**** BEGIN LOGGING AT Sat Feb 26 02:59:57 2011 Feb 26 06:26:21 Tartarus: pong Feb 26 06:28:53 Question about switching to bitbake 1.12.0: If I run bitbake --version from my OE build dir, I get "BitBake Build Tool Core version 1.12.0, bitbake version 1.8.12". Is this normal? Feb 26 07:12:23 also, I get the following error "TypeError: __init__() takes exactly 3 arguments (2 given)" Feb 26 10:39:42 gm Feb 26 10:39:52 morning Feb 26 11:02:42 there is any example of cross-cross-compiling using oe? objective build in host(x86 64bit) an avr toolchain(binutils gcc avrlibc) for use on mips Feb 26 14:56:47 ok, time to see if libcap2 is a dep dep in some images or something Feb 26 14:56:53 s/dep dep/new dep/ Feb 26 14:57:03 Not sure how 2.16 could have built on uclibc either Feb 26 14:57:44 yes, it is, there we go Feb 26 15:11:50 03Tom Rini  07master * rf993f993f5 10openembedded.git/conf/local.conf.sample: Feb 26 15:11:50 local.conf.sample: Update BBMASK example Feb 26 15:11:50 Since bitbake.conf has a real BBMASK (and has for some time) we shouldn't Feb 26 15:11:50 suggest that people start out by wiping that out. Feb 26 15:11:50 Signed-off-by: Tom Rini Feb 26 15:11:59 03Tom Rini  07master * r150102aaf8 10openembedded.git/recipes/libcap/ (libcap2/fix-uclibc-build.patch libcap2_2.20.bb): Feb 26 15:11:59 libcap2: Fix uClibc build Feb 26 15:11:59 On uClibc we have not Feb 26 15:11:59 Signed-off-by: Tom Rini Feb 26 15:14:47 03Tom Rini  07master * rc50a4ef00e 10openembedded.git/classes/sdk.bbclass: Feb 26 15:14:47 sdk.bbclass: Need to mask TOOLCHAIN_OPTIONS here as well Feb 26 15:14:47 Signed-off-by: Tom Rini Feb 26 15:15:12 OK, that should fix most of the failures I can see atm Feb 26 15:24:18 Hi all Feb 26 15:26:39 gm Feb 26 15:27:07 I have some packages that not compiled correctly,but if i do bitbake -c clean package,and bitbake package,it compiles ok. Is there any trick to make openembedded clean all packages unbuilded and build it again? Feb 26 15:39:14 mwester, in yet? Feb 26 15:39:35 Just stepped into my office here 5 mins ago. :) Feb 26 15:39:38 heh Feb 26 15:39:43 good timing. Feb 26 15:39:46 So, found a *slug* problem Feb 26 15:39:52 uh-oh Feb 26 15:39:56 Easy fix Feb 26 15:40:01 distro confs Feb 26 15:40:06 slugos.conf does TARGET_ARCH ?= Feb 26 15:40:14 ucslug does TARGET_ARCH = Feb 26 15:40:29 Is there a reason it's just not left to the machine to set, period? Feb 26 15:41:11 The failure I see is forcing sheevaplug to BE Feb 26 15:41:25 Heh! Feb 26 15:41:25 (on ucslug) Feb 26 15:41:39 I was wondering about that myself, but never got to figuring it out! Feb 26 15:41:57 That should be fixed then. Feb 26 15:42:01 Best is to just drop TARGET_ARCH from *slug*.conf Feb 26 15:42:15 Agreed. It's silly to specifiy it there in the first place. Feb 26 15:42:28 k, thanks. I'll let you go in and fix it Feb 26 15:42:32 Off to the next failure :) Feb 26 15:42:37 Will do. :) Thanks! Feb 26 15:55:26 03Mike Westerhof  07org.openembedded.dev * rd86370d18a 10openembedded.git/conf/distro/ (include/slugos.inc slugos.conf ucslugc.conf): SlugOS: remove TARGET_ARCH (set in machine conf instead), bump up distro to v6.0 Feb 26 15:55:54 Thanks Tartarus Feb 26 15:56:08 np Feb 26 15:57:07 there is any example of cross-cross-compiling using oe? objective build in host(x86 64bit) an avr toolchain(binutils gcc avrlibc) for use on mips Feb 26 15:58:04 btw Tartarus finally the build ended succesfully thanks a lot Feb 26 15:58:04 tuxbrain_away, the canadian-sdk stuff shows up to do that in general, but is for builds on build-host, targets $whatever, runs on mingw32 Feb 26 15:58:35 03Tom Rini  07master * r656cc1a317 10openembedded.git/classes/gnome.bbclass: Feb 26 15:58:36 gnome.bbclass: Add --disable-scrollkeeper to EXTRA_OECONF Feb 26 15:58:36 Virtually all gnome recipes pass this already, this just fixes Feb 26 15:58:36 a few such as zenity. Feb 26 15:58:36 Signed-off-by: Tom Rini Feb 26 16:01:53 tuxbrain_away: build native-sdk-image for machine "qemumips" and use that image to build avr toolchain inside qemu :) Feb 26 16:03:02 hm, building binutils for sheevaplug fails in configure: Feb 26 16:03:03 | configure.ac:32: error: Please use exactly Autoconf 2.64 instead of 2.65. Feb 26 16:03:11 what is the way to tackle this ? Feb 26 16:04:19 that was actually fatal? Feb 26 16:04:54 iirc, we usually just delete that check Feb 26 16:05:03 and make sure the reconf still looks sane, by hand, afterwards Feb 26 16:05:10 did anybody here get errors like this one after switching to bitbake 1.12.0: Feb 26 16:05:13 File "/usr/bin/bitbake", line 117, in main Feb 26 16:05:14 cooker = bb.cooker.BBCooker(configuration) Feb 26 16:05:17 TypeError: __init__() takes exactly 3 arguments (2 given) Feb 26 16:06:30 Tartarus: it was fatal, actually i found on the web someoene who just patched the check Feb 26 16:06:45 and i have not too much ideas on how the stuff should look like Feb 26 16:07:41 actually I am kinda surprised I am getting this, had expected that binutils is mature Feb 26 16:08:21 hvontres|home: I've seen that one logn time ago, not on 1.12.0 Feb 26 16:12:52 03Tom Rini  07master * rb789be061e 10openembedded.git/recipes/gimp/gimp.inc: (log message trimmed) Feb 26 16:12:52 gimp.inc: Drop sed-native from DEPENDS Feb 26 16:12:52 This was introduced forever ago and doesn't look to make sense Feb 26 16:12:52 anymore (there's no non-standard regexes being applied). This Feb 26 16:12:52 is also racy with 'sed' being invoked elsewhere. If we really Feb 26 16:12:52 can't rely on host sed we need to tackle this early and for Feb 26 16:12:53 everyone. Feb 26 16:22:45 Tartarus: hi, where do you have attr/xattr.h in uClibc? I was fixing this today also, but quite differently - by adding xattr into the DISTRO_FEATURES Feb 26 16:23:30 Tartarus: isn't it comming from some different lib, I think something like ext2fs related? Feb 26 16:28:44 ynezz, oh, good point, lemme double check Feb 26 16:29:39 yeah, arg :( Feb 26 16:29:46 must have come from something else Feb 26 16:31:28 but apparently the DISTRO_FEATURES += "xattr" doesn't work, but according to the uclibc-config.inc it should, dunno what's wrong Feb 26 16:32:55 hmmmm Feb 26 16:34:48 what conf did you do that in? Feb 26 16:35:08 micro-uclibc.conf Feb 26 16:36:09 ah, I missed the DISTRO_FEATURES in micro.conf Feb 26 16:36:11 argh Feb 26 16:39:30 rebuilding Feb 26 16:39:50 I think, that the minimal will need the same fix Feb 26 16:39:55 yeah Feb 26 16:39:59 and maybe a few others Feb 26 16:40:18 and what about jack/libvpx fixes? Feb 26 16:40:50 jack fails on one box and the libvpx on another Feb 26 16:40:59 libvpx is known Feb 26 16:41:22 I'll pull the workaround in for the release but khem wants to look at the real problem Feb 26 16:41:31 for libvpx? Feb 26 16:41:36 yeah Feb 26 16:41:44 your jack patches look fine Feb 26 16:41:45 it's git repo problem, asked on #vp8 Feb 26 16:42:03 say again? Feb 26 16:42:42 16:14:17 < wwcv> ynezz: not sure what's broken, but i'm getting the same. will look into it Feb 26 16:42:55 http://patches.openembedded.org/patch/256/ is the workaround Feb 26 16:42:59 or at least khem says it a workaround not a fix Feb 26 16:43:27 ah, it's different issue Feb 26 16:44:50 http://lists.linuxtogo.org/pipermail/openembedded-devel/2011-February/030430.html Feb 26 16:45:10 I thought I pulled that in Feb 26 16:45:30 than I missed it, sorry Feb 26 16:46:18 but I lost it? bah Feb 26 16:46:50 ah, seems so :) Feb 26 16:46:57 checked it now and it's missing Feb 26 16:47:21 * Tartarus checks what else he might have goofed Feb 26 16:47:24 :) Feb 26 16:47:27 ok, bbl Feb 26 16:47:37 ok, so I think I see it Feb 26 16:47:43 I wanted to push it, but patchwork missed it :( Feb 26 17:35:15 Tartarus: build finished fine, so this patch for micro-uclibc http://pastebin.com/ceZ0u2aS fixes that libcap2 build issue Feb 26 17:36:10 confirmed minimal as well Feb 26 17:38:16 ynezz, posted to the ML? Feb 26 17:38:52 nope Feb 26 17:39:38 can you? i Feb 26 17:39:43 'll git am them in and push Feb 26 17:39:48 grabbed your jack ones Feb 26 17:39:52 ok, no problem Feb 26 17:40:48 what about other -uclibc based distributions? Feb 26 17:41:51 btw, can you push such change without 2 ACKs? Feb 26 17:42:26 Khem's busy at SCALE atm Feb 26 17:42:33 and micro is rather unmaintained :| Feb 26 17:43:04 but seems good for some corner cases Feb 26 17:43:20 it's interesting and I wish it was more actively maintained Feb 26 17:45:21 sent Feb 26 17:49:47 Tartarus: btw I've rw access, so I could push it myself after I got 2 ACKs Feb 26 17:50:15 ah, right, in that case, sec.. Feb 26 17:50:29 pls, push it yourself, no problem Feb 26 17:50:42 03Petr Štetiar  07master * r884e3ae5d8 10openembedded.git/recipes/jack/ (files/remove-wrong-host-test.patch jack_0.118.0.bb): (log message trimmed) Feb 26 17:50:42 jack: fix wrong configure check on the build host Feb 26 17:50:42 | checking existing, conflicting JACK installs... Feb 26 17:50:42 | You appear to have at least one existing installation of JACK. Feb 26 17:50:42 | Complete or partial JACK installs exist in: /usr/lib Feb 26 17:50:43 [...] Feb 26 17:50:44 NOTE: package jack-0.118.0-r2: task do_configure: Failed Feb 26 17:50:52 I'm travelling, so bad inet access atm Feb 26 17:50:55 bbl Feb 26 17:50:55 k Feb 26 17:50:59 later & thanks Feb 26 17:51:04 03Tom Rini  07master * rb93d21c42d 10openembedded.git/conf/distro/minimal-uclibc.conf: Feb 26 17:51:04 minimal-uclibc.conf: Add xattr to DISTRO_FEATURES Feb 26 17:51:04 Signed-off-by: Tom Rini Feb 26 17:51:04 03Tom Rini  07master * rbf79b8822f 10openembedded.git/recipes/libcap/ (libcap2/fix-uclibc-build.patch libcap2_2.20.bb): Feb 26 17:51:04 Revert "libcap2: Fix uClibc build" Feb 26 17:51:04 Petr Štetiar pointed out that I just switched this to using Feb 26 17:51:05 attr/xattr.h from e2fsprogs or similar rather than correctly fixing this. Feb 26 17:51:05 This reverts commit 150102aaf8226fcb56101ad0fbeda00ca95aef72. Feb 26 17:51:06 03Petr Štetiar  07master * rb31a6b5b8a 10openembedded.git/recipes/jack/ (jack_0.103.0.bb jack_0.116.2.bb): Feb 26 17:51:06 jack: remove old versions Feb 26 17:51:07 The jack is not pinned anywhere and according to the git log it seems, that Feb 26 17:51:10 just the latest stable version is maintained and used. Feb 26 17:51:10 Signed-off-by: Petr Štetiar Feb 26 17:51:10 Signed-off-by: Tom Rini Feb 26 17:51:10 03Petr Štetiar  07master * r0478950960 10openembedded.git/recipes/webm/ (6 files in 2 dirs): (log message trimmed) Feb 26 17:51:11 libvpx: fix fetch errors, remove older releases and add security fix Feb 26 17:51:11 - remove old and unfetchable versions v0.9.0, v0.9.1, v0.9.2 Feb 26 17:51:12 - add v0.9.5 patch for CVE-2010-4203 security fix Feb 26 17:51:12 - fix fetch errors for v0.9.2 and v0.9.5: Feb 26 18:03:57 03Petr Štetiar  07master * r46169e02d9 10openembedded.git/conf/distro/ (micro-uclibc.conf micro.conf): Feb 26 18:03:57 micro-uclibc: add xattr to distro features to fix build failure of libcap2 Feb 26 18:03:57 cap_file.c:8:23: fatal error: sys/xattr.h: No such file or directory Feb 26 18:03:57 Signed-off-by: Petr Štetiar Feb 26 18:03:57 Signed-off-by: Tom Rini Feb 26 18:35:56 re Feb 26 18:40:00 evening Feb 26 18:40:07 hi jay7 Feb 26 18:40:17 rere Feb 26 18:47:14 hi woglinde Feb 26 18:47:19 he pb Feb 26 18:47:47 pb you have some kindergarten in uk too? Feb 26 18:49:06 woglinde: yes, though we don't usually use that name. Feb 26 18:49:44 pb whats called then? Feb 26 18:50:21 woglinde: the thing that very young children go to (up to about age 3) is usually a "nursery". from 3-4 you go to "preschool", and then at 4 you go to primary school. Feb 26 18:50:28 ah Feb 26 18:51:19 hm with 4 into school Feb 26 18:52:24 well.. all clean builds are fine on testing branch Feb 26 18:52:35 yeah, compulsory education starts at 4 in this country Feb 26 18:52:35 all incremental are failed by some reason Feb 26 18:52:43 mostly by uicmoc-native Feb 26 18:52:49 one by freetype Feb 26 18:53:00 ah, no Feb 26 18:53:04 cups, not freetype Feb 26 18:53:05 jay7 hm whats with freetype? Feb 26 18:53:11 aah good Feb 26 18:53:23 so my patch didnt destroy to much Feb 26 18:53:28 incremental builds are mess.. Feb 26 18:53:50 seems I should remove old TMPDIR and repopulate Feb 26 18:57:13 yes yes Feb 26 19:03:20 hm, I wonder if console-image should build with micro-uclibc Feb 26 19:03:34 hm Feb 26 19:03:37 should Feb 26 19:03:57 maybe somme stuff chokes about the directory layout Feb 26 19:04:03 which only has /bin and lib Feb 26 19:04:35 hm Feb 26 19:12:29 ynezz: have you found workaround for building different *libc variants with testbuilder? Feb 26 19:12:39 eFfeM_work: Thanks. Any idea how you fixed it? Feb 26 19:13:52 Jay7: nope, I just modify the local.conf when building Angstrom and add ANGSTROMLIBC variable Feb 26 19:14:02 tartarus ping Feb 26 19:16:33 well.. one pair of builds left Feb 26 19:17:22 ah they sorted the libcap stuff out Feb 26 19:17:32 good work ynezz saw the xattr featur too Feb 26 19:18:51 But it raises the issue: do we really expect that every software package will build with every libc implementation? Feb 26 19:19:17 There has to be a compromise somewhere, otherwise why wouldn't we just select the smallest libc and be done with it? Feb 26 19:25:25 mwester: Are there any build instructions for Slugos 6.0? I just got a second slug and would like to help test Feb 26 19:26:05 MACHINE=nslu2le (or nslu2be) and DISTRO=slugos --- and then bitbake slugos-image Feb 26 19:27:05 Install is the same as SlugOS 5.3: http://www.nslu2-linux.org/wiki/SlugOS/InstallandTurnupABasicSlugOSSystem Feb 26 19:27:33 mwester: I agree and looking for some suggestions Feb 26 19:27:36 mwester: Thanks. I'll give that a try as soon as I figure out how I busted my bitbake install Feb 26 19:28:11 ynezz, I wonder if perhaps we need a new variable similar to COMPATIBLE_MACHINE Feb 26 19:45:44 join #bitbake Feb 26 19:46:03 When I try to compile fbv it tries to find libjpeg.so on the host. How do I make it instead look in the right place? Feb 26 19:46:18 risca hm let me see Feb 26 19:46:41 Log file coming up Feb 26 19:46:59 hm no Feb 26 19:47:01 I meant Feb 26 19:47:10 I trying it myself Feb 26 19:47:47 No need for a log file, then? Feb 26 19:47:53 will see Feb 26 19:48:29 Well, here it is anyway ^^ Feb 26 19:48:30 http://oe.pastebin.com/ZKyTrb1Q Feb 26 19:50:12 hm dev with libtool4 worked Feb 26 19:51:50 risca local.conf please Feb 26 19:51:57 I'm using DISTRO="minimal", which I think is using libtool 2.4 Feb 26 19:52:07 I was looking at that right now xD Feb 26 19:52:15 hm Feb 26 19:53:30 http://oe.pastebin.com/ZyVRPPv1 Feb 26 19:53:32 There Feb 26 19:53:50 risca ah see it now Feb 26 19:53:56 compiler skips it here Feb 26 19:54:06 thanks for noticing Feb 26 19:54:38 I meant compiler skips the incompatible Feb 26 19:55:06 The compiler skips the incompatible libjpeg? Feb 26 19:55:12 yes Feb 26 19:55:32 I don't see in my log that it even tries the libs that should be compatible Feb 26 19:55:44 which compiler? Feb 26 19:56:01 The cross-compiler (arm-etc..) Feb 26 19:56:06 version Feb 26 19:57:06 gcc 4.5 I think Feb 26 19:58:08 When looking at http://oe.pastebin.com/ZKyTrb1Q I only see a "-L/usr/lib" and not any "-L/*path to TMPDIR*/sysroot/....etc..." Feb 26 19:58:44 rsica yes I saw it too Feb 26 19:58:53 I will check the Makefile Feb 26 19:59:28 arm-oe-linux-gnueabi-gcc (GCC) 4.5.3 20110214 (prerelease) <-- confirmed gcc 4.5.3 Feb 26 19:59:36 okay Feb 26 19:59:37 yes Feb 26 19:59:50 ah its actual the linker Feb 26 19:59:57 which skips the incompatible Feb 26 20:01:00 Yeah, figured as much Feb 26 20:04:56 Three Cheers for the bitbake devs :) I really like the new interface :) Feb 26 20:06:24 woglinde, So what would one do to make it find the compatible libs? Feb 26 20:22:06 gosh, the libelf download site is slow Feb 26 20:42:50 risca maybe we need to alter the configure or makefile Feb 26 20:50:00 risca I will fix it Feb 26 21:02:01 he stefan_schmidt Feb 26 21:02:15 hi woglinde Feb 26 21:14:19 woglinde, I will look into this problem a little later. Right now I have to fix some server issues Feb 26 21:17:54 rsica I already fixed it Feb 26 21:18:07 will push it with my xawtv changes Feb 26 21:56:26 re Feb 26 21:57:50 03Henning Heinold  07org.openembedded.dev * rc9c44dd2fb 10openembedded.git/recipes/fbv/ (4 files in 2 dirs): Feb 26 21:57:50 fbv: fix cflags and libs and enable libungif Feb 26 21:57:50 * remove missleading patch Feb 26 21:57:50 * bump PR Feb 26 21:57:50 risca stuff is pushed Feb 26 21:57:52 03Henning Heinold  07org.openembedded.dev * r106d2dcc40 10openembedded.git/recipes/fs/ (files/fontsproto.patch files/xtrans.patch fs_cvs.bb): fs: fix header inclusions and bump PR Feb 26 21:57:54 03Henning Heinold  07org.openembedded.dev * r2caafc8ebe 10openembedded.git/recipes/xawtv/ (20 files in 3 dirs): Feb 26 21:57:54 xawtv: update to version 3.95 Feb 26 21:57:54 * added usefull debian-patches Feb 26 21:57:54 * convert sys_siglist to strsignal Feb 26 21:57:54 fixes uclibc build Feb 26 21:57:54 * enable alsa Feb 26 21:57:55 * remove older version 3.93 Feb 26 21:58:10 woglinde_, thanks =) Feb 26 22:01:15 please try if it works on device Feb 26 22:08:03 hi ant Feb 26 22:09:06 hey Henning Feb 26 22:10:12 03Andrea Adami  07org.openembedded.dev * r43d262ae2a 10openembedded.git/recipes/linux/linux-kexecboot-2.6.24/poodle/defconfig: Feb 26 22:10:12 linux-kexecboot: rework 2.6.24 defconfig for Zaurus poodle. Feb 26 22:10:12 * add boot from CF Feb 26 22:10:12 * add vfat and ext3 filesystems Feb 26 22:10:12 Signed-off-by: Andrea Adami Feb 26 22:21:52 ant__: hey there :) Feb 26 22:22:27 ant__: looks like I should try building kexecboot for poodle tonight, right after my slug build finishes Feb 26 22:25:44 hvontres|home: hey :) Feb 26 22:26:37 * Jay7 is working on text UI mode for kexecboot right now Feb 26 22:27:41 good nite Feb 26 22:29:13 hvontres|home: hey Feb 26 22:29:24 I got poodle finally Feb 26 22:30:20 ant__: cool. Feb 26 22:32:20 ant__: any luck with getting 2.6.3x to boot on poodle yet? Feb 26 22:34:45 * mwester gives up getting bitbake 1.12 working on Fedora. Feb 26 22:34:48 :( Feb 26 22:35:16 hmm Feb 26 22:35:39 what's the problem? Feb 26 22:36:07 hvontres|home: not yet Feb 26 22:39:08 prints the progress bar, then hangs up competely. Several [defunct] python processes hang around. Nothing but "killall python" stop it at that point. Feb 26 22:39:11 mwester, 1.12 seems sane on my F13 box Feb 26 22:39:48 ant__: Yeah, I think the power management stuff needs to be re-done and I have no clue where to start :( Feb 26 22:39:50 I skip odd versions, but haven't got to getting F14 running on the build servers yet. Feb 26 22:40:10 But F12 isn't *that* old. Feb 26 22:40:30 hvontres|home: lumag is in same state with tosa :) Feb 26 22:40:47 and Pavel Machek with his spitz iirc :) Feb 26 22:41:02 we all are in stuck wrt power management.. Feb 26 22:42:03 there was some attempts in zaurus-devel@ Feb 26 22:44:49 * mwester finds a Fedora 12 VM set up to build SlugOS, and wonders how hard it would be to attempt an upgrade to F14. Feb 26 23:23:34 03Andrea Adami  07org.openembedded.dev * r50e7d72e63 10openembedded.git/recipes/linux/linux-kexecboot-2.6.26/poodle/defconfig: Feb 26 23:23:34 linux-kexecboot: rework defconfig for Zaurus poodle Feb 26 23:23:34 * sync kernel config with other Zaurus models Feb 26 23:23:34 Signed-off-by: Andrea Adami Feb 26 23:23:59 hvontres|home: ^^ 2.6.26 seems not booting :/ Feb 26 23:30:28 ant__: do you have a serial cable Feb 26 23:31:01 yes Feb 26 23:31:18 ant__: how far is it getting before it hangs? Feb 26 23:31:30 early, no bootlogo Feb 26 23:31:37 lcd black Feb 26 23:32:15 ant__: if you look at the serial output, does the kernel de-compress at all? Feb 26 23:32:30 I still have to check Feb 26 23:32:33 mom Feb 26 23:34:49 I'll have to rebuild with printk ;) Feb 26 23:35:53 ant__: BTW, I found that you can kexexboot to another kexecboot on your SD card... no re-flash needed :) Feb 26 23:37:20 ah, sure Feb 26 23:37:35 atm we have a nasty bug with kexecboot master Feb 26 23:38:11 with qvga Feb 26 23:38:23 * Jay7 is working hardly on adding bugs^Wfeatures to kexecboot :) Feb 26 23:49:23 bitbake 1.12 also fails on Fedora 13 - new install, up-to-date. Feb 26 23:50:08 Oh - no, my mistake (permissions problem). Feb 26 23:50:15 Let's try that again. Feb 26 23:50:53 Ok. So bitbake 1.12 work on Fedora 13, but not on Fedora 12. Feb 27 00:04:13 mwester: root@LKG0F902A:~# uname -a Feb 27 00:04:14 Linux LKG0F902A 2.6.37 #1 Sat Feb 26 15:24:40 PST 2011 armv5teb GNU/Linux Feb 27 00:04:34 mwester: thanks for all the work on Slugos :) Feb 27 01:12:51 hmmm, has anybody seen this error on autotools: "checking host system type... Invalid configuration `armeb-oe-linux-gnueabi': machine `armeb-oe' not recognized Feb 27 01:12:55 " Feb 27 02:50:09 Ok, that was painful. Feb 27 02:51:00 But, I've traced the problem with bitbake 1.12 down to the python version. 2.6.4 works, 2.6.2 does not. Feb 27 02:51:30 Fedora 13 and higher use 2.6.4, but the last update for Fedora 12 stopped at 2.6.2 Feb 27 02:52:29 So, the question is if the problem can be worked around in bitbake -- which would make it easier for a larger group of people to use bitbake 1.12, or if the problem is too difficult, in which case OE may have to support bitbake 1.10 longer then people would like. Feb 27 02:53:18 (Or I suppose, OE could just force the issue and insist that every OE user be bleeding edge) Feb 27 02:54:10 It is nice to see muiltiple cores parsing recipes. :) Feb 27 02:55:39 mwester: any clue where bitbake falls appart? Feb 27 02:56:19 In a place where I find myself unable to debug it -- immediately after it spawns the threads to do the multiprocessing. Feb 27 02:56:51 stracing shows the parent process opening all the semaphores, and such-like -- and then the children go [defunct], and the parent gets stuck waiting forever. Feb 27 02:57:10 I have no idea how to even begin to strace the child processes/threads. Feb 27 02:57:25 (python is not even a second language for me, I'm afraid) **** ENDING LOGGING AT Sun Feb 27 02:59:57 2011