[02:37] kyleN, [02:38] bspencer: yes? [02:38] to clarify the theme issues -- when you change the theme (somehow) a gconf key changes. But the UI doesn't [02:38] is that correct? [02:39] bspencer: mildy [02:39] and you are expecting us to change something so that you can just change the gconf key and the theme auto-switches magically. [02:39] no [02:39] ok. :) We're looking through bugs here. [02:39] wait [02:39] let me give you a URL [02:39] yes, when we change the gconf key, we want the theme to change properly [02:39] right now, some of it doesn't change but it should [02:40] what doesn't change is the panels in the marquee [02:40] https://bugs.launchpad.net/moblin-multimedia/+bug/206389 [02:40] what /does/ change? [02:40] they fails to detect the gconf key change, fail to switch to the new theme dir [02:40] what DOES change is gtk apps. they pick up the new gtkrc files in the new theme denoted by the new key [02:41] even without restarting them? [02:41] since marquee panels are (or should be) gtk apps, they should change too [02:41] yes, they do it dynamically [02:41] coool [02:41] for example, go to ubuntu desktop and change /desktop/gnome/interface/gtk_theme key manually to another valid theme dir [02:42] and all gtk apps maigcally redraw [02:42] I have that working in UME [02:42] except the darn marquees don't redraw [02:42] marquee panels, that is [02:43] does the mobile-basic-flash change? [02:43] (the icon backgrounds) [02:43] well, it is a flash movie, so no. We will actually change our gconf key from a user action in our flash movie [02:44] and by the way, the hicolor icon theme dir does not seem complete [02:44] hm. [02:44] and it is the default, so if you launch a gnome app that requires a valid gnome icon theme, it borks [02:44] I don't remember messing that too much [02:44] could be a corner case, but try installing pcmanfm and you will see [02:45] then change the icon theme to the "gnome" one via gconf key, and pcmanfm no longer borks [02:46] hm. I forget how the icon theme cascades... but I thought it would try hicolor last. I'd have to look that up again. So the "gnome" theme must have the icons and it doesn't fall back to hicolor [02:46] perhpas the fallbacks aren't configured correctly, I haven't checked [02:47] kyleN, did you check this bug: https://bugs.launchpad.net/moblin-multimedia/+bug/206389 [02:47] is it resolved? [02:47] looking now [02:47] working late? (9:47pm) [02:47] on a deadline [02:48] network is incredibly slow. am thinking smagoun just did something very BIG [02:48] and then cut and ran [02:49] ok i see it [02:49] here's the deal [02:49] we think moblin should provide a theme switching solution that works [02:49] however, that bug was erroneously pegged to moblin because it is really that our customer's flash icon for theme switching was not hooked up to anythin [02:50] so, I think the issue of you foks getting theme switching to work, including the marquee, and probably creating a UI for it, is still on the table [02:50] kyleN, yep, got it. [02:51] ok [02:51] I'll look at marquee and the UI is still pending agreement :) [02:51] cool [03:03] lool, telepathy-haze was removed from UME recently? [03:06] bspencer: telepathy-* was pulled out of being installed -- it drags in pidgin-data, which conflicts with pidgin-maemo, which is in the PPA. [03:09] true. [03:09] StevenK, I notice that I can still install pidgin-data manually [03:09] then everything works. [03:09] we updated the fsets to not include telepahty as it was breaking image creation [03:10] but... ubuntu-mobile-dev pkg depends on telepathy-devel [03:10] and I can't fix that. :) [03:10] so it is not possible to install the developer tools fset currently, unless you first manually install pidgin-data (an obscure link) [03:12] Argh [03:12] bspencer: But pidgin-maemo-data and pidgin-data conflict [03:12] So you can install it since you don't have pidgin-maemo-data installed? [03:13] In terms of ubuntu-mobile-dev -- good catch, I'll fix that now. [03:27] StevenK, ah -- I don't know about pidgin-maemo-data -- but I saw something about you guys making pidgin-maemo your default chat solution [03:27] should be cool. [03:28] and they are a lot more likely to sue Nokia first and not try to go after you guys [05:47] asac, StevenK [05:47] I'm working on the xulrunner upgrade [05:47] I installed xulrunner-1.9-dev, but where do I get xulrunner-gtkmozembed-embedding ? (configure complaining) [05:48] -PKG_CHECK_MODULES(MOZEMBED, mozilla-gtkmozembed >= 1.8.1) [05:48] +PKG_CHECK_MODULES(MOZEMBED, mozilla-gtkmozembed-embedding >= 1.8.1) [05:49] You don't, the pkgconfig name changed [05:49] bspencer: Give me a moment, I'll look it up [05:49] ok. [05:52] Hmm. mozilla-gtkmozembed-embedding should work [05:52] I was using libxul-embedding, if I recall correctly [05:53] I didn't see that as part of Hardy [05:54] at least the pkg name libxul-embedding [05:55] Ubuntu package, or pkg-config package? [05:57] when I grep /usr/lib/pkgconfig files, I have xulrunner-gtkmozembed... but nothing with "embedding" [05:58] hold on... maybe I have it.. [05:58] (hardy-lpia)root@liquified:~# dpkg -S /usr/lib/pkgconfig/libxul-embedding.pc [05:59] xulrunner-1.9-dev: /usr/lib/pkgconfig/libxul-embedding.pc [05:59] yep [05:59] so why [05:59] checking for MOZEMBED... configure: error: Package requirements (xulrunner-gtkmozembed-embedding >= 1.8.1) were not met: [05:59] No package 'xulrunner-gtkmozembed-embedding' found [06:00] (hardy-lpia)root@liquified:~# pkg-config --exists libxul-embedding && echo 'yes' [06:00] yes [06:00] oops -- s/xulrunner/mozilla [06:00] well, I ain't the sharpest marble [09:39] StevenK: awake? [10:23] asac: Hrm? [10:25] StevenK: mail send ;) [10:25] StevenK: got it? [10:29] I see it. [10:29] I'll test it tomorrow [10:30] thanks [10:31] maybe it didn't work for us, because there is something else broken in our packaging? [10:31] StevenK: ? [10:31] e.g. missing data files or something? === davmor2 is now known as davmor2_away === davmor2_away is now known as davmor2 [14:00] anybody tried to create menlow fset? [14:00] I mean to add [14:03] I get: Couldn't find any package matching "xf86-video-psb linux-menlow-default-2.6.22 psb-kmd-default-2.6.22 moblin-chat" [14:04] of course that xf86 is xorg-...psb [14:04] but the linux-menlow-default is definitely not in the repositories [14:05] persia lool any idea why? === asac_ is now known as asac [14:10] this is related to MIC [14:10] For xf86-video-psb, I suspect xserver-xorg-video-psb is preferred. I'm not sure about the others. [14:10] of course [14:11] but why don't we have the fset corrected in the moblin-image-creator.deb :) [14:11] persia do you think it makes any sense to add the moblin.org repositories? [14:11] to get the menlow kernels from intel? [14:11] :D [14:12] theseinfeld: I'm very much not a fan of mixing repositories. It may be that the image-creator needs tweaking to support different repos in the cases where core contents differ significantly [14:12] me neither [14:13] but there is no menlow in the ume for now... persia, do you know why? [14:13] That I don't. I'm looking at the MIC fset definitions now. [14:14] I didn't check the crownbeach [14:15] persia the poulsbo is also out :( [14:16] hey...i am looking at setting up a lpia VM image ..is there some pointers for doing this (like can i use QEMU or do i need kernel virtualisation?)..i saw lool was doing something with this? [14:17] ian_brasil: Both QEMU and kernel virtualisation will work for lpia emulation on another machine. You are most likely to have difficulties with the default video drivers available for UME, but if you work around that, it ought work. [14:18] persia: thx a lot [14:18] lpia is pretty much i386 :) [14:19] ian_brasil: I've added support to ubuntu-vm-builder in bzr to accept --arch lpia and the like [14:19] Something like ubuntu-vm-builder qemu hardy --arch lpia should do [14:20] I really need to finish these bits [14:20] lool: you rock :) [14:21] Well you wont get UME at this point! [14:21] why not? [14:22] It just builds a regular VM [14:22] ala debootstrap [14:22] but i will get an lpia chroot so i can run tests , right? [14:22] What remains to be done is to extract the list of packages from MIC, wrap all of this in a new script and add new packages for qemu config [14:22] ian_brasil: Yes, a lpia vm even [14:23] ian_brasil: I personally use "kvm" instead of qemu and run it with a -vnc :12 or so [14:23] And then connect to it via vnc [14:25] i will have a play and write up on the wiki how i get on then.. [14:25] persia, any ideas where the menlow are gone? [14:25] theseinfeld: Looks to me like the current fset definitions assume the moblin repo, at least in terms of package names. I'm not sure whether you can mix repos. [14:26] well, I mixed them in the past and ended up with dependecy hell [14:28] That would be about what I expect. Unfortunately, I'm not sure enough of the correct package names for menlow/poulsboro to upload an updated MIC with confidence. Maybe someone else understands it better. [14:28] I can give you the package names [14:29] linux-menlow-default linux-menlow-default-2.6.24 [14:29] linux-menlow-developer-2.6.24 [14:30] These are moblin names or ubuntu-mobile names? [14:30] psb-kmd-default psb-kmd-default-2.6.24 psb-kmd-developer-2.6.24 [14:30] moblin names [14:30] there are no ubuntu names since they are in the fset but not in the ppa/ports repos [14:30] :( [14:31] persia what do you think is the best way forward with this menlow thing? [14:32] That's the part I'm unsure about. [14:32] I am mostly interested in getting the poulsbo video kernel modules [14:33] The brute force method is to grab from moblin, verify packaging, and push to ubuntu-mobile. [14:33] if that is not in UME we don't have the next intel platform ready, so there is no point of talking anything but the Q1 [14:33] :D [14:33] On the other hand, I'm not sure if that's actually the best way to get something nice. [14:33] Well, it works on the SR8, basically (and other mccaslin devices). [14:34] yeah, old generation mccaslin devices but not the new menlow to come [14:34] :D [14:34] something that moblin guys are having now [14:35] Right. The better way is to review the new packages, and add them selectively as one understands the integration. This likely takes longer, and requires a bit of testing (preferably with the right hardware). This is especially true for driver-type stuff. [14:36] how about then just getting them in as they are in moblin and later include them? [14:36] test and include them... [14:37] s/xf86-video-psb/xserver-xorg-video-psb/ is something that was done before (for 0.39), but I'm not finding lots of info about the psb stuff, and I'm very much not tempted to touch the kernel without hardware. [14:37] you don't have that hardware from Intel? [14:37] can't you get it from them? [14:38] The moblin repos ought be mostly source-compatible. You might try a rebuild of their packages to test your hardware: if they work, that's a good argument for upload. [14:39] theseinfeld: Not I: I get my hardware retail :) [14:39] afk 30 seconds [14:40] persia i have to get something [14:47] persia, me back [14:48] Get:1 http://www.moblin.org gaston/main linux-menlow-default-2.6.24 2.6.24-11 [13.5MB] [14:48] :D [14:48] that's not a rebuild... [14:49] theseinfeld: You'll get a better test getting the source... [14:49] :D [14:49] and compile it? [14:49] yes [14:49] using chrooted? [14:50] or pbuilder, or sbuild. that's the idea, yes. [14:50] hmm... [14:50] ok [14:50] no problem [14:50] I can do that [14:50] i have my lpia pbuilder [14:50] if it works...what then? [14:51] then take the .deb from wherever you set the results directory to be, and install it on the hardware. [14:52] And if that works, see if you can determine what patch enables the construction of that binary package, and whether it can be applied to the Ubuntu linux package, At that point, you can file a bug :) [14:53] persia, do you think that they have a patch for that? [14:53] uhm... [14:53] theseinfeld: By "patch" I mean "change to the code". [14:53] yeah... I might end up comparing the whole kernel tree :) [14:53] lets see... [14:53] I let you know one of those days if this works [14:53] :D [14:54] and bug...etc... [14:54] so I will be back with those issues [15:12] is the next month deadline still feasable? [15:13] s/feasable/feasible/ [15:13] I mean, if there is no menlow... [15:13] :D [15:13] ? [15:39] persia: poulsboro, how cute :) [15:40] bye [15:41] did I say pousboro? [15:41] poulsboro? [15:41] I must be dead tired... [15:41] good night... [15:41] * theseinfeld out... [15:44] lool: i ran ubuntu-vm-builder kvm hardy --arch lpia and i get invalid architecture ..is there something i am missing? [15:47] ian_brasil: You need to build the package out of bzr [15:47] It wasn't uploaded (too close to release I guess) [15:48] ah ok... [15:48] ian_brasil: https://code.launchpad.net/~soren/ubuntu-jeos/trunk [15:49] ian_brasil: bzr branch lp:~soren/ubuntu-jeos/trunk; cd trunk; debuild -i [15:49] And you should get a .deb; install it and it should work [15:49] nice one === matt_c_ is now known as matt_c === robr__ is now known as robr === robr___ is now known as robr [18:03] ¡¡ agoliveira is giving a session in #ubuntu-classroom !! [18:08] Hi, when we will be able to install UME on Nokia N8xx ? [18:08] Probably never. [18:09] read the faq [18:09] okei === robr___ is now known as robr [18:44] wasabi: You never know. Mojo's Hasty Hippogriff ought contain many of the UME packages. Something may occur. [18:44] Yeah. I know. I hope it does. [18:44] But I wouldn't count on it. :0 [18:44] Bunch of closed stuff on the nXXX's [18:46] Yeah, well, it's the primary mojo target. [18:56] QUESTION : can the "desktop manager" app (despite of the touchscreen feature) be used on common desktops? [18:58] hi agoliveira! [18:58] pablom: Sorry but what desktop manager? Matchbox? It's the one used on UME. [18:58] thanks! [18:58] phoenix24: Hi. [18:58] and matchbox can be used with ubuntu desktop? [18:59] sorry for the english [18:59] agoliveira: thx will the log are published the usual place ? [18:59] pablom: Yes it can but it's not very apropriate as it it has some limitations like working on fullscreen only. [18:59] ok [19:00] rZr: I gues so, it's my first time doing that :) [19:00] what get my attention is that it is very user friendly [19:00] thanks a lot [19:00] pablom: That's the idea: to have something very userfriendly [19:00] agoliveira: i am frustrated i came at the end, any other one planned soon ? [19:01] I'll try it and give a look at the source too [19:01] i am talking about UME or embedded stuff [19:01] agoliveira:What about marketing? Are there efforts underway to push/assist UME to product manufactures? For example, I read somewhere that Nokia is interested. [19:01] forgot to say, im brazilian too [19:01] :) [19:02] rZr: Not currently but you bother jcastro for it :) [19:02] mybunche: There's a team called MSG that does this. I work with the cor4e team. [19:02] pablom: Bem vindo! :) [19:02] agoliveira:thanks [19:03] mybunche: No problem. [19:03] agoliveira, valeu. I did a modification to xfdesktop to make some things more user friendly, and I think matchbox-desktop is what Im looking for [19:03] btw do you have interessing pointer about optimizing filesystem for saving flash memories usage ? [19:04] rZr: Yes, that's a good point. [19:05] Great work. Looking forward to the future of UME. It 2am here, work tomorrow, need some sleep. Cheers! [19:05] i am working on this for tuning my "live system" on USB device [19:05] BTW, in about 2 weeks we will have the UDS (Ubuntu Developers Summit) If you have interesting ideas, you can drop by or contact us top bring them in and have them discussed for the next release. [19:06] ok [19:07] mesh networking is something to be tested w/ mobile computing , any plans about this ? [19:08] rZr: Not for now but it's an interesting idea. [19:08] I know that intel have mesh working with classmatepc, another project [19:08] it's more at the application level than the OS though [19:08] pablom: the OLPC project has been critized about this recently too [19:09] agoliveira: are there connections w/ OLPC project too, i know some debian friends are on it [19:10] rZr, good to know, mesh is a great feature on these cases [19:10] rZr: Not this project, no. [19:10] pablom: let me find you this article [19:10] rZr, thanks [19:12] * rZr is doing archeology on his wiki :) [19:13] gotcha : http://www.usenix.org/event/upsec08/tech/full_papers/patterson/patterson_html/ [19:15] rZr, i ll read it! thanks a lot [19:15] you're welcome [19:15] dinner time now [19:15] I'll hang around [19:16] i'm going now too, see ya [19:19] agoliveira: trying to run hildon-desktop I get http://pastebin.com/m3da6da93 [19:21] Riddell: How did you try it? Anyway DBUS is not running, you have to start it first. [19:22] agoliveira: clicking Terminal in image-creator, export DISPLAY=:1 (my xephyr) and running hildon-desktop [19:22] I did run dbus-launch, but that didn't hepl [19:22] help [19:24] Riddell: Try ume-xephyr-start [19:25] Riddell: Close your Xephyr first [19:26] agoliveira: did something, but not there yet http://pastebin.com/m192029c9 [19:28] Riddell: This is really weird. Did you run xhost + in another console to allow Xephyr to connect to the host? [19:29] yes [19:31] I had best results running Xephyr as my regular user, and attaching from within UME by starting dbus as root, su'ing to ume, setting the display, and launching. [19:32] persia: launching what? [19:32] * persia hunts down the script name [19:34] persia: That's how I do it too [19:34] Riddell: start-hildon from https://wiki.ubuntu.com/MobileAndEmbedded/HildonDesktopManualProcedure This is a deprecated procedure, so it may not be ideal. [19:34] start-hildon I think [19:34] How much time would have been saved would we have a solution to build vms, I wonder [19:35] /usr/bin/start-hildon: line 28: /usr/sbin/mas: No such file or directory [19:35] is about all that does [19:35] At least days * num_developers [19:35] I thought that was fixed. Hmmm... [19:36] lool: +1 [19:37] Riddell: Try /usr/share/ume-config-common/ume-gui-start [19:38] Hmm start-hildon must have been broken for a good while, wow [19:38] persia: http://pastebin.com/m343568c7 [19:39] Riddell: And you've set DISPLAY to match your Xephyr? [19:40] What happens if you try to e.g. start kdm in the Xephyr from your regular system? [19:40] persia: yes (xclock works fine) [19:41] But not matchbox? [19:41] matchbox? [19:41] I've never managed to get kdm to work in xephyr [19:41] Riddell: I think your KDE instincts don't let you run something based on Gnome :) [19:42] Riddell: Try stepping through the commands individually. Perhaps starting with /usr/bin/matchbox-window-manager (with lots of options). [19:43] This might help track down the source of the issue. [19:43] persia: matchbox-window-manager does seem to be running [19:43] but nothing is showing in xephyr [19:44] Riddell: OK. From your log, I'm fairly sure sapwood-server is also running. [19:44] it is [19:44] This just leaves /usr/bin/hildon-desktop Does it work if you call it directly? [19:45] Wait. Matchbox is running, but nothing is showing in Xephyr? The WM ought load something. [19:45] running hildon-desktop http://pastebin.com/m476b97af [19:45] xclock now loads full screen [19:46] xclock being full screen would be an indicator that matchbox is working. [19:47] Right. Seems you're having trouble connecting to your dbus. When I encountered that, I rebooted. You may be able to just kill some processes, but I remember dbus not liking to be killed. [19:48] Well dbus itself probably doesn't mind too much, but the apps connected to it... lalala [19:48] Let's just say they share the joy [19:49] can't I just restart dbus in my chroot? [19:49] lool: Isn't it supposed to run a separate dbus in the chroot? My memory of dbus not being easily killed comes from schroot 0.52, where it would leave extra dbus sessions running from every build. [19:50] persia: It should, but as I bind mount everything e.g. /tmp to access my host Xorg... [19:50] Riddell: You might give that a try. While nm is being most talkative about it, I suspect the "WARNING: not connected" followed by the segfault is the same thing (only by less defensive programmers) [19:51] lool: Ah. Yes. Too much bind-mounting can cause confusion. This is why I like virtualisation and extra hardware. [19:51] One of the devices I've never found, but want is a 5.25" bay computer: install two or three in a tower as test boxes without disturbing the primary workstation. [19:54] persia: Interesting [19:54] I think I'm more happy with vms which I can pack in my laptop [19:54] Let's run VMs on the Q1! [19:55] lool: Maybe. The issue there is resource contention. Depends on the use case, I suppose. I tend to break things. [19:56] I see [19:58] Riddell: I can replicate the issue. On the other hand, I also replicated the fumbling with the order of commands, which may be related. I'll try a couple other things, and let you know the procedure. [20:08] lool: Did you ever have success with getting apport to work? I can replicate, but I can't see the core, so have a little trouble figuring out why it crashes. [20:08] persia: Yes, it's fixed [20:08] persia: I've prepared a SRU [20:08] persia: naturally, retrace wont work with ppa packages [20:08] lool: OK. It's in -proposed? [20:08] persia: Don't think it was accepted to proposed yet [20:09] persia: It's in the bzr branch though [20:09] I'm not worried about retrace. I just want to have a core file somewhere, and I'm not getting it by default. [20:09] ~ubuntu-core-dev/apport/ubuntu [20:09] persia: The issue is with unionfs [20:10] With getting no core file? If I turn off apport, should I just get the normal ones then? [20:11] persia: Oh, I don't know about real core files [20:11] I thought you meant crash files from apport [20:12] They might be turned off when apport is setup [20:12] Ah. That's the thing I want. apport is just sometimes one way to get it. [20:12] But otherwise, check ulimit -c [20:12] persia: Here, core files seem disabled by default [20:12] (0) [20:12] ulimit -c unlimited would fix it I guess [20:13] Yeah. That was my issue. On the other hand, It's kinda late here, so I'm not surprised to be missing things :) [20:13] -rwxr-xr-x 1 ume ume 798720 2008-04-29 19:13 core [20:24] Right. I'm not up to digging the stack deeply on this now, but it's an issue with a bad "level" being passed to brightness_from_level in libbrightness from moblin-applets. Be good to get an apport bug filed against it for proper investigation later. [20:25] Riddell: My apologies, but I'm not likely to get you a working recipe just now. === cprov is now known as cprov-afk === rZr is now known as RzR