[01:59] <tripzero> what is UME going to do for a gps app?
[07:42] <dholbach> good morning
[07:46] <lool> morning
[07:46] <dholbach> hi lool
[07:52] <bpsew> morning
[09:54] <mvo> lool: for ubuntu-mobile image, should I use more than "http://ppa.lp.net/ubuntu-mobile/ubuntu hardy main" and the regular hardy archive on ports.u.c?
[10:00] <lool> mvo: Hmm theoritically this is pulled if you use a MIC platform
[10:00] <lool> mvo: If you don't, then you should probably include the ppa; yes
[10:01] <lool> mvo: Also, we'll move to a different archive soon
[10:01] <lool> mvo: Technically, the releases we have been doing were not based on ports + ppa, but on a snapshot of these hosted on snapshot.u.c
[10:02] <theseinfeld> lool when is the plan to move to the new archives?
[10:03] <mvo> lool:  ok, thanks - I was playing with simple-mobile-builder today again and with davmor2 we fixed some issue, it should be pretty good now. did the release include virtual images too?
[10:03]  * mvo wonders if it is worth building new ones based on the snapshot.u.c sources.list entries
[10:04] <davmor2> might be for RC2
[10:04]  * theseinfeld wonders on davmor2
[10:05] <lool> theseinfeld: when it's in place :)
[10:05] <theseinfeld> any ~ deadline?
[10:05] <lool> mvo: We didn't release vm images until now
[10:05] <theseinfeld> I actually made some vm images
[10:05] <davmor2> :)
[10:06] <theseinfeld> you need to do some hacking with the initram to get it working based on the project chrooted directory, other than that, it is quite straight forward
[10:06] <lool> mvo: I'd suggest you only care about a single config in smb, or use the MIC provided configs -- all supported MIC configs could be made available with the config copying machinery I added
[10:06] <lool> mvo: For the single default config, I'd recommend hardy+ppa for now and mobile archive afterwards
[10:07] <theseinfeld> you mean s/hardy/ports/ right?
[10:07] <lool> mvo: BTW I chatted with persia on the topic, and he was tempted to hack on MIC to add vm creation support from it
[10:08] <lool> mvo: he mentionned concerns of bug-to-bug compatibility between MIC and u-v-b
[10:08] <persia> I've since been convinced that this wouldn't be beneficial.
[10:08] <lool> theseinfeld: It's the same thing
[10:08] <davmor2> lool that would be cool :)
[10:08] <lool> theseinfeld: hardy is a dist/suite; ports.u.c is where we host binaries for non-official architectures
[10:08] <persia> As much as getting exactly the right set of packages together is desired, we can probably do that by abusing --add-pkg in u-v-m
[10:08] <lool> theseinfeld: archive.ubuntu.com where we host source in all cases and binaries for supported arches
[10:09] <lool> persia: This is implemented already :)
[10:09] <persia> lool: Completely?  Last time I has slightly different results, but I've not tried the latest simple-mobile-builder.
[10:09] <lool> mvo: Is your goal to spend time on it so that it's ready for release?  Otherwise, I'd suggest we work together after the release
[10:10] <lool> persia: Well it's fully implemented, but some things differ
[10:10] <lool> persia: There are currently some bugs and also s-m-b adds a fixed set of packages to the list of pkgs to isntall
[10:11] <lool> persia: Also, config is copied manually instead of being provided by an ume-config-vm package
[10:11] <mvo> lool: after release is fine for me, I just tweaked it a bit for davmor2 (he found some bugs)
[10:11] <persia> lool: It's the differences that worried me.  The argument that convinced me not to make a VM platform for MIC was that we weren't likely to do lots of VM testing of the final release, and should look at how we want to VM-test for intrepid instead.
[10:11] <lool> mvo: Ok; cool
[10:12] <lool> mvo: I have a bunch of bugs/changes I'd like to fix, I mentionned them to you shortly, but generally speaking I think it's quite good already; I just didn't have much time since UDS to work on it
[10:12] <mvo> and I found that neverball starts :) its just painfully slow, but other stuff like frozen-bubbles or gnash seems to be pretty usable
[10:12] <davmor2> mvo: rev10 is much better menu is correct etc.  I'll play around with it and make up a quick wiki page on the testing section
[10:12] <mvo> sure, I understand that, there is a rleease to make :)
[10:13] <lool> persia: Apart of squashfs and initramfs, I think the differences were globally minimal with this approach
[10:15] <persia> lool: Yes, although I find that the way the filesystems work for the MIC images can be part of some of the differences we see with our testing.  Depends on the test goal.
[10:15] <persia> I'll take another look at simple-mobile-builder with the new updates: maybe my opinion is outdated.
[10:16] <lool> persia: Exactly, it depends on what you want to test
[10:16] <persia> That's true too :)
[10:16] <davmor2> persia: lool:  As a tester using MIC I spent most of the day confirming if it was a bug or an issue with MIC.  Using the kvm image I lose the majority of that.
[10:16] <lool> persia: Arguably we could as well use released images for cdrom or USB in vms, but these wouldn't work
[10:17] <lool> (and we could make these work too IMO)
[10:17] <persia> lool: Let's make the daily images work for VM for intrepid, for wider (safe) testing.  For now, the small differences are probably acceptable.
[10:18] <lool> davmor2: Certainly we need to fix MIC issues for this release, but do we want to spend time creating vms which allow us to diagnose MIC bugs if they are only 10% of the bugs and we wont mainly use MIC in intrepid?
[10:18] <lool> Or do we want vms which reproduce 90% of the bugs by allowing us to run the same packaged apps?
[10:19] <lool> persia: I think I could argue that there are hardware bugs which you can't reproduce in vms anyway, so we wont ever reproduce it in full
[10:20] <persia> lool: You don't have to argue.  I've already ceased efforts to make MIC do VMs :)
[10:20] <lool> persia: Perhaps if we had a more layered approach (probably non-MIC) we could try reproducing more of the current build layout (squashfs and initramfs)
[10:20] <lool> For example, if MIC initramfs was properly packaged, such as casper, it would be trivial to include it in vms
[10:21] <davmor2> lool: true but then most of the apps don't run in mic which is the issue.  They are running in kvm version and finally it enables me to test UME again which then helps confirm issues with cgregan.  I have noticed issues that he hadn't and visa versa.  That currently isn't workable in MIC
[10:22] <lool> davmor2: I understand your point, but I don't understand what you're arguing for/against
[10:22] <persia> davmor2: There's two separate issues: one is VM vs. Xephyr.  I think we all agree VM is the solution.  The other is MIC vs. something else.  I think we ought use the same tool for everything, but can wait for intrepid.
[10:24] <davmor2> Okay cool :) Just making the point rather than arguing though.
[10:26] <lool> The only thing I'm arguing against is spending too much times on vms (for example by spinning them from MIC) at this point; simple vm builds are a good start already :)
[10:28] <persia> Four people.  Four different goals.  None at odds.  Discussion is good :)
[10:30]  * lool would just like to avoid waste of efforts; who am I to prevent people to work on something they'd like to do?  :-)
[10:32] <theseinfeld> so, in the future there might be no MIC?
[10:32] <theseinfeld> long time future :)
[11:02] <jussi01> persia: happy now? :D
[11:04] <persia> jussi01: Thanks.  I can now point at things like bug #217838 and get a response :)
[11:04] <jussi01> :)
[11:21] <davmor2> Guys on date and time if you change manual to ntp are you able to install the ntp packages?
[11:22] <davmor2> manual to automatic sorry
[11:29] <persia> Installing packages is a little tricky: there's no GUI for this by default.
[11:30] <ogra> what do you do with the default action then ? time-admin tries to pop up the installation dialog
[11:31] <davmor2> and then crashes :)
[11:31] <ogra> hmm, it surely shouldnt crash blindly
[11:32]  * persia tests to try to find a way to not crash
[11:33] <ogra> probably linking the app its trying to call to /bin/true or so might help 
[11:33] <davmor2> ogra: it does throw up a polite message
[11:33] <ogra> ah, k
[11:33] <persia> davmor2: I can't get a crash, although it's not currently useful.
[11:34] <davmor2> persia: might be an issue with vm
[11:35] <persia> davmor2: Maybe, but that's not the sort of thing I'd expect for VM vs MIC images.
[11:36] <davmor2> sorry not a crash it just flashes like it has crashed and then puts up a you can't do this message
[11:36] <persia> davmor2: Ah.  Yes.  That's exactly my experience.
[11:37] <persia> I'm glad to know we see the same thing: it makes it easier to be sure of any fix :)
[11:37] <davmor2> that's why I assumed it had crashed
[11:38] <persia> davmor2: I define "crash" as when the tool goes away completely (or worse, the windowing system goes away).
[11:38] <persia> And "hang" as when it stops accepting any input.
[11:38] <davmor2> okay we'll all get on the same page :)
[11:41] <dns53> can anyone tell me what the differences will be between ubuntu-mobile created by the image creator vs a normal install on a computer with the ubuntu-mobile pakcage
[11:42] <persia> dns53: There are some setup differences, and differences in how the initial base files are created.
[11:43] <persia> Also, if you previously installed a different sort of Ubuntu, you may end up with a different set of applications.
[11:43] <persia> Further, ubuntu-mobile wasn't quite ready for hardy, so there are some packages that are a little different.
[11:43] <persia> I seem to remember a couple outstanding bugs on the meta-package as well, perhaps masked because we tend to use MIC.
[11:44] <lool> Yes, for example the terminal.desktop file is shipped by mobile-basic-flash
[11:44] <dns53> well there is no entry in gdm, i created one and runs the script in the wiki
[11:45] <lool> dns53: We don't use gdm in UME
[11:48] <persia> dns53: This would be one of the cases where your package selection was different :)
[11:49] <persia> dns53: If you want to test, https://wiki.ubuntu.com/Testing/Cases/UMEinstall-kvm was recently published, if that helps.  It may change over time (it's new), so be warned.
[11:49] <dns53> i've tried the image-creator a few times but nothing has booted of my eeepc yet
[11:52] <persia> dns53: Which platform did you select in MIC?
[11:54] <dns53> micassin is suppose to work but the images created when booting of a usb drive fail on boot, not sure if it is the kernel, init, mounting the root file system
[11:58] <persia> dns53: Hrm.  I'm not sure then.  The McCaslin images work for me, but I have different hardware.
[11:59] <ogra> dns53, you could try if the classmate image works for you ... its pretty HW specific though, you likely need additional stuff 
[11:59] <ogra> http://people.ubuntu.com/~ogra/classmate/images/hardy/
[12:00] <ogra> (and not (yet) based on mobile)
[12:23] <davmor2> I'm still having issues with midbrowser open oo-trig.xls in OO.o.  But on a plus side ODR does open it now although it does look hideous.
[12:51] <persia> davmor2: Having Time&Date install the NTP software depends on having a working synaptic, which isn't in -mobile.  We could likely hack something more directly using libapt, but it wouldn't be quick.  I think I'd rather it didn't work than possibly crashing without the nice user message, etc.
[12:52] <mvo> I would suggest using the update-manager backend code, it should be trivial to extend for this purpose
[12:52] <mvo> (probably not hardy though)
[12:52] <persia> mvo: Calling python from C?
[12:52] <mvo> IIRC time-admin just does a system() or exec() so that should be ok
[12:52] <mvo> it does not link against libapt
[12:52] <persia> Or calling some special update-manager call that allows for arbitrary new package installation?
[12:53] <mvo> yes, that is what I was thinking
[12:54] <persia> mvo: Could be.  How about a call into gnome-app-install?  Is UpdateManager much better?
[12:55]  * persia hasn't looked at gnome-app-install other than the .desktop files
[12:55] <davmor2> persia: A simpler fix for now might be to put a warning in place and then add at the bottom open terminal and type sudo apt-get install ntp
[12:55] <ogra> persia, note that unionfs tears down the performance massively for g-a-i
[12:55] <davmor2> and put in a better fix for intrepid
[12:55] <persia> ogra: True.
[12:56] <ogra> how about gdebi ? 
[12:56] <ogra> you could just mae a call to gdebi-gtk
[12:56] <ogra> *make
[12:56] <persia> Doesn't gdebi expect a local .deb?
[12:56] <ogra> (and pull the package in advance somehow)
[12:56] <ogra> righ
[12:56] <ogra> t
[12:57] <persia> Could seed the ntp package, but it's a very, very ugly hack.
[12:57] <persia> (where "seed" means putting in place in advance: not adding to the seed)
[12:57] <mvo> if you have apturl installed, just system("apturl apt:ntp")
[12:57] <ogra> apt-get -d ntp && gdebi-gtk /var/cache/apt/archives/ntp*.deb or so
[12:58] <mvo> but iirc it uses synaptic as its backend too, so please ignore that last suggestion
[12:58] <persia> ogra: Right.  That's still dirty for both gst-package and gst-time
[12:58] <ogra> right
[12:58] <mvo> I'm happy to make a mode for gdebi (if you instal lthat) and/or update-manager, that is trivial
[12:58] <mvo> do you need it for hardy-final-lipa?
[12:58] <persia> lool?
[12:59] <persia> mvo: Really, it's not that I dislike synaptic, it's just that it's not there, and it wanted hildonmm :)
[13:00] <ogra> it would help me on the classmate as well
[13:00] <ogra> (to have a gdebi mode that pulls from the archive)
[13:00] <mvo> it does not need hildonmm
[13:00] <mvo> the gui uses plain gtk
[13:00] <mvo> no gtkmm 
[13:00] <ogra> even though on the cmpc g-a-i is a requirement anyway
[13:01] <ogra> so i have to keep /var on a separate partition anyway and dont run into the performance probs here
[13:01] <davmor2> dinner time talk to you latter
[13:13] <persia> synaptic is 25MB on-disk, including additional dependencies.
[13:13] <persia> Unfortunately, a simple install didn't fix Time&Date.
[13:13] <ogra> do you really have ot care fo diskspace in UME = 
[13:13] <ogra> ?
[13:13] <ogra> *to care for
[13:13] <persia> ogra: There's talk of 2GB devices.
[13:14] <persia> My MID is 4GB, but the sales people try to sell at 512MB (but that's not linux)
[13:14] <ogra> ah, well, dont let them do that ... i know how painful it is :)
[13:14] <persia> It needs a bit of time.
[13:14] <ogra> 2G are mean, not fish nor meat 
[13:15] <ogra> 4G are just perfect
[13:15] <persia> I think once storage gets a little cheaper, and more people use these things, it won't matter so much.
[13:15] <persia> Currently, to get to a price point where people != me want to buy them, they need to cut features.  In six months/a year it won't matter so much.
[13:26] <ogra> yeah
[13:27] <ogra> sadly we need to solve todays probs today :)
[13:27] <ogra> my minimal spec for cmpc *is* 2G 256M
[13:27] <ogra> which makes me cry every time i type it
[13:28] <persia> ogra: Wow!  My 3.5" is bigger than that.  I'm guessing the 256 is harder than the 2?
[13:29] <ogra> yeah
[13:29] <persia> Is the 2G at least fast-access, or filtered through a low-speed channel?
[13:29] <ogra> the 2G is fine with the new ATA devices ... the former cmpc model had the dusk on usb ... combined with no L2 cache and 256M thats just pain
[13:29] <ogra> *disk
[13:35] <stgraber> ogra: fortunately you didn't have to use FF2 with that :)
[13:36] <ogra> i did
[13:36] <ogra> for the gutsy image :)
[13:36] <ogra> which was a PITA
[13:36] <stgraber> did you put a mention "do not try to open a second tab as it'd likely crash your system (or make it so slow that you think it crashed)"
[13:38] <ogra> nah, gutsy was proof of concept
[13:38]  * stgraber still remembers of the localapps try we did with scottie :) we opened youtube, it went fine for a while and then ... no firefox :)
[13:38] <ogra> nobody expected it to actually work properly
[13:39] <stgraber> cmpc 2g + intrepid should rock with that swap-in-ram thing and other improvments (and fast I/O too :))
[13:40] <ogra> well, lets see ... i need to build on the netbook image in the future 
[13:40] <ogra> thats likely less highly optimized for the HW than my personal image was
[13:42] <persia> ogra: I thought you had a plan to base on -desktop?
[13:42] <ogra> well, netbook will be based on -desktop anyway
[13:42] <ogra> at least thats what i was told yet
[13:42] <ogra> i hope the meeting today can clearify a bit more
[13:42] <persia> ogra: Are you basing on netbook-remix, or a different netbook?
[13:43]  * ogra doesnt even know who produces the current image 
[13:43] <ogra> netbook-remix afaik
[13:43] <persia> ogra: There's a hardy remix with code at https://launchpad.net/~netbook-remix-team/+archive
[13:43] <ogra> persia, i was suggesting to use gnome for ume :) netbook will be -desktop anyway with a different launcher/desktop
[13:44] <persia> I hear it might not be getting updated for intrepid.
[13:44] <ogra> right, that means i need a lot of ppa stuff :/
[13:45] <persia> ogra: Why PPA stuff?  Can't it be done as a flavour?
[13:45] <ogra> well, i will still be bound to hardy as releasename 
[13:46] <ogra> i.e. my patches have to be against the hardy kernel ... my apps have o be synced with hardy instead of intrepid etc
[13:47] <ogra> i cant create a flavour in the hardy archive :)
[13:47] <persia> ogra: Ah.  Retrospective images.  I see.
[13:47] <ogra> :(
[13:48] <ogra> and that at least until january for four different classmate arches :/
[13:50] <ogra> urgh
[13:50] <ogra> netbook remix uses gpl3 ? 
[13:52] <lool> mvo, persia: while playing with in on menlow, I noticed update-manager will say "all updates" installed when there's no network
[13:52] <persia> Yes, and it doesn't keep track of the last update.
[13:56] <tripzero> ﻿ what is UME going to do for a gps app?
[13:59] <persia> tripzero: No selection has been made.  Have you had success with any of the various GPS apps in Ubuntu?
[13:59] <tripzero> have you looked into navit?
[13:59] <tripzero> http://navit-project.org
[14:00] <persia> My hardware doesn't have GPS.
[14:00] <tripzero> it's a little rough, but it's gtk, and it's maturing a lot
[14:00] <persia> Is it in Ubuntu?
[14:01] <tripzero> no
[14:01] <ogra> someone should packge it then :)
[14:01] <tripzero> the only one that is in ubuntu is gpsdrive, and it's not very good
[14:02] <persia> tripzero: That's likely a first requirement to get stuff into Ubuntu-Mobile.  You might want to try to get the app packaged first.
[14:02] <tripzero> persia, i've got an older version of it packaged
[14:02] <tripzero> i'll see if I can get it updated
[14:02] <persia> tripzero: That's probably a good first step.  You might work with the team in #ubuntu-motu to get the latest version packaged and on REVU for review.
[14:03] <persia> They likely won't do the packaging, but they can provide advice and help (as many people in that channel are also packaging things)
[14:04] <tripzero> persia: i'll get with the navit folks and get it packaged up then
[14:04] <tripzero> do you know if UME has selected a media player interface yet?
[14:04] <persia> tripzero: Thanks.  Having it in Ubuntu, and some reports on how well it works with the Ubuntu-Mobile environment would be the cleanest path to getting it in by default.
[14:09] <DNis> hi, anybody  as a detailed roadmap for ubuntu mobile please ?
[14:09] <DNis> i'm interested for an htc shift
[14:10] <persia> DNis: How do you mean?  There's some test images available.
[14:11] <persia> More than that is really what gets put into Ubuntu for the target architectures for the next release, and what is in the seeds.
[14:11] <persia> For the most part, we're likely to have something based on GNOME to some degree.
[14:11] <DNis>  persia : i mean a release date for an release candidate for example and release, and if there is test images for this device.
[14:12] <persia> DNis: There are no special test-images for that device.  Do you know what chip it runs?
[14:12] <DNis> htc shift run on an intel maccaslin platform, like teh samsung q1 ultra
[14:13] <DNis> with intel a110 processor
[14:13] <tripzero> persia: ﻿do you know if UME has selected a media player interface yet?
[14:13] <persia> DNis: There are some McCaslin images for testing, which might work for that.  They aren't quite perfect for my McCaslin (SR8), but they work.
[14:14] <DNis> persia: it's in the roadmap for first ubuntu mobile release but i wished i could test it :-)
[14:14] <persia> DNis: Take a look at cdimages.ubuntu.com
[14:14] <DNis> persia: thx i will take a look at it
[14:14] <persia> http://cdimages.ubuntu.com/moblin/releases/hardy/rc/
[14:15] <DNis> persia: is it simple to install ?
[14:16] <persia> DNis: Burn the image to USB, and boot off USB
[14:19] <DNis> persia: it install a grub boot loader ? we can make multi-boot ?
[14:22] <persia> DNis: I think so, but I've not tried a dual-boot configuration.
[14:24] <DNis> persia: ok thx for your answers
[14:24] <DNis> persia: it's in rc state so the release is very soon ?
[14:25] <persia> DNis: As soon as possible.  Still a few things to fix.
[14:28] <DNis> persia: ok thx a lot again
[14:29] <persia> DNis: Does it work?  Please share your experience, so we can update the wiki if the HTC Shift supported.
[14:32] <DNis> persia: mine is not already in my hand but shipping. but i will do if you want as soon as i have it
[14:36] <persia> DNis: Excellent.  Thanks.
[14:37] <DNis> persia: you are a developer on the project ?
[14:37] <persia> DNis: Yes, although that doesn't necessarily mean I'm right about everything :)
[14:39] <DNis> persia: ok ;-)
[14:53] <persia> mvo: No need to build a special mode for UpdateManager in a hurry.  synaptic will be fine (although thanks for all your help not using synaptic for update-manager)
[15:00] <lool> mvo: I don't think we need any changes from you right now; we'll just call as a backend
[15:00] <lool> (like persia just said)
[15:01] <lool> (sorry just got to the backlog of this afternoon)
[15:01] <persia> :)
[15:01] <mvo> ok
[15:23] <persia> Does anyone know where the path is getting set in the UME session?
[15:23] <ogra> if its like in desktop it should come from /etc/environment
[15:24] <persia> ogra: That's what I thought, but it doesn't.  There's a hack somewhere, which needs elimination.
[15:24] <ogra> yeah, sounds like
[15:25] <ogra> any ~/.*rc files
[15:25] <ogra> ?
[15:25] <persia> That's why time-admin can't install even when it has it's dependencies.
[15:25] <persia> Nope.  Not in ~/.?*
[15:25] <ogra> gnomerc is a typical desktop ENV mangling file 
[15:26] <ogra> /etc/bash.bashrc ?
[15:27] <persia> ogra: It's not in one of the standard places.  This is a hack :)
[15:27] <ogra> meh
[15:28] <persia> Unfortunately, unlike many hacks, `grep PATH $(dpkg -L ume-config-common)` doesn't help :(
[15:28] <ogra> grep -r PATH /etc/*
[15:28] <ogra> ?
[15:31] <persia> ogra: Thanks for pushing me to the simple, instead of my chasing the complex for hours :)
[15:31] <ogra> :)
[15:50] <lool> cgregan: What's the boot time constraint again?
[15:51] <ub511> is it possible to run ubuntu-mobile on via epia ?
[15:51] <cgregan> lool: https://blueprints.launchpad.net/astoria/+spec/3.3.7-performance
[15:52] <lool> cgregan: So we're slightly above with -18
[15:52] <lool> http://people.ubuntu.com/~lool/bootcharts/crownbeach/hardy-20080605-1.png
[15:52] <lool> Let me grab -19
[15:52] <lool> amitk: No meta for -19 yet?
[15:53] <stgraber> lool: in the new queue
[15:54] <lool> stgraber: thanks
[15:55] <lool> cgregan: NB this is with HT
[15:55] <lool> cgregan: It's probably worse without
[16:05] <lool> cgregan: Check the -3 png, it's with -19
[16:06] <cgregan> 32 seconds.
[16:06] <cgregan> that's pretty good
[16:35] <persia> mvo: I tracked down the larger issue with gst-package and synaptic.
[16:36] <persia> When the user is logged in normally, they get ENV_PATH from /etc/login.defs.
[16:36] <persia> This becomes ENV_SUPATH when they use sudo or gksu.
[16:36] <persia> Since moving to policyKit for g-s-t, /usr/sbin isn't in the path, so it doesn't find synaptic.
[16:37] <persia> I suspect other synaptic-callers might have the same issue, although I've not looked at them all.
[16:37] <ogra>  /usr/sbin should always be in the path ... thats not polkit specific 
[16:38] <persia> ogra: It's not defined in ENV_PATH in login.defs even on a standard Ubuntu Desktop install.
[16:38] <persia> Only for ENV_SUPATH.
[16:38] <persia> I thought the path came from /etc/environment, but appaprently that's not always so.
[16:39] <ogra> $PATH will override that
[16:39] <ogra> and its usually in environment ... 
[16:39] <ogra> just add the proper one there and it should work
[16:39] <ogra> else you have deeper probs
[16:39] <persia> ogra: At what point?  PATH isn't force defined anywhere in the login sequence, and login assigns either ENV_PATH or ENV_SUPATH depending on the login.
[16:40] <persia> I think there are deeper issues, as I can replicate the confusion on my desktop
[16:40] <ogra> well, just add the ubuntu standard value to /etc/environment
[16:41] <persia> ogra: It's already there
[16:41] <ogra> (note also that /etc/environment only gets read at boot)
[16:41] <ogra> re-login wont help
[16:41] <persia> Yeah, but it was there from image-creation time.
[16:41] <ogra> hmm
[16:41] <ogra> weird
[16:42] <persia> You're sure it's not with the switch to polkit?  With sudo and gksu I get the right path.
[16:42] <persia> s/with/related to/
[16:43] <ogra> well, i can call all apps with gksu here 
[16:43] <persia> I can call all apps with gksu on my image, just not as the default user.
[16:43] <persia> (not using gksu).
[16:43] <persia> When the app doesn't use policyKit, that works great.
[16:43] <ogra> do you have polkit installed ? 
[16:44] <persia> When the app does use PolicyKit, policyKit can't get the session, and so doesn't authenticat.
[16:44] <ogra> i remember you guys adding suport for consolekit
[16:44] <persia> Hmm.  I thought so.  I can at least authenticate properly.
[16:44] <ogra> do you see an active session with ck-list-sessions ? 
[16:45] <persia> Yep.  libpolkit2 & friends are installed, and ck-list-sessions shows a session.
[16:45] <ogra> active = TRUE
[16:45] <ogra> ?
[16:45] <persia> Yes.
[16:46] <ogra> and a proper x11-display-device pointing to a tty ? 
[16:46] <ogra> else polkit wont accept it as authenticated
[16:46] <persia> /dev/tty7
[16:47] <ogra> hmm, looks ok
[16:47] <ogra> http://paste.ubuntu.com/17252/
[16:47] <ogra> my session for coparison
[16:47] <persia> polkit seems to work, except that I don't get switched to ENV_SUPATH
[16:48] <persia> Only differences are that I don't have a real name, and I've been logged in 11 hours less.
[16:48] <ogra> are you in the admin group with that user ? 
[16:48] <ogra> (i would assume so since gksu works)
[16:48] <persia> Yep.
[16:50] <ogra> grep polkituser /etc/passwd ? 
[16:51] <ogra> (should be created by the postinst ... pulling straws here)
[16:52] <persia> polkituser:x:105:109:PolicyKit,,,:/var/tun/PolicyKit:/bin/false
[16:52] <lool> /var/*tun*?
[16:52] <StevenK> So typed out by hand
[16:52] <persia> /var/run (I'm not using IRC from that device, and copy&paste doesn't work there very well anyway)
[16:52] <lool> Nah, it's correct onmy desktop
[16:52] <ogra> looks right to me
[16:54] <ogra> ogra@osiris:~$ /usr/bin/polkit-auth|grep systemtools
[16:54] <ogra> org.freedesktop.systemtoolsbackends.set
[16:54] <ogra> do you get that value ? 
[16:54] <persia> Everything looks fine.  stracing time-admin shows it to be only checking ENV_PATH and not ENV_SUPATH before reporting that it can't find synaptic, and telling the user it can't install the packages.
[16:55] <persia> No!  It's empty.  That's likely it.
[16:55] <persia> What's supposed to set that, and how?
[16:58] <ogra_> so strace time-admin wasnt a good idea :P
[16:58] <persia> Ah.  It does show with --show-obtainable though :(
[16:59] <persia> And once obtained, it shows in the grep.
[17:00] <persia> ogra: strace told me why it couldn't run synaptic (shortened PATH), but not why the PATH was only ENV_PATH.
[17:00] <persia> I blame policykit, but that might be unfair.
[17:00] <ogra> right, sicne it works here 
[17:00] <lool> #startmeeting
[17:00] <MootBot> Meeting started at 11:02. The chair is lool.
[17:00] <MootBot> Commands Available: [TOPIC], [IDEA], [ACTION], [AGREED], [LINK], [VOTE]
[17:00] <ogra> it might be that gdm sets up something you dont have
[17:00] <persia> ogra: Later, and I'll look at it.
[17:01] <ogra> right
[17:01] <lool> No items from last week, moving to this week
[17:01] <lool> [topic]     *
[17:01] <lool> [topic]           netbook image classmate PC relation (who is resonsible for what, who builds which parts etc) OliverGrawert
[17:01] <MootBot> New Topic:      * 
[17:01] <MootBot> New Topic:            netbook image classmate PC relation (who is resonsible for what, who builds which parts etc) OliverGrawert 
[17:01] <lool> ogra: floor is yours :)
[17:01] <ogra> well, intel constantly asks me about specs etc in the conf calls 
[17:02] <ogra> i tried to actually find out which team is responsible over the last days and seem to run in circles
[17:02] <lool> Which specs?
[17:02] <agoliveira> ogra: You aways look like that :)
[17:02] <ogra> been pointed to ume first ... then to lexington and they pointed me to -desktop
[17:02] <lool> who pointed you for what at where? :)
[17:03] <ogra> lool, i.e. they were told at computex that the netbook iage will only exist for LTS not for later releases
[17:03] <ogra> adam le from intel was told that, dont ask me by whom 
[17:03] <ogra> so he forwarded that to me 
[17:03] <persia> The netbook-remix-image being worked on by that team is only targeted at LTS.
[17:03] <ogra> imho it would be really unfortunate to stick with hardy here
[17:04] <persia> No reason someone can't leverage that and track it for other releases.
[17:04] <ogra> well, that gets me into maitaining half of the cmpc software stack for 3 years in a ppa
[17:04] <lool> ogra: Well that's what we do so far; if however we have room for a customer interesting in intrepid support, I guess this could be discussed
[17:04] <ogra> persia, my prob is that i need to start development now ...
[17:04] <ogra> but need to know my build target
[17:05] <ogra> and i actually need an image and hardware to run it on ... but thats another story :)
[17:05] <lool> Aha
[17:05] <ogra> HW shoudl come next week ... one atom cmpc reserved for you guys
[17:05] <lool> ogra: Shouldn't this question be resolved by your management or canonical management in general rather than Ubuntu specs assignments?
[17:06] <ogra> (currently tareted to go to david, not sure thats the best though)
[17:06] <ogra> lool, well, i want to have a direct connection to my counterpart doing the lowlevel netbook stuff 
[17:06] <ogra> for tech work
[17:06] <ogra> but i cant find that counterpart anywhere 
[17:06] <ogra> thats the point of my agenda item :)
[17:07] <lool> Ok; I don't have the answer
[17:07] <ogra> well, that got clear to me when i ran through the supposedly involved teams :)
[17:07] <ogra> nobody seems to have an answer
[17:08] <lool> ogra: Raise it to management would be my personal recommendation
[17:08] <ogra> i.e. i dont even know if the actual netbook HW spec we agree on matches the minimal cmpc spec
[17:08] <lool> Dunno, I have seen neither :)
[17:08] <ogra> right, i was hoping rich would be here as well
[17:09] <ogra> well, cmpc is 2G disk, 256M ram
[17:09] <ogra> so you know 50% :)
[17:09] <lool> I'd be happy to discuss a technical decision, but I feel like you're the only one in touch with the problem at hand
[17:09] <ogra> but well, the topic cant go anywhere ... 
[17:09] <lool> Perhaps it's too early for other UME people to understand the questions at hand?
[17:10] <ogra> well, the question is simple
[17:10] <ogra> will UME build the netbook base image ? 
[17:10] <ogra> or rather "the mobile team" 
[17:10] <lool> The information I have so far allow me to tell you:
[17:10] <lool> I don't know
[17:10] <ogra> heh
[17:10] <ogra> great :)
[17:11] <lool> ogra: However, there are strategic meetings this week which might get us an answer
[17:11] <ogra> ah
[17:11] <ogra> thats a small ray of light on my horizon then :)
[17:12] <ogra> well, since we cant get more info up i think i'm done with my topic :)
[17:12] <lool> Ok; I don't think we have enough info to discuss this here and now, so I'm tempted to adjourn the topic and meeting
[17:12] <lool> I can raise it to davidm and ask for more info this week
[17:12] <ogra> yeah
[17:12] <lool> But we're busy with release
[17:12] <ogra> i thought he would be here 
[17:12] <lool> Thanks everybody
[17:13] <lool> #endmeeting
[17:13] <MootBot> Meeting finished at 11:15.
[17:13] <ogra> else i wouldnt have put it on the agenda ... i can ping you guys personally all the time :) david is harder to catch
[17:18] <bspencer__> short meeting today
[17:20] <agoliveira> bspencer__: Best type.
[17:21] <bspencer__> for me it was over before it began
[18:01] <TomaszD> hi
[18:01] <TomaszD> Invalid architecture: "lpia". Valid choices are: i386 amd64
[18:02] <TomaszD> I'm using this https://wiki.ubuntu.com/Testing/Cases/UMEinstall
[18:02] <TomaszD> can anyone help?
[18:03] <TomaszD> this appears after issuing sudo ./simple-mobile-builder kvm hardy --sources /usr/share/pdk/platforms/menlow-lpia-ubuntu-hardy-ppa/sources/hardy.list --sources /usr/share/pdk/platforms/menlow-lpia-ubuntu-hardy-ppa/sources/um-ppa-hardy.list --sources hardy-full.list
[18:03] <persia> TomaszD: Sorry about that.  You need the updated ubuntu-vm-builder in hardy-proposed.  Which arch do you have?
[18:03] <TomaszD> I'm using ubuntu amd64
[18:03] <TomaszD> ok I'll enable hardy-proposed
[18:03] <persia> NO need.  I'll get you a URL for the .dsb
[18:04] <persia> https://launchpad.net/ubuntu/hardy/i386/ubuntu-vm-builder/0.4-0ubuntu0.3
[18:04] <persia> Just download that, and you should be all set.
[18:05] <TomaszD> thanks persia !
[18:05] <persia> TomaszD: Thanks for helping to test the images.
[18:05] <TomaszD> no problem, got really curious
[18:06] <TomaszD> btw, it would be reasonable to update the instructions to include the link to this new package
[18:06] <persia> TomaszD: You could, but it ought drop into -updates in a couple days.
[18:06] <TomaszD> ahh, then it's ok
[18:07] <persia> TomaszD: Well, mostly.  Please share the URL to the deb with anyone else who has the problem.
[18:07] <TomaszD> yeah I will
[18:10] <mthaddon> hey guys, I just tried running the setup script as per https://wiki.ubuntu.com/Testing/Cases/UMEinstall and got "lpia invalid architecture" - I went into the simple-mobile-builder and changed lpia to i386 and it now seems to be building okay...
[18:11] <davmor2> mthaddon: did you follow all the steps
[18:11] <persia> mthaddon: You might not get everything you expect that way.  Try the updated VM builder from http://launchpadlibrarian.net/15004078/ubuntu-vm-builder_0.4-0ubuntu0.3_all.deb
[18:11] <mthaddon> ok, will give that a try, thx
[18:13] <persia> ogra: Looking at the GDM code, it doesn't seem to insert /usr/sbin in the PATH.  There's even a workaround in the gdmsetup .desktop file just in case /usr/sbin isn't in the path.  Any other ideas?
[18:14] <vadi2> Hi... the 'simple-mobile-builder' script fails for me (from https://wiki.ubuntu.com/Testing/Cases/UMEinstall). It says 'Invalid architecture: "lpia". Valid choices are: i386 amd64'.
[18:14] <persia> vadi2: Try the updated VM builder from http://launchpadlibrarian.net/15004078/ubuntu-vm-builder_0.4-0ubuntu0.3_all.deb
[18:14]  * persia updates the wiki page already
[18:15] <davmor2> vadi2: did you install ubuntu-vm-builder?
[18:15] <vadi2> I'm following https://wiki.ubuntu.com/Testing/Cases/UMEinstall, and it uses a simple-image-builder
[18:15] <vadi2> I installed this vm builder though. Now what?
[18:15] <davmor2> now run the script again
[18:16] <vadi2> Oh that worked, thanks
[18:20] <lool> persia: hardy-updates doesn't have lpia support I think
[18:21] <lool> persia: (for ubuntu-vm-builder)
[18:21] <lool> Unless it was added recently
[18:21] <ogra> persia, hmm, not really, i'll dig a bit
[18:21] <persia> lool: No, it doesn't, but that LP link does (hardy-proposed)
[18:21] <persia> ogra: Thanks.  I'm currently digging in policykit itself.
[18:21] <lool> persia: Ok; I was told it wouldn't be included, glad it was :)
[18:23] <persia> lool: Mind you, my bug changed from not working in -updates to working in the PPA to taking my system down hard with the current -proposed, but I've a special setup.
[18:24] <lool> How come so many people heard about simple-mobile-builder here? :)
[18:24] <lool> Oh https://wiki.ubuntu.com/Testing/Cases/UMEinstall
[18:26] <persia> lool: http://blog.qa.ubuntu.com/
[18:27] <persia> This is the call for testers for the RC, so we can get lots of eyes.
[18:27] <persia> Given that the RC won't work for most people, they will use simple-mobile-builder
[18:28] <persia> This isn't bad, because they get up-to-the-minute fixes, although the UI for creating the second test image is a little odd.
[18:29] <davmor2> lool: my fault :)
[18:29] <lool> eh
[18:29] <davmor2> check out planet ubuntu
[18:31] <ogra> sweet :)
[18:31] <davmor2> lool: hopefully get lots of eyes looking at the rc pinpointing issues and confirming :)
[18:31] <davmor2> was the plan
[18:31] <davmor2> maybe even patching too :)
[18:32] <davmor2> ogra: you like it?
[18:32] <ogra> yep
[18:32] <TomaszD> persia, now when I want to run ubuntu.kvm I'm getting:
[18:32] <TomaszD> tom@tom-pc:~/simple-image-builder/ubuntu-vm-hardy-lpia$ ./ubuntu.kvm 
[18:32] <TomaszD> open /dev/kvm: Permission denied
[18:32] <TomaszD> Could not initialize KVM, will disable KVM support
[18:32] <TomaszD> Ubuntu does not support running KVM without hardware acceleration. Sorry.
[18:33] <lool> TomaszD: adduser $you kvm; logout and login
[18:33] <lool> Or sg kvm
[18:33] <lool> Or you don't have hardware support
[18:33] <TomaszD> no,  I have a Core 2 Duo 2.2GHz
[18:34] <TomaszD> ok added myself , brb
[18:35] <davmor2> lool: I'll add that to the wiki page
[18:38] <TomaszD> ok, runs now, eats one core of my processor and nothing happens, just says "Booting from Hard Disk..."
[18:41]  * lool &
[18:41] <persia> TomaszD: That's different.
[18:42] <davmor2> it is different
[18:42] <persia> It's supposed to bring up a GUI
[18:42] <persia> TomaszD: Maybe try with qemu ?  It will be much slower, but it has a better chance of working.
[18:43] <TomaszD> wonder why it doesn't work :(
[18:43] <TomaszD> with kvm I mean
[18:44] <TomaszD> kvm should work fine
[18:44] <persia> Did it work with qemu?
[18:44] <vadi2> Where can I report issues? The ISO tracker doesn't list mobile
[18:44] <TomaszD> where do I change that
[18:45] <davmor2> it's on the wiki page TomaszD
[18:45] <davmor2> vadi2: it's on the blog third here down :)
[18:46] <vadi2> ops, thanks
[18:47] <TomaszD> same thing happens, the GUI doesn't come up
[18:48] <TomaszD> such a shame, I really wanted to try this out
[18:49] <davmor2> TomaszD: I'm going to guess that something has gone wrong then.  Try removing the folder called ubuntu-vm..... and run the script again and see if the second build works right.
[18:49] <TomaszD> davmor2, I'm going to repeat the whole procedure from the start
[18:50] <davmor2> TomaszD: good luck
[19:10] <TomaszD> davmor2, I think I know what the problem might be
[19:10] <TomaszD> the script 404s on hildon packages
[19:10] <TomaszD> Nie udało się pobrać http://ports.ubuntu.com/ubuntu-ports/pool/universe/h/hildon-desktop/libhildonwm0_2.0.11-1~svn15367-0ubuntu1_lpia.deb  403 Forbidden
[19:10] <TomaszD> Nie udało się pobrać http://ports.ubuntu.com/ubuntu-ports/pool/universe/h/hildon-desktop/libhildondesktop0_2.0.11-1~svn15367-0ubuntu1_lpia.deb  403 Forbidden
[19:10] <TomaszD> Nie udało się pobrać http://ports.ubuntu.com/ubuntu-ports/pool/universe/h/hildon-desktop/hildon-desktop_2.0.11-1~svn15367-0ubuntu1_lpia.deb  403 Forbidden
[19:10] <TomaszD> Nie udało się pobrać http://ports.ubuntu.com/ubuntu-ports/pool/universe/m/marquee-plugins/marquee-plugins_0.22-0ubuntu1_lpia.deb  404 Not Found
[19:10] <TomaszD> or 403s even
[19:10] <TomaszD> "Nie udało się pobrać" = couldn't download
[19:12] <TomaszD> hmm well the packages are there, I can download them via my web browser
[19:13] <davmor2> TomaszD: they maybe updating the packages
[19:13] <TomaszD> ok
[19:14] <TomaszD> davmor2, how would I go on by downloading these by hand and installing them inside the lpia image?
[19:15] <ogra> persia, hmm, did you ever try to build an image for vbox ? that just crashed strangely here
[19:15] <ogra> /usr/bin/ubuntu-vm-builder: line 1054: vm_target_conversion: command not found
[19:17] <TomaszD> ha, it starts! grub found the drive!
[19:17] <davmor2> TomaszD: if you can start kvm and get into shell you can try running sudo apt-get install packages that are missing 
[19:17] <TomaszD> davmor2, I'm almost there, I just don't know the login and password :]
[19:17] <davmor2> ume login pass is ubuntu
[19:18] <TomaszD> ok
[19:19] <TomaszD> omg, hildon-desktop is a metapackage, this will take a while
[19:19] <TomaszD> but it sure looks promising
[19:26] <TomaszD> how do I run the GUI from the console? startx doesn't exist
[19:27] <ogra> it should
[19:27] <persia> Now I'm really confused.  policykit explicitly sets PATH to "/usr/sbin:/usr/bin:/sbin:/bin" in several places to avoid various sorts of privilege escalation attacks.  Why would something running under policykit search the user path?
[19:27] <TomaszD> then I'm probably still missing a large piece of the install
[19:28] <persia> TomaszD: neither ubuntu-mobile nor hildon-desktop depend on an X server.  Try installing xserver-xorg
[19:28] <ogra> persia, hmm, i wonder how the env of that upstart script looks like before it does su ume
[19:29] <TomaszD> ok I will, there's a new kernel coming atm
[19:29] <ogra> persia, might be that there are some initscripts missing that need to be run before anyone logs in
[19:31] <ogra> hum, i wonder if its necessary that polkit runs *before* you catually log in :)
[19:31] <ogra> *actually
[19:31] <TomaszD> persia, installed xserver-xorg, rebooted for the new kernel, startx command not found
[19:31] <persia> ogra: You might have it.  Important services like hal and dbus aren't running when /etc/event.d/session is allowed to kick off.
[19:32] <ogra> in the desktop install it is S01policykit in rc2.d ... probably the upstart login thingie runs before 01
[19:32] <persia> policykit starts beforehand though, so I'm not sure.
[19:32] <persia> Hmm.  Right.  Maybe I need to ask somewhere else then.
[19:33] <ogra> pitti would probably know
[19:33] <ogra> he did a lot on polkit
[19:33] <TomaszD> ok nevermind, I'm still missing a chunk of -mobile..
[19:34] <persia> ogra: I need some more background on upstart first.
[19:34] <ogra> Keybuk then
[19:35] <persia> Docs first.  Then People.
[19:46] <TomaszD> sigh..
[19:46] <TomaszD> hildon-desktop, ubuntu-mobile, xorg-server...
[19:46] <TomaszD> startx doesn't bring up the GUI
[19:46] <persia> Right.  The docs explain how to set dependencies in upstrart events.  I set /etc/events.d/session to "start on started rc2", and it retains the behaviour.  Not upstart then.
[19:47] <persia> TomaszD: ume-config-common
[19:48] <TomaszD> same issue
[19:48] <persia> TomaszD: What does startx give you?
[19:48] <TomaszD> No valid fontpath could be found
[19:49]  * persia waits for bzr
[19:50] <TomaszD> I think I'll just go with the procedure for the third time, wipe everything... maybe this time it'll pull all the packages automatically
[19:50] <persia> TomaszD: Try `sudo aptitude install ubuntu-mobile ume-config-common gnome-icon-theme gnome-menus xserver-xorg-video-cirrus xfonts-base`
[19:51]  * ogra just boots a qemu image built 10mins ago
[19:51] <ogra> eeek
[19:51] <persia> You probably have most of those, but that's the PKGS hint in simple-mobile-builder that ought get you a minimal system.
[19:51] <ogra> persia, can you add -br to the startx call ? 
[19:51] <ogra> the grid hurts :)
[19:52] <ogra> trivial fix, big outcome :)
[19:52] <TomaszD> ok gnome-* and xfonts were missing
[19:53] <TomaszD> holy cow, it works!
[19:53] <ogra> gah, qemu indeed doesnt properly translate my laptop touchscreen
[19:53] <persia> ogra: What was -br supposed to do?  It didn't make a difference for me.
[19:53] <ogra> black background
[19:53] <davmor2> TomaszD: cool :)
[19:53] <ogra> its an Xorg parameter
[19:54] <TomaszD> thanks for all the help, will test this now
[19:54] <persia> ogra: Even with -br, I still get the grey grid.
[19:54] <ogra> persia, i think it needs to go behind the --dpi option
[19:54] <persia> TomaszD: Excellent.  Sorry for all the trouble getting it set up.
[19:55] <persia> ogra: I don't have a --dpi option in /etc/event.d/session
[19:55] <ogra> oh
[19:55] <TomaszD> goodness me the icons look atrocious, I bet it's the number one complaint :] they need to be svg or bigger 
[19:55] <ogra> i have it in the upstat-session script that got copied over in simple-image-builder
[19:55] <persia> ogra: Try fiddling with /etc/event.d/session  It's set to respawn, so killing X will get you a fresh login.
[19:55] <ogra> ok
[19:56] <ogra> god, qemu is slooow ...
[19:56]  * ogra wishes vbox would have worked
[19:56] <TomaszD> no kvm love?
[19:56] <persia> ogra: It's just a debootstrap and some package installs.  Surely you can do that in vbox, although maybe not with the same shell script.
[19:57] <ogra> TomaszD, sadly no ... 
[19:57] <persia> TomaszD: no login manager at all.
[19:57] <ogra> vbox is usually fine for me 
[19:57] <persia> OpenEmbedded has a QT-based environment stack, which it might be possible to port, if someone felt like doing it.
[19:58]  * persia has clearly mistaken "kvm" for "kdm".
[19:58] <persia> ogra: Have you tried kqemu-source?
[19:58] <TomaszD> persia, yeah I was about to ask what you mean
[19:59] <ogra> yeah, had that before but i decided to go back to the convenient vbox ... i mainly did ltsp stuff in vbox yet, for that the internal network functions there are just awesome
[20:00] <TomaszD> nice, my favourite game neverball is preinstalled, slideshow speed but nice to know it'll be there
[20:00] <ogra> no special config or setup needed ... just works and i'm lazy
[20:01] <ogra> hmm, if the terminal would open that would help
[20:01] <TomaszD> yeah, it doesn't
[20:01] <TomaszD> notepad app doesn't work as well, but that may be due to my incomplete install
[20:02] <TomaszD> and Galculator? Please, it's a Calculator :P
[20:02] <ogra> same for me with notepad
[20:02] <ogra> Starting Notepad .... is sitting in front of me
[20:03] <TomaszD> ok should photos and videos show a scary black screen with no content, no "Hello there, import photos" or anything like that?
[20:03]  * persia finds this odd, as this image has "Calculator" and "Mousepad", rather than "Galculator" and "Notepad"
[20:04] <ogra> i wonder if we're missing packages
[20:04] <TomaszD> I don't have a preferences menu item
[20:04] <ogra> and indeed i didnt keep a log :(
[20:04] <TomaszD> only "All"
[20:06] <davmor2> Guys quick query on file manager why are there two bars at the bottom (1 with GO and the one below with the rest of the buttons)
[20:06] <davmor2> the top bar cuts the bottom bars icon up
[20:06] <persia> TomaszD: If you pull down the menu from the top left, you ought to be able to select categories.  One of these ought be preferences.
[20:06] <persia> ogra: You can get a list from dpkg --get-selections, no?
[20:06] <ogra> if i could get a shell :P
[20:07] <TomaszD> persia, heh I know there should be more, the pulldown list only includes the All category
[20:07] <persia> Well, running stuff from a VT won't get you the environment benefits, but it will let you run CLI stuff.
[20:07] <TomaszD> my install is borked
[20:07] <ogra> well, i thought about an xterm :)
[20:07] <persia> TomaszD: Yes, your install is broken.  I'm not sure why it didn't work for you, especially twice.  It seems to work for some people.
[20:07] <TomaszD> I'll let install again while go away and do something else
[20:07] <TomaszD> *+it
[20:16] <ogra> mumble
[20:21] <ogra> meh, recovery mode doent work :/
[20:21] <ogra> *doesnt
[20:23] <persia> ogra: Right.  I'll go chase that now.
[20:23]  * persia gives up on policykit
[20:24] <TomaszD> ha, still downloading, looks like it'll be ok this time
[20:24] <ogra> persia, sulogin not installed i'D guess
[20:24] <ogra> hmm, no, its there
[20:25] <davmor2> TomaszD: I hope so, that way you'll see what we're all so happy about.
[20:26] <TomaszD> a general question though. How do you guys plan handling i18n ?
[20:26] <persia> TomaszD: Most of the applications are translated (inherited from GNOME).
[20:26] <persia> Just changing the locale ought do it.
[20:26] <TomaszD> (I'm ubuntu-l10n-pl admin)
[20:26] <TomaszD> what about hildon though
[20:26] <davmor2> ignoring it and make everyone speak english :D
[20:26] <persia> Check the contents of /etc/default/locale
[20:26] <TomaszD> and is there going to be an applet or preference for the language?
[20:27] <persia> davmor2: No, really we do have multilingual support, just not any way to select the language except from the command line.
[20:27] <TomaszD> hmm the script pulls kernel -16, but -17 is available for some time now
[20:27] <davmor2> persia: the big laughing face should of been somewhat of a give away ;)
[20:28]  * persia often misinterprets smilies
[20:30] <TomaszD> well it installed fine, I got straight into the gui
[20:30] <TomaszD> but no sign of Preferences
[20:30] <TomaszD> oh and I forgot to mention, the Chat button is missing an icon
[20:30] <TomaszD> which I presume is Pidgin
[20:31] <TomaszD> but it just sits there
[20:31] <davmor2> TomaszD: works here :-/
[20:32] <TomaszD> blimey, is my computer special, is my net connection special, what the hell is wrong
[20:33] <davmor2> TomaszD: your computer is in the wrong corner of the room ;)
[20:33] <TomaszD> I can move it, it's a laptop... :P
[20:33] <TomaszD> ok, bbl
[20:34] <ogra> persia, -br works fine for me appending it after the 96
[20:35] <ogra> exec su -l ume -c "/usr/bin/startx -- -dpi 96 -br"
[20:35] <persia> ogra: You had a --dpi to start?
[20:36] <ogra> yeah
[20:36] <persia> I'm nowhere near 96 DPI, so not at all tempted to use that.
[20:36] <persia> (and don't have the --dpi 96 in my image)
[20:37] <ogra> well, thats whats in my session script
[20:37]  * persia has a significantly stronger suspicion that the VM images do not match the MIC images.
[20:37] <ogra> yeah
[20:39] <ogra> how does your session script look like ? 
[20:40] <ogra> "/usr/bin/startx --" ?
[20:44] <persia> exec su -l ume -c "/usr/bin/startx -- -config xorg-samsungq1ultra,conf"
[20:44] <persia> Note that this isn't a Q1U, so using that xorg.conf breaks a couple things.
[20:44] <persia> My touchscreen works if I don't have the -config
[20:45] <persia> (well, calibration is broken, but that's a different issue)
[20:45] <ogra> put it in front of -config
[20:45] <ogra> exec su -l ume -c "/usr/bin/startx -- -br -config xorg-samsungq1ultra,conf"
[20:45] <persia> I did that.  No change.
[20:45] <ogra> thast pretty weird
[20:46] <persia> Yep.
[20:46] <ogra> startx definately hands over verything after -- to xinit
[20:46] <persia> Lots of things about the MIC build are weird.  Some of these weirrd things are also good.
[20:46] <ogra> which then appends it to X/Xorg as parameter 
[20:47] <ogra> works fine here ... it probably doesnt like the -config option
[20:47] <ogra> did you try adding it behind ? 
[20:47] <ogra> xec su -l ume -c "/usr/bin/startx -- -config xorg-samsungq1ultra,conf -br"
[20:47] <persia> After -config?  No.  I'll try that.
[20:48] <ogra> i doubt you use ppa X packages :)
[20:48] <ogra> so it must work one way or the other
[20:48] <persia> Nope.  Still gives the standard X screen for a bit.
[20:48] <ogra> hmm
[20:48] <persia> I'm using *lots* of PPA packages.
[20:48]  * persia is currently waiting for the PPA build of grub
[20:49] <persia> Oh.  PPA X packages?  No.  Only ubuntu-mobile PPA.
[20:50] <ogra> right
[20:50] <ogra> why the heck do you actually have to use the -config option ? 
[20:51]  * ogra doesnt see the benefit
[20:51] <persia> It has the special hacks for the Q1U, and is part of the standard MIC build for McCaslin.
[20:51] <ogra> yeah
[20:51] <ogra> but why not in the standard place
[20:51] <persia> For my hardware, it's actually a detriment, but I don't want to hack it out just to break someone else.
[20:52] <persia> I think the idea was that one would be able to set up lots of different hardware configurations in MIC.
[20:52] <ogra> right, just copy the right filein place or link it
[20:52] <persia> This might be fine for some person sitting in an office building images for all the toys on the workbench, but is less ideal for having a standard test build.
[20:53] <persia> Oh, sure, I could do that, but it'd just break again when I built the next image (I run a couple images a day).
[20:53] <persia> Same for all the other improvements you suggested a couple weeks ago ;)
[20:55] <ogra> ooooh
[20:55] <ogra> found the issue with the image builder :)
[20:55] <ogra> /usr/share/pdk/platforms/menlow-lpia-ubuntu-hardy-ppa/sources/hardy.list doesnt actually exist 
[20:56] <ogra> so the wikipage is wrong
[20:56] <persia> heh.  Try using the mccaslin lists.
[20:57] <ogra> /usr/share/pdk doesnt exist
[20:57] <ogra> wher is it supposed to come from  
[20:57] <ogra> ?
[20:57] <persia> It's in moblin-image-creator
[20:57] <ogra> ah
[20:57] <ogra> thats the issing bit 
[20:57] <ogra> *missin
[20:57] <ogra> bah
[20:58] <ogra> the wikipage doesnt talk about MIC at all
[20:58] <ogra> davmor2,^^^
[20:59] <davmor2> ogra the bit below does
[20:59] <hp2133> A friend just pointed me to here.  Any packages available yet?
[20:59] <davmor2> https://wiki.ubuntu.com/Testing/Cases/UMEinstall
[20:59] <davmor2> ogra: ^
[21:00] <ogra> davmor2, well it should tell you to install MIC at the top 
[21:00] <persia> davmor2: Right, but the point is that apparently MIC needs to be installed in order for the simple-vm-builder to pull the right feature sets
[21:00] <ogra> sudo ./simple-mobile-builder kvm hardy --sources /usr/share/pdk/platforms/menlow-lpia-ubuntu-hardy-ppa/sources/hardy.list --sources /usr/share/pdk/platforms/menlow-lpia-ubuntu-hardy-ppa/sources/um-ppa-hardy.list --sources hardy-full.list 
[21:00] <ogra> all the --sources files are missing without it installed
[21:00] <davmor2> hang on then 
[21:01] <davmor2> damn no mvo
[21:01] <ogra> and hardy-full.list should become ./hardy-full.list
[21:01] <persia> I know a bit about that code.  What did you want to ask?
[21:02] <davmor2> persia: if it could be pulled in or added to the app save having to pull in another app just for the repo list
[21:03] <persia> davmor2: Duplication of data is bad because it causes extra merge work.
[21:03] <persia> Also, I don't like feature sets anyway.
[21:04] <persia> And, given the bzr branch being pulled, that change would require mvo
[21:06] <davmor2> I guess the big problem here is that everyone who has tried it has mic installed already :(
[21:06] <persia> davmor2: Yep.  That's why we call for more testers.  We get to find out our missed assumptions.
[21:07] <davmor2> :) indeed :)
[21:07] <persia> Everyone also had ubuntu-vm-builder already installed, and we caught that quickly.
[21:07] <persia> We got the hint about /usr/share/pdk early, but failed to investigate it soon enough.
[21:09] <ogra> Get:1 http://ppa.launchpad.net hardy/main libdrm2 2.3.0.16-0ubuntu1~804um2 
[21:09] <ogra> yay
[21:09] <ogra> that looks better 
[21:17] <ogra> hmm
[21:17] <ogra> cp: cannot stat `./hardy-full.list': No such file or directory
[21:18] <ogra> seems it needs the full path :/
[21:18] <ogra> likely because it builds in TMPDIR instead of $(pwd)
[21:19]  * ogra starts another build with $(pwd)/hardy-full.list instead of ./hardy-full.list lets see if that helps
[21:26] <ogra> oooh
[21:27] <ogra> and there is a typo in hardy-full.list
[21:27] <ogra> W: Failed to fetch http://ports.ubuntu.com/ubuntu/dists/hardy/main/binary-lpia/Packages.gz  404 Not Found
[21:27] <ogra> must be http://ports.ubuntu.com/dists not http://ports.ubuntu.com/ubuntu
[21:58] <persia> So, update-manager has this nifty feature by which it downloads package changelogs from changelogs.ubuntu.com, which gets updated about every 4 hours.  Anyone have any good ideas how to display snapshot changelogs where the packages are newer than the default changelogs.ubuntu.com?
[22:05] <ogra> mvo would 
[22:07] <ogra> geez, now thats what i call an image
[22:08] <ogra> that looks way better and wuite different
[22:08] <ogra> *quite
[22:09] <persia> Right.  Did you already update the wiki?  Shall I?
[22:09] <ogra> doing now
[22:10] <persia> ogra: Thank you :)
[22:22] <ogra> persia, btw, my current image has -dpi 96 as well
[22:22] <ogra> but -.br doesnt work either
[22:23] <ogra> -br
[22:23] <persia> ogra: Hrm.  I'm knee-deep in update-manager-hildon right now, but I'll roll another image when I finish to see if I can get that.  Might be a mccaslin vs. menlow thing (I run mccaslin fsets)
[22:23] <ogra> ah
[22:23] <ogra> well, seems to have the same issue though
[22:23] <persia> -br not working means you're closer to me.  If you can figure out how to make it work, you'll make everyone happy :)
[22:23] <ogra> trying to find that 
[22:24] <ogra> probably someone added xressources files or soething differently weird
[22:24] <persia> Likely, or something odd in the selected X configuration.
[22:25] <persia> For the extra hacky bits, look at the ume-config-foo packages
[22:25] <ogra> xorg.conf looks ok
[22:27] <ogra> uuuuh
[22:29]  * persia translates "uuuuh" to "There's something mysteriously hackish going on"
[22:29] <ogra> nah
[22:29] <ogra> i just had a look at the session script 
[22:29] <ogra> that released an uuuh ... not specific to any special line or so :)
[22:29] <persia> You might want to change it to trigger on started rc2, rather than running in parallel with the rest of the rc2 startup.
[22:29] <ogra> just rebooting
[22:29] <persia> It makes it take longer to get to the desktop, but it guarantees dbus, hal, etc. beforehand.
[22:30] <ogra> ha
[22:30] <ogra> the grid comes from the framebuffer before the card gets initialized
[22:30] <ogra> its actually black with -br
[22:30] <ogra> the driver doesnt init properly
[22:31] <persia> Which driver?  -psb?
[22:31] <ogra> cirrus i think
[22:31] <ogra> whatever qemu uses by default
[22:31] <persia> cirrus
[22:31] <ogra> might as well just be vm slowness
[22:31] <ogra> but it definately turns black if the crosshair pointer is shown
[22:32] <persia> Strange.  On my hardware I get the gray screen anyway on boot.
[22:32]  * ogra tries a proper shutdown
[22:33] <ogra> do you have a movable mousepointer while you see that ? 
[22:33] <persia> Yes, the standard X pointer.
[22:33] <ogra> hmm
[22:33] <persia> (even with -br in various places)
[22:34] <persia> It only lasts about a second or two, but it's noticable.
[22:34] <ogra> i get a white screen without pointer now ... and that turns black as soon as the pointer shows up
[22:34] <persia> That sounds like the right solution.  What's your call in /etc/event.d/session?
[22:35] <ogra> so poering off the vm cleaned the videoram i assume
[22:35] <ogra> exec su -l ume -c "/usr/bin/startx -- -dpi 96 -br"
[22:36] <persia> Ah.  I've only tested reboots.  Testing a full shutdown now...
[22:36] <ogra> Öq
[22:36] <ogra> whoops
[22:37] <persia> Nah.  Oeq is a good way to put it :)
[22:37] <ogra> (that was a misfocused :q on a german keyboard in us kbd environment :) )
[22:37] <persia> : is bound to Oe?
[22:38] <persia> Or is it compound, so that O: generates Oe?
[22:38] <ogra> : is ö with us mapping on german kbd
[22:39] <persia> With a full power-down I get the black screen.  Definitely cached video-ram.  Sorry for the confusion.
[22:39] <ogra> well, i could have guessed that one earlier :)
[22:40] <persia> If I only provided enough information...
[22:40] <ogra> the photos app doesnt start apparently 
[22:42] <ogra> oh, no locale 
[22:43] <ogra> setting LANG makes it start
[22:44] <ogra> oh, funny
[22:44] <ogra> env had LANG="de_DE.UTF-8"
[22:45] <persia> ogra: Locale is now being set by MIC.  Just put something in /etc/default/locale, and you ought be fine.
[22:45] <ogra> seems simple-mobile-builder sets it to the build env locale
[22:45] <ogra> ah
[22:45] <persia> ogra: Where does it set it?
[22:45] <ogra> no idea but how else should LANG="de_DE.UTF-8" get in there
[22:46] <persia> Not sure.  Do you have an /etc/default/locale?
[22:46] <ogra> one of the tools involved must have set it unless qemu hands it over from the host
[22:46] <persia> How about a LANG entry in /etc/environment?
[22:46] <ogra> i have /etc/default/locale
[22:46] <ogra> with de_DE
[22:46] <persia> Interesting.  I'm glad it's there, but a little surprised.
[22:46] <ogra> well, tricky
[22:47] <ogra> de_DE isnt generated
[22:47] <persia> Must be set based on the build locale.  Nice touch to test multi-lingual.
[22:47] <ogra> it sets it but doesnt run locale-gen
[22:47] <ogra> which makes all the moblin apps fail
[22:47] <ogra> at least all moblin-media ones
[22:48] <persia> Ah.  That ought be fixed.  I'm not sure when/how MIC runs localegen, but it seems to work without running it manually.
[22:50] <ogra> # Keep perl inside the chroot quiet
[22:50] <ogra> HOSTLANG=$LANG
[22:50] <ogra> LANG=C
[22:50] <ogra> ha
[22:50] <ogra> /usr/bin/ubuntu-vm-builder
[22:50] <ogra> hmm
[22:50] <persia> heh
[22:50] <ogra> it seems to run locale-gen later
[22:52] <ogra> but it didnt for my image ... hmm
[22:53] <ogra> intresting, i see it generating them in the build log
[22:54] <ogra> ah i know why :)
[22:54] <ogra> my fault
[22:55] <ogra> i ran the command with LANG=C in front to get proper error messages ... the builder script copies /etc/default/locale /etc/default/console-setup and /etc/timezone from the build machine and runs locale-gen $LANG afterwards
[22:55] <ogra> bug !
[22:56] <ogra> (it should read /etc/default/locale to get the value)
[23:16] <ogra> oh, the terminal somehow sits on top of all windows now ... no way for me to get the menu open in front of it
[23:18] <persia> ogra: Click in the upper left hand corner
[23:19] <ogra> yeah, the show desktop button works 
[23:19] <ogra> but the behavior persists if the terminal win is back
[23:19] <vadi2> How can I uninstall the test enviroment?
[23:20] <persia> vadi2: How do you mean?  Just clean up from building the image?
[23:20] <persia> Or do you also want to purge all the packages you installed to make it work?
[23:20] <vadi2> ﻿persia: no, like, remove completely. I rebooted and got a kernel panic, and decided I'm not knowledgeable enough to participate in the testing if something goes wrong...
[23:20] <ogra> ah, the issue might be caused by the fact that xterm has no app menu :)
[23:21] <vadi2> ﻿persia: I'll be testing out the final thing when they ship a device with this though :)
[23:22] <persia> vadi2: OK.  Given that, I'd probably do the following.
[23:22] <persia> rm -r simple-image-builder
[23:22] <vadi2> ok, got that
[23:22] <persia> sudo aptitude purge ubuntu-vm-builder moblin-image-creator
[23:22] <persia> The rest is probably safer to leave, and you might want it later.
[23:22] <vadi2> Ok cool, thanks