[00:09] <xivulon> evand is anything in particular holding back the release of the new lupin hardy?
[00:15] <evand> lupin 0.19 is in hardy-proposed.
[00:22] <xivulon> evand does that mean that it is on the ISO? from the manifest it looks like 16 is in there.
[00:23] <evand> What ISO?  Are the 8.04.1 CDs being generated somewhere now?
[00:24] <xivulon> yep they are in http://cdimage.ubuntu.com/hardy/daily-live/
[00:26] <xivulon> and most things relevant to wubi seems to be in :)
[00:26] <xivulon> except casper/autopartition-loop/lupin
[00:31] <evand> hrm, not sure
[01:37] <compbrain> cjwatson: One last question for the day, Sometime last year you recommended I preseed a call to `debconf-set partman-auto/disk` so I could determine if a machine had a disk at /dev/sda or /dev/hda. Do you know what stage I would have called that? an early command is too early as disk drivers haven't loaded and such
[01:38]  * compbrain fails searching irclogs.u.c
[01:40] <cjwatson> compbrain: I might have suggested that you write out a partman auto.d script; nowadays it would actually be in /lib/partman/display.d/00local or something like that
[01:40] <cjwatson> doing that on the fly in partman/early_command
[01:41] <compbrain> partman/early_command is probably what I was looking for.
[01:41] <cjwatson> err, preseed/early_command I mean
[01:41] <compbrain> oh :d
[01:41] <cjwatson> there's no partman/early_command (maybe there should be!), that was a typo
[01:42] <compbrain> I ended up with a 2 line bash script ( [ -b /dev/sda ] && debconf-set ....)
[01:42] <compbrain> But I can't figure where I would have called that from where disks would have been detected but not partitioned
[01:43] <cjwatson> /lib/partman/display.d/00local would be suitable
[01:43] <cjwatson> it's awkward at the moment, you have to write that script out dynamically from preseed/early_command
[01:43] <cjwatson> anywhere else, you're relying on luck at best
[01:44] <cjwatson> perhaps you could file a bug on partman-base to ask for a preseeding hook there?
[01:46] <compbrain> In theory I can pipe some commands into that file, and they are executed prior to partmans logic, yes?
[01:46] <compbrain> Yea, i'll do that.
[01:47] <cjwatson> right
[01:47] <compbrain> Groovy
[01:51] <compbrain> #239348 filed
[01:55] <cjwatson> ta
[02:03] <compbrain> Giving this a whirl.
[02:29] <compbrain> Everything looks good, except it didn't want to purge the old lvm volume..
[04:21] <compbrain> Everything looks groovy. Thanks again cjwatson.
[06:02] <superm1> compbrain, alternatively if you know that your systems would support EDD, you can also pass edd=on on the kernel command line, and then use the EDD symlink made in /dev/disk/by-id instead of hda/sda
[06:07] <compbrain> EDD? I can't seem to expand that acronym
[06:08] <superm1> http://linux.dell.com/projects.shtml#edd
[06:09] <compbrain> ahh, ok. Thanks.
[06:09] <superm1> i'm not sure of your particular situation whether all machines you are working on will support it
[06:10] <superm1> if so, its a very useful technology for install time decisions like that
[06:10] <compbrain> I ended needing to add in a workaround in grub as well. Some of the machines i'm installing to have 6 internal disks, the installer tries setting up grub with root=hd(4,0) or something crazy like that.
[06:11] <superm1> yeah cjwatson added a patch for grub that adds EDD support to that too
[06:11] <superm1> so it can more intelligently decide on disks
[06:11] <compbrain> I'm going with the knowledge that the 100G disk in the system is the boot disk connected to the motherboard directly, install/boot from that
[06:12] <compbrain> [open /proc/partitions, sort by device size, excluding zero sized devices, pick the smallest and install to it] (all the other drives are 500G or larger)
[06:12] <superm1> geez that sounds like a hack.  what happens when you start to ship 120GB drives because the stock of 100GB drives runs out :)
[06:12] <superm1> ah
[06:13] <compbrain> hack, yeah. But some of this hardware is old, pre edd-capable iirc
[06:13] <superm1> as a more long term solution you may consider to investigate if EDD does a better job for you then.  hacks like that unfortunately don't always work forever :)
[06:13] <superm1> ah
[06:13] <compbrain> in this particular case i've got 5 machines of the same model, but with different disk configurations
[07:15] <ganes> cjwatson, please help to boot the by setting boot=casper in isolinux.cfg file
[07:23] <cjwatson> ganes: not without more information from you; you've asked this question at least five times now here (I mostly happened not to be here at the time), and so far haven't provided any real information on what's going wrong
[07:24] <cjwatson> I'm about to get on a train, so what you need to do is (a) provide a clear and detailed description of what you've changed and what symptoms you are seeing, (b) stick around rather than leaving
[07:24] <ganes> cjwatson, i customised the filesystem according to my need , but the cd is not bootng , the package casper is persent in the build cd but not booting after the creation of live cd
[07:25] <ganes> cjwatson, while booting i checked the casper.log file "unable to find the filesytem "
[07:26] <ganes> cjwatson, this is the log message i am getting
[07:45] <compbrain> ganes: Perhaps you can prepare a bug report detailing your setup and include the logs on launchpad
[07:46] <superm1> ganes, did you make sure that your live disk contained the .disk directory still and that the initrd contained the same UUID as the one in the .disk directory?
[07:46] <superm1> that is how it finds the matching live filesystem to use when booting casper
[07:47] <ganes> superm1, where is this .disk tell me the location & i will check
[07:47] <superm1> it is supposed to be in the root of your CD
[07:48] <superm1> it's very easily missed (as it's a hidden directory) when you are copying the contents of a CD over to customize
[07:48] <ganes> superm1, if not there what i have to do
[07:48] <superm1> ganes, redo your customizations to include it again
[07:48] <ganes> superm1, from where i have to copy
[07:48] <superm1> it's in the root of every functional live CD
[07:49] <superm1> it has to come from the same CD that you got the matching initrd however
[07:49] <ganes> superm1, ok thanks & i do the customisation & come back
[07:49] <superm1> as there is a UUID stored in both
[07:50] <ganes> superm1, store in both means one is at root dir & where the other one
[07:50] <superm1> casper/initrd.gz
[07:50] <superm1> there is a UUID inside that file
[07:50] <superm1> so if you customized your initrd at all, you will need to make sure the UUID in .disk/ and in that initrd match
[07:51] <ganes> superm1, as a .disk file inside the initrd
[07:51] <superm1> ganes, its actually a flat file inside the initrd.  the exact naming of it escapes me right now
[07:52] <superm1> there is a flat file in .disk as well, usually something like casper-uuid-generic
[07:52] <superm1> or something like that
[07:52] <ganes> superm1, ya i customised the initrd but i didnt see any .disk file there
[07:52] <ganes> ok
[07:53] <superm1> ganes, if you look at the casper package, there is a script included with it
[07:53] <superm1> casper-new-uuid
[07:53] <superm1> it's used to make a UUID in the initrd and a matching one for the .disk directory
[07:53] <superm1> you may end up needing to do that
[07:54] <ganes> ok thanks
[07:59] <ganes> superm1, i have the filesystem in one partition there if i install casper it is updating the initrd but not the partition itself not booting for that above thing is applicable
[08:00] <ganes> superm1, i have the filesystem in one partition there if i install casper it is updating the initrd but the partition itself not booting for that above thing is applicable
[08:04] <superm1> ganes, if you are replacing the initrd on the live disk, you have to ensure that the .disk/casper-uuid-generic file gets updated to that same UUID no matter what the circumstance
[08:11] <ganes> superm1, thank you very much
[08:11] <superm1> no prob. hope you can sort it all out :)
[10:23] <CIA-1> pkgsel: cjwatson * r104 ubuntu/ (9 files in 3 dirs): merge from Debian 0.20
[10:26] <CIA-1> pkgsel: cjwatson * r105 ubuntu/debian/po/ (am.po mr.po): msgmerge
[10:35] <CIA-1> pkgsel: cjwatson * r106 ubuntu/debian/changelog: releasing version 0.20ubuntu1
[10:40] <tjaalton> don't know if this is the right place to ask, but I'm trying to make an usplash theme to only output text and not use any throbber at all
[10:41] <tjaalton> but haven't figured out how to do that
[10:47] <tjaalton> which button should show the text, btw? I tried everything
[10:53] <xivulon> cjwatson re #235518, the original bug report was due to something else, and I am not aware of other issues, do you still wish the fix to be in the point release?
[10:53] <ganes> superm1, there is no any file with the name .disk i checked under the folder /root
[10:54] <cjwatson> ganes: it should be in the CD root, *not* in the initramfs
[10:55] <ganes> cjwatson, i boot into the feisty version ubuntu & checked the file
[10:55] <ganes> cjwatson, its not there
[10:55] <cjwatson> xivulon: well, does it go wrong in a simple non-English install test (say German)? if so, the fix should be included; if not, save it for intrepid
[10:55] <cjwatson> ganes: the CD root is not /root
[10:56] <cjwatson> the CD root == the .iso contents
[10:56] <cjwatson> typically on /cdrom
[10:56] <cjwatson> and I am absolutely certain that it has been present since Ubuntu began
[10:56] <ganes> cjwatson, it mount the cd iso image
[10:56] <ganes> cjwatson, ok
[10:57] <cjwatson> if you aren't finding it on a standard Ubuntu image, you're looking in the wrong place
[10:57] <ganes> cjwatson, no i booted & checked the file ., ok i do as you said now
[10:58] <cjwatson> checked *which* file?
[10:58] <cjwatson> if you're looking under /root as you said above, that's the wrong place.
[10:58] <ganes> cjwatson, ok
[11:01] <ganes> cjwatson, yeah its there ., where i have to paste this file in initrd
[11:01] <cjwatson> it MUST NOT BE IN THE INITRD
[11:01] <cjwatson> it belongs in the .iso contents
[11:02] <ganes> cjwatson, then what is the use of this file for me
[11:02] <cjwatson> /cdrom/.disk/casper-uuid* needs to match /conf/uuid.conf in the initrd
[11:03] <ganes> cjwatson, if this both get match will the cd boot by casper
[11:03] <cjwatson> yes
[11:03] <ganes> ok
[11:04] <cjwatson> the purpose is to ensure that if you copy a live CD to a partition on the hard disk, then booting a different live CD will actually boot the right thing
[11:05] <ganes> cjwatson, i copied the filesystem to hardisk & update the initrd file by installing my own squashfs & aufs ... after updating the initrd the system is not booting
[11:06] <ganes> cjwatson, reason for customisation is want to remove the usplash & i want to make it as bootsplash
[11:15] <tjaalton> umm, does the usplash in hardy support usplash_write at all?
[11:16] <tjaalton> usplash-test.sh doesn't output any text
[12:44] <xivulon> cjwatson: I haven't tested with german installations but I haven't seen any complaints either.
[12:44] <xivulon> I believe that evand did a few french installations...
[12:44] <xivulon> evand did you notice any issue re 235518?
[13:04] <ganes> cjwatson, if i boot with casper i am getting the error as follows "mounting /dev/loop* on /filesystem.squashfs cannot mount /dev/loop* on /cdrom/casper
[14:09] <evand> xivulon: Wubi installs have never failed for me because of the language selection (Spanish and French XP Pro).  I verified the French XP Home bug by modifying os-prober to output the same text.
[14:09] <evand> So I imagine the fix should be saved for Intrepid.
[14:09] <evand> But lets see how the new 8.04.1 CDs work for others.
[14:16] <xivulon> I have done then with the wubi changes
[14:17] <xivulon> I have a new branch since I did a fiew turnarounds and wanted a cleaner slate
[14:17] <xivulon> I am inclined to rename hardy -> trunk and save there the new changes
[14:18] <xivulon> once the other packages mentioned yesterday hit the iso we can start testing
[14:20] <xivulon> on second though I might keep the hardy name since the intrepid one will be basically a rewrite
[18:05] <CIA-1> ubiquity: evand * r2690 ubiquity/debian/po/pt_BR.po: Fix a pt_BR translation issue by hand; translation team notified.
[18:07] <CIA-1> ubiquity: evand * r2691 ubiquity/debian/po/pt_BR.po: Ugh. Proper fix for the above change.