[00:50] <JontheEchidna> something whispered a sweet-nothing to my MBR sunday morning
[00:50] <JontheEchidna> on the bright side, the 12.04.1 alternate CD works, as did an upgrade to quantal
[05:54] <Quintasan> mikhas: ping
[05:55] <mikhas> Quintasan, pong
[05:56] <Quintasan> mikhas: You have inquired about maliit packaging, right? I just moved house and I'm currently moving stuff around. The framework packaging is almost done
[05:56] <mikhas> ok
[05:56] <Quintasan> I didn't touch maliit-plugins module yet
[05:56] <Quintasan> framework packaging requires review and it should be ready for inclusion
[05:56] <mikhas> right, for the plugins, I think it is important to make it work with all the CONFIG+=enable… features
[05:57] <mikhas> and you might want to get someone who could look over the theming
[05:57] <Quintasan> I'll keep that in mind. I'll link you to the packaging branches when I'm done
[05:57] <Quintasan> mikhas: theming?
[05:57] <mikhas> Quintasan, to make the keyboard look blend into plasma
[05:58] <Quintasan> Yeah but how does that work in regard with packaging?
[05:58] <Quintasan> There are some themes that I can package or something?
[05:58] <Quintasan> Or you're suggesting we get someone to design a theme?
[05:58]  * Quintasan just woke up and his brain is not working 100%
[05:59] <mikhas> Quintasan, https://gitorious.org/maliit/maliit-plugins/blobs/master/maliit-keyboard/data/styles/README
[05:59] <mikhas> Quintasan, well, you will 1) have to add a new styling profile for plasma, 2) make it the default profile when compiling plugins
[06:01] <Quintasan> oh
[06:02] <Quintasan> mikhas: In other words we need to have a plasma profile and I need to set it up as the default profile
[06:02] <Quintasan> not even other words
[06:02] <Quintasan> I seem to get it
[06:02] <mikhas> :-)
[06:02]  * Quintasan bookmarks
[06:03] <Quintasan> I still have to unpack stuff though
[06:03] <mikhas> sure, take your time
[06:03] <mikhas> best to keep the bug report up-to-date
[06:03] <mikhas> we do this for Kyle ;-)
[06:32] <pkt> anyone familiar with ubiquity kde frontend? I 'm trying to debug bug #1055967
[07:32] <asac> hi. any idea what happened to twinkle in quantal?
[07:50] <Daviey> asac: removed due to being dead upstream, following debian bug 673900
[08:44] <Quintasan> yofel: sup with the certifciate?
[08:44] <Quintasan> Riddell: http://www.hardkernel.com/renewal_2011/products/prdt_info.php
[08:44] <Quintasan> Did we look at those?
[08:45] <yofel> Quintasan: should be the same, but I had to restore my server from backup after a disk failure
[08:45] <Quintasan> Quassel says it changed
[08:45] <Quintasan> Well, I'll accept it
[08:46] <Quintasan> I don't think I store any confidental info in my IRC :P
[08:52] <Riddell> Quintasan: I've not looked at them, what for?
[08:54] <Riddell> pkt: hmm, drat
[09:01] <Quintasan> Riddell: I rembember there was discussion if we are ever buying developer boards and if yes then what boards
[09:13] <jussi> Quintasan: I think a nice shotgun approach would be good...
[09:14] <Quintasan> jussi: shotgun approach to what? :P
[09:14] <jussi> 2 of those, 2 of those, 2 of these :P
[09:14] <Quintasan> oh
[09:14] <Quintasan> :O
[09:14] <jussi> Quintasan: dev board buying'
[09:15] <Quintasan> christ why is this board crashing
[09:16] <Riddell> Quintasan: I have a pandaboard, having a developer isn't too interesting compared to having an actual tablet or other hardware to run it on
[09:16] <Quintasan> True that
[09:16] <Riddell> but of course the problem is knowing if the kernel will run at all
[09:16] <Riddell> which is what vivaldi have been fighting since april
[09:17] <Quintasan> Now, find a tablet that has hardware acceleration and works with Kubuntu :D
[09:17] <jussi> Quintasan: theres a bunch of intel ones :P
[09:17] <Quintasan> jussi: You sure we can like, install quantal on them and I will have GLES acceleration working?
[09:18] <jussi> think so, iirc.
[09:18] <jussi> or at least some version of ubuntu
[09:18] <Quintasan> That would be like, the device we've been looking for?
[09:18] <sladen> any idea where I can find "Lars Knoll" on IRC?
[09:20] <jussi> they are very pretty though..
[09:21] <jussi> Quintasan: intel isnt quite arm... they are just a standard atom processor 
[09:23] <Riddell> sladen: lars on #qt-labs, what are you after?
[09:24] <sladen> Riddell: git blame suggests Jiang Jiang and Lars Knoll
[09:25] <sladen> Riddell: everything tagged as Lars was just a mass import?
[09:25] <sladen> Riddell: unless everything tagged as Lars was just a mass import?
[09:25] <Riddell> I don't know
[09:32] <pkt> Riddell: hi
[09:33] <pkt> drat?
[09:44] <Riddell> pkt: I'm just about to test the beta 2 candidates
[09:46] <pkt> I see
[09:47] <pkt> Riddell: did you see the bug I reported?
[09:47] <pkt> I don't think it needs any further logs, it is deterministic and the only helpful line in the logs is the one I pasted
[09:48] <pkt> there is also a permission denied error but it is bogus
[09:49] <pkt> it is for /proc/*/auxv IIRC
[09:49] <pkt> The real problem seems to be as mentioned, that something is trying to access the session bus while running as root
[09:50] <tsimpson> to whomever designed ubiquity: you SUCK
[09:51] <xnox> tsimpson: please use polite tone.
[09:52] <Riddell> pkt: what's the bug number?
[09:52] <tsimpson> xnox: can't too many hoops to get it in gdb
[09:52] <pkt> Riddell: bug  #1055967
[09:52] <xnox> tsimpson: debug-ubiquity. Plus why would you want gdb for a pure python app.... I have no clue.
[09:53] <xnox> tsimpson: debug-ubiquity at the kernel boot prompt.
[09:53] <tsimpson> xnox: because the crash is not in python
[09:53] <xnox> also not sure what gdb has to do with ubiquity desing.
[09:53] <tsimpson> and I'm not in an installer
[09:54] <xnox> tsimpson: than it's not ubiquity, but just a regular live-session of ubuntu/kubuntu/whatnot
[09:54] <Riddell> pkt: drat, you're right
[09:54] <pkt> Riddell: to the extent of my abilities I would be happy to help fix this problem
[09:54] <tsimpson> xnox: nope, running "ubiquity" is certainly ubiquity :)
[09:54] <xnox> tsimpson: try different cd boot options (esc at boot splash) and see how far you get.
[09:55] <xnox> tsimpson: ubiquity-dm != ubiquity
[09:55] <Riddell> this didn't happen on the images I tried on friday
[09:55] <xnox> by default it launches ubiquity-dm desktop environment with a greeter, but you can bypass it to boot into live session if you press esc at the boot splash and choose to....
[09:56] <pkt> Riddell: this is good info because it narrows it down
[09:57] <pkt> we only need to figure out what changed between friday and monday then
[09:57] <pkt> especially in core kde that would affect ubiquity
[09:58] <xnox> Riddell: you tried precise-daily images on friday, correct? (but is not in quantal)
[09:58] <xnox> Riddell: you tried precise-daily images on friday, correct? (bug is not in quantal)
[09:59] <pkt> yes it is precise, not quantal
[10:05] <tsimpson> xnox: I'm not on a live session, or an installer, I'm running "ubiquity" from an install. and I'm not talking about the UI, I mean the design of the software
[10:06] <xnox> tsimpson: ubiquity is not meant to be launched from normal/installed system.
[10:06] <xnox> =/
[10:06] <tsimpson> xnox: yes, I know
[10:06] <xnox> the same reason why udebs were not meant to be installed on the normal install.
[10:07] <xnox> you will have more luck launching it in a vm or lxc container/chroot.
[10:08] <tsimpson> I am in a VM
[10:12] <Riddell> pkt, xnox: mm no, I'm trying quantal images
[10:14] <Riddell> xnox: logs on bug 1056079
[10:16] <pkt> If I remember correctly this bug has happened again in the past
[10:16] <pkt> and some tricky code was added to save user ids and such
[10:16] <pkt> you can see the comments in kde_ui.py
[10:17] <pkt> unfortunately I didn't have time to debug it more
[10:17] <pkt> (to see why these protection measures fail)
[10:19] <tsimpson> seems that it's missing the DBus session address in the environment
[10:28] <pkt> tsimpson: it looked to me more like that it is trying to access session bus from code running as root
[10:28] <pkt> but exactly where it is doing this I didn't have time to figure it out
[10:29] <tsimpson> in the KApplication constructor
[10:29] <pkt> wasn't this fixed?
[10:30] <pkt> I thought it was doing a trick with saving its "rootiness" and becoming user before this constructor
[10:30] <pkt> for this reason
[10:31] <pkt> It is really great that you guys are looking into this in any case
[10:31] <pkt> I will try to join the fun after lunch
[10:34] <tsimpson> apparently, it's failing again
[10:44] <Riddell> hum, I'm pretty stuck for ideas on how to investigate this
[10:44] <Riddell> weird that it should affect both precise and quantal at the same time
[10:45] <pkt> so it affects quantal as well?
[10:45] <pkt> I haven't downloaded the quantal daily (not much bandwidth here)
[10:47] <Riddell> yes
[10:48] <pkt> do you think colin watson would be interested
[10:48] <Riddell> I expect so
[10:48] <pkt> I think he was the one that fixed the similar problem the previous time
[10:49] <Riddell> there was a dbus security update recently in both precise and quantal
[10:49] <pkt> aha
[10:49] <pkt> this could well be it
[10:49] <pkt> essentially ubiquity was exploiting a loophole
[10:51] <pkt> I wonder how this could be verified
[10:52] <pkt> maybe if we installed newer dbus in 12.04.1 livecd and see if ubiquity breaks?
[10:52] <xnox> ubiquity-dm is the thing that starts dbus before lauching QtCore to pain the wallpaper
[10:53] <xnox> s/pain/paint/
[10:53] <kubotu> xnox meant: "ubiquity-dm is the thing that starts dbus before lauching QtCore to paint the wallpaper"
[10:53] <Riddell> yeah the problem goes away using the previous build of dbus packages
[10:53] <pkt> at least this gives a clue :)
[10:54] <Peace-> hi
[10:55] <pkt> The essential question is whether the thing ubiquity was doing is legitimate (so the problem is the dbus fix) or not (which means ubiquity needs rewrite :P)
[10:57] <xnox> and since it didn't break != kubuntu images, maybe there is something different in the default user setup?
[10:58] <pkt> I think it is just because the gtk frontend somehow didn't need this trick
[10:59] <pkt> maybe "KApplication" is too much for something like an installer and it should be written on top of just Qt instead
[10:59] <pkt> but I doubt this will work for LTS
[11:00] <Peace-> Riddell: http://www.youtube.com/watch?v=wAB1whqH9n8
[11:05] <Riddell> Peace-: that's what we do for the default plasma set up in kubuntu
[11:07] <Peace-> well i did the snippets for plasma Riddell
[11:07] <Peace-> it was funny 
[11:21] <pkt> I think the "right fix" for precise is to take back the dbus patch
[11:21] <pkt> and fix the 2-3 vulnerable applications instead
[11:22] <pkt> (spice, xorg, etc we don't have pam_systemd I think)
[11:23] <pkt> reducing the functionality of dbus without warning doesn't look like LTS material
[11:38] <pkt> Btw, I can also confirm this for precise,
[11:38] <pkt> I installed the latest libdbus on top of 12.04.1 livecd
[11:38] <pkt> and boom goes ubiquity ...
[11:39] <pkt> It makes sense if you read what the patch is doing
[11:41] <Riddell> mm
[11:42] <pkt> In debian that patch for CVE-2012-3524 wasn't accepted I think
[11:43] <pkt> note the last phrase ;)
[11:51] <Riddell> 12:48 < mdeslaur> Riddell: they said that, and then decided that the best fix would be in libdbus itself
[11:51] <Riddell> 12:49 < mdeslaur> Riddell: this is an upstream patch, if you don't fix it now, you'll have to fix it soon when the new dbus hits
[11:51] <Riddell> says the security guy who added the patch
[11:52] <pkt> I see
[11:53] <pkt> The 100% right fix should probably be to have the parts of ubiquity that run as root in a different process
[11:54] <pkt> this way there would be no such dbus-related mess
[11:54] <pkt> but for precise would this be possible?
[11:57] <dpm> Hi Riddell, all kubuntu-docs templates have now been approved in LP. Do we have someone who'll take care of fetching the translations from LP and reuploading the package with them on the translation deadline day?
[11:58] <Riddell> dpm: I'll do it if Darkwing isn't back online by then
[11:58] <dpm> cool, thanks Riddell
[12:22] <pkt> In my personal kubuntu derivative I will solve this problem for now in the livecd-building code I think
[12:24] <pkt> I will install with "old" dbus and I will let the upgrade during the install to bring the new one
[12:24] <pkt> this seems like the easiest workaround
[12:38] <Riddell> pkt: what's your derivative?
[12:52] <pkt> Riddell: it is an educational project
[12:53] <pkt> it is like a framework for building ubuntu flavors starting from seeds
[12:53] <pkt> without need for debian-cd or ubuntu-cdimage that make my eyes bleed :P
[12:53] <pkt> I only use live-helper from debian
[12:54] <pkt> Of course this means no alternate cds for me right now but I don't care
[12:55] <pkt> I already have some code at http://github.com/squibbylinux but it is in a very alpha state right now
[12:56] <pkt> In the beginning I thought that the ubiquity bug was my own fault and this ate 1 day from my time :P
[12:57] <pkt> The example distribution is based on kubuntu because I love KDE and the work you guys are doing in kubuntu :)
[13:02] <yofel_> shadeslayer: ^
[13:14]  * shadeslayer looks
[13:14] <shadeslayer> pkt: I can most likely help
[13:14] <shadeslayer> looking at your live build stuff
[13:14] <pkt> shadeslayer: cool :)
[13:15] <pkt> the live-build patches are just bug fixes
[13:15] <pkt> I tried to get them to ubuntu but nobody cared :P
[13:15] <shadeslayer> actually, I'm not sure what you're trying to accomplish, so reading the backlog first :)
[13:15] <shadeslayer> and you want live-build patches upstreamed in debian
[13:16] <shadeslayer> there's #live-build on OFTC
[13:16] <pkt> no
[13:16] <shadeslayer> ugh, I mean, #debian-live
[13:16] <pkt> I think they are already in debian
[13:16] <shadeslayer> oh ok
[13:16] <pkt> I 'm just basing my work on precise so I backported
[13:16] <shadeslayer> ahhh
[13:17] <pkt> What I 'm trying to accomplish
[13:17] <pkt> is to build a livecd that is as close to official kubuntu (at first) as possible
[13:17] <pkt> but without debian-cd / ubuntu-cdimage
[13:17] <shadeslayer> fairly simple
[13:17] <shadeslayer> :)
[13:17] <pkt> only with live-helper
[13:17] <pkt> and after this is done
[13:17] <shadeslayer> yeah, I've accomplished that for quantal
[13:18] <pkt> I can start changing stuff :P
[13:18] <pkt> cool, do you have your stuff somewhere?
[13:18] <shadeslayer> yes, but it's a private repo at the moment :P
[13:18] <pkt> I see, no problem
[13:18] <shadeslayer> sec
[13:19] <shadeslayer> pkt: you want ubuntu-defaults-builder
[13:19] <shadeslayer> that has a nifty helper script that uses live build
[13:19] <pkt> I tried this one first
[13:19] <shadeslayer> ah ... and?
[13:20] <pkt> https://github.com/ubuntu-gr/ubuntu-defaults-el-gr
[13:20] <pkt> I got disappointed
[13:20] <pkt> too many bugs and limitations to fight against
[13:20] <pkt> So I decided to just copy the interesting parts
[13:21] <shadeslayer> nope, what I meant was : /usr/bin/ubuntu-defaults-image
[13:21] <shadeslayer> look at that
[13:21] <pkt> Yes, this is what I did
[13:21] <pkt> I got the interesting parts from that script
[13:21] <pkt> the "build" you see in squibby-build
[13:21] <pkt> and the stuff under auto/
[13:21] <pkt> are inspired from this code
[13:21] <shadeslayer> oh .. I just customized it to my requirements a bit
[13:21]  * shadeslayer looks
[13:22] <shadeslayer> https://github.com/squibbylinux/squibby-build/blob/master/squibby-builder/build < this right?
[13:22] <pkt> yep
[13:22] <shadeslayer> sed -i "s/^\\(LB_SYSLINUX_THEME=\\).*/\\1\"kubuntu-$SUITE\"/" config/binary < is wrong
[13:23] <shadeslayer> oh wait, not it's not
[13:23] <shadeslayer> ok, looks fine
[13:23] <shadeslayer> pkt: what's the issue with building from that script?
[13:23] <pkt> it is because ubuntu-defaults was meant for ubuntu
[13:23] <shadeslayer> you also want to set some things like PROJECT=kubuntu
[13:23] <pkt> why so?
[13:24] <pkt> it isn't used anywhere
[13:24] <shadeslayer> it is :D
[13:24] <pkt> you forget that I don't use the ubuntu-defaults scripts
[13:24] <shadeslayer> ahhh you completely used your own auto/config
[13:24] <pkt> yep
[13:25] <pkt> I wanted something very simple to build upon
[13:25] <shadeslayer> pkt: and you're building in a precise chroot/precise install ?
[13:25] <pkt> shadeslayer: yes
[13:25] <pkt> After I get the basics working, I will add some complexity back again
[13:26] <pkt> like support for PROJECT
[13:26] <pkt> because a friend of mine wants to use it for Greek schools
[13:26] <pkt> and they use ubuntu and gnome-fallback instead of KDE
[13:26] <pkt> but as they say "first make it work"
[13:27] <shadeslayer> thing is, ubuntu-defaults-builder has some things that are needed to build stuff .... and you might have missed something vital
[13:27] <shadeslayer> what I did was to add a project in auto/config and then set my PROJECT var
[13:27] <pkt> Well, I have a working livecd right now
[13:28] <pkt> so I don't think I missed anything
[13:28] <pkt> I even have some of the functionality from cdimage
[13:28] <pkt> if I missed something it must be fairly minor
[13:29] <shadeslayer> ok ... so ... you need my help with?
[13:30] <pkt> I don't know, maybe you have things that I 've missed
[13:30] <pkt> the main things missing right now
[13:30] <pkt> are EFI support
[13:30] <shadeslayer> ah
[13:31] <pkt> and making a package list automatically starting from the seeds
[13:31] <pkt> i.e., hacking germinate output into a package list
[13:31] <shadeslayer> right, for seeds, you want to read https://wiki.ubuntu.com/SeedManagement
[13:31] <pkt> I 'm also building a version of wubi but this is also pretty trivial job
[13:32] <shadeslayer> but the gist of it, everything in live will end up on the live cd but not on the install
[13:32] <pkt> mainly search/replace because wubi has hardcoded its distros
[13:32] <shadeslayer> everything in desktop will end up on the install
[13:32] <pkt> yes, I got this about live and desktop
[13:33] <pkt> I already have the package lists
[13:33] <BluesKaj> Hi all
[13:33] <shadeslayer> cool
[13:33] <pkt> just for now I built them manually
[13:33] <pkt> I just need to automate this
[13:33] <pkt> In general it is going well, I have no complaints :)
[13:33] <shadeslayer> ok, in order to generate the meta package, look at kubuntu-meta
[13:33] <pkt> yes
[13:34] <shadeslayer> it has a update script that you can run
[13:34] <pkt> I have squibby-meta :)
[13:34] <shadeslayer> which is what I do for my meta package
[13:34] <pkt> it is in the repo you saw
[13:34] <shadeslayer> ah looking
[13:34] <pkt> I have a ppa for all these things
[13:34] <pkt> and for the seeds
[13:34] <shadeslayer> looks good
[13:34] <shadeslayer> yeah, that's what I do
[13:34] <pkt> I prefer git so I keep them at github
[13:35] <pkt> but I have bzr branches because germinate likes them
[13:35] <pkt> and they just pull from git automatically
[13:35] <shadeslayer> right
[13:35] <pkt> (I like launchpad)
[13:35]  * shadeslayer has everything on github as well
[13:35] <pkt> great :)
[13:35] <shadeslayer> EFI support is fiddly
[13:35] <Darkwing> dpm, Riddell, I'll be online by then and I'll get that done. 
[13:36] <Darkwing> Riddell: Thanks for getting those up.
[13:36] <pkt> shadeslayer: I think it shouldn't need much
[13:36] <dpm> awesome, thanks Darkwing
[13:36] <pkt> I already know how to set up qemu to test it
[13:36] <Darkwing> dpm: Not a problem. :D
[13:37] <pkt> shadeslayer: so, it is basically a little hacking in the "binary" stage I think
[13:37] <shadeslayer> kinda, afaik the ubuntu images have a efi/ folder
[13:37] <pkt> yes
[13:37] <pkt> so the problem is basically to include that as well
[13:37] <shadeslayer> so you can extract that folder and put them in your binary includes
[13:37] <shadeslayer> and it'll land up on your ISO
[13:37] <pkt> yes, essentially that's it 
[13:38] <shadeslayer> however, I'm not sure if that even works
[13:38] <pkt> we will see
[13:38] <shadeslayer> atleast it didn't work on my Macbook Pro
[13:38] <pkt> hmm I think EFI for macs is different
[13:38] <pkt> this is why there is a different image for them right?
[13:38] <shadeslayer> aye, but that doesn't really work :P
[13:39] <shadeslayer> the way I do it is to make a efi/boot folder on my USB
[13:39] <pkt> I don't care Macbooks :P
[13:39] <shadeslayer> then loop over the ISO
[13:39] <shadeslayer> and boot that
[13:39] <shadeslayer> needs a special partitioning scheme and what not
[13:39] <pkt> I see
[13:39] <pkt> I would happily leave this for later
[13:39] <pkt> at first, Qemu PC is my target
[13:40] <shadeslayer> yeah, I don't recommend putting EFI boot as your priority ;)
[13:40] <pkt> There would be no point
[13:40] <pkt> Large part of the motivation is to teach in my LUG how a distro is being put together
[13:41] <shadeslayer> as for WUBI, from what I remember, the config script had a SUBPROJECT variable
[13:41] <shadeslayer> and you could set SUBPROJECT=wubi and it worked
[13:41] <pkt> So, if I focus on such details it would be a distraction
[13:41] <shadeslayer> but WUBI was not a priority for me, so I didn't look into it much
[13:41] <pkt> Yes, it is not a priority for me either
[13:41] <shadeslayer> :)
[13:41] <pkt> First is to fix this annoying ubiquity issue
[13:42] <shadeslayer> which is?
[13:42] <pkt> so that I can have "version 0.1" :P
[13:42] <pkt> just a sec to find the bug ID
[13:42] <pkt> bug #1055967
[13:42] <pkt> This one
[13:43] <shadeslayer> ahh ok
[13:43] <pkt> "Luckily" it isn't a problem with my scripts :P
[13:43] <shadeslayer> can't find a session bus
[13:43] <shadeslayer> how odd
[13:43] <pkt> we traced it
[13:43] <pkt> It is the last dbus update that causes it
[13:44] <shadeslayer> yeah, a bit odd that it doesn't find the sessions bus :)
[13:44] <pkt> It isn't odd
[13:44] <pkt> The problem is that the dbus security update
[13:44] <pkt> it essentially forbids using libdbus if you are seteuid
[13:44] <shadeslayer> well it's a bug, but why did dbus doesn't have a session bus
[13:44] <shadeslayer> ahh ok
[13:45] <pkt> It has a session bus
[13:45] <shadeslayer> but doesn't allow you to access it
[13:45] <Riddell> pkt: cjwatson asks about testing this http://paste.ubuntu.com/1226425/
[13:45] <pkt> it is ubiquity's problem for running kde code via sudo :P
[13:45] <shadeslayer> heh
[13:45] <Riddell> pkt: but I couldn't get it to work, are you able to test it?
[13:46] <pkt> Riddell: I can try, but I will need 10-20 minutes
[13:47] <Riddell> please do
[13:48] <shadeslayer> w00t, I get to be a crew member at UDS
[13:49] <pkt> Riddell: why is the patch for ubiquity-dm?
[13:49] <pkt> it doesn't work when I call ubiquity from terminal either
[13:49] <pkt> does ubiquity-dm run in this case?
[13:50] <Riddell> pkt: ubiquity-dm runs from the live cd
[13:50] <pkt> yes I 'm using a live cd to test
[13:51] <Riddell> to start X and the initial use of ubiquity
[13:51] <pkt> first it breaks at ubiquity-dm
[13:51] <pkt> but then if I go to desktop
[13:51] <pkt> and call ubiquity from there
[13:51] <pkt> it breaks again
[13:51] <pkt> so I 'm thinking that the problem is not just in ubiquity-dm
[13:51] <Riddell> yes, I agree
[13:51] <xnox> see ubiquity-wrapper and the funky way sudo is used when running under kde.
[13:52] <xnox> and related comments.
[14:05] <pkt> Riddell: I tried to test in any case by making ubiquity-dm the default-display-manager
[14:06] <pkt> but this doesn't work
[14:06] <pkt> when I disconnect I get kdm
[14:09] <pkt> I will try to build a livecd with the patch
[14:13] <shadeslayer> !find cairo.h
[14:16] <shadeslayer> heh
[14:16] <shadeslayer> http://opendesktop.org/content/show.php/kded-appmenu?content=153882&PHPSESSID=0752e1424180bef0b9f467bf04053917 < top comment
[14:19] <Riddell> shadeslayer: yeah we should get it packaged, it'll probably replace agateau's menu stuff at some point
[14:19] <shadeslayer> aye
[14:20] <shadeslayer> I kind of like my menu's that way
[14:20] <shadeslayer> except in some apps where I don't use a menu at all
[14:32] <shadeslayer> *sigh*
[14:38] <Riddelll> shadeslayer: why the long sigh?
[14:38] <shadeslayer> source files without headers
[14:38] <shadeslayer> license headers I mean
[14:39] <shadeslayer> why the extra 'l' :P
[14:39] <shadeslayer> now my brain keeps saying Riddellllllll
[14:40] <highvoltage> scumbag brain.
[14:40] <shadeslayer> :D
[14:41] <Riddelll> it's so I don't get confused with Riddell 
[14:42] <shadeslayer> ah :)
[15:07] <kubotu> ::qt-bugs:: [744812] FontConfig/Qt stack choke on Ubuntu Medium font meta-data (No medium in Inkscape and too bo... @ https://bugs.launchpad.net/bugs/744812 (by Francois Thirioux)
[15:51] <genii-around> A fellow in #k wants to contribute as a developer, anywhere particular i should point him at?
[15:51] <Tm_T> here in general I suppose
[15:51] <n0ti0nis> hello
[15:51] <Riddelll> hi n0ti0nis 
[15:52] <Tm_T> welcome (:
[15:52] <n0ti0nis> I'd like to contribute as a developer, I suppose a I need a mentor :)
[15:52] <n0ti0nis> Thank you! :)
[15:53] <Riddelll> n0ti0nis: what kind of stuff are you into?  we need packagers, testers, writers, coders etc etc
[15:53] <n0ti0nis> Mainly I'm a developer, I know few things on packages, but I'm here to learn
[15:54] <n0ti0nis> I know Python, C++, C, C#, PHP
[15:54] <n0ti0nis> I love QtCreator of course :D
[15:57] <Riddelll> n0ti0nis: first thing is to hang around here, this channel is where most of the devel chat happens
[15:59] <n0ti0nis> ok
[16:00] <n0ti0nis> is there a tasks list for beginners?
[16:00] <Riddelll> there are bugs http://goo.gl/ESmab  and features http://goo.gl/ZGGJP
[16:01] <Riddelll> today I'm looking at a very fiddly and annoying bug with the installer ubiquity and sudo passing environment variables
[16:02] <pkt> me too ...
[16:02] <pkt> sorry for taking so long with the testing btw
[16:03] <n0ti0nis> interesting
[16:04] <pkt> yes, join the fun :P :)
[16:04] <n0ti0nis> :D
[16:05] <Riddelll> n0ti0nis: there's packages that need updated, owncloud-client would please Blizzz 
[16:06] <n0ti0nis> uh I'll look at it
[16:07] <pkt> Riddelll: FWIW colin's patch doesn't seem to help
[16:07] <pkt> it seems ubiquity-dm now dies just there is no dr konqui
[16:07] <pkt> so it stays with a black screen at boot
[16:11] <Riddelll> pkt: yeah I reach the same conclusion
[16:12] <pkt> it is just so nasty
[16:13] <pkt> even if we read the code
[16:13] <pkt> it seems to me no amount of messing with uids can help
[16:13] <pkt> because the CVE patch doesn't include only the code for "mutilating" libdbus when seteuid
[16:14] <pkt> it also includes advanced code to detect if you are really seteuid or not
[16:14] <Riddelll> pkt: yeah all the stuff with uids in ubiquity is horrible
[16:14] <Riddelll> I have ubiquity running now by converting it from a kde app to a qt app
[16:15] <pkt> but now that the detection is strong I think that all this trickery became useles
[16:15] <pkt> yes
[16:15] <pkt> this is what I think the solution would be
[16:15] <pkt> at least this way there won't be as much dbus stuff happening behind your back
[16:15] <Riddelll> not very satisfying but.. yeah
[16:16] <pkt> I think there is a reason the gtk interface is gtk and not gnome
[16:16] <shadeslayer> 0.o
[16:16] <shadeslayer> that statement makes no sense to me
[16:17] <pkt> I would expect the same type of problem would happen with gnome as well
[16:17] <pkt> but I could be wrong as well, I don't know its internals well
[16:18] <pkt> The problem is libraries/support code doing d-bus calls as part of their initialization
[16:18] <pkt> for a process that wants to run as root this is not so nice
[16:19] <pkt> switching from KApplication to QApplication saves from this problem
[16:20] <pkt> shadeslayer: does it make a little more sense now?
[16:22] <n0ti0nis> owncloud needs libopts25 libqt4-core libqt4-gui ntp, by the way everythinf seems fine
[16:23] <shadeslayer> pkt: not really :P
[16:23] <pkt> ok then maybe I 'm just talking nonsense out of frustration :P
[16:24] <shadeslayer> or maybe it's because I'm a bit sick and am just having difficulty trying to understand what you're trying to say ;)
[16:24] <Riddelll> n0ti0nis: it's an old version we have in the archive
[16:25] <pkt> shadeslayer: we have a problem with the ubiquity kde interface
[16:25] <pkt> it is written in python (like rest of ubiquity)
[16:25] <shadeslayer> ok ....
[16:25] <pkt> like every kde application it uses KApplication class
[16:26] <shadeslayer> right
[16:26] <pkt> when we construct our KApplication it forces us to be running as root
[16:26] <pkt> (because it won't let us become root later)
[16:26] <pkt> but
[16:26] <BluesKaj> , stuff to do for a while , BBL
[16:26] <pkt> As part of its initialization, there are calls to the session bus
[16:27] <pkt> and this causes the grief
[16:27] <pkt> at least this is roughly how I understand this
[16:28] <shadeslayer> oh ok
[16:28] <Riddelll> it's even more complex because it gets root through sudo then drops back to user again
[16:28] <shadeslayer> @_@
[16:28] <pkt> yes, but this is just trickery
[16:28] <Riddelll> this trickery confuses the various security mechanisms
[16:28] <pkt> it is trying to "hide its rootiness" in an attempt to deceive libdbus
[16:29] <pkt> and that worked until now
[16:29] <shadeslayer> how does the gtk stuff do it?
[16:29] <pkt> gtk is just a windowing toolkit not a desktop environment
[16:29] <pkt> so it doesn't do session bus calls behind your back in general
[16:29] <pkt> same as qt
[16:30] <shadeslayer> well ... what I meant was ... how does ubiquity-frontend-gtk do it?
[16:30] <shadeslayer> i.e access dbus
[16:30] <Riddelll> it doesn't
[16:30] <pkt> because gtk doesn't access dbus behind your back
[16:30] <pkt> it doesn't have to do this trickery
[16:30] <shadeslayer> oh ... so why does the kde installer need to access dbus?
[16:30] <Riddelll> it doesn't
[16:31] <pkt> KDE itself is doing this
[16:31] <shadeslayer> ahhh
[16:31] <Riddelll> KApplication does it
[16:31] <shadeslayer> gotcha
[16:31] <Riddelll> as part of its normal self register on startup stuff
[16:31] <shadeslayer> right
[16:31]  * shadeslayer has to go for about a hour
[16:31] <shadeslayer> cya
[16:31] <pkt> cya
[16:33] <pkt> now that libdbus uses the kernel to understand if you are root or not, no amount of uid saving will trick it
[16:36] <pkt> does KApplication really, really need these session d-bus calls?
[16:38] <Riddelll> here's my quick incomplete version http://starsky.19inch.net/~jr/tmp/kde_ui.py
[16:38] <Riddelll> I'm out for a couple of hours
[16:39] <Riddelll> if pkt or n0ti0nis or anyone fancies getting rid of the kde classes in that it would be good
[16:39] <Riddelll> else I'll look at it when I get back
[16:39] <pkt> cool
[16:40] <pkt> I won't be able to look at it now, but if you are not done by tomorrow I can try to help
[21:52]  * ScottK considers causing some trouble.
[21:57] <ScottK> There we go.
[21:57] <ScottK> http://skitterman.wordpress.com/2012/09/25/desktop-search/
[22:49]  * yofel is tired and needed a moment until he got that :D
[22:49] <yofel> ScottK++
[22:49] <yofel> and I'm off to bed - good night
[23:14] <xnox> ScottK: funny in a 'Ha-ha' way =) good joke, but needs a little more context, a punch line and pictures of kittens.
[23:14] <xnox> =)