**** BEGIN LOGGING AT Wed Jun 02 02:59:57 2010 Jun 02 05:37:40 03Martin Jansa  07org.openembedded.dev * r12390b6da7 10openembedded.git/recipes/gnome/ (cheese_2.30.0.bb cheese_2.30.1.bb): Jun 02 05:37:40 cheese: add gst-plugins-base to DEPENDS Jun 02 05:37:40 Signed-off-by: Martin Jansa Jun 02 06:24:25 william_z: why is it accessing /usr/bin/autoreconf ? do you have autotools in ASSUME_PROVIDED Jun 02 06:27:20 tks khem, is ASSUME_PROVIDED in which dir? I haven't seen this file, after I update oe, it display this error Jun 02 06:28:05 william_z: hmm whats your host distro Jun 02 06:28:38 angstrom-2008.1 Jun 02 06:29:15 but my MACHINE = "smdk6410", not omap3530 Jun 02 06:30:54 william_z: ASSUME_PROVIDED is a variable Jun 02 06:31:19 william_z: e.g. set in local.conf, but can be set in any file that is parsed during startup Jun 02 06:32:08 in my local.conf , I don't set it Jun 02 06:32:16 #ASSUME_PROVIDED += "virtual/${TARGET_PREFIX}gcc virtual/libc" Jun 02 06:33:48 should I reset ASSUME_PROVIDED += "virtual/${TARGET_PREFIX}gcc virtual/libc"? Jun 02 06:34:51 william_z: thats not the problem it doesnt matter Jun 02 06:35:52 william_z: did it work before for you ? Jun 02 06:36:07 and are you trying a fresh build after update Jun 02 06:36:30 william_z: could you be so kind to paste the link to your log again, I must have joined afterwards. Jun 02 06:36:34 yes, it worked well before, Jun 02 06:38:05 AUTOV is 1.10 Jun 02 06:38:05 NOTE: Executing autoreconf --verbose --install --force --exclude=autopoint -I /media/disk-1/stuff/tmp/work/armv6-angstrom-linux-gnueabi/linphone-3.1.0-r1/linphone-3.1.0/m4/ -I/media/disk-1/stuff/tmp/sysroots/armv6-angstrom-linux-gnueabi/usr/share/aclocal-1.10 -I /media/disk-1/stuff/tmp/sysroots/armv6-angstrom-linux-gnueabi/usr/share/aclocal Jun 02 06:38:05 /usr/bin/autoreconf: unrecognized option `--exclude=autopoint' Jun 02 06:38:05 Try `/usr/bin/autoreconf --help' for more information. Jun 02 06:38:06 FATAL: autoreconf execution failed. Jun 02 06:39:19 william_z: thanks, do you have any autoreconf inside your tmp directory? Jun 02 06:39:19 yes, i did fresh build Jun 02 06:39:52 william_z: could you run it with strace and see where it is stating for autoreconf? Jun 02 06:39:55 no, I haven't find it] Jun 02 06:43:28 zecke, how to run it with strace? Jun 02 06:44:29 william_z: how do you run the script right now? Jun 02 06:44:38 I bitbake -b linphone_3.1.0.bb or bitbake linphone Jun 02 06:45:40 before I can bitbake vlc well, but now it can't work also Jun 02 06:45:45 william_z: you should bitbake linphone Jun 02 06:46:01 bitbaking a recipe may not pull in all deps Jun 02 06:47:44 tks, I bitbake linphone, some .bb files can't build success Jun 02 06:48:10 such as coreutils-native_7.2.bb Jun 02 06:48:30 RROR: Error in executing: /media/disk-1/stuff/openembedded/recipes/coreutils/coreutils-native_7.2.bb Jun 02 06:48:30 ERROR: Exception: Message:[Errno 2] No such file or directory Jun 02 06:48:57 william_z: you can just put strace -fF -o log.txt CMD in front and see where it is looking.. Jun 02 06:49:16 william_z: the other thing is use bitbake -g and see if something is depending on autoconf. Jun 02 06:49:23 william_z: which version of bitbake are you using? Jun 02 06:49:45 bitbake-1.8.18 Jun 02 06:49:49 yeah may be this BBCLASSEXTEND thing is biting him Jun 02 06:50:02 ok, I try it now Jun 02 06:50:18 william_z: can you try using bitbake from git master Jun 02 06:50:54 I'm not, how to bitbake from git master? Jun 02 06:51:50 git clone git://git.openembedded.net/bitbake Jun 02 06:51:57 william_z: you can use another release. Jun 02 06:52:23 nevermind Jun 02 06:52:25 sorry Jun 02 06:53:06 not at all, you're very kind Jun 02 06:54:13 william_z: yeah, but saying there is another tarball was crap. :) Jun 02 06:56:26 now I get bitbake from "git clone git://git.openembedded.net/bitbake " and try it Jun 02 06:56:59 ok good Jun 02 06:59:59 same error Jun 02 07:01:02 william_z: delete the tmpdir and rebuild Jun 02 07:01:08 from scratch Jun 02 07:02:38 ok , I try it now Jun 02 07:10:00 03Vitus Jensen  07org.openembedded.dev * rf00180ced5 10openembedded.git/recipes/linux/linux-n1200_2.6.27-rc9+git.bb: Jun 02 07:10:00 linux-n1200_2.6.27-rc9+git: change KERNEL_OUTPUT to a relative path Jun 02 07:10:00 deploy method in kernel.bbclass automatically adds {S} Jun 02 07:10:00 Signed-off-by: Vitus Jensen Jun 02 07:10:00 Signed-off-by: Eric BĂ©nard Jun 02 07:10:55 hi Jun 02 07:11:08 hi ericben Jun 02 07:11:24 in patchwork: what is needed to be able to change the status of a patch once it's applied ? Jun 02 07:11:30 hi khem Jun 02 07:11:42 oh time to bed now though Jun 02 07:11:56 khem: good night ! Jun 02 07:12:10 * khem -> sleep () Jun 02 07:12:42 morning Jun 02 07:12:47 n8 khem Jun 02 07:38:14 03Koen Kooi  07org.openembedded.dev * r2eb2c7af99 10openembedded.git/recipes/cairo/ (3 files in 2 dirs): cairo 1.8.10: add patch to get rid of some double precision FP math and convert to new style staging Jun 02 07:38:24 03Koen Kooi  07org.openembedded.dev * rd2f0ab2191 10openembedded.git/recipes/gnome/cheese_2.30.1.bb: cheese: remove scrollkeeper-native, it doesn't exist in OE so we shouldn't be adding it to DEPENDS Jun 02 07:38:24 03Koen Kooi  07org.openembedded.dev * r716a10bbfc 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Jun 02 07:38:24 opkg: bump SRCREV to 530 Jun 02 07:38:24 The commits since 527 are: Jun 02 07:38:25 * Don't try to print cmd_name before its initialised. From Amanous, Issue #48 Jun 02 07:53:00 03Koen Kooi  07org.openembedded.dev * re676ecc9ec 10openembedded.git/recipes/gnome/scrollkeeper_0.3.14.bb: scrollkeeper: add 0.3.14. Only -native portion works, but that's what we are interested in Jun 02 07:53:38 hi Jun 02 07:56:56 03Martin Jansa  07org.openembedded.dev * rfb641200b6 10openembedded.git/recipes/xinput-calibrator/ (8 files in 3 dirs): Jun 02 07:56:56 xinput-calibrator: add INC_PR, move machine specific file to separate recipe Jun 02 07:56:56 * sorry for making whole xinput-calibrator machine specific before, this Jun 02 07:56:56 way it should be better for upgrades (as proved by normal pointercal Jun 02 07:56:56 for tslib) Jun 02 08:03:40 good morning Jun 02 08:05:07 i try to compile gnome on oe but i get this error - can anyone tell me how to fix it? http://nopaste.info/a8eee231ca.html Jun 02 08:06:03 william_z: I kicked off a build too but mine is failing way earlier Jun 02 08:08:08 now I can't build it well Jun 02 08:08:34 and I did fresh build also Jun 02 08:11:40 hi, i have an issue with gcc cross, for my arch one of the files fails to compile due to -Os, replacing manually with -O3 repairs it, but can't really find where this -Os comes from, so no idea how really fix it, anyone an idea on where this is from ? Jun 02 08:12:50 03Martin Jansa  07org.openembedded.dev * r67304d1a9f 10openembedded.git/recipes/klibc/ (7 files): Jun 02 08:12:50 Revert "klibc: fix packaging" Jun 02 08:12:50 This reverts commit e020cd9b2c61fd19bb55ec3fd38abcfdbd3f651c. Jun 02 08:12:50 Acked-by: Koen Kooi Jun 02 08:12:50 Signed-off-by: Martin Jansa Jun 02 08:13:01 03Martin Jansa  07org.openembedded.dev * r49a99cc3c0 10openembedded.git/recipes/klibc/ (klibc-common.inc klibc-utils.inc): Jun 02 08:13:01 klibc: fix packaging without moving PACKAGES and FILES_ back to recipes Jun 02 08:13:01 Signed-off-by: Martin Jansa Jun 02 08:20:18 zecke, it can build now, I rename the tmp file, did fresh build Jun 02 08:24:19 03Holger Hans Peter Freyther  07org.openembedded.dev * r368b4b672f 10openembedded.git/recipes/m4/ (files/f13-compile.patch m4_1.4.14.bb): Jun 02 08:24:19 m4: Fix m4-native-1.4.14 build on Fedora13... Jun 02 08:24:19 The build error was a link failure not finding S_ISDIR Jun 02 08:24:19 which is a macro provided for the stat syscall.. Jun 02 08:25:04 william_z: I had to the above... and now the second thing that is getting compiled is autoconf-native. which should have autoreconf Jun 02 08:25:42 ERROR: '/media/disk-1/stuff/openembedded/recipes/coreutils/coreutils-native_7.2.bb' failed Jun 02 08:26:02 ERROR: Error in executing: /media/disk-1/stuff/openembedded/recipes/coreutils/coreutils-native_7.2.bb Jun 02 08:26:03 ERROR: Exception: Message:[Errno 2] No such file or directory Jun 02 08:26:13 zecke: I've seen the same on gentoo (so it's not fc13 specific) see 798066d5b2183acc617b5b9d18066f02680aa9d6 Jun 02 08:28:11 william_z: which file does it try? do you have done something funky like mounting your disk noexec? Jun 02 08:28:36 JaMa: which glibc? Jun 02 08:30:15 the openembedded is my disk-1 dir, when I use oe, I should mount the disk Jun 02 08:30:48 zecke: on host? Jun 02 08:30:55 zecke: on target eglibc-2.12 Jun 02 08:31:23 on host glibc-2.11.1 Jun 02 08:32:09 answer to own question of 20 min's ago: these flags are in config/mt-ospace Jun 02 08:37:44 which recipe should provide ${STAGING_DATADIR_NATIVE}/gettext/config.rpath? grep-native fails during configure because it doesn't exist. Can this be a concurrency issue or a missing dependency issue? Jun 02 08:43:43 janp: gettext-native Jun 02 08:47:05 ericben: should i create a patch so that grep-native depends on gettext-native? Jun 02 08:47:33 janp: maybe inherit gettext would be better solution Jun 02 08:47:48 but I'm not sure.. Jun 02 08:49:46 JaMa: the only grep-native would need from the inherit gettext is the depends variable, all the rest is not really needed Jun 02 08:51:06 janp: I don't what is the right way between DEPENDS & inherit Jun 02 08:51:38 sorry read 'I don't know' Jun 02 08:52:01 william_z: is it mounted noexec? Jun 02 08:53:17 sorry, what's the noexec mean? Jun 02 08:54:58 william_z: it is a mount flag, man mount, it means executables from this mountpoint are not allowed to be executed. Jun 02 08:55:14 ericben: Is multiple inherit possible? grep-native already inherits from native Jun 02 08:56:13 mount -t ntfs /dev/sda5 /media/disk-1 -o iocharaset=en_US.UTF-8 Jun 02 08:56:51 I think it is allowed to be executed Jun 02 08:58:01 all error is owning to gnu-config_20050701.bb file Jun 02 08:58:23 janp: yes, check other native recipe : grep -rn inherit recipes/*/*native*.bb|grep gettext Jun 02 09:00:53 Anyone using gtk+extra in their builds or has it always been broken? Jun 02 09:02:23 janp: inherit seems as preferred way (it will disable nls when not wanted) see 9ea3fb75e6b04bfee8b8d17bbc0ea0d6a6ff41e9 , but not sure if it's intended to work when called from -native recipe (as it also depends on non-native gettext so it can create maybe circular dependency) Jun 02 09:04:26 janp: seems it should work ok from: 12a05b660c12d2be2efd44db7372159ab62bd4fd Jun 02 09:11:39 JaMa: added gettext to inherit list, running test build now, so far no circular dependencies reported. Will post patch when bb grep-native is successfull. Jun 02 09:12:19 janp: thanks Jun 02 09:28:03 khem: there is one segfault in gdk-update-icon-cache, backtrace shows __bswapsi2 from libgcc2.c, BT: http://paste.pocoo.org/show/220954/ Jun 02 09:32:58 khem: and those Alignment traps are still here on armv4t (armv5t was fine yesterday) Jun 02 09:35:05 khem: and that bswapsi2 seems quite common (ie "ps a" works "ps u" segfault) http://paste.pocoo.org/show/220961/ Jun 02 09:58:17 ant__: hi Jun 02 09:58:24 hello Jun 02 10:00:18 JaMa: I tested gcc_4.5 and eglibc_2.12 (binutils 2.20.1) on armv5te / c7x0 Jun 02 10:00:35 kernel does not boot and lot of segmentation faults. Jun 02 10:00:52 nice .. bitbake -k world Running task 14200 of 70446. lets wait some weeks to see whats failing ... Jun 02 10:01:29 minimal/console-image comes to login using kernel (2.6.26) compiled with older gcc Jun 02 10:01:41 kernel and modules fwiw Jun 02 10:02:24 william_z: type mount and see how it is actually mounted Jun 02 10:02:39 william_z: and do you have good experience with NTFS? Jun 02 10:03:02 no, I haven't Jun 02 10:04:30 I just put oe in disk-1 dir, and it worked well before a long time Jun 02 10:04:42 william_z: type mount, and see the mount params Jun 02 10:06:16 ok Jun 02 10:08:57 the disk-1's type is fuseblk Jun 02 10:12:37 william_z: please paste the whole line for the mountpoint. or alternatively compile a hello world app and try to execute it from your disk-1 mountpoint. Jun 02 10:13:26 ok Jun 02 10:13:55 mount -t ntfs /dev/sda5 /media/disk-1 -o iocharaset=en_US.UTF-8 Jun 02 10:16:14 JaMa: build was successfull and the patch has been sent to the mailing list. Thx for the help. Jun 02 10:16:59 i try to compile gnome on oe but i get this error - can anyone tell me how to fix this? http://nopaste.info/a8eee231ca.html Jun 02 10:19:14 zecke, I build a hello world app, and it can execute on my board(s3c6410) Jun 02 10:21:59 I bitbake -b helloworld_1.0.0.bb ,but when I bitbake bitbake helloworld, it display the same error, Jun 02 10:22:37 william_z: Okay, you have not followed my thought. Jun 02 10:22:39 ERROR: Error in executing: /media/disk-1/stuff/openembedded/recipes/gnu-config/gnu-config_20050701.bb Jun 02 10:22:39 ERROR: Exception: Message:[Errno 2] No such file or directory Jun 02 10:22:54 william_z: in your current setup everything is failing with weird error messages Jun 02 10:23:10 william_z: which makes an issue with your setup very likely... Jun 02 10:23:42 william_z: one of the possible issues is that you are trying to run executables of a mountpoint that is not allowed to run executables. Jun 02 10:24:02 sorry , maybe I misunderstand your meaning Jun 02 10:24:23 william_z: you had two options here, paste the output of mount, look at the line for the mountpoint in question, search for noexec Jun 02 10:24:28 william_z: I asked twice, you didn Jun 02 10:24:36 you didn't understand it twice. :) Jun 02 10:24:58 sorry about that Jun 02 10:25:12 william_z: this brings me to option number two, build a simple application, have it on your mountpoint (/media/disk-1) and try to execute it. Jun 02 10:25:41 /dev/sda6 on / type ext3 (rw,relatime,errors=remount-ro) Jun 02 10:25:41 tmpfs on /lib/init/rw type tmpfs (rw,nosuid,mode=0755) Jun 02 10:25:41 proc on /proc type proc (rw,noexec,nosuid,nodev) Jun 02 10:25:41 sysfs on /sys type sysfs (rw,noexec,nosuid,nodev) Jun 02 10:25:41 varrun on /var/run type tmpfs (rw,nosuid,mode=0755) Jun 02 10:25:42 varlock on /var/lock type tmpfs (rw,noexec,nosuid,nodev,mode=1777) Jun 02 10:25:44 udev on /dev type tmpfs (rw,mode=0755) Jun 02 10:25:48 tmpfs on /dev/shm type tmpfs (rw,nosuid,nodev) Jun 02 10:25:50 devpts on /dev/pts type devpts (rw,noexec,nosuid,gid=5,mode=620) Jun 02 10:25:52 fusectl on /sys/fs/fuse/connections type fusectl (rw) Jun 02 10:25:54 lrm on /lib/modules/2.6.28-16-generic/volatile type tmpfs (rw,mode=755) Jun 02 10:25:56 securityfs on /sys/kernel/security type securityfs (rw) Jun 02 10:25:58 binfmt_misc on /proc/sys/fs/binfmt_misc type binfmt_misc (rw,noexec,nosuid,nodev) Jun 02 10:26:00 gvfs-fuse-daemon on /home/william/.gvfs type fuse.gvfs-fuse-daemon (rw,nosuid,nodev,user=william) Jun 02 10:26:01 *sigh* Jun 02 10:26:02 nfsd on /proc/fs/nfsd type nfsd (rw) Jun 02 10:26:04 /dev/sda5 on /media/disk-1 type fuseblk (rw,nosuid,nodev,allow_other,blksize=4096) Jun 02 10:26:31 oh, I get you Jun 02 10:26:45 ;) Jun 02 10:26:56 okay, so no noexec there... Jun 02 10:27:27 william_z: well, then. Currently gnu-config and coreutils-native fail with the same error... sadly they don't tell us which file... Jun 02 10:27:33 ant__: which kernel did you use? you have one with khem's patch (2.6.34 has it) Jun 02 10:27:45 william_z: you can use strace -fF -o log.txt NORMAL_CMD once again.. Jun 02 10:27:58 william_z: and then search it for stat calls towards the end of the file. Jun 02 10:28:02 william_z: isn't you ridsk-1 in ntfs ? Jun 02 10:28:17 fstype is fuseblk : what is the fs ? Jun 02 10:28:20 ant__: I have it running on spitz, but also with some changes to CFLAGS see angtrom patches in http://gitorious.org/~jama/angstrom/jama-shr-experimental Jun 02 10:28:53 yes,it's ntfs Jun 02 10:29:10 william_z : that can be your problem I think Jun 02 10:30:10 ericben: it would be still interesting to know in which way it is failing. Jun 02 10:30:19 but I use it a long time before, about one year in disk-1 dir Jun 02 10:32:42 bitbake -fF -o log.txt NORMAL_CMD helloworld? Jun 02 10:32:45 zecke: yes I agree with you but the configuration is quite non common and I don't know if ntfs allow all mods for files like ext or other fs Jun 02 10:33:24 bitbake: error: no such option: -F Jun 02 10:35:02 william_z: "strace -fF -o log.txt NORMAL_CMD" -- I don't see bitbake in what zecke told you Jun 02 10:36:12 when I execute app in /media/disk-1 dir, it can be executed Jun 02 10:37:12 bitbake strace -fF -o log.txt NORMAL_CMD helloworld Jun 02 10:37:12 Usage: bitbake [options] [package ...] Jun 02 10:37:12 Executes the specified task (default is 'build') for a given set of BitBake files. Jun 02 10:37:12 It expects that BBFILES is defined, which is a space separated list of files to Jun 02 10:37:12 be executed. BBFILES does support wildcards. Jun 02 10:37:13 Default BBFILES are the .bb files in the current directory. Jun 02 10:37:17 bitbake: error: no such option: -F Jun 02 10:40:27 tks zecke, likewise and ericben, I'll try it in another dir later Jun 02 10:41:45 thanks you again,it's very kind of you! thanks Jun 02 10:48:20 how can i show the build process in bitbake? Jun 02 10:48:26 and not hide it Jun 02 10:51:00 puchu: add options -D and -v to bitbake, as can be found in the help of 'bitbake -h' Jun 02 11:11:21 it seems the mesa_7.8.bb recipe is missing a dependency to libxt. Jun 02 11:33:48 janp: thanks Jun 02 11:34:40 ant__: what DISTRO setting are you useing for c7xx ? Jun 02 11:35:22 i try to compile gnome on oe but i get this error - can anyone tell me how to fix this? http://nopaste.info/a8eee231ca.html Jun 02 11:35:46 it hangs at gnome-applets and gnome-games Jun 02 12:18:55 puchu: maybe you can disable the doc in this package (even it that is not a fix) Jun 02 12:26:47 hi all, im super new to OE and i have a couple of questions, OE is primarily used as a recipe/make system to produce an image or distribution, correct? Jun 02 12:38:10 josh2: pretty much, yes Jun 02 12:40:04 thanks, so what happens if you need to do dev work? can you make changes to some source, rebuild, then commit changes all from within the OE environment, or do you have to do a special git pull? Jun 02 12:43:14 03Koen Kooi  07org.openembedded.dev * r85e20773aa 10openembedded.git/recipes/tasks/task-sdk-native.inc: task-sdk-native: add cvs to RRECOMMENDS for autopoint Jun 02 12:44:25 josh2, I tend to use stGIT on top of the openembedded metadata, that way I can organise changes into a set of patches that can be popped on and off. Jun 02 12:47:00 so why dont i ask the real question, im at a company that builds a custom linux file system for an omap3 (beagleboard ish) type product, so far everything is custom make scripts and SVN, someone suggested OE as an alternative, OE seems great for deployment, but im concerned about the dev workflow model Jun 02 12:48:02 03Koen Kooi  07org.openembedded.dev * rbf5e17fdbc 10openembedded.git/recipes/tasks/task-sdk-native.inc: task-sdk-native: conform to OE packaging guidelines, spotted by Martin Jansa Jun 02 12:48:07 josh2: for dev it is a bit underdeveloped. There are a couple of solutions available. Jun 02 12:48:35 josh2: My favorite is, build a toolchain (with the deps needed) and give that to dev's, they can install the -dev into their toolchain Jun 02 12:49:22 josh2: the old style is to use bitbake -cdevshell my-package. This will open a screen/xterm/gnome-terminal with the environment setup in the sourcedir, you can type make and such there. Or execute the scripts OE generated for execution Jun 02 12:50:06 josh2: or quite new is a bbclass, this way you can instruct OE to not do the do_fetch to fetch the sources but use a local git repository to get your sources Jun 02 12:50:39 that sounds interesting, does it work with svn? Jun 02 12:51:17 also, does the recipe system place nice with multiple toolchains, ie ARM, TI DSP, and integrity? Jun 02 12:52:44 josh2: right now, we only target the target app processor, I am not aware of any recipe compiling software for a DSP. Jun 02 12:52:58 josh2: in most cases, we would just package the already existing dsp binary. Jun 02 12:54:24 does the recipe system allow you to specify a toolchain, so you could build the linux kernel,fs, and packages with one toolchain, but build a separate code package with a different toolchain? Jun 02 12:55:16 josh2: Oh, it does not use git at all.. it will just take the source from a local tree, it is quite new and I have not used this method, there was a mail on the OE devel mailinglist on how to use it two days ago. Jun 02 12:56:08 josh2: in theory yes, in practice I have not seen it. What happens for the normal package is to get automatically DEPEND = "virtual/arm-linux-gcc" Jun 02 12:56:27 josh2: this means it would build the toolchain, and the default CC,LD settings already point to the resulting toolchain. Jun 02 12:56:58 josh2: so in theory, one could remove that and put virtual/ti-dsp-cc there, and inherit from another class to change the TARGET_CC, TARGET_CPP and similiar variables. Jun 02 12:57:02 josh2: yeah, oe does already use allow one to use a different toolchain for the kernel, and it would be fairly easy to extend that same mechanism to other arbitrary packages. Jun 02 12:57:08 josh2 you might want to create an sdk for your devs, there is info on it in the manual Jun 02 12:57:28 josh2: the closest thing we have right now as an example is the kernel, native and cross bbclass in classes/ Jun 02 12:58:06 cool, do you have a link to the instructions that went out to the devel mailing list a couple days ago? Jun 02 12:58:37 03Koen Kooi  07org.openembedded.dev * r1f7eee75cc 10openembedded.git/recipes/psplash/psplash-zap_1.0.bb: psplash-zap: fix typo, spotted by Martin Jansa Jun 02 12:58:40 03Koen Kooi  07org.openembedded.dev * r13a1e07895 10openembedded.git/recipes/tasks/task-sdk-native.inc: task-sdk-native: and fix typo, also reported by Martin Jun 02 12:58:59 btw, i tried using an external toolchain with oe, managed to build helloworld, but things like bash still trigger building gcc-cross-initial (even though I used ASSUME_PROVIDED); guess gcc-cross-initial is forced in by linux-libc-headers or so Jun 02 12:59:21 josh2 : I am also new but convinced that OE does compile for DSP Jun 02 13:00:12 eFfeM_work: base.bbclass will add a dependency on virtual/${TARGET_PREFIX}gcc unless you INHIBIT_DEFAULT_DEPS. Jun 02 13:00:28 and virtual/libc, I think Jun 02 13:00:44 josh2: http://lists.linuxtogo.org/pipermail/openembedded-devel/2010-May/020287.html was the question Jun 02 13:01:23 great thanks, one last question, from others ive talked to, they say to start with the angstrom distro and modify from there, since im much more light weight and custom, should i start from scratch? Jun 02 13:01:25 josh2 : read line 143 in http://cgit.openembedded.org/cgit.cgi/openembedded/tree/recipes/ti/ti-codec-engine_2.21.bb?id=08dcecdc77ceb6a3673605216493905859f11982 Jun 02 13:01:45 pb_ ah i think I didn't add INHIBIT_DEFAULT_DEPS, thought that ASSUME_PROVIDED would take care of not rebuilding Jun 02 13:02:28 gaston: neat Jun 02 13:08:33 eFfeM_work: I think it should so long as you have ASSUME_PROVIDED all the right things, but there are quite a few packages that you need to list. I suspect you probably overlooked one or other. Jun 02 13:09:04 josh2 : Something for you guys with cash to spend to consider as well: http://www10.edacafe.com/nbc/articles/index.php?section=ICNews&articleid=816211 Jun 02 13:10:39 pb_ guess so, the odd thing was that it started to compile a package (gcc-cross-initial) that said was provided, don't understand why it would Jun 02 13:10:49 and bitbake -g did not reveal a dependency Jun 02 13:15:09 pb_ tried to follow this: http://docs.openembedded.org/usermanual/html/commonuse_prebuilt_toolchain.html with some more ASSUME_PROVIDED Jun 02 13:15:29 josh2: you can use a overlay if you want to test/modify stuff for your needs Jun 02 13:15:34 eFfeM_work: this was quite old... probably 2005, when I had to target a EZX device. Jun 02 13:15:35 anyway in that doc redefining things like CC is no good plan as it breaks configure Jun 02 13:15:50 eFfeM_work: the preferred way is the external-toolchain bbclass, which no one has stepped up to document. :) Jun 02 13:16:11 zecke you answer my questions before I actually can ask them :-) Jun 02 13:17:01 peeked at external-toolchain but didn't really get that part going either (and using csl as an example didn't help much either) Jun 02 13:17:19 thanks for the info guys, im gonna build angstrom to get a feel for how things work Jun 02 13:17:40 anyway, i'm not almost in a position that I can build the compiler myself (using oe) Jun 02 13:17:47 (gcc for nios2) Jun 02 13:18:18 03Vladimir Zapolskiy  07org.openembedded.dev * r08ed57bc4a 10openembedded.git/recipes/xorg-xserver/ (xserver-kdrive-1.5.3/no_xkb.patch xserver-kdrive_1.5.3.bb): Jun 02 13:18:18 kdrive: fix compilation problem of xserver-kdrive Jun 02 13:18:18 This patch adds application of no_xkb patch for the last Jun 02 13:18:18 available version of xserver-kdrive in OpenEmbedded. Jun 02 13:18:18 Signed-off-by: Vladimir Zapolskiy Jun 02 13:36:15 03Koen Kooi  07org.openembedded.dev * r2cfe270632 10openembedded.git/recipes/powervr-drivers/libgles-omap3.inc: libgles-omap3: add RRECOMMENDS for x11-wsegl to x11 packages Jun 02 13:42:03 morning all Jun 02 13:42:09 hi rp Jun 02 13:50:19 eFfeM_work: you can take a view at the poky manual for the external toolchain thing Jun 02 13:56:03 hey RP Jun 02 14:01:32 I'm working on a srctree example and getting the following bitbake error: FATAL: Unable to start to 'None' UI due to exception: 27224. Jun 02 14:02:48 zecke, i think i peeked into it, but will revisit, the issue is that I have a pre-built external toolchain (so not an sdk that is made with oe) Jun 02 14:02:59 UI = knotty Jun 02 14:06:31 03Dmitry Eremin-Solenikov  07org.openembedded.dev * r330d29167f 10openembedded.git/recipes/qemu/ (qemu_0.12.3.bb qemu_0.12.4.bb): Jun 02 14:06:31 qemu: fix the fix for qemu-native vs. no zlib on host Jun 02 14:06:31 Make the qemu virtual:native build to pass staging dir via EXTRA_OECONF, Jun 02 14:06:31 not via QEMU_CFLAGS, which aren't passed to configure environment. Jun 02 14:06:31 Signed-off-by: Dmitry Eremin-Solenikov Jun 02 14:10:53 cbrake: ouch :/ Jun 02 14:11:06 cbrake: An incredibly useful bitbake error message :( Jun 02 14:12:01 RP: yeah, this one is proving elusive Jun 02 14:12:58 RP: a little more info: http://pastebin.ca/1875980 Jun 02 14:13:26 I uncommented the 'print event' in the knotty.py Jun 02 14:14:41 cbrake: I really don't understand those messages :( Jun 02 14:16:04 RP: it appears I'm getting an exception in knotty.py, so just trying to trace what is going on -- perhaps I need to add code to get a backtrace Jun 02 14:16:35 cbrake: It looks like we have a general expection hook which is swallowing the real error :/ Jun 02 14:16:43 RP: nod Jun 02 14:19:31 there is a hook that hides the traceback, but only when debugging isn't enabled. if we audit the exception handling in bitbake, that'll likely prove unnecessary, but.. Jun 02 14:19:35 * kergoth gets caffeine Jun 02 14:20:03 now, there are plenty of places where tracebacks are lost due to the way the exceptions are handled, hopefully that's not the case here Jun 02 14:20:13 Error in sys.excepthook Jun 02 14:20:56 zecke, reread the poky handbook (at least the toolchain part), but this seems mostly on how to use a toolchain made by poky externally, I want to use an existing C compiler etc within oe, so there does not seem too much of a match Jun 02 14:21:04 * kergoth 's never seen anything like that, there isn't much that happens in the hook Jun 02 14:21:07 unless ofc http://www.pokylinux.org/doc/poky-handbook.html is not the thing I should be reading Jun 02 14:21:20 kergoth: where is the hook? Jun 02 14:21:25 bitbake. Jun 02 14:21:29 bitbake/bin/bitbake Jun 02 14:22:42 eFfeM_work: There is external toolchain support in poky Jun 02 14:22:51 eFfeM_work: and it is known to work, I used it recently Jun 02 14:22:56 * cbrake tries to comment out exception hook Jun 02 14:23:22 same thing Jun 02 14:27:58 kergoth: yeah, its a threading issue -- if I add a sleep, then I see the exception Jun 02 14:28:00 NOTE: package autotools-demo-1.0-12-ga090fff-r0: task do_install: Succeeded Jun 02 14:28:00 Jun 02 14:28:00 ERROR: TaskFailed event exception, aborting Jun 02 14:28:00 ERROR: Build of /scratch/oe/oe-build/oe/openembedded.custom/recipes/autotools-demo/autotools-demo.bb do_package failed Jun 02 14:28:42 kergoth: now, how to improve bb so these exceptions get reported ... Jun 02 14:30:41 RP is there any doc on it ? Jun 02 14:32:54 RP, need to run, but feel free to leave a msg Jun 02 14:33:01 kergoth: seems like we need to do a join on the server thread before exitting Jun 02 14:33:04 everyone: have a nice day/evening Jun 02 14:44:36 eFfeM_work: no docs. POKYMODE="external-xyz" is what enables it in Poky iirc Jun 02 15:12:02 03Koen Kooi  07org.openembedded.dev * r7c9ff232dd 10openembedded.git/recipes/cairo/ (2 files in 2 dirs): cairo 1.8.10: use proper patch to get rid of DP FP math Jun 02 15:30:51 hi stefan Jun 02 15:31:01 hi woglinde Jun 02 15:31:03 hi all Jun 02 16:00:59 re kergoth Jun 02 16:01:05 hey Jun 02 16:09:13 kergoth: I just sent a patch to bitbake-dev that improves some of the exception handling Jun 02 16:10:19 cool Jun 02 16:10:26 thanks for working on it Jun 02 16:10:39 * kergoth 's pretty busy with work stuff at the moment Jun 02 16:10:48 well, sucking up mindshare, anyway Jun 02 16:10:52 * kergoth glares at OE-Signatures Jun 02 16:12:12 hmm Jun 02 16:12:43 RP: TaskStarted is fired in the task's new process, or is it from the client or server? Jun 02 16:13:14 hi gnutoo Jun 02 16:15:19 uh uh Jun 02 16:15:21 http://dragonegg.llvm.org/ Jun 02 16:44:35 Hmmm, is there a trick I am missing to build a toolchain and SDK with all the given packages in an image or is it (as I do now) just a case of manually keeping your toolchain tasks and image tasks in sync and keeping the big stuff from the image out of the toolchain/sdk tasks? Jun 02 16:47:04 DJWillis, it is manual unless you design your custom image and custom SDK to be in sync Jun 02 16:49:43 Tartarus: that is just what I expected (and do), just wondered if I was missing a trick ;) Jun 02 16:50:26 DJWillis, only trick is you can do 'require recipes/foo/mystuff.inc' if you have your image.bb in images/ and sdk.bb in meta/ Jun 02 17:38:41 03Khem Raj  07org.openembedded.dev * r1e641a55aa 10openembedded.git/recipes/qemu/ (qemu_0.12.3.bb qemu_0.12.4.bb): Jun 02 17:38:41 qemu: When using _append we dont need += Jun 02 17:38:41 Signed-off-by: Khem Raj Jun 02 17:38:44 03Khem Raj  07org.openembedded.dev * r0e3aefdda5 10openembedded.git/recipes/m4/ (files/m4-uclibc-sched_param-def.patch m4_1.4.14.bb): Jun 02 17:38:44 m4-1.4.14: Fix compilation on uclibc. Jun 02 17:38:44 * uClibc does not expose sched_param the same way Jun 02 17:38:44 as glibc even though it pretends like glibc all Jun 02 17:41:31 03Khem Raj  07org.openembedded.dev * ra9f8abee2d 10openembedded.git/recipes/pump/ (7 files in 2 dirs): Jun 02 17:41:31 pump: Pull patches from gentoo and debian and make it compilable with uclibc Jun 02 17:41:31 Signed-off-by: Khem Raj Jun 02 17:43:27 03Koen Kooi  07org.openembedded.dev * reb1cfd85ca 10openembedded.git/recipes/cairo/ (2 files in 2 dirs): cairo 1.8.10: even faster version of patch Jun 02 17:45:08 khem: hi Jun 02 17:46:06 hey JaMa whats up Jun 02 17:46:32 03Khem Raj  07org.openembedded.dev * r39a7a0152d 10openembedded.git/conf/machine/include/qemu.inc: Jun 02 17:46:32 qemu.inc: zap_root_password for qemu images. Jun 02 17:46:32 Signed-off-by: Khem Raj Jun 02 17:46:47 khem: seen my morning gcc issue? Jun 02 17:48:33 JaMa: __bswapsi2 one just scrolled and read it Jun 02 17:50:23 yeah.. if you have some hint what to try, I'll relaunch my build Jun 02 17:50:57 JaMa: this is first time I am seeing this issue. Jun 02 17:52:02 me too, and only on armv4t, the armv5t build from the same OE checkout went fine Jun 02 17:52:28 khem: it's from chroot, but I guess that after reboot it can only get worse Jun 02 17:53:41 hi does any one know how to fix this: http://pastebin.org/301449 Jun 02 17:54:13 it happens while bitbaking nano for atngw100 (angstrom, uclibc) Jun 02 17:54:14 soltys: inherit gettext Jun 02 17:54:16 maybe Jun 02 17:54:23 JaMa: __bswapsi2 is C function hmm Jun 02 17:55:04 JaMa: there is tjan inherit ant some hack in libiconv recipe.. Jun 02 17:55:04 soltys: I have huge pile of fixes for uclibc Jun 02 17:55:10 soltys: you have to wait a while Jun 02 17:55:18 I will push them onto a userbranch Jun 02 17:55:21 today Jun 02 17:55:28 then you can pull them Jun 02 17:55:36 it has fixed the issue you are seeing Jun 02 17:55:53 khem: ok I'll wait ;) Jun 02 17:55:58 JaMa: hmm Jun 02 17:56:15 khem: yeah and that file wasn't changed for long except branch create Jun 02 17:56:18 JaMa: can you disassemble libgcc and see this function Jun 02 17:57:19 something better than objdump? Jun 02 17:57:30 objdump -d Jun 02 17:58:12 mmt building binutils Jun 02 17:59:24 oh you could use cross-bintils too Jun 02 18:02:26 ah it will be faster then on freerunner :) Jun 02 18:04:29 khem: http://paste.pocoo.org/show/221117/ Jun 02 18:04:56 whole file http://paste.pocoo.org/show/221118/ Jun 02 18:09:30 JaMa: this is libgcc right ? Jun 02 18:10:22 JaMa: can you debug it a bit and set a breakpoint on __bswapsi2 Jun 02 18:10:32 and when it hits the bp then step couple of intrs Jun 02 18:10:38 instrs with si Jun 02 18:16:43 khem: yes lib/libgcc_s.so.1 from libgcc1_4.5-r0+svnr160043.5_armv4t.ipk Jun 02 18:18:38 khem: mmt (I can give you ssh access to my neo if it will help you resolve it faster) Jun 02 18:18:50 JaMa: eh sure Jun 02 18:19:19 03Stefan Schmidt  07org.openembedded.dev * rd502567189 10openembedded.git/recipes/u-boot/ (files/bug20/bug-uboot.patch u-boot_git.bb): u-boot_git: Add bug20 support Jun 02 18:19:39 JaMa: ok one more thing look into the build log of gcc-cross log.do_compile and search for _bswapsi2.o Jun 02 18:19:53 then see did it use -O2 or -Os to compile this file Jun 02 18:25:03 Hi, can anyone tell me what I'm doing wrong when bitbake gives me this error: ERROR: Openembedded's config sanity checker detected a potential misconfiguration. Jun 02 18:25:46 mindthomas: you have to continue reading Jun 02 18:26:38 I've followed this guide Jun 02 18:26:38 http://www.electronics.diycinema.co.uk/embedded/mini2440/configuring-open-embedded-to-bake-angstrom-images-for-the-mini2440/ Jun 02 18:26:59 I meant you have to continue reading the message bitbake is giving you Jun 02 18:27:09 is that only thing it says Jun 02 18:27:27 khem: hmm -g -Os -O2 -g -Os Jun 02 18:27:51 http://paste.pocoo.org/show/221131/ Jun 02 18:28:35 ERROR: Openembedded's config sanity checker detected a potential misconfiguration. Either fix the cause of this error or at your own risk disable the checker (see sanity.conf). Following is the list of potential problems / advisories: Jun 02 18:28:46 Please install following missing utilities: chrpath Jun 02 18:29:05 mindthomas: install chrpath on your host Jun 02 18:29:21 i'm new to linux, so what is that? Jun 02 18:29:32 and how to install? Jun 02 18:29:39 which distro are you running? Jun 02 18:29:44 Fedora 11 Jun 02 18:30:23 yum install chrpath Jun 02 18:30:33 JaMa: ok Jun 02 18:30:54 JaMa: Do you have buildtree for gcc 4.4.4 and 4.5 both lying around ? Jun 02 18:32:46 JaMa: I think I know the problem Jun 02 18:33:40 khem: no, using rm_work.. so only temp from them :/ Jun 02 18:33:50 JaMa: ok no problem. Jun 02 18:33:58 JaMa: can you try to rebuild gcc-cross Jun 02 18:34:03 with one change Jun 02 18:35:28 sure Jun 02 18:35:41 thanks khem, that worked... but what is yum? Jun 02 18:36:13 mindthomas: its a package manager for redhat based distro Jun 02 18:37:09 JaMa in recipes/gcc/gcc-configure-common.inc comment out OPTSPACE = "${@base_contains('TARGET_ARCH', 'powerpc', '', '--enable-target-optspace',d)}" Jun 02 18:37:32 and add OPTSPACE = "" in next line Jun 02 18:37:39 and then rebuild gcc-cross Jun 02 18:37:48 and also your image Jun 02 18:38:20 I think the problem is that __bswapsi2 is calling itself Jun 02 18:38:59 whole image? Jun 02 18:39:17 I mean dont nuke the tmp Jun 02 18:39:24 but just gcc-cross Jun 02 18:39:30 and rebuild it Jun 02 18:39:42 then you need new libgcc in your image Jun 02 18:40:24 -c fetch -f -b gcc-cros.. and now it's building Jun 02 18:40:28 so either you can just do bitbake and it will do it or you have to do manually main be bitbake -c clean ; bitbake your-image after rebuilding gcc-cross Jun 02 18:40:50 why not only opkg install -force-reinstall libgcc1? Jun 02 18:40:57 even better Jun 02 18:41:12 I just need updated libgcc on target Jun 02 18:41:13 if I can get it running again :) Jun 02 18:41:31 ok, that's why I was confused by rebuilding image.. Jun 02 18:41:41 oh well. Jun 02 18:41:51 sometimes I talk too much Jun 02 18:42:24 ahh I know why it said wrong gzip magic :) Jun 02 18:42:45 because on that port is now running ssh-for-you instead of http :) Jun 02 18:42:55 heh ok Jun 02 18:43:00 I will relinquish it Jun 02 18:50:43 hi khem Jun 02 18:58:15 hello woglinde Jun 02 19:00:50 hi Jun 02 19:00:58 I've some checksums issues with libxfce4util_4.6.1.bb Jun 02 19:01:00 so: Jun 02 19:01:09 *the checksums are in the recipe and matches Jun 02 19:01:14 *I've bitbake 1.10 Jun 02 19:01:20 *there is no SRC_URI but: Jun 02 19:01:33 * cbrake wonders what bitbake is running configure in a recipe that inherits qt4e before qt-embedded and qt-tools is built? Jun 02 19:01:36 there is inherit xfce46 Jun 02 19:01:47 I'll look into that Jun 02 19:02:34 cbrake: simplify the statement :) Jun 02 19:03:54 hey khem Jun 02 19:04:00 khem: so I'm building this recipe: http://cgit.bec-systems.com/cgit.cgi/qt-overo-palo-gpio/tree/qt-overo-palo-gpio.bb Jun 02 19:04:24 khem: and it starts the configure phase of this recipe before qmake is built Jun 02 19:04:47 cbrake: ah hmm Jun 02 19:04:56 srctree merges configure, compile, install into a single task, and pulls over the dependencies, at least its supposed to. maybe its failing to pull over the do_populate_staging deptask Jun 02 19:04:59 that must be a depends problem in the class Jun 02 19:05:15 as i said in the thread, it does so to avoid interleaving tasks amongst recipes sharing the same S, i.e. native/cross/target Jun 02 19:05:28 and as i also said, it could be made optional, and automatically disabled when B is used Jun 02 19:05:44 * kergoth_ hasn't had time to deal with it yet Jun 02 19:06:44 hey cminyard Jun 02 19:06:54 kergoth_: ahh, that sounds likely Jun 02 19:07:47 if you aren't using a native or cross version of the recipe, then you can confidently say you don't need it to do it Jun 02 19:09:07 http://dl.dropbox.com/u/112715/Files/inscm-1.tar.bz2 was what i was experimenting with when i created the class Jun 02 19:09:24 had shasum-native and m4-native going, so needed that code Jun 02 19:10:11 take a look at the class, the merge_tasks function Jun 02 19:10:31 this was also written prior to the new style staging stuff, so it may need tweaking Jun 02 19:10:58 it explicitly runs do_stage, which isn't likely what we want it doing Jun 02 19:20:01 kergoth_: looked briefly at your inscm tar ball -- I'm out of time today, so I'll summarize the issues I know about on the mail list Jun 02 19:20:43 cool, that would help, there are a lot of mails in the threads :) Jun 02 19:21:00 * kergoth_ hasn't touched the class since he created it :) Jun 02 19:21:10 kergoth_: heh, well its almost working :-) Jun 02 19:23:44 I really like the idea of storing the sources + recipe in a git repository. what i'd really like, though, is to be able to "import" recipes + patches from the OE repository into a local git repository in a project area, and have it obey the dependency digraph when importing, so you could pull over a segment of the repository in this other form Jun 02 19:24:01 or something along those lines, anyway Jun 02 19:24:42 03Khem Raj  07org.openembedded.dev * r396c84a300 10openembedded.git/recipes/uclibc/ (uclibc-git/uClibc.distro uclibc_git.bb): Jun 02 19:24:42 uclibc-git: Move to latest master and enable UCLIBC_HAS_BSD_RES_CLOSE Jun 02 19:24:42 Signed-off-by: Khem Raj Jun 02 19:28:29 i did a prototype of a git repository creator for a recipe, once Jun 02 19:28:51 kergoth_: ok, "srctree status" email sent Jun 02 19:28:52 it actually walked the history of the recipe and the patches the recipe referenced, in the oe repository, to try to reconstruct the history in the other repo, using topic branches for the patches Jun 02 19:28:56 but then my head hurt, and i gave up :) Jun 02 19:34:41 kergoth_: yeah, I can imagine that would make the head hurt! Jun 02 19:35:29 kergoth_: in some ways I really like the seperation of meta data and TMPDIR, so still trying to figure out a good model for doing both Jun 02 19:38:14 Hi. When running bitbake should the linux system kind of freeze, or be very slow? Like right now the screen is black caused by the screen saver, but if I move the mouse after a long time it gets to the type password screen Jun 02 19:39:29 mindthomas? Jun 02 19:39:45 screensaver has nothing to do with bitbake Jun 02 19:39:51 03Khem Raj  07org.openembedded.dev * r823aa27005 10openembedded.git/conf/machine/include/qemu.inc: Jun 02 19:39:51 Revert "qemu.inc: zap_root_password for qemu images." Jun 02 19:39:51 Its better not to put this was here but in the images Jun 02 19:39:51 if needed. Jun 02 19:39:51 This reverts commit 39a7a0152d5d9a38ab202e11eb2eaab163dc35fa. Jun 02 19:39:54 what build machine you have? Jun 02 19:40:28 woglinde: no, but Linux goes into screensaver after a while you know... Jun 02 19:40:42 i'm running Fedora 11 on an 64-bit quadcore 2.3GHz Jun 02 19:41:35 But the text in the terminal haven't also changed in a while now? Shouldn't it post alot of text? Jun 02 19:41:50 mindthomas hm Jun 02 19:41:57 The last line is: NOTE: Parsing finished. 0 cached, 8057 parsed, 331 skipped, 0 masked. Jun 02 19:42:06 mindthomas which tree are you using and which bitbake? Jun 02 19:42:33 openembedded? what do you mean? Jun 02 19:42:35 and which image do you building Jun 02 19:42:39 yes oe Jun 02 19:42:52 i'm new to linux ;) Jun 02 19:43:17 I'm trying to build an Angstrom image Jun 02 19:43:24 I followed this guide: http://www.electronics.diycinema.co.uk/embedded/mini2440/configuring-open-embedded-to-bake-angstrom-images-for-the-mini2440/ Jun 02 19:43:38 urgs the mini again Jun 02 19:43:41 *g* Jun 02 19:43:46 hmm Jun 02 19:43:57 http://www.electronics.diycinema.co.uk/embedded/mini2440/configuring-open-embedded-to-bake-angstrom-images-for-the-mini2440/ Jun 02 19:44:01 ups Jun 02 19:44:39 but shouldn't bitbake tell me more Jun 02 19:44:50 or is that last line ok, and should I then just wait? Jun 02 19:45:15 are you sing bitbake 1.8.16? Jun 02 19:45:21 no Jun 02 19:45:23 .18? Jun 02 19:45:26 does that matter? Jun 02 19:45:36 yes Jun 02 19:45:40 oh, why Jun 02 19:45:42 .16 is not good for stable Jun 02 19:45:46 ups dev Jun 02 19:45:49 hms Jun 02 19:46:01 what image are you building? Jun 02 19:46:03 console? Jun 02 19:46:36 everything Jun 02 19:46:37 bye the way Jun 02 19:46:43 everything? Jun 02 19:46:46 I ran the last command which is on the tutorial Jun 02 19:46:49 try http://www.angstrom-distribution.org/narcissus/ Jun 02 19:47:10 I have Jun 02 19:47:20 I can't make a good/small enough image Jun 02 19:47:33 don't know if it even works with that Jun 02 19:47:46 kill bitbake Jun 02 19:47:53 and try bitbake gcc-cross Jun 02 19:48:02 Isn't it possible to make an image to include X11 GNOME under 128mb Jun 02 19:48:09 - i just restarted linux Jun 02 19:48:43 anbd try to comment OESTATS_BUILDER = "myname" Jun 02 19:49:00 maybee bitbake tried to get the stats builder and failed Jun 02 19:49:16 so you would recommend me trying now Jun 02 19:49:17 local.conf Jun 02 19:49:19 after I restarted Jun 02 19:49:23 sure Jun 02 19:49:28 why you restarted linux? Jun 02 19:49:32 its not windows Jun 02 19:49:46 you only need to restart it when kernel has security bugs Jun 02 19:49:57 or usbstack goes beserk Jun 02 19:50:11 hehe, ha :) I'm a current windows guy, running Fedora 11 in VMWare ;) Jun 02 19:50:21 in the local.conf file Jun 02 19:50:30 there is a OESTATS_BUILDER line Jun 02 19:50:35 just like the one you wrote Jun 02 19:50:40 yes put # infront of it Jun 02 19:50:57 did you add myname without think? Jun 02 19:50:57 hmm, why? but ok Jun 02 19:51:07 no Jun 02 19:51:20 the line was excactly like you wrote Jun 02 19:51:50 maybee we the manual should changed it to Jun 02 19:52:06 OESTATS_BUILDER = "don't be stupid change it to your name" Jun 02 19:52:23 haha Jun 02 19:52:26 the username? Jun 02 19:52:30 or my real name Jun 02 19:52:39 username is okay Jun 02 19:52:41 but I as siad Jun 02 19:52:48 try it with out stats Jun 02 19:52:51 ok Jun 02 19:52:56 so just run bitbake again? Jun 02 19:53:00 did you check that you have network access Jun 02 19:53:05 inside vmware Jun 02 19:53:08 yes yes Jun 02 19:53:11 firefox works fine :) Jun 02 19:53:12 okay Jun 02 19:53:18 it also downloads the files in bitbake fine Jun 02 19:53:20 try bitbake gcc-cross Jun 02 19:53:32 so not: bitbake base-image ; bitbake console-image ; bitbake x11-image - as i did before? Jun 02 19:53:40 first something easy Jun 02 19:53:45 the rest can follow Jun 02 19:53:51 ok Jun 02 19:54:26 hm by the way why you want gnome? Jun 02 19:55:08 on the mini2440 Jun 02 19:55:34 you could text e17 too Jun 02 19:55:37 ups test Jun 02 19:55:45 which should be smaller Jun 02 19:55:59 hmm yeah Jun 02 19:56:08 but that isn't in the online builder Jun 02 19:56:16 anyways, after restarting i can't execute bitbake Jun 02 19:56:21 it says command not found? Jun 02 19:56:37 and yes, i'm executing it in the right directory Jun 02 19:57:42 you need to do the "source source-me.txt" again Jun 02 19:58:00 which sets the search path for oe and bitbake right Jun 02 19:58:14 ah ok :) Jun 02 19:58:22 have to learn what the different commands do Jun 02 19:58:49 right now i've found out SU, SUDO, VI, GEDIT, YUM and GIT Jun 02 19:58:51 <_chase_> mindThomas: Have you tried the angstrom setup scripts at http://gitorious.org/angstrom/angstrom-setup-scripts. Helps with setting up your environment. Jun 02 19:59:20 _chase_ the tutorial has its own script Jun 02 19:59:37 <_chase_> OK. I'm not familiar with what tuturial he is using. Jun 02 19:59:53 mindThomas there are lot more Jun 02 20:00:00 no stupid gui stuff Jun 02 20:00:39 what I just want to be able to do is install TCL and install and run Eggdrop and Apache on the mini2440 Jun 02 20:01:03 eggdrop needs tcl? Jun 02 20:01:36 yes Jun 02 20:02:05 i'm trying bitbake gcc-cross now Jun 02 20:02:29 NOTE: Handling BitBake files - looks like it's downloading? Jun 02 20:02:41 mindThomas: on a virtual expect bitbake console-image to take 6 hrs or something Jun 02 20:02:50 WHAT :O Jun 02 20:03:02 what if i installed it normally Jun 02 20:03:13 on my 64-bit quadcore 2.3GHz, with 4gb ram? Jun 02 20:03:18 right now the virtual only got 2gb? Jun 02 20:03:27 hm 1 hour Jun 02 20:03:44 handling bitbake files is parsing, not downloading Jun 02 20:04:10 oh, thanks for that information kergoth_ - good to know... So what exactly does BitBake do? Jun 02 20:04:41 how does one go about saying rtfm in a nice way? Jun 02 20:05:03 ^-- like that :) Jun 02 20:05:08 ;) Jun 02 20:05:14 yeah ok ;) Jun 02 20:05:35 sorry, i'm new to all this, so this it's ok response kergoth_ :) Jun 02 20:05:54 woglinde: now the bitbake is the same place as before Jun 02 20:06:15 just wrote the "NOTE: Parsing finished. 0 cached, 8057 parsed, 331 skipped, 0 masked Jun 02 20:06:44 hm okay Jun 02 20:06:49 type crtl +c Jun 02 20:06:54 to kill it? Jun 02 20:06:56 run bitbake -DDD gcc-cross Jun 02 20:06:58 yes Jun 02 20:07:26 if vmware is really this slow Jun 02 20:07:29 consider dual boot Jun 02 20:07:52 but then I have to reformat, right? Jun 02 20:08:25 or use another disk Jun 02 20:09:56 now I'm running the debug of bitbake Jun 02 20:10:24 and then it writes - a lot of lines ;) Jun 02 20:11:07 yes Jun 02 20:11:20 so you see something is done Jun 02 20:11:40 ? Jun 02 20:13:58 The it says NOTE: Parsing finished - the same, and then it stops Jun 02 20:14:06 but it's doesn't looks like the program has ended! Jun 02 20:14:15 it's just like before! Jun 02 20:16:00 hm Jun 02 20:16:04 sorry Jun 02 20:17:27 kergoth_: read the fine manual Jun 02 20:17:30 :) Jun 02 20:20:00 so you don't know why it doesn't work :( Jun 02 20:20:04 ok then Jun 02 20:31:38 khem: that'd work :) Jun 02 21:00:18 hi pb Jun 02 21:06:23 hi Jun 02 21:06:25 hi woglinde Jun 02 21:14:48 hi all Jun 02 21:14:53 hey likewise Jun 02 21:14:55 hi likewise Jun 02 21:17:36 hi woglinde, khem Jun 02 21:31:29 khem: I tested gcc_4.5 and eglibc_2.12 (binutils 2.20.1) on armv5te / c7x0 - kernel 2.6.26 Jun 02 21:31:38 khem: kernel does not boot and lot of segmentation faults. Jun 02 21:32:03 khem: minimal/console-image comes to login using kernel + modules (2.6.26) compiled with older gcc Jun 02 21:32:09 ant__: for kernel you need to backport the fix I committed to 2.6.34 Jun 02 21:32:15 yep Jun 02 21:32:22 I did all vanilla Jun 02 21:32:34 yeah vanilla wont work Jun 02 21:32:55 but basically each bin seems to have alignments faults Jun 02 21:33:38 and yes, I have __bswapsi2 in do_compile of gcc-cross Jun 02 21:33:42 hmm yeah thats coming from ld Jun 02 21:33:46 ld.so Jun 02 21:34:36 can I just use the new kernel and debug the userland? Jun 02 21:34:42 s/new/old/ Jun 02 21:35:27 yes Jun 02 21:35:29 I have serial cable here, if you need I can track a full boot until console Jun 02 21:35:39 with stderr+stdout Jun 02 21:35:47 that would be nice Jun 02 21:35:54 ok, just a few mins Jun 02 21:40:52 khem: I'm logging the 3rd reboot..do you need the first for any reason? Jun 02 21:41:23 ant__: no third is fine Jun 02 21:43:16 khem: http://fr.pastebin.ca/1876264 Jun 02 21:49:37 khem: seems all coreutils are affected Jun 02 21:49:47 good nite Jun 02 21:51:52 ant__: yes if you look at the address it happens at same address Jun 02 21:52:03 good night woglinge Jun 02 21:53:23 khem: here I need your or pb__ 's magic to decode the string (or an hour readings) Jun 02 21:55:57 ant__: ldr r2, [r5, r8] Jun 02 21:56:00 thats what it is Jun 02 21:56:24 I just see a repeated pattern Jun 02 21:56:57 other instruction is str r1, [r5, r8] Jun 02 21:57:18 I guess r5+r8 ends up in non 4byte aligned address Jun 02 22:00:15 hm.. I see ... -O2 -g -Os -DIN_GCC ...in do_compile Jun 02 22:00:28 of gcc-4.5 Jun 02 22:01:00 yeah thats problem too Jun 02 22:03:46 well, echo CRTSTUFF_CFLAGS = '-O2 -g -Os -DIN_GCC Jun 02 22:04:00 this is the first occurrence Jun 02 22:05:35 khem: is crtstuff Jun 02 22:06:16 gcc/crtstuff.c Jun 02 22:07:29 ant__: yes the fix/workaround is to disable --enable-target-optspace Jun 02 22:07:44 and rebuild libgcc Jun 02 22:08:39 ant__: can you upload your libc.so.6 and librt.so somehwere Jun 02 22:08:45 sure Jun 02 22:13:49 khem: he. megaupload is killing epiphany... Jun 02 22:14:45 libc.so.6 http://rapidshare.com/files/394584432/libc-2.12.so.html Jun 02 22:15:29 librt.so http://rapidshare.com/files/394584540/librt-2.12.so.html Jun 02 22:18:34 ant__: I can not access rapidshare Jun 02 22:18:42 np, mom Jun 02 22:19:12 or megaupload Jun 02 22:19:58 mom Jun 02 22:21:10 can you ssh ? Jun 02 22:21:13 yes Jun 02 22:43:09 What recipe should I be using for a basic console-only installation? Jun 02 22:43:51 I tried console-image but aborted it when it started building cups Jun 02 23:26:03 bgamari: try minimal-image Jun 02 23:28:35 good night Jun 02 23:29:10 khem: Thanks **** ENDING LOGGING AT Thu Jun 03 02:59:57 2010