[02:44] <alex-weej> http://alex-weej.blogspot.com/2006/03/human-user-group.html
[02:48] <Hobbsee> alex-weej: look at the callender
[02:48] <alex-weej> Hobbsee: ?
[02:49] <alex-weej> yes i know i posted it in March, i still think it's a good idea! :P
[02:49] <Hobbsee> alex-weej: it's the day after christmas day, or still christmas day - most people arent going to be looking @ irc
[02:49] <alex-weej> oh right :P
[02:49] <alex-weej> well... you saw it, right? :P
[02:50] <Hobbsee> sure, but i dont deal in gnome
[02:52] <mjg59> alex-weej: I'm not entirely convinced by the assertion. Real users go in the 1000+ range. System users shouldn't.
[02:52] <mjg59> So why worry?
[02:53] <alex-weej> well when i setup a package
[02:53] <wasabi_> heh, did somebody just change back to spacial nautilus? :)
[02:53] <wasabi_> my naut just switched back after an update
[02:53] <alex-weej> and it creates a new user with uid=1004
[02:53] <Lathiat> it shouldnt
[02:53] <alex-weej> i don't want to see that jazz in my facebrowser!
[02:53] <alex-weej> Lathiat: that's not the point
[02:53] <mjg59> alex-weej: It shouldn't
[02:53] <Hobbsee> fix the package so it shouldnt?
[02:53] <mjg59> And that *is* the point
[02:53] <alex-weej> i don't know whether you guys realise this
[02:54] <alex-weej> but there are only 999 numbers less than 1000.
[02:54] <wasabi_> Actually there are infinity, when you consider negative values.
[02:54] <mjg59> And you're going to have over 999 system users on a single system?
[02:54] <wasabi_> Just making sure you know. ;0
[02:54] <alex-weej> mjg59: again, not the point
[02:54] <mjg59> alex-weej: ...
[02:54] <alex-weej> ideally, we'd be using UUIDs anyway :/
[02:55] <wasabi_> oh geeze
[02:55] <Lathiat> is that a joke?
[02:55] <alex-weej> it's a bit crack getting an ext3 file system off a mate's computer and seeing all the file permissions set to local users that makes no sense
[02:55] <wasabi_> nautilus is closing widnows behind itself now too
[02:55] <Lathiat> ahh
[02:55] <alex-weej> Lathiat: absolutely not :o)
[02:55] <Lathiat> so you mean, much longer UIDs
[02:55] <mjg59> alex-weej: Any package that creates a system user in the 1000+ range is breaking Debian policy
[02:55] <mjg59> So it's buggy
[02:55] <mjg59> So you report it and we fix it
[02:56] <mjg59> And that is *entirely* the point
[02:56] <wasabi_> If you WANT to argue about having only 999 system users, then we can certainly do so.
[02:56] <alex-weej> mjg59: but debian rules the world. you can't fix third party packages.
[02:56] <wasabi_> =)
[02:56] <alex-weej> wasabi_: i'd say a more useful argument is based on the case i suggested above
[02:56] <mjg59> alex-weej: If .debs don't conform to Debian policy, they're broken
[02:56] <wasabi_> You mean third party apps?
[02:56] <mjg59> It's not our job to fix them
[02:57] <wasabi_> If a third party app overwrites all of /usr, we're not going to fix it, either.
[02:57] <Hobbsee> alex-weej: and they dont tend to make it in the archives, either
[02:57] <alex-weej> when uid's no longer become simply sequential numbers, you CAN'T rely on it being somewhere above 1000.
[02:57] <wasabi_> Sure you can.
[02:57] <mjg59> alex-weej: When it's an issue, we'll deal with it
[02:57] <alex-weej> it is an issue now. see my example!
[02:57] <wasabi_> Wish teh third party software?
[02:57] <alex-weej> with the file system
[02:58] <wasabi_> Huh?
[02:58] <mjg59> It's an issue because you've installed some piece of shit package produced by either an incompetent or someone ignorant of the fact that it's behaving like that
[02:58] <alex-weej> uid 1003 on a foreign system does not correspond to uid 1003 on my system
[02:58] <wasabi_> That would be correct.
[02:58] <alex-weej> if they were both UUIDs, this would not be a problem
[02:58] <wasabi_> ...
[02:58] <alex-weej> as they would never match anyway
[02:58] <wasabi_> You want to fix that, go whine someplace else.
[02:58] <wasabi_> Specifically to the guys behind POSIX.
[02:58] <mjg59> uids are 32-bit numbers.
[02:58] <Lathiat> thats a much bigger issue
[02:58] <mjg59> Ubuntu can't fix that.
[02:58] <wasabi_> Yes, that issue is huge.
[02:59] <Lathiat> also, UUIDs don't guarantee that'l *never* happen
[02:59] <mjg59> A 32-bit number does not usefully identify every UNIX account on the planet
[02:59] <Lathiat> it just increases the problem space ;)
[02:59] <alex-weej> no, but they make the chances unreasonably slim
[02:59] <alex-weej> pfft
[02:59] <alex-weej> "Windows does it"
[02:59] <wasabi_> alex-weej: Don't get ME wrong anyways, I am all for UUIDs.
[02:59] <wasabi_> I have other issues which are impossible with 32 bit UIDs.
[03:00] <wasabi_> Which I am keenly interested in solving. I just know it's not something a few guys in the ubuntu channel can tackle right now. ;0
[03:00] <alex-weej> lol
[03:00] <alex-weej> i wasn't implying that it was
[03:00] <mjg59> It's not a problem that gets fixed by creating a human group
[03:00] <alex-weej> true
[03:01] <alex-weej> but the human group means that one less thing is relying on numerical sequentiality of uids
[03:01] <mjg59> In Debian-based systems, real users have UIDs that are 1000 or above. Other users don't.
[03:01] <wasabi_> Wait, what relies on numerical sequentialiy?
[03:01] <mjg59> So, given that users are already clearly defined as human or not-human, there's no need for a human group
[03:02] <wasabi_> ranges != sequentially. EVen Windows has "ranges"
[03:02] <wasabi_> Special SID prefixes which correspond to system users, etc.
[03:02] <wasabi_> There's nothing significantly different from <1000 and that.
[03:03] <alex-weej> except there are probably more than 999 different system services that might require specific well-known UID available for windows :P
[03:03] <mjg59> alex-weej: Oh, hey, if that's your concern, we only offer 100 specified UIDs
[03:03] <wasabi_> alex-weej: Nope, not really.
[03:04] <mjg59> Oh, wait
[03:04] <wasabi_> alex-weej: There's like, 30.
[03:04] <mjg59> Sorry, 4100
[03:04] <mjg59> 5100
[03:06] <alex-weej> ok
[03:07] <alex-weej> i can live with debian's policy of humans having 1000 <= UID < whatever
[03:07] <wasabi_> Anyways, there are other more worthy things to complain about.
[03:07] <wasabi_> LIke that the face browser actually ENUMERATES using getpwent from 1000 to the end. ;)
[03:07] <alex-weej> like Firefox?
[03:09] <alex-weej> tbh fuck it all
[03:09] <alex-weej> we should be logging people into ubuntu with OpenID!
[03:09] <wasabi_> No way.
[03:09] <wasabi_> But it should be an option. :)
[03:09] <alex-weej> wasabi_: that WAS a joke. :P
[03:13] <alex-weej> yeah anyway going back to the whole UID jazz just for a second
[03:14] <alex-weej> genuine question: what exactly makes stuff like sysvinit replaceable within the scope of ubuntu, but not stuff like that?
[03:16] <mjg59> Changing uid to something other than a 32-bit value would break POSIX compliance
[03:16] <mjg59> And hence NFS, NIS, LDAP, Hesiod and so on
[03:18] <alex-weej> right
[03:18] <alex-weej> that's less than good
[03:19] <alex-weej> it does seem crack when you put it that way
[03:19] <alex-weej> but still useful :F
[03:28] <wasabi_> The problem is a real world one.
[03:28] <wasabi_> Windows SIDs work, for distributed environments, you know.. Active Directory.
[03:28] <wasabi_> And they're pretty much a requirement for that line of work, imo. So, we'll fix it eventually.
[03:28] <wasabi_> Smart UID allocation or mapping policies aren't bad though.
[03:29] <wasabi_> 32 bits really is enough for a single organization. Just not for every individual on the face of the planet.
[04:01] <alex-weej> wasabi_: *nod*
[05:25] <bluefoxicy> I'm being hit up for i686 again
[05:25] <bluefoxicy> Someone wants my help swaying the devs to deprecate i586 and i486 support
[05:26] <bluefoxicy> I told him to profile archlinux against archlinux i486 instead of against Ubuntu
[05:28] <Hobbsee> why dont you do the benchmarks yourself, and then talk?
[06:05] <bluefoxicy> Hobbsee:  I did some microbenchmarks with nbench a while back and got zip.
[06:05] <bluefoxicy> I guess I could do some realistic benchmarks
[06:05] <Hobbsee> that's the only way you'll get input on it
[06:06] <bluefoxicy> maybe build libz + gzip and then compress tons of crap (random data?  The analysis that runs takes a long time to figure out it's not getting anywhere)
[06:07] <bluefoxicy> but that still leaves actually justifying dropping i486/i586 support OR adding another arch repo; I'd need some pretty impressive bench mark numbers for that, and it's just not going to happen.
[06:07] <bluefoxicy> doh.. now I'm curious as to the exact numbers I'll get :/
[06:08] <Hobbsee> haha
[06:10] <bluefoxicy> speaking of making things faster, did anyone ever pin down the firefox issue
[06:10] <bluefoxicy> it used to be that it would load in like 2 seconds (any other distro; official builds); unless you got it from Ubuntu, then it took 14-25 seconds to load
[06:11] <wasabi_> Um. When making a new package upload, what's the proper way to signify that it should be placed in multiverse?
[06:11] <wasabi_> It's been awhile.
[06:11] <bluefoxicy> I believe that goes on the first line in the changelog entry
[06:11] <wasabi_> distro should go there.
[06:12] <bluefoxicy> yer right
[06:12] <Hobbsee> wasabi_: i'm not sure.  same procedure as universe, i'd expect
[06:12] <Hobbsee> one of the distro people would know
[06:13] <wasabi_> Heh. Last time I did this there was no procedure other than asking somebody to NOT put it in universe.
[06:20] <Hobbsee> ah
[06:20] <Hobbsee> okay then
[11:51] <neutrinomass> Comments on bug 77009 ? (or should I raise this on the mailing list? )
[11:51] <Ubugtu> Malone bug 77009 in venkman "failed deps on iceweasel | iceape-browser | icedove" [Undecided,Unconfirmed]  http://launchpad.net/bugs/77009
[12:18] <Mithrandir> neutrinomass: looks like something which should be discussed on ubuntu-devel
[12:22] <neutrinomass> Mithrandir: ok, I'll send a mail sometime today 
[12:25] <Mithrandir> Hobbsee: how's your xmas been?
[12:27] <Hobbsee> Mithrandir: good :)
[02:13] <adam0509> Sorry, to distrurb, but do someone know about this package : xserver-xorg-input-joystick  ?
[02:14] <Hobbsee> adam0509: probably not today, as it's still a public holiday in most countries.  and your question is very board
[02:14] <Hobbsee> *broad
[02:15] <adam0509> ok thx
[05:38] <bddebian> Heya
[05:39] <_ion> hi
[05:40] <bddebian> Hello _ion
[05:52] <Adri2000> cjwatson, Mithrandir, any archive admin: I have just seen that the binary package homebank-data is in main, while homebank is in universe. homebank-data should be in universe too. I don't know why homebank-data landed in main, but see yourself: http://archive.ubuntu.com/ubuntu/pool/main/h/homebank/ :p
[05:57] <fdoving> Adri2000: confirmed. report a bug? 
[05:59] <Adri2000> yeah, will do if no archive admin answers
[06:01] <bhale> Adri2000: they are on holiday and unlikely to answer until after the new year..
[06:01] <bhale> subscribe ubuntu-archive to your bug report
[06:09] <Adri2000> done, bug 77179, fdoving: you can confirm it :)
[06:09] <Ubugtu> Malone bug 77179 in homebank "homebank-data should be in universe, not in main" [Undecided,Unconfirmed]  http://launchpad.net/bugs/77179
[06:12] <fdoving> Adri, confirmed.
[06:13] <Adri2000> thanks
[06:17] <wasabi> Does anybody have a good overview of the plan regarding udev in the initramfs and mdadm/evms/lvm integration?
[06:18] <wasabi> And it's current state.
[06:45] <wasabi> Latest kernel is fun. Locks on boot at Checking aperture.
[06:57] <Mithrandir> wasabi: "make udev call evms_activate/vgchange/raidstart when discovering block devices".
[06:57] <wasabi> heh
[07:31] <cjwatson> wasabi: multiverse> it's Section: multiverse/whatever in debian/control; but you *can* just let an archive admin override it
[07:39] <cjwatson> wasabi: (you can also mark it non-free/blah or contrib/blah as appropriate per Debian policy, and we'll take that as a signal that it should go in multiverse)
[07:39] <wasabi> Ahh. Makes sense. So, seeing as I already uploaded it, and it's in NEW, who do I contact to make sure it gets in the right place?
[07:39] <wasabi> (i did not prefix section)
[07:40] <Mithrandir> it'll be done by hand by whoever NEWs it.
[07:57] <cjwatson> wasabi: no need
[07:57] <cjwatson> er, as Mithrandir said
[07:58] <cjwatson> wasabi: if it goes in the wrong place, you can always file a bug afterwards and subscribe ubuntu-archive to have it moved
[07:58] <wasabi> k
[07:58] <wasabi> thanks. =)
[08:18] <alex-weej> is there some kind of online ubuntu hardware compatibility database?
[08:21] <lotusleaf> alex-weej: well there's !hwdb
[08:21] <alex-weej> lotusleaf: ta
[08:21] <lotusleaf> alex-weej: and some tidbits on the wiki
[08:22] <alex-weej> hmm
[08:22] <alex-weej> dunno if hwdb is what i think it is
[08:22] <alex-weej> have you ever seen Wine's AppDB?
[08:23] <alex-weej> "Submissions Total: 441642 Today: 235416"
[08:23] <alex-weej> wtf? half the hwdb submissions submitted today?
[09:23] <Zober2> hi guys
[09:23] <Zober2> raid array in dmraid is not recognized by gparted, is there something i need to do before gparted will see the raid array?
[11:27] <_MMA_> I hope its just me but is there something in place that wont let me go to Feisty from Edgy?
[11:27] <_MMA_> I tried to put in a Feisty daily and use update-manager to upgrade but it said: "Authentication the upgrade failed. There may be a problem with the network or the server.
[11:28] <_MMA_> I was testing the -lowlatency kernels but messed up my partition and had to reinstall.
[11:36] <Adri2000> !seen mvo
[11:37] <Adri2000> thanks nickserv