[00:56] <CIA-33> ubiquity: cjwatson * r3461 ubiquity/debian/changelog: releasing version 1.99.21
[08:43] <CIA-33> ubiquity-slideshow-ubuntu: evand * r145 html/ (Makefile debian/changelog debian/control):
[08:43] <CIA-33> ubiquity-slideshow-ubuntu: The buildds do not install recommends by default, so add a build
[08:43] <CIA-33> ubiquity-slideshow-ubuntu: dependency on python-utidylib (LP: #430719).
[09:45] <davmor2> evand: I got an av/fw package called c.o.m.o.d.o everytime a .py file is created in %temp% it gets flaged by the av is this a bug or just an over protective av?
[09:45] <evand> sounds like a buggy virus scanner.
[09:45] <evand> what product?
[09:46] <davmor2> it might actually be the fw part of it thinking about the message, I'll grab you the link
[09:48] <davmor2> evand: http://www.comodo.com/home/download/download.php?prod=firewall  I went for the fw/av combo
[09:52] <evand> davmor2: http://virusscan.jotti.org/en/scanresult/84d58605e4881b4d6ff54579e3ccb03b7196bb82 and http://www.virustotal.com/analisis/6ebe1e5ad3f813f28f32d2d6763ebd3cc5a975b9a2c3e3158d1595a78656b5c1-1253177472
[09:53] <evand> I'd say your virus scanner is broken.
[09:57] <davmor2> evand: I'm assuming that because the fw and av are linked that one is triggering the other.  eg .py wants access to system files so the fw flags it at the point the av say this maybe a threat
[09:58] <evand> no idea, but I am certain that wubi is virus free and that most antivirus programs do not flag it
[10:02] <davmor2> cool I'll just take it as being comodo then :)
[10:07] <davmor2> cjwatson_ evand: Wubi install now works again but on reboot system drops into grub cli.  Could not find bootloader configuration
[10:19] <davmor2> also cjwatson_ I don't think it is quiet supressing gdm startup on exit brilliantly desktop background flashed up 3-4 times before finally dying
[10:31] <evand> davmor2: can you please try http://people.canonical.com/~evand/wubi/karmic/wubi-r153.exe and see if it gets you any further?
[10:32] <davmor2> evand: wilko
[10:33] <davmor2> I think this is the old grub 2 issue though as the system is installed but I'll have a quick look anyhow
[10:49] <davmor2> evand: cross your fingers
[11:00] <davmor2> evand: it's still trying to start the desktop after the install.   Grub has come up after all I get in the menu list is the memory test listings
[11:22] <evand1> success!  I've finally, finally got javascript-based reflection working for the firefox slide
[11:37] <cjwatson> davmor2: right, so my lupin script didn't work - guess I'd better have a look
[11:38] <cjwatson> evand1: can I use that wubi.exe with a CD, or will it download from the network?
[11:38] <evand1> cjwatson: you can use it with a CD
[11:38] <cjwatson> good dgood
[11:38] <cjwatson> my ISP is going to be unhappy enough with me this month already
[11:39] <evand1> hahaha
[11:45] <davmor2> cjwatson: I dropped it on the cd
[11:46] <davmor2> cjwatson:  no idea why that would be the case
[11:47] <davmor2> cjwatson: wubi-r153.exe is the one where I only get the memory test show up in grub
[11:48] <davmor2> cjwatson: for the wubi that is really on the cd https://bugs.launchpad.net/bugs/431285 is the bug I opened.
[11:54] <davmor2> cjwatson: also the screen flashes the desktop image 4-6 times before finally rebooting.  Flashes between the desktop and cli.
[11:56] <cjwatson> you will need a current wubi build, certainly; I don't know how old the one on the CD is but there have been recent relevant commits
[11:56] <cjwatson> lupin ships the script that hooks into grub2 menu generation
[12:04] <davmor2> cjwatson: r153 had the same issue with the flashing screen on exit
[12:05] <cjwatson> flashing screen> not my problem :)
[12:06] <cjwatson> that's either a desktop thing or some problem with upstart
[12:06] <cjwatson> probably the latter
[12:06] <cjwatson> I doubt it's particularly related to wubi although of course it's possible that wubi happens to tickle something
[12:08] <davmor2> cjwatson: I'm going to have a look and see if install ubuntu from the menu does the same thing before I bug it, but it isn't pretty.
[12:09] <cjwatson> oh, I'm sure - it's just not something I know about or that's really an installer thing
[12:09] <cjwatson> at least not as far as I know
[12:48] <lool> cjwatson: Just pushed http://paste.ubuntu.com/272740/ does this look sane to you?
[12:48] <lool> Also is it ok to upload this now?  I dont care to get it in A6 or anything but would like to forget about it
[12:49] <cjwatson> lool: seems ok, but please wait until after a6
[12:49] <cjwatson> if it's in bzr, it won't be forgotten
[12:49] <lool> It is
[12:49] <lool> I closed the bug in a subsequent commit
[12:49] <lool> Sorry, didn't setup CIA
[12:49] <cjwatson> I've 'bzr up'ed, so it'll show up in my unreleased-packages script
[12:49] <lool> Oh cool
[12:50] <lool> cjwatson: I wish you tell me about your setup/workflow for this unreleased-packages stuff next time we meet
[12:50] <cjwatson> meh, it's just a hack :)
[12:50] <lool> I'm sure you're using plenty of nice scripts and ideas and would like to borrow  :-)
[12:50] <cjwatson> http://paste.ubuntu.com/272744/
[12:51] <lool> I'd love to sit for a couple of hours watching you work
[12:51] <cjwatson> we could probably all do with more of that with each other ...
[12:51] <cjwatson> I'm not that efficient, though, I flit around a lot
[12:52] <lool> cjwatson: what's your toplevel between src/debian and the source package name?
[12:52] <lool> It's ~/src/<distro>/<something>/<source package>/debian/changelog
[12:54] <cjwatson> actually it's ~/src/<distro>/<source package>/<some directory name, usually either source or source-upstream-version>/debian/changelog
[12:54] <lool> aha
[12:55] <cjwatson> I have the extra level of nesting there because dpkg-buildpackage dumps stuff in .. so it keeps things reasonably tidy
[12:55] <lool> Ack; I do the same with my scratch shell function which creates ~/scratch/$source/; cds to that and apt-get source $source
[12:56] <lool> (cwd is preserved since it's a shell function)
[12:56] <cjwatson> mm, I've always been meaning to write that but can never be bothered, I just do it by hand
[12:56] <lool> http://paste.ubuntu.com/272747/
[12:56] <lool> (zsh)
[12:57] <lool> I put that in ~/.zshfns and have fpath=(~/.zshfns $fpath) in .zshenv andautoload -U scratch in .zshrc
[12:57] <cjwatson> thanks, will bashify and snarf
[13:19] <lool> When I run the alternate installer over serial console, it shows up in black and white; is there a way to tell it about my TERM being screen in xterm?
[13:19] <cjwatson> I doubt it has the right terminfo bits for that
[13:20] <lool> Ah ok; it only has linux and the serial console one
[13:20] <lool> vt102
[13:20] <lool> I see a minor issue in the black and white mode: first letter in choice lists are white on white; I guess they are supposed to be bold which maps to white
[13:20] <lool> Is this worth reporting?
[13:23] <lool> I guess nobody cares about b&w / vt102
[13:23] <lool> cjwatson: Actually on some boards one can connect to d-i over SSH; I guess that means it should support xterm if people are connecting from xterm to d-i over ssh?
[13:29] <cjwatson> maybe, I think we only ship ansi bterm dumb linux vt102 right now
[13:29] <lool> cjwatson: Mind adding debian-installer-utils to your list of packages to push?  We didn't properly match the new name of the mmc /sys ID_PATH
[13:30] <cjwatson> di-utils-terminfo is responsible
[13:30] <cjwatson> sure
[13:30] <lool> cjwatson: Can I change the TERM after d-i started?
[13:30] <lool> perhaps in preseed
[13:30] <cjwatson> I doubt it
[13:30] <cjwatson> well, the one you connect to over ssh is a new instance
[13:30] <cjwatson> so you could set it for that I suppose
[13:30] <lool> ah
[13:30] <cjwatson> but I don't know if there's a canned way; I don't expect so
[13:31] <ogra> cjwatson, if i make changes to cdrom-detect.postinst on disk in d-i from the shell, they dont affect d-i at all it seems ... is that because the stuff is already loaded ?
[13:32]  * ogra just tried lools fix in a running environment 
[13:32] <lool> ogra: Jan  1 00:16:15 kernel: FAT: codepage cp437 not found
[13:32] <lool> Jan  1 00:16:15 anna[10729]: mount: mounting /dev/mmcblk0p2 on /media failed: In
[13:32] <lool> valid argument
[13:32] <lool> ogra: It's a kernel issue
[13:32] <ogra> gah
[13:32] <lool> probably these should be in the fs udeb or something
[13:32] <ogra> if we had a fs udeb :P
[13:33] <ogra> i think we only have the het one which contains sunrpc and nothing more
[13:33] <ogra> *net
[13:33] <lool> We have much more in the alternate
[13:33] <cjwatson> ogra: they do affect d-i, although only if cdrom-detect.postinst is not already running
[13:33] <lool> /lib/modules/2.6.31-100-imx51/kernel/drivers/net/wireless/airo.ko
[13:33] <lool> /lib/modules/2.6.31-100-imx51/kernel/drivers/net/wireless/atmel.ko
[13:33] <lool> /lib/modules/2.6.31-100-imx51/kernel/drivers/net/wireless/orinoco/orinoco.ko
[13:33] <lool> etc.
[13:33] <cjwatson> ogra: changes to templates files require more work
[13:33] <lool> /lib/modules/2.6.31-100-imx51/kernel/fs/isofs/isofs.ko
[13:33] <lool> /lib/modules/2.6.31-100-imx51/kernel/fs/configfs/configfs.ko
[13:33] <lool> ogra: ^
[13:34] <ogra> well, i only added the one line to .postinst, but lool has a valid point
[13:34] <lool> fat-modules
[13:34] <ogra> lool, weird, why arent these in netboot then ?
[13:35] <lool> because netboot doesn't need them
[13:35] <lool> it's supposed to retrieve udebs over the net
[13:36] <ogra> well, not the NIC drivers i would guess :)
[13:37] <lool> We do have a nic-usb-modules-2.6.31-100-imx51-di_2.6.31-100.7_armel.udeb but it doesn't have asix
[13:37] <lool> and it's not in the image
[13:38] <ogra> i have an atmel and musb adapter here ... but only sunrpc in in the netboot image
[13:39] <lool> ogra: Ah the atmel one is in udeb
[13:39] <ogra> yep
[13:39] <ogra> but the udeb isnt in netinst
[13:39] <lool> nic-modules-2.6.31-100-imx51-di_2.6.31-100.7_armel.udeb
[13:40] <ogra> needs to be added i suppose
[13:40] <ogra> and if there is a separate one for firmware, that too
[13:40] <ogra> (i think i remember there is one ... from nslu2)
[13:45] <lool> Hmm
[13:46] <lool> we only have udebs for dove not for dove-z0
[13:46] <lool> Oh wait wrong source package
[13:46] <ogra> heh
[13:46] <ogra> we urgently need to file removal requests
[13:46] <lool> Right we have them
[13:46] <ogra> there is so much mess in the archive
[13:47] <ogra> all these obsolete binaries from going back and forward with the kernel names
[14:05] <lool> ogra: Do we need nic in dove or is it builtin?
[14:05] <lool> I think it is but am not sure
[14:06] <ogra> it is
[14:06] <ogra> i never isntalled any modules and still use tftpboot
[14:06] <ogra> works with all kernels up to the recent one
[14:06] <lool> Does it make sense to allow using USB network adapters if we have builtin support for the builting ethernet?  I think not
[14:07] <lool> ogra: So what's the driver for your chip?  You said atmel.ko?
[14:07] <lool> ogra: It's odd because it shows up in ./lib/modules/2.6.31-100-imx51/kernel/drivers/net/wireless/atmel.ko
[14:07] <lool> instead of ./lib/modules/2.6.31-100-imx51/kernel/drivers/net/usb/
[14:07]  * ogra tries to find his rarely used usb wlan key
[14:08] <lool> I can see the contention between putting stuff in /wireless or /usb
[14:08] <lool> It makes it hard to decide which udebs to include
[14:09] <ogra> well, wireless derfinately needs firmware
[14:09] <ogra> some normal usb NICs too
[14:09] <ogra> but not all of them
[14:09] <ogra> sorry, i cant find my atmel atm
[14:10] <ogra> i have another zdrwXXX too somewhere
[14:10] <lool> ogra: What's the question mark after nic-firmware in d-i?
[14:10] <lool> build/pkg-lists/netboot/armel/ixp4xx.cfg:nic-firmware ?
[14:11] <cjwatson> it means that it's optional
[14:11] <ogra> right
[14:11] <cjwatson> I think this is documented in build/README, or else in the kernel-wedge documentation
[14:13] <lool> cjwatson: Thanks build/README has it
[14:14] <lool> I guess we want to keep the firmware optional to leave the ability to build free images easily
[14:14] <ogra> well, we used it by default in ixp4xx
[14:14] <cjwatson> yes
[14:14] <ogra> simply because its useless without
[14:14] <cjwatson> it doesn't matter though, if the package is present it makes no difference
[14:15] <ogra> yeah
[14:19] <CIA-33> usb-creator: evand * r204 trunk/ (debian/changelog usbcreator/misc.py):
[14:19] <CIA-33> usb-creator: Move logging back to the home directory, now that usb-creator is run as a
[14:19] <CIA-33> usb-creator: regular user (LP: #431266).
[14:22] <lool> Does someone know which modules are required for USB keyboard support?
[14:22] <lool> (Assuming USB host controller driver is already there)
[14:22] <lool> usbhid.ko?
[14:23] <lool> I'll give that a go
[14:23] <ogra> yep
[14:24] <ogra> but there is more i think
[14:25] <ogra> lool, hmm, checking my babbage, usbhid should actually be enough
[14:30] <TommyT> I am here to provide more data for bug #430333 (but I am a newb on irc).
[14:33] <lool> http://paste.ubuntu.com/272839/ is what I have for now for imx51
[14:33] <cjwatson> TommyT: will be with you in a minute (also, don't tell people you're a newbie and they probably won't notice, that's my advice)
[14:34] <ogra> lool, sweet
[14:34] <cjwatson> lool: haven't checked in detail but fine by me
[14:37] <cjwatson> TommyT: are you sitting at the rescue:grub> prompt right now?
[14:44] <TommyT> oops sorry I missed your post. I will switch irc clients and be right bac
[14:45] <TommyT> I'm in front of the rescue:grub> prompt now
[14:46] <cjwatson> TommyT: ok, can you type: ls (hd0,1)
[14:46] <cjwatson> TommyT: then: ls (hd0,1)/
[14:47] <cjwatson> (note trailing slash)
[14:47] <cjwatson> and let me know the output
[14:47] <TommyT> yes ... it has lost+found/ var/ etc/ ....
[14:48] <TommyT> including vmlinuz & initrd.img
[14:49] <cjwatson> ok, now: ls (hd0,1)/boot/
[14:50] <TommyT> grub/ System.map-2.6.31-10-generic abi-2.6.31-10-generic ... config ... memtest ... vmcore... etc.
[14:50] <cjwatson> if possible, give me everything
[14:50] <cjwatson> try not to abbreviate
[14:51] <TommyT> ok add to that config-2.6.31-10-generic memetest86+.bin vmcoreinfo-2.6.31-10-generic vmlinuz-2.6.31-10-generic initrd.img-2.6.31-10-generic
[14:52] <cjwatson> hmm, ok
[14:52] <cjwatson> let me shortcut a bit
[14:52] <cjwatson> TommyT: insmod normal
[14:53] <cjwatson> what does that do?
[14:53] <TommyT> back to the prompt
[14:53] <cjwatson> normal
[14:53] <cjwatson> (as in, type 'normal')
[14:53] <TommyT> unknown command 'normal'
[14:53] <cjwatson> echo ${root}
[14:54] <TommyT> hd0,1
[14:54] <cjwatson> hmm!
[14:54] <cjwatson> TommyT: ls (hd0,1)/boot/grub/normal.mod
[14:54] <cjwatson> what the heck is wrong here
[14:54] <TommyT> I haven't typed your command yet but if it's like before the damage will be sporadic
[14:55] <TommyT> when I typed the command it came back
[14:55] <TommyT> normal.mod
[14:55] <cjwatson> right, the thing is, the 'normal' module is the thing that handles the normal menu path
[14:55] <cjwatson> and it seems like loading that module isn't actually providing the 'normal' command
[14:56] <cjwatson> which is, to put it mildly, thoroughly weir
[14:56] <cjwatson> d
[14:56] <TommyT> If it'
[14:56] <cjwatson> especially when there's no error reported
[14:56] <TommyT> if it is like before, the nodes etc are trashed.
[14:56] <cjwatson> yeah, but no error ...
[14:56] <cjwatson> so what I want to find out is exactly *how* they're trashed
[14:56] <cjwatson> bear with me a bit
[14:58] <cjwatson> TommyT: BTW, the manual partitioning problems you had were almost certainly bug 430724, fixed in today's images
[14:59] <cjwatson> but I don't think that's *directly* relevant to this grub problem
[14:59] <TommyT> cool, thanks.
[14:59] <cjwatson> so, I think what I probably want to do, if possible, is to have you boot into a karmic live CD and then run a few iterations of a grub filesystem testing utility
[15:00] <cjwatson> how comfortable are you with running a program I supply over the Internet?
[15:00] <cjwatson> or, well, if possible, the easiest thing would actually be for me to have SSH access to the machine - but I'd entirely understand if you didn't want to give that to some stranger
[15:01] <TommyT> I will do whichever (though I'm not sure how easily I can open a path to the machine through my firewall)... but remember this is a netbook and will take awhile. Should I get a newer iso to boot from or will this old one do?
[15:02] <cjwatson> I think the old one will be OK - I'm actually going to need to build custom grub-fstest programs with extra debugging
[15:03] <cjwatson> though we can start out with the stock upstream one
[15:03] <cjwatson> while I can't really tell from the rescue shell, it sounds as if it thinks normal.mod is a zero-sized file or something
[15:03] <cjwatson> hmm, though actually that should generate a warning
[15:04] <TommyT> I'm booting from the image I have (from Monday). It will take it about 5 mins
[15:06] <TommyT> Actually, should I have booted into a shell?
[15:06] <TommyT> Once the GUI comes up I will switch to a VT if necessary.
[15:07] <cjwatson> doesn't matter
[15:07] <cjwatson> a terminal will be fine
[15:08] <TommyT> OK it came up faster this time. I'm at Terminal
[15:08] <cjwatson> ideally, have networking configured as well
[15:08] <cjwatson> at least outgoing, so that you can dump output places
[15:08] <TommyT> yes it found my lan
[15:08] <cjwatson> first off, grab http://people.canonical.com/~cjwatson/ubuntu-paste - you can use that program to send stuff to paste.ubuntu.com easily
[15:09] <cjwatson> you'll need to chmod +x it of course
[15:11] <TommyT> ok. done
[15:14] <cjwatson> TommyT: ok, now download http://people.canonical.com/~cjwatson/tmp/tommyt/grub-fstest
[15:14] <cjwatson> TommyT: once you've done that, first, a quick sanity check:  sudo ./grub-fstest /dev/sda1 ls /
[15:15] <TommyT> file not found... checking spelling...
[15:15] <TommyT> OH i missed tommyt
[15:15] <TommyT> sorry
[15:15] <TommyT> is capitalization right? still not found
[15:16] <TommyT> oh I got it now.
[15:16] <cjwatson> definitely exists
[15:17] <TommyT> sorry I keep reaching for the wrong keyboard
[15:18] <TommyT> ok I ran your "sanity check" and there was no output
[15:19] <TommyT> do I need to mount /dev/sda1
[15:19] <cjwatson> no, you don't
[15:19] <cjwatson> does /dev/sda1 actually exist?
[15:19] <TommyT> it shows when I ls /dev/sd*
[15:20] <cjwatson> ok:  sudo ./grub-fstest /dev/sda1 ls -l / 2>&1 | ./ubuntu-paste
[15:20] <cjwatson> that should print a URL
[15:23] <TommyT> http://paste.ubuntu.com/272872/
[15:25] <cjwatson> hmm, so that's fine
[15:25] <cjwatson> sudo ./grub-fstest /dev/sda1 ls -l /boot/grub/normal.mod 2>&1
[15:27] <TommyT> 34656  normal.mod
[15:27] <cjwatson> huh, that's fine, I wasn't really expecting that
[15:28] <cjwatson> sudo ./grub-fstest /dev/sda5 cmp /boot/grub/normal.mod /boot/grub/normal.mod
[15:28] <cjwatson> err, replace /dev/sda5 with /dev/sda1 there
[15:29] <TommyT> error: open error
[15:29] <TommyT> is there a grub on the iso?
[15:30] <TommyT> there is no grub on the iso
[15:30] <cjwatson> uh, what do you mean?
[15:30] <cjwatson> ok, let's just confirm your open error
[15:30] <TommyT> the USB device I boot from has a grub directory but nothing in it
[15:30] <cjwatson> sudo ./grub-fstest -d all /dev/sda1 cmp /boot/grub/normal.mod /boot/grub/normal.mod 2>&1 | ./ubuntu-paste
[15:30] <cjwatson> yeah, that's expected
[15:31] <cjwatson> CDs use isolinux to boot, at present
[15:33] <TommyT> http://paste.ubuntu.com/272881/
[15:34] <ogra> cjwatson, lool, http://paste.ubuntu.com/272882/ do you think thats ok to create an SD card netinstall image for imx51 ? (i added the needed functionallity to redboot-install)
[15:36] <cjwatson> TommyT: this output is actually very odd. Could you run the same command again, just so that I can confirm it's consistent?
[15:37] <cjwatson> TommyT: then also run: sudo ./grub-fstest /dev/sda1 ls -l /boot/grub/normal.mod 2>&1 | ./ubuntu-paste
[15:37] <cjwatson> also double-check typing
[15:38] <TommyT> http://paste.ubuntu.com/272886/ verified typing
[15:39] <lool> ogra: You probably want to add dependencies in your target
[15:39] <lool> ogra: I dont get the purpose of create_blank_image
[15:40] <ogra> lool, i just stole from ixp4xx ... i cam make it of=$@ indeed
[15:40] <ogra> and move it down
[15:40] <lool> ogra: No I mean the target
[15:40] <lool> the temp file is ok
[15:40] <TommyT> second command: http://paste.ubuntu.com/272888/
[15:41] <ogra> lool, i dont get it ... it has to be called before $(SOME_DEST)/$(EXTRANAME)di-imx51.bin so its in the traget list before it
[15:41] <cjwatson> oh, drat
[15:41] <cjwatson> TommyT: sorry, I forgot a bit:  sudo ./grub-fstest -d all /dev/sda1 ls -l /boot/grub/normal.mod 2>&1 | ./ubuntu-paste
[15:41] <cjwatson> need the debug information there
[15:41] <lool> ogra: Why dont you just drop create_blank_image and create the zero file before calling redboot-install?
[15:42] <ogra> lool, thats what i just offered a min ago :)
[15:42] <lool> ogra: You said of=$@ and I said not to
[15:42] <ogra> ah, you want me to keep the temfile
[15:43] <ogra> *temp
[15:43] <lool> Yes
[15:43] <ogra> http://paste.ubuntu.com/272890/
[15:43] <ogra> like that ?
[15:43] <lool> Otherwise if one kills the build at an inappropriate time, a blank image might be generated on second run
[15:43] <TommyT> cjwatson: http://paste.ubuntu.com/272889/
[15:43] <lool> ogra: You dropped the deps...
[15:44] <lool> ogra: You need a bdep on redboot-tools too
[15:44] <ogra> whoops http://paste.ubuntu.com/272891/
[15:44] <ogra> indeed
[15:44] <ogra> i didnt want to have the dep change in the paste atm
[15:44] <ogra> the deps line is insanely long :)
[15:45] <lool> That looks ok
[15:45] <ogra> cool
[15:45] <lool> Dunno about the name
[15:45] <ogra> i'll add the dep and commit if nobody objects
[15:46] <ogra> which name ? the filename or the description ?
[15:46] <lool> ogra: You might miss a mkdir though
[15:47] <lool> ogra: have a look at orion5x/netboot.cfg target mv2120:
[15:47] <lool> it creates a netboot.img
[15:47] <ogra> a mkdir ?
[15:47] <lool> ogra: you definitely want a babbage subdir there, perhaps even babage-2
[15:47] <ogra> $(TEMP) surely exists
[15:47] <lool> ogra: /imx51 is for the subarh independent images (like the imx51 kernel applying to e.g. lange boards)
[15:48]  * ogra looks at orion
[15:48] <lool> /imx51/babbage would be appropriate for a babbage specific file
[15:48] <ogra> ok
[15:48] <lool> ogra: if you're brave create one subdir + netboot.img per redboot.bin
[15:49] <ogra> that needs a) further redboot-install changes and b) verification that the kernel even works on B1
[15:49] <lool> ogra: sounds like an afternoon of fun!
[15:50] <ogra> wow, the orion stuff looks a lot saner that ixp4xx
[15:50] <lool> Yeah
[15:50]  * ogra rewrites i'll keep the B1 stuff for that extra afternoon of fun when i need it :P
[15:51] <ogra> especially since it looks like the karmic kernel will drop B1 support
[15:53] <cjwatson> TommyT: this is really, really confusing me. I can keep going if you like with progressively bigger hammers, but would ssh access be at all possible?
[15:54] <TommyT> I will try. I presume I have to install it so I'm looking at that now, then I will see if I can open the firewall
[15:55] <cjwatson> you don't have to install
[15:55] <cjwatson> actually, it's better if you don't
[15:55] <TommyT> oops
[15:55] <cjwatson> well, you have to install the openssh-server package - you don't have to install the operating system, which is how I first read your comment :)
[15:55] <TommyT> that's what I meant yes openssh
[15:55] <ogra> lool, so how about http://paste.ubuntu.com/272897/ ?
[15:57] <TommyT> what username/pass works on the live iso?
[15:57] <davmor2> TommyT: ubuntu/blank
[15:57] <davmor2> as in no pass just ubuntu
[15:58] <TommyT> OK sorry I should have said what user/pass works to ssh into a live iso that sshd has been installed on, or do I need to create an account?
[15:58] <cjwatson> yes, though to make that work with ssh you have to either (a) edit /etc/ssh/sshd_config and add 'PermitEmptyPasswords yes', then 'sudo /etc/init.d/ssh restart' or (b) 'passwd ubuntu' and make it something else
[15:58] <TommyT> right that answers it
[15:58] <cjwatson> TommyT: same answer as davmor2 gave, but with my modifications :)
[15:59] <cjwatson> I suggest not quoting the login IP address in a public channel though, nor the password if you set it ...
[16:02] <lool> ogra: Do you mind if I'm picky?   I would name the target babbage (single target for all babbage boards) the temp dir should be $(TEMP)/babbage/ and the public publish dir babbage-TO2
[16:02] <lool> or babbage-2.x
[16:02] <ogra> i'll add the dash
[16:02] <ogra> but $TEMP is from d-i i think
[16:03] <lool> ogra: You dont want to use a generic name like you do
[16:03] <ogra> why should i add an extra babbage there
[16:03] <lool> ogra: e.g. kurobox uses $(TEMP)/kuroboxpro/kernel.uboot
[16:03] <lool> ogra: an extra babbage?
[16:03] <ogra> well, i use what d-i generates
[16:04] <ogra> so the only place using $(TEMP)/babbage/ would make sense is for teh temp_image.img
[16:04] <lool> ogra: Err yes that's entirely the point
[16:04] <ogra> oh, ok
[16:04] <lool> Or instead of temp_image.img use a better name
[16:04] <ogra> heh
[16:04] <lool> e.g. babbage-TO2-sd.img or something
[16:04] <ogra> babbage-2.img :)
[16:05] <lool> ogra: Do we need all the console=?
[16:05] <lool> I think the default of console=tty0 should just work
[16:06] <ogra> http://paste.ubuntu.com/272903/
[16:06] <lool> 20 MB seems really large
[16:06] <ogra> i'D like to keep a serial one for debug, but i can drop it
[16:06] <lool> Is that what we use?
[16:06] <ogra> its the default and would need serious redboot-install hacking
[16:06] <ogra> which i'd like to avoid
[16:06] <lool> FIS_SIZE="$((16 * 1024 * 1024))"
[16:07] <lool> That's what I used in debian-cd
[16:07] <lool> Bah are you saying that redboot-install just has some extra padding to avoid computing the actual size?
[16:07] <ogra> FIS_SIZE="$((16 * $MIB))"
[16:07] <ogra> same in redboot-install
[16:08] <ogra> hmm, why did it complain when i used less than 20
[16:08] <lool> I think you forget the partition table
[16:08] <lool> and vfat
[16:08] <lool> Actually strike the comment on the partition table
[16:08] <lool> ogra: Anyway exactly 16 MB should be exactly what you want  :-)
[16:09] <ogra> Error: The location 16777216B is outside of the device /var/build/images/netboot/netboot.img
[16:10] <lool> ogra: That's because you need the fix in debian-cd
[16:10] <ogra> 17 works
[16:11] <ogra> oh, and i just see i need to port it to blkid
[16:11] <ogra> still uses vol_id
[16:11] <ogra> lool, are you fine with 17 for now ?
[16:12] <lool> good lord no
[16:12] <ogra> sigh
[16:12] <ogra> i really didnt plan to invest that much time into it
[16:12] <TommyT> cjwatson: sorry I haven't ever tried going into the firewall and this old version of DD-WRT may not support secure forwarding
[16:13] <ogra> i'll turn it to 16 once i merged your debian-cd fixes
[16:14] <ogra> and in karmic+1 we can add a special mode to it that only creates a fis partition for the exact size if its run in that special setup where we want a netinst
[16:15]  * ogra wonders if lool fell over 
[16:16] <lool> I'm still here
[16:16] <lool> And I really don't see why you prefer hiding bugs with large padding
[16:16] <ogra> bugs ?
[16:16] <lool> Well yes
[16:17] <lool> Why wouldn't it work at 16 MB otherwise?
[16:17] <ogra> its not the purpose of redboot-install to generate netboot images
[16:17] <ogra> it *should* work with fis_size+kernel_size+inird_size ... but it doesnt
[16:17] <lool> Uh no
[16:17] <lool> fis_size
[16:17] <lool> the kernel and the initrd are in the FIS
[16:18] <lool> ogra: So if it doesn't, aint it a bug?
[16:18] <ogra> err, redboot_size i mean
[16:18] <lool> No
[16:18] <ogra> its a whishlist item, really
[16:18] <lool> redboot is in the FIS too
[16:18] <lool> God
[16:18] <ogra> redboot-install is for creating upgradeable SD images, remember ?
[16:18] <ogra> thats why we picked a big enough hardcoded fis size
[16:18] <lool> I'm sure you'd like it if GRUB wrote only one extra byte of it's MBR
[16:22] <ogra> i'll just leave it alone then until redboot-install grew multiple heads and just write the wikipage to get kernel and initrd via http ...
[16:22] <lool> Multiple heads... for bogus arithmetics...
[16:22] <ogra> for multiple purposes which it wasnt intended to
[16:23] <lool> Tss
[16:23] <ogra> its great if it can do that at some point but it doesnt yet and i dont want to invest a day to rewrite it
[16:24] <cjwatson> TommyT: OK, I'm beginning to get stuck here - I'm going to have to add more debugging to the code and try again
[16:25] <cjwatson> TommyT: one last thing with the existing code, can you try:  sudo ./grub-fstest -d all /dev/sda1 crc /boot/grub/normal.mod
[16:25] <cjwatson> actually
[16:25] <cjwatson> sudo ./grub-fstest -d all /dev/sda1 crc /boot/grub/normal.mod 2>&1 | ./ubuntu-paste
[16:35] <TommyT> cjwatson: http://paste.ubuntu.com/272922/
[16:36] <lool> ogra: surprize, I just applied my one liner and 16 MB passes...
[16:37]  * ogra bzr pulls
[16:37] <lool> ogra: Now that actually let me to see some things which need to be fixed in redboot-install: set -e, vol_id (I just workedaround it not computing UUID); and there's a warning in a fis usage at the end
[16:38] <ogra> vol_id is bogus
[16:38] <ogra> i just ripped it out
[16:38] <lool> http://paste.ubuntu.com/272927/
[16:38] <ogra> we dont have the second partition at all in redboot-install
[16:39] <lool> ogra: What was it used for?
[16:39] <cjwatson> TommyT: ah, that's slightly more informative. How big is this disk again?
[16:39] <TommyT> it's 32gig
[16:39] <lool> ogra: Keep in mind that we want to use it in debian-cd...
[16:39] <ogra> lool, originally it came from writing the whole image
[16:39] <ogra> yes, you can call blkid outside of redboot-tools
[16:40] <ogra> its only to pass UUID= on the cmdline
[16:40] <CIA-33> usb-creator: evand * r205 trunk/ (3 files in 2 dirs):
[16:40] <CIA-33> usb-creator: Use GIO instead of gnomevfs. Only lookup GNOME device names and
[16:40] <CIA-33> usb-creator: icons as needed.
[16:40] <ogra> for the default cmdline
[16:40] <lool> ogra: I personally wish redboot-install wouldn't hardcode the pathnames to the .bins as we want to be able to run it against multiple boards
[16:40] <cjwatson> TommyT: ok, so it's trying to read something from an offset 68GB in, which isn't going to work
[16:41] <ogra> lool, yes, thats a bug
[16:41] <cjwatson> TommyT: sudo stat /boot/grub/normal.mod | ./ubuntu-paste
[16:41] <cjwatson> err
[16:41] <lool> ogra: so the one liner and off by one is the following:
[16:41] <cjwatson> sorry, that won't work
[16:41] <lool> -parted -s "$DEV" mkpart primary fat32 "512B" "${FIS_SIZE}B"
[16:41] <lool> +parted -s "$DEV" mkpart primary fat32 "512B" "$(($FIS_SIZE - 1))B"
[16:41] <lool> ogra: but please dont merge it
[16:41] <ogra> lool, but moot until we have multiple bins we can actually use :)
[16:41] <lool> merge the dbeian-cd changes instead
[16:41] <cjwatson> TommyT: right, I'll prepare a new grub-fstest
[16:41] <ogra> thats what i told you i will do on the weekend
[16:41] <TommyT> cjwatson: ok
[16:43] <lool> ogra: I know but you expected me to push it and I wont because I prefer the debian-cd to go in, not this one liner
[16:43] <ogra> ok
[16:45] <CIA-33> usb-creator: evand * r206 trunk/ (debian/changelog usbcreator/backends/devicekit/backend.py):
[16:45] <CIA-33> usb-creator: Remove the device in the DeviceKit-disks backend when it's removed
[16:45] <CIA-33> usb-creator: from the system.
[17:00] <cjwatson> TommyT: ok, I've uploaded a new grub-fstest to the same location, http://people.canonical.com/~cjwatson/tmp/tommyt/grub-fstest
[17:00] <cjwatson> TommyT: with that, could you please try this again: sudo ./grub-fstest -d all /dev/sda1 crc /boot/grub/normal.mod 2>&1 | ./ubuntu-paste
[17:01] <cjwatson> TommyT: (I've just been called to dinner, but will see responses when I get back)
[17:01] <TommyT> cjwatson: lunch soon here too...
[17:08] <TommyT> cjwatson: unexpected response: [('date', 'Thu, 17 Sep 2009 16:04:29 GMT'), ('transfer-encoding', 'chunked'), ('connection', 'close'), ('content-type', 'text/html; charset-utf-8'), ('server', 'Apache/2.2.8 (Ubuntu) mod_python/3/3/1 Python/2.5.2 mod_ssl/2.2.8 OpenSSL/0.9.8g mod_perl/2.0.3 Perl/v5.8.8')]
[17:09] <TommyT> cjwatson: I'm also getting warnings about a full disk. df shows /dev/loop0 mounted at /rofs is at 100%
[17:10] <TommyT> cjwatson: nevermind... it's complaining about something else. I think.
[17:11] <CIA-33> ubiquity-slideshow-ubuntu: evand * r146 html/ (221 files in 12 dirs): Updated translations from Launchpad.
[17:16] <CIA-33> ubiquity-slideshow-ubuntu: evand * r147 html/debian/changelog: Add Dylan's changes to debian/changelog
[17:22] <CIA-33> ubiquity-slideshow-ubuntu: evand * r148 html/debian/changelog: releasing version 6
[17:41] <CIA-33> usb-creator: evand * r207 trunk/ (debian/changelog usbcreator/frontends/gtk/frontend.py):
[17:41] <CIA-33> usb-creator: Only set the non-size columns to expand to fill available space in the GTK+
[17:41] <CIA-33> usb-creator: frontend. Set a minimum width of 75px for all columns.
[18:28] <cjwatson> TommyT: that should be a transient error - try again?
[18:38] <cjwatson> TommyT: if it still doesn't work, you may need to reboot the live CD, download ubuntu-paste and grub-fstest again, and run that command again - there shouldn't be more setup required than that
[18:40] <cjwatson> TommyT: BTW, I really appreciate your help with this - you've no idea how rare it is for people to be willing to help out as remote-hands on this kind of slow debugging process
[18:59] <CIA-33> ubiquity: mterry * r3462 trunk/ (debian/changelog ubiquity/components/ubi-timezone.py):
[18:59] <CIA-33> ubiquity: gtk: Don't assume there will be a country shortlist for the selected
[18:59] <CIA-33> ubiquity: language. Fixes the lack of a default zone for Polish. LP: #425011
[19:06] <CIA-33> usb-creator: evand * r208 trunk/usbcreator/ (backends/windows/backend.py install.py misc.py): ifdef out some Linux-specific code.
[19:29] <CIA-33> ubiquity: mterry * r3463 trunk/ (debian/changelog gui/qt/app.ui):
[19:29] <CIA-33> ubiquity: kde: Let back/forward/quit buttons activate when highlighted and the
[19:29] <CIA-33> ubiquity: user presses Enter. Part of bug #46600
[19:50] <TommyT> cjwatson: rebooted & the two files are still in my home directory (I have persistence turned on with the iso ?) so I tried yet again, and that particular command barfs in exactly the same way. (I went back through the history and some still work.)
[19:51] <TommyT> cjwatson: I will try download again
[19:51] <cjwatson> maybe you can try uploading it somewhere else some other way
[19:51] <TommyT> yes I can do that.
[19:51] <cjwatson> sudo ./grub-fstest -d all /dev/sda1 crc /boot/grub/normal.mod >grub-fstest.out 2>&1
[19:51] <cjwatson> and stuff grub-fstest.out somewhere
[19:56] <TommyT> cjwatson: http://paste.ubuntu.com/273026/
[20:31] <TLF> hello, may I ask
[20:31] <TommyT> cjwatson: just to let you know I will be away again sometime in next 20mins.
[20:32] <TLF> the migration assistant just looks on "real" partitiomns (e.g. /dev/sda1) or it looks in mounted partitions, too?
[20:32] <TLF> For example, I have a working and mountable image of an OS, will migration assistant look there if I mount it?
[20:33] <TLF> Thanks (this question is assuming the new and separate from installer migration assistant is not going to be released in alpha6)
[20:39] <cjwatson> TommyT: ok, I'm looking
[20:51] <cjwatson> TommyT: one more thing for today, I'd like the output of: sudo debugfs -R 'stat /boot/grub/normal.mod' /dev/sda5 | cat
[20:52] <cjwatson> TommyT: except make that /dev/sda1 not /dev/sda5
[20:52] <cjwatson> TommyT: (the weird "| cat" at the end is to stop it sending the output to a pager)
[20:52] <TommyT> ok
[20:53] <cjwatson> I want to compare that inode's flags according to debugfs with the ones grub is seeing
[20:53] <cjwatson> because the latter look a bit suspicious to me
[20:57] <TommyT> I'm just going to type it
[20:57] <cjwatson> ok, please be absolutely accurate
[20:57] <TommyT> debugfs 1.41.9 (22-Aug-2009)
[20:57] <TommyT> stat: Attempt to read block from filesystem resulted in short read while reading inode 6817
[20:58] <cjwatson> wuh, even debugfs doesn't like it?!
[20:58] <cjwatson> maybe it is screwed then, but how come it's reproducible ...
[20:58] <TommyT> I should go soon.... BTW my USB stick is flashing constantly with r/w activity. Normal?
[20:59] <cjwatson> that's the same inode that grub-fstest is reading
[20:59] <cjwatson> not sure what the USB stick activity is, would rather not distract myself with that :)
[21:01] <TommyT> just in case I didn't close the port I just turned off the LAN connection.
[21:01] <TommyT> sorry but I will have to run for awhile. I will leave irc up so let me know if you want to try anything else
[21:02] <cjwatson> ok, I don't think I'll do any more tonight
[21:02] <cjwatson> my assessment is that the filesystem is indeed busted, but I would love to find out why it's being *created* that way
[21:02] <cjwatson> it looks almost as if bits of text have been written over the inode
[21:05] <cjwatson> though nothing very coherent, but lots of bytes are suspiciously ASCII lower-case letters when there's really no reason to expect them to be
[21:22] <CIA-33> usb-creator: evand * r209 trunk/ (3 files in 3 dirs): Add the missing retry dialog to the GTK+ frontend.
[21:35] <evand> cjwatson: regarding bug 431993 and the current state of clear_partitions, do you recall why we whitelist /usr/local as a mounted partition, but not as a directory under the root filesystem?  I imagine this is just a glaring error on my part, but I want to be sure I'm not missing part of the puzzle.
[21:36] <evand> TLF: it defers to os-prober to point it at operating systems.  If os-prober detects this OS, then migration-assistant should as well.
[21:37] <evand> cjwatson: other than it not being as easy as rm -rf /usr :)
[21:40] <cjwatson> evand: I think that's just a mistake - likewise /var/local
[21:40] <evand> indeed
[21:40] <evand> and /usr/src as the bug points out
[21:40] <cjwatson> and I tendyeah
[21:40] <cjwatson> maybe it would be worth going through the Contents file and looking for directories in which we actually ship files, and nuking those
[21:40]  * evand dusts off the FHS
[21:41] <cjwatson> though the cross-compiler directories would be inconvenient
[21:42] <evand> Contents file?
[21:43] <cjwatson> it's in /ubuntu/dists/karmic/ etc. on mirrors
[21:43] <cjwatson> map from file names to packages
[21:43] <cjwatson> apt-file and auto-apt use it
[21:44] <evand> I'm afraid of doing explicit deletion as if we ever ship a new directory and forget to upload clear_partitions, bad things happen
[21:44] <evand> err, well, can happen
[21:56] <cjwatson> bad things either way though
[21:57] <cjwatson> dunno - FHS directories except src are probably safe, at least
[21:58] <evand> yeah, it's a tough situation.  Hrm.
[21:58] <davmor2> cjwatson: any idea what tool mdz used on bug 43706
[21:58] <evand> I'll take a look at the Contents file and see what I can come up with
[22:00] <cjwatson> davmor2: to look at the /cow filesystem, you mean?
[22:01] <davmor2> to see how much memory is being used I think is what I need to check
[22:01] <cjwatson> davmor2: it at least used to be the case that if you booted with showmounts as a kernel parameter you could then look through /cow
[22:01] <cjwatson> which represents everything on the filesystem that's been changed with respect to the read-only squashfs
[22:02] <cjwatson> memory in general, well, all sorts of things, you could start with 'free' :-)
[22:02] <davmor2> would top be a good measure too
[22:03] <cjwatson> there are a variety of tools like that, but you have to be careful about interpretation
[22:03] <cjwatson> not all the numbers mean what you might intuitively expect
[22:04] <davmor2> cjwatson: sounds about right which is why I was asking before leaping in
[22:04] <cjwatson> in particular (a) Linux aggressively *tries* to use as much memory as it can to cache things, because memory is faster than disk; of course it'll evict the cache if it actually needs more memory
[22:05] <cjwatson> and (b) numbers for individual processes in top need to be interpreted carefully because they often include memory that's actually shared with other processes as well as memory-mapped regions (X especially tends to confuse people)
[22:05] <cjwatson> X maps your video memory and that's accounted to it in top
[22:05] <davmor2> so stick with free then :)
[22:06] <cjwatson> oh and a quirk of the X protocol means that other processes get to leak memory into X's process space
[22:06] <cjwatson> effectively
[22:06] <cjwatson> well, it depends what you want to measure and in how much detail
[22:08] <cjwatson> if it's just "would the computer perform reasonably if it only had 256M of memory", then the best way to do *that* is to boot with mem=256M as a kernel parameter
[22:08] <cjwatson> then, as far as Linux is concerned, your machine will only *have* 256M of memory
[22:17] <davmor2> cjwatson: cool I'll try that one first
[22:17]  * evand already knows the answer to that :)
[22:17] <davmor2> is it no
[22:17] <evand> I'd wager
[22:18] <cjwatson> I often use 256M for virtual machines; it's not great but it's not *too* bad
[22:18] <davmor2> does live say it needs 378M or something like that
[22:19] <davmor2> 384 I was close
[22:21] <evand> I keep them at 768M, unless I'm on the desktop, then I rarely bother in the first place
[22:21] <cjwatson> yeah, my laptop only has 1GB
[22:21]  * evand shakes his fist at Intel and the 3.3GB
[22:21] <davmor2> cjwatson: the machine I'm running the test on only has a 1 gig
[22:22] <evand> limit
[22:22] <evand> ah, that would do it
[22:22] <cjwatson> davmor2: yeah, but you don't run the tests in a vm, do you?
[22:22] <davmor2> nope
[22:22] <davmor2> machine as in hw
[22:23] <davmor2> I do use vm but not that often
[22:26] <davmor2> ouch Firefox is heavy
[22:26] <davmor2> you don't notice it on a 64bit machine with 4 gig of ram
[22:27] <evand> I've been happy with chromium
[22:28] <superm1> davmor2, for http://iso.qa.ubuntu.com/qatracker/result/3044/282 did you choose the live env or only-ubiquity mode?
[22:29] <davmor2> live in order to test live cd worked
[22:30] <superm1> okay
[22:30] <superm1> have you done any tests today that would have produced the same results as bug 432140 then today by chance?
[22:32] <davmor2> no every thing has been mostly stable in the loosest possible mean of the word
[22:41] <davmor2> ouch I take it back FF is light weight in comparison to OO.o
[22:41] <davmor2> still waiting for it to open