**** BEGIN LOGGING AT Mon Aug 27 02:59:58 2012 Aug 27 08:01:09 infinity, ah, great, i thought they were disabled by higher order or some such ... my bamboo-feeder tests every new omap3 server image that comes down the drain automatically now so we can see about 1h later if it is installable, at least vm testing is constantly covered with that for omap3 Aug 27 08:05:32 * ogra_ sighs about gvfs .... the archive skew killed all arm builds, fun Aug 27 09:03:56 meh Aug 27 09:04:16 rsalveti: ogra_: http://paste.ubuntu.com/1169507/ - pvr-omap4 dkms/pvr-omap4 fails to compile Aug 27 09:04:34 during installation of the pvr-omap4 package Aug 27 09:04:58 what release ? Aug 27 09:05:06 quantal Aug 27 09:05:12 can ubuntu still accept the patch from bugs.debian.org/676533 for the 12.10 release ? Aug 27 09:05:15 wrong pvr version ;) Aug 27 09:05:49 xranby, 1.9 was uploaded on the weekend, you want that one Aug 27 09:06:04 not sure where its stuck ... probably sitting in binary NEW Aug 27 09:06:11 ah Aug 27 09:06:18 i thought it had been processed by now Aug 27 09:07:05 suihkulokki, i dont see why not, can you file a bug ? Aug 27 09:09:46 ogra_, your bamboo feeder is awesome I must say :p Aug 27 09:09:55 heh, thanks Aug 27 09:10:09 i just finished a virtual version that uses oamp3 qemu VMs Aug 27 09:10:27 once i'm done i'll package that as an arm testing suite Aug 27 09:10:58 xranby, https://launchpad.net/ubuntu/+source/pvr-omap4/1.9.0.5.1.1-0ubuntu1/+build/3742590 ... waits for libdri2 Aug 27 09:11:21 which version of libdri2 ? Aug 27 09:11:30 there is a libdri2-1 in the archive Aug 27 09:12:24 and a libdri2-dev Aug 27 09:12:53 lets see Aug 27 09:13:02 * ogra_ gives the pvr build back Aug 27 09:13:34 * lilstevie wishes there was a qemu tegra Aug 27 09:16:41 ogra_: thats really odd. why do your builders fail to find libdri2-1 when i can install it on my panda.. Aug 27 09:16:48 * xranby checking /etc/apt* Aug 27 09:17:50 might be that pvr needs it in main Aug 27 09:18:07 iirc restricted doesnt allow using universe b-deps Aug 27 09:40:43 ogra_: i tested a manual build of the sourcepackage: pvr-omap4_1.9.0.5.1.1-0ubuntu1_armhf.deb pvr-omap4-dbg and pvr-omap4-dev got built fine Aug 27 09:40:49 and they install correctly Aug 27 09:41:25 i am still missing libegl1-sgx-omap4 and libgles2-sgx-omap4 that did not got produced from the build Aug 27 09:41:31 by the build Aug 27 09:43:26 well from the looks of it.. this new package should not build those two package.. Aug 27 09:43:37 * xranby are still puzzled why es2_gears fails Aug 27 09:55:08 ogra_: https://bugs.launchpad.net/ubuntu/+source/openssl/+bug/1042162 Aug 27 09:55:10 Ubuntu bug 1042162 in openssl "Please enable arm assembler optimisations in openssl " [Undecided,New] Aug 27 09:56:05 suihkulokki, merci :) Aug 27 11:28:25 How can I find the changes from mainstream kernel to the ubuntu modified one? Point is; I have a custom 2.6.37 omap kernel and I'm looking for the ureadahead patch. Aug 27 11:28:48 there are no arm specific changes for omap3 afaik Aug 27 11:28:59 we just use what mainline provides Aug 27 11:29:26 and ureadahead works and has meaning? Aug 27 11:29:27 only omap4 and ac100 use custom kernels Aug 27 11:29:59 ureadahed doesnt support Sd cards yet so makin it work would be moot Aug 27 11:30:25 (and i dont think there are any non mainlined kernel changes required for ureadahead) Aug 27 11:30:39 oh, what is there to support? I mean being a block device like all other block devs Aug 27 11:31:42 * sveinse admitting not fully knowing what ureadahead really does Aug 27 11:34:22 The goal is to reduce the bootup time. In our product, the kernel is booted within 4 secods. But the main (qt4/qws) app fires up 30 seconds after that. All this time is spent in initramfs and booting, so I'm looking for optimizations ... Aug 27 11:35:09 I was under the impression that precise could boot very quickly (due to ureadahead), and this is what I'm hunting for Aug 27 11:59:02 well, did you try bootchart ? did it show anything obvious? Aug 27 12:05:36 how come the Aug-27 image of armhf-omap4 is only 22 MB? Aug 27 12:05:52 ogra_, working on it Aug 27 12:06:23 jimerickson, messed up buildds Aug 27 12:06:36 just sotiung it with the IS team Aug 27 12:06:55 ok thanks Aug 27 12:24:19 ogra_, are the live image build logs somewhere public? Aug 27 12:24:38 I remember getting mails last year with links to some but I forget where they pointed at Aug 27 12:37:16 ogra_: You have experience running omap3, right. How is the X performance in OMAP3515? Do you know FH? Aug 27 12:37:26 http://people.canonical.com/~ubuntu-archive/cd-build-logs/ubuntu/quantal/ and http://people.canonical.com/~ubuntu-archive/livefs-build-logs/quantal/ Aug 27 12:38:02 janimo, ^^^ Aug 27 12:38:27 sveinse, no idea, i havent run omap3 desktop in a long time Aug 27 12:38:57 ogra_, thanks Aug 27 12:42:26 We're currently running framebuffer Qt/QWS on a VGA screen. There's a discussion internally if we gain something by running a single window Qt on X. E.g. video support and such Aug 27 13:07:20 sveinse: 30s boot is long indeed Aug 27 13:07:42 sveinse: tried with other card or usb storage for rootfs? Aug 27 14:12:48 Perhaps a stupid question, and totally OT here. But how do I cross compile for i386 on amd64? I'm used to using arm-linux-gnueabi-gcc, but there is no such i386-linux-gnu-gcc AFAICS Aug 27 14:13:45 * ogra_ would just use a chroot Aug 27 14:14:19 * lilstevie uses pbuilder (so chroot like ogra_ said) Aug 27 14:17:21 thanks. Not as trivial as I'd hoped, since you must move everything inside the chroot for it to work. But ok Aug 27 14:20:03 pity. I hoped it was possible to bind mount a directory from the main FS into the chroot there a second... But naah. Aug 27 14:20:36 sure you can Aug 27 14:20:45 I don't see why you couldn't Aug 27 14:21:05 unless something has changed recently Aug 27 14:21:11 * ogra_ does bind mount /dev in his chroots all the time Aug 27 14:21:39 * lilstevie used to bind mount ~/image_patches when modifying omap images for the tf101 Aug 27 14:21:41 no reason why you shouldnt be able to user other dirs ... Aug 27 14:21:51 *use Aug 27 14:22:13 ogra_ very true. I do the same when I come to think of it. I think it might be my chroot script which is too "intelligent" in respect of mount... Aug 27 14:25:33 You're going to frown, but is it possible to create a arm chroot and replace gcc & binutils & make (etc.) with arm-linux-gnueabi-gcc toolset from the host? Aug 27 14:26:05 yeah, people call that OBS :P Aug 27 14:26:12 I mean with binfmt+qemu I there is no principal difference from running i386 vs armel as long as all libs are available Aug 27 14:26:41 right, thats what opensuse build serivce does :) Aug 27 14:26:57 lilstevie, do you know if the sensors on the transformer require some android specific userspace binaries to work well? It seems they need at least some ioctls to get started Aug 27 14:26:58 but using self rolled gcc's for everything Aug 27 14:27:24 janimo, I would have no clue, I really do not see much point for most sensors under ubuntu Aug 27 14:27:44 well, at least screen orientation via accel would be nice Aug 27 14:27:50 also light sensor for screen dimming Aug 27 14:27:56 at least some x86 laptops have the latter Aug 27 14:27:57 janimo, however I do believe the mpl stuff should work without much userspace interaction Aug 27 14:28:05 but probably using platfrom specific code Aug 27 14:28:12 eh, autobright would be nice Aug 27 14:28:13 lilstevie, mpl or mpu? Aug 27 14:28:37 but accel is a lot of effort from past experience with xorg and accelerometers Aug 27 14:28:40 mpl Aug 27 14:28:43 that is the brand Aug 27 14:28:46 ah Aug 27 14:29:05 I thought Motion Processing Unit as that's how the gyro+accel are referred to in many places Aug 27 14:29:40 I have no experience with X and accel, no idea if it works on x86 or some exisiting ubuntu hw at all Aug 27 14:29:56 janimo, years ago we tried it with the iPhone 3G Aug 27 14:29:58 indeed no use in sensor capabilities if userspace support is not here Aug 27 14:29:59 but that was karmic Aug 27 14:30:11 and it was horrible Aug 27 14:30:28 there's no unified accel interface/low level lib in linux either Aug 27 14:30:39 only platform specifics in each app afaik Aug 27 14:31:05 but I mean I guess if you could hook it in with xrandr it might be a little bit more bearable Aug 27 14:47:54 can someone try bug 501277 on ARM, it looks fixed on other archs, but I notice the arm* files look like they're missing Aug 27 14:47:55 Launchpad bug 501277 in gdb "gdb "catch syscall" doesn't work, missing syscalls/amd64-linux.xml" [Undecided,Fix released] https://launchpad.net/bugs/501277 Aug 27 15:06:21 janimo, there seems to be a kernel fix for the vm.min_kbytes issue with USB WLAN ... might be good to have that in the ac100 kernel Aug 27 15:14:10 ogra_, did not know about that bug, is it in marvin24 branch? Aug 27 15:14:21 if so I plan on tracking that close to release time anyway Aug 27 15:14:28 janimo, huh ? Aug 27 15:14:42 ogra_, vm.min_kbytes fix for USB WLAN Aug 27 15:14:54 we added the sysctl.d file so often in the past that i'm surprised you dont remember it Aug 27 15:15:07 oh, it was not me who added it I guess Aug 27 15:15:19 or if so, it is one of those ugly things I force myself to forget Aug 27 15:15:28 Heh. Aug 27 15:15:35 well, its the issue where you get oopses in demsg all the time ... Aug 27 15:15:36 anyway I certainly did not know it relates to USB WLAN Aug 27 15:15:50 It relates to some drivers, ish. Aug 27 15:15:51 iirc you added it for mx5 :) Aug 27 15:16:02 But apw did express concern with the way it was "fixed" in that driver in the ac100 tree. Aug 27 15:16:10 I think I forgot all my mx5 dealings altogether Aug 27 15:16:16 right, and someone (ericm or cooloney) has a fix for that by setting the right value from the kernel Aug 27 15:16:43 infinity, so i have an intresting morning with celbalrai Aug 27 15:16:51 I will certainly add it to ac100 if there's a patch already Aug 27 15:16:52 s/have/had/ all fixed now Aug 27 15:17:00 ogra_: Yeah, you nick hilighted me in passing, I went and read a bit of it. Aug 27 15:18:06 seems we accidentially ended up with an armel chroot ... at the same time the new live-build grew support for building foreign arches in qemu-static chroots ... guess what happened when the machine reported armel to live-build :) Aug 27 15:18:47 ogra_: It was always an armel chroot, which was a non-issue until today. Aug 27 15:18:57 ogra_: Or, rather, until livecd-rootfs 2.76. Aug 27 15:19:06 well, until new live-build i guess then Aug 27 15:19:09 ogra_: Thanks for hunting that down. Aug 27 15:19:21 well, i wanted images :P Aug 27 15:19:41 Passing --architecture to live-build as a result of -A is definitely not what we want. :/ Aug 27 15:20:04 yep Aug 27 15:20:15 Or maybe it's that --architecture is DTWT. Pick one. Aug 27 15:20:30 But we can worry about that later. Thanks for fixing for now. Aug 27 15:23:57 * ogra_ twiddles thumbs waiting for the publisher Aug 27 15:24:34 ubuntu@panda:~$ apt-cache madison libdri2-1|grep main Aug 27 15:24:34 ubuntu@panda:~$ Aug 27 15:24:37 *twiddle* Aug 27 15:24:48 Hrm? Aug 27 15:24:52 Was there an MIR when I wasn't looking? Aug 27 15:24:59 mterry has moved it it seems Aug 27 15:25:09 bug 1042151 Aug 27 15:25:10 Launchpad bug 1042151 in libdri2 "MIR: libdri2 needs to be in main for pvr-omap4 (restricted) to build" [High,Fix committed] https://launchpad.net/bugs/1042151 Aug 27 15:25:29 Ahh, I was going to work with rsalveti on that today, I guess he got impatient. Aug 27 15:25:33 Oh, or you got impatient. Aug 27 15:25:34 and yeah, i added a MIR :) not sure you looked though Aug 27 15:26:08 And no, mterry can't move things. "Fix Committed" is the cue for someone (like me) to fix it. ;) Aug 27 15:26:11 * infinity fixes. Aug 27 15:26:22 just waiting for it to move so https://launchpad.net/ubuntu/+source/pvr-omap4/1.9.0.5.1.1-0ubuntu1/+build/3742590 can finally build Aug 27 15:26:28 aha Aug 27 15:26:48 i thoguht the MIR team did the actual movin too nowadays Aug 27 15:27:07 Well, if the team member who approves happens to be an AA, sure. Aug 27 15:27:10 But not all of them are. Aug 27 15:27:18 ah Aug 27 15:27:25 yeah, i know asac isnt Aug 27 15:27:36 not that he does many MIR reviews lately :) Aug 27 15:28:49 * ogra_ wonders why lubuntu-desktop takes so much longer than ubuntu-server ... the resulting image is 200MB smaller Aug 27 15:29:07 likely more suqashfs fiddling Aug 27 15:29:12 Yeah, but desktoppy crap has all those irksome postinsts and triggers that, like... Do stuff. Aug 27 15:29:22 oh, yeah that too Aug 27 15:30:56 I am looking forward to lubuntu on the ac100. Had not used lxde before but if it is snappier than unity-2d was I am all for it Aug 27 15:31:31 i was playing with it over the weekend and ended up with a completely new desktop due to playing around :) Aug 27 15:32:16 my test machine is now running a funny awn setup +openbox and pcmanfm Aug 27 15:32:44 htop shows below 100M used right after boot :) Aug 27 15:33:29 intrestingly i needed xcompmgr to get awn working .... t does seem to work realyl fast and flawless Aug 27 15:33:40 ven better than metacitys builtin compositor Aug 27 15:33:46 *even Aug 27 15:33:59 * ogra_ sighs about his typing Aug 27 15:35:10 infinity, so what do we do to get pvr installed by default (i'm not even sure that works with dkms packages) ... ? Aug 27 15:35:29 we cant just depends on it with ubuntu-desktop ... Aug 27 15:36:02 we could add it to ship and set the "install third party drivers" box in ubiquity by default, but that will also pull in other stuff Aug 27 15:36:07 ogra_: We can't install it by default, see the tech-board decision about the "install non-free" checkbox. Aug 27 15:36:30 ogra_: But I really have no idea what we can/should do, given that nothing works without it. :/ Aug 27 15:36:33 infinity, well, the desktop wont start if we dont have it Aug 27 15:36:55 and the license seems to even permit it Aug 27 15:38:21 infinity, that decision was for flash though ... i cant imagine they would block us if its for a driver Aug 27 15:39:20 bug 723831 btw Aug 27 15:39:21 Launchpad bug 723831 in ayatana-design "Installer – The option to 'install third-party software' when installing Ubuntu should be selected by default (aka "make Youtube work")" [High,Confirmed] https://launchpad.net/bugs/723831 Aug 27 15:40:20 ",We have always made some clearly stated concessions for hardware drivers (graphics, wlan) which http://www.ubuntu.com/project/about-ubuntu/licensing points out. But we unanimously feel that Flash does not at all fall into this category" Aug 27 15:40:27 ogra_: The decision has to do with freeness, not flash specifically. Aug 27 15:40:28 pitti in comment #19 Aug 27 15:40:51 ogra_: Though, I can sort of see the "maybe we can hand-wave drivers past this" thing. Aug 27 15:41:00 yeah Aug 27 15:41:33 I haven't even looked at how the drivers-common thing works, since it's replaced jockey. Aug 27 15:41:46 i implemented the arm bits Aug 27 15:42:07 it checks /proc/cpuinfo and works down a list of possible drivers Aug 27 15:42:20 but that doesnt help in the installer Aug 27 15:42:29 it wont run before the desktop is up Aug 27 15:42:33 catch22 Aug 27 15:44:16 Right. Well, we could just add it to the squashfs, then, in the subarch section. Aug 27 15:44:29 Same place we add the u-boot variant and kernel. Aug 27 15:44:34 how would dkms handle that ? Aug 27 15:45:00 i.e. would we end up with a reboot trigger in the squashfs etc Aug 27 15:45:15 does the build work at all ... Aug 27 15:45:18 Dunno, never tried to dkms in a chroot, but I suspect it'll just pick the highest-versioned kernel installed in the root and be happy. Maybe. Aug 27 15:45:30 hehe, high hopes Aug 27 15:45:40 Reboot triggers end up in the livefs anyway, that's nothing new. Aug 27 15:45:47 They get cleared at boot. Aug 27 15:45:48 but yeah, i guess putting it into livecd-rootfs is the best Aug 27 15:46:31 As long as it doesn't fall over on trying to guess anything about the running kernel (which, if it does, is a bug in dkms anyway, I'd say), it should DTRT, ish. Aug 27 15:47:12 well, lets just add it once pvr is there Aug 27 15:47:18 and see what happens Aug 27 15:47:46 preferably at a time where lamont is up to recover the machine though :) Aug 27 15:48:08 I don't see how it would break the builder. Aug 27 15:48:27 well, i dont know how qemu did it on the weekend either ... Aug 27 15:48:47 fact is celbalrai was dead for two days Aug 27 15:49:39 lubuntu-armhf-ac100 on celbalrai.buildd finished at 2012-08-27 15:46:44 (success) Aug 27 15:49:40 yay Aug 27 15:56:41 ogra_: FTBFS. Enjoy. Aug 27 15:56:46 https://launchpadlibrarian.net/113731856/buildlog_ubuntu-quantal-armhf.pvr-omap4_1.9.0.5.1.1-0ubuntu1_FAILEDTOBUILD.txt.gz Aug 27 15:57:16 bah Aug 27 15:57:31 Missing an entry in pvr-omap4.dirs, I assume. Aug 27 15:57:32 must be a buildd error :P Aug 27 15:57:56 the buildd tarball should just ship /build/buildd/pvr-omap4-1.9.0.5.1.1/debian/pvr-omap4/usr/lib/pvr-omap4-egl/ld.so.conf Aug 27 15:58:00 :P Aug 27 15:58:05 Uh huh. Aug 27 15:58:24 rsalveti, https://launchpadlibrarian.net/113731856/buildlog_ubuntu-quantal-armhf.pvr-omap4_1.9.0.5.1.1-0ubuntu1_FAILEDTOBUILD.txt.gz Aug 27 15:58:34 * ogra_ defers friendly :) Aug 27 15:58:34 Oh, I'll just fix it. :P Aug 27 16:02:12 hrm, the only thing i realyl dislike about lubuntu is that it defaults to chromium Aug 27 16:02:28 which is broken on arm for 2/3 of a cycle Aug 27 16:24:34 ogra_: Test-build successful and uploaded. Aug 27 16:24:36 rsalveti: ^ Aug 27 16:31:00 ogra_: And built just in time, should hit this upcoming :03 publisher run, if you want to play with it. Aug 27 16:48:50 * rsalveti back on-line Aug 27 16:48:52 infinity: what was the issue? Aug 27 16:49:20 rsalveti: http://launchpadlibrarian.net/113734702/pvr-omap4_1.9.0.5.1.1-0ubuntu1_1.9.0.5.1.1-0ubuntu2.diff.gz Aug 27 16:50:35 hm, that's interesting, that rules didn't change from previous driver Aug 27 16:50:43 and built fine when I tested with pdebuild Aug 27 16:50:57 * rsalveti checking the diff again Aug 27 16:53:13 yup, nothing changed at that perspective from the previous one, which built fine as well Aug 27 16:55:36 might be a result of the new compat/source version Aug 27 16:55:47 rsalveti: Weird. Was that directory in pvr-omap4.dirs before? Or maybe made in another target and the build-arch/build-any split broke you? Aug 27 16:55:48 infinity: thanks anyway, sorry for not catching it before Aug 27 16:56:51 not so sure, will try to investigate it more to try to understand what happened there Aug 27 17:04:38 infinity: ogra_: I also pushed 2 fixes at the xserver and modesettings driver this weekend, so it should finally work fine at this point Aug 27 17:04:58 the video-omap one should work out-of-the-box and the pvr will be enabled once the pvr-omap4 package is installed Aug 27 17:18:14 oh, curious, why did my panda not have the headers instaled Aug 27 17:19:51 ogra_: with which kernel? Aug 27 17:19:56 3.5 Aug 27 17:20:04 came with a dist upgrade Aug 27 17:20:17 rsalveti, hmm, ouch Aug 27 17:20:24 who installs it by default? Aug 27 17:20:30 seems you still depend on dos2unix ... thats still in universe Aug 27 17:20:52 we can fix that in a patch if needed Aug 27 17:20:57 let me check that Aug 27 17:21:07 so we don't need another mir Aug 27 17:21:51 ogra_: but can you at least try to see if it all works for you? Aug 27 17:24:03 rsalveti, on it, thoguht where was that compiz PPA again ? Aug 27 17:25:49 we requested the team to publish it for arm as well, don't know if that was done, and we'd also need to change it to use the official branch Aug 27 17:25:56 which I'm not that sure it happened already Aug 27 17:26:16 ogra_: check if you're at least able to run es2gears or glmark2-es2 Aug 27 17:26:33 k Aug 27 17:27:09 ogra_: infinity: talking about glmark2, got a patch to fix the ftbs with the latest mesa: bug 1040895 Aug 27 17:27:11 Launchpad bug 1040895 in glmark2 "FTBFS when building against latest mesa (9.0 based)" [High,In progress] https://launchpad.net/bugs/1040895 Aug 27 17:27:29 mesa updated the abi (actually fixing the abi) Aug 27 17:27:46 but one function was changed to return int to void, which broke glmark2 Aug 27 17:32:34 rsalveti: Sure, I can sponsor that. Aug 27 17:33:13 and I'm actually afraid of other issue that might happen with latest mesa Aug 27 17:33:19 they updated the headers for most of the standards Aug 27 17:33:50 we'll probably know easily in a rebuild Aug 27 17:34:36 * infinity throws it at sbuild. Aug 27 17:35:16 rsalveti, hmm, didnt look great Aug 27 17:35:37 * ogra_ reboots the panda Aug 27 17:35:40 ogra_: what happened? Aug 27 17:35:50 ogra_: if you built the new module, a reboot is needed Aug 27 17:35:59 some dri message "couldnt open pipe window" or some such Aug 27 17:36:07 why ? Aug 27 17:36:17 cant you just make dkms modprobe it ? Aug 27 17:36:26 did you have the older one installed already? Aug 27 17:36:30 nope Aug 27 17:36:40 hm, then I believe it was supposed to work Aug 27 17:36:54 was a virgin milestone install, just upgraded regulary until unity-2d was removed Aug 27 17:36:55 can you provide your xorg.log and the output at your console? Aug 27 17:37:15 hmm, i rebooted already, i dont think xorg keeps a backup Aug 27 17:37:46 yup, there's a 1.log Aug 27 17:37:59 thats for display 1 Aug 27 17:38:01 the log.old actually Aug 27 17:38:05 but there is a .old ;) Aug 27 17:38:16 * ogra_ saves it Aug 27 17:38:24 * rsalveti trying to get his brain started, didn't sleep at the plane Aug 27 17:38:43 oh, where are you ? Aug 27 17:38:49 SD. Aug 27 17:38:59 going to plumbers, at Washington atm Aug 27 17:39:05 ah Aug 27 17:39:08 got a 4 hours delay at the first flight Aug 27 17:39:08 Oh, almost in SD. :P Aug 27 17:39:09 * ogra_ is envious Aug 27 17:39:13 rsalveti: Why a day early? Aug 27 17:39:20 infinity: 1k $ less Aug 27 17:39:24 Ahh. Aug 27 17:39:28 jetlag :P Aug 27 17:39:35 It was cheap here, regardless. Aug 27 17:39:43 Though a day early wouldn't have hurt my feelings. Aug 27 17:40:16 I believe there might be some other conferences and stuff to do tomorrow as well Aug 27 17:40:39 OMAP: Failed to load module "omap_pvr" (module does not exist, 0) Aug 27 17:40:41 hmpf Aug 27 17:41:28 ogra_: check your dkms log Aug 27 17:41:36 or rebuild the module/reinstall pvr-omap4 Aug 27 17:41:36 and lsmod agrees Aug 27 17:41:38 ok Aug 27 17:42:02 maybe this issue with the lack of headers interfered with your module Aug 27 17:42:12 getting dkms to skip the build part of it Aug 27 17:42:14 or similar Aug 27 17:42:24 Well, yeah, it won't build unless you have headers installed. Aug 27 17:42:27 That's a bit of a no-brainer. Aug 27 17:42:32 well, it wasnt an "issue" theyjust came in when i installed omyp4pvr Aug 27 17:42:42 god ! Aug 27 17:42:42 But installing the headers SHOULD cause DKMS builds. Aug 27 17:42:58 this kbd is really not for my fingers it seems Aug 27 17:43:36 "oh my 4" seems like an appropriate subarch name. Aug 27 17:44:42 haha Aug 27 17:45:45 rsalveti: Is the "de $Place" appelation on the end of names still really common down there, or do you just like to show off and have a longer name? :P Aug 27 17:46:03 I'd totally be "Adam Conrad of Calgary", but it sounds stupid in English. Aug 27 17:48:57 infinity: lol Aug 27 17:49:27 it's still common to have sur names from both families Aug 27 17:49:33 but the father part is the last Aug 27 17:49:48 Ahh. So, not the placename bit, like in some cultures. Aug 27 17:50:02 well, we have kind of everything around here hehe :-) Aug 27 17:50:47 ;) Aug 27 17:50:55 infinity, oh, btw, riku asked about bug 1042162 do we have any objections ? (i assume not) Aug 27 17:50:56 Launchpad bug 1042162 in openssl "Please enable arm assembler optimisations in openssl " [Undecided,New] https://launchpad.net/bugs/1042162 Aug 27 17:51:48 hm, interesting Aug 27 17:54:10 (EE) OMAP: Failed to load module "omap_pvr" (module does not exist, 0) Aug 27 17:54:13 hmm, no go Aug 27 17:55:05 ogra_: is pvr-omap4 installed? Aug 27 17:55:46 indeed it is Aug 27 17:55:49 that comes from the pvr-omap4 package Aug 27 17:55:55 which the video-omap one tries to load Aug 27 17:57:04 ubuntu@panda:~$ dpkg -L pvr-omap4|wc -l Aug 27 17:57:04 14 Aug 27 17:57:10 try starting X by hand, and give me the output Aug 27 17:57:14 not much in there Aug 27 17:57:16 Xorg -verbose 10 Aug 27 17:57:44 in fact only ld.so.conf, alt_ld.so.conf and copyright Aug 27 17:57:55 is it supposed to be like that ? Aug 27 17:58:35 nops Aug 27 17:58:51 ubuntu@panda:~$ apt-cache show pvr-omap4|grep Installed Aug 27 17:58:52 Installed-Size: 47 Aug 27 17:58:53 YAY Aug 27 17:59:01 maybe some other issue with infinity's change Aug 27 17:59:05 the new was of space saving Aug 27 17:59:15 way Aug 27 17:59:30 but at least it seems the video-omap is loaded automatically already Aug 27 18:00:11 oh, probaby i shoudl remove that xorg.conf i created when testing -omap a month ago :P Aug 27 18:00:42 yup, there's nothing at the pvr-omap4 package Aug 27 18:00:47 please :-) Aug 27 18:00:47 still loads OMAP :) Aug 27 18:01:04 so that side is fine ;) Aug 27 18:01:37 Hello! What is the right method to use qemu-user-static and chroot for compiling/installing packages offline like an emulator. I'm trying to compile OpenCV from source on BeagleBoard XM and its taking long. Which one is right method here - http://paste.ubuntu.com/1170326/ ? Aug 27 18:02:08 ogra_: weird, some other issue is happening with this driver Aug 27 18:02:54 yeah, 47 bytes doesn definitely not look right Aug 27 18:04:21 orated, after installing apt-get install qemu-user-static just use the qemu-debootstrap command ... after bootstrapping everythinng behaves like any other chtoor in the world Aug 27 18:05:05 i really wouldnt try to build on an SD card (as 2.) seems to suggest) ... just do it in a chroot on the host Aug 27 18:06:31 * ogra_ is out ... if there is anything to sponsor later i'll drop in for a moment Aug 27 18:07:03 One moment ogra_ ... I'd like to confirm the steps please Aug 27 18:07:54 apt-get install qemu-user-static Aug 27 18:07:56 cp /usr/bin/qemu-arm-static Aug 27 18:07:57 qemu-debootstrap Aug 27 18:07:59 ? Aug 27 18:08:11 I've not done bootsrapping before Aug 27 18:08:32 ogra_: there's a way to use tofrodos instead of dos2unix, which is already in main Aug 27 18:08:58 ogra_: should have a patch later in the day, and will test with my panda again Aug 27 18:12:44 rsalveti: What are you using dos2unix for? Aug 27 18:13:43 infinity: pvr build system is using it to generate the build files Aug 27 18:14:02 rsalveti: Just to strip ^M or something? Aug 27 18:14:24 rsalveti: Pretty sure that can be done with something as simple as tr. Aug 27 18:14:27 Seems like it could just use tr or something. Aug 27 18:14:46 I might even have an old shell script for that. Aug 27 18:15:44 afaik it was using during it while generating a few files, so not just a simple convert thing Aug 27 18:15:50 need to dig at the dirty pvr build system Aug 27 18:16:21 Ah, here we go. "tr -d '\015' <$1 >$2" Aug 27 18:16:31 rsalveti: Well, it might also be sane to tr the files in question during the package build and ship them translated, and patch out the part where it does it during the module build. Aug 27 18:16:41 rsalveti: Shipping icky DOSy files in the package seems wrong. Aug 27 18:16:58 GrueMaster: What do you have against '\r'? :) Aug 27 18:17:06 yup, but it's a pain to maintain as a package patch Aug 27 18:17:15 as it changes mostly at every update Aug 27 18:17:27 rsalveti: No, I mean do it in debian/rules during binary-arch. Aug 27 18:17:31 Nothing. Like I said, old script (i.e. 1996 from my AIX days). Aug 27 18:17:44 rsalveti: Don't mangle the source package, just the results in debian/pvr-whatever Aug 27 18:18:08 sure, but would also need a patch to remove the dos2unix calls Aug 27 18:18:10 rsalveti: So what's shipped in the binary packages is sane and reasonable (and readable). Aug 27 18:18:33 rsalveti: Yeah, but removing the dos2unix bits is at least maintainable. ;) Aug 27 18:18:40 rsalveti: And you have to patch those bits one way or the other. Aug 27 18:18:45 hurricanes and earthquakes, sounds like a perfect timing to visit us Aug 27 18:19:30 infinity: yup, agree Aug 27 18:22:53 something to do at my next flight Aug 27 18:24:05 bbl Aug 27 18:25:20 Hi infinity! When I was searching for offline package compilation/installation I found old log - http://irclogs.ubuntu.com/2012/06/23/%23ubuntu-devel.txt - and as you can see, step1 here -http://paste.ubuntu.com/1170326/ -is based on your reply in the log. Could you please guide me which method is right so that I can complete OpenCV source compilation and other relevant installations? Aug 27 18:27:38 orated: Assuming you have a rootfs (like ubuntu-core), then those directions seem fine. Aug 27 18:28:01 orated: If your goal is package building, though, I recommend mk-sbuild from ubuntu-dev-tools, and just using sbuild/schroot. Aug 27 18:33:59 infinity: Umm, before someone suggested bootstrapping and now mk-sbuild ;) I'm really confused now. I've only done basic chroot before from liveUSB/CD. I'm not aware of bootstrapping or sbuild/schroot. I'm having difficulty completing source installation of OpenCV on BB XM, that I thought of trying emulator. What's the difference in trying - Aug 27 18:34:00 apt-get install qemu-user-static, cd /media/rootfs, cp /usr/bin/qemu-arm-static usr/bin/, cp /etc/resolv.conf etc/resolv.conf, chroot . ? Aug 27 18:35:19 orated: mk-sbuild will end up using debootstrap on the backend. Most of these tools do. Aug 27 18:35:53 orated: But yeah, if you have a rootfs unpacked in /media/rootfs (like ubuntu-core, for instance), then those quick and dirty directions work fine. Aug 27 18:37:58 infinity: Well, then could you explain a bit more how to use mk-sbuild in my case? Aug 27 18:38:44 https://wiki.edubuntu.org/SecurityTeam/BuildEnvironment <-- A good writeup. Aug 27 18:39:02 http://wiki.debian.org/mk-sbuild <-- A bit less verbose Aug 27 18:40:53 Thanks, I'll try that **** ENDING LOGGING AT Tue Aug 28 03:00:03 2012