=== Jack87|Away is now known as Jack87
=== doko_ is now known as doko
[07:32] <doko> janimo, does the linux-ac100 3.0.8-3.1 kernel have the mmap patch for arm?
[07:33] <janimo> doko, yes
[07:33] <janimo> forgot to close the bug
[07:33] <doko> precise
[07:33] <doko> ok, thanks
[07:33] <janimo> so ac100 is fine on both oneiric and precise
[07:33] <doko> cool, thanks
[09:54] <hrw> guys: any new information about moving arm(el,hf) from ports.ubuntu.com to normal repo?
[10:00] <ndec> hrw: +1 ;-)
[10:37] <janimo> hrw, maybe ask on ubuntu-devel?
[10:38] <janimo> wrt arm moving from ports to regular archives
[12:24] <ogra_> hrw, https://lists.ubuntu.com/mailman/listinfo/ubuntu-release is the place to ask about arm moving to archive.u.c ... but i wouldnt bet on it happening in the forseeable future
[12:25] <hrw> ogra_: here I can usually meet ubuntu/ARM people which may know more and often respond faster then ML
[12:25] <hrw> ;)
[12:26] <ogra_> well, all i know is that i'm told off every time i ask (which i do once every year at least) because we dont want to piss off our mirros
[12:26] <ogra_> *mirrors
[12:27] <ogra_> a.u.c is generally mirrored ... ports isnt (apart from a bunch of inofficial mirrors)
[12:29] <hrw> k
=== chuck_ is now known as zul
[15:31] <[cc]smart> i'm looking at http://cdimage.ubuntu.com/daily-preinstalled/current/precise-preinstalled-desktop-armhf+ac100.bootimg and find that the md5 sum doesnt match the MD5SUMS file, did retries, find the tar.gr match the MD5 sum. should be concerned ?
[15:54] <ogra_> [cc]smart, just take an older one, someone tried to do a manual build that failed, so whats in current/ might actually be messed up
[15:58] <[cc]smart> ogra_, unfortunately, previous one shows the same symptoms. happen to kknow an alternative source for hf bootimages ?
[15:58] <ogra_> no
[15:58] <ogra_> what are the samptoms ?
[15:58] <ogra_> *sym.
[15:58] <[cc]smart> thx anyways
[15:58] <[cc]smart> mismatch in md5sum
[15:58] <ogra_> the md5 of the bootimg file really shouldnt matter
[15:59] <ogra_> as long as the tarball one is right
[15:59] <[cc]smart> ok, i will in this case try current
[16:06] <RoyK> erm... small issue here - I have ubuntu installed on a pandaboard, and in this small program, I try to call getuid(), but it seems to block on the getuid32() call: getuid32()                              = 0
[16:06] <RoyK> any idea?
[17:00] <GrueMaster> rsalveti: Do you know if the correct settings for the 4460 ever made it into the kernel/u-boot for Precise?
[17:01] <ogra_> there are in the tree
[17:01] <ogra_> not sure they made it into the packages yet
[17:01] <GrueMaster> Ok.
[17:02] <GrueMaster> I've been limiting my 4460 uptime as I didn't want to have a core meltdown.
[17:02] <ogra_> i just put a cup of cold coffee on top ...
[17:02] <ogra_> so its warm in the evening :P
[17:03] <GrueMaster> Chip is too small and my coffee mug too big.  I'm afraid the mug would topple over or the board would shatter from the weight.
[17:04] <ogra_> take "light coffee" then :)
[17:04] <rsalveti> GrueMaster: they should all be in already
[17:04] <GrueMaster> I have a small heatsink on it now.
[17:04] <rsalveti> but depends a bit on the last rebase from ppisati
[17:04] <ogra_> right
[17:04] <rsalveti> but you should need only a newer kernel
[17:04] <GrueMaster> rsalveti: Excellent.  I just hadn't heard what the status was.
[17:21] <ppisati> the last P/omap4 should have all we need for 4460
[17:21] <ogra_> awesome
[17:21] <ppisati> GrueMaster: do you have overheating with P kernel?
[17:21] <GrueMaster> I haven't tested on it yet.  Doing that now.
[17:22] <GrueMaster> I have been doing most of my testing on my other pandas.  I just hadn't heard the status of this fix, that's why I asked.
[17:23] <ppisati> k
[17:23] <GrueMaster> I do occasionally fire up the 4460, but I don't run it for extended periods (the other systems are on 24/7 reimaging and testing core daily).
[17:27] <ndec> if you have DVFS and thermal management, you should be good
[17:27] <ogra_> i think it has a builtin fuse mechanism to shut down the power if it gets to hot
[19:11] <shadeslayer> lilstevie: hey! :)
[19:13] <shadeslayer> lilstevie: Any new breakthrough's on the transformer situation?
[19:54] <GrueMaster> Hrm.  Using netinstall for precise armhf (possibly armel as well) to install Ubuntu Desktop to USB, Network Manager refuses to handle eth0.  System has connectivity, but doesn't list it in the applet.
[19:54] <rcn-ee> it's not still in /etc/network/interfaces ?
[19:55] <GrueMaster> Not sure how to file a bug on this one.  needs more testing.
[19:56] <GrueMaster> rcn-ee: No, it is there.  Just not showing up in the applet, even after manually adding it to the wired connections list.
[19:56] <GrueMaster> ifconfig shows a link as well.
[19:56] <rcn-ee> you should drop it from there then, as last i check network-manager won't manage anything already listed in /etc/network/interfaces..
[19:57] <infinity> GrueMaster: That doesn't sound like something that would be even remotely arch-specific.
[19:57] <infinity> And yes, eth* shouldn't be in interfaces(5) at all on a fresh install.
[19:57] <infinity> Unless you manually configured the adaptor.
[19:57] <GrueMaster> No, I doubt it is even platform specific.  Just that I am currently testing armhf.
[19:57] <infinity> Did you statically configure it in the installer?
[19:58] <GrueMaster> Probably in my preseed.
[19:58] <infinity> That would do it.
[19:58] <infinity> network-manager won't show interfaces with a static config in inerfaces.
[19:59] <GrueMaster> ah.
[19:59] <GrueMaster> Well.  That sucks.
[19:59] <infinity> It's by design.
[19:59] <infinity> interfaces is considered canonical, network-manager and connman and other high-level tools shouldn't be allowed to override it.
[20:00] <GrueMaster> Need the static conf in the preseed for automation, but need to remove it post-install.  Hrm.
[20:00] <infinity> Eh?
[20:00] <infinity> You can netboot with dhcp.
[20:00] <infinity> I see no reason for a static config.
[20:00] <GrueMaster> It is dhcp.
[20:00] <rcn-ee> GrueMaster,  d-i preseed/late_command ;)
[20:01] <GrueMaster> rcn-ee: My late_command is gettng a bit long as it is.
[20:01] <infinity> Oh, you mean you're telling d-i it's a dhcp interface, and it's writing out interfaces(5) with "iface eth0 inet dhcp"?
[20:02] <infinity> Fixing that could be complicated. :/
[20:02] <GrueMaster> yep.
[20:03] <infinity> Could just be one of the subtle d-i versus ubiquity differences we get to live with.  But there might be some clever way around it.
[20:03] <infinity> (The problem is that we need precisely that behaviour for anything !desktop)
[20:03] <infinity> So, there would have to be a package in ubuntu-desktop that actually reverts that setup.
[20:03] <infinity> And only on install.
[20:03] <infinity> So, a udeb produced from n-m or something equally icky.
[20:03] <GrueMaster> http://paste.ubuntu.com/793062/ is my preseed for precise.
[20:03] <rcn-ee> ah, mine was too, eventually i just added another script to "/usr/lib/finish-install.d/08-something" to the patched netinstall initramfs  and run a big script in a chroot..
[20:05] <GrueMaster> I used tasksel post-install to install ubuntu-desktop.  Will try a preseed with ubuntu-desktop added.
[20:06] <infinity> GrueMaster: I suspect it will end up the same either way.
[20:06] <infinity> GrueMaster: Does it do the wrong thing if you change choose_interface to auto?
[20:06] <GrueMaster> yea, probably.
[20:07] <GrueMaster> It used to.  Not sure if it works properly.  Will run a quick spin to find out.
[20:08] <infinity>     - Call /usr/lib/NetworkManager/ifblacklist_migrate.sh from
[20:08] <infinity>       finish-install if present, so that 'auto dhcp' interfaces are disabled
[20:08] <infinity>       if network-manager is in use.  May be preseeded away with
[20:08] <infinity>       netcfg/network-manager.
[20:08] <infinity> Hrm.
[20:08] <infinity> Looks like netcfg has code to deal with this.
[20:09] <infinity> And, supposedly, the code triggers by default...
[20:09] <GrueMaster> Obviously well documented (considering the pain I went through to get this far with the preseed).
[20:10] <infinity> Well, the preseed is an inverse.  You only need to preseed to PREVENT the mangling.
[20:10] <infinity> It's meant to happen by default.
[20:10] <infinity> So, maybe something's preventing netcfg's finish-install from running correctly.
[20:10] <GrueMaster> Well, so far it seems to at least connect during netinstall with auto.
[20:11] <GrueMaster> I'll know in ~15 minutes if this changes things.
[20:11] <infinity> Hrm.  It might be that /usr/lib/NetworkManager/ifblacklist_migrate.sh is missing.
[20:11] <GrueMaster> Seems to me I had to add that in for different platforms, and never changed it when I started using different preseeds for different systems.
[20:11] <GrueMaster> Oh?
[20:12] <infinity> Well, that's a wild guess.  Haven't looked at the precise n-m package yet.
[20:12] <infinity> Oh!
[20:12] <infinity> Wait.
[20:13] <infinity> You used tasksel after the install was complete?
[20:13] <infinity> As in, after d-i was done and you were dumped at a loging?
[20:13] <infinity> login*
[20:13] <infinity> If so, then yeah, preseeding ubuntu-desktop will fix it.
[20:13] <GrueMaster> yea.  My bad?
[20:13] <GrueMaster> ok.
[20:13] <infinity> Cause this finish-install magic in netcfg is in the udeb (ie: in d-i), but it's checking if n-m is in the target.
[20:13]  * GrueMaster feels sheepish.
[20:13] <infinity> So, n-m needs to be installed before d-i is done.
[20:14] <GrueMaster> Well, better to raise the question here than in a launchpad bug report.
[20:14] <infinity> Heh.
[20:15] <infinity> S'ok, every time someone asks something like this, I research and learn a little bit more.
[20:15] <infinity> Eventually, my head will explode, and paint the walls with d-i internals.
[20:15] <infinity> It'll be pretty.
[20:15] <GrueMaster> Yea, if you like crimson & gray walls.
[20:16] <infinity> Pfft, it's d-i.  Yellow and blue walls.
[20:17] <infinity> (I still refuse to admit that kirkland changed the terminal color palette for Ubuntu's d-i, yellow and blue for life!)
[21:01] <badZeppelin> hi folks! I have managed to install sgx opengl drivers on my beagleboard now but only the powervr demo apps make use of those sgx libraries when I check with ldd
[21:01] <badZeppelin> stock opengl apps like es2gears point to some other library that uses mesa software rasterizer
[21:02] <badZeppelin> does anyone have a clue how I make all applications use the ones in /usr/lib