
[03:32] <wasabi> brb
[03:32] <Spads> iwj: have you tried the cdrom/ target specifically?
[03:32] <Keybuk> wasabi: there's a whole bunch of specs for MTV registered
[03:32] <ogra> iwj, ah, i always use rsync rsync://cdimage....
[03:32] <Spads> or rather the cdimage/ target
[03:33] <Keybuk> udev-lvm, udev-evms, udev-device-mapper, udev-mdadm
[03:33] <ogra> but you are right, new connections are refused here as well
[03:33] <iwj> Spads: Ah, if I say   ...::cdimage  it works.
[03:33] <Spads> yeah
[03:33] <wasabi> Keybuk: Yeah.
[03:33] <Spads> cdimage doesn't have a default stanza and it doesn't list
[03:33] <wasabi> I registered myself to show up on everything having to do with md and lvm and evms.
[03:33] <cjwatson> tfheen: sorry, I'll start comparing seed revnos against an sftp update from now on
[03:34] <tfheen> cjwatson: my fault just as much.  Ubuntu livefs-es are running now at least.
[03:35] <elmo> I've increased the limits on cdimage, try again now
[03:35] <mdz> iwj: ^^
[03:37] <iwj> elmo: Thanks.
[03:41] <tfheen> mdz: ubuntu and edubuntu livefs-es didn't pick up the seed changes because I forgot to ask Colin to do a publisher run.  He's doing so now, which will make the next build reasonable.
[03:42] <mdz> tfheen: so I should abort this test install, ok
[03:42] <tfheen> mdz: yes.  Again, sorry.
[03:43] <tfheen> I'll say when we have reasonable images.
[03:43] <tfheen> alternate should be good, since it doesn't require a publisher run, but livefs-es need it.
[03:46] <bddebian> Howdy
[03:53] <doko_> mvo: ping
[03:53] <mvo> doko_: pong
[04:04] <tfheen> cjwatson: publisher done?
[04:12] <cjwatson> tfheen: yes
[04:13] <tfheen> ubuntu livefs building again.
[04:25] <pschulz01> Greetings...
[04:25] <pschulz01> What is being done about files that regularly need updating?
[04:26] <mdz> they are regularly updated
[04:26] <pschulz01> I have submitted a couple of bugs.. regarding usb.ids and pci.ids.
[04:26] <pschulz01> Neither of these have been updated for 6 months.
[04:26] <cjwatson> those are in packages we sync from Debian
[04:27] <cjwatson> our default system doesn't actually care about them except for display, AFAIK
[04:27] <Keybuk> correct
[04:27] <cjwatson> so they are not a high priority
[04:27] <cjwatson> for hardware detection, the kernel's module device tables are used
[04:27] <elmo> cjwatson: so what you're saying is that they are not a priority that need prioritising?
[04:27] <juliux> hi all, upgrade vom dapper edubuntu to edgy edubuntu is here working without problems
[04:28] <juliux> also the update for the nvida driver
[04:28] <Spads> elmo: dream the impossible dream
[04:28] <Nafallo> pschulz01: man update-pciids
[04:28] <CarlFK> grep irc /var/log/installer/syslog  "pkgsel: Entity 'irc-server' not defined" which may be a proximate error to something a few 100 lines up (that I am trying to find the beginning of)
[04:28] <Keybuk> elmo: we synergistically leverage contributions from our Debian colleagues thereby reprioritifying them to a matter of collaboration
[04:28] <CarlFK> from edgy-alternate, for those that don't know me :)
[04:29] <fabbione> you guys give me headake!
[04:29] <cjwatson> irc-server the what?
[04:29] <cjwatson> CarlFK: full log please
[04:29] <pschulz01> Nafallo: What about usb.ids.. aahh there's one of those as well :-)
[04:29] <CarlFK> cjwatson: that's what I said. coming iup
[04:29] <KurtKraut> There is a serious bug in Edgy's Ubiquity (67082) that makes the root filesystem unrecognized. This makes Ubuntu unable to install. Anyone could tell me is this will be fixed before 26th ?
[04:29] <elmo> iwj: ping
[04:30] <KurtKraut> This problem so so frequent that it was reported 3 times in different IDs.
[04:30] <cjwatson> KurtKraut: that makes it considerably less frequent than many other problems. :-)
[04:30] <cjwatson> KurtKraut: AFAIK that problem only happens if you go back and forward between gparted and the mount points page and make different selections each time
[04:31] <cjwatson> which makes it merely bad rather than a showstopper
[04:31] <cjwatson> I haven't seen it in test installs so far
[04:31] <KurtKraut> cjwatson, hmm... yes, you're right. But I guess when a bug makes the system unable to install, people tend to think 'of, this still beta and someone will fix', and doesn't report it.
[04:32] <cjwatson> I'll believe you when the bugs about failed installs stop rolling into my inbox
[04:32] <CarlFK> cjwatson:  http://dev.personnelware.com/carl/temp/Oct24/a/syslog
[04:32] <CarlFK> 3.5mg - want it zipped?
[04:33] <KurtKraut> cjwatson, ok, but let us be straight to the point... any chance this problem being fixed before Edgy release ?
[04:33] <cjwatson> KurtKraut: contrary to what one person in that bug report claims, bug 67254 is a different problem
[04:33] <Ubugtu> Malone bug 67254 in qtparted "Kubuntu : partition display problem" [Undecided,Unconfirmed]  http://launchpad.net/bugs/67254
[04:33] <cjwatson> KurtKraut: I'm afraid not at this point
[04:33] <crimsun> fabbione: pong
[04:33] <cjwatson> we've already frozen solid and will only stop for absolute showstoppers
[04:34] <cjwatson> if we have to reupload ubiquity for some showstopper reason, I'll consider it
[04:34] <cjwatson> but at present I don't actually know what the problem is, and, without having reproduced it locally, attempting to fix it is risky
[04:34] <KurtKraut> cjwatson, I do understand. But what should a bug have to be considerated a 'showstopper' for instance ?
[04:35] <fabbione> crimsun: sorry.. i need 5 minutes right now.. i will ping you back soon
[04:35] <cjwatson> KurtKraut: for a start, there would have to be no documentable workaround
[04:35] <cjwatson> KurtKraut: I'm not attempting to say that this isn't an important bug, but I'm afraid I just don't think it's release-critical - we have finite and increasingly little time available
[04:36] <cr3> on several occasions, I have encountered images > 700M under cdimage.ubuntu.com/daily-live/current. is that normal considering this is a daily-live directory?
[04:36] <iwj> elmo: pong ?
[04:36] <cjwatson> cr3: well, to start with, it's only worth reporting if there isn't a .OVERSIZED file present
[04:37] <cjwatson> cr3: secondly, I just lowered the size limit a little bit. According to Media Motion who ship our pressed CDs, the real limit is actually a little over 700MB, and our scripts were told the figure they gave us
[04:37] <cjwatson> cr3: but we've since heard of some rare cases where it really does have to be <= 700 * 1024 * 1024, so I just dropped the size limit to take account of that
[04:38] <KurtKraut> cjwatson, I'm talking here with the guy who reported the bug. He asked me to come here and ask because he is not fluent in english (neither I am :P)
[04:38] <cr3> ok, so I'll get some dvd-rws to test for now, no problem
[04:38] <elmo> iwj: I've had several dapper -> edgy upgrades where the default font choice in gnome went from subpixel -> best shape which makes my fonts in gnome-terminal look like crap.  is this known and if so, is there a bug about it somewhere?
[04:38] <cjwatson> KurtKraut: the primary thing I need is /var/log/syslog and /var/log/partman attached to the bug
[04:38] <cr3> cjwatson: thanks for the explanation though, I'll notice the .OVERSIZED file next time
[04:38] <cjwatson> as a rule, screenshots don't help me much
[04:40] <cjwatson> CarlFK: no thanks, unzipped is fine
[04:40] <KurtKraut> cjwatson, I'm asking him to reproduce the problem and post these files you've asked.
[04:41] <cjwatson> KurtKraut: thanks
[04:41] <cjwatson> there is a duplicate with log files, so it's not critical
[04:41] <cjwatson> also a precise description of exactly what steps he took in the partitioner would be good
[04:42] <KurtKraut> cjwatson, he's taking notes of the steps he done and I'll translate it to english
[04:42] <pirast> is it planned to make xorg autodetect hardware?
[04:42] <pirast> otherwise ill create a spec
[04:43] <CarlFK> cjwatson: lots of "dpkg: warning - ignoring pre-dependency problem !"  too
[04:43] <cjwatson> CarlFK: unimportant
[04:43] <CarlFK> ok
[04:44] <cjwatson> CarlFK: oh, that's not an installer bug, that's some documentation file that's broken
[04:44] <cjwatson> CarlFK: it's probably an ubuntu-docs bug
[04:45] <cjwatson> (from the looks of the filenames)
[04:45] <cjwatson> CarlFK: to be more verbose, "ignoring pre-dependency problem" is entirely normal when you're bootstrapping a chroot from nothing and half the pre-deps simply aren't there yet, but will be later
[04:46] <Keybuk> pirast: it's being developed upstream; we'll pick it up when they release a version with that in
[04:46] <CarlFK> ok - if you arn't worried, I'm not either :)
[04:46] <cjwatson> debootstrap extracts all the required packages first (dpkg -x) so it doesn't actually cause a practical problem
[04:46] <tfheen> .. and the race is on, king and royal are both almost done with royal about 40MB in the lead.  The end of the race is about 100M away, so this is exciting.
[04:46] <pirast> Keybuk, k.. thanks :-)
[04:46] <iwj> elmo: Err, it's not know by me.  There was some fontconfig subpixel configuration change I was involved in but aside from that you might be better off asking seb128 ...
[04:46] <iwj> s/know&/n
[04:46] <Keybuk> pirast: the spec would roughly end up being "hurry up and wait"
[04:47] <cjwatson> iwj: (that was the reason I directed elmo to you - it sounded like it might be a related problem)
[04:47] <Keybuk> tfheen: you should listen to workrave, and take breaks, y'know
[04:47] <tfheen> Keybuk: I've been outside today and all.  Just release bounciness.
[04:47] <pirast> Keybuk, yeah.. I won't create a spec then :-) One more question, is there any timeplan when the upstream feature is done?
[04:47] <iwj> elmo: is the problem that the symlinks in /etc/fonts/conf.d related to *-bitmaps.conf are wrong ?
[04:48] <tfheen> .. king is catching up, about 30MB behind now, but royal is _almost_ there..
[04:48] <tfheen> and royal won.
[04:48] <tfheen> there, cron.daily-live running for ubuntu, edubuntu livefs-es building.
[04:49] <CarlFK> cjwatson: can you give me a 30 second look at:  http://dev.personnelware.com/carl/temp/Oct24/a/late_command.sh.txt  and give me a tip on how to make the "# chroot $TARGET" not confuse the installer?
[04:49] <Keybuk> pirast: wrong channel to ask that I'm afraid, you'd need to speak to the Xorg guys directly
[04:49] <cjwatson> CarlFK: 'chroot /target' spawns an interactive shell and then sits there; it does not mean "run the rest of this script in the chroot"
[04:49] <pirast> Keybuk, yeah, thanks anyway :-) i though that you'd know :-)
[04:50] <cjwatson> CarlFK: just use apt-install instead of chroot /target apt-get install and you should get progress bars and everything
[04:51] <iwj> elmo: If it's the subpixel thing, it's related to fontconfig-config 2.3.2-7ubuntu2, which is an attempt at #56682.  It wasn't possible to preserve all previous handmade configuration.
[04:51] <iwj> bug 56682
[04:51] <Ubugtu> Malone bug 56682 in fontconfig "Raster fonts appear in Edgy" [Medium,Fix released]  http://launchpad.net/bugs/56682
[04:52] <cjwatson> CarlFK: from the looks of things you're confusing a shell script with a series of commands typed at a shell prompt :)
[04:52] <cjwatson> (the 'exit' down near the bottom of the script needs to go as well)
[04:52] <CarlFK> cjwatson: I got the feeling that the 'apt-get update' run in the installer enviroment wasn't updating the /target, but now that I think about it, it should.  either way - I don't need more of your time now :)
[04:52] <CarlFK> cjwatson: the exit was just to skip the rest of the script 
[04:52] <cjwatson> CarlFK: apt-get update> you were correct to start with; it won't
[04:53] <cjwatson> there is no "apt-get" binary in the installer environment
[04:53] <elmo> iwj: I don't think it's related to that
[04:53] <cjwatson> CarlFK: you don't need to run apt-get update there, though; just preseed apt-setup properly instead of munging sources.list by hand
[04:53] <cjwatson> CarlFK: apt-setup has a facility for preseeding "local" mirrors - see the installation guide for examples
[04:54] <elmo> iwj: well depending on what a sane /etc/fonts/conf.d is meant to look like
[04:54] <cjwatson> CarlFK: in edgy final apt-setup has been fixed so that you shouldn't need to manually force in the official mirrors anyway
[04:55] <iwj> elmo: Well, by default you'd have 30-debconf-no-bitmaps.conf 10-defoma.conf as symlinks and a bunch of random *.conf files, including no-* and yes-* in some cases.
[04:56] <elmo> iwj: http://people.ubuntu.com/~james/tmp/fonts-dir.txt
[04:56] <mdz> tfheen: what's the ETA for a new candidate?
[04:56] <slomo_> cjwatson: does ubiquity copy the complete filesystem or only the files belonging to a package and then run the maintainer scripts the same way as for a normal installation of the package?
[04:57] <cjwatson> slomo_: it just copies the complete filesystem, and then runs some manual fixups
[04:57] <iwj> elmo: Hmm.  20-debconf-sub-pixel.conf is part of why it's not the way you want it to be.
[04:57] <cjwatson> slomo_: in general any maintainer script that might do anything hardware-specific needs to be re-run by ubiquity
[04:59] <slomo_> cjwatson: including /var? then we need a fixup for dbus in feisty... but well, we can care about that then :)
[04:59] <tfheen> mdz: ubuntu is just done
[04:59] <tfheen> sfllaw: ^^
[04:59] <Keybuk> slomo_: what does dbus put in /var?
[04:59] <cjwatson> slomo_: all files
[04:59] <iwj> elmo: Oh, no, the problem is that you asked for subpixel and it turned that off, right ?
[05:00] <cjwatson> slomo_: all files on the squashfs, that is, not on the live session
[05:00] <tfheen> sfllaw: as in, ubuntu desktop for all arches is up.
[05:00] <iwj> elmo: This is controlled by the debconf value fontconfig/subpixel_rendering
[05:00] <cjwatson> slomo_: and does dbus in edgy strictly speaking require this?
[05:00] <elmo> iwj: yes
[05:00] <slomo_> cjwatson, Keybuk: a machine uuid that should be unique among machines and is recommended to stay the same way after reboots (/var/lib/dbus/machine-uuid iirc)
[05:00] <elmo>   fontconfig/subpixel_rendering: Always
[05:00] <CarlFK> cjwatson: thanks for the tips - things are working pretty good for me as is - I may ask for more after edgy is up and away 
[05:00] <slomo_> this will be created by the init script if not present already
[05:00] <elmo> hmm
[05:01] <cjwatson> slomo_: that would need to be reconfigured, yes
[05:01] <cjwatson> CarlFK: ok
[05:01] <Keybuk> slomo_: is this so Fedora can find out how many users they don't have? :p
[05:01] <Keybuk> why does a machine-local bus need a world-unique id?
[05:01] <BenC> tfheen: ping
[05:01] <tfheen> BenC: You sent me a contentless ping.  This is a contentless pong.  Please provide a bit of information about what you want and I'll respond when I am around.
[05:01] <slomo_> Keybuk: no... for ssh -X kind of sessions it's needed
[05:02] <Keybuk> slomo_: that can be a per-boot id though, no?
[05:02] <Keybuk> it doesn't need cross-reboot persistance?
[05:03] <slomo_> Keybuk: it can... but the dbus guys recommend to keep it always the same as people will most probably begin to assume that something called "machine-uuid" is always the same
[05:03] <Keybuk> why have the file at all?
[05:03] <BenC> tfheen: re bug 62135, I have a trivial fix and upload ready for it
[05:03] <Ubugtu> Malone bug 62135 in mesa "Support for Intel 965 (GMA X3000) doesn't work" [High,Confirmed]  http://launchpad.net/bugs/62135
[05:03] <Keybuk> system bus starts, generate a uuid, and just uses it
[05:04] <cjwatson> yeah, just pass it as a command-line argument?
[05:04] <cjwatson> or do clients need to know it?
[05:04] <tfheen> BenC: please follow the procedure at https://wiki.ubuntu.com/StableReleaseUpdates
[05:04] <BenC> tfheen: No uploads allowed for edgy release?
[05:04] <tfheen> BenC: no
[05:05] <tfheen> BenC: unless you find something earth-shattering.
[05:05] <slomo_> Keybuk: no idea... the only way why it makes sense to store it is to have it persistent... well, let ubiquity remove that file and we're done and don't need to worry about it anymore...
[05:05] <cjwatson> not for main, anyway
[05:05] <cjwatson> slomo_: ideally, it wouldn't require any hacks in ubiquity at all
[05:05] <BenC> tfheen: it's super trivial, and without it, i965 users will have broken GL (meaning all GL apps will just show a black/gray screen)
[05:05] <cjwatson> but removing the file's fine if that's the only way
[05:05] <Keybuk> slomo_: but why does it need to be persistent? :p
[05:05] <whiprush> thom: I'm around if you want to talk about that spec
[05:06] <BenC> tfheen: It's a one line change to a Makefile to set CC=gcc-3.4 for i965_dri.so only, and build-dep on gcc-3.4 for i386 and amd64
[05:06] <tfheen> BenC: well, it's just too late.  Release is in less than 48 hours.
[05:07] <slomo_> Keybuk: no idea... better ask the dbus guys... i can't think of a reason other than having it because it could be implemented ;)
[05:07] <Keybuk> the dbus guys are on scary crack most of the time
[05:07] <sjoerd> Keybuk: it doesn't have to be persistent
[05:08] <Keybuk> then why is it persistent?
[05:08] <Keybuk> why go to the trouble of having it persistent if it doesn't need to be?
[05:08] <BenC> tfheen: Tried last week to get with you about it, but didn't get a response on IRC..I'll update the bug report
[05:09] <sjoerd> because they rather had it in /var/lib instead of /var/run, because it might be usefull to be presistent..
[05:09] <Keybuk> but why write it at all?
[05:09] <Keybuk> if it doesn't need to be persistent, it doesn't need to be in a file
[05:09] <Keybuk> that means you don't need any of the reading or writing code
[05:09] <sjoerd> oh, the sessions busses need to read it
[05:09] <Keybuk> can't they get it from the system bus?
[05:09] <slomo_> the session bus could ask the system bus, no?
[05:09] <Keybuk> they are kinda connected, after all
[05:10] <thom> whiprush: i spammed you in private message :-)
[05:10] <Robot101> slomo_: bus daemons don't open bus connections to other buses
[05:11] <sjoerd> The only guarantee dbus needs at this time is that the key stays the same untill the box is rebooted...
[05:11] <sjoerd> and that the key is different from the one on other boxes
[05:11] <Keybuk> right... so it could generate it when the system bus starts
[05:11] <cjwatson> BenC: I know it's no good now, but if you'd just uploaded it then we'd have seen it in the unapproved queue and could have decided out of band :-/
[05:12] <BenC> cjwatson: Makes sense, didn't think about that
[05:12] <sjoerd> Keybuk: i totally agree with you and tried pushing to put in in /var/run.. So it is indeed created on boot when the system bus is first started.. But the dbus upstream guys didn't like it :(
[05:13] <tfheen> BenC: oh well, sorry about that. 
[05:13] <BenC> good thing is, it isn't an install issue, so it's quite alright being in edgy-updates
[05:13] <whiprush> thom: not registered so I don't think you can see my reply, so yes, that sounds awesome.
[05:13] <BenC> tfheen: understandable, you're pretty busy...I'll get it into edgy-updates
[05:15] <thom> whiprush: ah, arse
[05:15] <thom> ok
[05:15] <sjoerd> Keybuk: please send your comments to the dbus list, it might not be too late....
[05:15] <whiprush> thom: sec, let me freenode myself.
[05:19] <Keybuk> sjoerd: they don't like me on there :)  I challenge their fragile little world
[05:20] <sjoerd> then we're doomed to use /var/lib blah :)
[05:21] <Robot101> sjoerd: if you think havoc's talking shit, tell him :P
[05:21] <Robot101> ask him what the use case is for persistent uuids
[05:21] <sjoerd> Robot101: i've tried last night on #dbus and on the dbus list, i had my share for now
[05:22] <Robot101> let me have a go :D
[05:22] <sjoerd> :)
[05:22] <Keybuk> sjoerd: we're not doomed to do anything
[05:22] <Keybuk> we have a distro
[05:22] <Keybuk> HAND ME THE BUG SPRAY!

[05:22] <sjoerd> hehe
[05:22] <poningru> rofl
[05:23] <Keybuk> we're probably going to have to patch out system service activation
[05:23] <whiprush> thom: I apparently can't use a nickserv for more than a week without losing the password. Got IM?
[05:23] <Keybuk> along with whatever hal tries to do as well
[05:23] <Keybuk> so we can get messy in there <g>
[05:23] <sjoerd> hehe
[05:23] <sjoerd> the system service activation scares me too..
[05:24] <sladen> whiprush: change the password to something memorable, like "ubunflu", or "you suck" :)
[05:25] <Keybuk> that dbus list thread turned out to be mostly me talking to a wall, from what I could tell :-/
[05:25] <sladen> whiprush: seveas might be helpful in getting it back
[05:29] <JanC> are the final translations already in the edgy repositories?
[05:30] <tfheen> edubuntu livefs-es done; live images building.
[05:30] <thom> whiprush: heh, sure. 
[05:31] <iwj> elmo: Hmm.  Well the fontconfig machinery is working I think but I'm not sure I really understand the gnome side very well.
[05:31] <pepsiman> JanC: looks that way
[05:31] <ogra_> tfheen, thanks 
[05:31] <JanC> then something is wrong with them I fear  :-/
[05:32] <tfheen> JanC: oh?
[05:33] <cjwatson> JanC: if it's just missing translations, then those can be fixed by post-release updates
[05:33] <JanC> I get several complaints about wrong or missing Dutch translations that are okay on my system (using the daily language pack builds) and have been okay for weeks to months
[05:34] <cjwatson> JanC: nl isn't one of the language packs on the live CD, which may be the reason
[05:34] <pepsiman> daily langpacks stopped building at translation freeze too, they should be the same
[05:34] <JanC> that's weird
[05:35] <highvoltage> 
[05:35] <JanC> I'll look if I can compare them
[05:35] <highvoltage> 
[05:35] <pitti> tfheen, sfllaw: Testing/Current says 'out of date' for all disks; is desktop 24.4 ready to test or are we waiting for some bug fixes?
[05:36] <tfheen> pitti: desktop 24.4 is current, but sfllaw is asleep.  Please fix the wiki page and go ahead with testing.
[05:36] <pitti> yesmasta
[05:36] <pitti> tfheen: alternate as well?
[05:37] <tfheen> pitti: yes.
[05:37] <pitti> rock
[05:38] <iwj> tfheen: Oh, good.
[05:38] <ogra_> tfheen: either lithium is slow with copying or i miss a powerpc live iso
[05:38] <ogra_> ah, it was only slow
[05:38] <iwj> tfheen: You mean 24.1 ?
[05:38] <iwj> cdimage/daily/20061024.1
[05:39] <cjwatson> iwj: daily-live
[05:39] <iwj> Ah.
[05:39] <iwj> Err, no, not `ah'.  I mean, for the alternate.
[05:40] <cjwatson> sfllaw: I'm taking Ubuntu desktop i386 and giving Keybuk Kubuntu desktop i386
[05:40] <cjwatson> since we're both in the office and I've already started on the former anyway
[05:40] <Keybuk> note that Keybuk's ability to test large numbers of candidates is limited to his access to the vmware box here
[05:41] <pitti> cjwatson: can you please ping me when you have finished editing TEsting/Current?
[05:41] <cjwatson> whoever updated Testing/Current forgot to tag old PASS/FAIL with the old build numbers
[05:41] <cjwatson> pitti: just did
[05:41] <ogra_> heh, i wanted to ask the same
[05:41] <pitti> ah, thanks
[05:42] <ogra_> pitti: could you add 20061024.2 as the live candidate for edubuntu ? (its out of date currently)
[05:42] <pitti> ogra_: sure
[05:42] <ogra_> thanks :)
[05:44] <pitti> tfheen: server 20061023 is considered current?
[05:44] <JanC> *fuck*
[05:44] <pitti> JanC: now? here?
[05:45] <JanC> this is the second time during edgy that the Dutch tranlations have been reverted to an old faulty state   :-(
[05:45] <JanC> this isn't funny anymore  :-(
[05:46] <pitti> Riddell: Testing/Current says Kubuntu live 20061024.2, but there's only 24.1
[05:46] <cjwatson> JanC: but you said the language packs were fine?
[05:46] <cjwatson> iwj: I think 24.1 is fine for alternate but I'm not sure
[05:47] <cjwatson> iwj: yeah, looks ok
[05:47] <JanC> the daily build that I probably still have installed yes, the package in edgy apparently not...  :-(
[05:47] <cjwatson> JanC: should be fixable post-release if pitti knows the problem
[05:47] <pitti> JanC: edgy's packages are from last Friday; I don't know of any problem with them
[05:47] <JanC> It's probably all reverted in launchpad too
[05:48] <iwj> cjwatson: Oh, good.  Because that's what I've got on this CD :-).
[05:56] <jordi> does anyone know how big aprox. is a ubuntu mirror of the three latest distros?
[05:56] <tfheen> pitti: no.
[05:56] <jordi> say, in this moment, edgy, dapper and breeze, for i386
[05:56] <jordi> s/breeze/breezy/
[05:57] <pitti> tfheen: ok, then I put server to 'out of date'
[05:58] <tfheen> pitti: please just make it 20061024 if that's not available yet.
[05:58] <pitti> tfheen: oh, ok
[05:59] <fabbione> tfheen: did you roll out new server images for all arches or only i386/amd64?
[05:59] <tfheen> pitti: as in "build running"
[05:59] <tfheen> fabbione: I started all arches now, but I can just do i386/amd64 if you'd like
[05:59] <fabbione> tfheen: that's fine
[05:59] <fabbione> i can restest sparc very quickly
[06:00] <tfheen> fabbione: fixed to just do i386 + amd64
[06:00] <fabbione> tfheen: ok...
[06:00] <JanC> right, so rosetta has "forgotten" our translations fixes again...   :-(
[06:01] <JanC> this isn't funny anymore  :-(
[06:01] <cjwatson> #launchpad maybe?
[06:02] <tfheen> I'm off for a roleplaying session for a few hours.  If there's anything urgent, I have my cellphone.
[06:02] <pitti> argh, rescue mode bombs out -- this worked until RC
[06:04] <ogra_> pitti: yeah, same here
[06:04] <fabbione> pitti: how does it bomb out?
[06:05] <pitti> with a read dialog box saying that it failed to start
[06:05] <pitti> ogra_: ^ FYI
[06:05] <ogra_> bug 60423
[06:05] <Ubugtu> Malone bug 60423 in rescue "freeze after exiting rescue shell" [Undecided,Unconfirmed]  http://launchpad.net/bugs/60423
[06:05] <ogra_> i just saw that one
[06:06] <ogra_> but it starts fine and also runs a shell in the target system ... for me its only the reboot/shutdown thats failing
[06:06] <fabbione> oh interesting
[06:12] <cjwatson> yeah, I haven't diagnosed 60423 yet, but sounds different from pitti's problem
[06:13] <pitti> cjwatson: indeed it's different, I filed bug 68005 about it
[06:13] <Ubugtu> Malone bug 68005 in debian-installer "rescue mode fails" [Medium,Unconfirmed]  http://launchpad.net/bugs/68005
[06:13] <pitti> 60423 is around for ages, but it doesn't break the actual rescue mode
[06:14] <pitti> on the bright side, cdrom-check now reboots properly \o/
[06:14] <pitti> tfheen: did you deliberately fix bug 28033? it's still open, but doesn't happen to me any more (both on amd64 and powerpc)
[06:14] <Ubugtu> Malone bug 28033 in cdrom-checker "cdrom-checker-menu never reboots" [Medium,Confirmed]  http://launchpad.net/bugs/28033
[06:18] <Riddell> pitti: fixed
[06:19] <wasabi_> Heh. We do have a ton of specs don't we.
[06:19] <wasabi_> There should be some note someplace that making a spec with no intention or ability to implement it is mean.
[06:24] <cjwatson> pitti: Oct 24 16:01:15 rescue-mode: no partitions found!
[06:24] <iwj> pitti, cjwatson: alternate rescue wfm on i386.
[06:24] <pitti> ugh, it asked me for four partitions
[06:24] <pitti> cjwatson: ah, wait, that was on the other system
[06:24] <pitti> cjwatson: sorry
[06:25] <cjwatson> pitti: do you still have that image running?
[06:25] <pitti> cjwatson: unfortunately not, already wiped
[06:25] <pitti> I'll try it again after the current installation
[06:26] <cjwatson> pitti: if you get the same error, please tell me the output of 'list-devices partition' on tty2 (and dig into that with 'set -x' and such if you can)
[06:26] <pitti> cjwatson: I *thought* the image had an installation, but I was wrong
[06:26] <jdong> is bug 63558 going to get any attention?
[06:26] <Ubugtu> Malone bug 63558 in usplash "Latest usplash leaves my consoles corrupted" [Low,Confirmed]  http://launchpad.net/bugs/63558
[06:26] <pitti> cjwatson: yup, will do
[06:26] <jdong> it appears to affect all ATI video cards, and results in unusable/garbled VT's
[06:27] <jdong> this is a regression... usplash worked fine on these cards up to bzr revno 90 or so
[06:27] <sladen> jdong: I thought mjg59 had default with the usplash+ATI issue, is it still there?
[06:27] <pitti> argh, that's the second attempt to fix other people's edits; please respect locks
[06:28] <jdong> sladen: as of this morning my ATI laptop still has garbled VT's when using usplash....
[06:36] <pitti> cjwatson: works fine now
[06:37] <cjwatson> pitti: hmmmmmmmmmmm
[06:39] <cjwatson> pitti: ah, if there were indeed no partitions, that would be it :)
[06:39] <pitti> cjwatson: since I'm only secondary tester for amd64 I do these in vmware; I wiped the disk *brown paperbag*
[06:40] <cjwatson> heh, ok
[06:51] <cjwatson> just running the publisher again to publish some universe stuff
[07:00] <fabbione> let see what happens..
[07:00] <fabbione> this is a valid config from netboot/netinstall :)
[07:00] <fabbione> BUM!
[07:01] <pitti> oh, the progress bar in usplash is now at the very top screen border -- is that intended?
[07:01] <mdz> pitti: it isn't here
[07:02] <pitti> might be a vmware glitch, but it worked earlier
[07:03] <fabbione> ogra: ping?
[07:04] <lucas> mmh, I have an horizontal line about 1cm below my mouse cursor on edgy with the fglrx driver
[07:04] <lucas> does this ring a bell ?
[07:06] <mvo> lucas: I have seen this before, is this a setup with working dri maybe (check your xorg.0.log)?
[07:06] <fabbione> cjwatson: ping?
[07:06] <pitti> Kamion: current amd64/alternate does not install language-support-en; is that known?
[07:07] <pitti> cjwatson: ^
[07:17] <pitti> Kamion: oh, if the publisher isn't running, then no wonder why the postgresql-8.1 dapper-security upload doesn't come through; any chance to slip it in?
[07:18] <siretart> is usplash on amd64 delibrately back/white, and the progress bar at the top of the screen?
[07:19] <sladen> siretart: sounds like what pitti mentioned at 18:01
[07:20] <siretart> sladen: oh. (1901 for me, but anyway)
[07:20] <siretart> thanks
[07:21] <siretart> pitti: no, I see that on my amd64 as well
[07:21] <pitti> must have changed recently
[07:22] <siretart> pitti: is usplash black/white for you as well?
[07:23] <Toadstool> siretart: I have the same issues on my amd64 laptop
[07:24] <tfheen> pitti: http://launchpad.net/bugs/28033 > yes, I fixed it.  At least I think I fixed it.
[07:24] <Ubugtu> Malone bug 28033 in cdrom-checker "cdrom-checker-menu never reboots" [Medium,Confirmed]  
[07:24] <tfheen> siretart: it's greyscale on amd64 by design
[07:24] <pitti> mvo, sfllaw, other testers: please verify that language-support-en gets installed (check OO.o help); it doesn't for me
[07:27] <siretart> tfheen: ah. I see. I fear bug reports about usplash on amd64 being 'broken' because of that. is there a rationale for that behavior?
[07:27] <sladen> siretart: I suspect that it's to do with palette handling
[07:28] <siretart> I like the black/white one. really. Its just that I fear bugs because of confused users
[07:28] <siretart> perhaps we just make the logo read: ubuntu - amd64 edition or something like that
[07:31] <cjwatson> pitti: I ran the publisher a moment ago, which I think included postgresql-8.1
[07:31] <pitti> cjwatson: cheers
[07:31] <cjwatson> pitti: haven't encountered alternate not installing language-support-en; bug me
[07:32] <pitti> cjwatson: done, bug 68017, and just followed up
[07:32] <Ubugtu> Malone bug 68017 in debian-installer "does not install language-support-en" [Undecided,Unconfirmed]  http://launchpad.net/bugs/68017
[07:32] <pitti> cjwatson: might again be due to saying 'no' to 'download additional language support'
[07:32] <ogra> fabbione: ?
[07:32] <cjwatson> pitti: indeed
[07:32] <cjwatson> DDTT
[07:33] <cjwatson> pitti: it's a dup of another one of your bugs, I'm afraid ;-)
[07:33] <fabbione> ogra: does ltsp-build-chroot REALLY forces you to have an edubuntu CD ?
[07:33] <ogra> nope
[07:33] <pitti> cjwatson: yeah, it just occured to me too late
[07:33] <ogra> but edubuntu is the only flavor that uses it by default
[07:34] <ogra> fabbione: you need to use expert to run it on non-edubuntu systems
[07:34] <fabbione> ogra: one sec.. i need to show you the exact error message
[07:36] <fabbione> Oct 24 18:58:51 ltsp-client-builder: no CD-ROM found ! Not installing ltsp chroot
[07:36] <ogra> oh, right
[07:37] <fabbione> ogra: note that this was a netboot/netinstall
[07:37] <ogra> yep
[07:37] <ogra> it needs teh CD
[07:37] <fabbione> that's a bug
[07:37] <ogra> s/teh/a/
[07:37] <pitti> cjwatson: do you want OEM configurator bugs filed directly against oem-config or d-i?
[07:37] <pitti> cjwatson: (it messes up the locale)
[07:38] <ogra> fabbione: feel free to file it ... but it was never intended that its used for non Cd installs ... so its rather a whishlist bug :)
[07:38] <cjwatson> pitti: oem-config
[07:38] <pitti> mdz: re usplash> might be special for the amd64/nvidia usplash (which is just gray)
[07:38] <cjwatson> pitti: hopefully oem-config should get a lot less crap with simplify-oem-install or whatever the spec is called; planning on merging it into ubiquity
[07:39] <fabbione> ogra: what package would that be? and no it's not wishlist.. i might have a local mirror from where to install a bunch of servers.. why on earth would i want to download an extra cd?
[07:39] <pitti> the original locale was de_DE.UTF-8, but it changed it to de_DE
[07:39] <pitti> ...@euro
[07:40] <ogra> fabbione: it was written for edubuntu CD installs in first place ... i simply never had a request for it to work in netboot installs ... 
[07:43] <janimo> pitti: do all apps put their tranlations in language-pack-XX-base by default? I wonder why system-config-printer has no .mo files in there
[07:44] <pitti> janimo: it might not have been imported in time, no idea
[07:44] <fabbione> ogra: well you still didn't answer my question.. against what package you want the bug?
[07:45] <iwj> pitti: Sorry, I think I may have just trashed your edit on TestingCurrent.
[07:45] <pitti> yup, just saw it
[07:46] <iwj> tfheen: Is there going to be a 20061024 DVD ?
[07:47] <iwj> pitti: Should I redo my edit or are you merging ?
[07:47] <pitti> iwj: already fixed
[07:48] <iwj> Ta
[07:51] <mdz> we believe we've isolated bug 60938 and it justifies building a new livefs
[07:51] <Ubugtu> Malone bug 60938 in linux-restricted-modules-2.6.17 "ath_hal missing from modules.dep" [High,Confirmed]  http://launchpad.net/bugs/60938
[07:58] <Keybuk> publisher running again
[08:04] <amep> I am testing the Edgy 20061024 (15:51) AMD64 desktop install disk. When I start it I get an error message from GNOME: "There was an error starting the GNOME Settings Daemon. [...]  Activation of org.gnome.SettingsDaemon timed out [...] "
[08:05] <amep> Where could I get more info to attach to a bug report? The message itself doesn't seem that useful.
[08:17] <Keybuk> update-initramfs scares me
[08:18] <Keybuk> I swear that it just picks a random kernel off the disk
[08:19] <mvo> pitti: language-support-en is installed here (edubuntu/live install)
[08:19] <mvo> firefox has no launchpad-integration support anymore? known?
[08:19] <mvo> oh, sorry
[08:19] <mvo> I take that back
[08:20] <pitti> mvo: I answered 'no' on 'download complete language support' (since I had no network), that was the problem
[08:21] <mvo> pitti: aha, ok
[08:22] <Keybuk> pitti: "why doesn't it do this if I tell it not to" ?
[08:22] <pitti> Keybuk: that's not the point, language-support-en should always be installed
[08:22] <pitti> Keybuk: the problem is that the question is confusing and doesn't do just one thing
[08:23] <pitti> 'download complete support for your [German]  language' != 'do not install any language-* at all, even if available on the CD'
[08:25] <Keybuk> I like the fact that message is always in English
[08:25] <Keybuk> just to drive the point home
[08:27] <Nafallo> haha
[08:29] <ogra> fabbione: ltsp-client-builder is a part of ltsp ... file it against that ...
[08:29] <Keybuk> faster pussycat, publish...publish...
[08:30] <shackan> Keybuk, it does? (update-initramfs just updated the generic kernel, while I'm using 686)
[08:30] <Keybuk> shackan: it did? oh good
[08:31] <Keybuk> we don't know why that does that yet
[08:31] <Keybuk> other than it might be locale specific
[08:31] <shackan> scary
[08:31] <nixternal> heh, now i know why my kernel went back to generic yesterday..messing with usplash, i did update-initramfs
[08:40] <Keybuk> kernel went back to != this bug
[08:40] <Keybuk> that's just a grub order
[08:44] <pitti> ogra: ouch, sorry, just broke your lock
[08:45] <pitti> bah, this note should really be flashing red, not in the same gray as the normal one
[08:45] <ogra> pitti: unless you test edubuntu that should do no harm
[08:45] <pitti> no, I didn't
[08:53] <Riddell> \sh_away: I don't think Kamion is from the US
[08:57] <sfllaw> $
[08:57] <mdz> 
[09:00] <Keybuk> 
[09:01] <abattoir> 
[09:02] <Nafallo> mdz: moved? :-)
[09:02] <mdz> no more so than usual
[09:02] <Nafallo> oki :-)
[09:11] <pitti> mdz: so we're going to have new live CDs? shall I mark the current ones as obsolete in Testing/Current, so that testing focuses on alternates for now?
[09:11] <Keybuk> we'll need new alternates too
[09:12] <pitti> Keybuk: hmkay; there go 3.5 hours of testing, but the bug seemed bad enough
[09:18] <tfheen> iwj: there'll be a dvd, yes.
[09:21] <iwj> tfheen: Right.  Well, I think I'll pick that up tomorrow.
[09:44] <fdsd> hey guys, what is more reliable? rsync -av or cp -av on a dying drive?
[09:47] <jcole> why doesn't ubuntu have this feature -> http://www.thecodingstudio.com/opensource/linux/screenshots/original/Fedora%20Core%206/44.gif
[09:50] <LaserJock> jcole: rough guess, people haven't coded/ported it yet?
[09:52] <mvo> sfllaw: do you have a test-system with a intel ICH6 or ICH7 ide chipset?
[09:53] <Burgwork> jcole: that is system-config-xfree86
[09:53] <mvo> sfllaw: I wonder if you can reproduce #66779 on it? the bug might be releated to the hardware that is used
[09:53] <Burgwork> jcole: it needs to be tested and then ported
[09:55] <Arador> anyone has a clue about #67957 , is something expected to happen in dapper -> edgy so that it only behaves OK in clean installs?
[09:57] <jcole> Burgwork: oh ic, this is some special redhat app... not gnome
[09:59] <Riddell> how do I do the validity check on the power pc CDs?
[09:59] <Burgwork> jcole: system-config might be a good place to start with that umbrella stuff
[10:02] <mantiena-baltix> pitti: hi
[10:02] <mantiena-baltix> pitti: what you think about updating language packs today or tomorow ?
[10:06] <pygi> mantiena-baltix: I think it won't happen :P
[10:07] <mantiena-baltix> pygi: so, Polish and Lithuanian f-spot and other language fixes will not be into final edgy version...
[10:07] <pygi> mantiena-baltix: seriously doubt it ^_^ But I'm not the one in charge :P
[10:09] <pitti> mantiena-baltix: it won't happen
[10:09] <pygi> yay, I was correct for once :P
[10:09] <pitti> mantiena-baltix: next round of updates will be at start of November
[10:15] <rmjb> hello, if I delete my initrd from /boot will I get it back by running update-initramfs -c?
[10:15] <sfllaw> mvo: Yeah, we do.
[10:15] <sfllaw> mvo: Let me ask cr3.
[10:17] <mvo> sfllaw: thats great, thanks. I will try to reproduce the problem here again with my local test-machine
[10:18] <cjwatson> Riddell: boot with 'check'
[10:18] <sfllaw> mvo: It's the same machine as was reported by Marc.
[10:18] <keescook> should there be an "Examples" folder in the edubuntu live CD?
[10:19] <Riddell> cjwatson: hmm, that's what I guessed and tried on the DVD, I'll give it another shot though
[10:20] <sfllaw> mvo: I can try to reproduce and run tests, if you like.
[10:20] <mvo> sfllaw: great! I really wonder if a install works fine on this
[10:20] <ogra> hmm, my laptop hardlocks with the broadcom driver loaded 
[10:20] <ogra> (amd64) 
[10:20] <sfllaw> mvo: Are there any options you want me to pass in to get more verbose logs?
[10:22] <mvo> sfllaw: no, if it is really a kernel problem then the dmesg output should be enough
[10:22] <cjwatson> tfheen: any idea how the rebuilds are going?
[10:23] <sfllaw> mvo: There's already a dmesg from that machine in the bug report.
[10:24] <cjwatson> keescook: yes, should be on the desktop
[10:24] <cjwatson> not a huge deal if it isn't, although it would be a bug
[10:24] <ogra> not on edubuntu, no
[10:24] <cjwatson> oh, you're right, example-content isn't in edubuntu-desktop
[10:24] <cjwatson> keescook: ogra's correct, ignore me
[10:25] <keescook> cjwatson, ogra: okay, good.  no new bugs.  :)
[10:26] <ogra> if we ship on 2 CDs and its still missing then it will be a bug :)
[10:27] <ogra> hrm ...
[10:27] <ogra> dropping the sbcm firmware makes the amd64 install boot just fine ...
[10:27] <ogra> *bcm
[10:30] <ogra> woah
[10:30] <ogra> who translated my FUSE entry in the users and groups tool to german ...
[10:30] <ogra> that became a horrible long sentence ...
[10:36] <sladen> * bdale and various people are around west ken/shepard's bush for dinner
[10:39] <mvo> cjwatson: I would like to reopen bug #66779 because I got it with two different CDs (one burned with low-speed and check ok on a brand-new cdrom)
[10:39] <Ubugtu> Malone bug 66779 in ubiquity "Ubuntu 6.10 Desktop for i386 installation crashes on zlib_inflate_codes" [Undecided,Rejected]  http://launchpad.net/bugs/66779
[10:39] <cjwatson> mvo: please reassign it to the kernel; I can't fix it
[10:39] <cjwatson> interesting that it's certain chipsets
[10:40] <cjwatson> we may have to release-note that :(
[10:41] <mvo> cjwatson: the data is still sparse, but it looks like it. unfortunately (if the assumption is true) it is the kind of chipset you find in most modern laptops
[10:41] <cjwatson> mvo: does the alternate CD work on those laptops?
[10:42] <cjwatson> if so we have a documentable workaround
[10:42] <mvo> cjwatson: not sure, I need to check that. I'm waiting for sfllaw to confirm on a lap machine
[10:42] <sfllaw> mvo: Apparently, it's unreproducable now.
[10:42] <sfllaw> 16:41 < cr3> sfllaw and EtienneG: go right ahead with that VAIO, I haven't been 
[10:43] <sfllaw>              able to reproduce the problem. I've been testing 20061024.1 which 
[10:43] <sfllaw>              has not had a problem so far.
[10:43] <mvo> sfllaw: it happens for me during the fs-copy (at around ~40-50%)
[10:44] <cjwatson> it's probably somewhat random
[10:44] <cjwatson> I was suspecting failure of squashfs to deal with short reads earlier, but I was really just thinking out loud
[10:44] <cjwatson> careful analysis of the various oopsen might well help
[10:49] <Sp4rKy> hi
[10:50] <Sp4rKy> please, i want modify the initramfs file which is in the live cd (casper/initrd.gz), what's the way ? gzip -d initrd.gz , cpio -i < initrd , and after ?
[10:51] <Sp4rKy> for recreate it ?
[10:57] <keescook> pitti: you already made note of the busted "rescue" mode, yes?
[10:57] <pitti> keescook: in the sense that it hangs after exiting the rescue shell, yes
[10:58] <pitti> that bug is there for ages
[10:59] <cjwatson> BenC: I'd appreciate your feedback on those zlib_inflate* oopsen as soon as you can
[10:59] <cjwatson> mvo: could you tell mdz about your findings when he returns? I believe that he and Scott are having dinner
[11:06] <keescook> pitti: on edubuntu (PPC), the rescue mode just flat fails to start.
[11:06] <keescook>  /devicepath/install/powerpc/vmlinux: Unknown or corrupt filesystem
[11:07] <Sp4rKy> please, some help about the way to compil an initramfs systeme
[11:07] <pitti> keescook: I didn't see that yet, please discuss with cjwatson and file a bug
[11:10] <tfheen> cjwatson: all rebuilds with the exception of dvds are done.
[11:12] <BenC> cjwatson: ok
[11:17] <tfheen> iwj: dvds are building now, they'll be done in an hour or so, iirc
[11:18] <visit0r> I tested upgrade from dapper to edgy, now my DVB devices are not detected...
[11:18] <visit0r> known issue?
[11:19] <visit0r> in addition, a whole bunch of python packages are kept back
[11:19] <keescook> cjwatson: I didn't see any obvious duplicates, so I filed bug 68077.
[11:19] <Ubugtu> Malone bug 68077 in Ubuntu ""rescue" fails to load kernel on PPC edubuntu Desktop CD" [Undecided,Unconfirmed]  http://launchpad.net/bugs/68077
[11:26] <visit0r> never mind, sudo /etc/init.d/udev restart made them appear again
[11:27] <jdong> wasn't there some talk earlier today about dysfunctional rescue mode?
[11:27] <jdong> probably a different issue
[11:27] <pitti> yes, indeed
[11:27] <jdong> never mind my rambling :)
[11:27] <pitti> jdong: that was me trying to use rescue mode on a totally blank disk
[11:28] <jdong> pitti: LOL, ah , I see
[11:28] <jdong> I guess rescue mode ain't that magical
[11:28] <keescook> this disk passes disk check, I'll snag the .3 alternate and see how it behaves too
[11:32] <KurtKraut> cjwatson, are you there ?
[11:38] <pitti> tfheen: it is correct that alternate 20061024.2 does *not* yet have the new initramfs-tools?
[11:38] <tfheen> pitti: which new initramfs-tools?
[11:38] <pitti> tfheen: 0.69ubuntu19, with the fix for bug 60938
[11:38] <Ubugtu> Malone bug 60938 in linux-restricted-modules-2.6.17 "ath_hal missing from modules.dep" [High,Confirmed]  http://launchpad.net/bugs/60938
[11:39] <tfheen> pitti: I have not approved any upload which fixes that bug.  At least I can't remember having done so
[11:39] <pitti> the images are quite fresh, that's why I wondered
[11:39] <honzasch> Hi, first time here. May I have just a small comment to some glitches in Czech localization of edgy installer or not good place or time now?
[11:39] <pitti> tfheen: somebody must have, I just got it though dist-upgrade
[11:39] <tfheen> honzasch: I'm sorry, but it's way too late to change that now.
[11:39] <pitti> tfheen: and mdz wanted new images for this bug
[11:40] <keescook> pitti: I didn't have the 0.69ubuntu19 update on my 20061024.2 edubuntu (it downloaded after install)
[11:40] <honzasch> OK, no problem.
[11:40] <pitti> tfheen: the new desktops (24.5) have that new version
[11:40] <pitti> keescook: right, that's what I'm saying; the newly spun Ubuntu alternates don't have it, but desktops have
[11:41] <tfheen> pitti: I can't see any discussion of that here, and mdz have not said anything about this to me, by email or by phone.
[11:42] <pitti> tfheen: so you didn't spin the 20061024.2 alternate images?
[11:43] <pitti> anyway, /me goes to bed now and tests the new images tomorrow, assuming that there will be yet another alternate rollout
[11:55] <Lure> tfheen: 07:51	mdz	we believe we've isolated bug 60938 and it justifies building a new livefs
[11:55] <Ubugtu> Malone bug 60938 in linux-restricted-modules-2.6.17 "ath_hal missing from modules.dep" [High,Confirmed]  http://launchpad.net/bugs/60938
[11:55] <Lure> tfheen: from http://people.ubuntu.com/~fabbione/irclogs/ubuntu-devel-current.html
[11:57] <tfheen> Lure: thanks.
[11:57] <tfheen> I'm so not happy about that, but I'll discuss that with mdz in the morning.