**** BEGIN LOGGING AT Tue May 18 02:59:57 2010 May 18 05:01:12 IMAGE_FSTYPES="ext3 jffs2 yaffs2" , yet a yaffs2 image is not being created?? only jffs2, ext3 and tar.gz images are being created , no yaffs2 ?? May 18 05:02:57 Spyzer: strange and no error ? May 18 05:03:12 no :( May 18 05:03:38 i haven't done a git pull in a long time in the openembedded directory, should I?? May 18 05:03:55 long time? May 18 05:03:58 how long May 18 05:04:05 1 month... May 18 05:04:29 may not help May 18 05:04:38 hmm :( May 18 05:04:39 but try May 18 05:05:35 do i have to give IMAGE_FSTYPES = "yaffs2" or "yaffs" ?? May 18 05:07:09 yaffs2 May 18 05:07:18 Spyzer: where id you put your IMAGE_FSTYPES? maybe try bitbake -e image | tee env.log and check value in env.log May 18 05:08:52 khem: i have made my own recipes in openembedded direcotry so on git pull i get something like "error: Your local changes would be overwritten by merge. Aborting." ?? May 18 05:09:14 you can stash them May 18 05:09:25 git stash;git pull --rebase;git stash pop May 18 05:09:25 please tell me how do i?? May 18 05:09:34 ok May 18 05:18:50 instead of my build dir i just mistakenly did a a bitbake -e console-image in my openembedded dir. Will this create problems ?? May 18 05:55:50 np May 18 05:55:55 no May 18 06:10:48 khem: let's keep talking here, since it will be more oe-related chatting as for now May 18 06:10:54 Hi, I need help on generating "helloworld" Image for "x86" MACHINE and "celinux-test" DISTRO. May 18 06:12:09 Nikunj: whats your question May 18 06:12:48 Nikunj: Please dont PM random people. Ask your questions here. May 18 06:15:14 Question is : I am getting an error when i run "$bitbake /path/to/helloworld-image.bb " the error is " NOTE: Handling BitBake files: \ (6783/8369) [81 %]NOTE: :list indices must be integers, not NoneType while evaluating: http://projects.linuxtogo.org/frs/download.php/14/prismstumbler-0.7.4pre1.tar.gz file://gpsapi.patch;patch=1 ${@['', 'file://wireless.patch;patch=1'][bb.data.getVar('PREFE May 18 06:18:23 Nikunj: what version of bitbake are you using? May 18 06:19:36 Khem : its 1.8.18 May 18 06:19:58 Nikunj: Can you try using master May 18 06:20:23 Khem : master? May 18 06:20:30 git clone git://git.openembedded.net/bitbake May 18 06:20:35 and use that one May 18 06:20:58 Khem : Ok, let me try.. May 18 06:21:49 my qemu compiling task is failing again and again ?? May 18 06:22:10 i am doing bitbake console-image for MACHINE=mini244 and DISTRO=angstrom0 May 18 06:22:20 DISTRO=angstrom-2008.1 May 18 06:22:56 Khem : It says "git clone http://git.openembedded.net/bitbake Initialized empty Git repository in /home/nikunj/oe/BB/bitbake/.git/ Cannot get remote repository information. Perhaps git-update-server-info needs to be run there?" May 18 06:23:09 Nikunj: btw. celinux-test is quite old, I would recommend using something more recent and maintained like minimal for distro May 18 06:23:13 when i check the sources directory theres is no such fileas qemu-0.14.tar.gz?? May 18 06:23:38 Khem: hmm.. ok. May 18 06:23:42 sorry qemu-0.12.4.tar.gz i mean?? May 18 06:23:53 Nikunj: you did not do it right look at the command I gave you May 18 06:23:59 its git protocol not http May 18 06:24:27 Spyzer: whats the error May 18 06:25:02 Khem: Yes, I tried both..I am behind the firewall..tried with "git:// " also.. but it says "Initialized empty Git repository in /home/nikunj/oe/BB/bitbake/.git/ fatal: Unable to look up git.openembedded.net (port 9418) (Name or service not known) fetch-pack from 'git://git.openembedded.net/bitbake' failed." May 18 06:25:02 Spyzer is it qemu-native ? May 18 06:25:16 no May 18 06:25:42 how do i forcibly donwload any package?? May 18 06:25:43 looks like you're having trouble with your dns May 18 06:25:45 Spyzer: go into your src directory and delete all qemu-* files May 18 06:26:00 Nikunj: hmm May 18 06:26:37 Nikunj: try "ping git.openembedded.net", see if it works May 18 06:26:38 khem: i just hope that after updating openembedded the yaffs2 image gets created now May 18 06:26:43 Khem : btw, "http_proxy" variable is also set.. May 18 06:27:06 doesn't matter if ping is reaching it or not, just the name resolution May 18 06:28:11 Nikunj: If u r behind a proxy then i doubt u can download git and cvs based solutions until u have the availability of a dns server May 18 06:28:18 Lint2 : nope.. ping is not reaching..says " unknown host " , both with "git://" and "http://" May 18 06:28:53 Nikunj: do you have a local dns server runing?? May 18 06:29:32 can you tell us what's your dns settings? May 18 06:30:12 Nikunj: try pinging 208.67.222.222 May 18 06:31:40 03Roger Monk  07org.openembedded.dev * r725f882c29 10openembedded.git/recipes/ti/ti-eula-unpack.inc: (log message trimmed) May 18 06:31:40 ti-eula-unpack: Force Console Mode, add Error Prints May 18 06:31:40 * ti-eula-unpack runs installjammer installers to install ti components May 18 06:31:40 * installjammer can run in GUI mode or console mode May 18 06:31:40 * ti-eula-unpack passes options and string paths to the installer May 18 06:31:41 which need to be run in console mode May 18 06:31:42 * GUI mode is automatically selected if an X11 Display is found which May 18 06:32:00 Spyzer: It got stuck.. May 18 06:32:39 Lint2 : How to check dns settings.. sorry I am a newbie. network manager is not opening.. May 18 06:32:39 u mean its not pinging?? May 18 06:32:45 Spyzer: it should get created I just tried it here May 18 06:33:23 Nikunj: cat /etc/resolv.conf and then checkout the nameserver lines if available May 18 06:33:29 Nikunj: try this as root: cat /etc/resolv.conf May 18 06:33:32 Spyzer : yeah..the prompt is idle... May 18 06:33:52 Nikunj: Ask your network admin to either allow access or set the access somehow because you need internet access to get OE working May 18 06:33:54 khem: yes i am trying that too May 18 06:34:46 Lint2 and Spyzer : yes its showing "nameserver" entries. May 18 06:35:01 Nikung: even though u might have internet access with squid running u might not be allowed to access a dns server, ask ur network admin to give u opendns server access May 18 06:35:12 Nikunj:and what r those?? May 18 06:35:24 I would recommend google's dns May 18 06:35:43 Spyzer: nameserver 43.88.62.92 nameserver 43.68.8.1 May 18 06:35:55 you just have to use "nameserver 8.8.8.8" May 18 06:36:21 So should i edit resolve.conf file..? May 18 06:36:23 it's the nameserver I've just used to start downloading oe here (I'm doing that right now) May 18 06:36:29 yep May 18 06:36:36 kool.. letme try.. May 18 06:36:38 comment the old ones May 18 06:36:45 with a # May 18 06:41:51 Lint2 : I have done the settings..did the network restart also.. same results.. not able to ping. May 18 06:43:50 Lint2 : I think since dhcp is enabled.. its automatically taking same old dns settings.. what i did is gone now.. May 18 06:44:17 that's likely May 18 06:44:30 I never saw that happen before, but makes sense May 18 06:44:34 oh, wait May 18 06:44:46 you executed some network initialization script? May 18 06:44:52 you shouldn't do that May 18 06:44:58 just change the dns back to 8.8.8.8 May 18 06:45:09 and it should work rightaway, no re-init needed May 18 06:51:22 Lint2 : same results..!! :( May 18 06:52:05 what's the command you're trying to execute? (the git cmd) May 18 06:54:02 git clone git://git.openembedded.net/bitbake May 18 06:55:23 hold on May 18 06:55:32 * khem -> sleep () // good night all May 18 06:56:07 'night khem May 18 06:56:11 thanks for the help May 18 06:56:15 I'm still downloading it May 18 06:56:20 Khem : bunch of thanks.. May 18 06:56:45 try this: git clone git://140.211.169.165/bitbake/ May 18 06:56:51 Lint2: yeah have fun look into conf/machine and conf/distro May 18 06:57:17 I'll surely do if I'm not asleep by 20 min May 18 07:01:01 Lint2 : "fatal: unable to connect a socket (Connection timed out) fetch-pack from 'git://140.211.169.165/bitbake/' failed." May 18 07:03:36 you have to use a proxy to access web, right? May 18 07:03:54 yeah May 18 07:04:17 humm May 18 07:04:39 I guess there's no way to access such thing, so May 18 07:05:09 earlier I thought that you had direct access to it (often called 'nat' connection), but that error tells me that you don't May 18 07:05:30 and sadly you can't use http proxy to transport git May 18 07:05:52 bottom line..? May 18 07:06:00 why don't you try just to find a release of it somewhere in the net? May 18 07:06:27 which release I think I have the latest stable version of bitbake. May 18 07:07:30 I don't get it.. what did you mean? (english is not my spoken language) May 18 07:07:35 I mean May 18 07:07:52 try to find a file online (via http) that have all the files that this git command will generate for you May 18 07:08:22 hmm ok. May 18 07:09:23 there was a link on the site saying something for people between proxies, if I'm not mistaken May 18 07:10:20 ok..let me check May 18 07:15:09 03Khem Raj  07org.openembedded.dev * rad5668a7ac 10openembedded.git/recipes/uclibc/ (26 files in 13 dirs): (log message trimmed) May 18 07:15:09 uclibc-git: Add libm.so to dummy shared objects May 18 07:15:09 * Generate dummy libm.so in uclibc-initial because May 18 07:15:09 ppc gcc-cross libgcc links with libm and it May 18 07:15:09 needs the notion of libm when building gcc-cross-intermediate May 18 07:15:10 * Cleanup machine config files and rename from .config to .machine May 18 07:15:11 * Bump SRCREV to latest master May 18 07:15:18 Lint2 : I am goin for a break..will catch you in sometime..? May 18 07:26:02 morning May 18 07:27:59 hi hrw May 18 07:29:16 hi May 18 07:43:26 good morning May 18 08:55:20 03Eric Bénard  07org.openembedded.dev * r39b030e6da 10openembedded.git/classes/cross.bbclass: (log message trimmed) May 18 08:55:20 cross.bbclass: fix build of meta-toolchain May 18 08:55:20 * without this bitbake meta-toolchain fails because it doesn't find who May 18 08:55:20 provides gdb-cross-sdk May 18 08:55:20 * found thanks to a hint from Phil Blundell on IRC May 18 08:55:21 * tested with Angstrom/ARMv5te May 18 08:55:21 Signed-off-by: Eric Bénard May 18 08:55:27 03Eric Bénard  07org.openembedded.dev * rf2f3cff6cf 10openembedded.git/classes/module-base.bbclass: May 18 08:55:27 module-base.bbclass: keep PR if MACHINE_KERNEL_PR is not set May 18 08:55:27 Signed-off-by: Eric Bénard May 18 08:55:27 Acked-by: Phil Blundell May 18 09:12:21 please help me out anyone, I am doing IMAGE_FSTYPE = "yaffs ext3 jffs2" and yaffs doesn't get created only jffs2 and ext3 get created, please help me..... May 18 09:13:00 are the tools for yaffs on your system? May 18 09:13:06 like?? May 18 09:13:09 and does OE know about them? May 18 09:13:22 Spyzer: 07:07:08 < khem> yaffs2 May 18 09:13:44 Spyzer: are you following the answers? or do you still have yaffs instead yaffs2? May 18 09:14:33 i have written yaffs2 in IMAGE_FSTYPES May 18 09:14:41 as khem specified :( May 18 09:14:56 but do i need to have any yaffs tools before hand May 18 09:15:05 11:12:20 < Spyzer> please help me out anyone, I am doing IMAGE_FSTYPE = "yaffs ext3 jffs2" and yaffs doesn't get created only jffs2 and ext3 get created, please help me..... May 18 09:15:13 doesn't look like yaffs2 to me May 18 09:15:26 Spyzer: yesterday I told you to check conf/machine/include/neuros.inc & conf/machine/neuros-osd2.conf which is the only machine using yaffs2 in OE actually May 18 09:15:35 that was written in a hurry so "2" was left out i have alwasy written yaffs2 in the original .conf file ;) May 18 09:16:12 so can't i set mini2440 to use yaffs?? May 18 09:16:15 Spyzer: yaffs2-utils-native should be built before yaffs2 image automatically May 18 09:17:03 ericben: may i elaborate the root of problem to u?? May 18 09:17:09 if u have time May 18 09:17:57 good morning May 18 09:18:02 morning May 18 09:21:00 i added yaffs2 in mini2440.conf file in OE/conf/machine directory will that help making yaffs2 images for mini2440 May 18 09:21:01 ? May 18 09:21:57 ericben: have u burnt or downloaded any images to mini2440?? May 18 09:22:05 Spyzer: I don't have mini2440 May 18 09:22:27 Spyzer : there is one important parameter do pass to mkyaffs2image which is the layout# May 18 09:22:35 actually it's set to 1 in bitbake.conf May 18 09:22:56 maybee you need to set it to 0 in your machine's conf file like in neuros.inc May 18 09:23:10 all of this really depends on how yaffs2 is handled on your machine May 18 09:23:15 I think May 18 09:24:53 * has anyone downloaded fs images to mini2440?? May 18 09:33:37 i got disconnected so i ask again :) May 18 09:33:44 * has anyone downloaded fs images to mini2440?? May 18 09:34:21 Spyzer: http://www.angstrom-distribution.org/narcissus/ May 18 09:40:26 hrw: it is not showing option for yaffs2 image type?? May 18 09:40:45 so? May 18 09:40:51 use ubifs? May 18 09:42:27 actually the main problem i am facing is that, after i build kernel and fs for mini2440 and the moment i try to burn it(through minicom) minicom shows download a yaffs image to the device which i don't have May 18 09:42:43 also after ignoring and downloading a "jffs2" image instead of yaffs May 18 09:42:59 the kernel boots up with error imnages like image nmot maountable May 18 09:43:06 and kernel panic not syncing errors May 18 09:55:59 03Koen Kooi  07org.openembedded.dev * r48849e915d 10openembedded.git/contrib/angstrom/extract-feedarch.sh: extract-feedarch: script to extract machine and feedarch from angstrom-feed-config packages May 18 09:55:59 03Koen Kooi  07org.openembedded.dev * r34f67a4260 10openembedded.git/contrib/angstrom/sort.sh: angstrom feed sorter: add more machines May 18 09:58:37 florian: on this page http://narnia.cs.ttu.edu/drupal/node/131 is a way to burn images(and that too jffs2) on mini2440 but he uses a flash sd card. I used the minicom itself to burn the images May 18 09:59:08 is there any problem in using minicom instead of supervivi terminal?? May 18 10:01:41 * florian is using minicom too May 18 10:02:59 florian: do u use the three options of donwloading the vivi and downloading the kernal and the fs image as shown in minicom or do u use the supervivi terminal by pressing "q" on minicom?? May 18 10:03:18 i just use the default options in minicom May 18 10:04:03 Spyzer: What I did is more or less exactly this: http://labs.kernelconcepts.de/Publications/Micro2440/ May 18 10:11:03 armv7a-angstrom-linux-gnueabi/usr/lib/libattr.la: No such file or directory while building recipes/gettext/gettext_0.17.bb do_compile failed May 18 10:11:20 What would be the solution to avoid this error May 18 10:12:15 Spyzer: Guessing from your error message one likely missing step was to set the bootargs. May 18 10:13:27 When i compile the package gettext_0.17.bb iam facing this problem May 18 10:13:37 idth.lo wait-process.lo xmalloc.lo xstrdup.lo xerror.lo xmalloca.lo xreadlink.lo xsetenv.lo xstriconv.lo xvasprintf.lo xasprintf.lo acl.lo argmatch.lo asnprintf.lo file-has-acl.lo fnmatch.lo gettimeofday.lo printf-args.lo printf-parse.lo safe-read.lo safe-write.lo strnlen.lo tempname.lo vasnprintf.lo wcwidth.lo libcroco_rpl.la libglib_rpl.la | /bin/grep: /media/disk/project/open_embedded_base/build/overo-oe/tmp/staging/armv7a-angstrom- May 18 10:13:55 /bin/sed: can't read /media/disk/project/open_embedded_base/build/overo-oe/tmp/staging/armv7a-angstrom-linux-gnueabi/usr/lib/libattr.la: No such file or directory May 18 10:17:31 Will you please share more information about the bootargs May 18 10:19:17 libattr.la was missing how to get that file in staging folder May 18 10:26:03 hello good morning, does someone know if is it possible to change the var CROSS_COMPILE to use a custom toolchain to build a bb file (kernel recipe)? May 18 10:29:10 03Martin Jansa  07org.openembedded.dev * re2a7e630d4 10openembedded.git/recipes/linux/ (3 files in 2 dirs): May 18 10:29:10 linux-openmoko-2.6.34: add new version May 18 10:29:10 Signed-off-by: Martin Jansa May 18 10:29:21 03Martin Jansa  07org.openembedded.dev * r96b22b8e5a 10openembedded.git/recipes/freesmartphone/ (cornucopia.inc fso-specs_git.bb libfso-glib_git.bb): May 18 10:29:21 freesmartphone: bump cornucopia SRCREV for fixed +CMT format and other fixes May 18 10:29:21 Signed-off-by: Martin Jansa May 18 10:36:15 ERROR: 'org.openembedded.dev/recipes/gettext/gettext_0.17.bb' failed May 18 10:36:34 Did any one facing problem in compiling gettext_1.17.bb ? May 18 10:37:25 Ananth: I have but it fails while building libiconv pulled by gettext.. May 18 10:37:41 Ananth: but it was few days ago.. May 18 10:38:13 and it stil fails.. May 18 10:40:44 For me it still fails and in the staging process i cannot find a needed lib file (libattr.la) May 18 10:40:49 Any idea ? May 18 10:41:14 nope. May 18 10:55:24 seems it's managed by the TARGET_PREFIX, thank you anyways May 18 10:55:54 florian: in many tutorials given on the internet about downloading kernel and fs images to mini2440 they use supervivi terminal instead of default options provided in vivi(thru minicom) why is that?? May 18 10:56:25 Spyzer: No idea... never used it. May 18 10:57:23 Spyzer: btw.. its absolutely sufficient to ask a question one time in one channel. asking the same question the same people in different channels is useless. May 18 10:58:34 ohh sorry May 18 10:59:55 florain i meant that why do we have to load everything using commands like load ram 0x31000000 239016 u May 18 11:00:19 why can't we just use the donwload kernel, download viv option given in the vivi boot menu?? May 18 11:01:09 u gave me the link http://labs.kernelconcepts.de/Publications/Micro2440/ there also they are following the same procedure, instead of default options in vivi they go into the supervivi terminal and give specific memory values?? May 18 11:01:10 No idea... the only thing I used it for was to get U-Boot into NAND. May 18 11:01:29 and then... what abt the kernel and fs?? May 18 11:02:31 how did u manage to get the kernel and a gpe-image or something on the board?? May 18 11:05:26 also my mini2440 has 256Mb nand flash, so do the memory addresses given on many mini2440 supervivi tutorials work for me as well?? May 18 11:10:15 florian: how did u manage to download gpe-image or kernel to the board?? May 18 11:11:05 Spyzer: Like in this description I mentioned... May 18 11:14:08 so u used a flash sd card to download those to the board, but i don't have one so is there any other way u might now. I mean through usb or serial cable itself, using ./s3c2410_boot_usb itself?? May 18 11:14:14 *now know May 18 11:16:30 Spyzer: never tried this... of course you can use the network connection and tftp May 18 11:17:28 hmm.. when the uboot gets deployed on the board can i set up an ip of the board?? May 18 11:17:38 yep May 18 11:19:38 florian: here http://code.google.com/p/mini2440/wiki/MiniBringup is a guide to flash mini2440 through tftp, but don't u think the memory addresses given there can change for mini2440 with 64 M and the other with 256 M?? May 18 11:19:59 i mean can i safely use this guide ?? May 18 11:25:13 florian how'll i exactly use tftp?? i mean in the sc card's case we had "fatload mmc ...." what shall be the way in tftp?? May 18 11:27:44 Spyzer: you have some command like "tftp $address $filename" May 18 11:29:31 i know that much :), but how do i use that command with fatload?? May 18 11:29:49 afterall fatload will load the downloaded information to 0x31000000 May 18 11:31:52 use the same address for downloading via tftp May 18 12:00:50 U don't need fatload if you use tftp command (because you don't have to do anything with fat anymore) May 18 12:08:37 Spyzer : I am not able to solve the problem..!! error with bitbake.. May 18 12:41:26 Hi! We have massive problems with ppc603 and g++ 4.2.4 converting long long to double. The conversion yields wrong results. May 18 12:43:32 Unfortunately qt-embedded 4.5.2 is full of such conversions. ppc405 uses g++ 4.1.1 but that versions produces incorrect values, too. May 18 12:44:04 Has anybody seen similar problems? Some go away with -O3 but not all. May 18 12:45:29 vps, make a small example and post it to the list May 18 12:45:37 OK, will do! May 18 12:46:24 BTW: fp-int-convert-long-double.c from the gcc testsuite fail too :-o May 18 12:51:15 double check the site/ppc files, someone might have messed up on one of the tests May 18 12:51:25 therefore forcing gcc to do the wrong thing May 18 12:54:52 XorA: site/ppc? You are talking about oe/recipes/gcc/gcc-4.2.4/ ? May 18 12:56:22 vps: no, the forced autoconf tests in the site/ directory in root of OE May 18 12:59:32 XorA: will look at it. Not that those easy to understand :-D May 18 13:02:17 vps: yeah they can be a bit obtuse, but it seems odd that gcc fails that badly :-D May 18 13:02:30 vps: worth asking khem as well May 18 13:12:00 The strange thing is that gcc is working fine. It's just C++ code which fails. I think I will post a complete report to the list May 18 13:53:45 hi gconf_2.28.0.bb does not build here, http://tinderbox.openembedded.org/packages/574400/ anybody else have this? configure: error: C compiler cannot create executables May 18 13:53:54 other recipes are ok May 18 14:02:58 how do I make bitbake fetch the latest svn rev in a recipe? May 18 14:05:04 vonami|work: SRCREV = ${AUTOREV} May 18 14:05:42 JaMa: I'll try, thanks May 18 14:06:03 or better put some sane revision in recipe and add SRCREV_pn-yourRecipeName = ${AUTOREV} in local.conf or some other config you're including May 18 14:06:14 ao2: works fine here (armv5te/angstrom/x86 build system) May 18 14:06:39 vonami|work: see shr-autorev.inc or fso-autorev.inc for examples May 18 14:07:10 eriben, thanks, I'll try cleaning up tmp/ then May 18 14:28:17 morning kergoth May 18 14:55:40 florian: linuxtogo disk is out of space May 18 15:03:37 hi, it seems i have an upgraded mini2440 board and so when i boot into uboot it says nand_info() cannot allocate oob_buf May 18 15:04:31 here http://code.google.com/p/mini2440/wiki/MiniBringup its given that this error can be handled by replacing 'git clone git://repo.or.cz/u-boot-openmoko/mini2440.git' with 'git clone git://gitorious.org/hiteg-development-boards/mini2440.git' May 18 15:04:38 how do i do this in openembedded?? May 18 15:06:33 i got it May 18 15:06:52 but i didn't get that what shall be the branch?? May 18 15:13:07 in other words how do i get uboot 2k version from oe bitbake?? May 18 15:17:47 Does anyone know of an easy way to create a package that includes an image file generated from an image task? May 18 15:24:13 if i do bitbake u-boot_2009.1.bb will it generate a "compatible" u-boot for my mini2440?? May 18 15:26:38 anyone please help me, uboot is giving me error like nand_scan(): Cannot allocate oob_buf, anyone please tell me how to make a newer uboot for mini2440..... May 18 15:26:40 please May 18 15:32:46 please tell me how do i build u-boot-nand-2k for mini2440 through oe?? May 18 15:50:56 3~/quit() May 18 16:04:29 03Michael Jones  07org.openembedded.dev * r8bd97afd23 10openembedded.git/docs/usermanual/ (8 files in 2 dirs): May 18 16:04:29 docs/ spelling fixes May 18 16:04:29 Just a few things in the documentation I noticed while reading up on OE. I don't May 18 16:04:29 understand why 'git diff' included "\ No newline at end of file" for some of May 18 16:04:29 these, or how I could instruct it not to, but I also didn't want to pick apart May 18 16:04:30 my patch outside of git. May 18 16:04:31 Signed-off-by: Michael Jones May 18 16:31:42 hi, while trying to install prelink_20090925 with opkg, it complains of a missing dependency which is elfutils-dev : it there any reason for prelink to depends on a -dev package ? May 18 16:52:33 good morning all May 18 16:53:01 hi khem May 18 16:53:10 eriben: hello May 18 16:54:23 vps: fp-int-convert-long-double.c failure hmmm. Can you use gcc 4.4.4 and retry with it and also post your gcc -v output for the ppc compiler May 18 16:54:56 vps: It could be a bug in long double conversion routines May 18 17:04:31 heh libiconv needs gettext and gettext needs libiconv May 18 17:04:36 another catch-22 May 18 17:09:33 khem: how to build a gcc cross tool chain?? May 18 17:09:39 sorry i got disconnected May 18 17:10:16 is there something wrong with the toolchain oe builds? May 18 17:11:53 no, kergoth_ the question i asked has indirecvt relationship with my problem. My problem is that i am unable to use the "normal" uboot bcoz my mini2440 is an upgraded version, so i need to specifically build a new uboot-nand-2k version and hence i ask , how to bitbake a gcc-cross-toolchain so that i can use it externally May 18 17:13:25 Spyzer: if you already built an image then you have the toolchain May 18 17:13:40 if not you can build one using bitbake meta-toolchain May 18 17:13:57 the toolchain is in /cross May 18 17:17:52 concerning the fact that prelink depends on elfutils-dev, the problem is that the libs of elfutils are installed in elfutils-dev May 18 17:18:11 is this fix to elfutils_0.143.bb the right one ? May 18 17:18:13 +FILES_${PN} += "${libdir}/libdw-0.143.so ${libdir}/libelf-0.143.so ${libdir}/libasm-0.143.so" May 18 17:18:58 at least that works with this fix (prelink now depends only on elfutils) May 18 17:19:44 Question: How do you turn off the GNU_HASH check per recipe again? Not the LDFLAGS check but actually turn it off. May 18 17:21:57 DJWillis: INSANE_SKIP_${PN} = True ? May 18 17:22:14 DJWillis: not sure May 18 17:22:59 Nope, I can't remember either, got a recipe that builds it's own 'plugins' but they are not strictly Linux ELF SO's so the sanity check barfs. May 18 17:24:35 DJWillis: I think that's INSANE_SKIP, in insane.bbclass it skip all the checks at line 430 May 18 17:25:18 eriben: yes thats right solution May 18 17:25:30 eriben: thanks for the pointer, yep, just trying that. May 18 17:26:16 anyone for my question concerning elfutils ? May 18 17:27:57 khem: export CROSS_COMPILE=what?? should i give the directory path to the toolchain?? May 18 17:29:03 Spyzer: full path to cross compiler (leave out gcc in the end) like arm-linux-gcc then arm-linux- May 18 17:29:39 kergoth: if one has disabled NLS do we need gettext on the system May 18 17:29:51 ok May 18 17:30:33 eriben: odd, I wonder why the SO's of elfutils are ending up in the dev package, your solution looks ok. May 18 17:32:46 DJWillis: it seems that when libs end with .so they end in -dev and that usually .so is a link to a binary ending with x.y.z but elfutils its the reverse : .so is the binary and .so.1 the link May 18 17:33:29 at least a few libs I looked at are splited this way May 18 17:35:21 Normally plain .so ends up in the -dev and the versioned .so ends up in the versioned lib package so you can install more than one version of the .so simultaneously. May 18 17:36:37 broonie: ok, so the real problem is in the way elfutils is versioning its libs, I'll try to look for a better fix May 18 17:44:15 please help me out, i followed the instructions for uboot here http://wiki.linuxmce.org/index.php/Mini2440 and yet when i am trying to do go 0x31000000 in mini2440 it is not executing?? May 18 17:51:26 hi all May 18 17:53:05 I get a "Requested 'glib-2.0 >= 2.23.6' but version of GLib is 2.22.1" error for the gtk+-native recipe May 18 17:53:13 what's the best way to solve this issue? May 18 17:53:42 I've read some previous discussion about this in the mailing list, but it wasn't quite conclusive. May 18 17:54:50 mini2440 players, please listen to me... May 18 18:11:06 I just deleted all glibc-2.0-*-native recipes to get rid of the "Requested 'glib-2.0 >= 2.23.6' but version of GLib is 2.22.1" problem - May 18 18:11:16 hi i built a mini2440 uboot as given here http://wiki.linuxmce.org/index.php/Mini2440 and tried to load it in the device as given here http://mini2440.spaceopera.org/site/developmentenviroment/stepbystep/index.html but i get stuck at line "go 0x31000000" it gives such error: May 18 18:11:16 - hope I don't cause any catasthropes :) May 18 18:11:31 argument 0 = 0x00000000 May 18 18:11:31 argument 1 = 0x00000000 May 18 18:11:31 argument 2 = 0x00000000 May 18 18:11:31 argument 3 = 0x00000000 May 18 18:11:42 and then mini2440 hangs May 18 18:11:44 please help May 18 18:13:01 i was building a uboot of my own rather than using the uboot deployed by oe because i have a 256M flash nand which requires this sort of a special treatment May 18 18:21:49 Spyzer: you should always execute first from ram to see if uboot binary is ok May 18 18:25:55 yeah i do that... but finally i have to write uboot into the device. i followed the steps mentioned in the tutorials and yet i am stuck, can u possibly help me May 18 18:29:00 Spyzer: I have never handled mini2440 May 18 18:29:18 but may be the addeess it got flashed to is wrong May 18 18:29:25 or you missed a step in procedure May 18 18:29:43 reflash the working uboot May 18 18:33:24 today i am as frustrated as hell khem, i have just tried every possible thing i can to burn my_oen_customized image in the mini2440 but no progress, i just can't even boot the uboot :( :'( May 18 18:43:40 03Denys Dmytriyenko  07org.openembedded.dev * r9074fefdb8 10openembedded.git/recipes/meta/external-toolchain-csl.bb: May 18 18:43:40 external-toolchain-csl: set the individual licesnes for packages May 18 18:43:40 * glibc is LGPLv2.1 May 18 18:43:40 * libgcc/libstdc++ is GPL (v2/v3) with GCC Runtime Library Exception May 18 18:43:40 Signed-off-by: Denys Dmytriyenko May 18 18:47:09 Spyzer: try go 0x33f80000 May 18 18:47:16 Spyzer: try go 0x33f80000 May 18 18:49:19 i have tried that as well ericben May 18 18:49:43 please share anything u know May 18 18:51:40 please... i am dying May 18 18:53:12 ericben: what "can" be the possible problem here?? May 18 18:53:42 if i know the problem only then i can even try to search for a solution else its all blind May 18 18:53:57 I know nothing of this board but there seems to be plenty ressources on the net : http://github.com/tekkamanninja/U-boot-2009.08_tekkaman seems to have fixed the 256MB nand problem May 18 19:09:00 ericben: the link u gave i downloaded the source but when I am trying to set CROSS_COMPILE, May 18 19:09:21 I am unable to do so as in the Makefile it is hardlinked with a preset May 18 19:09:53 cross_compile how do i change this, i tried to replace it by my compile toolchain by giving the ful path but no use May 18 19:10:49 btw if i accidentally delete a file in the dir, how do i get it again thru git?? May 18 19:11:12 Spyzer: change CROSS_COMPILE to your toolchain's prefix May 18 19:11:44 i did but am still getting the errors May 18 19:12:28 Spyzer: study the error message and fix the problem they tell May 18 19:12:51 i accidentally deleted the makefile how do i get it again thru git?? May 18 19:15:45 ericben: please i am newbie with git, ho do i get a file again when i have accidentally deleted it, i am doing git pull but it says already up to date, git checkout shows that Makefile needs to be downloaded but doesn't do anything?? May 18 19:16:06 git chekout Makefile May 18 19:17:22 thnx May 18 19:18:14 hi everyone, i have a question to git as well. I'm trying to specify my local git repository in SRC_URI with file:/// , and it doesn't work with message ERROR: SRCREV was used yet no valid SCM was found in SRC_URI. Is there way to specify local git repo? May 18 19:20:51 re May 18 19:24:07 ahh, ok, file:// is used for patches and staff... hmmm.... I don't setup apache for git serving May 18 19:27:08 lgopher: SRC_URI = "git:///home/ebenard/full_path_to_git_dir/test;proto=git May 18 19:27:23 03Khem Raj  07org.openembedded.dev * rde76b53196 10openembedded.git/recipes/gcc/ (2 files in 2 dirs): May 18 19:27:23 gcc-4.4.4: Fix build failure on powerpc/uclibc. May 18 19:27:23 * For uclibc __to_type is typedef'ed to const __ctype_touplow_t* May 18 19:27:23 which is a short int. So the mangled alias breaks because on May 18 19:27:23 glibc __to_type is int* and mangled alias is set accordingly May 18 19:27:24 Signed-off-by: Khem Raj May 18 19:30:36 03Khem Raj  07org.openembedded.dev * rdbb9b89e02 10openembedded.git/recipes/util-linux-ng/util-linux-ng.inc: May 18 19:30:36 util-linux-ng.inc: Dont add -lintl when USE_NLS="no" May 18 19:30:36 Signed-off-by: Khem Raj May 18 19:33:00 03Khem Raj  07org.openembedded.dev * r59599c1473 10openembedded.git/recipes/gnutls/gnutls.inc: May 18 19:33:00 gnutls.inc: Obtain gettext in DEPENDS from inheriting gettext bbclass May 18 19:33:00 . May 18 19:33:00 Signed-off-by: Khem Raj May 18 19:34:12 ericben: still doesn't work, tries to get package from angstorm.org and blows obviously SRC_URI = "git:///home/oe/projects/${PN}.git;protocol=git;branch=master" May 18 19:34:40 lgohper: here that works May 18 19:35:05 ericben: is test in the end important ? May 18 19:35:30 no thats the name of the directory containing the git tree May 18 19:35:45 is ls /home/oe/projects/${PN}.git a directory on your PC ? May 18 19:36:08 yes and it's git bare repository May 18 19:36:22 and test is the project in my case May 18 19:47:16 ericben: first problem was that it didn't like .git in the folder name in the end. After I changed this, it first goes to angstrom and after 5th attemps it finaly gets the git repository, packs it in tgz in downloads and then dies with git read-tree 1 May 18 19:47:16 fatal: Not a valid object name 1 May 18 19:51:26 did you set a SRCREV ? May 18 19:51:51 ericben: not, I want to build from head May 18 19:52:21 then SRCREV = "${AUTOREV}" May 18 19:53:43 ericben: that was the trick, thanks. May 18 19:55:46 ericben: the only annoying thing left is that it tries to get first tar from angstrom. May 18 19:56:23 lgopher: maybe there is a setting to disable mirrors, I have the same thing here but didn't look how to disable it May 18 19:57:45 BB_GENERATE_MIRROR_TARBALLS = "0" I think May 18 19:59:02 though even with that setting, I keep getting tarballs of the checkout. What I don't get is tarballs of the entire git repo. May 18 19:59:54 wait, that answers one of your previous questions... May 18 20:01:08 03Denys Dmytriyenko  07org.openembedded.dev * r36b166cd2d 10openembedded.git/recipes/powervr-drivers/ (5 files): (log message trimmed) May 18 20:01:08 powervr-drivers: move GFX SDK bin from FILESPATH to DL_DIR May 18 20:01:08 * Don't pollute the recipes/files directories with downloads - avoid git May 18 20:01:08 reporting the file as untracked May 18 20:01:08 * The same GFX SDK binary is required by libgles-omap3 and omap3-sgx-modules May 18 20:01:09 recipes and instructions tell to put them in separate FILESPATH directories, May 18 20:01:10 leading to confusion or duplication (it's half a GB file!) May 18 20:01:24 I'm not sure what I did to prevent it from hitting Angstrom for everything first...CVS_TARBALL_STASH="" maybe May 18 20:03:16 Zygo: thanks for both tips, trying this rigth now. May 18 20:05:17 Zygo: ericben: yes it was CVS_TARBALL_STASH="" in the bb goes direct to git. Thanks ! May 18 20:06:04 lgopher: interesting, thanks May 18 20:22:08 ~seen zecke May 18 20:22:11 zecke is currently on #oe (4h 23m 51s), last said: 'ao2: it is supported.'. May 18 20:23:48 woglinde: hallo May 18 20:23:50 i have somehow atleaso fone "go 0x33f80000" but now in the uboot there is no option coming as "nand createbbt" May 18 20:27:15 re May 18 20:27:18 hi khem May 18 20:34:39 woglinde_: what are u working on these days May 18 20:34:59 diplomathesis May 18 20:35:01 and java May 18 20:36:13 java ? May 18 20:36:27 whats ur diplomthesis May 18 20:36:39 yes openjdk May 18 20:36:44 1.8 May 18 20:36:56 args icedtead 1.8 May 18 20:36:59 I meant May 18 20:37:01 I see May 18 20:37:13 so u will get Masters after that May 18 20:37:40 diplomathesis is about tracking of mobile wireless clients May 18 20:37:52 to collect information for routing hints May 18 20:38:05 e.g in which direction the client is moving May 18 20:46:56 03Koen Kooi  07org.openembedded.dev * r23fe8f3c1a 10openembedded.git/contrib/angstrom/sort.sh: angstrom feed sorter: sort more archs May 18 20:51:49 woglinde_: hmm sounds cool May 18 20:52:02 woglinde_: are you working on the software side May 18 20:57:47 khem yes May 18 20:57:56 my framework is python based May 18 20:58:01 with all the goodies May 18 20:58:11 pyro, scapy, daemon May 18 20:58:20 twisted May 18 20:58:38 woglinde_: ok May 18 20:58:58 so its going to be a softwarestack on both clients and servers May 18 20:59:27 only for the stationaries router May 18 20:59:38 because you are in control of them May 18 21:00:07 I made it all debian packages May 18 21:00:11 hm remindes me May 18 21:00:16 to put them into oe May 18 21:01:56 woglinde_: yeah OE is better :) May 18 21:02:25 na May 18 21:02:32 our testbed is running debian May 18 22:25:35 woglinde_: hey May 18 22:36:00 he zecke May 18 22:47:41 03Andrea Adami  07org.openembedded.dev * r06fff17991 10openembedded.git/recipes/kexecboot/linux-kexecboot.inc: May 18 22:47:41 linux-kexecboot: sed out conflicting getline in scripts/unifdef.c May 18 22:47:41 * mainline since 2.6.29-rc8 May 18 22:48:11 ant still inthere? May 18 22:48:30 the crappy getline May 18 22:48:40 the idea is the actual .dev will be next stable soon, with 2.6.2x May 18 22:48:55 then we'll move to semi-working 2.6.34 May 18 22:49:21 woglinde_: interestingly the problem emerged today... May 18 22:49:33 linux.inc changes? May 18 23:01:32 hey zecke May 18 23:01:32 ant_home: whats the problem May 18 23:02:30 ho May 18 23:02:53 khem: indirectly https://patchwork.kernel.org/patch/11166/ May 18 23:03:18 this kernel is including linux.inc May 18 23:05:09 I can't patch all the old kernels...I sed in the .inc May 18 23:05:17 won't harm May 18 23:07:03 well, I could add a test for kernel version <2.6.29 for the sake of precision May 18 23:07:30 but as I said, nothing happens for k. >=2.6.29 May 18 23:08:40 oh we have that patch in OE May 18 23:08:46 for quite a while now May 18 23:09:05 the question is: what triggered the issue in linux-kexecboot? May 18 23:09:18 must be very recent May 18 23:09:25 commit May 18 23:10:59 ant_home: ah it was for other packages using kbuild May 18 23:11:02 like uclibc May 18 23:11:40 I've seen the same issue trying klibc > 1.515 against 2.6.26. Failure on make_headers_install May 18 23:11:58 but they say upstream 2.6.2x is dead... May 18 23:12:20 and they are working for linux-next May 18 23:13:07 ant_home: did u upgrade your build box May 18 23:13:18 well, is Gentoo stable May 18 23:13:24 because this happens with glibc 2.10 onwards May 18 23:14:14 [ebuild R ] sys-libs/glibc-2.10.1-r1 May 18 23:14:24 ok May 18 23:14:32 let see when May 18 23:14:54 actually if you are using unifdef from OE then it has the fix May 18 23:15:04 uclibc has also it May 18 23:15:16 but if kernel is using its own May 18 23:15:25 then it needs similar fix May 18 23:15:43 recipes/uclibc/uclibc-0.9.30.1/gcc-4.4-fixlet.patch May 18 23:17:10 khem: on the same host I could not build linux-rp but not linux-kexecboot May 18 23:17:20 err.. opposite May 18 23:17:31 so I suspect linux-rp.inc is ok May 18 23:17:46 linux.inc could have got some change May 18 23:18:06 now rebuilding from scratch :/ May 18 23:18:44 anyway, for x86 glibc-2.10.1-r1 is stable since 10 Jan 2010 May 18 23:18:54 so no, it's not this May 18 23:19:43 Gentoo host May 18 23:21:02 khem: :D May 18 23:21:16 4 days ago...linux.inc: add protptype for 'kernel-headers' package May 18 23:21:28 yeah thats will be it May 18 23:21:45 make_headers_install May 18 23:21:50 it could have used installed unifdef May 18 23:22:52 feel free to polish it ;) May 18 23:24:34 ant_home: so u can build linux-rp but not linux-kexecboot May 18 23:24:52 well, I already committed a patch May 18 23:24:55 but yes May 18 23:25:28 linux-rp.inc is almost unmaintained May 18 23:26:22 khem: btw, how it is we have qemu machines in linux-kexecboot? May 18 23:26:30 do qemu survive to kexec now? May 18 23:26:37 *does May 18 23:26:53 ant_home: no idea never tried that May 18 23:27:14 someone tried and miserably failed :) May 18 23:29:08 heh May 18 23:33:48 khem: actually any kernel using linux.inc and being <=2.6.29 would be broken May 18 23:34:00 apart kexecboot May 18 23:34:09 and linux-rp May 18 23:34:11 yeah thats when getline is introduced into unifdef May 18 23:35:22 so perhaps the sed should happen inn linux.inc May 18 23:35:50 this change touched most machines :/ May 18 23:36:00 ant_home: Was going to ask about doing it there May 18 23:36:20 this is core stuff May 18 23:36:37 I don't feel prone to touch it ;) May 18 23:36:44 yeah thats when getline is introduced into unifdef]'']12222223456788888 May 18 23:37:04 iirc 2.6.29-rc8 May 18 23:37:30 khem: cat on the keyboard? May 18 23:37:58 ant_home: heh its my daughter May 18 23:38:07 I went to fetch some fruits for her May 18 23:38:11 I^_^ May 18 23:39:26 Tartarus: khem you know what the issue really is? May 18 23:39:47 is that base_version_less_or_equal is still unable to parse complex PV May 18 23:39:50 conflict with newer glibc May 18 23:39:53 just always change it May 18 23:40:12 Tartarus: hey May 18 23:40:19 Or grab the patch that exists in a few versions subdirs May 18 23:40:19 Tartarus: how is ppc thing May 18 23:40:22 did it boot May 18 23:40:35 khem: Busy with internal work, haven't had the cycles to get someone to try, sorry May 18 23:40:46 my qemu boots fine at O2 but does not boot on Os May 18 23:41:02 odd May 18 23:41:12 I tried uclibc/eglibc same results init kills the kernel May 18 23:42:30 Tartarus: I think it could be debugged easily but again time problem:) May 18 23:44:25 Tartarus: khem: sent msg to the ML May 18 23:48:57 'nite **** ENDING LOGGING AT Wed May 19 02:59:57 2010