[00:01] <slangasek> heh
[00:01] <hwilde> so is anybody looking at usb device reassignment for intrepid ?  everytime I reboot the ttyUSBs are all random
[00:01] <hwilde> I have to run a script in rc.local to try and sort them out and create symlinks
[00:01] <slangasek> not that I'm aware of
[00:02] <hwilde> yeah I can't find any bug posts or anything even related
[00:02] <slangasek> in general, device names are not guaranteed to be persistent across reboots unless you use udev rules to link them
[00:02] <slangasek> I don't know about USB ttys though, I've never even had one of those let alone two
[00:02] <hwilde> right, but udev doesn't seem to know the diff
[00:11] <hwilde> slangasek, you know udev ?
[00:13] <hwilde> slangasek, if you can make udev rules that figure out these usbs you would be my hero  http://pastebin.com/m5ac22ce0
[00:13] <slangasek> I only know udev on a cargo-cult basis
[00:20] <hwilde> where can I find the udev devs :)
[00:20] <slangasek> well, the Ubuntu udev maintainer just timed out of the channel... :)
[00:20] <hwilde> doh!
[00:21] <hwilde> hey probably read my pastebin and locked up
[00:22] <lucent> slangasek: ah, is there no udev/DBUS/HAL script to use to kick network-manager into gear before you login?
[00:22] <slangasek> lucent: the only keystore network-manager currently supports is gnome-keyring
[00:23] <slangasek> which makes logging in to password-protected networks ... problematic
[00:23] <lucent> oh gosh.
[00:23] <lucent> gnome isn't even there yet, so uh, I see now.
[00:23] <lucent> it would require some kind of "IMEH CHARGIN MY LAZERZ" followed by udev/DBUS/HAL hack
[00:24] <lucent> that's not a typical use case for a desktop system, is it?
[00:25] <slangasek> are you defining "desktop" as "running GNOME", or as "not a laptop"?
[00:25] <slangasek> because roaming Kerberos authentication is totally a valid use case
[00:25] <hwilde> the problem with udev is that it only sees the USB_BUS and USB_DEV and those swap every reboot... udevinfo does not have enough specific info to define custom rules   http://pastebin.com/m5ac22ce0
[00:25] <lucent> I'm defining wifi on boot as the use case
[00:25] <lucent> how common is that anyways?
[00:26] <hwilde> wifi on boot pre-login
[00:26] <slangasek> lucent: not as common as it ought to be, but it's *my* chosen configuration. :)
[00:28] <slangasek> Kerberos is pretty important at the enterprise level, both in the context of AD and for more traditional Kerberos setups
[00:28] <lucent> there's no sensible default for this IMO
[00:28] <slangasek> sure there is: have a keystore that doesn't have a password, that root is allowed to read
[00:28] <lucent> unless you're going to hack GDM to have a wifi selector
[00:28] <slangasek> they're friggin' wep keys, not state secrets
[00:29] <slangasek> (unless you work for the goverment, then, ok, maybe the wep keys are state secrets)
[00:29] <lucent> it's not just the keystore though that is a problem
[00:29] <lucent> it's the usability
[00:29] <hwilde> you could have two step login,  first local to start the network support, then over AD or whatever
[00:29] <lucent> what if you have a wep + VPN that you log into before getting network access to the domain?
  I don't consider "which network should we connect to?" to be a per-user setting
[00:30] <slangasek> so I don't think it should use per-user password stores
[00:30] <slangasek> hwilde: this is contrary to the goal of Single Sign-On
[00:30] <lucent> it's per-user because of the use of keystores
[00:30] <hwilde> but hten how does admin login to the secure network when you break your enterprise login :)
[00:30] <slangasek> hwilde: I *can* do that, but it's bloody annoying
[00:30] <hwilde> slangasek, consider me the devil's advocate then until I have consistent usb devices :)
[00:31] <slangasek> heh
[00:31] <lucent> slangasek: IMO the course of action is a new package which includes a boot-time configuration interface for networking
[00:32] <slangasek> I don't want to have to interact at all with the system in order to get it to come up on the network either, if that's what you mean
[00:32] <lucent> specifically "extra weird" networking like wifi on boot, or vpn tunnels
[00:32] <slangasek> system password store + n-m --> bring the network up without me having to touch it
[00:33] <slangasek> n-m actually already does this for interfaces that don't require passwords
[00:33] <slangasek> but if it's a password-encrypted network, you have to log in and run the applet to get anywhere
[00:33] <lucent> secured network access on boot without user input?  That's not security!
[00:33] <slangasek> iz broken
[00:33] <slangasek> er
[00:33] <hwilde> might as well autologin at that point
[00:33] <slangasek> sure it is
[00:33] <slangasek> um no
[00:33] <lucent> seriously, hwilde
[00:34] <lucent> let's think this through here
[00:34] <lucent> how do we provide secured network access on boot, before gdm
[00:34] <lucent> so that gdm can login to resources on a secured network
[00:34] <hwilde> if this is restricted to enterprise networks, there should only be one network to connect to.
[00:34] <slangasek> hwilde: ahem, you already admitted above that you do the same thing on your system, only without the use of n-m
[00:35] <slangasek> there's no reason it would be restricted to enterprise networks
[00:35] <hwilde> my system is an autonomous robot so yeah it starts up automagically.
[00:35] <slangasek> it's restricted to machines that are part of an enterprise *environment*, which may include both enterprise wifi and public roaming
[00:36] <hwilde> slangasek, but there is only *one* network in that enterprise environment that you need pre-login connectivity aka where is the active directory
[00:36] <slangasek> i.e., if I have a company-issued laptop, I want to get Kerberos authentication at the gdm screen, whenever I'm connected to a network
[00:36] <slangasek> hwilde: no, why would you assume I would only use Kerberos when I'm directly connected to the network? :)
[00:37] <hwilde> I dunno, I guess I am missing why everyone objects to the method you have worked out
[00:37] <slangasek> it's my single sign-on, and I use it for everything from verifying my own password (company laptop, remember), to accessing fileshares (when I'm in range), to authenticating my email client
[00:37] <hwilde> if you can't login, you can't have access to the network,   so why not startup networking before login?
[00:38] <slangasek> hmm?  I don't claim that everyone objects to that
[00:38] <slangasek> it just doesn't work currently with n-m
[00:38] <hwilde> ohhhh
[00:38] <slangasek> because no one's implemented a system-level password store, because GNOME people think about these things upside-down. :)
[00:39] <lucent> I'm asking how do we provide secure network layer access before gdm so we can login
[00:39] <hwilde> typically they respond to popular demand
[00:39] <hwilde> but that is an important piece to taking market share away from windows in enterprise environments
[00:39] <lucent> maybe a power-on password?
[00:40] <lucent> USB thumbdrive as a key?
[00:40] <slangasek> lucent: if you really have a network that only a subset of your users are allowed to connect to, fine -- don't put that key in the system password store? :)
[00:40] <hwilde> what if you use an encrypted rsa/dsa key to authenticate to the network at first, then you can to kerberos login?
[00:40] <slangasek> and then it would work the same way that it works today
[00:40] <lucent> handing out physical layer access to your corporate network is not very bright
[00:41] <lucent> there should be a way to add a hook for VPN access too
[00:41] <slangasek> handing out physical access to your corporate laptop is probably not very bright either then ;)
[00:41] <lucent> shit gets stolen.  I mean, things happen
[00:42] <hwilde> encrypted partition
[00:42] <lucent> Okay, encrypted partition, good.  Now you need a password at boot to access your system?
[00:42] <hwilde> no the encrypted partition has the network key that gets online pre-login
[00:42] <slangasek> in that case, yes
[00:42] <hwilde> then the user can authenticate with AD or kerberos or whatever
[00:42] <lucent> in this scenario, having a package which implements a boot-time selectable network pre-config makes sense
[00:43] <lucent> without a key
[00:43] <lucent> instead you still need a means to be able to take user input though
[00:43] <lucent> like the VPN passphrase
[00:43] <lucent> unless you're storing VPN passphrases (not possible with OTP and not likley to be permitted by corporate)
[00:44] <lucent> the problem remains that you need user interaction to get network connectivity before you log in
[00:44] <slangasek> having to VPN in before you log in == doesn't work with Windows AD member machines either
[00:45] <lucent> slangasek: I am not familiar with Windows AD, what is this?
[00:45] <slangasek> AD == Active Directory?
[00:45] <hwilde> consider yourself lucky...
[00:45] <lucent> oh okay
[00:46] <hwilde> gtg,  if you see Keybuk or any udevs ask them if mine is a lost cause:    http://pastebin.com/m750bddea
[00:46] <slangasek> well, hmm, I don't /think/ you can start a VPN connection from Windows before logging in, but maybe you can run the VPN client as a system service too
[00:46] <lucent> The only big corp I worked for that had any level of secure login was Google, Inc.
[00:48] <lucent> so... not a lot of Microsoft anything
[00:48] <lucent> VPN pre-login would be a cool enhancement
[00:49] <slangasek> right, I won't ask you about what kind of VPNing they do, there are other people on the channel I can bug about that who probably have a better chance of being able to legitimately tell me :)
[00:49] <slangasek> but Kerberos is certainly a class of technology that's in Google's ballpark
[00:50] <lucent> I was a datacenter Monkey, I can't tell you any more than this :)
[01:16]  * toresbe thinks he's found what should've been filed as an RC bug...
[01:17] <toresbe> apt-get upgrade to hardy breaks on python-irclib
[01:17] <lucent> uh...
[01:17] <lucent> aren't we supposed to use the upgrade-manager?
[01:18] <toresbe> sorry. "upgrade-manager" breaks on python-irclib. :)
[01:18] <toresbe> as do subsequent invocations of dpkg --configure -a.
[01:19] <sistpoty> toresbe: any details? (e.g. output of "upgrade-manager" *g*)=
[01:20] <wgrant> toresbe: Let me guess... python-central complains about files being in the directory that aren't owned by it?"
[01:20] <toresbe> wgrant: won't remove local files, something to that effect
[01:20] <wgrant> It's like the python-opengl one I found; it needs C/R on python2.[34]-irclib
[01:20] <sistpoty> hey wgrant... where's fujitsu? *g*
[01:20] <wgrant> Right, that's right.
[01:20] <wgrant> sistpoty: Back in Hardy.
[01:20] <sistpoty> heh
[01:21] <wgrant> toresbe: Please file a bug, and I'll get it SRUed shortly.
[01:21] <toresbe> I'm in console mode trying to prepare for a reboot (my grub setup is kinda' awkward and roundabout)
[01:21] <jdong> whoopsies
[01:21] <jdong> note to self:
[01:21] <toresbe> wgrant: I don't know what an SRU is, but I'll do my best.
[01:21] <toresbe> :)
[01:21] <wgrant> toresbe: Stable Release Update.
[01:21] <jdong> do *NOT* make a loopback device on a LV and then use it to extend the LV itself
[01:21] <toresbe> wgrant: cool
[01:21] <toresbe> jdong: haha
[01:21] <wgrant> toresbe: For now, it's fairly easily recoverable.
[01:21] <jdong> well that was my attempt at perpetual motion.
[01:22] <toresbe> wgrant: I just removed the package, worked a treat ;)
[01:22] <wgrant> toresbe: apt-get remove python2.4-irclib
[01:22]  * lucent rolls through a Microsoft Windows XP install in kvm
[01:22] <wgrant> python-central is so unbelievably fragile.
[01:22] <lucent> this is so difficult, versus Hardy install
[01:22] <wgrant> toresbe: Great.
[01:22] <toresbe> heh. :)
[01:22] <toresbe> anyway, the rest of it is working great. Thanks, everyone in here, for another neato release.
[01:23] <wgrant> python-central-related upgrade failures should be top priority, as they can bring down an entire upgrade very easily.
[01:23] <lucent> Hardy installer is missing a feature that is in Windows XP installer:   It's not telling me that I'm an idiot.
[01:23] <lucent> :P
[01:23] <toresbe> haha
[01:23] <wgrant> lucent: Heh. Indeed.
[01:23] <toresbe> I'm thoroughly bummed about the Heron T-shirt being limited-edition. It's so good-looking I'd like one for myself.
[01:23] <wgrant> toresbe: They had Gutsy ones until very near Hardy's release.
[01:23] <wgrant> toresbe: I'm sure you can get one.
[01:23] <lucent> oh, I agree toresbe, that design is ill (really good)
[01:24] <toresbe> lucent: straight up, yo.
[01:24] <lucent> I was looking and thinking, if I had money right now, I would buy a T and wear it to burning man
[01:24] <lucent> I don't know, does it have "Ubuntu" words on it? I don't want the words if I'm wearing something there
[01:24] <lucent> the design is really cool and I like that
[01:25]  * toresbe wonders if he has met lucent at debconf, which would explain why he felt the need to explain "ill" to him.
[01:25] <toresbe> :)
[01:25] <lucent> har, no *nix gatherings since a long time now
[01:26] <wgrant> toresbe: Wait, are you upgrading from Dapper?
[01:26] <lucent> I watch those things from a safe, moderate-intelligence required, distance.  They're really full of people who are overwhelmingly intelligent.
[01:26] <toresbe> lucent: Hey, I'm an idiot, I have a great time there.
[01:27] <toresbe> lucent: I do lots of stuff that require no computer skills. Last debconf, I fixed a church organ with duct tape.
[01:27]  * jdong moans about the word idiot's high standards these days
[01:27] <toresbe> wgrant: no, Gutsy.
[01:27] <wgrant> Hmm, that's not what I thought it was, then.
[01:27] <lucent> toresbe: groovy!   I'm into tinkering with other people's code and engineering work, never my own
[01:27] <wgrant> toresbe: Did it mention any file in particular?
[01:29] <toresbe> wgrant: nope... sorry, I probably should've made allowance for letting some devs poke around before I removed the package.
[01:29] <wgrant> toresbe: I should be able to reproduce it here.
[01:29] <toresbe> wgrant: cool.
[01:29]  * toresbe reboots, takes opportunity to stuff in a video capture card he's been wanting to try
[01:29] <toresbe> up 11 days, pfah
[01:30] <wgrant> toresbe: Did you have python2.4-irclib installed?
[01:39] <toresbe> toresbe@fortran:~$ dpkg-query -f '${Status},${Package}\n' -W *irclib*
[01:39] <toresbe> purge ok not-installed,python-irclib
[01:39] <toresbe> install ok installed,python2.4-irclib
[01:39] <toresbe> wgrant: see above :)
[01:39] <wgrant> That's impressive.
[01:40] <wgrant> That should have been removed when you upgraded from Dapper.
[01:40] <wgrant> So it is the bug I thought.
[01:40] <toresbe> hah
[01:40] <toresbe> That bug is *so* three releases ago? :)
[01:40] <toresbe> Ugh! Sad to see that gnome-terminal hasn't gotten any less useless on this machine. Maybe it's a config file issue or something, but irssi screen redraws take several 1/10ths of a second.
[01:41] <toresbe> nearly a whole second
[01:41]  * toresbe goes back to uxterm
[01:41] <toresbe> Oh, and another cute one...
[01:41] <wgrant> toresbe: I use gnome-terminal fine with screen here.
[01:42] <toresbe> it works well on my work Hardy machine with the same nVidia driver... probably just a weird issue.
[01:42] <toresbe> http://gunkies.org/stuff/hardy-ohdearme.png :)
[01:42] <wgrant> ...
[01:42] <wgrant> Nice.
[01:43] <wgrant> The /Music isn't just wrapped off the first one, is it?
[01:43] <toresbe> Good question. I'll kill gnome and try again
[01:44] <toresbe> Hrm. Could well be, but resizing is not possible. :)
[01:44] <toresbe> So most probably it's doing some kind of automated sizing which may or may not be to blame
[01:46] <toresbe> Actually, a quick question which might be construed as a user support question: Anyone here have any experiences with, and/or advice for a video4linux previewing application?
[01:46] <toresbe> I used to use tvtime, but this is just an svideo grabber.
[01:52] <LaserJock> wgrant: oh no, what happened to your nick?!
[01:52] <LaserJock> say it ain't so ;-)
[01:52] <wgrant> LaserJock: It got left in Hardy.
[01:54] <LaserJock> I thought maybe Colin and Tollef had pulled you to the dark side
[01:55] <wgrant> Not as far as I'm aware.
[01:55] <jdong> you'er not supposed to be aware
[01:55] <jdong> like that pill thingie they use in night clubs
[01:55] <wgrant> Hahah.
[02:10] <slangasek> pill... thingie?
[02:10] <slangasek> nevermind, I didn't ask
[02:11] <sladen> *babies*
[02:12] <sistpoty> alcohol works as countermeasure, I've heard from s.o. :P
[03:52] <MattJ> Don't know if this is the right place to ask, but are there plans to have the .25 kernel in the repositories at some point?
[03:55] <wgrant> MattJ: For Intrepid, yes.
[03:56] <wgrant> For Hardy, no.
[03:56] <MattJ> Not even in backports?
[03:56] <lamont> wgrant: except that I expect it'll be .26 or later for intrepid...
[03:56] <wgrant> lamont: But it's 2.6.25 now.
[03:56] <lamont> sure
[03:56] <wgrant> MattJ: Backporting kernels is truly evil.l
[03:57] <MattJ> ^^
[03:58] <MattJ> iwlwifi is not working for me, and I need it to
[03:58] <wgrant> Getting 2.6.25 is not the soluition to that.
[03:58] <wgrant> Filing a bug and getting it fixed is.
[03:58] <MattJ> I found someone with a similar/same problem on lkml, they say it is fixed in .25
[03:59] <lamont> wgrand++
[03:59] <MattJ> Since the issue is not a security issue, is it likely to be fixed in Hardy?
[03:59] <wgrant> lamont: Pfft, I thought it would be easier to type than Fujitsu.
[04:00] <wgrant> MattJ: Very probably for 8.04.1
[04:00]  * MattJ files bug
[04:00] <wgrant> A lot of kernel regressions are scheduled for fixing then.
[04:00] <MattJ> :)
[04:00] <wgrant> 8.04.1 should happen in very early July.
[04:00] <pwnguin> MattJ: the proper technique there is to bisect the kernel
[04:01] <MattJ> and any idea of the short term solution for me? Gutsy (with ipw3945) worked ok
[04:01] <wgrant> MattJ: File a bug and see what the kernel devs say.
[04:01] <MattJ> k, will do so now
[04:01] <wgrant> Perhaps use the Gutsy kernel for now.
[04:02] <MattJ> I used to use the Gutsy kernel on Feisty, how funny :)
[04:03] <wgrant> MattJ: Is there anything special about your config? I've used Hardy's kernel on several ipw3945 boxes recently, and it seems to work.
[04:04] <MattJ> ipw was the old (pre-Hardy) driver, Hardy uses the newer iwlwifi (from the kernel)
[04:04] <MattJ> Nothing special, this is a fresh install
[04:05] <MattJ> A lot of people have had problems even detecting networks, but using the hardy backports modules seems to help those
[04:05] <MattJ> My problem is that it won't set up an ad-hoc network
[04:05] <MattJ> which I use every day
[04:10] <MattJ> The hardest part of filing a bug is knowing which package to file under
[04:10] <wgrant> MattJ: linux, probably.
[04:12] <MattJ> That warns me about filing a bug for upstream linux, etc.
[04:15] <wgrant> MattJ: The linux package. https://launchpad.net/ubuntu/+source/linux/+filebug
[04:15] <MattJ> Ah, thanks :)
[04:19] <sladen> MattJ: I filed this as high about 6weeks before release
[04:19] <sladen> MattJ: but the more dups the better
[04:19] <MattJ> sladen: Oh, where?
[04:20] <MattJ> I couldn't find any with my specific problem
[04:20] <sladen> MattJ: file _your_ bug
[04:21] <sladen> if it's a dup, it'll get duped, but at least there will be an independent report in your words
[04:21] <MattJ> How do duplicates help? :)
[04:21] <MattJ> If you insist
[04:21] <dsas> MattJ: It helps persuade someone that the problem is important
[04:22] <MattJ> If it were my bug tracker I would be annoyed at someone knowingly filing a duplicate :)
[04:22] <sladen> some people do
[04:22] <sladen> for me (personally), a good indicator of the relative size of a problem are the number of dups
[04:23] <sladen> and with them you get separate individual reports (multiple angles onto a problem) and not just misleadings "me toos" are actually turn out to be something else and need weeding out
[04:25] <sladen> MattJ: [after you've filed yours], start at  bug
[04:26] <sladen> MattJ: [after you've filed yours], start at  bug #205390 and recursively follow the dups
[04:26] <ubotu> Launchpad bug 205390 in ubuntu "8.04 beta; wireless broken after upgrade (ipw3945/iwl3945/ifrename/udev) (dup-of: 183968)" [Undecided,New] https://launchpad.net/bugs/205390
[04:26] <ubotu> Launchpad bug 183968 in udev "mac80211 "master" interface matches existant persistent network rules" [Critical,Fix released] https://launchpad.net/bugs/183968
[04:37] <MattJ> sladen: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/222302
[04:37] <ubotu> Launchpad bug 222302 in linux "iwl3945 can't set parameters for ad-hoc network" [Undecided,New]
[04:38] <MattJ> I don't have the wlan0_rename issue other people have
[04:43] <MattJ> I'll investigate installing the Gutsy kernel tomorrow
[04:43] <MattJ> Thanks for the help :)
[04:51] <hdevalence> has anyone ever heard of an idea where you would use a torrent to download debs for version upgrades instead of having everybody upgrade veeerrrryyy slowly when a new version is released?
[04:52] <hdevalence> because torrents of course scale opposite to http/ftp downloads, so if you did that upgrading would go much faster
[04:53] <mgolisch> ?
[04:53] <mgolisch> how would that be faster?
[04:53] <mgolisch> i mean you would need a dedicated torrent for each file
[04:53] <mgolisch> that would unmanageable
[04:53] <hdevalence> mgolisch: no
[04:54] <hdevalence> no, you don't
[04:54] <mgolisch> how else?
[04:54] <hdevalence> you can make a torrent of all the deb files and then only download the ones you want
[04:54] <mgolisch> if it where just one you would have to wait for the whole thing to complete
[04:55] <mgolisch> or did i get something wrong?
[04:55] <mgolisch> i mean you cant control in what order the chunks are downloaded
[04:56] <hdevalence> then how is it in KTorrent I can proitize files and choosenot to download them?
[04:56] <hdevalence> ick
[04:56] <hdevalence> prioritize files and choose not to
[04:56] <mgolisch> hm no idea maybe iam wrong
[04:58] <hdevalence> basically I'd be interested to know if it's been done, because it's something I think would be interesting to work on but I don't want to put in lots of effort if it's already made
[05:01] <awalton__> hdevalence, it's been done, but never very successfully. a new implementation would be nice.
[05:02] <awalton__> hdevalence, apt-torrent: http://sianka.free.fr/
[05:03] <hdevalence> i was thinking it would be neat to try getting it to use the alternate install cd torrent
[05:04] <hdevalence> but I don't know if that's the best way
[05:05] <hdevalence> anyways, I need to sleep
[05:05] <nixternal> or you can use Conary and download only the files that have changed :)
[05:35] <LaserJock> anybody around running Hardy and have gnumeric installed?
[05:36] <beuno> LaserJock, I can probably install it.  What seems to be the problem?
[05:36] <LaserJock> pi doesn't worh
[05:37] <LaserJock> *work
[05:37] <LaserJock> bug #222062
[05:37] <ubotu> Launchpad bug 222062 in gnumeric "functions don't work in formulas" [Undecided,New] https://launchpad.net/bugs/222062
[05:38] <beuno> LaserJock, installing. Might take a minute or two while the repos try not to die with all the hammering they're getting these days
[05:41] <beuno> LaserJock, what are you doing exactly?   =PI() in a cell?
[05:42] <LaserJock> hmm
[05:42] <LaserJock> I just used PI
[05:42] <beuno> just type in PI?
[05:42] <LaserJock> ah, PI() does work
[05:42] <beuno> it doesn't for me  :)
[05:42] <beuno> oh
[05:42] <beuno> yes it dows
[05:43] <beuno> er, does
[05:43] <beuno> I went for pi() because I saw the word function somewhere   :p
[05:43] <LaserJock> well, pi isn't a function
[05:43] <LaserJock> it's a constant
[05:43] <LaserJock> but ... hmm
[05:44] <beuno> right, I wouldn't of tried that if the bug didn't mention "functions"
[05:44] <beuno> LaserJock, http://www.gnome.org/projects/gnumeric/doc/gnumeric-PI.shtml
[05:45] <LaserJock> well, we need to figure out if it's a change in behavior
[05:47] <beuno> LaserJock, http://www.mailarchives.org/list/gnumeric-list/msg/2005/00002
[05:47] <beuno> tehre are emails from 2005
[05:47] <beuno> of users using pi()
[05:47] <beuno> (yes, my typing sucks today)
[05:47] <LaserJock> hmm
[05:47] <LaserJock> so what went wrong for the bug reporter I wonder
[05:49] <LaserJock> beuno: thanks for the help, I've followed up on the report
[05:49] <beuno> LaserJock, np
[05:50] <beuno> sometimes it just takes someone totally unaware of everything to look at it differently  :)
[05:50] <beuno> I force random people at the office to look at code when I'm stuck
[05:50] <beuno> and it usually is something terribly obvious
[06:15] <edugonch_> Hello, why I get compiler errors in this statement ----> this->_txt_Password->signal_insert_at_cursor().connect(
[06:15] <edugonch_>         sigc::mem_fun(*this, &LoginForm::onInsert_txt_Password));
[07:15] <DaBonBon> someone told me that snd-hda-intel was kind of disabled in hardy, to convenience the other sound card drivers. is that true?
[07:15] <DaBonBon> specifically,  --with-cards=hda-intel was not put during alsa configuration
[08:11] <infinity> DaBonBon: Can't imagine how that's possibly true, given than my hda_intel stuff is working great.
[08:14] <wgrant> Mine's working better than ever as well.
[08:29] <DaBonBon> strange, because my card is working really pathetic
[08:30] <wgrant> DaBonBon: Can you please be more descriptive?
[08:30] <DaBonBon> wgrant: yes, first of all, sound sometimes works, and sometimes doesn't. headphone automute doesn't work. i get too many channels in alsamixer
[08:31] <DaBonBon> all this used to happen pre 1.0.15 release of alsa, but after installing ubuntu-backport-modules in gutsy, which had alsa 1.0.15, all the problems were solved
[08:31] <wgrant> DaBonBon: Well, there are so many types of HDA cards that it's not surprising.
[08:31] <DaBonBon> now again, in hardy, i get those problems
[08:31] <wgrant> Please file a bug.
[08:31] <DaBonBon> wgrant: no, this card if fully supported
[08:31] <wgrant> Apparently not.
[08:31] <DaBonBon> wgrant: well, someone told me that snd-hda-intel has been deliberately disabled to facilitate other sound cards
[08:32] <DaBonBon> wgrant: yes, because it works perfectly on alsa 1.0.15, atleast in gutsy and other distros
[08:32] <wgrant> That sounds very odd, as Intel HDA is one of the most common types these days...
[08:32] <DaBonBon> exactly!
[08:32] <wgrant> And how on earth would disabling something facilitate support for other card?
[08:33]  * laga hands out tinfoil hats
[08:33] <DaBonBon> i guess i'll file a bug
[08:33] <DaBonBon> wgrant: here is a part of the logs --
[08:34] <DaBonBon> http://rafb.net/p/LQQ7ff80.html
[08:34] <DaBonBon> i don't know till how much extent is it true
[08:35] <wgrant> Ah.
[08:35] <wgrant> Well, file a bug, anyway.
[08:38] <DaBonBon> wgrant: what package should i file it against?
[08:40] <wgrant> DaBonBon: I think it's in linux-ubuntu-modules-2.6.24, but possibly linux.
[08:48] <DaBonBon> ok, i must leave.. thanks for the help, wgrant ..
[10:39] <Trewas> is hardy-proposed supposed to be completely empty (I am looking at bug 208666 and fixed audacious)?
[10:39] <ubotu> Launchpad bug 208666 in xmms-crossfade "audacious crashed with SIGSEGV in g_type_check_instance_cast()" [Medium,Fix committed] https://launchpad.net/bugs/208666
[10:43] <jeromeg> Trewas: why would it be supposed to be empty ?
[10:43] <Trewas> well, it seems to be empty :)
[10:44] <jeromeg> Trewas: what mirror are you using ?
[10:45] <Trewas> I was lookign at http://archive.ubuntu.com/ubuntu/dists/hardy-proposed/
[10:47] <jeromeg> Trewas: https://launchpad.net/ubuntu/+source/xmms-crossfade/0.3.14-1build1
[10:47] <jeromeg> it has been built only a few ours ago
[10:47] <jeromeg> it should be available soon in the archive
[10:50] <Trewas> ok, I noticed that the last message in that bug was only a few hours old so I wouldn't have been surprised if audacious was not yet there, but as hardy-proposed is completely empty I wondered if some switch had not yet been flipped or something
[10:52] <jeromeg> Trewas: maybe it's because hardy was released two days ago ;)
[10:52] <jeromeg> hardy-proposed is to handle SRU
[10:52] <Trewas> well yes, but the bug mentions that people should test the new package from hardy-proposed :)
[10:52] <jeromeg> yes, yes
[10:53] <jeromeg> it will be available soon
[10:53] <jeromeg> or it should at least ;)
[10:53] <Trewas> hehe, I'll wait and see
[10:55] <beniamino> is there *any* documentation on update-manager? /usr/share/doc contains only a one-line readme on update-manager-core
[12:34] <emgent> heya
[13:34] <twi_> what happened to binary-ppc on the mirrors?
[13:36] <geser> wasn't ppc moved to ports.ubuntu.com?
[13:37] <twi_> ah
[14:10] <crimsun> asac: a lot of people seem to be clamoring for simply enabling nspluginwrapper for i386.  What are your thoughts on doing that for 192888 instead of adjusting pulseaudio?
[14:10] <crimsun> asac: (for hardy-proposed)
[14:23] <ion_> ubotu: A link to bug #192888 for the lazy bastards among us, thank you.
[14:23] <ion_> Okay then.
[14:28] <Amaranth> bug 192888
[14:28] <ubotu> Launchpad bug 192888 in pulseaudio "firefox crashes on flash contents when using libflashsupport" [High,In progress] https://launchpad.net/bugs/192888
[14:29] <Amaranth> I would love to get nspluginwrapper on i386, flash screws up all the time even without pulseaudio
[14:30] <ogra> not for me ... with the first proposed fix to drop hal
[14:30] <ion_> amaranth: <metoo/>
[14:30] <ogra> (pulse-hal stuff indeed)
[16:48] <johanbr> Is https://launchpad.net/bugs/211205 marked private or something? The error message I get is not very informative.
[16:49] <stgraber> johanbr: this bug contains a backtrace so only QA members can read it
[16:49] <johanbr> Apparently it is marked private. At least the bot just told me so.
[16:49] <johanbr> stgraber: Maybe it'd be a good idea for launchpad to actually say that.
[16:50] <johanbr> And should private bugs really be referenced in a public changelog?
[16:51] <tuxice> features in intrepid?
[16:51] <tuxice> !intrepid
[16:51] <ubotu> Intrepid Ibex is the code name for Ubuntu 8.10, due October 2008 - For more info, see: https://wiki.ubuntu.com/IntrepidIbex
[16:51] <ion_> It supports keyboards.
[16:53] <tuxice> huh?
[16:56] <ScottK> johanbr: It's not private any more.
[16:58] <johanbr> ScottK: thanks
[16:59] <ScottK> You're welcome.
[17:10] <kennethr> when I click on the clock in gnome-panel, gnome-panel freezes.  How do I debug?
[17:13] <kennethr> How do I troubleshoot a hanging gnome-panel on ubuntu?
[17:18] <johanbr> kennethr: Change the session property for the panel from respawn to normal, then kill the panel and start it under gdb or something.
[17:19] <kennethr> johanbr: thanks, I'll start there.
[17:22] <kennethr> johanbr: how do I make use of the gnome-panel-dbg package?  Is it usefull in this case?
[17:22] <johanbr> Yes. You should have that installed to get a proper backtrace.
[17:23] <johanbr> kennethr: And probably also libgtk2.0-0-dbg and  libglib2.0-0-dbg at least.
[17:25] <kennethr> johanbr: when I get into gdb and say run, then I reproduce the hang, how do I get back to the gdb prompt to get a backtrace?
[17:25] <johanbr> Try just ctrl-c.
[17:26] <kennethr> perfect...thanks
[17:30] <scorcher7> Hi, I just uploaded a patch to launchpad that fixes launchpad bug #173772 in atomix. I have never written a patch before and the wiki says to find a dev. to review the patch for inclusion.
[17:30] <ubotu> Launchpad bug 173772 in atomix "about dialog won't close" [Undecided,Confirmed] https://launchpad.net/bugs/173772
[17:30] <scorcher7> Is that something one of you can help me with?
[17:36] <hunger> scorcher7: You have come to the right place, but I guess most devs are out for the weekend.
[17:39] <scorcher7> hunger: Should, I just come back on Monday and ask the same question?
[17:39] <hunger> scorcher7: You might get lucky and somebody might come around, but your chances are way better during the week.
[17:41] <scorcher7> hunger: Thanks for the tip.
[17:43] <hunger> scorcher7: You are welcome.
[19:36] <gustavold> hi, I dont know if it is offtopic... I've upgraded from gutsy to hardy... now every time I push a button in the numpad the X crashes
[19:38] <Ng> mdke: does the doc team want to know about mistakes in gnome documentation, or should I go and hassle the gnome folks? :)
[19:43] <dsas> Ng: Hassling the gnome folks is appreciated. You can always file a bug on the gnome documentation packages and it *may* get fixed here. (at least I think that happened in at least one release)
[19:43] <Ng> dsas: okidoki
[19:43] <dsas> perhaps gutsy we did a little, temporary fork
[19:45]  * Ng wonders how keen they will be to update docs for a point release
[19:54] <ffm_> What's the proper way to make a spec?
[19:54] <ffm_> Should I just add a page on the wiki, or should I start a forum thread about it?
[19:56] <LaserJock> ffm_: a wiki page is good
[19:57] <ffm_> LaserJock, Care to look at the one I've made? FontInstallSpec. (I'm not probably able to implement it, I only know Python ATM)
[19:57] <Nafallo> hmm
[19:57] <Nafallo> #ubuntu have support for hardy surely?
[19:58] <Nafallo> but maybe not edgy.
[19:58] <ffm_> Nafallo, What?
[19:58] <LaserJock> Nafallo: yeah, that should be updated
[19:59] <Nafallo> yay! edgy is EOL today :-)
[20:15] <Kim^J> Nafallo: Yo!
[20:17] <dmsuperman> What kind of libraries/examples exist that I could use to learn to write a keylogger/keystroke sender app in either C#, PHP, or (preferably not) C++
[20:18] <Kim^J> dmsuperman: The question is, why?
[20:18] <Kim^J> And I think you need to use some Xlib.
[20:18] <dmsuperman> Have you ever used AutoHotKey in windows?
[20:19] <Kim^J> dmsuperman: Yes.
[20:19] <dmsuperman> I basically want to emulate that, at least the text expansion portion
[20:19] <dmsuperman> there aren't any good existing solutions
[20:19] <dmsuperman> at least that I could find after much googling
[20:19] <Kim^J> dmsuperman: Check out xbindkeys source code and check what they use.
[20:19] <Kim^J> I think you WILL need to learn some C/C++ and use that.
[20:20] <mdke> Ng: it's probably helpful to file a bug on the ubuntu gnome-user-docs package too though, so that we can fix it in ubuntu then pass the fix upstream
[20:21] <dmsuperman> Alright. I learned the basic syntax of C++ quite a while ago, and I know C# now, so it shouldn't be too much of a leap I hope
[20:21] <Kim^J> Heh...
[20:21] <Kim^J> How much in C# do you know? (The more, the worse.)
[20:22] <dmsuperman> Not much
[20:22] <dmsuperman> I'm mainly a PHP guy, but I'm taking a few classes in C#
[20:22] <Kim^J> Good, smaller leap then.
[20:22] <dmsuperman> but over the years I've dabbled in various compiled languages, so i'm no stranger
[20:22] <dmsuperman> yeah
[20:22] <Kim^J> Well, get to work then. :)
[20:22] <dmsuperman> I take it C++ is the common denominator language in linux programming?
[20:23] <Kim^J> dmsuperman: Naaah, C is.
[20:23] <dmsuperman> ah
[20:23] <LaserJock> C/C++
[20:23] <Chipzz> dmsuperman: errr, xlib programming is mostly C programming
[20:23] <Kim^J> LaserJock: Well, moslty C.
[20:24] <Chipzz> not much C++ involved
[20:24] <Chipzz> if any at all
[20:24] <LaserJock> Kim^J: there's an awful lot of C++ out there
[20:24] <highvoltage> python must count for something too? or are you excluding interpreted languages?
[20:24] <LaserJock> and it's very similar to C
[20:24] <dmsuperman> I'd love to do it in PHP but I doubt it's very possible
[20:24] <Kim^J> LaserJock: I know, I code C/C++ every day. :)
[20:25] <Kim^J> highvoltage: For the thing he's doing, Python might not be the right choice. :)
[20:25] <LaserJock> so I usually put them together in any case
[20:25] <LaserJock> Kim^J: why not?
[20:25] <LaserJock> there are a lot of python libraries
[20:28] <Kim^J> Well, it looks like there's a xlib for Python.
[20:28] <LaserJock> python would be more likely than PHP or C# anyway :-)
[20:28] <dmsuperman> I don't know python though :(
[20:28] <dmsuperman> is it difficult to pick up coming from something like PHP?
[20:28] <LaserJock> it's not hard to pick up
[20:29] <Kim^J> dmsuperman: Good for you! (I think so, yes.)
[20:29] <dmsuperman> Kim^J, You say good because I _don't_ know it?
[20:29] <LaserJock> python is probably about the easiest programming language to learn, as a gross generalization
[20:29] <Kim^J> dmsuperman: Yes, because the less people know Python, the less programs get programmed in Python. Which I think is great.
[20:30] <ffm> Wait, is this a VHLL/HLL flamewar?
[20:30]  * ffm wants in.
[20:30] <dmsuperman> no
[20:30] <dmsuperman> wait
[20:30] <dmsuperman> stop
[20:30] <Kim^J> VHLL/HLL?
[20:30] <dmsuperman> I'm gonna learn python now haha
[20:31] <ffm> Kim^J, Very High Level Language (Python) vs High Level Language (C)
[20:31] <Kim^J> Very-High-Level-Language/High-Level-Language?
[20:31] <Kim^J> ffm: Ah.. :)
[20:31] <ion_> Python, ew. :-) </flame>
[20:31] <dmsuperman> psh
[20:31] <dmsuperman> I compile my code on paper
[20:31] <Kim^J> Naah, it's just that me and Python doesn't get along, I miss alot of things from C++, things doesn't behave the way I want it to, etc.
[20:31] <ffm> How hard is it to add functionality to a C gui app using python?
[20:32] <ffm> I know very little C, but have been meaning to learn it so I can contribute more to Ubuntu..
[20:32] <Kim^J> I don't what to contribute too.
[20:33] <ion_> jdong: Your hostname is shouting at me. :-)
[20:33] <highvoltage> how rude.
[20:33] <ffm> Kim^J, Well, you can implement my new spec...
[20:33] <LaserJock> ion_: he's an elitist MITer apparently ;-)
[20:33] <ffm> ;)
[20:33] <Chipzz> python indeed is very easy to learn - but make sure that you use consistent indentation settings across editors of you use more than one
[20:33] <Kim^J> ffm: Where?
[20:34] <dmsuperman> holy crap, my alternate install cd iso just peaked at 28mbps. I have advertised 15mbps connection 8-)
[20:34] <ffm> Kim^J, https://wiki.ubuntu.com/FontInstallSpec . It's very new (not even in launchpad)
[20:34] <Kim^J> Chipzz: There's only one editor! (But I won't tell since it's going to start a flamewar.)
[20:34] <Chipzz> Kim^J: no need to do so - we all know it's vim ;)
[20:34] <highvoltage> *nod*
[20:34]  * Chipzz ducks and runs away :)P
[20:35] <LaserJock> emacs forever!!
[20:35] <jdong> ion_: lol  I'm sorry to hear that ;-)
[20:35] <dmsuperman> Eclipse ftw
[20:35] <LaserJock> oh wait, I'm in vim
[20:35] <Kim^J> Chipzz: Yep
[20:35]  * dmsuperman runs
[20:35] <Kim^J> ffm: Well, I don't know to much about fonts and stuff, where can I look for more info? (You simply want to move the file right?=
[20:35] <Kim^J> r)
[20:36] <highvoltage> gooooogle
[20:37] <ffm> Kim^J, Basically, I'd like to add a button that says "Add Font" to gnome-appearance-properties, opens up a file picker dialog, and copys the file to ~/.fonts (or the system-wide equivalent)
[20:38] <Kim^J> ffm: Ok ok, well, I could do that, but I'm thinking of a download the ttf, double click it and it asks you to add that font...
[20:38] <Kim^J> Well, could do both. :)
[20:39] <ffm> Kim^J, Yeah.
[20:40] <LaserJock> wouldn't that kind of project be more appropriate in Gnome itself?
[20:40] <Kim^J> LaserJock: Well, maybe, we can create it for Ubuntu, give it to Gnome.
[20:40] <Kim^J> That is: Ubuntu --> Gnome is possible, it mustn't be Gnome --> Ubuntu all the time. .)
[20:41] <mjg59> Kim^J: Projects tend to be happier if you work with them, rather than just presenting them with a finished product
[20:41] <Kim^J> Ok...
[20:41] <LaserJock> well, I don't want to make a big deal about it, but a consistent critique of Ubuntu is that it does stuff and then they never end up going upstream
[20:41] <Kim^J> Tell that to ffm, not to me. :)
[20:43] <Chipzz> LaserJock: that works the other way around to imo
[20:43] <Chipzz> ie, upstream not looking at what has been done in ubuntu
[20:43] <Chipzz> PackageKit for example
[20:44] <LaserJock> well, yes, there is that kind of thing
[20:44] <LaserJock> but in general upstreams are supposed to be upstream
[20:44] <Chipzz> I wonder if hughise even bothered to look at what's been done in ubuntu
[20:44] <ffm> Kim^J, As I said, I can't implement it, I can just draft. Want to work on it here, or upstream?
[20:44] <Kim^J> ffm: I can work on it here.
[20:44] <ffm> LaserJock, The whole point of FOSS is that we can make changes, and the decentralized bit helps.
[20:45] <Kim^J> Just need to clear out some details on Nautilus and fonts. :)
[20:45] <ffm> Kim^J, Thanks.
[20:46] <LaserJock> ffm: well, I don't want to debate the whole point of FOSS too much, but I rather think its that we are free to join with upstream, not fork everything
[20:46] <ffm> LaserJock, We arn't forking, exactly...
[20:46] <LaserJock> ffm: you may get a lot more help upstream
[20:47] <ffm> LaserJock, If we make a successful feature, we can tell upstream about it. Plus, I like launchpad.
[20:47] <ffm> LaserJock, Ok, then what type of specs should be made in ubunu?
[20:47] <LaserJock> ffm: well, I'm not saying you can't
[20:47] <Kim^J> LaserJock: What's the point of arguing this? We ARE going to tell Gnome about it.
[20:47] <LaserJock> but in general, packaging and integration changes
[20:47] <LaserJock> things that are Ubuntu-specific
[20:48] <LaserJock> it's not telling Gnome about it
[20:48] <LaserJock> it's about working with Gnome on it
[20:48] <Kim^J> Why? So Ubuntu must wait even longer to get it?
[20:48] <LaserJock> why would it?
[20:48] <LaserJock> it won't take any longer
[20:48] <Kim^J> Do you think SuSE works WITH Gnome all the time? No, but the things gets to Gnome anyway.
[20:49] <Kim^J> s/gets/get/
[20:49] <LaserJock> yes, openSUSE does work with Gnome all the time
[20:49] <LaserJock> that's why Novell employees Gnome people
[20:49] <LaserJock> *employs
[20:50] <Kim^J> I think it goes MUCH faster to develop on own platform, test, use, extend, then make it upstream.
[20:50] <LaserJock> but why not do the *exact* same thing upstream?
[20:50] <LaserJock> it seems like a waste of effort
[20:50] <Kim^J> What?
[20:50] <Kim^J> It's being sent to Gnome anyway, the whole thing, done and ready to compile.
[20:51] <LaserJock> but why not work *with* Gnome is the point
[20:51] <LaserJock> you can get more testers, more knowledge, more help
[20:51] <Kim^J> But we are going to.
[20:51] <LaserJock> seems like a win-win situation
[20:51] <LaserJock> only after it's done
[20:52] <LaserJock> perhaps the Gnome guys have a much better way to implement it, etc.
[20:52] <LaserJock> or perhaps they've started work on something similar, who knows
[20:52] <LaserJock> but it's probably worth a look
[20:53] <Kim^J> Whatever...
[20:53] <ffm> LaserJock, The people on #gnome on GIMPnet are dead.
[20:54] <ffm> 140 ppl in chan, noone responds.
[20:54] <Kim^J> No life in ##gnome @ Freenode...
[20:54] <LaserJock> ffm: it's a weekend
[20:54] <Kim^J> LaserJock: So? Weekends = More time to IRC.
[20:54] <ffm> LaserJock, And I'm in school on weekdays.
[20:54] <LaserJock> for working people it's often a break
[20:55] <LaserJock> ffm: there's an interesting new technology called email, you may have heard of it ;-)
[20:55] <ffm> LaserJock, Bah!
[21:00] <uyann> word of advice, make sure users are adequately warned before using the entire disk for ubuntu during installation
[21:01] <ffm> uyann, You mean removing the windows partitions?
[21:01] <ffm> uyann, Or filling it to the brim?
[21:01] <uyann> I had both windows and ubuntu installed on my machine
[21:02] <uyann> I decided to install the latest ubuntu today
[21:02] <uyann> I was happily clicking on next next etc...
[21:02] <pwnguin> so in your state of clicking next, do you think you would have read and understood a warning?
[21:02] <uyann> because the default was set to use the entire disk I clicked on next
[21:02] <uyann> yes if the button was not "Next", I may have stopped
[21:03] <uyann> and if a red warning sign appeared.. absolutely
[21:03] <ffm> uyann, You'd be surprised.
[21:03] <Kim^J> uyann: Naaah, normal Windows-users would just click on that too without reading. ;)
[21:03] <uyann> I understand that Ubuntu is being made to be installed as easy as possible.. but it's really dangerous...
[21:03] <uyann> :/
[21:04] <ffm> uyann, Some people will only do so if required to do something like type "I understand all data on my disk will be lost"
[21:04] <LaserJock> uyann: do you remember if it had anything else before formating the drive?
[21:04] <uyann> yes
[21:04] <uyann> I remember the button "Advanced"
[21:04] <pwnguin> ffm: i recall a packages somewhere requiring that sort of interaction
[21:04] <uyann> and then I clicked on next and it started repartitioning
[21:04] <LaserJock> uyann: did it give you a summary of what it was going to do?
[21:05] <LaserJock> it's been a while since I used the GUI installer
[21:05] <pwnguin> uyann: maybe if we forced users to back up their data before installing
[21:05] <highvoltage> uyann: word of advice, if you boot from /any/ CD, try to be careful with just clicking on 'next' ;)
[21:05] <uyann> pwnguin, no, do not force users to backup
[21:05] <pwnguin> then there's no good solution
[21:05] <uyann> yeah, I understand it's my fault..
[21:05] <pwnguin> things can and will (in a small number of cases) go wrong
[21:06] <pwnguin> best practice is to make backups
[21:06] <pwnguin> i think the installer even says at the beginning to make a backup
[21:06] <LaserJock> I think a nice little thing with a stop sign or something that says "The entire hard drive is going to be erased" or some such
[21:06] <uyann> 4 seconds just passed when it was repartitioning
[21:06] <uyann> I believe I can still rescue the partition
[21:06] <uyann> LaserJock, I agree
[21:07] <pwnguin> LaserJock: so what if you resize the partition?
[21:08] <LaserJock> I think that's a bit more self explanatory
[21:08] <uyann> 'data may be lot by proceeding with this operation' or something to that effect
[21:08] <LaserJock> but if I remember right the installer says "Use entire hard disk" or something like that
[21:08] <uyann> yes
[21:09] <uyann> it was set to the default
[21:09] <LaserJock> "Use" does not imply "Erase"
[21:09] <hdevalence> i think the warning needs to be big, red, and scary
[21:09] <LaserJock> which is where I can see confusion
[21:09] <hdevalence> otherwise it'll be ignored
[21:10] <pwnguin> there's no situation i can imagine where "data may be lost by proceeding with this installation" SHOULDN'T be shown under this argument
[21:10] <pwnguin> repartitioning can cause data loss
[21:10] <pwnguin> reformatting the entire drive will cause data loss
[21:10] <pwnguin> the warning doesn't solve anything
[21:10] <pwnguin> backups do
[21:11] <LaserJock> no, I don't think we necessarily need a "your data may be lost"
[21:11] <LaserJock> but users need a reasonable expectation of what's gonna happen to their disk
[21:11] <LaserJock> with perhaps special attention to defaults
[21:11] <LaserJock> for the "Next" clickers ;-)
[21:12]  * uyann pour moi
[21:12] <pwnguin> whats going to happen is theres a chance for data loss, and even if warned, they'll blame anyone but the only person who could have backed it up
[21:12] <LaserJock> if you take the time to pick something other than the default then I assume you can read ;-)
[21:13] <pwnguin> heh
[21:13] <pwnguin> maybe switch next and back on that frame
[21:13] <pwnguin> so if they just click without looking, it'll be a loop
[21:13] <LaserJock> hah
[21:14] <LaserJock> I always use the Manual partitioning so I have no idea what the guided partitioning does
[21:14] <pwnguin> on a more serious note, why can't the ubuntu live cd provide at least an optional "back drive up" system?
[21:15] <LaserJock> s/can't/has nobody yet written/
[21:15] <pwnguin> i tried guided a few times, it's messy and sorta fluctuates. at one point the resize and install to a new partition guided thing was removed because it was prone to breakage
[21:16] <highvoltage> pwnguin: you can create a spec for something like that
[21:16] <ffm> I think there should be no default.
[21:16] <ffm> You should have to choose one of them.
[21:16] <highvoltage> ffm: wouldn't that, in itself be a default? :)
[21:16] <LaserJock> ffm: yeah, that would make sense actually
[21:16] <ffm> A) Wipe disk and use sensible defaults <bold> ALL DATA WILL BE LOST</bold>
[21:16] <pwnguin> highvoltage: maybe if there was an Ubuntu SoC it could have been a pick
[21:17] <LaserJock> uggg
[21:17] <LaserJock> that would mean it would never get done ;-)
[21:17] <pwnguin> colorfilter got done
[21:18] <pwnguin> its interesting though how debian never seems to get "this should be easier / safer" complaints
[21:18] <ffm> pwnguin, They don't bill themselves as designed for normal humans either.
[21:18] <LaserJock> they don't?
[21:19] <LaserJock> I'm pretty sure I've seen that kind of thing on debian-devel
 I think there should be no default. <<< I agree, a simple change that may work....
[21:19] <ffm> Why don't we bug the people in #ubuntu-installer?
[21:20] <ffm> LaserJock, Do I have to make a spec just for that?
[21:20] <LaserJock> ffm: I'm guessing not a lot of people are around in #ubuntu-installer because of the weekend
[21:20] <LaserJock> I would think just for the no-default thing that would be a bug report
[21:20] <LaserJock> safer handling of partitioning UI in general would probably be a spec
[21:21] <LaserJock> but you'd want to probably talk to Colin about it first
[21:21] <pwnguin> i still like my circular reference of doom idea ;)
[21:22] <pwnguin> if there's no default, how will our poor, mildly informed user make a choice?
[21:22] <uyann> pwnguin, hehe.. "Do you wish to go back an change a setting?" "Next"
[21:22] <uyann> s/an/and
[21:22] <pwnguin> the default is the default for a reason -- it's the most likely to work
[21:22] <uyann> it'll work, ofcourse
[21:22] <uyann> but it'll erase the entire disk
[21:22] <LaserJock> well, it's also the most damaging
[21:23] <pwnguin> (or does the most undamaging ;) )
[21:23] <LaserJock> I doubt you'd say that if you accidentally wiped a drive you cared about ;-)
[21:24] <pwnguin> if i had drives i cared about
[21:24] <pwnguin> i'd back them up
[21:25] <LaserJock> maybe
[21:25] <LaserJock> but a great majority of users don't
[21:25] <pwnguin> then we should help them
[21:25] <uyann> yup, a majority of people don't backup and don't read instructions
[21:26] <LaserJock> and it's a rather lame excuse for failure
[21:26] <uyann> I trusted Ubuntu. I'm now scarred for life. I feel like I've been raped by it.
[21:26] <pwnguin> as long as failure is possible, not pushing backups is a failure itself
[21:26] <uyann> lol
[21:26] <LaserJock> that's like a care mechanic saying "oh, you have a backup car right?" when he destroys your engine
[21:27] <LaserJock> sure, it's good to have one
[21:27] <pwnguin> when he destroys your engine he gets you a new one
[21:27] <LaserJock> but you should be able to reasonably expect that the mechanic isn't going to destroy your car
[21:27] <pwnguin> there is no new data copy unless you make it
[21:27] <pwnguin> a warning sign doesn't change your argument
[21:28] <LaserJock> sure
[21:28] <LaserJock> if we can lower the number of people who run in to problems then that's good
[21:28] <pwnguin> going back to the mechanic "i destroyed your engine; you read the sign on the door right?"
[21:28] <LaserJock> *if* there was a sign on the door then it would be my problem, yes
[21:29] <LaserJock> hence why people put signs up like that all the time
[21:29] <pwnguin> but you can't assess that risk
[21:29] <pwnguin> the mechanic can
[21:29] <LaserJock> so we (the mechanic) has a responsibility to reasonably warn users
[21:29] <LaserJock> *have
[21:30] <LaserJock> that's my point
[21:30] <pwnguin> and where possible, mititage the risk entirely
[21:30] <LaserJock> sure
[21:30] <LaserJock> but that doesn't exist right now
[21:30] <LaserJock> so it's rather mute
[21:30] <pwnguin> wow, i misspelled that pretty badly
[21:30] <pwnguin> mitigate
[21:31] <_hdevalence> has it been considered to use something like debtorrent for distribution upgrades?
[21:32] <LaserJock> I believe it has been considered to use debtorrent period, I believe
[21:33] <_hdevalence> I have to say, it's REALLY annoying when the servers are swamped with everyone trying to upgrade
[21:33] <LaserJock> uh, yeah
[21:33] <_hdevalence> and I'm sure that server bandwith for that many upgraders is not cheap
[21:34] <LaserJock> sure
[21:34] <pwnguin> well, everyone and their mom mirrors
[21:34] <_hdevalence> well, using my local mirror, I was getting speeds from 30-60 KB/s
[21:34] <LaserJock> I haven't really noticed all that much of a slow down this time
[21:35] <LaserJock> it was waay worse in the past I think
[21:35] <LaserJock> but yeah, it would help to have things like debtorrent, and package diffs
[21:36] <pwnguin> i wonder about package diffs
[21:36] <pwnguin> seems like you'd need to keep several of those around
[21:37] <_hdevalence> I think debtorrent would be good enough. package diffs are more complicated
[21:37] <LaserJock> pwnguin: it's a lot of load on the servers I think any way you do it