=== asac_ is now known as asac [04:51] persia: http://paste.ubuntu.com/47069/ if you can have a look at this script it would be great [05:11] ian_brasil: Perhaps it could take the new image file name as an argument as well? [05:11] * persia imagines grow-image ubuntu-mid.img 10000 [05:12] persia: ok [05:12] ian_brasil: Also, it looks like you're coying one image into another. I think you need to mount -o loop the old image, and copy -arP the contents of the old image into the new image. [05:13] Of course, I'm guessing. If it works, then it works :) [05:32] persia: http://paste.ubuntu.com/47075/ [05:33] ian_brasil: How much bigger do you want the image? [05:34] StevenK: On Friday we were discussing the two use cases of someone who wants to download and play (but doesn't have much bandwidth) vs. someone who wanted to be able to modify an image at their heart's content with ogra's script, and no fear of running our of space. This script seemed like a good way to address both use cases. [05:36] StevenK: i want to run dogtail tests on the ui and this means pulling lots of things in i believe https://bugs.edge.launchpad.net/moblin-ui-framework/+bug/155927 [05:36] Launchpad bug 155927 in moblin-ui-framework "cannot enable assistive technologies in ubuntu mobile" [Wishlist,Triaged] [05:39] ian_brasil: And I'd rather not increase the dailies to add 100MB or so. [05:40] StevenK: yes, persia suggested the script as an easy to maintain way for people to increase the img size [05:41] so you don't have to do this [05:53] persia: actually when i run ogras script on the new increased size image i do not get the chroot just the question whether i want to re-roll it [05:59] the larger image runs in kvm fine however [06:01] ian_brasil: Odd. [06:01] On the other hand, having it boot is an excellent thing :) [06:02] * persia tries [06:19] ian_brasil: usage notes: 1) it stomps on a local /images directory: you might want to use a tmpdir 2) I'm used to commands in the format command source destination, but this seems to be command destination source. 3) The final result ends up in a subdirectory, which is a little confusing. [06:20] Lastly, it appears that the result image contains a directory containing the source image. I think you want `sudo cp -arP images/tmpfs/* images/fobpath/` [06:25] ian_brasil: After changing images/tmpfs to images/tmpfs/*, edit-image works for me again. [06:31] persia: You don't, since * will miss the .disk directory on the image [06:35] Right. Then it needs something else. Perhaps a tar construction? [06:55] sudo cp -arP images/tmpfs/*.??* images/fobpath/ [06:55] maybe [07:00] persia: when you say works do you mean you increased the size and then ran ogras script to get the chroot? === asac_ is now known as asac [10:12] ian_brasil: Sorry for the delay. Yes, tha is precisely what I mean. [10:20] hey persia [10:20] sorry I didn't had time to look to update bluez-gnome [10:20] I'm Baptiste Mille-Mathias [10:26] crevette: No worries. It's something that needs testing as much as anything else. What features have you been testing? input? comms? audio? [10:27] persia, my test was weak, I just tested the empty plugin name in the UI [10:27] I didn't look at other stuff [10:28] crevette: Heh :) That seems common. While it's *excellent* that the display bug is fixed, I'm not sure it's best to update if it breaks other things. [10:29] I'd rather have a bad display than have it not work with bluetooth keyboards & mice. [10:29] I didn't tried to do the merge stuff also I didn't what is differnet from a plain update form latest ubuntu [10:29] persia, the package is just bluez-gnome oso id shouldn't impact except the applet and the properties UI [10:29] sorry for typo [10:30] crevette: "shouldn't": I agree. Still, needs testing :) [10:30] persia, so https://bugs.launchpad.net/ubuntu/+source/bluez-utils/+bug/258738 could be moved to bluez-gnome then ? [10:30] Launchpad bug 258738 in bluez-utils "intrepid regression: bluetooth services without name in tab "Services"" [Undecided,Confirmed] [10:31] perhaps we can also Marcel to give some hints [10:33] crevette: Moving the bug to the right package sounds like a good idea :) [10:33] Also, who is Marcel? [10:34] Marcel Hottman, upstream developper [10:34] Oh, right. I think we probably want to actually test what we have before we do that :) [10:34] It's just a matter of getting all the right people to test. [10:35] persia, it is just i'm not sure if it is bluez-gnome or bluez-utils fault as per comment https://bugs.launchpad.net/ubuntu/+source/bluez-utils/+bug/258738/comments/5 [10:35] Launchpad bug 258738 in bluez-utils "intrepid regression: bluetooth services without name in tab "Services"" [Undecided,Confirmed] [10:35] crevette: Right, but you also said that updating bluez-gnome to 0.28 fixed it for you, right? [10:36] yep [10:36] So, it's probably a bug because of different versions of bluez-utils and bluez-gnome. Given that bluez-gnome is out-of-date, I think bluez-gnome is the place to fix it. [10:39] okay so I move the component to bluez-gnome [10:51] crevette: Do you need assistance merging the code, or do you welcome that task? [10:55] persia, I need guidance [10:56] crevette: OK. Have you ever merged Ubuntu changes previously. [10:56] and I don't knwo when I'll be able to do it, because I'm a new father, and my time is limited [10:56] Yes indeed. That can limit time :) Congratulations! [10:57] thanks [11:02] persia, I did merge but a while back [11:02] at least more than one year [11:02] crevette: If you have time in about 20 minutes, I could walk through the merge with you. If not, I'll chat with slytherin (who has also been working on it), and we'll process the merge. [11:02] I seen there is a doc in the wiki regarding merge [11:03] I won't be available tonight before 21:15 GMT tonight [11:05] crevette: OK. We'll try it in the meantime, and hopefully have something for you to test then. Have a good day :) [11:05] thanks [11:05] anyway I'm interested to have a good bluetooth stack in GNOME [11:06] I did some patches on upstream (like gnome-user-share) to improve user experience [11:06] but my free time is shorten now [11:14] crevette: Understood. Any way you can help is appreciated :) [11:14] persia, bluez 4 is planned for intrepid+1 ? [11:15] I seen Fedora 10 wich will be issued in same time than intrepid will have it, but I don't know it it implies a lot of changes in dependents packages [11:17] I don't know that anyone is specifically planning BlueZ 4 for jaunty, but I'd be surprised if nobody updated it. [13:23] ian_brasil, https://code.launchpad.net/~ubuntu-mobile-dev/ubuntu-mobile/mobile-scripts has a grow_image.sh script as well now [13:23] i wrote it in a way that no root access is needed (uses mtools) [14:33] hello [14:33] I wonder whether ubuntu mid is meant to be an mobile phone OS [14:34] or is it just for netbooks and alike? [14:34] christian_: Nope. It's for little tiny devices. Like the Nokia N810 or the Sharp Zaurus. But it doesn't run on those, because they are ARM. [14:34] For netbooks and the like, there's a new flavour being developed: "Ubuntu Mobile", but it will be new for intrepid. [14:35] persia: when i try to run ogras script on the new image i get this error [14:35] /tmp/image/casper/filesystem.squashfs: No such file or directory [14:35] chroot: cannot run command `mount': No such file or directory [14:35] chroot: cannot run command `mount': No such file or directory [14:35] chroot: cannot run command `mkdir': No such file or directory [14:35] chroot: cannot run command `mount': No such file or directory [14:35] cp: cannot create regular file `/tmp/mergemount/etc/': Is a directory [14:35] chroot: cannot run command `su': No such file or directory [14:35] chroot: cannot run command `rm': No such file or directory [14:35] chroot: cannot run command `umount': No such file or directory [14:35] chroot: cannot run command `umount': No such file or directory [14:35] chroot: cannot run command `umount': No such file or directory [14:35] chroot: cannot run command `rm': No such file or directory [14:35] christian: I suppose if you've a phone with a high enough resolution, and the right processor, you might be able to run Ubuntu MID, but it doesn't have a dialing app, so it wouldn't be very useful for a phone. [14:35] ogra: ^^ ? [14:36] persia: thanks [14:36] ian_brasil, did you use my new grow_image.sh ? [14:36] ah sorry, please ignore...i just saw ogras script [14:36] :) [14:37] i'll add a shrink_image.sh later as well that strips out empty blocks again ... and a wrapper to call the three in a row [14:37] that rocks [14:37] the edit_image.sh definately needs error checking and cleanup routines [14:38] lease check for stale mounts if it fails [14:38] before running it again [14:38] ogra: ok, i will be hammering it today so if there is something wrong i imagine it will show up [14:39] thanks for that :) [15:16] ogra: i just grew an image and then ran edit_squashfs on it and i get this [15:16] /tmp/image/casper/filesystem.squashfs: No such file or directory [15:16] chroot: cannot run command `mount': No such file or directory [15:16] chroot: cannot run command `mount': No such file or directory [15:16] and it offers to re-roll but not the chroot [15:17] did you check for stale mounts before ? [15:17] how can i do that? [15:17] (that includes /proc and /sys [15:17] mount [15:17] :) [15:17] ha ha [15:17] there might be mounts from the failed attempt hanging around [15:18] i really need to add more error checking to edit_image [15:36] ogra: i just rebooted but still do not get a chroot [15:36] can you do the steps of the script manualy step by step ? [15:36] that should expose what exactly doesnt work [15:37] ok [15:40] it is falling over here [15:40] sudo mount -o loop -t squashfs /tmp/image/casper/filesystem.squashfs /tmp/squashfs [15:40] you have squashfs support? [15:40] modprobe squashfs [15:41] another option is that you ran out of loop devices ... [15:42] lool: yes [15:42] * ogra isnt really a friend of loop mounting ... thats why the grow_image script doesnt mount anything at all [15:42] does /tmp/image/casper/filesystem.squashfs exist ? [15:42] and /tmp/squashfs [15:44] /tmp/image/casper/filesystem.squashf this does not exist [15:44] so you have an empty image :/ [15:45] or the image isnt mounted and thats only a subsequent error [15:46] mount|grep /tmp/image [15:46] what does that return ? [15:47] /home/ian/Dev/modified_images/ubuntu-mid.img on /tmp/image type vfat (rw,loop=/dev/loop0) [15:47] ok [15:47] and /tmp/image/casper exists ? [15:47] ogra: no [15:47] weird [15:48] /tmp/image/ldlinux.sys [15:48] just that [15:48] thats after you used grow_image ? [15:49] i assume so [15:49] dpkg -l mtools [15:49] ? [15:50] ogra@osiris:~/Devel/branches/mobile-scripts$ grep Requires image-tools/grow_image.sh [15:50] # Requires: mtools, syslinux [15:50] :) [15:51] Desired=Unknown/Install/Remove/Purge/Hold [15:51] | Status=Not/Installed/Config-f/Unpacked/Failed-cfg/Half-inst/t-aWait/T-pend [15:51] |/ Err?=(none)/Hold/Reinst-required/X=both-problems (Status,Err: uppercase=bad) [15:51] ||/ Name Version Description [15:51] +++-==============-==============-============================================ [15:51] ii mtools 3.9.11-0ubuntu Tools for manipulating MSDOS files [15:54] hmm [15:55] do you have enough diskspace ? [15:55] i have 50GB free [15:55] the FS you wokr on should have 2x the imagesize plus the amount you grow [15:57] now it is working [15:59] weird [16:30] mterry, tremolux, DebbieFoghorn, bfiller: have any of you tried this morning's install? Looks badly broken to me! ;( [16:30] tonyespy: wrong channel :) [16:32] couple of questions: why does apt-get dist-upgrade in a chroot pull in firefox as well as midbrowser?..also is the preferred way to enter s3 sleep to use the script or pm-suspend? [16:32] bfiller: doh [16:37] ian_brasil, pm-suspend imho ... as it takes care for several extra things ... but opinions here differ a bit ... i think lool isnt a pm-utils fane :) [16:37] *fan [16:37] no idea about the dist-upgrade ... might be through a recommends [16:39] ogra: ok [17:43] ogra: Oh I'm not against pm-utils as a concept, but it's reputed slow [17:43] because we call to many useless scripts [17:43] We should aim at having the kernel suspend properly, and that will be increasingly the case with kernel modesettings [17:43] on a MID you will only need to use a subset [17:44] you will always have misbehaving modules needing a special treatment [17:44] ogra: You mean restricted modules? [17:46] lool, no, just modules [17:46] the r8169 is a typically misbehaving free module [17:46] due to the fact that new HW comes out and gets added [17:47] you will always have kernel releases where modules wont behave properly and you will need workarounds [17:47] ogra: It feels to me it would be a cleaner approach to fix the modules or flag them in kernel rather than in user space [17:47] or HW that doesnt do the right thing (think IBM) [17:47] right, an many cases the kernel is the right place to fix [17:48] but if you are in a freeze and the fix is to intrusive a scripted workaround might be better [17:49] and somethimes you simply have the case where a fix breaks on one HW while it fixes the other and you have to special case [17:49] The problem is that we pile mass of checks which are run for all hardware despite being hardware specifics -- just checking whether you have the hw by poking hal or running scripts which parse xml configs takes ages [17:49] While in the kernel, quirks are at least driver specific [17:50] most of the stuff pm-utils does *is* to apply these quirks [17:50] Nobody really knows how accurate the userspace quirks are these days [17:50] like vbetol stuff [17:50] For instance some thinkpad quirks have been applied to some models and removed for others, but the difference is actually the bios version [17:51] right [17:51] but tats nothing you can easily solve in kernel [17:54] if you look at /usr/lib/pm-utils/sleep.d/ ... most of the things in there are actually userspace fices [17:54] *fixes [17:56] it strikes me that it would be more efficient to source the functions once [17:56] * ogra lols about the comments in 95led and 05led [17:57] they are nearly the same scripts and both of them suggest to merge with the other [17:57] you wont get rid of stuff like 90clock [17:57] or 10NetworkManager [17:58] 94cpufreq is definately a time hog [17:59] but can be dropped once the module behaves [18:00] 95anacron even cals the initscript ... thts likely slow as well [18:01] * ogra urghs about 98smart-kernel-video [18:01] woah [18:01] ogra: You will if instead of piling shell foo, you implement it in C and over dbus [18:01] smart_kernel_nvidia || smart_kernel_fglrx || smart_kernel_intel ... [18:01] sigh [18:01] Anyway, it's in the works [18:01] ogra: See :) [18:02] so on intel it will wait for the return onf the nvidia and fglrx functions before even starting to think [18:02] thats a waste [18:05] wow, 99video is a real monster === asac_ is now known as asac [18:42] for the time being the script is a nice workaround? ..i am thinking of extending the shutdown applet rbelem wrote with this script..you think it is a good idea?