**** BEGIN LOGGING AT Thu Aug 05 02:59:57 2010 Aug 05 04:05:42 http://mislav.uniqpath.com/2010/07/git-tips/ Aug 05 04:32:25 03Roman I Khimov  07org.openembedded.dev * r2f17293d31 10openembedded.git/recipes/snort/ (5 files in 2 dirs): (log message trimmed) Aug 05 04:32:25 snort: merge Altell packaging fixes Aug 05 04:32:25 Snort was independently packaged by Altell some time ago, so merge things Aug 05 04:32:25 we've done into OE.dev: Aug 05 04:32:25 * fix packaging of snort dynamic libs Aug 05 04:32:26 * fix DEPENDS, libnet is not needed is not specifically used Aug 05 04:32:26 * fix RDEPENDS, no need to specify libs expicitly, barnyard is rather a Aug 05 04:32:37 03Roman I Khimov  07org.openembedded.dev * r57788916d7 10openembedded.git/recipes/libprelude/ (4 files in 2 dirs): (log message trimmed) Aug 05 04:32:37 libprelude: new recipe Aug 05 04:32:37 Libprelude is a library that guarantees secure connections between Aug 05 04:32:37 all sensors and the Prelude Manager. Libprelude provides an Application Aug 05 04:32:38 Programming Interface (API) for the communication with Prelude Aug 05 04:32:38 sub-systems, it supplies the necessary functionality for generating and Aug 05 04:32:39 emitting IDMEF events with Prelude and automates the saving and Aug 05 04:41:51 03Roman I Khimov  07org.openembedded.dev * r8796aaeac2 10openembedded.git/recipes/libgcrypt/libgcrypt_1.1.91.bb: (log message trimmed) Aug 05 04:41:52 libgcrypt: remove version 1.1.91 Aug 05 04:41:52 * dated December 2003 Aug 05 04:41:52 * even back then it was an unstable development release Aug 05 04:41:52 * not using libgcrypt.inc Aug 05 04:41:52 * not pinned by any public distro Aug 05 04:41:52 * has legacy staging Aug 05 04:42:14 03Roman I Khimov  07org.openembedded.dev * r38da3c44d6 10openembedded.git/ (6 files in 4 dirs): Aug 05 04:42:14 openldap: remove versions 2.2.29 and 2.3.11 Aug 05 04:42:14 Both are almost 5 years old and seem to be unmaintained, also not supported Aug 05 04:42:14 by upstream authors and probably containing tons of known vulnerabilities. Aug 05 04:42:14 Signed-off-by: Roman I Khimov Aug 05 04:49:44 03Roman I Khimov  07org.openembedded.dev * r7fe3d2f8ac 10openembedded.git/recipes/libgpg-error/ (3 files in 2 dirs): Aug 05 04:49:44 libgpg-error: remove versions 0.6 and 0.7 Aug 05 04:49:44 * dated November 2003 and March 2004 respectively Aug 05 04:49:44 * unstable releases even back then Aug 05 04:49:44 * not pinned by any public distro Aug 05 04:49:45 * version 0.6 has legacy staging Aug 05 04:49:45 Signed-off-by: Roman I Khimov Aug 05 05:03:10 kergoth, i think im getting a do_rootfs failure as a result of your commit acc720fa80227d08fd15764117e0c34e2387e804 - http://pastebin.com/hh5cCwzA Aug 05 05:04:21 as my commit explains, the issue is the opkg/opkg-collateral packages, they blow away the opkg config file we need to install more packages with :\ Aug 05 05:04:36 probably best to create the config file outside of the rootfs and point it to it explicitly Aug 05 05:05:02 hmm Aug 05 05:06:48 actually, the problem as i see it is that angstrom-2008.1.conf has IPKG_VARIANT = "opkg angstrom-version ${ANGSTROM_FEED_CONFIGS}" Aug 05 05:06:56 your commit just exposed that as dumb Aug 05 05:07:11 heh, yeah, that's a pretty lame way to add those Aug 05 05:08:33 hrm Aug 05 05:09:27 okay, maybe I was wrong about why its doing it.. it definitely refuses to install anything after it installs opkg, though. must be something to do with the /usr/lib/opkg files, not the config file Aug 05 05:09:41 since it does point to an opkg.conf that's in sysroots, not the rootfs Aug 05 05:10:31 guess we should really be looking at the opkg package to see why it blows away the existing database (assuming thats what it does) Aug 05 05:10:38 it probably wipes the package list database Aug 05 05:10:49 :) Aug 05 05:11:31 a subsequent opkg update also appears to hang, or attempts to prompt with a non-interactive stdin Aug 05 05:11:39 * kergoth was getting really annoyed with opkg yesterday Aug 05 05:12:32 there is no prompting in opkg anymore Aug 05 05:13:11 ah Aug 05 05:13:19 * kergoth shrugs, will revisit later Aug 05 05:13:24 the offline root it specifies, /mnt/oe/tmp/rootfs/minimal-gpe-image-tmp is not right Aug 05 05:13:38 well, sort of Aug 05 05:13:53 * grg thinks Aug 05 05:15:26 isn't the -tmp dir for -t, not -o? Aug 05 05:15:45 yeah. i pasted the wrong one Aug 05 05:16:36 i just checked the package list and it contains angstrom-version. The ipk file is in the right spot too... Aug 05 05:29:34 03Vladimir Sorokin  07org.openembedded.dev * r95fb07577f 10openembedded.git/recipes/corosync/ (corosync_1.2.6.bb files/fix-define-semun-union.patch): Aug 05 05:29:34 corosync: new version 1.2.7 Aug 05 05:29:34 * bugfix release, considered as safe upgrade Aug 05 05:29:34 Signed-off-by: Vladimir Sorokin Aug 05 05:29:34 Signed-off-by: Roman I Khimov Aug 05 05:38:21 kergoth: and it's set to empty when ONLINE_PACKAGE_MANAGEMENT == "none", isn't it? Aug 05 05:39:53 kergoth, strace says opkg is looking here for package lists: /mnt/oe/tmp/rootfs/minimal-gpe-image//var/lib/opkg/oe-qemumipsel Aug 05 05:41:07 it should be looking in /mnt/oe/tmp/rootfs/minimal-gpe-image/usr/lib/opkg/lists/oe-qemumipsel Aug 05 06:10:23 Hi all Aug 05 06:10:49 I was looking at the OE recipies Aug 05 06:11:09 there is a variable named NATIVE_INSTALL_WORKS Aug 05 06:11:29 I just want to know that in which scenrio one want to meke it true Aug 05 06:11:40 make it true Aug 05 06:42:57 can anybody tell me what is the benefit of using NATIVE_INSTALL_WORKS varaible? Aug 05 06:45:07 03Roman I Khimov  07org.openembedded.dev * rd3d01844cb 10openembedded.git/recipes/suricata/libhtp_0.2.3.bb: (log message trimmed) Aug 05 06:45:07 libhtp: new recipe Aug 05 06:45:07 The HTP Library is an HTTP normalizer and parser written by Ivan Ristic Aug 05 06:45:07 of Mod Security fame for the OISF. This integrates and provides very Aug 05 06:45:07 advanced processing of HTTP streams for Suricata. The HTP library is Aug 05 06:45:08 required by the engine, but may also be used independently in a range Aug 05 06:45:08 of applications and tools. Aug 05 06:45:12 03Roman I Khimov  07org.openembedded.dev * r5e02e8de46 10openembedded.git/recipes/pf_ring/ (3 files in 2 dirs): Aug 05 06:45:12 pf_ring: add recipes for PF_RING kernel module and userspace library Aug 05 06:45:12 PF_RING is a new type of network socket that dramatically improves Aug 05 06:45:12 the packet capture speed. Aug 05 06:45:12 Signed-off-by: Roman I Khimov Aug 05 06:45:12 03Roman I Khimov  07org.openembedded.dev * r64b73f5d3b 10openembedded.git/conf/bitbake.conf: Aug 05 06:45:13 bitbake.conf: trust server certificate when doing svn over https Aug 05 06:45:13 Fixes fetching from https:// svn repos with self-signed certs. Aug 05 06:45:14 Signed-off-by: Roman I Khimov Aug 05 06:45:42 gm Aug 05 06:46:08 barnyard2: new recipe Aug 05 06:46:08 Barnyard2 is a fork of the original barnyard project, designed Aug 05 06:46:08 specifically for Snort's new unified2 file format. Aug 05 06:46:08 Signed-off-by: Roman I Khimov Aug 05 06:50:28 Noor: NATIVE_INSTALL_WORKS is needed when a recipe needs a special do_install Aug 05 06:51:10 what do you mean by special do_install Aug 05 06:51:17 Noor: e.g. when you use BBCLASSEXTEND but the default do_install does not do the job Aug 05 06:51:33 you have to provide explicit do_install in the recipe Aug 05 06:52:10 NATIVE_INSTALL_WORKS tell bitbake of explicit do_install Aug 05 06:52:28 khem are you sure that is correct? Aug 05 06:52:55 03Roman I Khimov  07org.openembedded.dev * r929d154c3f 10openembedded.git/recipes/pf_ring/libpfring_svn.bb: Aug 05 06:52:55 libpfring: remove deprecated patch options Aug 05 06:52:55 Sorry, didn't check this one when picking from local branch. Aug 05 06:52:55 Signed-off-by: Roman I Khimov Aug 05 06:52:59 eFfeM_work: well thats what it should do, do you have other experiences Aug 05 06:53:43 well if I look in bison there is a NATIVE_INSTALL_WORKS = "1" in the recipe (actually the inc file) and there is no do_install in the recipe Aug 05 06:54:02 in the latest source this varaible is effecting do_populate_sysroot function which handles staging stuff Aug 05 06:54:13 Noor: yes Aug 05 06:54:31 do_populate_sysroot is the one thats affected Aug 05 06:54:46 it was my impression (but I am not a wiz in that area) that NATIVE_INSTALL_WORKS meant that the std install (the one for the non-native recipe) is also to be used for native Aug 05 06:57:11 peeked in the code: Aug 05 06:57:12 elif bb.data.getVar('NATIVE_INSTALL_WORKS', d, 1) == "1": Aug 05 06:57:12 legacy = False Aug 05 06:57:31 so it just turns off legacy staging Aug 05 06:57:39 yeah Aug 05 06:57:51 exactly Aug 05 06:58:04 afaik this means it uses do_install i.s.o do_stage and if there is a do_install in your recipe it will be taken otherwise the default rule Aug 05 06:58:12 (e.g. if you have autotools) Aug 05 06:58:34 yes Aug 05 06:58:36 khem, you wrote: Aug 05 06:58:37 (08:50:52 AM) khem: you have to provide explicit do_install in the recipe Aug 05 06:58:37 (08:51:29 AM) khem: NATIVE_INSTALL_WORKS tell bitbake of explicit do_install Aug 05 06:58:58 I forgot to mentioned do_install instead of do_stage Aug 05 06:58:58 the first explicit is not necessarily true Aug 05 06:59:17 for autotools recipes there is no need to provide an explicit do_install Aug 05 06:59:19 its true if someone had do_stage Aug 05 06:59:26 that's why I reacted Aug 05 06:59:27 ofc Aug 05 06:59:52 unless do_stage functionality is already in the auto do_install Aug 05 07:00:04 * khem seems not to be typing well today after a long day Aug 05 07:00:29 khem, guess it is about bedtime for you Aug 05 07:00:34 9 am here Aug 05 07:03:00 khem: is it possible that some version of gcc-4.5 doesn't support -fforward-propagate? (it's used in ncurses.inc and someone send me config.log where it shows gcc version 4.5.1 20100722 (prerelease) (GCC) and then fails later with "cc1: error: unrecognized command line option "-fforward-progagate"" Aug 05 07:04:01 JaMa: not possible Aug 05 07:04:54 khem: same gcc with same CFLAGS etc works here.. Aug 05 07:05:08 yeah he must be using old compiler Aug 05 07:05:53 khem: http://paste.pocoo.org/show/246163/ here is config.log if you notice something Aug 05 07:06:01 03Roman I Khimov  07org.openembedded.dev * rcc8dd8133d 10openembedded.git/recipes/perl/ (3 files): Aug 05 07:06:01 lib(html-mason|log-dispatch|module-build)-perl: remove do_stage no-ops Aug 05 07:06:01 Signed-off-by: Roman I Khimov Aug 05 07:06:01 03Roman I Khimov  07org.openembedded.dev * rf0a40c3524 10openembedded.git/recipes/perl/libparams-validate-perl_0.86.bb: (log message trimmed) Aug 05 07:06:01 libparams-validate-perl: update to version 0.95 Aug 05 07:06:02 * fixes some bugs Aug 05 07:06:03 * fixes build failure like: Aug 05 07:06:03 | Validate.xs: In function 'get_type': Aug 05 07:06:04 | Validate.xs:212: error: duplicate case value Aug 05 07:06:04 | Validate.xs:209: error: previously used here Aug 05 07:06:42 hmm nice native-sdk-image boots on qemuarm with autotools_staging gone :) Aug 05 07:06:49 I need few acks Aug 05 07:08:40 JaMa: heh there is a typo Aug 05 07:08:43 khem, cool Aug 05 07:08:44 -fforward-progagate Aug 05 07:08:52 should be propagate Aug 05 07:09:06 khem I'll try to build the specials tonight and if they work I can ack them Aug 05 07:09:37 don't think I manage to build all the recipes with the s/autotools_stage/autotools/ Aug 05 07:09:49 eFfeM_work: ok Aug 05 07:09:53 thx Aug 05 07:09:59 morning Aug 05 07:10:16 so if recipe is inherited from autotools you dont need to use this NATTIVE_INSTALL_WORKS Aug 05 07:10:27 JaMa: with gcc 4.5 you can eliminate all -fforward-propagate fixes Aug 05 07:10:27 but if the specials work, I'm happy to ack the big one too (code review does not indicate an issue) Aug 05 07:10:28 otherwise you have to make it true Aug 05 07:10:55 eFfeM_work: it all works well for me. Aug 05 07:11:01 I have few other fixes lined up Aug 05 07:11:08 Noor: not fully sure, but it is a quick test so if it fails without and works with, add it Aug 05 07:11:10 after them Aug 05 07:11:12 khem, figured it would Aug 05 07:11:24 khem, good work Aug 05 07:11:40 there were 2 I ran into seemed not a fall out from this work though Aug 05 07:11:55 btw, I would like to be able to do a do_install_append and append to the automake do_install Aug 05 07:13:10 JaMa: one needs 181244a3a629b6dacd5f221f9e9380ba9a078bb0 Aug 05 07:13:23 I have patches here to remove them Aug 05 07:13:35 but before that we have to switch to 4.3 Aug 05 07:13:41 oops 4.5 Aug 05 07:14:11 khem: I see, thanks, no idea why he doesn't have latest (as he said...) Aug 05 07:15:09 np Aug 05 07:15:24 * khem sleeps Aug 05 07:16:47 gnight Aug 05 07:24:13 good morning Aug 05 07:38:15 eFfeM_work: and for those native recipes who are not inherited from autotools this varaible should be in the recipe. it will make sure that do_install function is used insted of do_stage. correct me if I am wrong. Aug 05 07:38:35 Noor: correct Aug 05 07:39:03 ~seen blinvt' Aug 05 07:39:06 woglinde: i haven't seen 'blinvt'' Aug 05 07:39:09 ~seen blindvt' Aug 05 07:39:10 woglinde: i haven't seen 'blindvt'' Aug 05 07:39:20 Thanks eFfeM_work Aug 05 07:39:35 yw Aug 05 07:39:59 ~seen blindvt` Aug 05 07:40:00 blindvt` is currently on #oe #uclibc. Has said a total of 50 messages. Is idling for 17h 47m 11s, last said: 'B_Lizzard, i've sent the urgent bits to the list now, fwiw'. Aug 05 07:40:12 woglinde: guess you want that Aug 05 07:40:33 (it is a backquote) Aug 05 07:41:15 hm found it Aug 05 07:41:18 anyway Aug 05 07:41:35 ffmpeg libavocdec needs -D_C99_SOURCE for one file Aug 05 07:48:58 I need to add a squashfs image (say A-img) to another image (say B-img) (so an image within an image), but I am faced with the following issue: I added a DEPENDS ="A-img" to my B-img recipe; then bitbake B-img will also build A-img, but not the do_rootfs task. However, if I do a bitbake A-img the rootfs task *is* executed; how can I achieve that this also happens if I build A-img ??? Aug 05 07:49:12 oops last A-img should read B-img Aug 05 07:56:51 03jkridner@beagleboard.org  07org.openembedded.dev * r24a990c800 10openembedded.git/recipes/images/beagleboard-test-image.bb: Aug 05 07:56:51 beagleboard-test-image: add cpuburn-neon Aug 05 07:56:51 Signed-off-by: Jason Kridner Aug 05 07:56:51 Signed-off-by: Koen Kooi Aug 05 08:10:48 hi ant Aug 05 08:11:01 mornin' Aug 05 08:11:13 eFfeM_work: something like do_configure[depends] += "${INITRAMFS_IMAGE}:do_rootfs" Aug 05 08:11:50 ant_work: thanks, will try this Aug 05 08:14:45 ant_work: works like a charm! thanks alot!!!!! Aug 05 08:17:35 yw Aug 05 08:18:20 iirc you want to flash nand from a booted initramfs? Aug 05 08:19:32 ant_work: almost, the image for flash is too big for initramfs, so just made a regular image that contains the squasfs image to be flashed Aug 05 08:20:25 factory then can boot from the image with the squash img in it and it will install the squash image to the flash Aug 05 08:25:28 hi all, I am building Angstrom for a DM365 device Aug 05 08:25:35 is there someone who can suggest me a suitable oe-devel stable commit? Aug 05 08:26:12 other solution could be to put the initramfs.cpio.gz and the image to flash on a SD/CF and customize the init Aug 05 08:26:23 eltazar use dev branch Aug 05 08:26:35 eltazar all other will give you pain Aug 05 08:26:59 ok for dev branch Aug 05 08:27:20 eFfeM_work: or even embed the 'flasher'-initramfs in the kernel Aug 05 08:27:21 it is right the HEAD commit, or there is a stable one before it? Aug 05 08:28:14 ant_work: iirc that was not possible because of size limitations for either initramfs or kernel itself Aug 05 08:28:15 eltazar whats your goal? Aug 05 08:28:46 ant_work: the image to be flashed is 6.7 MB Aug 05 08:28:59 I need a stable rootfs to start developing a device Aug 05 08:29:39 eltazar use http://www.angstrom-distribution.org/narcissus/ Aug 05 08:29:39 eFfeM_work: from a klibc userland you could do all, keeping the image elsewhere (block dev, nfsroot, ...) Aug 05 08:30:13 eFfeM_work: you would just boot a 'special' kernel embeding a small initramfs Aug 05 08:30:24 no need to embedd the image Aug 05 08:30:38 ant_work: understood, but then I still would need to cook up an initramfs Aug 05 08:30:41 this seemed simpler Aug 05 08:31:00 I need also to be able to re-create the entire system, also with an OE overlay Aug 05 08:31:02 initramfs is very easy in OE nowadays Aug 05 08:31:17 maybe Narcissus can be more suitable for a demo Aug 05 08:31:26 basically i now can reuse the kernel that does the upgrade and flash it in the device; so no need to have a kernel with and without initramfs in it Aug 05 08:31:36 eltazar hm Aug 05 08:31:48 eltazar you cab try the stable/2009 branch Aug 05 08:32:03 eltazar but it maybee misses some features Aug 05 08:32:24 for the dm365 kernel Aug 05 08:32:39 eFfeM_work: if the kernel is initramfs-enabled you can just pass the 'service' initrd from the bootloader Aug 05 08:32:48 without creating a special kernel Aug 05 08:33:00 true Aug 05 08:33:09 morning all Aug 05 08:33:13 hi rp Aug 05 08:34:09 eFfeM_work: anyway much of our work was about having the kernel in /boot and not on the machine Aug 05 08:34:28 ugrade is much easier Aug 05 08:35:32 we have the kernel in a separate partition just as raw data and we use bootm to start it (NOR flash, not nand) Aug 05 08:35:48 upgrade is dd a new kernel to the partition Aug 05 08:35:58 and ofc ping pong strategy to avoid breaking things Aug 05 08:36:07 woglinde: stable/2009 is fine for Angstrom too? Aug 05 08:36:52 yes Aug 05 08:37:18 ok, thank you for your help, I'll give it a try Aug 05 08:38:09 eFfeM_work: you always boot from the same block device? Aug 05 08:43:28 ant_work: no, i have a u-boot script and env vars that choose between two kernels and rootfs-es Aug 05 08:51:46 \quit Aug 05 08:53:22 eFfeM_work: assumed u-boot works fine on your machine that's ok Aug 05 08:54:27 ant_work: if u-boot does not work we definitely have an isse (and u-boot will never be upgraded, actually it might well be that the env never changes (we might store the bit to decide what kernel to boot in some eeprom) Aug 05 08:55:23 no idea about the kind of device, but I assume the final user should not flip too much between kernels Aug 05 08:55:35 and rootfs's Aug 05 08:57:27 03Martin Jansa  07org.openembedded.dev * rddc5b66ddc 10openembedded.git/conf/bitbake.conf: (log message trimmed) Aug 05 08:57:27 bitbake.conf: subversion cmds, remove --trust-server-cert and add --force, --accept Aug 05 08:57:27 * --trust-server-cert is supported only in >=svn-1.6 Aug 05 08:57:27 http://subversion.apache.org/docs/release-notes/1.6.html Aug 05 08:57:27 * --force allows to overwrite existing dirs Aug 05 08:57:27 ie Aug 05 08:57:27 matchbox-panel-2-icon-themes_0.0.1.bb Aug 05 08:58:20 how do I enable debugfs in my kernel? Aug 05 08:58:39 ant_work: user cannot flip, user does not even have access to console or uboot, Aug 05 08:59:05 this is a real embedded device, flipping will only be done after an upgrade (if the initial reboot is ok) Aug 05 09:01:03 eFfeM_work: any idea about subversion versions in other LTS distros? I <1.5 still used somewhere? Aug 05 09:01:17 s/I/Is/ Aug 05 09:02:16 JaMa: no real idea, use ubuntu 10.04 at work and opensuse 11.2 at home, however at work we still have some people using 8.04 Aug 05 09:02:32 no idea what svn they have, can check Aug 05 09:03:28 svn, version 1.1.1 (r11581) in RHEL4 :) Aug 05 09:03:35 Hi, folks, have a short question. I have a custom kernel bb, which builds and runs fine. But, when I try to set the prefference in the local.conf with PREFERRED_PROVIDER_virtual/kernel = "linux-custom", I get the error, multiple providers available for virtual/kernel and the default kernel from angstrom builds. How can I override this? I set the DEFAULT_PREFERENCE_beagleboard = "1" in my kernel recepie. Aug 05 09:03:48 JaMa: wrt Gentoo, seems binutils issue has gone after khem's rework Aug 05 09:03:50 I hope nobody is using it for OE Aug 05 09:04:27 ant_work: because --enable-shared after http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=c52193ee037fe3d7c469256585ab70cb6c774ea8 Aug 05 09:04:28 JaMa: was also thinking of rhel4 Aug 05 09:04:48 and it is used by some, that is why we ended up with linux-libc-headers-native Aug 05 09:05:15 eFfeM_work: maybe we can try workaround it with subversion-native .. Aug 05 09:05:30 jama, could work Aug 05 09:06:07 see this thread on linux-libc-headers-native: http://www.mail-archive.com/openembedded-devel@lists.openembedded.org/msg06930.html Aug 05 09:07:17 iirc Tom Rini mentioned RHEL4 Aug 05 09:13:58 RHEL4? I have problems to get oe to build on RHEL5 due to python 2.4 Aug 05 09:14:59 * JaMa hopes that even people with RHEL4 have backported newer svn Aug 05 09:20:58 jama do we use libxcb/x11 wrapper in oe? Aug 05 09:21:30 wrapper? Aug 05 09:22:35 cbrake, Looks like there should be a fix rolling out for the Ubuntu 10.04 gzip problem relatively soon. Aug 05 09:23:20 oh..still open bug? Aug 05 09:23:43 Yup. Aug 05 09:24:08 I run "bitbake virtual/config -c menuconfig" and disable the device drivers for sound, but I still get CONFIG_SOUND=y in the generated defconfig. any ideas? Aug 05 09:25:05 tasslehoff: generated defconfig? or tmpdir/work/machine/linux/.config? Aug 05 09:25:39 -c menuconfig is not updating defconfig in OE metadata.... Aug 05 09:26:06 JaMa: ooh, in tmpdir, but I still picked the wrong one. thanks! Aug 05 09:34:30 var_dump Aug 05 09:34:59 Whoops! (Focus followed mouse, which moved...) Aug 05 09:42:40 03Enrico Scholz  07org.openembedded.dev * r0ecaae77eb 10openembedded.git/recipes/libfakekey/libfakekey_svn.bb: Aug 05 09:42:40 libfakekey: get rid of legacy staging Aug 05 09:42:40 Legacy staging is not needed; things are done by normal installation Aug 05 09:42:40 already. Aug 05 09:42:40 Signed-off-by: Enrico Scholz Aug 05 09:42:50 03Enrico Scholz  07org.openembedded.dev * reb74503cb6 10openembedded.git/recipes/libmatchbox/libmatchbox.inc: Aug 05 09:42:50 libmatchbox: get rid of legacy staging Aug 05 09:42:50 Legacy staging is not needed; things are done by normal installation Aug 05 09:42:50 already. Aug 05 09:42:50 Signed-off-by: Enrico Scholz Aug 05 09:49:40 03Enrico Scholz  07org.openembedded.dev * rfcae56e05c 10openembedded.git/recipes/ncurses/ (ncurses-5.7/tic-hang.patch ncurses_5.7.bb): (log message trimmed) Aug 05 09:49:40 ncurses: added 5.7 recipe Aug 05 09:49:40 It uses ideas from the old 5.4 recipe but brings a lot of new features Aug 05 09:49:40 like widec libs, pkgconfig files or splitted library packages. Aug 05 09:49:41 As lot of packages depend on ncurses, whole distribution must be Aug 05 09:49:41 probably rebuild when updating to this new version. Aug 05 09:49:42 Due to this and the complex recipe, it has a negative DEFAULT_PREFERENCE Aug 05 09:54:26 03Martin Jansa  07org.openembedded.dev * r91ac9b51b0 10openembedded.git/recipes/eds/eds-dbus_git.bb: Aug 05 09:54:27 eds-dbus: reenable calendar, was disabled in svn->git transition commit (no idea why) Aug 05 09:54:27 * compiles fine Aug 05 09:54:27 * dates depends on libecal Aug 05 09:54:27 Signed-off-by: Martin Jansa Aug 05 09:59:25 03Koen Kooi  07org.openembedded.dev * rc80c701f9f 10openembedded.git/recipes/intltool/intltool.inc: intltool: remove legacy staging Aug 05 10:27:41 RP__: about packaged staging, two days ago rebuild from pstage broke on 'sh: quilt not found' Aug 05 10:27:57 somehow quilt native was not yet built Aug 05 10:28:49 ant_work: I'm busy working on a replacement for packaged-staging ;-) Aug 05 10:28:56 first packages erroring out: gnu_config and linux_libc_headers Aug 05 10:29:05 great Aug 05 10:29:06 ;) Aug 05 10:29:10 ant_work: In fact I just sent out an email about it Aug 05 10:29:26 he..that's why I posted here :) Aug 05 10:30:00 ant_work: For that reason I'll be concentrating on the new one, the old one has too many issues :( Aug 05 10:31:03 I'l let you concentrate Aug 05 10:31:16 * ant_work heads out for a snack Aug 05 10:31:18 send those massage girls to RPs Aug 05 11:06:12 kergoth: Re acc720fa80227d08fd15764117e0c34e2387e804 This fails for me since you unconditionally call 'opkg-cl ${IPKG_ARGS} install ${PACKAGE_INSTALL_PKGMGR}' and i (as in micro-base-image) don't have ONLINE_PACKAGE_MANAGEMENT Aug 05 11:07:12 call back the massage girls, 35 mins is more than enough, time for RP to go back to work :-) Aug 05 11:08:16 blindvt`: Graham Gower sent better patch and proposed revert.. and I agree with Graham (same issue as you) Aug 05 11:08:31 what i don't quite get in that whole class is why you don't just say for i in ${SOMETHING}; What's the purpose of [ ! -z ] all over the place (let alone that this should nowadays be [ -n "" ]) Aug 05 11:09:53 JaMa, ah, great. Didn't read mail yet, sorry for the noise Aug 05 11:11:37 blindvt`: np, good that you confirmed it... Aug 05 11:15:30 JaMa, the correct fix is to move 'opkg-cl ${IPKG_ARGS} install ${PACKAGE_INSTALL_PKGMGR}' into the 'if [ "${ONLINE_PACKAGE_MANAGEMENT}" != "none" ]; then' guarded block Aug 05 11:15:42 obviously ;) Aug 05 11:16:51 test [!-z ""] should be 'is not null'. Dpn't think '-n' would work with undefined Aug 05 11:17:24 hm Aug 05 11:17:37 blindvt I have made a patch for isnan Aug 05 11:17:40 in uclibc Aug 05 11:18:31 blindvt`: The unary operator -z tests for a null string, while -n or no operator at all returns True if a string is not empty. Aug 05 11:18:36 http://www.ibm.com/developerworks/library/l-bash-test.html Aug 05 11:19:00 bashism Aug 05 11:19:05 yep Aug 05 11:19:28 I asked yesterday too... Aug 05 11:19:46 someone had grepped the recipes for a particular bashism Aug 05 11:19:57 can't find the log Aug 05 11:20:21 was a couple of weeks ago Aug 05 11:22:58 test -n is required by SUSv4, no longer bashism, FYI. http://www.opengroup.org/onlinepubs/9699919799/utilities/test.html Aug 05 11:23:53 *g* Aug 05 11:23:54 ant_work, don't get me wrong but some IBM infotainment thing is not relevant as opposed to the standard ;) Aug 05 11:24:43 well, I still think IBM links are one class better than 'first-googled-page' :) Aug 05 11:24:56 ant yes mostly Aug 05 11:25:11 woglinde, what's wrong with isnan as it is now? It expands to the correct type-variant Aug 05 11:25:16 but the informations can become quite old too Aug 05 11:25:35 blindvt in needs ISO_C99 Aug 05 11:25:46 in bits/mathcalls.h Aug 05 11:25:56 woglinde: the issue of distinguish btw null/undefined/zero is way old Aug 05 11:26:06 ant *g* sure Aug 05 11:26:43 blindvt I and I changed the defconfig's for arm ti GENERIC_ARM Aug 05 11:26:55 so cpuflags from uclibc isnt used Aug 05 11:27:10 will push the patches soon Aug 05 11:27:16 all done aginst git Aug 05 11:30:10 Folks, have a short question. I have a custom kernel bb, which builds and runs fine. But, when I try to set the prefference in the local.conf with PREFERRED_PROVIDER_virtual/kernel = "linux-custom", I get the error, multiple providers available for virtual/kernel and the default kernel from angstrom builds. How can I override this? I set the DEFAULT_PREFERENCE_beagleboard = "1" in my kernel recepie. Aug 05 11:33:26 show recipe? Aug 05 11:33:37 one sec. Aug 05 11:33:49 hi hrw Aug 05 11:35:00 http://gist.github.com/509602 Aug 05 11:36:40 hrw: I have impression that machine.conf setting overwrites the local.conf setting, is this the case ? Aug 05 11:37:50 hello, whats the suggested branche for oe when building for armv4t device? Aug 05 11:38:04 yes Aug 05 11:38:14 Seppoz_work depends on what you want Aug 05 11:38:26 armv4t builds even fine with the dev branch Aug 05 11:38:26 id like to build a x11-gpe-image Aug 05 11:38:36 well i tried the dev branche Aug 05 11:38:38 yesterday Aug 05 11:38:52 and stucked at some locale-en-us or so Aug 05 11:38:58 the abse image built just fine Aug 05 11:39:01 *base Aug 05 11:39:03 lgopher: you need to change machine config Aug 05 11:39:24 speppoz intressting becasue locale-en-us is needed for base too Aug 05 11:39:29 or better generated Aug 05 11:40:01 yes, at a build earlyer at the same day i get stuck at missing python xcbgen Aug 05 11:40:16 it says opkg cannot find locale-en-us Aug 05 11:40:26 package doesnt exist Aug 05 11:40:27 hrw: ok, I have an overlay and wanted to avoid to copying the machine files, but will do so now. Thanks. Aug 05 11:40:33 i also cant fine it in the recepies Aug 05 11:41:03 is it ok to use ubuntu 10.04 for oe? i aw most ppl using fc11 Aug 05 11:41:09 *saw Aug 05 11:41:45 speeoz tar can make some anger Aug 05 11:41:53 but in general it works Aug 05 11:43:09 i did a git pull be4 i went to work today so maybe it was just some issue in the particular revision, but ran into another error earlyer Aug 05 11:43:20 with some other lib Aug 05 11:43:23 seppoz dev is in flux Aug 05 11:43:29 what is suggested distro for bb Aug 05 11:43:29 so pull often Aug 05 11:43:35 angstroem Aug 05 11:43:55 should i switch to the stable 2009 branche? Aug 05 11:44:39 id really love to build angstrom 2010.7 tho Aug 05 11:45:23 Seppoz_work: let me guess... you have GLIBC_LIMIT_LOCALE (or sth like that) in local.conf Aug 05 11:45:36 i dont think so no Aug 05 11:45:45 lemme check tho Aug 05 11:46:05 i *think* i have somewhat like that Aug 05 11:46:10 saying en_us Aug 05 11:46:16 now that u mention it Aug 05 11:46:31 i saw that is some other ant thought it might be usefull Aug 05 11:46:36 should i drop that line? Aug 05 11:46:44 ideal local.conf contains: BBFILES, DISTRO, MACHINE - **nothing** more Aug 05 11:47:21 I know that in internet (and in metadata) users can find ~0.5MB long local.conf files which will give lot of switches Aug 05 11:47:24 mine doesnt contain much more afaik just *some* lines like builder and so Aug 05 11:47:38 i use the default i got from the angstrom script Aug 05 11:47:42 pastebin it Aug 05 11:47:54 give me a sec plz gotta boot the system Aug 05 11:48:59 speoz try http://www.angstrom-distribution.org/narcissus/ Aug 05 11:49:18 i did try that Aug 05 11:49:25 it dont build essential packages Aug 05 11:49:33 o.O? Aug 05 11:49:38 btw what device it's? Aug 05 11:49:45 hi ynezz Aug 05 11:49:47 beagle Aug 05 11:49:47 last i built missed the net package in busybox.. so no ifconfig :) Aug 05 11:49:48 he said Aug 05 11:50:02 ah, can't find it in backlog Aug 05 11:50:04 no, its mini2440 im building for Aug 05 11:50:07 hi woglinde Aug 05 11:50:11 ieeehks Aug 05 11:50:15 mini2440 Aug 05 11:50:22 yup Aug 05 11:50:23 all mini2440 have some problems Aug 05 11:50:26 I never saw Aug 05 11:50:52 yes pretty sucky board IMO, just trying to help out a friend Aug 05 11:51:10 Seppoz_work hm I didnt meant hardware Aug 05 11:51:20 hw is fine Aug 05 11:51:25 sw support is just fail Aug 05 11:51:49 it should be similar to GTA01 Aug 05 11:51:55 the manufactor told me he dont wanna employ ppl for *stupid uropean qestions* Aug 05 11:51:56 lol Aug 05 11:52:05 at least the CPU is same isn't it? Aug 05 11:52:08 :) Aug 05 11:52:12 yes its the same Aug 05 11:52:19 i have a working kernel now Aug 05 11:52:25 and *generally* Aug 05 11:52:28 antrom works out Aug 05 11:52:35 id just like to build 2010.7 Aug 05 11:52:38 ah Aug 05 11:52:41 and x11 Aug 05 11:52:48 which is very useless imo on the 2440 Aug 05 11:52:53 but my friend wants it Aug 05 11:53:00 ssh -X Aug 05 11:53:01 haha Aug 05 11:53:34 well i got that rnning earlyer, proplem.. no ifconfig in busybox and no tslib x11 input drivers Aug 05 11:53:47 rest worked out fine Aug 05 11:53:52 also matchbox worked Aug 05 11:54:38 http://pastebin.com/sxeRajjy Aug 05 11:54:45 thats the local.conf im having *here* Aug 05 11:55:01 the one at home is same just 1 more line with some en_us Aug 05 11:55:13 that i will remove as sonn as im at hiome ;P Aug 05 11:55:35 u see aything wrong here? Aug 05 11:57:41 ENABLE_BINARY_LOCALE_GENERATION = "0" Aug 05 11:57:46 is that wrong maybe? Aug 05 11:57:58 no Aug 05 11:58:17 ok! so maybe the only proplem is really that line i have there Aug 05 11:58:48 btw why MACHINE ?= "mini2440" Aug 05 11:59:30 ant_work its proably from this sucky outdated blog site Aug 05 11:59:37 ah, argh Aug 05 12:00:02 is that a proplem? Aug 05 12:00:07 should i use = Aug 05 12:00:12 no Aug 05 12:00:15 the script is from the angstrom site Aug 05 12:00:15 its just comsetic Aug 05 12:00:40 i mean its generated by the script from the angstrom site Aug 05 12:01:00 I think someone does # MACHNE = x; bitbake blah Aug 05 12:01:28 for multimachines, using a shell script Aug 05 12:02:09 so they override local.conf Aug 05 12:02:13 ? Aug 05 12:07:42 hello everybody Aug 05 12:09:54 GLIBC_GENERATE_LOCALES = "en_GB.UTF-8" Aug 05 12:09:59 thats what i have defined Aug 05 12:10:11 should i drop that line? Aug 05 12:11:06 depends on you if you need locales Aug 05 12:11:13 but ENABLE_BINARY_LOCALE_GENERATION = "0" Aug 05 12:11:17 disable it anyway Aug 05 12:11:42 well Aug 05 12:11:42 iirc en_US is somehow required Aug 05 12:11:47 the build needs them Aug 05 12:11:55 ???? Aug 05 12:12:04 it seems the image needs it Aug 05 12:12:05 Seppoz_work: don't set any Aug 05 12:12:06 the locale Aug 05 12:12:33 so drop ENABLE_BINARY_LOCALE_GENERATION = "0" and GLIBC_GENERATE_LOCALES = "en_GB.UTF-8" ? Aug 05 12:12:40 and let the system handle it Aug 05 12:12:46 or wahts the common way Aug 05 12:13:13 that's the common way. therwise you can try building both en_ Aug 05 12:13:42 ant_work: so u say i should drop both config lines right? Aug 05 12:13:53 i do a simple sample helloword and i would like build by bitbake, so i create hello.bb and i do :bitbake -b hello.bb there is task .... started task .......completed but i dont find an executable file i think hello.o Aug 05 12:13:54 yes Aug 05 12:13:59 any idea Aug 05 12:14:01 thanks! Aug 05 12:14:04 or you can keep ENABLE... 0 Aug 05 12:14:20 this should also work Aug 05 12:14:22 what is it doing exactly? Aug 05 12:14:32 probablythey interfere Aug 05 12:14:46 even if you set ENABLE.. to 0 Aug 05 12:14:52 so don't do it :) Aug 05 12:15:02 kamel: hello.o is not the execuable Aug 05 12:15:25 ant_work: ok thx, i will try to drop them as sonn as im at home Aug 05 12:15:33 i mean an object Aug 05 12:15:41 Seppoz_work: fwiw I did *months ago* builds enabling only a specific subset Aug 05 12:15:43 .ipk or .o like this Aug 05 12:16:00 i have a nother qestion, once i have an image built, and i change a .bb recepie will oe notice that? Aug 05 12:16:18 like modifying the x11-image.bb Aug 05 12:16:31 than doing bitbake x11-image Aug 05 12:16:38 will it take the new .bb? Aug 05 12:17:02 Seppoz_work after bitbake -b hello.bb we should find hello.ipk Aug 05 12:18:10 Seppoz_work: it should just work. You can always do -c clean and rebuild Aug 05 12:18:31 ant_work: whole image? :) Aug 05 12:18:58 yes, a single recipe Aug 05 12:19:19 not the files in the image :) Aug 05 12:19:23 i can do bitbake -c Aug 05 12:19:25 ? Aug 05 12:19:37 will it clean all? Aug 05 12:19:46 first just try bitbake x11-image Aug 05 12:20:05 im just interessted Aug 05 12:20:06 if no-go, then Aug 05 12:20:12 bitbake -c clean x11-image Aug 05 12:20:16 will -c also clean the toolchain Aug 05 12:20:21 no Aug 05 12:20:33 what do i do if i "git pull" Aug 05 12:20:41 why image? just bump PR in changed package Aug 05 12:20:51 afte git pull do i rebuild the whole thing? Aug 05 12:20:54 I suppose he's ading tasks Aug 05 12:21:24 well, I don't have crystal ball :) Aug 05 12:21:31 Seppoz_work: if there are new PR some packages could be rebuilt Aug 05 12:21:39 NOT? i have 2 if you want one Aug 05 12:21:39 xD Aug 05 12:21:40 depends on which pull Aug 05 12:21:55 will bitbake notice automatically? Aug 05 12:21:59 yes Aug 05 12:22:05 sounds cool! Aug 05 12:22:15 so i can just bb x11-image Aug 05 12:22:19 how would you do it otherwise? manually ? Aug 05 12:22:21 and it will rebuild what cnaged Aug 05 12:22:32 i have no idea sry ;P Aug 05 12:22:56 any one have an idea Aug 05 12:22:59 np, takes time Aug 05 12:23:08 kamel: show the .bb Aug 05 12:23:29 kamel: not here, but pls try to use pastebin Aug 05 12:24:03 ynezz can you give me a link of pastebin Aug 05 12:24:19 ant_work: the first time i use bb at all :( i was building all packages by hand be4 which i figured tok forever ;P Aug 05 12:24:32 http://pastebin.com/ Aug 05 12:25:21 kamel: http://tinyurl.com/ah9j6f Aug 05 12:25:39 hehe Aug 05 12:25:53 that site is astill funnyest ever Aug 05 12:25:55 :P Aug 05 12:26:13 hi kergoth Aug 05 12:29:40 how i pastebin Aug 05 12:29:40 email? Aug 05 12:29:46 ~pastebin Aug 05 12:29:48 [~pastebin] A "pastebin" is a web-based service where you should paste anything over 3 lines so you don't flood the channel. Here are links to a few : http://www.pastebin.com , http://pastebin.ca , http://channels.debian.net/paste , http://paste.lisp.org , http://bin.cakephp.org/ , http://asterisk.pastey.net/ , or install pastebinit with yum or aptitude. Aug 05 12:33:17 ieiehlss Aug 05 12:33:25 when I use BBCLASSEXTEND in a recipe Aug 05 12:33:30 do_stage isnt called anymore Aug 05 12:33:53 but we need it until the java situation is fixed Aug 05 12:35:05 hm ah I can use do_install Aug 05 12:40:45 hi gnutoo Aug 05 12:44:17 RP__: so the new packages staging does 'private staging'? (was cleaning a package will remove whatever it put into staging.) Aug 05 12:45:01 ant_work: no, that is one of the reasons for staging packages in the first place Aug 05 12:45:50 what about the staged dependencies? Aug 05 12:46:45 i.e. you remove bluez but still dbus, hal, ... Aug 05 12:47:01 hm are we going to drop hal too= Aug 05 12:47:27 is there a way to manage a kind of "include file" sentence in the local.conf ??? Aug 05 12:48:01 ant_work: You can remove each in turn. You could script removal of a set of them Aug 05 12:48:19 something like emerge --depclean in Gentoo Aug 05 12:48:26 ceyusa: sure, just include xyz Aug 05 12:48:34 woglinde, hi Aug 05 12:48:38 RP__: great! :D Aug 05 12:48:42 ceyusa: you probably want require Aug 05 12:49:02 and what does it do? Aug 05 12:49:18 ceyusa: include is optional, require is fatal error if not found Aug 05 12:49:28 RP__: nice! Aug 05 12:49:45 RP__: thanks! Aug 05 12:51:12 re Aug 05 12:57:52 http://pastebin.com/FGPUcn08 Aug 05 12:58:47 dont use CC=gcc Aug 05 12:58:55 ok Aug 05 12:59:17 but there is same problem i dont find .ip Aug 05 12:59:19 k Aug 05 12:59:23 and dont use oe as your scm Aug 05 12:59:24 .ipk Aug 05 12:59:30 for ftttddf Aug 05 13:00:05 i dont understand Aug 05 13:00:24 i ont use oe? Aug 05 13:01:29 SRC_URI = "file://hello.c" Aug 05 13:01:57 yes it is not correct? Aug 05 13:02:15 only for testing Aug 05 13:02:40 we will not accept a recipe which has this in Aug 05 13:02:45 or smiliar Aug 05 13:02:49 to other files Aug 05 13:02:51 ok just i would like test a simple helloword Aug 05 13:04:43 so where we can finf .ipk Aug 05 13:05:06 deploy Aug 05 13:05:14 in your defined oetmp Aug 05 13:07:09 what's line 11 and 12? Aug 05 13:07:46 ag right Aug 05 13:07:49 thats wrong Aug 05 13:08:38 kamel: ^ Aug 05 13:08:47 yes Aug 05 13:09:37 i declar a compiler and flag ...... Aug 05 13:09:50 linebreak Aug 05 13:10:39 D="${WORKDIR}" is wrong too Aug 05 13:11:22 you can give me a simple exemple juste print helloword Aug 05 13:11:34 .bb example Aug 05 13:11:58 why you want a bad example? Aug 05 13:12:10 look at the 2k recipes we already have Aug 05 13:12:14 and read the manual Aug 05 13:12:23 even when it is a bit outdated Aug 05 13:12:35 or poky manual Aug 05 13:13:40 lik please Aug 05 13:13:44 link Aug 05 13:14:11 /topic please Aug 05 13:15:04 or find /oe/recipes -iname "*.c" Aug 05 13:15:10 and look at recipes Aug 05 13:15:13 ynezz ieehks Aug 05 13:15:14 no Aug 05 13:15:15 for example shasum Aug 05 13:15:25 shasum is maybee the only one Aug 05 13:15:46 dont use file://mytocompilestuff for other recipes Aug 05 13:15:49 thanks Aug 05 13:21:59 there is just a package called helloworld :-) Aug 05 13:22:22 effem that too Aug 05 13:24:04 03Henning Heinold  07org.openembedded.dev * r52d472c597 10openembedded.git/recipes/uclibc/ (uclibc-git/isnan.patch uclibc_git.bb): uclibc: try to fix isnan and bump PR Aug 05 13:24:05 03Henning Heinold  07org.openembedded.dev * racd7fb9deb 10openembedded.git/recipes/uclibc/ (3 files in 3 dirs): uclibc: use GENERIC_ARM, because we override the cpu flags anyway, bump PR Aug 05 13:24:06 03Henning Heinold  07org.openembedded.dev * rc352aa642b 10openembedded.git/classes/java-library.bbclass: Aug 05 13:24:06 java-library.bbclass: workaround for staging when BBCLASSEXTENDED is used Aug 05 13:24:06 * when BBCLASSEXTENDED is used, do_stage is not called for Aug 05 13:24:06 target stuff, so workaround it with the oejarinstall in do_install too Aug 05 14:03:46 is it possible to use eglibc in oe? Aug 05 14:05:13 multiple distros do, afaik Aug 05 14:05:17 so i'd say yes :) Aug 05 14:05:53 ty Aug 05 14:07:37 np Aug 05 14:10:14 will it be faster if i write/read data to internal nand flash or to external sd card? Aug 05 14:10:45 depends entirely on the details of your hardware. you'd need to do some measurements. Aug 05 14:11:03 any good tool for that? Aug 05 14:11:58 bonnie, though that might be overkill in your case. Aug 05 14:12:01 time + dd, perhaps Aug 05 14:12:31 umm i really need exact speed meassurements Aug 05 14:12:47 write speed/read speed Aug 05 14:12:56 well, sure. either of those approaches will give you that. Aug 05 14:13:52 dd also? didnt notice that Aug 05 14:13:53 thanks Aug 05 14:14:08 in combination with time, yeah Aug 05 14:14:27 dd itself just does the reading or writing; time tells you how long it took, and dividing the amount of data by the time will give you the speed. Aug 05 14:15:26 ahh ok thx Aug 05 14:20:22 OT, but http://mislav.uniqpath.com/2010/07/git-tips/ has a few I didn't know about until last night Aug 05 14:21:08 you can get individual read/write numbers via the kernel's taskstats -- /proc//io, but you'd have to check it just before the process exited, or use the netlink interface to get sent the info when it exits Aug 05 14:21:43 * kergoth was working on per-OE-task IO load measurement Aug 05 14:22:31 03Koen Kooi  07org.openembedded.dev * r658594deb5 10openembedded.git/recipes/eds/eds-dbus_git.bb: eds-dbus: add missing dep on libical Aug 05 14:32:04 hi kergoth Aug 05 14:32:09 hey Aug 05 14:32:10 re Aug 05 14:42:13 hi woglinde Aug 05 14:42:26 jo pb Aug 05 14:43:55 pb still drive around with the digger? Aug 05 14:45:02 no, digging finished now. it's going back to the hire company today in fact. Aug 05 14:45:30 poured the concrete into the trenches yesterday. Aug 05 14:59:31 kergoth, thx for the note on overwriting file from a package - the particular config file from pulse doesn't unfortunately use localfiles Aug 05 14:59:59 tharvey: if you change it to do so, you can use FILESPATHBASE to get it to look for it in a different location. Aug 05 15:02:14 * RP notes its TSC time Aug 05 15:03:23 has somebody build the sdcc-native recipe Aug 05 15:03:42 kergoth, right... I use and love the FILESPATH stuff but unfortunately the config file in this case is from the source tarball and not a local file Aug 05 15:03:58 it is giving me an error in configuration cannot find required program bison Aug 05 15:04:50 if sdcc needs bison shouldn't this be in its DEPENDS list Aug 05 15:05:31 tharvey: might be worth changing it to do so, would make it easier to provide machine/distro specific versions, if its something that would be useful to override -- not very familiar with the pulseaudio stuff Aug 05 15:05:32 *shrugs* Aug 05 15:06:01 Noor: if it's trying to run bison, it wants bison-native, not bison, but yes, it should be in its deps Aug 05 15:06:59 kergoth, agreed... there are actually several patches I would like to get into oe when I get past some hurdles with this project, I'll put that on the list Aug 05 15:11:35 yeah .... but sdcc-native recipe does not have bison-native in its DEPENDS .... i'll fix it Aug 05 15:11:47 and send a patch to list if appropriate Aug 05 15:11:57 sure Aug 05 15:12:04 :) Aug 05 15:16:21 after adding bison-native in dependency list sdcc-native's configure has started complaining about flex :O Aug 05 15:16:34 sure Aug 05 15:16:51 often both is used Aug 05 15:16:58 OK Aug 05 15:17:58 but you know now the trick Aug 05 15:18:05 re kergoth Aug 05 15:18:12 hey Aug 05 15:18:46 woglinde: which trick ... have I learnt something ? Aug 05 15:19:15 after adding bison-native in dependency list sdcc-native's configure has started complaining about flex :O Aug 05 15:19:25 u mean bison and flex are used side by side Aug 05 15:19:34 often Aug 05 15:23:44 why do I never remember how to do apt pinning? Aug 05 15:23:48 every single time i have to go googling Aug 05 15:24:11 *g* Aug 05 15:24:23 because you dont do it every day Aug 05 15:25:07 yeah.. guess its something I should shove into evernote or something.. but then, I always forget to search it for anything and end up googling anyway, so I guess thats pointless :) Aug 05 15:28:05 *g* Aug 05 15:28:19 depends how fast you find the right information Aug 05 15:28:27 not an outdated false blog entry Aug 05 15:28:29 that's true. some stuff takes some digging Aug 05 15:28:43 I made a tips.txt for some stuff Aug 05 15:28:46 of course, google's ability to limit shit to just the past year, or past 3 months, or whatever, is awesome Aug 05 15:28:48 a wiki might be better Aug 05 15:31:11 morning Aug 05 15:31:36 03Khem Raj  07org.openembedded.dev * rc55e51d03a 10openembedded.git/recipes/popt/ (5 files): Aug 05 15:31:36 popt: Use BBCLASSEXTEND instead of native recipes. Aug 05 15:31:36 * Use INC_PR Aug 05 15:31:36 Signed-off-by: Khem Raj Aug 05 15:31:36 Acked-by: Martin Jansa Aug 05 15:31:36 03Khem Raj  07org.openembedded.dev * re951cbf49d 10openembedded.git/recipes/upnp/gupnp_0.12.5.bb: Aug 05 15:31:37 gupnp_0.12.5.bb: Get rid of legacy staging Aug 05 15:31:38 Signed-off-by: Khem Raj Aug 05 15:31:38 Acked-by: Martin Jansa Aug 05 15:31:39 03Khem Raj  07org.openembedded.dev * r2f2a0aa2c2 10openembedded.git/recipes/upnp/gupnp_0.13.1.bb: Aug 05 15:31:40 gupnp_0.13.1.bb: Get rid of legacy staging Aug 05 15:31:40 Signed-off-by: Khem Raj Aug 05 15:31:41 Acked-by: Martin Jansa Aug 05 15:32:17 hrms Aug 05 15:32:21 03Khem Raj  07org.openembedded.dev * rc6c9b7e640 10openembedded.git/recipes/gtk+/gtk+_2.20.1.bb: Aug 05 15:32:21 gtk+_2.20.1.bb: Ask for ARM_INSTRUCTION_SET to be 'arm' Aug 05 15:32:21 * It does not compile in thumb mode with gcc 4.5 Aug 05 15:32:21 other version of gcc were not tested probably they Aug 05 15:32:22 many have problem too. For now always build using Aug 05 15:32:23 arm mode. Aug 05 15:32:23 Signed-off-by: Khem Raj Aug 05 15:32:24 03Khem Raj  07org.openembedded.dev * r725dbb78be 10openembedded.git/classes/ (autotools_stage.bbclass recipe_sanity.bbclass): Aug 05 15:32:24 autotools_stage.bbclass: Delete Aug 05 15:32:25 * This class should not be needed anymore Aug 05 15:32:25 Signed-off-by: Khem Raj Aug 05 15:32:26 Acked-by: Richard Purdie Aug 05 15:32:26 Acked-by: Martin Jansa Aug 05 15:32:26 03Khem Raj  07org.openembedded.dev * r9d612c4bb9 10openembedded.git/recipes/ (138 files in 83 dirs): Aug 05 15:32:29 recipes: inherit autotools instead of autotools_stage Aug 05 15:32:29 Signed-off-by: Khem Raj Aug 05 15:32:29 Acked-by: Martin Jansa Aug 05 15:32:29 Acked-by: Richard Purdie Aug 05 15:32:30 03Khem Raj  07org.openembedded.dev * r189ef00c64 10openembedded.git/recipes/openobex/openobex_1.5.bb: Aug 05 15:32:39 debian qemu forgot the thumb-got1 patch Aug 05 15:32:47 in the latest qemu package Aug 05 15:43:04 03Elena Grandi  07org.openembedded.dev * rc41336e849 10openembedded.git/recipes/xapian/ (3 files): xapian-core: add BBCLASSEXTEND = "native" and remove -native recipes. Aug 05 15:44:00 woot Aug 05 15:44:03 clean up Aug 05 16:12:59 hi stefan Aug 05 16:13:08 hi woglinde Aug 05 16:13:11 hi all Aug 05 16:42:18 I wish to build uImage alone using bitbake Aug 05 16:42:29 will bitbake -k uImage will work ? Aug 05 16:43:46 uImage isn't a recipe Aug 05 16:43:52 you probably want bitbake virtual/kernel Aug 05 16:47:34 i accidentally deleted deploy folder. how to get all the images from MLO to file system image ? Aug 05 16:47:42 iam using overo MACHINE Aug 05 16:49:39 I did bitbake -k omap3-desktop-image and i got only omap3-desktop-image-overo.tar.bz2 Aug 05 16:50:21 Ananth: then you use tar (as root) to extract the filesystem Aug 05 16:51:36 Ananth: oh wait, you mean you want a single image file formatted with a filesystem (e.g. ext2) Aug 05 16:55:25 Now i did bitbake -k omap3-console-image and bitbake -k omap3-desktop-image and got omap3-console-image-overo.tar.bz2,omap3-desktop-image-overo.tar.bz2 Aug 05 16:56:01 I wish to know how can i get MLO,u-boot binary and uImage binary in deploy folder Aug 05 16:57:04 03Khem Raj  07org.openembedded.dev * r8d9c7304ec 10openembedded.git/recipes/qemu/ (14 files in 2 dirs): Aug 05 16:57:04 qemu: Add 0.12.5 Aug 05 16:57:04 * Remove 0.12.4 in favor of it. Aug 05 16:57:04 Signed-off-by: Khem Raj Aug 05 17:05:41 he khem Aug 05 17:05:57 khem is the thumb1-got1 patch sill in qemu 12.5 Aug 05 17:06:01 Hello Aug 05 17:06:11 debian 12.5 removed it Aug 05 17:06:16 woglinde: yes Aug 05 17:06:34 woglinde: hmm I will dig into it Aug 05 17:06:35 okay Aug 05 17:06:42 mostly its needed Aug 05 17:06:55 but I can try my uclibc based images which are thumb Aug 05 17:07:06 otherwise I cannt excute a helper tool for openjdk Aug 05 17:07:18 stumbled over it today Aug 05 17:07:44 because I had ASUSSME_RPOVIDED for qemu-native in Aug 05 17:09:36 woglinde: you cannt if qemu has this patch Aug 05 17:09:43 or you need the patch Aug 05 17:09:49 for jdk helper Aug 05 17:13:39 I need teh patch Aug 05 17:13:44 overwise the cp15 error comes Aug 05 17:13:48 for armv6 Aug 05 17:14:32 yes Aug 05 17:14:35 ok Aug 05 17:15:14 bbl, bye Aug 05 17:15:28 03Roman I Khimov  07org.openembedded.dev * r917fe8395c 10openembedded.git/recipes/libpcap/libpcap_1.1.1.bb: Aug 05 17:15:28 libpcap: remove deprecated patch options Aug 05 17:15:28 Signed-off-by: Roman I Khimov Aug 05 17:15:29 03Roman I Khimov  07org.openembedded.dev * r445360e9dc 10openembedded.git/recipes/geda/geda-gattrib_1.4.3.bb: Aug 05 17:15:29 geda-gattrib: remove deprecated patch parameters Aug 05 17:15:29 Signed-off-by: Roman I Khimov Aug 05 17:15:32 03Roman I Khimov  07org.openembedded.dev * r52550b17b3 10openembedded.git/recipes/barnyard2/barnyard2_1.8.bb: Aug 05 17:15:32 barnyard2: remove deprecated patch options Aug 05 17:15:32 Another pick-from-local-branch error, sorry. Aug 05 17:15:32 Hopefully, last one,merged patch syntax changes there. Aug 05 17:15:32 Signed-off-by: Roman I Khimov Aug 05 17:15:32 03Roman I Khimov  07org.openembedded.dev * r0418175eb5 10openembedded.git/recipes/yaffs2/yaffs2-utils.inc: Aug 05 17:15:32 yaffs2-utils: remove deprecated patch parameters Aug 05 17:15:33 Signed-off-by: Roman I Khimov Aug 05 17:27:55 03Roman I Khimov  07org.openembedded.dev * r35655a4dcb 10openembedded.git/recipes/libtiff/ (tiff-3.7.2/configure.patch tiff_3.7.2.bb): Aug 05 17:27:55 tiff: remove version 3.7.2 Aug 05 17:27:55 * released March 2005, obsolete Aug 05 17:27:55 * not pinned by any public distro Aug 05 17:27:55 * uses legacy staging Aug 05 17:27:56 Signed-off-by: Roman I Khimov Aug 05 17:43:51 03Roman I Khimov  07org.openembedded.dev * rd4171d825e 10openembedded.git/recipes/cramfs/ (cramfs-native_1.1.bb cramfs_1.1.bb): Aug 05 17:43:52 cramfs: remove -native recipe, replace with BBCLASSEXTEND Aug 05 17:43:52 Removes another do_stage instance along the way. Aug 05 17:43:52 Signed-off-by: Roman I Khimov Aug 05 17:46:20 Hi. I am getting an error in bitbake saying unreferenced name 'os' (not sure of the exact error as I am not in front of the machine). Aug 05 17:46:39 03Roman I Khimov  07org.openembedded.dev * r71c6f94e7d 10openembedded.git/recipes/poptop/poptop_1.3.0.bb: Aug 05 17:46:40 poptop: remove version 1.3.0 Aug 05 17:46:40 Release dated 2005 and the recipe in unmaintained, 1.3.4 recipe is alsmost Aug 05 17:46:40 completely different now. So, I see no point in keeping it around. Aug 05 17:46:40 Signed-off-by: Roman I Khimov Aug 05 17:48:59 How can I fix it? Aug 05 18:00:41 guys something is fishy here... my OE build now already has 25GB Aug 05 18:00:48 fresh build Aug 05 18:01:19 what's fishy? Aug 05 18:01:26 is that normal? Aug 05 18:01:54 /dev/sda1 39814552 30278948 7513104 81% / Aug 05 18:02:01 7GB was os Aug 05 18:02:04 rest is ow Aug 05 18:02:06 *oe Aug 05 18:02:12 i mean what the heck is going on Aug 05 18:02:53 is that normal? Aug 05 18:03:29 yes Aug 05 18:03:38 My OE_HOME got about 30GB Aug 05 18:03:42 WTF Aug 05 18:03:54 for a final 60MB image? Aug 05 18:03:58 yes Aug 05 18:04:09 ummm Aug 05 18:04:21 actually, I got two images for two different machines Aug 05 18:04:50 well i have one Aug 05 18:04:57 thats just insane Aug 05 18:05:06 is the Locale stuff the proplem? Aug 05 18:05:30 what blows it up so much? Aug 05 18:09:04 seems it started tidying up ;P Aug 05 18:09:10 the build now starts shrinking Aug 05 18:09:13 very funny ;P Aug 05 18:09:33 from 35 to 15 GB Aug 05 18:11:31 03Koen Kooi  07org.openembedded.dev * r01177e8ed8 10openembedded.git/recipes/libglade/libglade.inc: libglade: kill legacy staging Aug 05 18:11:42 03Koen Kooi  07org.openembedded.dev * r07fb388f27 10openembedded.git/recipes/eds/eds-dbus_git.bb: eds-dbus: work around crosscompile badness Aug 05 18:12:42 03Koen Kooi  07org.openembedded.dev * r1acdd18379 10openembedded.git/recipes/libdisplaymigration/libdisplaymigration_0.99.bb: libdisplaymigration: kill legacy staging Aug 05 18:13:41 Seppoz, I don't think you realize just how much goes into a 60mb image. Aug 05 18:13:46 a kernel tree alone is hundreds of megs Aug 05 18:14:59 03Koen Kooi  07org.openembedded.dev * r883097a830 10openembedded.git/recipes/gpe-ownerinfo/ (gpe-ownerinfo.inc gpe-ownerinfo_0.28.bb): gpe-ownerinfo: kill legacy staging Aug 05 18:16:18 kergoth_: well indeed, the kernel was like 700mb here im just wondering what it does that it takes 35 gb Aug 05 18:16:27 does it build all the packages now? Aug 05 18:16:30 ? Aug 05 18:16:31 or just the depends? Aug 05 18:16:36 what do you mean? Aug 05 18:16:47 it's always built everything from source, and has always emitted highly granular packages for everything it builds Aug 05 18:16:53 i mean does bitbalke bake *all* the recepies now? Aug 05 18:16:56 uh Aug 05 18:16:59 or just the depends Aug 05 18:17:01 if you were baking all recipes Aug 05 18:17:04 itd take like 1.5 weeks. Aug 05 18:17:12 and I'm not kidding. Aug 05 18:17:12 hahahaha Aug 05 18:17:15 there are 7,000 recipes Aug 05 18:17:28 wondering how much hdds id need Aug 05 18:17:28 xD Aug 05 18:18:00 watch the tasks it runs, you can see what it has to build to get there, and for each of those it has to download the sources, unpack them, patch them, configure, compile, package, etc. Aug 05 18:26:19 i do ;) Aug 05 18:27:57 if you do INHERIT += "rm_work", it'll clean up as it goes, when the recipe is complete, it'll wipe its unpacked sources and all the rest Aug 05 18:28:26 you can also change bitbake's scheduling to complete recipes faster, so it gets to the rm_work sooner, but that will slow down your builds if you do Aug 05 18:48:02 whow, full house in here Aug 05 18:48:16 does anybody know of build problems concerning openssl? Aug 05 19:00:59 03Frans Meulenbroeks  07org.openembedded.dev * r1964b28eee 10openembedded.git/recipes/glib-2.0/ (32 files in 11 dirs): Aug 05 19:00:59 glib-2.0: removed several old versions Aug 05 19:00:59 There were about 26 recipes in this dir; removed all recipes older than 2 years that Aug 05 19:00:59 were not pinned (as per TSC guidelines). Aug 05 19:00:59 There are still 17 recipes left.... Aug 05 19:00:59 Signed-off-by: Frans Meulenbroeks Aug 05 19:05:16 03Frans Meulenbroeks  07org.openembedded.dev * rfcb22c7a2a 10openembedded.git/recipes/intltool/ (9 files): Aug 05 19:05:16 intltool: removed old versions Aug 05 19:05:16 removed older unpinned versions Aug 05 19:05:16 Signed-off-by: Frans Meulenbroeks Aug 05 19:06:45 ERROR: QA Issue with gpe-conf: /OE/build/tmp_angstrom_2008_1/work/armv4t-angstrom-linux-gnueabi/gpe-conf-0.2.8-r0/packages-split/gpe-conf/usr/share/applications/gpe-conf-cardinfo.desktop: error: value "Application;SystemSettings;GPE;" for key "Categories" in group "Desktop Entry" contains an unregistered value "SystemSettings"; values extending the format should start with "X-" Aug 05 19:06:51 is this a known issue? Aug 05 19:10:19 ERROR: QA Issue with gpe-terminal: /OE/build/tmp_angstrom_2008_1/work/armv4t-angstrom-linux-gnueabi/gpe-terminal-1.1-r2/packages-split/gpe-terminal/usr/share/applications/gpe-terminal.desktop: warning: boolean key "Terminal" in group "Desktop Entry" has value "0", which is deprecated: boolean values should be "false" or "true" Aug 05 19:16:02 03Frans Meulenbroeks  07org.openembedded.dev * r6d414e9726 10openembedded.git/recipes/wxwidgets/ (wxbase_2.6.4.bb wxwidgets_2.6.4.bb): Aug 05 19:16:02 wxwidgets/wxbase: remvoed 2.6.4 Aug 05 19:16:02 old, unpinned, using legacy staging Aug 05 19:16:02 Signed-off-by: Frans Meulenbroeks Aug 05 19:22:12 03Frans Meulenbroeks  07org.openembedded.dev * r6b7c493e24 10openembedded.git/recipes/zziplib/zziplib-native_0.12.83.bb: Aug 05 19:22:12 zziplib-native: removed Aug 05 19:22:12 no one has a dependency on this one and the leading non-native Aug 05 19:22:12 recipe is already at 0.13.49 Aug 05 19:22:12 Signed-off-by: Frans Meulenbroeks Aug 05 19:30:32 03Frans Meulenbroeks  07org.openembedded.dev * r21df95a924 10openembedded.git/recipes/lirc/ (11 files in 2 dirs): Aug 05 19:30:32 lirc: removed old recipes Aug 05 19:30:32 the lirc recipes still used legacy staging (apart from the latest version) Aug 05 19:30:32 nobody pinned these so removed them (and the associated Aug 05 19:30:32 lirc-modules recipes) Aug 05 19:30:32 Signed-off-by: Frans Meulenbroeks Aug 05 19:31:42 03jkridner@beagleboard.org  07org.openembedded.dev * rb953e67533 10openembedded.git/recipes/angstrom/ (3 files in 2 dirs): Aug 05 19:31:43 angstrom-uboot-scripts: fixed tests for revision Aug 05 19:31:43 Tests in u-boot scripts require the 'test' function. Aug 05 19:31:43 Signed-off-by: Jason Kridner Aug 05 19:31:43 Signed-off-by: Koen Kooi Aug 05 19:31:44 03jkridner@beagleboard.org  07org.openembedded.dev * rf2636a6278 10openembedded.git/recipes/u-boot/ (3 files in 2 dirs): (log message trimmed) Aug 05 19:31:44 u-boot: beagleboard: default 3MP camera support Aug 05 19:31:44 This turns on the pull-ups so that the I2C will work by default as well Aug 05 19:31:45 as sets the default camera type. Currently, the kernel cannot probe for Aug 05 19:31:46 I2C devices, but that should be the long term fix. We put the change Aug 05 19:31:46 here into the default environment, instead of the boot.scr, such that Aug 05 19:31:47 there won't be a need to update your boot.scr in the future for the Aug 05 19:35:40 in u-bppt, how do i load images that are bigger than my ram to my nand? Aug 05 19:35:44 *uboot Aug 05 19:37:45 Seppoz: boot into linux (e.g. over tftp), erase nand (or your partition whereever you want the image to go, put the image on e.g. usb disk and after booting do a cp of the image to the partition Aug 05 19:38:17 but i cant use the linux from nand to do that right? Aug 05 19:38:53 if you have a kernel that is larger than your ram, no Aug 05 19:39:13 then you have a problem (and probably a kernel that is way too big) Aug 05 19:39:20 but its not possible from uboot directly is it? Aug 05 19:39:46 not really, would be very tricky Aug 05 19:40:02 read file partially (which means reading sectors) and copy in steps Aug 05 19:40:03 bad idea Aug 05 19:40:18 well i *think* Aug 05 19:40:22 i did that earlyer Aug 05 19:40:44 well if the file is not contiguously stored you have a problem Aug 05 19:40:49 just cant remeber how anymore Aug 05 19:40:49 xD Aug 05 19:40:57 lol yea indeed Aug 05 19:41:22 Q: what is gpe exactly? i dont need that for x right? Aug 05 19:41:25 x11-gpe-image Aug 05 19:41:37 its enought to have x11-image right? Aug 05 19:41:45 no idea Aug 05 19:41:47 gpi is just some gui Aug 05 19:41:49 ok Aug 05 19:41:56 i iwll know in a sec i guess Aug 05 19:45:55 Angstrom 2010.7-test-20100805 mini2440 Aug 05 19:45:58 nice nice Aug 05 19:58:44 03Frans Meulenbroeks  07org.openembedded.dev * r4518dba157 10openembedded.git/recipes/gphoto2/ (6 files): Aug 05 19:58:45 gphoto2: removed 2.1.x > 4 years old; not pinned Aug 05 19:58:45 Signed-off-by: Frans Meulenbroeks Aug 05 19:59:39 how do i add e.g xf86-input-tslib to my image? Aug 05 20:01:41 03Koen Kooi  07org.openembedded.dev * r7d4247c9d8 10openembedded.git/recipes/u-boot/u-boot_git.bb: u-boot: fix beagle patch names Aug 05 20:08:40 03Jason Kridner  07org.openembedded.dev * r7058b8b5dd 10openembedded.git/ (3 files in 3 dirs): Aug 05 20:08:41 linux-omap-psp_2.6.32: Move camera sensor reset Aug 05 20:08:41 Camera reset needed to be done after applying power. Aug 05 20:08:41 Signed-off-by: Jason Kridner Aug 05 20:08:41 Signed-off-by: Koen Kooi Aug 05 20:12:11 umm, ERROR: QA Issue with i2c: No GNU_HASH in the elf binary: '/OE/build/tmp_angstrom_2008_1/work/armv4t-angstrom-linux-gnueabi/i2c-1.0-r2/packages-split/i2c/usr/bin/i2c' ... whats the common wy to fix this? already tryed bitbake -c clean i2c Aug 05 20:35:42 re Aug 05 20:38:21 wb Aug 05 20:38:57 woglinde: whats the best way to resolve hashing issues? Aug 05 20:39:13 add our LDFLAGS Aug 05 20:39:23 what u mean? Aug 05 20:39:29 most of the time stupid software ignores external LDFLAGS Aug 05 20:39:35 ERROR: QA Issue with i2c: No GNU_HASH in the elf binary: '/OE/build/tmp_angstrom_2008_1/work/armv4t-angstrom-linux-gnueabi/i2c-1.0-r2/packages-split/i2c/usr/bin/i2c' Aug 05 20:39:46 yes Aug 05 20:39:49 i did try to clean it but sems i have to change its recepie maybe? Aug 05 20:40:01 look into git log Aug 05 20:40:09 you will find examples how to solve it Aug 05 20:40:17 ahh u think it had a proplem pulling them? Aug 05 20:40:20 ahh ok! Aug 05 20:40:24 will have a look thx Aug 05 20:40:53 no Aug 05 20:41:03 I meant look at older commits Aug 05 20:41:11 which fixes other packages Aug 05 20:41:36 for instance c86e303133eb530ea12a4e57cec5d3af62d141a0 Aug 05 20:41:38 ok! Aug 05 20:49:14 woglinde: whats the common way to get gtk+ into the image? Aug 05 20:49:24 just DEPEND=gtk+ Aug 05 20:49:40 DEPENDS = "task-base xf86-input-tslib gtk+" Aug 05 20:49:43 look at the .bb undeer images Aug 05 20:49:51 ty Aug 05 20:50:28 IMAGE_INSTALL += "task-base-extended Aug 05 20:50:42 u sure i need extended? Aug 05 20:50:53 according to the demo image of gtk all it needs is boot Aug 05 20:51:20 *sigh* Aug 05 20:51:25 that was only an example Aug 05 20:51:31 ahh im sorry :( Aug 05 20:51:37 either use one app which depends on gtk Aug 05 20:51:45 or put it directly there Aug 05 20:52:13 i *think* i did the 2end one hehe trying to build no3w Aug 05 20:52:15 *now Aug 05 20:52:17 see what happens Aug 05 20:52:19 there == your custom image Aug 05 20:53:33 what ya mean? Aug 05 20:54:28 recipes/images/native-sdk-image.bb Aug 05 20:54:35 you can make it similar Aug 05 20:54:43 with stuff you wantesd Aug 05 20:54:47 want Aug 05 20:55:46 yes i did that before, i love it Aug 05 20:56:04 i can now bitbake woglinde-cake Aug 05 20:56:06 lol Aug 05 20:56:11 hihi Aug 05 20:57:06 bitbake universe Aug 05 20:57:07 ;P Aug 05 21:04:01 jo ant Aug 05 21:04:55 hello Aug 05 21:46:01 huh, I read the woglinde-cake above as woglinde-cmake :) Aug 05 22:09:14 khem: ping? Aug 05 23:10:39 where is the cross compiler located after build? Aug 05 23:12:10 bitbake -e | grep \^STAGING_DIR_HOST Aug 05 23:14:04 no output Aug 05 23:14:15 do i have to bake anything special to get it? Aug 05 23:14:22 i mean it bust be there somehow Aug 05 23:14:27 but where ;P Aug 05 23:15:23 uh, did you think it built the stuff with magic? Aug 05 23:15:25 no shit its there Aug 05 23:15:46 it's down in tmp/sysroots Aug 05 23:16:15 STAGING_DIR_HOST="/OE/build/tmp_angstrom_2008_1/sysroots/armv4t-angstrom-linux-gnueabi" Aug 05 23:16:33 also, you must have copied and pasted wrong, that command does result in output Aug 05 23:16:43 yes, there you go Aug 05 23:17:03 well Aug 05 23:17:07 i was there befor Aug 05 23:17:11 but the bin is empty Aug 05 23:17:13 ermm, wrong arch. get STAGING_DIR_NATIVE instead Aug 05 23:18:29 sec Aug 05 23:19:13 STAGING_DIR_NATIVE="/OE/build/tmp_angstrom_2008_1/sysroots/i686-linux" Aug 05 23:19:36 you don't need to paste it here Aug 05 23:19:39 i know what its going to look like Aug 05 23:19:42 its for your benefit, not mine Aug 05 23:20:18 OE/build/tmp_angstrom_2008_1/sysroots/i686-linux/usr/armv4t/bin <- is that what i want? Aug 05 23:21:53 seems so THX Aug 05 23:26:08 yep, exactly Aug 05 23:26:09 np Aug 05 23:28:14 works thx Aug 05 23:28:19 very much hidden ;) Aug 05 23:28:35 is this all statically linked or relative? Aug 05 23:28:50 could i move it to /opt/cross? Aug 05 23:28:52 nvm Aug 05 23:28:59 ill just let it where it is :) Aug 05 23:38:07 is there a list of distros that i can build using qt? Aug 05 23:38:10 *oe Aug 05 23:38:58 Seppoz, if you want to build a standalone toolchain, bitbake meta-toolchain and take the tarball it gives you Aug 05 23:39:17 really? now THAT sounds interessting Aug 05 23:39:23 going to try that, thanks! Aug 05 23:39:31 ls openembedded/conf/distro/*.conf :P Aug 05 23:39:55 ehehe Aug 05 23:41:11 so guy said he built gento with OE Aug 05 23:41:15 is that even possible? Aug 05 23:41:23 *some guy Aug 05 23:42:01 lol whats iphone-compat? Aug 05 23:44:12 03Andrea Adami  07org.openembedded.dev * r34f2f07769 10openembedded.git/recipes/klibc/ (klibc-common.inc klibc.inc): Aug 05 23:44:12 Revert "klibc: get rid of legacy staging" Aug 05 23:44:12 * committed broken version with evil install Aug 05 23:44:12 * WIP Aug 05 23:44:12 This reverts commit b32ce2a63973fdf72d1ffe94b8418ec214cb2b20. Aug 05 23:46:54 *evil* shug Aug 05 23:47:09 stop making evil stuff mr Adami! Aug 05 23:47:22 :) Aug 05 23:49:00 Seppoz: I agree! :) Aug 05 23:49:07 :P Aug 05 23:49:31 you should be able to use Aug 05 23:49:32 QWSDisplay::setTransformation(QTransformedScreen::Rot90, 0); Aug 05 23:49:34 in your app Aug 05 23:49:48 or Aug 05 23:49:53 might even be better! Aug 05 23:49:54 /myapp -qws -display transformed:Rot90 Aug 05 23:50:00 -qt-gfx-transformed Aug 05 23:50:07 u have to recompile using this flag tho Aug 05 23:50:15 grrrrrrrrr Aug 05 23:50:17 sooooory! Aug 05 23:50:44 im very sorry Aug 05 23:50:50 03Denys Dmytriyenko  07org.openembedded.dev * r7c376183bb 10openembedded.git/recipes/gstreamer/ (6 files): Aug 05 23:50:51 gst-ffmpeg: use immediate var expansion of TOOLCHAIN_PATH for PATH_prepend Aug 05 23:50:51 Fixes the recursive var expansion when TOOLCHAIN_PATH is derived from PATH. Aug 05 23:50:51 Signed-off-by: Denys Dmytriyenko Aug 06 01:21:21 http://github.com/kergoth/bitbake/compare/master...vercmp - thoughts? **** ENDING LOGGING AT Fri Aug 06 02:59:58 2010