**** BEGIN LOGGING AT Fri Mar 05 02:59:58 2010 Mar 05 05:43:08 hi all Mar 05 05:43:24 how to build clutter 0.8 with egl native on angstrom/OE Mar 05 07:19:24 03Michael 'Mickey' Lauer  07org.openembedded.dev * r1b059c5061 10openembedded.git/recipes/freesmartphone/msmcommd_git.bb: msmcommd: catch up with upstream changes Mar 05 07:25:45 03Philip Balister  07org.openembedded.dev * r2cdfa33e57 10openembedded.git/recipes/omniorb/ (3 files in 2 dirs): Mar 05 07:25:45 omniorb : Update to version 4.1.4 Mar 05 07:25:45 * Recipe from majo.huber@googlemail.com Mar 05 07:25:45 * Recipe needs updating to BBCLASSEXTEND, but I wanted to get it in first. Mar 05 07:29:53 omni! Mar 05 08:09:13 good morning Mar 05 08:43:01 morning Mar 05 08:50:39 03Koen Kooi  07org.openembedded.dev * r28014bdfa3 10openembedded.git/recipes/gnome/gnome-python-extras_2.25.3.bb: gnome-python-extras: adjust to pygtk -> pygobject codegen move Mar 05 09:00:50 03Koen Kooi  07org.openembedded.dev * rdbb73cac5d 10openembedded.git/recipes/gnome/gconf-editor_2.28.0.bb: gconf-editor: add 2.28.0 Mar 05 09:00:51 03Steve Sakoman  07org.openembedded.dev * r34c20e2d53 10openembedded.git/recipes/powervr-drivers/ (libgles-omap3.inc libgles-omap3_3.01.00.02.bb): libgles-omap3: fix for unwanted *.desktop file, shorten app names in menus Mar 05 09:05:40 03Steve Sakoman  07org.openembedded.dev * rd9a1554931 10openembedded.git/recipes/libusb/ (files/libusb-shortok.patch libusb1_1.0.4.bb): Mar 05 09:05:40 libusb: add patch to fix "BOGUS urb flags" issue Mar 05 09:05:40 issue occurs when using gphoto2 to download photos from digital cameras Mar 05 09:29:30 03Sebastian Spaeth  07org.openembedded.dev * r41789577e9 10openembedded.git/ (2 files in 2 dirs): Mar 05 09:29:30 libgsm0710mux: bump & move SRCREV from sane-srcrev-fso.inc to recipe Mar 05 09:29:30 This version introduces the config option device_ack_timeout (by default 5), which is needed to fix a race condition on Freerunner phones. Mar 05 09:29:30 Signed-off-by: Sebastian Spaeth Mar 05 09:29:32 03Sebastian Spaeth  07org.openembedded.dev * r8f601461fa 10openembedded.git/recipes/freesmartphone/ (3 files in 3 dirs): Mar 05 09:29:32 fso-abyss: Update om-gta0X config files to include device_ack_timeout = 2 Mar 05 09:29:32 The default timeout of 5 seems too long and it failed getting a channel. Let's try 2 as this worked on my Freerunner. Mar 05 09:29:32 Signed-off-by: Sebastian Spaeth Mar 05 09:32:42 03Martin Jansa  07org.openembedded.dev * rbc30aaf5b8 10openembedded.git/recipes/tasks/task-shr.bb: Mar 05 09:32:42 task-shr: fix typo and add task-cli-tools, task-cli-tools-debug Mar 05 09:32:42 * remove what is already included in those tasks Mar 05 09:32:42 Signed-off-by: Martin Jansa Mar 05 09:32:43 03Martin Jansa  07org.openembedded.dev * rce7359e12e 10openembedded.git/recipes/tasks/task-shr-feed.bb: Mar 05 09:32:43 task-shr-feed: add estardict Mar 05 09:32:44 Signed-off-by: Martin Jansa Mar 05 09:32:44 03Vaudano Luca  07org.openembedded.dev * r12659be7a7 10openembedded.git/recipes/estardict/estardict_0.1.bb: Mar 05 09:32:45 estardict: new recipe for elementary based offline dictionary reader Mar 05 09:32:45 Signed-off-by: Martin Jansa Mar 05 09:45:48 03Koen Kooi  07org.openembedded.dev * rf3bf0f5e55 10openembedded.git/recipes/gtk+/ (gtk+-2.18.6/no-demos.patch gtk+_2.18.6.bb): gtk+: reinstate no-demos patch for gtk+-native, it fails to build without it Mar 05 09:49:16 03Holger Hans Peter Freyther  07org.openembedded.dev * r5a79ac8040 10openembedded.git/ (11 files in 5 dirs): (log message trimmed) Mar 05 09:49:17 gpdf: Remove all gpdf recipes from the repository Mar 05 09:49:17 * gpdf contains an embedded copy of xpdf which has had several Mar 05 09:49:17 security issues in the past. Mar 05 09:49:17 * gpdf is not maintained for some years and the xpdf copy is not Mar 05 09:49:17 updated. Mar 05 09:49:17 * Remove the gpdf checksums and PREFERRED_VERSIONS Mar 05 09:53:07 good morning Mar 05 09:58:11 hi florian, welcome back ;-) Mar 05 09:59:11 hi mckoan Mar 05 09:59:46 mckoan: Back home again? Mar 05 10:00:17 florian: yes, I came back yesterday Mar 05 10:01:13 * florian early in the morning Mar 05 10:16:45 03Sebastian Spaeth  07org.openembedded.dev * r86a928fcda 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh+git://git@git.openembedded.net/openembedded into org.openembedded.dev Mar 05 10:16:47 03Sebastian Spaeth  07org.openembedded.dev * r97d0c4acea 10openembedded.git/recipes/connman/mokonnect_svn.bb: Mar 05 10:16:47 mokonnect: rdepend on connman-plugin-udhcp Mar 05 10:16:47 Without dhcp support, it is not able to retrieve an IP address, which is not Mar 05 10:16:47 what we want. There is an udhcp and a dhclient plugin though, making udhcp the Mar 05 10:16:48 default for now, not sure if we need to provide something more flexible in the Mar 05 10:16:48 future. Mar 05 10:16:49 Signed-off-by: Sebastian Spaeth Mar 05 10:20:21 03Sebastian Spaeth  07org.openembedded.dev * r40a4545008 10openembedded.git/recipes/connman/mokonnect_svn.bb: Mar 05 10:20:21 mokonnect: also bump PR due to previous RDEPEND change Mar 05 10:20:21 Signed-off-by: Sebastian Spaeth Mar 05 10:27:40 03Sebastian Spaeth  07org.openembedded.dev * r5b51c0a7b8 10openembedded.git/recipes/ffalarms/ffalarms_git.bb: Mar 05 10:27:40 ffalarms: directly RDEPEND on alsa-scenarii-shr Mar 05 10:27:40 virtual/alsa-scenarios is usually (and correctly) provided by fsodeviced Mar 05 10:27:40 directly. However, the current ffalarms code is explicitely looking for the Mar 05 10:27:40 location of alsa-scenarii-shr, so it makes sense to just depend on that until Mar 05 10:27:41 the source code has been updated. Mar 05 10:27:42 Signed-off-by: Sebastian Spaeth Mar 05 10:35:58 03Koen Kooi  07org.openembedded.dev * rb3045d2a28 10openembedded.git/recipes/shadow/shadow.inc: shadow: use u-a on binaries that conflict with util-linux-ng Mar 05 10:35:59 03Koen Kooi  07org.openembedded.dev * rbbd436a98d 10openembedded.git/recipes/util-linux-ng/util-linux-ng.inc: util-linux-ng: use u-a on binaries that conflict with shadow Mar 05 11:01:24 03Koen Kooi  07org.openembedded.dev * r8495e5b2e2 10openembedded.git/conf/distro/include/sane-srcrevs.inc: opkg*: bump SRCREV to get fixes for conffile reporting and support mixed arch feeds Mar 05 11:09:16 03Koen Kooi  07org.openembedded.dev * r8cd06fd4e9 10openembedded.git/recipes/gnome/eog_2.28.2.bb: eog: fix build when gnome-python gets detected Mar 05 11:39:06 Hi, all! How could I flash .ubi image, produced by OE using u-boot on SheevaPlug? I get error UBIFS error (pid 1): ubifs_get_sb: cannot open "ubi0:rootfs", error -19 and so can't mount rootfs :( Mar 05 11:40:14 slapin_nb: how do you mount it? Mar 05 11:40:51 slapin_nb: I guess that you're using wrong rootfs name (check ubinize.cfg) Mar 05 11:41:42 slapin_nb: maybe something like ubi0:sheevaplug-rootfs is used Mar 05 11:57:27 hi mickey|office Mar 05 12:05:54 JaMa, bootargs=console=ttyS0,115200 mtdparts=orion_nand:0x400000@0x100000(uImage),0x1fb00000@0x500000(rootfs) ubi.mtd=1,512 root=ubi0:rootfs rootfstype=ubifs Mar 05 12:09:35 morning Mar 05 12:10:02 slapin_nb: I am using ubi.mtd=2 without mtdparts, but I am using only rootfs mtd partition for ubifs Mar 05 12:10:30 slapin_nb: then you should change root= or prepare ubi partition with name "rootfs" and write .ubifs there Mar 05 12:14:54 hrw, JaMa: Thanks a lot! Volume name change fixed this problem. Is there some configuration variable to set to set UBI volume name? Mar 05 12:15:32 no idea - I ubiformatted sheeva on sheeva Mar 05 12:16:22 slapin_nb: variable in OE? Mar 05 12:16:49 JaMa, yeah. Mar 05 12:17:04 slapin_nb: UBI_VOLNAME Mar 05 12:17:16 slapin_nb: check grep ubinize conf/bitbake.conf Mar 05 12:17:33 JaMa, thanks a lot! Mar 05 12:17:47 yw Mar 05 12:34:18 anyone experience with oe and lttng (or systemtap) ? Mar 05 12:36:49 eFfeM: RP is probably the one you need to nag about lttng Mar 05 12:37:27 ok, tnx, actually got my kernel as far that it patches so guess I made some progress :-) Mar 05 12:41:42 zecke: good morning. aren't you supposed to be on an aeroplane? Mar 05 12:43:04 pb_: not yet.. 16:30 I will leave for Heathrow, then at 21h I will leave for bangkok, and then I will go to TPE. *yeah* Mar 05 12:43:47 ah, quite an epic journey Mar 05 12:44:03 yeah, it was the cheapest one way route from iceland we could find. Mar 05 12:50:17 * zecke breaks sudo before going to the airport Mar 05 12:53:16 hehe Mar 05 12:54:42 03Holger Hans Peter Freyther  07org.openembedded.dev * r22277253ee 10openembedded.git/ (conf/checksums.ini recipes/sudo/sudo_1.7.2p1.bb): Mar 05 12:54:42 sudo: Upgrade from 1.7.2p1.bb to 1.7.2p4 Mar 05 12:54:42 This addresses CVE-2010-0426 Mar 05 12:54:42 http://www.vuxml.org/freebsd/018a84d0-2548-11df-b4a3-00e0815b8da8.html Mar 05 12:59:17 pb_: which version of gdk-pixbuf-csource do you have? my staging/x86_64-linux/usr/bin/gdk-pixbuf-csource -v -> 2.12.11 says that .png is uknown format (while building gtk+ demos) Mar 05 13:00:06 anyone feels like helping with upgraing GNUtls and OpenSSL? Mar 05 13:00:24 or dnsmasq? Mar 05 13:00:42 JaMa: the one made by gtk+-native seems to be broken Mar 05 13:00:57 JaMa: the one made from gdk-pixbus-csource-native seems to work Mar 05 13:02:11 XorA: thanks, will try to rebuild gdk-pixbus-csource-native Mar 05 13:02:31 zecke: gnutls 2.8.x or newer? Mar 05 13:02:36 * XorA is not sure why and I havent had time to dig deep enough Mar 05 13:03:09 hrw: 2.8.3 would be a start Mar 05 13:03:51 hrw: http://portaudit.FreeBSD.org/856a6f84-8b30-11de-8062-00e0815b8da8.html is fixed by 2.8.3 Mar 05 13:04:12 zecke: ok, can update gnutls but what about its patches? I do not have idea what they do so prefer to rather drop them then update (unless will be small amount of work) Mar 05 13:04:37 XorA: hmm I sent e-mail about problem with atk-1.0 includes when gdk-pixbus-csource-native overwrittes some staged files from gtk+-native, but cannot find it now Mar 05 13:05:28 JaMa: best solution would be fix the gtk+-native one Mar 05 13:05:42 JaMa: but I couldnt understand the bug Mar 05 13:07:27 XorA: bug 5405? Mar 05 13:07:42 !oebug 5405 Mar 05 13:08:33 that might explain it Mar 05 13:08:54 hm, odd cairo configure complains it cannot find pixman, yet pixman is in the DEPENDS lib and exists in work Mar 05 13:08:55 make gtk+-native PROVIDE gdk-pixbuf-csource-native and try again? Mar 05 13:09:58 XorA: but then this issue with gdk-pixbuf-csource really needs to be resolved in gtk+-native Mar 05 13:10:35 JaMa: gdk-pixbuf-csource-native was only supposed to be a stopgap until gtk+-native worked Mar 05 13:10:56 JaMa: but your the guy on the shop floor at the moment so Ill trust your opinion more Mar 05 13:11:34 "shop floor"? :) sorry I don't understand.. Mar 05 13:11:51 JaMa: the guy doing the work (as in factory floor) Mar 05 13:11:57 and I don't know about gtk/gdk Mar 05 13:12:54 it will be a few weeks before I get time to poke at it again Mar 05 13:43:18 03Koen Kooi  07org.openembedded.dev * r115947e4ab 10openembedded.git/recipes/sylpheed/claws-mail_3.6.1.bb: claws-mail: convert to new-style staging Mar 05 13:50:15 03Koen Kooi  07org.openembedded.dev * r61b548b995 10openembedded.git/recipes/hal/hal.inc: hal: convert to new style staging Mar 05 14:19:27 there are serveral nor flash partitioin in embedded linux. Such as :one is cramfs(ro) ,another is jffs2(rw). and i try to use oe to build system and manage the firmware of the system. So can the oe generate two image files: rootfs(cramfs) and data(jffs2)???? Mar 05 14:21:31 IamWhat, yes, you can generate your own image. Look at u-boot recipes and do analogously. Mar 05 14:23:21 IamWhat, also, nothing prevents you to make .bb which depends on several images. Mar 05 14:30:10 slapin_nb: thank you,i will see,but i have a question: does oe put all package into a directory(rootfs?) in a distro linux...So can oe output some package into a dirtectory(rootfs ) and put the others package into another dirtectory(data)?? Mar 05 14:31:26 IamWhat: each recipe emits binary packages which are then used by image recipes to produce images. you can make as many image recipes as you like, or take the packages from the build and construct the images manually, if you so desired Mar 05 14:51:47 kergoth: i don't like construct the images manually.because if the embedded linux don't use any package manager(such as ipkg or deb?) and some package upgrade ,i should rebuild the package and construct the images manually...It seems to mistake something....So i try to make it build firmware automaticly. But i don't know whether i can use oe do these? Mar 05 14:52:04 then you aren't reading. Mar 05 14:52:17 you've been told multiple times that you can build whatever images you want to build Mar 05 14:52:32 look at the recipes in openembedded/recipes/images/ for examples Mar 05 14:55:54 kergoth: i am a novice,thank you for your helps. Mar 05 14:57:19 Hi all, vala.bbclass contains a do_stage_append() function and bb complains about legacy staging. Mar 05 14:57:39 not knowing a lot about "new world staging", what is the best way to fix that? Mar 05 14:57:54 spaetz: do what do_stage_append does in do_install_append Mar 05 14:58:04 spaetz: do_install is now used for staging Mar 05 14:58:19 ok, will just do that. Thanks a lot. Mar 05 14:58:24 spaetz: means that SDK/native/staging should all be identicle Mar 05 14:58:35 I see Mar 05 14:58:37 makes sense Mar 05 14:58:47 and save a whole shitload of build time Mar 05 14:59:07 03Jeremy Lainé  07org.openembedded.dev * r51e8173bf6 10openembedded.git/recipes/qt4/qt4-tools-native_4.5.2.bb: qt4-tools-native_4.5.2.bb: use Qt's builtin freetype, otherwise build fails Mar 05 14:59:44 and use DATDIR not STAGING_DATADIR of course Mar 05 15:00:20 XorA: what about do_stage() where is ie smpeg-config modified to return staging paths? Mar 05 15:00:41 looking at the class I suspect vala-native may need to tweaks/fixes Mar 05 15:00:53 JaMa: you can add hooks for that sort of thing, if they're really necessary Mar 05 15:01:17 XorA: I'm just looking to smpeg_svn.bb where is do_stage I bit more difficult to convert Mar 05 15:01:19 god, something still uses smpeg Mar 05 15:01:30 XorA: python-pygame IIRC Mar 05 15:02:18 kergoth: any example? or it wasn't really needed yet? Mar 05 15:02:33 kergoth: or I can probably grep through poky repo i guess Mar 05 15:02:55 always check poky incase RP fixed it already :-) Mar 05 15:03:16 * JaMa checked a while ago and smpeg recipe wasn't found there at all Mar 05 15:04:02 SYSROOT_PREPROCESS_FUNCS is what you'll want to look for Mar 05 15:04:22 space separated list of functions to be executed Mar 05 15:04:24 sigh: install -d ${STAGING_DATADIR_NATIVE}/vala/vapi Mar 05 15:04:39 what would I use rather than STAGING_DATADIR_NATIVE? Mar 05 15:04:40 spaetz: ${D}${datadir} Mar 05 15:04:46 k, thanks a bunch Mar 05 15:05:09 in install everything goes into ${D}, and you use the target dirs (libdir, bindir, datadir, etc) Mar 05 15:06:05 kergoth: thanks Mar 05 15:06:16 np Mar 05 15:06:37 testing now. If it works fine, I'll mail a patch to the list Mar 05 15:07:00 * spaetz won't touch classes/* without ACKs Mar 05 15:07:11 :-D wise plan Mar 05 15:07:21 indeed, good rule of thumb Mar 05 15:07:50 mickeyl is vala gos Mar 05 15:07:50 ups god Mar 05 15:07:58 so I hink you need his ack anyway Mar 05 15:08:03 JaMa: from a quick git grep (well, somewhat quick, anyway :) cups, icu, and mysql have preprocess funcs set up in the oe repo. Mar 05 15:08:09 yep, proab Mar 05 15:08:17 oops Mar 05 15:09:49 ah neat I had missed that bit of documentation Mar 05 15:11:35 ~curse devs which store libtool macros in own ones Mar 05 15:11:36 May the fleas of a thousand camels infest your most sensitive regions, devs which store libtool macros in own ones ! Mar 05 15:27:20 morning kergoth Mar 05 15:41:36 NOTE: Running task 17 of 17 (ID: 2, /home/hrw/devel/oe/openembedded/recipes/gnutls/gnutls_2.8.5.bb, do_build) Mar 05 15:42:06 but newer gnutls do not have m4 macros so other apps can break Mar 05 15:59:20 Do OE-generated images have a default user (besides root)? Mar 05 15:59:35 no Mar 05 16:00:25 IOW, all processes run as root in a default image? Mar 05 16:00:45 no Mar 05 16:00:51 daemons can run as own users Mar 05 16:01:18 03Marcin Juszkiewicz  07org.openembedded.dev * rc7bf984e0d 10openembedded.git/recipes/gnutls/ (7 files in 2 dirs): Mar 05 16:01:18 gnutls: added 2.8.5, adapted to new staging Mar 05 16:01:18 Signed-off-by: Marcin Juszkiewicz Mar 05 16:01:22 zecke: pull and test? Mar 05 16:02:30 hrw: have you ever got kernel >=2.6.30 to compile for avr32? Mar 05 16:03:40 XorA: did not tried recently Mar 05 16:04:01 hrw: I dont care about recently, I just want to know if it has ever been done Mar 05 16:04:03 16:51 hrw@home:angstrom$ MACHINE=atngw100 bitbake -b openembedded/recipes/linux/linux_2.6.33.bb Mar 05 16:04:06 checking Mar 05 16:04:18 ops, toolchain first.. did tmp clean Mar 05 16:04:21 pb_: hey. interesting idea with staging using the split packages.. sort of a dogfooding :) Mar 05 16:04:28 * kergoth thinks Mar 05 16:04:29 * XorA has tried 64/32bit fedora/debian .dev/stable and the compile always fails Mar 05 16:04:30 kergoth: yeah, exactly Mar 05 16:05:42 kergoth: btw, can you think of any reason why setting STAGING_DIR_task-populate_staging = "${WORKDIR}/something" wouldn't work? seems like that would be an easy way to deal with all those recipes that still have legacy do_stage methods. Mar 05 16:05:56 I tried it just now and the override didn't seem to have any effect but I can't quite figure out why. Mar 05 16:07:53 hmm, not offhand.. as far as I know that should work Mar 05 16:09:49 Is there any way to have bitbake keep the modules and kernels, like it does the images it generates? Mar 05 16:10:13 hrw: are those users added via postinst or while doing do_rootfs? Mar 05 16:10:39 james_l: I don't think the modules and kernels are ever discarded. Mar 05 16:10:53 james_l: what do you mean? Mar 05 16:11:03 03Koen Kooi  07org.openembedded.dev * r41fa478402 10openembedded.git/recipes/gstreamer/ (2 files in 2 dirs): gst-plugins-base: add support for rowstride conversion and NEON audioconvert functions Mar 05 16:12:10 Laibsch: that directly contradicts what I'm seeing. I set say 2.6.30 as default kernel for tosa, rebuild, and then 2.6.29 is removed from deploy/images/.../tosa/ along with the modules This has happened for every version I've tried. Mar 05 16:12:40 james_l: are you talking about the loose files, or the .ipk packages? Mar 05 16:12:50 if the latter, they have probably just gone to the morgue, you can fish them out again if you want. Mar 05 16:13:12 if the former, that is weird; I can't think of anything that ought to be deleting those files. Mar 05 16:14:53 The loose files, I don't see any ipks anywhere either. Mar 05 16:15:12 Also, broken symlink so you know what I mean: Mar 05 16:15:21 lrwxrwxrwx 1 james james 25 2010-02-26 17:01 /home/james/OE/angstrom-dev/deploy/glibc/images/tosa/zImage-tosa.bin -> zImage-2.6.30-r5-tosa.bin Mar 05 16:15:37 Only that symlink and the zImage-kexecboot are there. Mar 05 16:15:55 Nothing zImage related (or modules) related is there. Mar 05 16:19:32 hrw: thanks, I'm trying to upgrade samba right now... too many versions.. I started with the bad one Mar 05 16:19:52 hrw: and while fetching I saw netatalk as dep... which of course has known security issues too Mar 05 16:22:24 Laibsch: postinst scripts are also called during do_rootfs Mar 05 16:22:56 kergoth: hrm, it seems that underscores just don't work in OVERRIDES. Mar 05 16:23:00 that is a bit of a shame Mar 05 16:23:11 hrw: I see. Thanks. Mar 05 16:23:13 huh, thats .. weird, and rather arbitrary Mar 05 16:23:26 i could see filtering out the separator, but other than that.. Mar 05 16:23:39 pb_: any ideas? Mar 05 16:23:49 well, i guess i see why, but still annoying Mar 05 16:23:52 hmm. Mar 05 16:24:05 * kergoth gets caffeine :) Mar 05 16:25:25 james_l: not offhand. I think you will have to inspect the relevant packages for things that delete from DEPLOY_DIR. afaik, there should not be any. Mar 05 16:26:50 kergoth: yeah, I think the _seen_overrides thing in data_smart is to blame. it is annoying. Mar 05 16:26:57 How is the LD cache being generated at image creation time? Mar 05 16:27:35 Laibsch: iirc, it isn't in most cases. I guess anybody who wants it must be using qemu for that. Mar 05 16:27:49 I see Mar 05 16:27:55 I think most distros either do without it altogether, or create it on first boot. Mar 05 16:27:59 I was told that was a required thing to do Mar 05 16:28:05 OK Mar 05 16:28:07 thanks Mar 05 16:28:55 it isn't required, ld.so will work perfectly happily without it (and, in the case of micro, is probably faster without it in most cases) Mar 05 16:30:47 Laibsch: Poky has/had ldconfig-native but I do not remember how good it was working Mar 05 16:31:01 Laibsch: I also had patch which used ldconfig with qemu long time ago Mar 05 16:45:18 bye all Mar 05 17:18:37 have a nice weekend Mar 05 17:51:20 How are bootloaders handled in OE? Is that a distro choice, a machine setting, an image setting? Mar 05 17:52:10 looks like it is a machine setting Mar 05 18:04:42 I believe all of them are per-machine. Mar 05 18:10:15 hmmm. Mar 05 18:10:46 RP: any idea, offhand, why an INHIBIT_AUTOTOOLS_DEPS_virtclass-native wouldn't behave as one would expect? Mar 05 18:10:55 * kergoth 's investigating, but figured he'd check Mar 05 18:12:04 Yes, it's all per machine, with a uboot machine feature, or so, to have the uboot userland tools installed at least Mar 05 18:12:11 otherwise, iirc, they just live in deploy Mar 05 18:12:34 03Koen Kooi  07org.openembedded.dev * rce6921da10 10openembedded.git/recipes/gnutls/gnutls.inc: gnutls: 2.8.5 lacks the .m4 files, so don't abort if cp fails Mar 05 19:02:06 Tartarus: thanks Mar 05 19:16:02 something in OE is breaking bloomin kernel compiles and I cant figure out what :-( Mar 05 19:17:00 XorA: that thing with "base_version_less_or_equal(KERNEL_VERSION, 2.6.17, , apm-wifi-suspendfix" ? Mar 05 19:17:20 JaMa: no Mar 05 19:17:37 JaMa: I think we are setting something in OE that is fucking Kconfig Mar 05 19:17:56 I can hack around it, but I would rather find the cause Mar 05 19:21:03 03Koen Kooi  07org.openembedded.dev * rde69350beb 10openembedded.git/conf/distro/include/angstrom-2008-preferred-versions.inc: angstrom: prefer gnutls 2.4.2, 2.8.5 induces too much breakage Mar 05 19:28:46 contrib/source-checker/oe-checksums-sorter.py is not in my bitbake/contrib folder. All that is in there is bbdev.sh and subfolder vim. Mar 05 19:29:00 hi Mar 05 19:29:04 Why am I missing that - and what else am I missing? Mar 05 19:29:04 XorA: like that: Mar 05 19:29:06 | arch/avr32/kernel/entry-avr32b.S: Assembler messages: Mar 05 19:29:06 | arch/avr32/kernel/entry-avr32b.S:190: Error: attempt to .org/.space backwards? (-6) Mar 05 19:29:09 | arch/avr32/kernel/entry-avr32b.o: Bad value Mar 05 19:29:11 XorA: ? Mar 05 19:29:13 | arch/avr32/kernel/entry-avr32b.S:874: FATAL: Can't write arch/avr32/kernel/entry-avr32b.o: Bad value Mar 05 19:29:24 XorA: thats 2.6.33 for atngw100 Mar 05 19:30:59 I'm trying to add a checksum to a new recipe, as per this: http://blogs.elphel.com/2009/12/openembeddedangstom-new-package-workflow-eyeos/ Mar 05 19:31:17 robtow: check tmp/ dir after fetch Mar 05 19:31:27 But my contrib folder seems curiously empty, and does not have oe-checksums-sorter.py Mar 05 19:31:53 hrw: yes, Mar 05 19:31:53 oh? Mar 05 19:32:22 hrw: the problem is somehow OE stops KBUILD_CFLAGS += -pipe -fno-builtin -mno-pic Mar 05 19:32:23 KBUILD_AFLAGS += -mrelax -mno-pic Mar 05 19:32:28 being evaluated Mar 05 19:33:13 XorA: nice catch. where they are set? Mar 05 19:33:29 arch/avr32/Makefile Mar 05 19:33:30 hrw - I see checksums.ini in tmp Mar 05 19:33:59 robtow: .dev tree? Mar 05 19:34:23 hrw - not following you. Mar 05 19:34:32 Sorry Mar 05 19:35:08 Trying to update gPhoto2 recipe to next version; as per workflow in http://blogs.elphel.com/2009/12/openembeddedangstom-new-package-workflow-eyeos/ Mar 05 19:35:39 robtow: OE from org.openembedded.dev tree generates new style checksums in tmp/ dir (and checksums.ini too). stable/2009 does only checksums.ini Mar 05 19:36:25 hrw: OE toolchain outside OE will build avr32 kernel inside OE it breaks Mar 05 19:36:29 hrw: bloody annoying Mar 05 19:37:16 hrw. Hmm. So I need to change my OE tree? Mar 05 19:37:16 * XorA goes to eat Mar 05 19:39:00 git clone git://git.openembedded.org/openembedded Mar 05 19:39:05 ??? Mar 05 19:39:13 robtow: which one you use now? Mar 05 19:40:24 hrw - MY OE is what I got from Mar 05 19:40:33 ok, so it is .dev Mar 05 19:41:08 The README file does not distinguish Mar 05 19:41:41 hrw - so the contrib dir does not have the subfolder source-checker in it. Mar 05 19:41:58 So How do I add the checksum info? Mar 05 19:43:12 robtow: strange - my .dev tree has contrib/source-checker/ Mar 05 19:43:21 hmmm Mar 05 19:43:34 Strange, indeed. Mar 05 19:43:35 robtow: anyway. after do_Fetch fail due to lack of checksum PLEASE take a look at tmp/ dir contents Mar 05 19:43:51 Yes, checksums.ini is in tmp Mar 05 19:44:03 what is a name of archive you are fetching? Mar 05 19:44:24 twO: Mar 05 19:44:28 http://downloads.sourceforge.net/gphoto/libgphoto2-2.4.8.tar.bz2 Mar 05 19:44:49 http://downloads.sourceforge.net/gphoto/libgphoto2-2.4.8.tar.bz2 Mar 05 19:45:15 well, same one twice Mar 05 19:47:18 and you tmp/ does not have libgphoto2248tarbz2.sum file? Mar 05 19:48:15 yes, it does. Mar 05 19:49:04 What do I do with that file? Mar 05 19:50:17 add to the recipe in sane way Mar 05 19:52:17 hrw - the previous recipe does not have anything that looks like that in it. So I fear I am not able to be sane. Mar 05 19:53:34 Thanks for trying to be helpful, tho. Mar 05 19:57:56 robtow: it's new way of setting checksums directly in recipe Mar 05 19:58:13 robtow: instead of global checksums.ini Mar 05 19:58:21 jaMa -- Hmmm. I almost see! Mar 05 20:00:06 robtow: look at gnutls.inc and gnutls_2.8.5.bb Mar 05 20:17:48 Wilco. Mar 05 20:20:59 gnutls_2.8.5.bb is not anywhere in my tree Mar 05 20:21:50 doing git pull Mar 05 20:22:42 Now it is in my tree. Mar 05 20:23:16 So is source-checker Mar 05 20:23:27 Mar 05 20:42:27 http://www.youtube.com/watch?v=dzPJWvE5RfU - one BUG + 8 usb webcams Mar 05 20:47:18 03Marcin Juszkiewicz  07org.openembedded.dev * ra42d240cc5 10openembedded.git/recipes/gnutls/gnutls.inc: Mar 05 20:47:18 gnutls: install m4 macros only in <2.8.5 versions Mar 05 20:47:18 2.8.5 does not have M4 macros which were available for older versions. Mar 05 20:47:18 Debian does not ship them too. Mar 05 20:47:18 Signed-off-by: Marcin Juszkiewicz Mar 05 20:58:34 khem, did i send you gmp-5.0.1 and mpfr-2.4.2-p3 ? Mar 05 20:59:27 khem, i don't see them applied yet and they're not in my outbox. just askin'.. Mar 05 21:01:21 kergoth, AFAICS you didn't apply that cosmetic s/ != None// to bitbake. Anything i should do? insist later on, forget it? Mar 05 21:02:07 kergoth, same for the "no bb files" thing, btw Mar 05 21:02:34 what you should do is what you just did.. remind me :) Mar 05 21:04:47 kergoth, and with pristine bb master i also get a backtrace for un-provided tasks. Typo of mine: bitbake micro-uclibc-image (instaed of micro-image; i configured micro-uclibc as distro, but ok, i may be confused wrt the naming conventions) leads to a proper Error but then to a backtrace which is rather unelegant in my poor user eyes :) Mar 05 21:05:01 hmm, okay, that should get fixed Mar 05 21:06:53 sounds like errors (non multiline errors or at least at sequence points) should just raise a cooker error if the error-buffer is not empty (or sprinkle lots of raises to the respective spots, i wouldn't know) Mar 05 21:10:03 JaMa, k, will do Mar 05 21:10:52 JaMa, erm, you're not Marcin, sorry.. Mar 05 21:13:00 well Mar 05 21:13:16 I've got Angstrom running on the SBC6020 :D Mar 05 21:13:50 someone at the embest forum had the correct sources :D Mar 05 22:37:55 * XorA is really getting fscked off with the kernel build system Mar 05 22:38:23 XorA: do the unset FOO in do_compile, etc, yet? Mar 05 22:38:42 Tartarus: yes, even stuck an export in there to show me what is left Mar 05 22:38:58 unset CFLAGS CPPFLAGS CXXFLAGS LDFLAGS MACHINE CC LD AS Mar 05 22:39:25 I just cant work out why compiling outside OE gives a different result that inside OE Mar 05 22:46:42 OMG Mar 05 22:46:46 OMG OMG OMG Mar 05 22:51:44 finally bloody linux.inc has a sed which seds away essential stuff Mar 05 22:56:15 sed -i -e s:-mno-pic::g arch/avr32/Makefile Mar 05 22:56:20 that? Mar 05 22:58:15 hrw|gone: yes Mar 05 22:58:40 hrw|gone: so much pain for one simple little line Mar 05 23:09:45 NOTE: Running task 794 of 806 (ID: 9, /home/hrw/devel/oe/openembedded/recipes/linux/linux_2.6.33.bb, do_install) Mar 05 23:09:58 XorA: you are great Mar 05 23:10:02 * hrw|gone -> bed Mar 05 23:10:12 hrw|gone: https://www.xora.org.uk/cgi-bin/cgit.cgi/openembedded/log/?h=xora/avr32-updates Mar 05 23:14:13 Hrm, gtk+ is breaking on me. It seems to be breaking in the demo section, but as far as I can tell the last commit was supposted to remove the demos from gtk. Mar 06 00:54:49 ok, now it's fixed **** ENDING LOGGING AT Sat Mar 06 02:59:57 2010