**** BEGIN LOGGING AT Tue Oct 06 15:14:14 2009 **** ENDING LOGGING AT Wed Oct 07 02:59:57 2009 **** BEGIN LOGGING AT Wed Oct 07 02:59:57 2009 Oct 07 06:08:35 http://dooble.sourceforge.net/maemo_dooblescratchbox2.PNG Oct 07 06:08:37 lol Oct 07 06:09:42 windows vmware ubuntu xephyr maemo Oct 07 06:35:13 hi mike^, Oct 07 06:35:42 I successfully compiled clutter 1.0 with EGL support Oct 07 06:36:18 Now am getting excellent performance specially with animations etc Oct 07 07:19:44 siji: Cool **** BEGIN LOGGING AT Wed Oct 07 14:49:23 2009 Oct 07 18:09:57 Hi, how do I create a ubuntu filesystem for ARM, anything simliar to OpenEmbedded? Oct 07 18:18:10 anyone? Oct 07 18:23:19 Vink___: https://wiki.ubuntu.com/ARM/RootfsFromScratch Oct 07 18:24:22 Can I compile the rootfs with customized packages? I want to add/ remove a lot of stuff Oct 07 18:26:46 ojn: Can I compile the rootfs with customized packages? I want to add/ remove a lot of stuff Oct 07 18:28:33 Vink___: It comes with alot of packages already, and the rest you can install with apt-get just like under regular ubuntu/debian Oct 07 18:29:18 If you feel a need to waste computrons on rebuilding your own software every time, then ubuntu isn't the obvious choice. gentoo or OE might be a better base. **** ENDING LOGGING AT Thu Oct 08 02:59:58 2009 **** BEGIN LOGGING AT Thu Oct 08 02:59:58 2009 Oct 08 13:55:14 is the current karmic booting on the beagleboard? I'm getting "init: procps main process (670) terminated with status 255" on boot and it stops there Oct 08 14:50:00 i am looking for a glib-2.20.0 and gstreamer-0.10 packages to be installed by apt-get into my beagleboard. The list of packages with the package name inside is long and I can't find out which is the package I need. can someone help with that? Oct 08 15:10:59 well, I took the most wider package, so I am sure it depends on the basic library I need :-) -life are easier when you memory Oct 08 16:40:14 beyossi: What do you need? Oct 08 16:40:22 beyossi: shared library? development headers? Oct 08 16:40:39 beyossi: apt-cache search libglib should give you a list with short descriptions Oct 08 16:41:02 dpb: I run karmic on beagleboard Oct 08 16:41:12 dpb: Do you use an initrd? Oct 08 16:41:33 dpb: Try with an initrd; would also be nice to see your full boot log on serial console Oct 08 19:03:37 lool: thanks, i found. Oct 08 19:04:44 my application compilation looks for asm/hwcap.h Oct 08 19:04:44 any idea what library to install?, is it libasm-dev? Oct 08 19:25:39 found. you need linux-libc-dev **** ENDING LOGGING AT Fri Oct 09 03:00:00 2009 **** BEGIN LOGGING AT Fri Oct 09 03:00:00 2009 Oct 09 06:36:51 lool: I don't use a initrd. Is there some howto how to do an initrd for an arm rootfs? I used rootstock to build the rootfs. Oct 09 06:37:10 lool: here's the boot log if you're interested: http://koti.kapsi.fi/dpb/beagle.txt Oct 09 06:40:08 lool: what kernel do you use? what is recommended? Oct 09 08:15:42 dpb: For beagle, there are debian packages on the web Oct 09 08:16:08 dpb: http://www.rcn-ee.com/deb/kernel/beagle/karmic/ Oct 09 08:16:17 dpb: I run a self built kernel Oct 09 08:17:07 dpb: Your kernel cmdline looks good, not sure why you dont get further output Oct 09 08:17:16 dpb: I didnt reboot in a while since the new mountall arrived Oct 09 08:17:45 dpb: It might be that it's borken without an initrd (that should be supported though); to create one, use update-initramfs -c in the chroot Oct 09 08:18:23 perhaps using qemu if you dont have working ARM hardware Oct 09 08:18:42 dpb: I'm abroad right now, but could probably help more next week Oct 09 08:36:27 lool: ok, thanks Oct 09 08:51:40 lool: hrmm, I wonder why rcn-ee.com and rcn-ee.net has different kernels... .net has much more newer kernels too, with the same exact path otherwise... ;o Oct 09 08:52:45 Eh Oct 09 08:52:47 No idea Oct 09 08:52:53 not running that site Oct 09 08:52:58 I'm using my own kernels Oct 09 08:53:08 dpb: Otherwise, the linuxomap tree mostly works Oct 09 08:53:14 (tmlind'd) Oct 09 08:53:18 (tmlind's) Oct 09 10:06:21 grmbl, cdrdao seems to use $(uname -m) at buildtime Oct 09 10:25:04 ogra: fix it! Oct 09 10:25:15 ogra: I think Jamie was looking at it already Oct 09 10:25:22 ogra: Did you see the conversation on the debian patch already? Oct 09 10:25:34 no, i didnt know there is a debian patch Oct 09 10:25:44 ogra: You fail basic grep then :) Oct 09 10:25:59 i just looked at the FTBFS list and started a testbuild with the armv5tel files copied Oct 09 10:26:10 22:38 < JamieBennett> caused by a missing RULES/ file for armv7l Oct 09 10:26:11 22:38 < JamieBennett> has arm4l Oct 09 10:26:30 22:46 < lool> JamieBennett: debian/patches/10-rules-armel.patch Oct 09 10:26:32 funnily fakeroot debian/rules clean does run ./configure for about 40 times Oct 09 10:26:43 22:48 < lool> JamieBennett: Our armel architecture is v6 and Debian's is v4, the build shouldn't have to know about v4 or v7; it would be nice to foce the build to use the target arch instead of the detected buildd arch Oct 09 10:26:57 there is armv5tel Oct 09 10:27:07 Did you read debian/patches/10-rules-armel.patch ? Oct 09 10:27:14 nope Oct 09 10:27:26 ogra@dove:~/devel/cdrdao-1.2.2$ less debian/patches/10-rules-armel.patch Oct 09 10:27:26 debian/patches/10-rules-armel.patch: No such file or directory Oct 09 10:27:30 pfft Oct 09 10:27:42 22:50 < lool> JamieBennett: minimum change: add v7l because that's what our buildds are, good change: make the build use the target arch instead of buildd arch, perfect world change: fix the build to detect what it needs to know about instead of trying to support all arches in the world Oct 09 10:27:51 dpatch ... heh Oct 09 10:28:17 I don't know what you're looking at, I see the pathc just fine Oct 09 10:28:23 dpkg-source: info: extracting cdrdao in cdrdao-1.2.2 Oct 09 10:28:23 dpkg-source: info: unpacking cdrdao_1.2.2.orig.tar.gz Oct 09 10:28:23 dpkg-source: info: applying cdrdao_1.2.2-18ubuntu2.diff.gz Oct 09 10:28:27 vi debian/patches/10-rules-armel.patch Oct 09 10:28:38 its .dpatch :) Oct 09 10:28:43 It's not Oct 09 10:29:02 cdrdao (1:1.2.2-18ubuntu2) karmic; urgency=low Oct 09 10:29:25 ogra: We are not Debian :-) Oct 09 10:29:33 well, thats what the archive gave me Oct 09 10:29:41 cdrdao (1:1.2.2-17) unstable; urgency=low Oct 09 10:29:58 seems my apt cache isnt up to date Oct 09 10:30:13 ogra: My version being newer, I suspect you suck ;-P Oct 09 10:30:20 pfft Oct 09 10:30:25 ogra: You're so yesterday-ish Oct 09 10:30:30 heh Oct 09 10:30:48 Was pushed Oct 3 Oct 09 10:30:57 yeah, i'm working remotely Oct 09 10:31:10 my laptop has my approx instance for my LAN Oct 09 10:31:25 so the dove board doesnt find the up to date source Oct 09 10:31:54 haha Oct 09 10:32:57 As I was saying you suck :) Oct 09 10:33:13 I'm running a script to disable/enable my mirror when I'm traveling Oct 09 10:33:18 But mirror is on the LAN Oct 09 10:35:02 ogra: I should mock you again for your slowness Oct 09 10:35:09 :P Oct 09 10:35:29 he cdrdao code should really be fixed to not use uname though Oct 09 10:35:32 *the Oct 09 10:36:33 I agree Oct 09 10:36:46 Actually I said it earlier :) Oct 09 10:36:57 .oO(hungry .... it smells so good and i didnt have dinner) Oct 09 10:37:00 That said it's probably easier to fix it like Debian for karmic Oct 09 10:37:05 indeed Oct 09 10:37:06 ogra: You had 3 beers Oct 09 10:37:17 liquid bread doesnt count Oct 09 10:37:20 ogra: And at least 2 cigarettes Oct 09 10:37:23 more Oct 09 10:37:29 i go a smoker room :) Oct 09 10:37:33 *got Oct 09 10:37:33 OMG Oct 09 10:37:48 there is a bunch of other packages that use the same build system and will need same fixes Oct 09 10:38:00 basicly all packages from the same upstream author... Oct 09 10:38:03 who evet invented that should be shot Oct 09 10:38:08 *ever Oct 09 10:38:10 ah, k Oct 09 10:38:17 that explains something Oct 09 10:39:18 suihkulokki: (I omitted the part of the logs where I name the upstream and complain about him ;) Oct 09 10:40:05 he-who-name-whe-shall-not-mention Oct 09 10:43:39 do you fear he shows up at your door ? Oct 09 10:44:00 thats just because of the umlauts, no ? **** ENDING LOGGING AT Sat Oct 10 02:59:56 2009 **** BEGIN LOGGING AT Sat Oct 10 02:59:56 2009 **** BEGIN LOGGING AT Sat Oct 10 06:05:56 2009 Oct 10 10:56:06 well, at least webchat works Oct 10 10:58:24 hey JamieBennett Oct 10 10:58:47 ogra_n900: Guess you got it working then ;) Oct 10 10:59:14 nope, webchat.freenode.net FTW Oct 10 10:59:28 :) Oct 10 12:11:12 yay Oct 10 14:25:08 ogra_n900, you already have a n900? Oct 10 14:27:53 playya, yup, nokia gave them away at the maemo summit Oct 10 14:30:21 playya_: You dont? Oct 10 14:30:43 davidm has one, ogra has one, I have one Oct 10 14:30:49 Everybody has one these days Oct 10 14:31:01 I bet jamiebennett has one too Oct 10 14:31:56 * armin76 doesn't Oct 10 14:32:51 hehehe Oct 10 17:31:43 lool, no. I#m a poor student Oct 10 17:32:46 i hope ogra will be available in göttingen. i need a photo of the device in my hands ;) Oct 10 17:37:41 omg rootstock finally works! Oct 10 17:37:47 w00f :P Oct 10 17:55:58 roope@ruoska-ng:~/Tyokansio/TBuntu/images$ sudo qemu-system-arm -M versatilepb -kernel ./vmlinuz-2.6.28-versatile -hda TBuntu9.10.img -m 512 -append "root=/dev/sda mem=256M ro" Oct 10 17:55:58 Segmentation fault (core dumped) Oct 10 17:56:12 >_< Oct 10 17:57:43 stragce it Oct 10 17:57:46 strace it rather Oct 10 17:57:59 I've been making karmic rootstocks all day Oct 10 17:58:11 most work Oct 10 17:58:30 although I'm not using the stock versatile kernel (I have my own extensions) Oct 10 18:00:19 after cahnging -m 512 to -m 256 it booted, but now shows kernel panic Oct 10 18:00:58 feature :D Oct 10 18:01:08 yes **** ENDING LOGGING AT Tue Oct 20 20:16:14 2009 **** BEGIN LOGGING AT Tue Oct 20 22:16:15 2009 **** ENDING LOGGING AT Tue Oct 20 22:56:54 2009 **** BEGIN LOGGING AT Tue Oct 20 23:36:15 2009 Oct 21 01:34:48 NC, the beta livecd could successfully install to my SATA as long as network is configured. thanks! Oct 21 01:38:04 orga/lool, could we boot karmic from NFS rootfs? I did not see any wiki page talking about NFS. by this way, I could recover the rootfs easily once I make it crash. Oct 21 01:38:26 I mean, with Canonical provided prebuilt kernel. Oct 21 01:52:09 eggonlea: NFS Root on arm is currently untested, but "should" work. Oct 21 01:52:20 what platform? Oct 21 01:55:46 dove Oct 21 01:56:29 i saw NFS as module in .config Oct 21 01:57:15 is it done by initramfs? Oct 21 02:02:01 GrueMaster : I can confirm that NFS root works with the current rootstock Oct 21 02:02:08 I"m using it for Cortex A9 development Oct 21 02:02:16 cool Oct 21 02:02:25 just make sure you compile it -into- your kernel Oct 21 02:02:39 don't leave it as a module. Because initrd is -not- generically supported yet Oct 21 02:06:00 yes, I can use NFS rootfs with rootstock and uImage compiled with NFS built-in Oct 21 02:06:31 but could we use Canonical stock prebuilt kernel for that? Oct 21 02:43:30 You can, but I don't reccomend it Oct 21 02:43:40 (sorry, cooking dinner, so I'm not at the terminal very often) **** ENDING LOGGING AT Wed Oct 21 02:59:59 2009 **** BEGIN LOGGING AT Wed Oct 21 02:59:59 2009 Oct 21 07:56:23 We didn't try nfsroot with initrd I'm afraid Oct 21 07:57:30 what's your time, lool? Never thought I could met you at this time. :P Oct 21 07:59:05 why I want to try it is I do NOT know what's the difference between the rootfs installed from livecd and rootstock. Oct 21 07:59:37 eggonlea: I'm always here at this time but I might not be talking here :-) Oct 21 07:59:42 eggonlea: It's 10am here Oct 21 08:00:03 I popup between 8:30 and 9:30 almost everyday Oct 21 08:00:09 eggonlea: Paris time Oct 21 08:00:17 eggonlea: Where are you based? Oct 21 08:00:46 Shanghai, PRC Oct 21 08:00:58 so I always miss your hot discussion Oct 21 08:01:52 Eh Oct 21 08:04:46 lool: I have an open question. Oct 21 08:05:20 Will ubuntu continue use uboot as bootloader, or will transit to kexec with linux as bootloader? Oct 21 08:05:36 any brainstorming, discussion or plan? Oct 21 08:07:04 eggonlea: We wanted to do this in this cycle, but didn't manage to Oct 21 08:07:29 we saw the most two benefits to use linux as boot loader: 1) powerful scripts support and 2) reuse driver code instead of porting everything from kernel to uboot Oct 21 08:07:39 eggonlea: https://blueprints.launchpad.net/ubuntu/+spec/mobile-arm-karmic-softboot-loader Oct 21 08:07:47 https://wiki.ubuntu.com/Specs/ARMSoftbootLoader Oct 21 08:07:53 eggonlea: I fully agree Oct 21 08:08:09 eggonlea: There's one drawback in the form of slower boot times Oct 21 08:08:47 aha~. I need spend more time in walking through the LP/wiki Oct 21 08:09:05 eggonlea: That's what we discussed at last UDS Oct 21 08:09:19 eggonlea: Now we should probably look at discussing that at next UDS Oct 21 08:09:22 In November Oct 21 08:09:28 eggonlea: Anybody coming from your group? Oct 21 08:10:07 we're planning to apply for it Oct 21 08:10:20 but my visa is on "check" Oct 21 08:11:52 anyway, maen should have sb. there Oct 21 08:15:59 Cool Oct 21 08:17:25 btw, I noticed the deadline of the end of Oct. Could I apply for it without confirmation in case my visa won't be "clear" in time? Oct 21 08:35:46 eggonlea: Yes certainly Oct 21 08:36:14 eggonlea: We try to estimate the number of meals etc. Oct 21 08:36:45 eggonlea: One thing to do in the weeks before UDS is registering your attendance on Launchpad with the days you are exactly coming Oct 21 08:37:12 eggonlea: You should also subscribe to the specs you're interested in (not a lot at this point; do that one week before UDS for instance Oct 21 11:51:43 join the www.worldcommunitygrid.org and help conquer cancer Oct 21 13:06:28 Mysterious: Please drop the unrelated messages from this chan Oct 21 13:33:23 lool: lol, never Oct 21 13:35:01 Mysterious: sorry? Oct 21 13:41:14 lool: ehm (15:06:27) lool: Mysterious: Please drop the unrelated messages from this chan Oct 21 14:03:39 Mysterious: Yes, why "never"? Oct 21 21:45:13 Looks like Marvell announced Dove today, calling it Sheeva PJ4? Oct 21 21:45:19 http://www.linuxfordevices.com/c/a/News/Marvell-Armada-100-500-600-and-1000/?kc=rss Oct 21 21:45:54 Well, I guess that's the name of the core. Armada 500/600 for the SoC. Oct 21 21:46:27 ojn: heya Oct 21 21:46:29 cool Oct 21 21:46:56 plars: hey Oct 21 21:47:42 ARM/Freescale are handing out Pegatron Netbooks tomorrow too, but I missed it in time to register for that session so none for me. :( Oct 21 21:49:24 ojn: handing them out? how nice! Oct 21 21:49:56 I don't have one of those yet Oct 21 21:50:08 Heh Oct 21 21:50:37 ojn: where is this happening? Oct 21 21:50:37 plars: ARM Techcon3 Oct 21 21:50:37 Santa Clara, Wed-Fri this week Oct 21 21:50:37 ah Oct 21 21:50:52 forgot that was this week Oct 21 21:51:05 ARM announced and is pushing their new GPU stuff as well Oct 21 21:51:19 Oh, and they're announcing Sparrow tomorrow **** ENDING LOGGING AT Thu Oct 22 02:59:56 2009 **** BEGIN LOGGING AT Thu Oct 22 02:59:57 2009 Oct 22 07:57:27 ojn: Eh too bad you missed the give away Oct 22 11:39:55 lool Oct 22 12:01:40 saeed__: Yup Oct 22 12:04:38 how can cross build kernel package for dove? Oct 22 12:05:08 I tried: CROSS_COMPILE=arm-marvell-linux-gnueabi- CONCURRENCY_LEVEL=3 fakeroot make-kpkg --arch arm --initrd --append-to-version=-custom kernel_image kernel_headers --cross_compile - Oct 22 12:19:51 Don't use make-kpkg Oct 22 12:20:02 saeed__: http://kernel.ubuntu.com/ Oct 22 12:20:27 saeed__: debuild/dpkg-buildpackage -e CROSS_COMPILE=arm-marvell-linux-gnueabi- should work Oct 22 12:22:09 saeed__: https://help.ubuntu.com/community/Kernel/Compile Oct 22 12:22:31 saeed__: Or http://idlethread.blogspot.com/2009/01/recipe-of-day-cross-compiling-armel.html Oct 22 13:58:40 hmm, Mesa 7.6, anyone know of any plans to enable the GLES1/GLES2 part of it and provide packages? Oct 22 14:05:14 Stskeeps: What for specifically? Oct 22 14:05:53 I think we dont have userspace using GLES ATM; I think Clutter could use it but there are some bugs Oct 22 14:05:56 lool: in this particular case, building against GLES2 headers and then drop in stuff like SGX Oct 22 14:06:15 Right Oct 22 14:06:29 (i'm aware there's some potential pitfalls in that) Oct 22 14:07:33 I dont think we enable GLES support in mesa in karmic but I guess we could next cycle Oct 22 14:07:37 :nod: Oct 22 14:08:26 then there's also the issue of package naming, like, if it's libgles2-dev and so on :P Oct 22 14:09:31 Looks like we just need to --enable-egl in a separate build Oct 22 14:09:52 yeah, that was my thought Oct 22 14:10:54 the effect is being able to do GLES stuff in public and fallback on sw rendering, .. and then be able to drop in GLES libs for your device's chip :P Oct 22 14:26:06 hi Oct 22 14:26:31 is there howto install ubuntu arm on a Nokia Internet Tablet? Oct 22 14:27:16 thopiekar: just create rootfs with rootstock and install Mer hw-support stuff :) Oct 22 14:28:00 thanks, Meiz_webchat Oct 22 14:41:27 does anyone of you know how large the size of the rootfs should be to install it on the flash memory? Oct 22 15:47:17 hmm I wanted to create now a rootfs with rootstock but it uses jaunty.. Oct 22 15:47:33 I created myself a karmic chroot Oct 22 15:47:43 with kubuntu-netbook installed.. Oct 22 15:48:17 is there a way to get that chroot as a rootfs? Oct 22 19:11:26 ojn : where did ARM/Freescale hand out the pegatron netbooks? Oct 22 21:33:16 hi, what is the /lib/vfp directory?? , why it is installed on my target? some of my .so files resides there as well as at /lib/ - and they are different. any idea? Oct 22 21:36:25 /vfp is probably for binaries compiled with vfp support Oct 22 21:36:33 https://lists.ubuntu.com/archives/jaunty-changes/2009-March/007966.html Oct 22 21:38:50 related: http://www.mail-archive.com/ubuntu-bugs@lists.ubuntu.com/msg1395037.html Oct 22 21:44:17 Thanks. but i didn't understand why it was added?, why all the files are in different size than the same files in /lib or /usr/lib ? Oct 22 21:45:09 probably because they use vfp instructions to speed up floating-point calculation Oct 22 21:48:58 well, on my host, when cross-compiling, my libraries were taken from /lib and /usr/lib, but on the target applications use /lib/vfp , I guess this is a good reason for problems. Is there a recommendation what to use? Oct 22 21:53:12 um, shouldn't you have a separate tree with libs for the target system? Oct 22 21:54:16 you should probably either disable vfp when compiling, or install/crolss-compile vfp libs on the host system... Oct 22 21:57:14 actually, on the host I was used a copy of the target filesystem (using relative symbolic links) and pointing the LIBDIR to either /lib or /usr/lib. But then when executing on the target the application was loading the libs from /lib/vfp... Oct 22 21:57:50 can I be sure that every library installed in /lib or /usr/lib is also installed in vfp ? Oct 22 21:58:14 i think it will fall back to /lib if it's not found in vfp... Oct 22 22:01:43 so i need to define both paths for linking?, is there a way to define it with priorities?. I have to say that this issue sounds not convenient at all. i will try playing with that... Oct 22 22:02:25 maybe try running the program with LD_HWCAP_MASK=0x0 and see if it helps Oct 22 22:04:02 beyossi: We built two passes of glibc in jaunty Oct 22 22:04:08 One for vfp and one for non-vfp Oct 22 22:04:20 The versioning between the two on an installed system is very strict Oct 22 22:12:23 lool: for example, libdl.so.2 is taken from the /lib/vfp library (monitoring using ldd), while the same file name exist at /lib/libdl.so.2 - the libdl.so is at: /usr/lib/libdl.so and it points to ../../lib/libdl.so.2 . so i don't understand why the program is using /lib/vfp Oct 22 22:20:25 black magic Oct 22 22:21:26 i guess ld is checking hwcaps and adding /vfp to search path if hardware supports it Oct 22 22:23:21 but as far as I understand, a program will look for libdl.so (which exists only once) and not directly to libdl.so.2, and my libdl.so points to a specific place Oct 22 22:27:51 btw, I renamed library /lib/vfp to /lib/vfp2 and then everything points back to the 'old places' - latter on I will try building my entire target with the vfp Oct 22 22:27:59 Thanks for your help Oct 22 22:32:17 Another issue I face is related to executing QT on my beagleboard which has a PowerVR GPU. after executing the app it fails with: "powervr: driver not found", does anybody know what drivers it looks for exactly?, anyway to debug it? **** ENDING LOGGING AT Fri Oct 23 02:59:57 2009 **** BEGIN LOGGING AT Fri Oct 23 02:59:57 2009 Oct 23 06:20:13 how many shipments of ubuntu are allowed by canonical? Oct 23 06:29:22 darkham: what do you mean? Oct 23 06:31:51 amitk: in my shipit account i've a message about too many shipments. I requested 2 copy of ubuntu from 7.10 until now Oct 23 06:33:34 * amitk has no clue about shipit. This is probably the wrong irc channel anyways. See here: http://www.ubuntu.com/getubuntu/shipit-faq Oct 23 06:35:46 amitk, i asked because the faq haven't any information of that Oct 23 06:36:15 but i understand it's not properly a well known stuff... Oct 23 06:36:43 They say to contact info@shipit.ubuntu.com for questions Oct 23 06:48:25 amitk, thanx Oct 23 07:32:21 beyossi: Yes that's normal Oct 23 07:33:06 beyossi: You should just leave the /lib/vfp alone, it's fine Oct 23 07:33:17 beyossi: the runtime linker will pick it up if it's there Oct 23 07:44:07 hi, guys, I could not find a suitable channel to ask this question: does ubuntu have any image engine like Android's SKIA? I'm wondering how to optimize image (jpeg/png/gif/etc.) codec for ARM. Oct 23 07:47:25 eggonlea: afaik skia just calls libpng, libjpeg, libgif etc Oct 23 07:52:17 suihkulokki: I mean, Android provides a unified entrance for all applications. all applications should use SKIA to get optimization automatically. Oct 23 07:52:54 suihkulokki: e.g. we could replace libjpeg with any other optimized codec (if any). Oct 23 07:53:38 eggonlea: or you could just optimize libjpeg and and get faster jpeg not only in android, but in *all* linux applications Oct 23 07:54:33 suihkulokki: yes, indeed. what I want to confirm is that: is libjpeg the one all applications reply on. Oct 23 07:55:28 ok, all might be a bit extreme (I'm sure that there are exceptions), but it is pretty universally used Oct 23 07:55:34 suihkulokki: jpeg/png/gif/tiff/etc. and any others we should take care of? Oct 23 07:58:23 eggonlea: of bitmap formats bmp and pcx still appear sometimes, but I'm not sure they are common enough to warrant optimizing work Oct 23 07:59:22 then again, I am actually not a bitmap data format expert, that's just my personal experience :P Oct 23 08:07:00 thanks! I'll take a look at the above three codec first. Oct 23 15:37:06 Hey all .. Oct 23 15:37:16 Where did ARM/Freescale give away the pegatron netbooks? Oct 23 15:37:28 I saw something about it yesterday in channel Oct 23 16:17:54 Martyn: http://dpaste.com/111120/ Oct 23 16:35:17 Martyn: at techcon Oct 23 16:41:44 martyn: it was nettops, not netbooks. My bad. Oct 23 16:41:52 and it was at techcon, yes Oct 23 16:41:56 * ojn has one next to him now Oct 23 16:42:10 *groan( Oct 23 16:42:13 I could have used one of those Oct 23 16:42:15 oh well Oct 23 16:42:47 I'll need to bug my contact at ARM to see if I can get a hold of one Oct 23 16:43:11 I have so much advanced hardware now, but I don't have the /current/ state of the art since it's one half generation behind where we are Oct 23 16:43:29 and frankly, I want to be able to work on what's coming out now, as well as what will be out in a year Oct 23 16:48:10 martyn: Philippe Robin was the person doing the first "pitch" session where they were handed out. I guess he might be a good point of contact Oct 23 16:55:43 Martyn: it's always hard to get a hold of competitor's products before they ship in volume though. :-) Oct 23 17:38:28 Martyn: ask one for me :) Oct 23 22:44:37 hey all, are there any ARM netbooks out now that run Ubuntu well, or is at all developmental? Oct 23 22:45:48 only the sharp one Oct 23 22:46:12 http://www.electricpig.co.uk/2009/08/28/sharp-netwalker-the-future-of-netbooks/ Oct 23 22:46:35 not sure where genesi stands yet Oct 23 22:46:37 https://www.genesi-usa.com/products/smartbook Oct 23 22:47:26 thx Oct 23 22:47:46 it's weird cause I have been seeing tons of demo's of ARM netbooks for the last 3-6 months, but haven't found anything I can actually buy Oct 23 22:48:22 right, most are still not on the market Oct 23 22:48:30 for say the sharp netwalker, is it fully supported, or is there a lot of manual hackery to get it working? Oct 23 22:48:36 there are definately some to come Oct 23 22:48:46 it has ubuntu preinstalled Oct 23 22:48:59 oh, that's 5" though Oct 23 22:49:11 but as the article says, its not clear if you will ever get it outside of the asian market Oct 23 22:49:20 its a shame Oct 23 22:49:29 kind of want a 9-11", but have been waiting months Oct 23 22:49:33 since Atom is a POS Oct 23 22:49:34 i think genesi is a better baet, but i'm not sure the HW is out yet Oct 23 22:49:55 *bet Oct 23 22:49:59 anyone else that seems closer? or is genesi the main vendor to watch right now for it? Oct 23 22:50:13 in terms of shipping in the foreseeable future Oct 23 22:50:51 as i said there are more vendors to come but i dont know when or which or where they stand yet, genesi a a sure bet that wil happen at some point Oct 23 22:51:07 ya Oct 23 22:51:23 lastly, where is Ubuntu in terms of auto installing on ARM? Oct 23 22:51:36 does it all work, or is still reliant on manually bootstrapping the installer and such? Oct 23 22:51:57 the images we offer work on the hw they are built for Oct 23 22:52:14 sadly the HW is all development boards yet Oct 23 22:52:25 that's a shame Oct 23 22:52:35 they are live images and work identical to any other ubuntu live image Oct 23 22:52:50 how far would you say it is before the installer supports everything, without needing hw-specific images? Oct 23 22:53:01 never Oct 23 22:53:16 there is no single hw standard for arms, so kernel needs to be built for each system specifically Oct 23 22:53:16 is ARM that platform specific? Oct 23 22:53:17 you will always have SoC specific images Oct 23 22:53:20 yes Oct 23 22:53:26 i see Oct 23 22:53:59 So build a custom kernel, and then userspace is platform-agnostic across ARM? Oct 23 22:54:03 we wil surely able o loosen that a bit ... i.e. the imx51 image might at some point support all imx51 boards Oct 23 22:54:13 but you will never see a generic armel image Oct 23 22:54:18 ya Oct 23 22:54:54 custom kernel means that you likely end up with broken userpace unless your kernel build is a properly packaged ubuntu kernel package Oct 23 22:55:11 right Oct 23 22:55:20 so as a developer, how would you handle that? Oct 23 22:55:26 delegate kernels to each vendor? Oct 23 22:55:29 there are more and more things in userspace that rely on working initramfs for example Oct 23 22:55:35 or do different vendors fund you to support their particular hardware? Oct 23 22:55:50 well, we currently have one kernel tree per SoC Oct 23 22:56:00 and have one set of images per SoC Oct 23 22:56:35 do you have enough developers to maintain many? Oct 23 22:56:37 http://cdimage.ubuntu.com/ports/daily-live/current/ Oct 23 22:56:47 maintenance isnt the issue Oct 23 22:57:13 oh ok Oct 23 22:57:21 bringing it up to the point where its a usable image is the hard part ... i.e. building the first image for a new HW Oct 23 22:57:42 once thats done its only about maintaining the status Oct 23 22:58:12 oh Oct 23 22:58:21 so is the plan to focus on 1 or 2 ARM platforms? Oct 23 22:58:54 not necessarily Oct 23 22:59:19 if the community steps up and i.e. maintains a beagleboard image we'll surely wont complain Oct 23 23:00:02 right Oct 23 23:00:56 well anyway, thanks for the info Oct 23 23:01:42 but yes, atm the canonical arm team maintains only two images for teo platforms that re able to run an ubuntu desktop smoothly ... there might be more next release who knows Oct 23 23:02:01 s/teo/the **** ENDING LOGGING AT Sat Oct 24 02:59:56 2009 **** BEGIN LOGGING AT Sat Oct 24 02:59:56 2009 Oct 24 05:35:21 anybody know if the novatel mifi is linux based? Oct 24 19:04:58 Hrmm, is rootstock supposed to work with karmic? qemu-system-arm doesn't seem to exist in karmic... Oct 24 19:19:10 dpb, exists here.. Oct 24 19:20:21 http://packages.ubuntu.com/search?searchon=contents&keywords=qemu-system-arm&mode=exactfilename&suite=karmic&arch=any Oct 24 19:20:34 dpb, install qemu-kvm-extras Oct 24 19:31:07 dpb: It does Oct 24 19:31:24 dpb: Package is qemu-arm-static Oct 24 19:34:47 dpb, just install the rootstock package Oct 24 19:35:35 dpb, it pulls in everything it needs Oct 24 19:43:00 ah, there's a package for that too! :) Oct 24 19:52:02 indeed ;) Oct 24 20:54:24 ogra: why does rootstock use apt-get instead of aptitude in the second stage package install? Oct 24 20:55:06 who uses aptitude ? Oct 24 20:55:22 * ogra knows nobody that uses aptitude and is actually no fan of it Oct 24 20:56:03 in lucid i'll switch everything to tasksel btw Oct 24 20:56:22 and knit a gui to sit on top Oct 24 20:56:47 aptitude seems to have better detection on what to install, for example "xorg xserver-xorg-video-fbdev" <- apt-get doesn't care, it'll install ALL the non-used video drivers. aptitude is more sane and sees that video-fbdev provides all that's needed Oct 24 20:57:35 still i wont default to aptitude .... Oct 24 20:58:04 i'll happily take any patches that add an option to use aptitude for people actually wanting that Oct 24 20:58:25 i dont really like software that tries to be smarter than the user Oct 24 20:59:02 though apt-get or aptitude wont be a topic anymore once tasksel is the default ... Oct 24 20:59:16 yeah, aptitude sees what I want, apt-get doesn't. it's more stupid than the user. Oct 24 20:59:38 I don't even know what tasksel is... <.< Oct 24 21:00:00 sudo tsaksel Oct 24 21:00:04 *tasksel Oct 24 21:02:10 and tasksel will run the graphical curses aptitude UI if I want to select my packages... Oct 24 21:03:07 nope Oct 24 21:03:09 So in future versions people are forced to use either a group of packages defined by ubuntu, or use the horrible curses aptitude UI? Oct 24 21:03:27 it will just offer you the list of avilable ubuntu tasks to select from Oct 24 21:04:04 it wont use curses Oct 24 21:04:07 So people can't give it packages they want anymore? Oct 24 21:04:10 it will get a gui Oct 24 21:04:31 right, not for the basic setup, ou can easily modify it later Oct 24 21:05:23 with qemu-arm-static you can just chroot into the rootfs and modify it as if you would run it on an arm Oct 24 21:05:34 (you can do that already) Oct 24 21:05:59 the thing is that i want it to be as close to debian-installer as possible Oct 24 21:06:36 if i wouldnt have to jump through several hackish hoops to make that work i would even actually run d-i in a chroot **** ENDING LOGGING AT Sun Oct 25 02:59:56 2009 **** BEGIN LOGGING AT Sun Oct 25 02:59:56 2009 **** ENDING LOGGING AT Mon Oct 26 02:59:58 2009 **** BEGIN LOGGING AT Mon Oct 26 02:59:58 2009 Oct 26 03:11:00 Martyn: after looking at the pictures online of it, it seems like the Pegatron that was handed out is actually using the same design as the Genesi Efika MX. Just slightly different stuffing options (no HDMI, less flash) Oct 26 06:33:04 ojn: link? :) Oct 26 06:49:31 http://www.genesi-usa.com/products/efika Oct 26 06:49:34 (ah, google) Oct 26 06:58:57 Martyn: oh well, i thought he meant pictures of the pegatron nettop, not of the efika :) Oct 26 08:48:29 ojn: Both the pegatron nettop you got and the efika thingies are built by pegatron AIUI Oct 26 08:48:34 It's very similar stuff Oct 26 14:18:31 ogra: if you'll still accept that aptitude patch, here is one: http://koti.kapsi.fi/~dpb/tmp/rootstock-aptitude.patch.txt Oct 26 14:18:49 for lucid or a PPA package Oct 26 14:18:59 for karmic it's to late to add new features Oct 26 14:19:36 there's new versions somewhere? I didn't find a link on the rootstock page... Oct 26 14:19:51 (we release on thu. the archive is deep frozen) Oct 26 14:20:45 sure, but is 0.1.3 the newest rootstock anyway? Oct 26 14:22:28 should be, yes Oct 26 14:22:34 http://bazaar.launchpad.net/~ogra/project-rootstock/trunk/changes Oct 26 14:23:12 Gives me Internal Server Error Oct 26 14:23:35 hmm, intresting, here too now Oct 26 14:23:46 https://code.launchpad.net/~ogra/project-rootstock/trunk Oct 26 14:25:41 Anyway, if you're willing to add that patch to upstream, that'd be great. Oct 26 14:26:04 will, do, though i dont think its of any use once the whole thing is switched to tasksel Oct 26 14:31:16 yeah.. Oct 26 14:31:37 when are you planning on doing that? Oct 26 14:31:51 lucisd Oct 26 14:31:54 *lucid Oct 26 14:32:08 ok Oct 26 14:32:33 it should bre integrated as closely with the normal ubuntu installer as possible Oct 26 14:33:05 so we can make sure the resulting setup is as close as possible on one that would be installed from debian-installer or ubiquity Oct 26 14:33:21 s/on/to Oct 26 14:34:11 for me, Ubiquity is a firefox extension.. damn people using same names for stuff.. :S Oct 26 14:34:25 yeah, complain at FF Oct 26 14:34:49 the ubuntu installer is a few years older than the FF extension :) Oct 26 14:35:11 Yeah, I just looked at that on wikipedia.. :) Oct 26 17:01:58 Morning Oct 26 17:19:52 Martyn: Evening ;) Oct 26 17:20:01 Morning Martyn / JamieBennett Oct 26 17:20:18 Hey hi. Oct 26 17:20:29 * Martyn is in the middle of doing a cortex-A9 build. This takes for fucking EVER Oct 26 17:20:39 :) Oct 26 17:20:59 * JamieBennett is winding down for dinner before doing a whole lot more testing later Oct 26 17:24:34 Martyn, Cortex-A9 ? Oct 26 17:24:43 I thought it's coming 2010.. Oct 26 17:24:49 what board? Oct 26 17:24:55 We already have models and prototype hardware Oct 26 17:25:11 PBX-A9 in my case, plus software sim (Fastmodel) of our new SoC Oct 26 17:25:20 We have quad core :) Oct 26 17:25:35 sounds good :) Oct 26 17:27:29 Meizirkki: Evening for you too no? Oct 26 17:28:26 Evening yes, but http://www.total-knowledge.com/~ilya/mips/ugt.html :D Oct 26 17:28:44 :) Oct 26 17:29:09 It must be dinner time, I have this page open at the moment and I don't know why :) - http://aht.seriouseats.com/archives/2009/10/the-burger-lab-the-fake-shack-how-to-make-shake-shack-burger.html Oct 26 21:47:04 Anyone got a link to upgrade instructions jaunty->karmic for a redboot-based freescale board (i.e. babbage-like)? Oct 26 21:47:17 They provided some URLs at the conference but I didn't write them down and the slides are not yet up :) **** ENDING LOGGING AT Tue Oct 27 02:59:56 2009 **** BEGIN LOGGING AT Tue Oct 27 02:59:56 2009 Oct 27 09:04:55 ojn: We dont I'm afraid Oct 27 10:09:24 hi, i tried to install mythtv on jaunty on arm, but it complains about unmet dependencies for mythtv-frontend and mythtv-backend, any idea how to get mythtv on my system ? Oct 27 10:24:54 eFfeM: It might be that some packages failed to build to have the set of packages in mythtv installable; this might be better in karmic (didn't check) but karmic requires armv6 + vfp Oct 27 10:25:38 i know karmic requires armv6, and I am on a sheevaplug which is armv5 Oct 27 10:25:56 apparently people managed to do a manual build Oct 27 10:27:09 one of the issues seems to be that mythtv does not like qt4 so if libqt4-dev is installed it won't build Oct 27 10:27:13 http://computingplugs.com/index.php/Compiling_Mythtv-0.21_on_the_Sheeva_Plug Oct 27 10:27:33 doing this now ... Oct 27 16:59:07 ogra: I revamped most of the bits in https://wiki.ubuntu.com/ARM/BabbageKarmicInstall which had become inaccurate; mind reviewing? The link is broken but should work after release Oct 27 17:02:30 lool, looks ok, i was planning to add another subpage about how to switch install modes (oem, only-ubiquity, or if you feel like, integrity check) ... redboot-cmdline makes that easy Oct 27 17:28:36 ogra: That would be useful as I was playing with redboot stuff at the end of last week to little effect. Oct 27 17:32:43 what did you try ? Oct 27 17:33:14 * ogra just discovered a bug in redboot-cmdline ... it doesnt like slashes in cmdlines :) Oct 27 17:34:25 ogra: I discovered another, there is no documentation :P Oct 27 17:34:32 -h Oct 27 17:34:33 :P Oct 27 17:34:41 it only has two options Oct 27 17:34:59 err, three Oct 27 17:35:39 (the -o isnt actually needed on sanely set up redboot partitions) Oct 27 17:35:46 I was trying to change boot options with redboot-install but ... Oct 27 17:35:50 I gave up Oct 27 17:36:19 with redboot-install you also install a redboot binary, kernel and initramfs :) Oct 27 17:36:35 its purpose it to create a bootable partition Oct 27 17:36:43 doh :) Oct 27 17:36:45 you just used the wrong tool Oct 27 17:37:05 i agree it should grow manpages though :) Oct 27 17:37:16 its on my lucid todo already Oct 27 17:38:25 I was also hitting a brick wall with fconfig :( Oct 27 18:07:09 * ojn feels brave and does a dist-upgrade from jaunty to karmic on the pegatron Oct 27 18:07:21 ojn: tell us how that goes :-0 Oct 27 18:07:41 Only some people at Canonical have the hardware and they just use a karmic chroot on an external harddrive Oct 27 18:07:52 ojn: The biggest issue is lack of serial console, you need a special adapter I think Oct 27 18:08:12 i think you might run into kernel upgrade issues Oct 27 18:08:22 userspace should be flawless though Oct 27 18:08:44 lool: yeah, of course the system didn't come with that. I should be able to replace the kernel using the fconfig tools though, right? Oct 27 18:08:54 if it breaks it will break in initramfs setup or anything else related to the bootloader Oct 27 18:08:56 ojn: Yes; did you get kernel source? Oct 27 18:09:07 ojn: Our kernels don't support this particular board Oct 27 18:09:08 (or kernel package) Oct 27 18:09:11 lool: no idea Oct 27 18:09:22 ojn: Actually fconfig is for redboot configuration; you want "fis" to replace the kernel Oct 27 18:09:25 lool: really? Hmm. Oct 27 18:09:28 ojn: Oh wait Oct 27 18:09:33 ojn: Your kernel is actually in flash on these Oct 27 18:09:49 yeah, we wont break it ... Oct 27 18:09:50 ojn: fis/fconfig are userspace tools to replace stuff on a SD Oct 27 18:10:02 lool: Ugh. double fail. Oct 27 18:10:02 but the packages might fail to install/upgrade Oct 27 18:10:35 as i said, bootloader, initramfs or kernel ... one of the three will likely break Oct 27 18:10:37 Ok. I'll just archive the hardware until there's reasonable hardware support then. Oct 27 18:11:27 back to the ti boards instead. At least I can upgrade those myself. Oct 27 18:12:09 its unlikely that we'll ever support initramfs less setups ... which i think the pegatron uses Oct 27 18:12:26 (in the distro releases that is) Oct 27 18:13:04 ojn: put gentoo on it and gimme access *g* Oct 27 18:13:13 Right. I used initram-less setups all the time on ppc Oct 27 18:14:04 armin76: Want me to put speed holes in it first? Oct 27 18:14:06 right, ubuntu uses initramfs for way to many things in userspace ... the general distro stuff will go on using it pretty sure Oct 27 18:14:43 oem images that are tailored for a certain task (like the jaunty on your pegatron) might not ... but then you have a cut down in features Oct 27 18:14:47 ojn: yes :D Oct 27 18:15:39 ogra: what's the approach to deal with the platform proliferation anyway? I'm hoping both marvell and freescale allow multi-board kernels, even if arm overall doesn't allow multi-platform (marvell+freescale for example) kenrels at this time? Oct 27 18:15:48 ojn: Did I mention that your kernel probably lacks a flash driver? Oct 27 18:16:00 lool: yeah, looks like this is a piece of useless junk. Oct 27 18:16:04 ojn: I think the driver is ready "upstream" now but I dont know whether it's public Oct 27 18:16:17 thats why its free :) Oct 27 18:16:17 ojn: *cough* sorry; that should fix itself pretty soon though Oct 27 18:16:25 lool: that's sorta moot since I can't even boot a different kernel as far as I can tell Oct 27 18:16:32 lool: how? without a flash driver, how do I replace the kernel? Oct 27 18:16:46 armin76: sorta defeats the purpose, doesn't it? Oct 27 18:17:07 ojn, afaik the lange patches arent fully compatible with the babbage patches ... which likely keeps us from having a generic imx51 kernel Oct 27 18:17:54 cool, my only-ubiquity install passed fine :) Oct 27 18:17:58 ogra: that will just be a world of pain then Oct 27 18:18:16 but i'm sure people will work on fixing it over time. or they will go insane. Oct 27 18:18:34 rightr Oct 27 18:18:44 luckily thats in the realm of the kernel team :P Oct 27 18:18:51 i only use their packages :) Oct 27 18:19:08 so they have to find a way :) Oct 27 18:19:14 ojn: So below the board you have DIP switches Oct 27 18:19:21 yeah. i'm guessing it's just not top prio for the first platforms Oct 27 18:19:21 ojn: You can make it boot on micro SD Oct 27 18:19:29 ojn: Which is also below the board Oct 27 18:19:30 lool: ah, cool Oct 27 18:19:37 yeah, I noticed that Oct 27 18:19:43 ojn: That's the setup which Efika MX / Genesi uses Oct 27 18:19:48 can't you netboot it? Oct 27 18:19:57 tftpboot, that is Oct 27 18:20:07 not sure of redboot has tftp client? Oct 27 18:20:23 it can get kernels from http at least Oct 27 18:20:29 but you need to configure it Oct 27 18:20:37 which means you need access to fconfig Oct 27 18:20:45 which in turn doesnt support flash atm Oct 27 18:20:49 or at least a serial console? :) Oct 27 18:20:59 can't you do it on the bootloader itself? Oct 27 18:21:39 ojn, well, if you have a proper redboot setup, redboot-tools will just work Oct 27 18:22:09 ogra: ok. I will tread lightly until I have a serial console available so I at least can recover from misconfiguration. Oct 27 18:22:26 ojn, which means you need neither a serial console nor flash support ... use an SD card use redboot-install and redboot-cmdline with your own kernel and initramfs Oct 27 18:22:44 ogra: heh, ok Oct 27 18:22:52 use the DIP switch method lool pointed out Oct 27 18:22:54 serial console is cool Oct 27 18:23:23 ojn: does it have vga? Oct 27 18:23:27 then use an SD with redboot-install, your initramfs and kernel ... then you can use userspace/rootfs as you like Oct 27 18:24:21 armin76: yes, it has vga but the framebuffer driver initializes (late) in linux Oct 27 18:24:44 ogra: thanks, will take a look. need to order another microsd, the one I have I use on the overo. Oct 27 18:26:15 https://wiki.ubuntu.com/ARM/BabbageImageFromScratch should largely get you going (you will likely need your own redboot binary too for lange though) Oct 27 18:26:17 why didn't they use uboot? Oct 27 18:26:32 uboot is underway ... but still immature Oct 27 18:26:43 and redboot works ... Oct 27 18:27:32 u-boot isn't exactly great either Oct 27 18:27:33 and is now configurable without second machine, at least on our images/installs Oct 27 18:27:56 ogra: whats the difference between the babbage and lange? Oct 27 18:28:23 different SoC same arch ... so different kernel patches and bootloaders Oct 27 18:28:42 neither was published in time to make karmic Oct 27 18:29:19 i want lange support aside babbage in lucid though ... so i hope to get kernel and redboot patches for next round Oct 27 18:31:50 Anyone here with a nokia n900, btw? Oct 27 18:32:06 ogra has one Oct 27 18:46:34 I think I scared him away. :) Oct 27 18:53:48 ojn: try #maemo for n900 users Oct 27 18:54:05 ojn, yes, it could run ubuntu, if you truely wanted :P Oct 27 18:54:41 it wouldn't even be that hard Oct 27 18:54:57 we have multiboot already, and when you have something poking the watchdogs.. Oct 27 19:12:51 yeah, that was a bit offtopic. Wasn't asking with anything ubuntu in mind for it, just looking for someone who has one Oct 27 19:15:49 we have multiboot already, and when you have something poking the watchdogs.. Oct 27 19:15:52 er. Oct 27 19:17:18 ojn: I have an n900 too, so does ogra and Stskeeps at least Oct 27 19:18:18 JamieBennett: cool. see /msg. :) **** ENDING LOGGING AT Wed Oct 28 02:59:58 2009 **** BEGIN LOGGING AT Wed Oct 28 02:59:59 2009 Oct 28 10:36:04 Hi, has anyone else experienced a sudden, unexpected logout from the live session in the recend imx51 live images? I experienced this in 20091027.1 and now also 20091027.2 Oct 28 10:37:07 In both cases jockey-gtk crashed, but I don't know whether that's the cause... the bug report for this (#403955) suggests this crash is harmless though. Oct 28 11:01:25 dmart, nope, works fine here Oct 28 11:01:37 i'm just doing an oem install Oct 28 11:02:54 Unless it really is somehow related to the jockey-gtk issue, I don't have the faintest idea what happened :( There's no extra apport crash dump. Oct 28 11:03:45 might be triggered by jockey looking for hardware Oct 28 11:04:24 i'll try it in my next install test Oct 28 11:30:27 * ogra puts up https://wiki.ubuntu.com/ARM/BabbageInstallVariants (which exposes a bug in redboot-tools :/ ) Oct 28 11:31:11 if anyone is intrested in trying oem installs or ubiquity only installs on the babbage, see above Oct 28 12:48:02 hi folks Oct 28 13:19:49 * ogra tries an install to prepartitioned SD Oct 28 13:20:43 * kblin tries to patch a debian package for the first time in years Oct 28 13:22:42 dmesg Oct 28 13:23:05 hrm, wrong kbd Oct 28 13:23:15 top Oct 28 13:23:41 * ogra sighs Oct 28 13:24:34 dmart, jockey runs fine here, no crashes Oct 28 13:25:13 ogra: hmmm, it doesn't seem to happen every time for me, just sometimes Oct 28 13:25:31 strange Oct 28 13:31:15 dmart: I had all the issues you describe at various times Oct 28 13:31:29 (crash of jockey or other stuff, and logout of the session albeit only once) Oct 28 13:31:54 lool: Did it seem to you that those two things were related? Oct 28 13:31:59 whoa, the deps for devscripts are insane Oct 28 13:35:01 dmart: No Oct 28 13:35:15 dmart: I also get an occasional indicator-applet crash Oct 28 13:35:36 I think the occasional crashes are just bugs in the apps which are visible on armel due to races rather than specific platform porting issues Oct 28 13:35:48 And I think the logout thing is either a xorg bug or a kernel bug Oct 28 13:38:16 lool: I get the indicator-applet crash sometimes too; it was to intermittent for me to be able to report it yet, though. Do you know if there is a bug on this? Oct 28 13:47:49 dmart: I filed one today Oct 28 13:47:56 http://launchpad.net/bugs/462538 Oct 28 13:47:58 Launchpad bug 462538 in gnome-panel "gnome-panel crashed with SIGSEGV in free()" [Undecided,New] Oct 28 13:55:26 lool: Hmm, I had one of those, too. It seemed to respawn gnome-panel rather than killing the login though: Oct 28 13:55:28 http://launchpad.net/bugs/458109 Oct 28 13:55:30 Launchpad bug 458109 in gnome-panel "gnome-panel crashed with SIGSEGV in free()" [Medium,Incomplete] Oct 28 13:56:10 http://launchpad.net/bugs/461912 Oct 28 13:56:14 Launchpad bug 461912 in gnome-panel "gnome-panel crashed with SIGSEGV in gdk_window_impl_x11_get_colormap()" [Medium,Triaged] Oct 28 17:21:37 plars, when you did manual partitioning on Dove, how did you create the partitions? Oct 28 17:23:12 NCommander: I did it a few different ways Oct 28 17:23:26 NCommander: first time, I did a 100M /boot, and the rest as ext4 for / Oct 28 17:23:41 NCommander: when I got the warning message, I only had a single 20G / which was ext2 Oct 28 17:23:56 NCommander: are there any other possibilities you think I should try? Oct 28 17:24:09 plars, you used "New Partition Table" right? Oct 28 17:24:37 NCommander: no, don't think so... usually just remove the partitions that are there manually and create my own Oct 28 17:24:59 plars, yeah, I might have just found a bug. I'm reconfirming the behavior happens on a fresh start Oct 28 17:25:23 NCommander: with new partition table? Oct 28 17:25:40 plars, yeah. If you create it, then add a new uboot partition, the mountpoint gets stuck on / versus /boot Oct 28 17:25:48 even if you change it manually Oct 28 17:25:55 (or it changes to /home!) Oct 28 17:26:11 strange, I can try to confirm here too if you'd like Oct 28 17:26:14 But I was abusing partman on that run, and might have gotten it stuck in some strange error mode, hence why I'm trying on a fresh start Oct 28 17:33:41 plars, crud, it happens on a fresh install. Can you try and confirm? Oct 28 17:33:59 will do Oct 28 17:34:14 plars, wait, I might have figured out what's causing this strange behavior Oct 28 17:34:24 plars, (its still a bug, but not quite as serious as I first realized) Oct 28 17:34:53 plars, if you make the partition exactly the same size as the one you just erased, there is some sorta memory effect Oct 28 17:36:03 plars, but the mountpoint still ends up wrong :-/ Oct 28 17:36:13 ps -ef |grep ubiq Oct 28 17:36:18 bah Oct 28 17:36:22 wrong keyboard Oct 28 17:36:32 wow it's taking a long time to launch Oct 28 17:39:46 plars, ? Oct 28 17:39:49 It came up pretty fast for me Oct 28 17:41:33 at what point do you see it? while you are still in the partitioner? or after? Oct 28 17:42:01 plars, while still in the partitioner Oct 28 17:42:10 so far, so good for me Oct 28 17:42:19 I had a single 20G ext2 / last time Oct 28 17:42:24 I did new partition table Oct 28 17:42:26 plars, basically, do Manual partitioning, new partition table, then create a u-boot partition first, with a size of 200-300 MiB at the start with a mountpoint of /boot Oct 28 17:42:34 oh wait Oct 28 17:42:47 yes, I see the mountpoint is wrong on my uboot partition Oct 28 17:42:51 it has / instead of /boot Oct 28 17:42:56 yup Oct 28 17:42:57 :-/ Oct 28 17:43:17 plars, it used to work though Oct 28 17:43:29 Because I can get it do that if I delete all partitions, and then try and add it Oct 28 17:47:40 ah Oct 28 17:47:42 I see something here Oct 28 17:48:06 so I tried deleting it, and created it over 3 times... every time now it comes up and forces itself to be /home Oct 28 17:48:09 but Oct 28 17:48:27 if I delete it and recreate it as ext2 instead of uboot type, it works and puts itself in as /boot Oct 28 17:48:32 plars, yup Oct 28 17:49:02 plars, I think I know what's causing this behavior, I think I missed an edit when I addedd support for the partman-uboot mountpoints Oct 28 17:50:44 so it works 90% of the time Oct 28 18:00:53 NCommander: did you open a bug on it already? Oct 28 18:01:23 plars, not yet, I'm doing a full install because it seems hitting New Partition Table actually zaps the partition table the moment you hit it Oct 28 18:01:54 plars, and I want to confirm that this bug only occurs when the partition table is empty (and doesn't occur when you manually remove partitions) Oct 28 18:02:15 (and I want to install the dev environment for ubiquity on armel) Oct 28 18:05:56 plars, if you want to file a bug, feel free, otherwise I'll do it in a little bit Oct 28 18:06:27 I can do it... against ubiquity? or where? Oct 28 18:08:25 plars, ubiquity. Use apport in the installer environment; it will properly attach all the logs Oct 28 18:27:15 plars, what's the bug number? Oct 28 18:27:43 give me a sec... it's opening now but swapping between 3 different installs at the moment Oct 28 18:27:58 plars, heh. Isn't install testing week fun? Oct 28 18:28:09 it's crazy, very little sleep :) Oct 28 18:28:25 plars, reminds me of UDS ;-) Oct 28 18:28:43 I got WAY more sleep during UDS Oct 28 18:29:11 respins dropping at 11PM makes for a long night Oct 28 18:34:21 NCommander: bug #462798 Oct 28 18:34:23 Launchpad bug 462798 in ubiquity "selecting 'new partition table' confuses the partitioning" [Undecided,New] https://launchpad.net/bugs/462798 Oct 28 18:36:00 plars, thanks. I marked the ISO test as failing since I couldn't complete installation :-/ **** ENDING LOGGING AT Thu Oct 29 02:59:57 2009 **** BEGIN LOGGING AT Thu Oct 29 02:59:57 2009 Oct 29 16:45:20 i hear someone was working on maemo in ubuntu. stil true? Oct 29 16:48:44 pwnguin, Have a look at Mer http://wiki.maemo.org/Mer Oct 29 16:50:01 ive seen mer Oct 29 16:50:10 i just grabbed a vbox image last night Oct 29 18:56:27 Martyn: 9.10! Oct 29 19:13:59 lool: Do you happen to have info on what the different DIP switches on the pegatron/efika control? No silk screen info. :( Oct 29 20:55:02 Is ubuntu 9.10 supported on the ARMv5te architecture, or is what I've been hearing true that it requires ARM6 or later with vfp? Oct 29 20:55:25 ogra: ^ Oct 29 20:58:17 Karmic is ARMv6 and above Oct 29 20:58:49 (big difference between ARMv6 and ARM6 :-) Oct 29 20:59:13 right, v6 is what I meant. As in, won't run on v5te without vfp Oct 29 21:00:11 any particular reason ? debian still supports v5te iirc Oct 29 21:01:11 I run ubuntu on my sheevaplug (which is ARMv5te-based without a vfp), and was hoping to be able to do a straight dist-upgrade instead of having to completely wipe and load a new disi Oct 29 21:01:32 *dist Oct 29 21:02:33 darkstar62: unfortunately that's not possible. It is a bit of a bummer since the sheevaplug is a neat platform, too bad Marvell had been lagging behind on the architecture for so long. Oct 29 21:04:08 is it lack of resources to maintain the two that's the problem? I'm kinda surprised there would only be the one release for v5 Oct 29 21:04:49 eFfeM: v6 is needed for SMP (A9) Oct 29 21:05:04 No lock primitives in v5 I think? Oct 29 21:05:39 hmm, I could be wrong on that though Oct 29 21:05:49 since there's v5 smp chips. :) Oct 29 21:05:57 ojn no idea about whether there are lock primitives Oct 29 21:06:13 but guess this means non-smp support is drooped ? Oct 29 21:06:45 I don't see why that would be dropped Oct 29 21:07:27 as you said v6 is needed for smp Oct 29 21:07:33 yes? Oct 29 21:07:39 there are v6 non-smp systems too Oct 29 21:07:45 sure Oct 29 21:08:22 but if you can have systems w/o smp then why could armv5te not be supported ? Oct 29 21:10:05 No idea. It probably makes sense from a business decision though; especially if the chip vendors are paying for the ports. It's bad marketing for them to say "get this distro we sponsored that doesn't take advantage of our hardware". Oct 29 21:10:14 indeed Oct 29 21:16:26 well, if that's the case, could the community work with the ubuntu devs to support armv5? Oct 29 21:16:56 Is anyone stopping you? Oct 29 21:17:00 I have no idea, but anyone is free to do their own spin Oct 29 21:17:23 I was thinking more alongside ubuntu to have a ubuntu version for armv5 instead of a spin-off or fork Oct 29 21:17:32 I'd just run debian on it though. since it's headless, and alot of the differentiation between ubuntu and debian is on the desktop, you're not missing all that much Oct 29 21:18:01 mm, true Oct 29 21:18:39 I moved back to debian when ubuntu abandoned ppc (for my ppc boxes), and I'm not missing much. Oct 29 21:18:55 ojn: thing is the plug comes with ubuntu :) Oct 29 21:19:07 yeah, I'd have to completely re-image Oct 29 21:19:17 which I can do, I had just hoped I wouldn't have to Oct 29 21:19:22 ah Oct 29 21:19:28 bummer Oct 29 21:20:08 yep, ironic :) Oct 29 21:20:09 darkstar62: compiling the entire archive for armv5 would be a bit pointless don't you think? Oct 29 21:20:28 (since debian already supports it) Oct 29 21:21:11 for headless-operations I'd agree Oct 29 21:21:35 but there are some shivaplug users using usb video/keyboard/mouse combos to use it as a desktop Oct 29 21:21:50 debian has desktop as well :P Oct 29 21:21:55 true Oct 29 21:22:06 it would require an archive for armv5, optimised v6 and (who knows) in the future armv7 and so on Oct 29 21:22:13 but there again, the shiva plug comes with ubuntu, so they'd have to completely reinstall Oct 29 21:22:34 sounds like something you should talk to the shiva plug providers about Oct 29 21:22:40 darkstar62: you have to blame marvell for that :) Oct 29 21:22:41 did they coordinate with canonical? Oct 29 21:22:59 my guess is no, since it's not supported officially by canonical Oct 29 21:23:04 right Oct 29 21:23:06 so they're on their own Oct 29 21:23:12 and yeah, definitely it's marvell's issue on that Oct 29 21:24:34 amitk: well, really all it would take is to provide something at a low enough common denominator. With glibc multilib support some of the perfomance can be recovered (but not all). Oct 29 21:24:37 well, I guess if there isn't much call for an armv5 port outside some shivaplug users, it doesn't make sense to setup a whole system for just that Oct 29 21:25:57 amitk: i think the next after karmic is armv7 only Oct 29 21:26:16 armin76: would make sense to be able to move to thumb2 Oct 29 22:46:28 armin76: some folks with ubuntu cloaks said otherwise, btw Oct 30 01:44:19 * pwnguin just realized there's no c in persia's name Oct 30 01:44:28 now it makes a lot more sense why he's in .jp Oct 30 01:45:17 Actually, the two are completely unrelated (although it does make transliteration easier) **** ENDING LOGGING AT Fri Oct 30 02:59:58 2009 **** BEGIN LOGGING AT Fri Oct 30 02:59:58 2009 Oct 30 09:21:07 does any body use ZSYNC to download 9.10? It reports 302 error. Seems ubuntu.com changed the file location and zsync cannot deal with http redirection. Oct 30 09:22:17 I could use zsync for 9.10 alpha, beta and rc. Oct 30 09:22:52 but failed on this final one. Oct 30 09:27:45 [answer my own question] just use redirected cdn.release instead of the release server. :P Oct 30 09:30:15 note that the servers are under heavy load during the first 10 days after release usually Oct 30 09:31:28 /nod Oct 30 09:32:32 hope all are using bt or zsync instead of downloading .iso/.img directly with multithread sw Oct 30 14:48:42 oh well, I'm giving up for now. ARM hardware doesn't seem to be quite ready for prime-time yet Oct 30 14:50:38 can't build products on systems that decide to sometimes do the right thing Oct 30 15:08:05 kblin: you are aware that how well usb works on beagleboard is not representantive of how all other arm works with usb? Oct 30 15:08:33 suihkulokki: yeah, I tried a shevaplug as well Oct 30 15:10:35 that was even worse :) Oct 30 15:11:22 kblin: hmm? apart for the bootloader not supporting usb, usb has been solid for me on sheeva Oct 30 15:11:47 does IPv6 work for you? Oct 30 15:12:01 no, the sheeva had other issues Oct 30 15:12:13 not the least dying on me after two weeks Oct 30 15:12:59 and the globalscale support really sucks, especially if you're outside the US Oct 30 15:14:12 haven't had a ipv6 setup for a while Oct 30 15:15:56 the default kernel shipped with the sheeva claims to support ipv6, but is missing the kernel module Oct 30 15:16:37 which severly screws up applications that try to support ipv6 Oct 30 15:18:06 I didn't get around to try a custom kernel, as my plug died before I could test Oct 30 16:06:30 hi, i'm getting the mobile phone toshiba TG01, with a snapdragon processor Oct 30 16:07:58 I want to install ubuntu on the phone, but i can't seem to find what to install on it. Oct 30 16:09:02 the snapdragon is based on ARM that is why I ended up here Oct 30 16:09:32 can someone tell me if I can install ubuntu-arm on the phone Oct 30 16:09:35 thanks Oct 30 16:10:36 unfortunately there's much more variety between ARMs than between x86-based chips Oct 30 16:11:07 hard to give definite answers there Oct 30 16:13:11 tordne: after a lot of work messing around with all sorts of stuff, you would in the end have a phone that you can't make calls with. Is that really what you want? Oct 30 16:14:20 @ojn: yes Oct 30 16:14:27 @no :-D Oct 30 16:14:49 see, that's why open source is great Oct 30 16:15:04 it's all about choice.. even if the choice is shooting yourself in the foot Oct 30 16:15:07 I don Oct 30 16:15:20 Linux is free only if your time has no value... :) Oct 30 16:16:29 @ojn: no I wouldn't mind to mess up the phone Oct 30 16:17:38 @ojn: but i like linux, and can program a bit... maybe this is a start for me to realize something Oct 30 16:40:05 @ojn: I think I found something better "Mer linux" Oct 30 16:41:45 Once I've created a tarball from the rootfsfromscratch, how do I install that to an sd card such that it is bootable? I tried just creating a big ext2 filesystem and extracting inside it but the arm board seemingly didn't boot from it Oct 30 16:42:18 (I had already figured that it wouldnt work but gave it a shot anyway) Oct 30 16:43:06 BeardedChimp: It depends on your board. Some firmware expects to read the kernel (or even the firmware itself) from a FAT partition at the beginning of the media. It's more of a board than a distro question Oct 30 16:45:20 I have an image of their edge install floating about, I'm guessing I could work it out by having a peak inside? Oct 30 16:45:56 probably, yes Oct 30 16:50:43 http://pastebin.com/m2e5b4365 I think this is the relevant bit from the docs, bit beyond me at the minute Oct 30 16:53:54 BeardedChimp: hope you didn't used karmic as rootfs Oct 30 16:54:32 No :) Oct 30 16:54:48 I figured that would be suicide Oct 30 16:55:18 The board comes with sarge installed on it, but with support dropped for it, it aint great Oct 30 16:55:35 dist-upgrade to etch? Oct 30 16:56:01 tbh im looking for >.27 kernel Oct 30 16:56:07 Which is why I went for ubuntu **** ENDING LOGGING AT Sat Oct 31 02:59:57 2009 **** BEGIN LOGGING AT Sat Oct 31 02:59:57 2009 Oct 31 09:31:55 ojn: I don't have info on the different DIP switches, albeit I expect them to be similar to babbage boards; I have access to babbage board info but it's inconvenient for me right now, and I have access to people with pegatron boards too Oct 31 09:32:04 I could look at this next week Oct 31 09:32:41 ojn: Would you mind pining JamieBennett / ogra on this topic? They are both related to this kind of projects more than I am **** ENDING LOGGING AT Sat Oct 31 11:39:06 2009 **** BEGIN LOGGING AT Sat Oct 31 11:39:30 2009 Oct 31 15:46:42 lool: thanks, will do **** ENDING LOGGING AT Sun Nov 01 02:59:56 2009