[01:02] * bspencer attempts to fix hardy image failure by pushing cairo and pango source packages to PPA... === \sh_away is now known as \sh === dns_ is now known as dns53 === asac_ is now known as asac === \sh is now known as \sh_away === \sh_away is now known as \sh === \sh is now known as \sh_away === \sh_away is now known as \sh === \sh is now known as \sh_away === \sh_away is now known as \sh [16:06] hi all [16:06] nice to meet you [16:06] why no any msg left by you? === \sh is now known as \sh_away [17:01] lool: ping [17:02] sabotage: pong [17:02] sabotage: I'm in the conference call [17:02] I was looking into the blueprint for theming action from last meeting [17:02] and I find this: https://blueprints.launchpad.net/ubuntu/+spec/mobile-theming [17:02] but I can not edit it, no permission to create the blueprint [17:03] lool: sorry, get back to me when not on a call [17:04] sabotage: Are you a member of ubuntu-mobile? [17:04] yes (pretty sure) [17:04] sabotage: I think Tollef is the owner and we are stuck [17:05] Mithrandir: Can you make ubuntu-mobile owner of the specs you currently own [17:06] confirmed, I am a member of ubuntu-mobile [17:06] sabotage: probably an ownership issue [17:06] ok...I also did not have the ability to register a new blueprint so... [17:07] maybe has to do with my permissions? [17:07] sabotage: I've mailed Mithrandir to please set ubuntu-mobile as owner [17:07] lool: thanks [19:07] no bfiller ? [19:09] lool: have you or anyone on your team tried elisa on a Q1? I get a nasty message about my graphics hardware not being configured right, have you seen it? [19:10] "Your graphical hardware is not properly configured to support 3D software such as Elisa Media Center" [19:11] I have elisa-0.3.5-1, and other 3D apps work fine === \sh_away is now known as \sh [19:12] smagoun: I didn't see it [19:13] smagoun: I know some hardware isn't supported by pigment ATM with a SVN only fix, but I don't think this was associated with an error message [19:15] lool: found a workaround. Naturally it has nothing to do w/ the error message :( [19:15] https://code.fluendo.com/elisa/trac/ticket/1165 [19:16] smagoun: wow [19:17] smagoun: If you find a nicer patch, I'll bridge it to the archive or upstream; they are on #elisa BTW [19:17] s/nicer/nice [19:17] s/nice/ [19:18] lool: ok, thanks [19:53] bfiller: The problem with the mobile-basic-flash in the PPA is that the version is 0.35, which is incorrect (there is no such thing as 0.35, it hasn't been released). Sounds like there's another problem on Intel's side, there are two 0.34 tags in git (hardy-v0.34 and RELEASE_0.34) that contain very different code. [19:54] bfiller, ping [19:54] smagoun: ok [19:54] bspencer: what's up [19:54] bfiller, howdy. we finished first round of changes for freedesktop.org [19:55] bspencer: great [19:55] it makes changes and may have unexpected results [19:55] bspencer: such as? [19:55] I added two gconf keys: onlyshowin_filter and onlyshowin_ignore to help reduce the short term impact [19:56] if an application doesn't supply OnlyShowIn=GNOME;Mobile, then it doesn't appear [19:56] if it has a category that doesn't map to gnome-menus default, it will show up in "Other" [19:56] the nice "Mobile" category is gone [19:56] the default home screen now comes up to Accessories as the first thing the user sees [19:57] bspencer: Did you discuss this at FreeDesktop or GNOME? It would rock to get applications writer set this key properly in their .Desktop files [19:57] galculator doesn't have an icon [19:57] ...the list is long [19:58] bspencer: so if onlyshowin_ignore is true, then all apps will be displayed in "Other" even if they don't have the onlyshowin property set? [19:58] bfiller, not exactly... [19:58] onlyshowin_filter = bool FALSE = don't filter anything [19:59] if FALSE, then it will show all .desktop file apps. (e.g. bluetooth_analyzer, etc.) [19:59] if TRUE, it filters and you get a very reduced set [19:59] bspencer: I see [19:59] so I added the next key: onlyshowin_ignore = [list of desktop files to show regardless] [20:00] onlyshowin_ignore defaults to something like [browser, moblin-media(x3), galculator, mousepad, ...] [20:00] onlyshowin_ignore should be removed later when everyone is onboard. [20:00] open for suggestions [20:01] I kind of think that while using .desktop is nice, using OnlyShowIn is a hassle [20:01] but it is standard [20:01] bspencer: It's required for systems where you install KDE and GNOME packages for instance [20:01] bspencer: it sounds like a good solution. The key is making the .desktop files correct [20:02] one thing undefined is what to show on the first-time, because "Accessories" is lame, and "All" now shows the Preferences apps so it is too cluttered [20:02] and I was hesitant to fork the gnome_menu and add a new "Mobile" category [20:03] bspencer: What about adding a Main or Home category? [20:03] what we would need is either to have a mobile-basic-flash specific config, or add a new category group to /etc/xdg/menus/application.menu combined with .desktop file changes to add themselves to this group [20:03] lool, yes, right that is an option. [20:04] but how does an app-writer know if he should be there? That is a UI-specific decisoin. [20:04] s/UI/Home-screen implementation [20:04] bspencer: I think adding a new category group to application.menu combined with .desktop would make sense [20:05] bspencer: the same way apps have conditional mobile code for ui they could do the same for .desktop? [20:05] bspencer: Why not simply not use All but some sensible Applications combination of all categories? [20:05] That is, Home would only show the .desktop files in categories which we have decided should be on the home screen [20:06] lool, bfiller you're going in separate directions [20:06] lool's idea is easier and is mobile-basic-flash specific, true? [20:06] bfiller: I think bspencer is correct that it's hard for people to tell whether to list their app in this category [20:06] And it would be lots of patching [20:07] bspencer: We could set a per category flag whether to include it or not [20:07] lool, that would be simple enough. We could add another gconf key even with a list of Home categories. [20:07] lool: but aren't apps going to have modify their .desktop files already to add the onlyshowin property to be shown in the mobile UI? [20:08] Actually we can probably express it already; something like Settings [20:08] bfiller: But that's going to be done as part of hildonization [20:08] bfiller: While deciding whether to be shown or not by default is a bit hard to ask to an upstream [20:09] and one mobile UI might choose to only show 4 apps, while another (ours) would show all apps except Preferenes [20:09] bfiller: It might be simpler to have a specific .menu file rather than gconf keys IMO [20:09] s/bfiller/bspencer [20:09] So many b*s! [20:10] lool: as part of hildonization for .desktop files why would it be difficult to add "Category=Mobile"? [20:10] bspencer: Hmm that's a good point; but the configuration place already exists [20:10] bspencer: Let us simply define a Menu which corresponds to what to show on the home screen and this should be customized by each implementation [20:11] bfiller: That's fine, it's exactly what I agree should be done; what's hard is to ask people to add X-ShowOnHomeScreen: yes or similar [20:11] won't everyone just put that, even if they shouldn't be on the home screen? [20:11] Exactly, which is why I'm saying a .desktop file level solution isn't ok [20:12] I think the solution is to define a menu and display whatever it matches on the home screen; the home menu so to say; it could include everything except settings by default [20:12] And be customized for $ui or $customer [20:12] But it's a single file to patch [20:12] lool: I agree. I think I'm confused with the difference between Category field in .desktop file and Categories that are shown in the Mobile UI [20:12] Changing 20 apps .desktop files for a customer would be painful [20:13] bfiller: Mobile UI doesn't show categorgies but Menus (I hope!) [20:13] categorgies hmm don't want to know what I was thinking of [20:13] it does what Ubuntu does now, same categories [20:13] and follows the same rules -- if no app there, it doesn't show. If no matching category: Other [20:13] bspencer, lool : I get it now. I agree with lool's solution :) [20:14] bspencer: Categories or menus? [20:14] lool, uh... [20:15] bspencer: If you open /etc/xdg/menus/applications.menu on an Ubuntu system [20:15] * bspencer opens... [20:15] It's a tree of Menus [20:15] Defining filters on .desktop files [20:16] The filters can be based on the categories in the .desktop files [20:16] lool, oh, right. Yes -- that is the Name we show [20:16] Cool [20:16] bspencer: So simply define a Home menu, make it include everything except settings [20:16] which is a little odd sometimes. [20:16] And if someone needs to specialize, he only has to change this file [20:17] for example, if you put "Internet" in your category, it won't show up in "Internet". You have to put "Network" [20:17] > lool, define a Home menu [20:17] bspencer: The menu is the Internet menu, and it includes anything related to Network [20:17] in a separate file ? [20:17] bspencer: I don't know whether that's possible, but I suspect it is [20:17] installed separately by mobile-basic-flash. Or try to patch gnome-menu? [20:18] lool, yes I think a separate one is possible, I see settings.menu and preferences.menu there. [20:18] I'll look into it [20:18] bspencer: In Debian, we have gnome-applications.menu, gnome-screensavers.menu, etc. [20:18] yep [20:18] bspencer: I'm pretty sure there is a cascading effect of menus. I believe it will look in a special direcotry in the user's home dir [20:18] You could have mobile-ui-foo.menu which would say how menus are layered in this UI [20:19] ok. I think I have enough to go forward here. [20:19] bspencer: preferences.menu is used in System -> Preferences under GNOME [20:19] And screensavers, I'm sure you know where it's used :à) [20:19] bspencer: Cool [20:20] bfiller, let's chat about hardy [20:20] bfiller, two things: broken, and smagoun's email [20:20] I tried to build hardy image but hildon-desktop wouldn't install (last night). [20:20] something about cairo and pango version mismatch [20:20] is anyone building hardy and has this been addressed, or do I need to? [20:21] rusty and I hypothesized that if we pushed a newer source version of pango and cairo to PPA it might work. [20:21] lool, I guess I should bug you about it too. ^^ [20:21] bspencer: we build a nightly hardy image of a hardy snapshot and PPA. It did not have the problem you are referring to. [20:21] bfiller: I'm trying to do a build with mccaslin-lpia-ubuntu-hardy-ppa-snapshot but it seems to have troubles creating the project, it loops forever on a set of packages with apt-get install, have you seen this? [20:22] lool: did the nightly core image work correctly [20:22] bfiller, hm. Is the snapshot the same as I would get using the "snapshot" option in MIC? [20:22] ToddBrandt, welcome to the same conversation :) [20:22] bspencer: I'd need more info [20:23] bspencer: What APT repos are you using and what was the exact failure? [20:23] Ah ToddBrandt has the same issue [20:23] ahh, nice [20:24] ATM, I don't have a screen attached to the host where I run MIC on; if you give me the exact APT error, I could figure it out perhaps :) [20:24] incidentally, mccaslin-lpia-ubuntu-hardy-ppa installs without error [20:24] bspencer: not sure what image-creator snapshot is - we are using hardy snapshot from a few days agaon [20:24] s/agaon/ago [20:25] bspencer: we (Lexington/MSG) have our own snapshots that don't correspond to the ones from the core team [20:25] ToddBrandt, bspencer: would one of you be so kind to provide [20:25] me the error? [20:25] lool, yep, fetching... [20:25] bspencer: I'm guessing something in hardy was reve'd that is causing hildon-desktop from ppa to fail? [20:25] lool, I can give you the whole output log if you want [20:25] 1 sec [20:26] ToddBrandt: The important part is the apt-get install failure [20:26] Oh heck, I'll run image-creator in ssh -X [20:27] lool: I thought moblin should be pointing at core's hardy snapshopt + PPA? [20:27] bfiller: Did I contradict this earlier? [20:28] lool, while my hardy image builds... let me double-check something we just agreed on [20:29] if I add a new .menu file to the xdg/menus location, the only control I have is specifying which categories to show in my "Home" category. I can't specify specific apps. [20:29] lool: no you didn't, just not sure if the moblin guys are pointing at the right place [20:29] I'm grabbing the error from the log [20:30] bspencer: I don't know what the .menu files allow; perhaps you can name app? [20:30] so if a home screen wants to filter the apps, that would be something they do on their own, not dictated by us. [20:30] lool, ok, sure. [20:30] I'll look into it. [20:31] bfiller: Right; I usually double check what people are using as APT sources [20:31] bspencer: It would be part of writing the config; should be a one hour job [20:32] lool, bfiller: http://moblin.pastebin.org/25522 [20:32] I'm seeing a different error today. Something about mobile-basic-flash not being configured. [20:32] * bspencer reads ToddBrandt's log [20:33] ToddBrandt: And with what config you say this is happening? [20:33] lool: mccaslin-lpia-ubuntu-hardy-ppa-snapshot [20:33] Ah crap, I am creating a -ppa one, not a -snapshot [20:34] mccaslin-lpia-ubuntu-hardy-ppa works fine, mccaslin-lpia-ubuntu-hardy-ppa-snapshot is busted [20:35] ToddBrandt: So what happens is pretty simple: packages uploaded to the ppa are built against the latest version of hardy and this might create dependencies on versions which have been uploaded to hardy after the snapshot [20:35] ToddBrandt: The solution is to find what is required and also push it to the ppa [20:35] bspencer: oh wait, I see the error you mentioned, that apparantly happened even earlier and may have cause the pastebin error I showed [20:35] It's something which davidm and rustyl have discussed in depth I think [20:35] heya there [20:37] lool: this error happens during project creation, the one I sent before was target creation [20:37] http://moblin.pastebin.org/25523 [20:37] lool: this is the command I use image-creator --command=create-project --platform-name=mccaslin-lpia-ubuntu-hardy-ppa-snapshot --project-name=atomichardy --project-desc=test --project-path=/usr/src/projects/atomichardy [20:39] http://moblin.pastebin.org/25525 [20:39] # [20:39] Setting up mobile-basic-flash (0.38) ... [20:39] # [20:39] I/O warning : failed to load external entity "/etc/gconf/schemas/mobile-basic-flash.schemas" [20:39] # [20:39] Failed to open `/etc/gconf/schemas/mobile-basic-flash.schemas': No such file or directory [20:39] That's pretty bad [20:40] I wonder when it was uploaded.... [20:40] bspencer: Hmm that's gutsy [20:40] I haven't uploaded to hardy since we changed the schema [20:40] because it wasn't ready yet. :-\ [20:42] * lool builds for -snapshot now [20:42] bspencer: ok, this may seem wierd, but I just tried to build a mccaslin-lpia-ubuntu-hardy-ppa-snapshot project outside the Intel firewall and it worked [20:42] so how does mobile-basic-flash v0.38 get into hardy? [20:42] * bspencer is perplexed [20:42] we may have some bad links in there [20:44] ToddBrandt: It failed for me [20:44] bspencer: there is no mobile-basic-flash v0.38 in hardy or in the PPA [20:44] lool, same error? [20:44] lool: the MSG group's solution to the Hardy/PPA mismatch is to have stable snapshots of both Hardy and PPA. Both then get reved together to prevent mismatch. [20:45] http://paste.ubuntu.com/6112/ [20:46] lool, that is what I expected to see [20:46] lool, that is what I came to complain about. Not sure about the other issue. I'll look into it. [20:46] :) [20:46] perhaps cairo and pango have been updated in hardy [20:46] lool: oh never mind, I got the same error, but the project built far enough that I can build moblin-applets so I'm cool with that for now [20:47] So hildon-desktop fails to install with this combination [20:48] This is mostly thanks to GNOME 2.22 [20:49] pango1.0 1.20.0-1 was uploaded the 2008-03-11; hildon-desktop was uploaded to the ppa 2008-03-21 [20:50] When it built, it started to require libpango1.0-0 >= 1.20, but only 1.19.4-0ubuntu1 is available [20:52] I've uploaded pango1.0_1.20.0-1ubuntu1~804um1.dsc to the um ppa now; once it's built, this will be solved [20:52] Same issue with cairo, libcairo2 >= 1.5.12 is now required [20:53] 1.5.8 is available [20:53] lool, much thanks [20:54] lool: I uploaded that version of hildon-desktop that is causing the problem. I did not know to check for newer dependecies in Hardy vs. Snapshot. What should I have done differently? [20:55] should Skype run on UME? any special way to install? [20:56] cairo_1.5.14-0ubuntu1~804um1 pushed [20:56] bspencer: This is all I see for now; ping me again if you see it later on [20:56] bfiller: Hmm [20:56] lool, great. [20:56] RwL, I haven't tried Skype, but most apps I have tried do work. Give it a shot. [21:01] bfiller, I hear you have a fix in the works to dismiss the startup banner when a dialog pops up. [21:02] bspencer: yes I do. It's working on my modified version of m-b-f. I need to grab your latest m-b-f and test and make a patch. Hopefully I'll get this done this week. [21:03] bfiller, I'd like that patch. The firefox dialog is always right there behind the banner "Restore previous session" ... [21:03] bspencer: I also have a (better) fix that will solve the singleton app issue *without* having to modify the .desktop file [21:03] bspencer: I will get these both to you when they are complete for sure [21:03] bfiller: I just uploaded moblin-applets 0.49.1 with that build bug fix [21:04] ToddBrandt: great, thanks [21:04] bfiller, you mean a singleton app for any app regardless of their .desktop ? Great! [21:05] bspencer: yes, my previous fix assumed their was a valid StartupWMClass defined in the .desktop, otherwise it would not work. The new way assumes everything is a singleton. It could be modified to check for the "SingleInstance" property in the .desktop. I think this is standard.. [21:11] <\sh> guys, I need some people to test wine on lpia :) [21:12] What in wine needs to be tested? [21:13] <\sh> GrueMaster: e.g. installing an windows app and how it hits the cpu on real lpia devices [21:14] <\sh> GrueMaster: use windows rar as an example...and no I don't want to test "windows games" on it [21:14] \sh: winmine !!! [21:14] :) [21:14] No games? Awe.... [21:14] <\sh> crevette: that will work out of the box ;) [21:14] or minesweeper whatever it is called [21:14] :) [21:14] <\sh> GrueMaster: well, I don't know anything about the ram on those devices..so it's hard to tell what is a good test case ,) [21:15] * \sh needs to get hands on a nokia device [21:15] \sh: Nokia is arm [21:15] My understanding is that the actual units will have a min of 256M. [21:15] <\sh> or whatever is cheap [21:15] lpia is Intel and we have like 256MB [21:15] <\sh> and has intel in it [21:15] \sh: You can actually run lpia on a regular PC; it's 32-bits [21:15] launch your laptop with mem=256 [21:16] or in a vmware guest [21:16] I'll check the wine DB and see what apps work well. No need to test apps that don't work. [21:16] <\sh> lool: are hct vendor devices (which runs win mobile normally) a perfect candidate? [21:16] <\sh> GrueMaster: you'll find lpia packages of wine on my blog. [21:16] <\sh> GrueMaster: http://www.sourcecode.de/content/wine-low-power-intel-architecture-lpia [21:17] \sh: These are arm too [21:17] I think [21:18] <\sh> well, the fun part is to simulate a cpu with less then 3GHz ... I guess those devices are not so fast as my desktop or any vmware running on my desktop [21:19] <\sh> HTC shift is lpia... [21:20] \sh: Got it. I'll try to test it later this afternoon (it's 2:20pm now). [21:21] <\sh> GrueMaster: thx a lot...if that works..I'll need to convince people to add lpia to the P-a-s file for debian and ubuntu [21:21] \sh: There's a pas for ubuntu?! [21:22] <\sh> lool: we share it with debian [21:23] Hmm ok, didn't know that [21:23] Makes a couple of things clearer to me now [21:23] a pas ? [21:23] Packages arch specific [21:23] a file where you say what to not build on some arches [21:24] ah okay, thank you, I can back to sleep now [21:24] :) [21:25] <\sh> crevette: http://cvs.debian.org/srcdep/Packages-arch-specific?rev=1.742&root=dak&view=markup [21:27] <\sh> {elmo,lamont,infinity} needs to be bribed ;) [21:29] <\sh> hmm htc shift x9500 with umts and hsdpa for 1169 euro on amazon...my wife will kill me if I hit the "buy now" button [21:31] ToddBrandt, === danny is now known as lardbit [21:36] bspencer: howdy [21:38] ToddBrandt, what platform did you create to get the mobile-basic-flash error you posted in your pastebin? [21:38] snapshot one [21:38] mccaslin-lpia-ubuntu-hardy-ppa-snapshot [21:38] ToddBrandt, I can't reproduce it [21:38] I can reproduce it with non-snapshot gutsy image, just not hardy snapshot [21:39] did you create a target? [21:39] The error is on target creation [21:39] oh, ok. [21:39] I thought I did [21:39] yea both the project and target fail with errors [21:40] http://moblin.pastebin.org/25523 is for the project build error [21:40] http://moblin.pastebin.org/25522 is for the target error [21:41] bspencer: and actually I don't see the mobile-basi-c-flash error in either, I think lool is the only one to see it so far [21:42] * bspencer needs a clue for installing schemas [21:42] there's too many ways [21:42] and I don't do it right apparently [21:42] :( [21:45] bspencer: is your latest in mobile-basic-flash master? [21:45] yep [21:45] ToddBrandt, I have the step to install the schema when the user runs "make install" [21:45] instead of forcing them to do it with a postinst script [21:45] and I used dh_gconf [21:46] but somewhere it screwed up, so I'm redoing it [21:46] heh, I think that's why I bailed on it [21:46] I thought it was working yesterdaqy when I checked in. I verified that I could install the package and make an image too, but for some reason today it is not working [21:46] dh_gconf is a gnome thing [21:47] ToddBrandt, in schemasdir = $(GCONF_SCHEMA_FILE_DIR) [21:48] where does GCONF_SCHEMA_FILE_DIR get defined? [21:48] is that part of AM_GCONF_SOURCE_2 ? [21:48] (in my configure.ac, but not in yours anywhere I can see) === \sh is now known as \sh_away [21:49] I use the gconf-schemas command, it knows on its own [21:50] There's probably some reason for me to use dh_gconf in the rules file to install the schemas but I'll be damned if I could ever get it to work properly [21:55] error trying to install Skype: package architecture (i386) does not match system (lpia) [21:56] I downloaded the Ubuntu 7.4+ deb... [21:56] should I be installing it via apt-get instead? [21:57] RwL, :( You'll have to tweak the source to get it to install [21:57] because that package is blocking based on arch check [21:58] RwL: or just manually install forcing the arch if the idea is just testing. Should work the same. [21:58] RwL: dpkg --force-architecture IIRC [21:58] aha! thanks, will check & try [21:59] ToddBrandt, if I put postinst in debian directory, do I need to do more to get it to run ? [21:59] or is just having it sufficient ? [22:01] ToddBrandt, how can I remove a schema from my gconf to reset the system ? [22:01] bspencer: once you call gconf-schemas --register, it's registered, every user on the system will then get those settings by default [22:01] there changes get put in their .gconf dir [22:02] yeah, I copied your pattern, put the gconf-schemas in a postinst file [22:02] agoliveira: that bypassed the architecture check but there are apparently dependency problems too. should I just try to get those via apt-get? (libqt4-core and libqt4-gui were the ones listed) [22:02] do I need to call that somewhere (I don't see it anywhere so I assume that it just gets run) [22:02] I just remove them [22:02] ToddBrandt, right, how? [22:02] however the correct way to do it is with the --unregister [22:03] gconf-schemas --unregister [22:03] I think I should fix that in my postrm [22:03] hm,... postrm :) [22:03] yep, if someone can explain to me why that's bad I'll change it [22:03] but until that day, it works so well [22:04] ToddBrandt, but why is OLD_DIR=/etc/gconf/schemas and not /usr/share/gconf/schemas ? [22:04] RwL: Yes, if you solve the dependencies manually it should install [22:04] because that's a bug I think, I need to use --unregister [22:05] it's probably leaving those things on the system after uninstall, which is bad [22:05] got it [22:05] I'll try [22:07] argh! I just now realize that horace made changes last night [22:07] thus the broken schema. Mine was working! :( [22:08] bfiller: I added the moblin-settings-xsettings.c code to moblin-settings-daemon, however in order to ensure I've done it properly can you tell me what you use to test? [22:09] bspencer: ahhhh, so then it shall me I who copies your code :D [22:10] bfiller: and if it's on hardy I have no way to test that at the moment, I can't create a target image with MIC [22:11] * ToddBrandt just realizes that bfiller isn't even here [22:11] kyleN: I added the moblin-settings-xsettings.c code to moblin-settings-daemon, however in order to ensure I've done it properly can you tell me what you use to test? [22:24] ToddBrandt, gconf-schemas --unregister doesn't remove them when I run gconf-editor [22:24] * bspencer looks at how to restart gconf... [22:25] gconf-editor only looks at the changes, if the user has changes made to those keys then they'll stay [22:25] ToddBrandt, oh. do I remove ~/... something? [22:25] * bspencer looks [22:26] bspencer: you have to remember that gconf is actually a joke on us linux guys by the windows developers who created the registry. They're laughing at us right now... [22:26] ToddBrandt, if it worked like the registry I'd have no complaints. (sad but true) [22:26] bspencer: worst case you could delete the directory path under /root/.gconf and /home/ume/.gconf but that's getting hacky [22:27] is there a non-hack way? I just want ot verify that my install actually creates the schema correctly in my target [22:27] but I can't get rid of the old schema [22:27] Actually, why don;t you use gconftool-2? [22:27] ok. I'm usin git [22:27] I gconftool-2 --shutdown [22:27] you can run that as whatever user, then tell it to kill those keys [22:27] still has schema [22:28] bspencer, run man gconftool-2 [22:28] --recursive-unset [22:28] Recursively unset all keys at or below the key/directory names on the command line. [22:29] there's also --unapply-schema which II've never tried [22:38] kyleN: I added the moblin-settings-xsettings.c code to moblin-settings-daemon, however in order to ensure I've done it properly can you tell me what you use to test? [22:39] ToddBrandt, needless to say, none of those commands gets rid of my gconf keys :( [22:40] but I'll just make a new target and try it there [22:41] I think user gconf changes aren't meant to be removed [22:41] I think gconf removal is impossible [22:41] yea [22:41] :) [22:41] which seems wrong [22:41] but there it is [23:05] ToddBrandt, bspencer: MIC should be able to go further with the -snapshot platform [23:05] lool: did you fix it? [23:06] Yes, well the uploads I did to the ppa fixed it [23:07] ahh, kewl [23:07] I'll try it now, thanks! [23:08] ToddBrandt: It will probably fail in the q1full fset subsequently though; it just did for me [23:08] How can I help fix that? [23:08] i.e. how do I figure out which packages to upload from the errors, and then how do I upload them? [23:08] ToddBrandt: So for instance, the error I now get is: [23:09] http://paste.ubuntu.com/6123/ [23:09] ToddBrandt: I then chroot /srv/chroots/mccaslin-lpia-hardy-ppa-snapshot/targets/q1ufull/fs /bin/bash [23:09] Reproduce with apt-get install ubuntu-mobile [23:09] oh lord, moblin-applets is breaking it [23:09] Then I try "apt-get install ubuntu-mobile cheese" [23:10] and I see: [23:10] The following packages have unmet dependencies: cheese: Depends: libebook1.2-9 (>= 2.22.0) but 2.21.92-0ubuntu1 is to be installed [23:10] At which point it's clear that libebook1.2-9 is too old now [23:10] yea, that's the about-=me dialogs dep [23:10] So I'll snapshot eds [23:10] It's going to take a while [23:11] wait, so I need to upload the latest ebook to hardy? [23:11] can you explain what snapshot eds does? [23:11] * ToddBrandt admits his ignorance [23:12] ToddBrandt: Oh nothing in particular, I took EDS from hardy, and pushed it to th eppa [23:12] With a proper changelog entry and version [23:12] I'm uploading evolution-data-server_2.22.0-0ubuntu2~804um1.dsc now [23:13] When it's built, we should see the install continuing [23:13] ahh, so that should take care of the libebook dep [23:13] Yes [23:13] would I have access to upload things like that? [23:13] i.e. could I be of help instead of just someboedy whining? [23:13] Sure, you could do it as well [23:13] Anybody in ubuntu-mobile can [23:13] gotcha [23:13] Another fix is to take a newer snapshot [23:14] oh yea, with dput [23:14] And the proper fix is to build against the snapshot, but this requires a separate archive [23:15] lool: ok, so I find the hardy package that will satisfy the dependency, then I doanload it, pull out its source, and create the .source/.build.dsc files that are needed for spu, then I dput it to moblin-ppa? [23:15] spu = dput [23:15] Yes [23:15] gotcha [23:15] If you look at my cairo of pango1.0 uploads, you should see the delta with hardy is only the changelog [23:16] https://launchpad.net/~ubuntu-mobile/+archive/+builds => building now [23:16] ok