**** BEGIN LOGGING AT Sat Dec 18 03:00:01 2010 Dec 18 03:23:21 ka6sox, barely Dec 18 03:23:32 Crofton,I hear ya Dec 18 07:13:56 bah...git clone git://git.openembedded.org/bitbake.git /oe/bitbake is very unresponsive today :/ Dec 18 07:17:27 git pull hangs too Dec 18 08:04:23 well.. tonight's build fails by 2 reasons: uuid absence from e2fsprogs and mkfs.jffs2 absence (only for ben-nanonote) Dec 18 08:05:00 that was about angstrom-2008.1 Dec 18 08:05:25 tiff was failed on minimal/akita Dec 18 08:08:16 hm.. Dec 18 08:08:31 all logs on tinderbox are truncated Dec 18 08:13:40 hm.. we have problems with git Dec 18 08:13:55 git pull hangs as ant_ was reported Dec 18 08:14:31 * JaMa did git pull OK (git+ssh://) Dec 18 08:29:11 hi Dec 18 08:29:17 Jay7: same problem here with git clone Dec 18 08:33:15 * Jay7 have reported e2fsprogs failure to ML Dec 18 08:33:32 git:// or http://? Dec 18 08:33:49 Jay7: git:// Dec 18 08:33:58 now trying http Dec 18 08:34:02 eFfeM: morning Dec 18 08:34:02 on another host I also cannot pull, but here it still works Dec 18 08:34:05 gm Dec 18 08:34:08 gm Dec 18 08:34:10 hi eFfeM Dec 18 08:34:11 oops Dec 18 08:34:16 eFfeM: have replied to your post about mkfs.jffs2 Dec 18 08:34:23 I'm affected too Dec 18 08:34:31 seems we have similar targets :) Dec 18 08:35:19 ericben: http://paste.pocoo.org/show/306300/ still works Dec 18 08:35:40 Jay7: thanks, I worked around it by manually building mtd-utils-native Dec 18 08:35:55 good that the ML is up again Dec 18 08:36:24 JaMa: it fails when not using git+ssh Dec 18 08:36:58 JaMa: or at least when doing a fresh clone as anonymous Dec 18 08:37:13 03Koen Kooi  07org.openembedded.dev * r7e7399ec33 10openembedded.git/recipes/e2fsprogs/e2fsprogs.inc: Dec 18 08:37:13 e2fsprogs: fix do_install breakages Dec 18 08:37:13 there is no 'uuid' binary, only 'uuidd': Dec 18 08:37:13 koen@dominion:/OE/org.openembedded.dev/recipes/e2fsprogs$ ls /OE/angstrom-dev/work/armv7a-angstrom-linux-gnueabi/e2fsprogs-1.41.12-r26/image/sbin Dec 18 08:37:13 badblocks* dumpe2fs* e2fsck.e2fsprogs* e2label* filefrag* fsck.ext3* fsck.ext4dev* mke2fs.e2fsprogs* mkfs.ext3* mkfs.ext4dev* resize2fs* uuidd* Dec 18 08:37:13 debugfs* e2freefrag* e2image* e2undo* fsck.ext2* fsck.ext4* logsave* mkfs.ext2* mkfs.ext4* mklost+found* tune2fs* Dec 18 08:37:14 Signed-off-by: Koen Kooi Dec 18 08:37:30 holy koen Dec 18 08:38:28 that's a mighty long commit message for koen :-) Dec 18 08:39:20 well.. but I have another 3 strange errors.. Dec 18 08:39:44 I should recheck this Dec 18 08:42:02 eFfeM: btw, have you noticed that logs on tinderbox was truncated last time? Dec 18 08:42:19 all my are except uuid's Dec 18 08:42:20 eFfeM: thanks for mpd Dec 18 08:43:43 JaMa: tinderbox has stange behaviour sometime Dec 18 08:44:00 ericben: I would say most of time :) Dec 18 08:44:07 even on my setup (far less used than on OE) I found problems Dec 18 08:44:23 I solved most of them by tweaking the web server configuration Dec 18 08:44:24 ericben, I just updated some things on melo...if its acting really strange we can reboot it. Dec 18 08:45:22 Jay7: I had some mtd-utils .o and binaries in my recipes dir yesterday no idea how they came here, I'll see what the autobuilder says about it Dec 18 08:45:38 eFfeM: that was known problem with bitbake Dec 18 08:45:42 JaMa: was easy, just copied the recipe, updated checksums and it works Dec 18 08:45:43 it was fixed already Dec 18 08:46:16 Jay7: then maybe that caused it to think that mtd utils was build but could not find the binary Dec 18 08:46:21 ka6sox: actual problem is git clone which doesn't work Dec 18 08:46:29 eFfeM: well.. may be Dec 18 08:46:40 ka6sox: if you have time to look at it that would be great Dec 18 08:46:44 ericben, cloning from where? Dec 18 08:46:46 eFfeM: was your build with that error clean or incremental? Dec 18 08:47:34 incremental, it was my system at home Dec 18 08:47:35 if its ours then I need to fix that. Dec 18 08:47:57 home builds are incremental, work builds are often on the autobuilder and started from scratch Dec 18 08:48:31 ka6sox: problem with our main git Dec 18 08:48:42 git.oe.org Dec 18 08:49:06 check please that git-daemon is running Dec 18 08:49:11 aha. Dec 18 08:49:13 jas. Dec 18 08:51:33 okay check now pleas? Dec 18 08:51:58 working Dec 18 08:52:02 10x Dec 18 08:57:06 Jay7, yw Dec 18 08:57:43 so now we wait till tinderbox attempts to build something. Dec 18 09:04:04 ka6sox: thanks, clone works fine now Dec 18 09:04:18 ericben, np... Dec 18 09:21:55 ERROR: Task /home/ebenard/oetest/testbuilder/openembedded/recipes/images/qt4e-demo-image.bb (do_rm_work) has circular dependency on /home/ebenard/oetest/testbuilder/openembedded/recipes/images/qt4e-demo-image.bb (do_rm_work_all) Dec 18 09:21:59 strange Dec 18 09:22:11 seems a random problem as it worked fine yesterday Dec 18 09:22:45 there is a lot of random problems Dec 18 09:23:16 looks like race sometimes Dec 18 09:23:51 Jay7: seems reproducable I got it again Dec 18 09:24:56 ericben: git bisect is your friend again :) Dec 18 09:29:12 ERROR: Error parsing /storage/oe/openembedded/recipes/freesmartphone/frameworkd_git.bb: database is locked Dec 18 09:29:13 o_O Dec 18 09:51:14 hm, buidling with bitbake head gives me Dec 18 09:51:15 old_fds.append(os.dup(fd)) Dec 18 09:51:15 OSError: [Errno 9] Bad file descriptor Dec 18 09:52:40 no! I don't like to see that error here! :) Dec 18 09:55:03 i moved back quite a bit in bb git and then the problem seems gone Dec 18 09:55:16 will report later, need to go no Dec 18 09:55:17 now Dec 18 10:29:39 Jay7: after trying to bisect the problem, I cleaned all the driectories and started again and ... no more problem ... Dec 18 11:07:08 ericben|away: heh.. Dec 18 11:07:56 ericben|away: btw, I've added builds interactive shell mode Dec 18 11:08:05 will commit after some testing Dec 18 11:08:34 this allow to get shell with all env preset for some of configured build Dec 18 11:09:15 eFfeM: I've easy fix for mkfs.jffs2 bug Dec 18 11:09:25 just clean mtd-utils[-native] Dec 18 11:09:37 have cleaned both and rerun build Dec 18 11:09:42 and it was succeed Dec 18 11:10:19 I think I understand from where problem was got Dec 18 11:10:35 we have build mtd-utils with old unfixed bitbake Dec 18 11:10:56 so it is market as built but files are out of right directory Dec 18 11:11:08 clean build should be fine then Dec 18 11:12:21 Jay7: correct that is also my inderstanding Dec 18 11:12:35 but bitbake git head does not seem to work for me Dec 18 11:12:36 well.. mtd-utils-native only Dec 18 11:18:37 Jay7: also noticed that see also the msg of grg on ML Dec 18 11:18:46 btw bitbake head also does not work for me, reported to ML Dec 18 11:25:42 wow, gtk+-2.20.1 is now failed :) Dec 18 11:25:46 minimal/akita Dec 18 11:26:25 strange that is was fine for collie.. Dec 18 11:30:28 gtk+ failure: http://tinderbox.openembedded.net/packages/1367691/ Dec 18 11:30:38 will check with clean build Dec 18 11:35:34 hi, is set sysroot the only thing to do with cross-gdb + gdbserver? Dec 18 11:35:50 I've that: Dec 18 11:35:51 warning: no loadable sections found in added symbol-file /home/gnutoo/embedded/sshfs/usr/lib/.debug/libz.so.1.2.3 Dec 18 11:37:34 gdb-cross runs on my laptop wchih is also the build system Dec 18 11:37:53 and the "gdb sysroot" is available trough sshfs Dec 18 11:46:50 howto clean ? Dec 18 11:47:13 or clean cache Dec 18 11:49:05 touch local.conf ? Dec 18 11:53:57 echo > GNUtoo|laptop Dec 18 11:54:14 thats is nice :) Dec 18 11:59:19 hm.. stupid question - can I safely remove $TMPDIR/work directory? Dec 18 11:59:43 will this affect next incremental build badly? Dec 18 12:00:25 it shouldn't imho Dec 18 12:01:59 Jay7 Yes Dec 18 12:03:09 l1nux, :) Dec 18 12:12:15 Jay7, do you use autorev? Dec 18 12:12:44 ah only the work Dec 18 12:12:48 not the whole tmpdir Dec 18 12:12:57 I think it'll do something like rm_work Dec 18 12:15:31 heh small meego upgrade :) Overall download size: 219.1 MiB. After the operation, additional 378.8 MiB will be used. Dec 18 12:16:06 ups that should go to #openmoko-cdevel Dec 18 12:22:43 * Jay7 is gone for some long time Dec 18 14:42:20 kergoth_: commit 416dfe211ff0c37d746efe113bb6064b7e57d43 fails for me, see ML Dec 18 14:53:52 maybe i don't understand a thing ... weekly a tag is created into testing-next branch ... so what is the testing branch for ? Dec 18 14:55:16 it is an easy point for testers to sync on, for my testing I pull from this branch, it is not intended to commit to this branch Dec 18 14:55:35 is the ML live now then? :) Dec 18 14:55:45 eFfeM: with bitbake master try 'bitbake -u goggle xyz' Dec 18 14:56:19 the new UI seems working Dec 18 14:56:44 Tartarus: ML works Dec 18 14:56:55 ant__: bitbake master git head gives me errors, see ML Dec 18 14:57:05 try the new UI Dec 18 14:57:32 eFfeM, thanks ... i'm trying to build but after 2010.12 release date i didn't find any branch that work for my target :( ... (avr32) ... Dec 18 14:57:42 kergoth will repair later the old-good console Dec 18 14:58:01 ant__: is xyz the recipe name Dec 18 14:58:26 yes, or a task Dec 18 14:59:15 I'm pretty sure I would reproduce your breakage here so I tried straight the UI way ;) Dec 18 14:59:29 ant__: kergoth_ no success, I briefly get a window then it stops again, still get the errno 9 Dec 18 14:59:34 afk for 15-20 mins Dec 18 15:03:20 kergoth_, foerster : the new UI is impressive but somehow I'd prefer to have stout/stderr somewhere Dec 18 15:11:58 re Dec 18 15:13:06 trying a fresh clone to see if it is better Dec 18 15:13:38 no difference Dec 18 15:15:02 Try a git bisect yet? Dec 18 15:15:20 Dunno if kergoth is up yet, but he always asks me if I've done one yet when I break thigns for him :) Dec 18 15:16:53 Tartarus: yes, see above commit 416dfe211ff0c37d746efe113bb6064b7e57d43 apparently it cannot dup an fd (EBADF) Dec 18 15:17:10 Oh, so you've narrowed it down to a commit, good :) Dec 18 15:18:23 hello Dec 18 15:18:32 * khem yawns Dec 18 15:19:15 bye Dec 18 15:26:35 khem yawns away people :-) Dec 18 15:26:48 heh Dec 18 15:26:51 seems so Dec 18 15:35:01 eFfeM: here bitbake master works on Gentoo (python-2.6 & 3.1) 32 bit. pygtk-2.17 is installed Dec 18 15:42:32 this is opensuse 11.2 Dec 18 15:44:31 no idea what pygtk i have Dec 18 15:49:38 khem, tried minimal on autobuilder, sane-toolchain says libtool 2.4 and HAS SYSROOT true, but for sheeva I still get a build fail on testing-next Dec 18 15:49:54 what fails Dec 18 15:50:22 khem, libtool in do_install Dec 18 15:50:22 /usr/lib/libncurses.so: file not recognized: File format not recognized Dec 18 15:51:23 then arm-oe-linux-gnueabi-libtool: install: error: relink `libgettextsrc.la' with the above command before installing it Dec 18 15:52:27 hmmm Dec 18 15:52:34 thats gettext ? Dec 18 15:52:42 yes Dec 18 15:53:08 it seemed the problem was gone with enabling sysroot Dec 18 15:53:14 and this was on a clean build Dec 18 15:53:49 retrying from head, if needed will kick testing again Dec 18 15:54:32 eFfeM: in testing next you still need to set HAS SYSROOT manually or cherry pick 7ea9be2b5d3ac4fb5da74506c2516b3412726495 Dec 18 15:54:53 JaMa: ah ok, thought that was in testing Dec 18 15:55:44 unfortunatelly I've pushed it too late for testing_next, so there will be quite a lot of failures :/ Dec 18 15:56:36 ant__: does it work after you actually set python to 3.1 with eselect python? Dec 18 15:56:48 bitbake won't work with python 3 right now Dec 18 15:56:52 just wont Dec 18 15:57:12 ant__: that's why I'm surprised by what ant__ said Dec 18 15:57:16 kergoth I'm using python 2.6.2 Dec 18 15:57:35 i started a branch to start working on python3, but there's a fair bit more to be done Dec 18 15:57:41 just did the low hanging fruit Dec 18 15:57:42 Tartarus: wrt commit fa6b202f9c954df43aa26ec8c9fa8b717d4074c1 ideally in the native case it should not have a depends on ncurses Dec 18 15:57:44 I've even tried it with patches from your python-3 branch Dec 18 15:57:52 yeah, that wont be enough :) Dec 18 15:57:58 hi, what's the advised way to debug a program built with openembedded, because I've that issue: Dec 18 15:58:00 we need to get it working at some point here Dec 18 15:58:13 especially since some distros are shipping with it as default now (even if they are the crazy ones ;) Dec 18 15:58:17 warning: no loadable sections found in added symbol-file /home/gnutoo/embedded/sshfs/usr/lib/.debug/libz.so.1.2.3 Dec 18 15:58:21 here's my setup: Dec 18 15:58:24 yup you said it to me about week ago so I've stopped trying :) Dec 18 15:58:26 gdb-cross+gdbserver Dec 18 15:58:30 kergoth any idea on my problem with 416dfe211ff0c37d746efe113bb6064b7e57d43? Dec 18 15:58:31 hehe Dec 18 15:58:34 filesystem exported trough sshfs Dec 18 15:58:42 and set sysroot /path/to/sshfs/rootfs Dec 18 15:59:02 eFfeM: nope. all it does is changes stdin/stdout/stderr to /dev/null, i don't see how its possible to mess that up Dec 18 15:59:12 eFfeM: though i did have the wrong file mode in that commit, its fixed in a later one Dec 18 15:59:16 make sure you've tested with latest master Dec 18 16:00:08 * kergoth afk Dec 18 16:01:04 eFfeM, thanks for fixing mpd Dec 18 16:01:25 kergoth master head also has the problem Dec 18 16:01:32 GNUtoo|laptop: np, my pleasure Dec 18 16:03:11 JaMa: thought what was in the .inc file for gettext was enough, will set sysroot usage explicitly Dec 18 16:03:37 GNUtoo|laptop: hmmm Dec 18 16:04:20 eFfeM: which .inc? Dec 18 16:04:59 sane-toolchain.inc Dec 18 16:05:21 just saw your commit message, i thought what was in there was enough Dec 18 16:05:50 so after seeing that commit it's clear why it wasn't enough before, right? Dec 18 16:05:53 GNUtoo|laptop: it seems that in addition to .debug more got stripped from libz Dec 18 16:05:56 and it does not matter that it was too late for the testing, this round is a total fail anyway due to the e2fsprogs breakage (which was fixed by koen earlier today) Dec 18 16:06:08 JaMa: do u have libz so in the rfs ? Dec 18 16:06:39 khem, that's not what #gdb told me some time ago Dec 18 16:06:45 JaMa: yes, it is understood, missed the bitbake.conf setting (actually I think we should probably move to 2.4 as the default) Dec 18 16:06:54 khem whats .la file in sysroot/target should now looks like? Dec 18 16:06:56 khem, they told me gdb was trying to load the .debug library as a normal lib Dec 18 16:07:00 khem: yes -rwxr-xr-x 1 root root 51K Dec 7 13:59 /usr/lib/libz.so.1.2.3 Dec 18 16:07:07 but I think your version seem more realistic Dec 18 16:07:42 JaMa: ok can you pastebin readelf -e output on it ? Dec 18 16:07:48 sure, mmt Dec 18 16:07:54 khem, yes of course Dec 18 16:08:01 woglinde: they should have =/usr/lib etc. Dec 18 16:08:05 JaMa: no, still on python-2.6 as (sane) default Dec 18 16:08:06 khem, and it's not the only lib, a lot of other libs are like this Dec 18 16:08:20 woglinde: which means libtool will append sysroot to those libdirs Dec 18 16:08:34 GNUtoo|laptop: I am not surpised Dec 18 16:08:39 I still have /usr/lib in it Dec 18 16:08:47 khem, what could it be? Dec 18 16:08:54 am I doing something wrong? Dec 18 16:09:01 GNUtoo|laptop: stripped out symbol tables Dec 18 16:09:32 hm maybee I should build from scratch Dec 18 16:09:43 khem, ah ok Dec 18 16:09:46 GNUtoo|laptop: JaMa as a finger excercise (if that is good english) I'm trying to see if I can get mpd 0.16 to compile under poky Dec 18 16:09:47 how can I fix it? Dec 18 16:10:06 eFfeM, what about getting gmpc to compile? Dec 18 16:10:13 khem: http://paste.pocoo.org/show/306451/ Dec 18 16:10:36 khem could you please pastebin libz.la? Dec 18 16:10:44 GNUtoo|laptop: I thought it did compile Dec 18 16:10:56 eFfeM, maybe it doesn't anymore Dec 18 16:11:24 eFfeM, http://pastebin.com/MDm4Mpqq Dec 18 16:11:43 JaMa: ok so as expected its missing .symtab Dec 18 16:11:45 03Martin Jansa  07master * re667a1fb42 10openembedded.git/ (2 files in 2 dirs): Dec 18 16:11:45 sane-toolchain-eglibc: remove ggdb3 work around, those alignment traps are gone now Dec 18 16:11:45 Signed-off-by: Martin Jansa Dec 18 16:11:52 GNUtoo|laptop: ^^ that should make debugging a bit easier again, eglibc-dbg is small again :) Dec 18 16:12:01 JaMa: now pastebin readelf -e of same from .debug Dec 18 16:12:07 ok I'll paste libz.la Dec 18 16:12:59 khem: libz-dbg http://paste.pocoo.org/show/306454/ Dec 18 16:14:11 http://pastebin.com/fUhR4Tm6 my libz.la Dec 18 16:15:14 JaMa: ok so that seems to be ok to me Dec 18 16:15:50 GNUtoo|laptop: what happens if you use gdb on target Dec 18 16:15:52 does it work ? Dec 18 16:16:07 I have to set sysroot to / Dec 18 16:16:12 and it does exactly the same Dec 18 16:16:14 same error Dec 18 16:16:15 etc... Dec 18 16:16:22 if I don't set sysroot to / Dec 18 16:16:26 hmmm on target too ? Dec 18 16:16:27 it says: Dec 18 16:16:51 warning: Unable to find dynamic linker breakpoint function. Dec 18 16:16:55 and fails with: Dec 18 16:17:04 /usr/lib/libfsotransport.so.0: No such file or directory. Dec 18 16:17:06 etc... Dec 18 16:17:16 lots of libs like that Dec 18 16:17:42 sorry Dec 18 16:17:48 didnt saw the pastebin Dec 18 16:18:13 GNUtoo|laptop: doing a bb -cclean gmpc, bb gmpc, but for now a bb gmpc said: Dec 18 16:18:14 NOTE: package gmpc-0.20.0-r0: task do_rm_work_all: Succeeded Dec 18 16:18:14 NOTE: Tasks Summary: Attempted 2980 tasks of which 2975 didn't need to be rerun and 0 failed. Dec 18 16:18:37 eFfeM, ah ok so the problem is on my side only Dec 18 16:19:01 GNUtoo|laptop: seems like ld.so is stripped Dec 18 16:19:12 ah ok ouch Dec 18 16:19:18 JaMa: can you post same readelf for ld.so Dec 18 16:19:35 GNUtoo|laptop: from the log and these lines: | ./gmpc-meta-watcher.h:57:17: error: expected '=', ',', ';', 'asm' or '__attribute__' before '*' token Dec 18 16:19:45 I'd say it is missing a type and probably a .h file Dec 18 16:20:02 ok I'll look Dec 18 16:20:18 khem: I've a bit different ld.so than what GNUtoo|laptop has (with my last sane-toolchain-eglibc commit) Dec 18 16:20:53 GNUtoo|laptop: I recall having roughly the same error message on 0.17.96 so I added 0.20 and that worked for me Dec 18 16:20:54 hmmm then GNUtoo|laptop you can provide it Dec 18 16:20:57 might have a clean build since Dec 18 16:21:03 khem: but fwiw http://paste.pocoo.org/show/306459/ Dec 18 16:21:08 might have done a clean build since Dec 18 16:21:24 what should I do? Dec 18 16:21:38 arm-readelf --all ld.so? Dec 18 16:21:50 JaMa: doesnt look good Dec 18 16:21:59 JaMa: do u have .debug for it too ? Dec 18 16:22:03 yes Dec 18 16:22:40 khem: http://paste.pocoo.org/show/306462/ Dec 18 16:22:43 GNUtoo|laptop: readelf -e /lib/ld-linux* Dec 18 16:22:49 on target Dec 18 16:23:17 JaMa: ok I understand the proble Dec 18 16:23:19 m Dec 18 16:23:47 khem: ah I also have newer binutils than GNUtoo|laptop build Dec 18 16:23:47 now I know why my son was so quite yesterday he hammeted the 'm' key on this machine Dec 18 16:23:51 hm lightspark has now ppc support Dec 18 16:24:13 what is lightspark Dec 18 16:24:22 nice Dec 18 16:24:30 JaMa: I think something changed in the way Dec 18 16:24:32 a free flash player Dec 18 16:24:34 like gnash Dec 18 16:24:41 khem alternative flashplayer Dec 18 16:24:47 with some intressting stuff Dec 18 16:25:01 flash is dead people should use HTML5 Dec 18 16:25:14 tell this the video portals Dec 18 16:25:24 khem: I'm using binutils-2.21 locally, GNUtoo|laptop build has probably 2.20.1 if he is using SHR builds or build SHR as is Dec 18 16:25:43 looking which version it'll install Dec 18 16:25:50 Installing binutils (2.20.1-r13.4.6) to root... Dec 18 16:26:10 the problem with the video portals is that they insist on wanting DRM Dec 18 16:26:20 yup whole image was built with that version Dec 18 16:26:36 GNUtoo|laptop: dou have an image where it worked ? Dec 18 16:26:52 no, last time I debugged someting was some years ago Dec 18 16:26:58 with wesnoth Dec 18 16:27:34 http://pastebin.com/nuXATd6Z Dec 18 16:28:29 GNUtoo|laptop: I understand the problem Dec 18 16:28:47 I will see if I can fix it Dec 18 16:28:55 today or tomorrow Dec 18 16:29:04 nice!!!! Dec 18 16:29:07 thanks a lot!!!!! Dec 18 16:29:29 * GNUtoo|laptop still has to read the elf introduction Dec 18 16:33:59 GNUtoo|laptop: gmpc rebuild was fine Dec 18 16:34:11 * eFfeM scurries away to get some food Dec 18 16:34:40 eFfeM, ok Dec 18 16:34:47 * GNUtoo|laptop will look Dec 18 16:40:31 hmmm I don't find typedef for GmpcMetaWatcher in the code Dec 18 16:41:46 what is .gob? Dec 18 16:42:14 kergoth, foerster: about the goggle UI, how I'm I supposed to exit? CTR C ? Dec 18 16:42:16 glib meta-object stuff? Dec 18 16:43:24 kergoth, foerster;: I mean at the end, after successful build (Build completed is the first line of the UI) Dec 18 16:48:57 ant__: click the exit button at the top of the window Dec 18 16:49:18 fwiw there are 2 Sl+ processes Dec 18 16:49:28 after exiting? Dec 18 16:49:34 before Dec 18 16:49:57 strabge, I have BB_NUMBER_THREADS = 4 Dec 18 16:50:26 number threads of for runqueue execution Dec 18 16:50:38 the ui is in a process, the server is in a process Dec 18 16:50:49 the server may spawn up to BB_NUMBER_THREADS workers Dec 18 16:51:03 ah, up to Dec 18 16:52:22 I noiced that some tasks like do_unpack seem to left the log aaround Dec 18 16:53:01 foerster: any idea on my bb issue (bad file descr) Dec 18 16:53:21 eFfeM: no, that's odd. kergoth did all of that stuff. Dec 18 16:53:44 he is setting stdin/out/err to /dev/null (which was opened as append) Dec 18 16:53:54 are you running on a normal terminal? Dec 18 16:53:56 eFfeM: fwiw this mornig I could not pull bitbake master, I did it afternoon Dec 18 16:54:23 try to clone it again, just in case Dec 18 16:54:23 foerster: yes Dec 18 16:54:50 foerster: gnome terminal 2.28.0 Dec 18 16:55:04 and bash, Dec 18 16:55:07 eFfeM: what distro? Dec 18 16:55:12 opensuse 11.2 Dec 18 16:55:14 * kergoth 's never seen this, or heard this Dec 18 16:55:24 hmm Dec 18 16:55:43 gnome-terminal works fine here Dec 18 16:56:02 well I'm not a python wiz but I'm more than happy to add some python code to see which file descriptor it is (if you paste me the lines) Dec 18 16:56:18 2.30.2 fwiw Dec 18 16:56:20 i know what file descriptor it is. Dec 18 16:56:28 its redirecting sys.stdout, sys.stdin, and sys.stderr to /dev/null Dec 18 16:56:29 its that simple Dec 18 16:57:21 i know but it might be it is only for one of those Dec 18 16:57:47 JaMa|Off: Did you sort out the gitpv patchset from otavio Dec 18 16:58:00 when you do please cleanup patchwork too Dec 18 16:58:59 ciao Dec 18 16:59:07 kergoth, foerster: well, the "Unpacking.." string is not removed after task do_unpoack ends Dec 18 16:59:18 kergoth if I redirect stdout and stderr to a file I still get the error Dec 18 16:59:19 why would it be? Dec 18 16:59:27 eFfeM: and stdin? Dec 18 17:00:26 kergoth bitbake helloworld < /dev/null also gives the error Dec 18 17:00:33 dunno Dec 18 17:00:38 works fine for everyone else, i have no idea Dec 18 17:00:39 and also if I use a file Dec 18 17:00:42 * kergoth re-reads the code Dec 18 17:01:37 well the dup(2) man page says: EBADF oldfd isn't an open file descriptor, or newfd is out of the allowed range for file descriptors. Dec 18 17:01:51 khem: only acked, I expect otavio to push it himself and cleanup it when he do Dec 18 17:02:26 hmmm Dec 18 17:02:27 ok even better Dec 18 17:02:40 kergoth: why do we have rm_work_all ? Dec 18 17:02:49 the task does nothing as such Dec 18 17:02:51 khem: because of recrdeptask Dec 18 17:03:02 you can't dep on yourself Dec 18 17:03:07 ugh right Dec 18 17:03:08 iirc, anyway Dec 18 17:03:21 eFfeM: i have an idea, hold a second and i'll get you a patch to test Dec 18 17:03:46 why cant be just inserm rm_work between two tasks in chain Dec 18 17:04:02 or is rm_work the last task ? Dec 18 17:04:03 because rm_work runs after build, to make sure its truly after every other task that may exist Dec 18 17:04:13 hmm Dec 18 17:04:27 eFfeM: what python version? Dec 18 17:04:32 so basically when it comes to tasks of same recipe its the last in the list Dec 18 17:04:59 yep Dec 18 17:05:39 kergoth: so how does rm_work_all help ? is is a global task which then rdeps on individual rm_work Dec 18 17:07:22 eFfeM: http://pastebin.com/QU9fHTqA Dec 18 17:07:31 that will at least print some info (hopefully) Dec 18 17:08:24 kergoth: I was trying to read rm_work Dec 18 17:08:27 bbclass Dec 18 17:08:36 eFfeM: that goes near the end in lib/bb/utils.py, line 805 Dec 18 17:08:46 rm -rf $dir may be we can use rm -rf $dir& Dec 18 17:10:07 eFfeM: apply http://kergoth.pastey.net/144033 and run again if you would Dec 18 17:10:16 foerster: heh :) Dec 18 17:10:40 argh...wiki / main site seems not responding again... Dec 18 17:10:52 kergoth: yours is much more elaborate :) Dec 18 17:11:16 with mine his build will continue, just with warnings :) Dec 18 17:11:45 foerster: thinking about dropping none/xmlrpc, can always resurrect from history if necessary, but they're currently incompatible with the UIs Dec 18 17:11:46 heh Dec 18 17:11:48 eFfeM, do you know how glib stuff works? Dec 18 17:11:51 * kergoth does locally Dec 18 17:12:13 how i can solve the bitbake stopping after parsing bug ? Dec 18 17:12:41 kergoth: where are looged thing like" cp: cannot stat `/oe/build/tmp/work/i686-linux/stagemanager-native-0.0.1-r15/sysroot-destdir//oe/build/tmp/*': No such file or directory" ? Dec 18 17:12:58 with new UI? Dec 18 17:13:12 ant__: that won't be seen anymore with current master, or shouldn't be. that was a python function using os.system, we kill off all stdout/stderr from the server now, if the python function wants to have output, it should use bb.msg Dec 18 17:13:18 eFfeM, beginning of src/gmpc-meta-watcher.gob http://pastebin.com/4dUZX4bj Dec 18 17:13:27 ant__, talking to me ? Dec 18 17:13:50 eFfeM, and the error : | ./gmpc-meta-watcher.h:57:17: error: expected '=', ',', ';', 'asm' or '__attribute__' before '*' token Dec 18 17:14:11 and ./gmpc-meta-watcher.h +57 has: Dec 18 17:14:20 gremlin[it]: no, is abou goggle UI Dec 18 17:14:32 (and old broken keys on kb :) Dec 18 17:14:38 GmpcMetaWatcher * gmpc_meta_watcher_new (void); Dec 18 17:14:41 03Chris Larson  07master * r23c6122437 10bitbake.git/lib/bb/server/ (none.py xmlrpc.py): Dec 18 17:14:41 server: drop none and xmlrpc Dec 18 17:14:41 These are no longer used, and are no longer compatible with the current UIs. Dec 18 17:14:41 Signed-off-by: Chris Larson Dec 18 17:14:45 03Richard Purdie  07master * rde2368ce38 10bitbake.git/lib/bb/ (build.py data_smart.py): Dec 18 17:14:45 build: Add support for pre and postfuncs for tasks Dec 18 17:14:45 Signed-off-by: Richard Purdie Dec 18 17:14:45 Signed-off-by: Chris Larson Dec 18 17:15:16 kergoth: there are other things - like all UIs referencing xmlrpclib for exceptions, etc Dec 18 17:15:22 those can all be removed Dec 18 17:15:25 * kergoth nods Dec 18 17:15:38 those are useless whether the old servers are in the repo or not :) Dec 18 17:15:48 right, since they're not used at all Dec 18 17:16:24 gotta go, later Dec 18 17:17:03 kergoth: shouldn't the QA at least appear? Dec 18 17:17:27 that depends on what the class is doing Dec 18 17:17:44 afaik those all use bb.note, etc, which works fine Dec 18 17:17:49 sry was afk Dec 18 17:17:54 gremlin[it]: you have something in ASSUME_PROVIDED and trying to set PREFERRED_VERSION for same Dec 18 17:18:12 bitbake goes into endless loop in such a case Dec 18 17:18:30 python 2.6.2 Dec 18 17:19:24 khem, cpu is low while it "wait" ... pressing CRTL-C make in continue ... but is really annoying Dec 18 17:20:13 kergoth with your pathc I get Dec 18 17:20:14 WARNING: ', mode '' at 0x8454020> (fd 0) could not be duplicated: [Errno 9] Bad file descriptor Dec 18 17:20:15 kergoth: heh, I was building console-image which is apparently QA-free :) Dec 18 17:20:41 GNUtoo|laptop: no glib knowledge here Dec 18 17:20:47 ok Dec 18 17:20:54 eFfeM: so your stdin got closed. Dec 18 17:20:57 strange Dec 18 17:21:27 * kergoth is using 2.6.6 Dec 18 17:21:41 GNUtoo|laptop: this is because you do not have GmpcMetaWatcher Dec 18 17:21:50 seems similar wiht what I have with the previous version Dec 18 17:22:16 what''s GmpcMetaWatcher exactly? Dec 18 17:22:37 ant__: well, unpack and patch messages work fine, so if the qa messages aren't, then the classes need to be fixd Dec 18 17:22:58 let see with cups and other opie-stuff, building right now... Dec 18 17:23:01 i think it is a generated file, I also did someting with gob2 you might want to see if I have that patch Dec 18 17:23:50 kergoth with your patch the stuff parses then spits out the build configuration then it hangs Dec 18 17:24:43 03Chris Larson  07master * r8672abb230 10bitbake.git/lib/bb/server/process.py: Dec 18 17:24:43 server: fix NULL file mode Dec 18 17:24:43 Signed-off-by: Chris Larson Dec 18 17:25:05 don't know what to tell you. i can't reproduce anything anywhere near like what you're seeing Dec 18 17:25:12 maybe its a bug in multiprocessing in old python Dec 18 17:25:15 * kergoth looks for 2.6 changelog Dec 18 17:27:16 kergoth i think you gave me an argument to upgrade to 11.3, guess that will bring me a newer python Dec 18 17:27:33 http://svn.python.org/view?view=rev&revision=73708 looks interesting Dec 18 17:27:46 something about a double close of stdin resulting in bad file descriptor.. Dec 18 17:27:55 http://bugs.python.org/issue6337 Dec 18 17:28:04 not sure if thats it, but it sounds related Dec 18 17:28:54 trying Dec 18 17:29:29 if it turns out that fixes it, might want to bump the minimum python version in master to 2.6. instead of 2.6 Dec 18 17:31:28 kergoth i have bug 6337 Dec 18 17:31:39 so guess we nailed it down Dec 18 17:31:53 going to update to 11.3 Dec 18 17:33:35 cool. any idea which python version that fix made it into? Dec 18 17:36:08 nope Dec 18 17:36:49 but it was apparenlty fixed on jun 24 2009 Dec 18 17:37:12 kergoth: no, log.qa_package is displayed even if it exists Dec 18 17:37:18 is not Dec 18 17:37:28 i know, why would it be? Dec 18 17:37:36 the messages go both to the ui *and* to the log for python functions Dec 18 17:37:41 they aren't hidden from the ui Dec 18 17:37:53 so showing it again would just be showing you the same thing twice Dec 18 17:38:38 imho those are worth a double evidence, on file and on screen Dec 18 17:39:20 who would dig in workdir to search QA? Dec 18 17:40:25 kergoth: wat could be filtered are the annoying 'Angstrom doesn't support ..' instead ;) Dec 18 17:40:39 ant__: i don't feel like being spammed, thanks. Dec 18 17:40:52 oh, is saturday, sorry Dec 18 17:40:53 yes, that's on angstrom, talk to them Dec 18 17:40:53 RIP Dec 18 17:41:08 i didn't say you had to dig in workdir to find qa Dec 18 17:41:09 (or move to minimal :P) Dec 18 17:41:11 are you really this dense? Dec 18 17:41:21 if the task is using correct output mechanisms, the messages are output during the build Dec 18 17:41:27 if not, they aren't, and the task should be fixed. Dec 18 17:41:31 it's that simple Dec 18 17:44:18 * kergoth reviews qa classes Dec 18 17:46:08 hrm, wonder if we should revert the kill of stdin/stdout/stderr for this release and postpone it till the next, given it does silence crappy classes Dec 18 17:48:47 kergoth that will also keep down complaints from people with old python versions Dec 18 17:49:14 it'll mean our non-default UIs will be not particularly useful for the release Dec 18 17:49:25 since some messages will go to the UI and others to the console you ran the ui from Dec 18 17:49:56 add a config param to redirect ? Dec 18 17:50:02 or option Dec 18 17:54:25 03Chris Larson  07master * rc3aec3e887 10bitbake.git/lib/bb/ (server/process.py utils.py): Dec 18 17:54:25 Revert "server: kill stdin/stdout/stderr" Dec 18 17:54:25 This is a useful thing, but requires that tasks behave correctly, sending Dec 18 17:54:25 their output via the usual bitbake messaging mechanisms, rather than via Dec 18 17:54:25 stdout. We'll postpone this until after the release. Dec 18 17:54:25 Signed-off-by: Chris Larson Dec 18 18:01:00 03Chris Larson  07master * r24857e2ceb 10bitbake.git/lib/bb/shell.py: Dec 18 18:01:00 shell: remove Dec 18 18:01:00 bb.shell is unused and incompatible with current bitbake. We can resurrect it Dec 18 18:01:00 in a more functional form when we add the text input to the goggle and ncurses Dec 18 18:01:00 UIs. Dec 18 18:01:01 Signed-off-by: Chris Larson Dec 18 18:02:00 * kergoth started a 'next' branch with the bits we want to hold off on until the release Dec 18 18:11:57 I'm trying to find the sources for a kernel on a Gumstix Overo board. Its running some OE based Angstrom and have ipkg at my disposal. Can you get the sources with it, similar to apt-get source? Dec 18 18:18:10 sveinse: well you need to checkout overo repo Dec 18 18:20:32 khem: OK, I found a release note saying "b943494aa13db729684eb9eb760ae7d04460821f Merge branch 'overo' of git://gitorious.org/gumstix-oe/mainline into overo" Dec 18 18:20:48 I guess its just to clone that one and go from there Dec 18 18:29:51 kergoth: quick thought - can't we capture the output of each task into a stringIO, convert it to a log message, and also warn about deprecation of printing to stdout? Dec 18 18:30:10 sveinse: it will download the sources onces you clone the metadata Dec 18 18:30:46 bitbake virtual/kernel will download src and build the kernel binary too Dec 18 18:31:01 you need to set up the build env Dec 18 18:31:12 go thru the getting started guide if you need help Dec 18 18:31:17 you can also ask here Dec 18 18:31:52 Its a rather lengthy operation since it needs to build a toolchain for me as well, right? Dec 18 19:33:47 eFfeM, what should I do then? you have a newer gmpc version? Dec 18 19:50:51 kergoth I might have one more bb issue, it remains stuck at parsing recipes, I've seen this earlier today one more time Dec 18 19:51:16 GNUtoo|laptop: I have the version which is in the repo Dec 18 19:52:25 ok Dec 18 19:52:28 me too Dec 18 19:52:44 GNUtoo|laptop: I moved gob2 to 2.0.17 maybe that also influenced it, see the git log of the .17 version Dec 18 19:52:58 gmpc_0.20.0.bb Dec 18 19:53:02 might be I did a rebuild I seem to recall those type files were generated by gob2 Dec 18 19:53:07 ok I'll pull and retry Dec 18 19:53:13 yes they were Dec 18 19:53:15 it log gob2_2.0.17.bb Dec 18 19:53:19 thanks a lot for the infos Dec 18 19:53:28 hope it helps Dec 18 20:06:43 sveinse, also opkg info gives where some source is Dec 18 20:09:49 still failing after rebuild of gob2 Dec 18 20:10:09 kergoth touched local.conf so it thought the cache was out of date, then it works properly again, also on the next cached build Dec 18 20:10:24 I'll look which one is picked Dec 18 20:10:48 not sure if gob2 helps I am by no means a glib wiz Dec 18 20:11:13 ha just what I tought looking at the recipes Dec 18 20:11:27 gob2_2.0.14.bb gob2_2.0.17.bb gob2-native_2.0.14.bb Dec 18 20:11:32 that's an ls in gob2 dir Dec 18 20:11:50 I bet the cause is obvious to you know Dec 18 20:12:05 but how to fix? Dec 18 20:12:21 remove gob2-native_2.0.14? Dec 18 20:12:30 so it pick the BBCLASS_EXTENDS? Dec 18 20:39:11 mhhh DISTRO="minimal" + LIBC="uclibc" is different than DISTRO="minimal-uclibc" ??? Dec 18 20:43:36 gremlin[it], look at the config files, but you've got to grep for overrides too Dec 18 20:45:48 foerster|away: StringIO won't work, os.system() writes directly to file descriptors, not python file objects. we'd have to write to the log file and also be reading the data from the log file as its written Dec 18 20:48:07 GNUtoo|laptop, seem isn't the same ... uclibc in the first way don't have largefile ... in the latter probably yes (I'm checking) ... Dec 18 20:48:28 ok Dec 18 20:51:45 GNUtoo|laptop, i don't like to have two to specify a distro in uclibc flavor with different results ... Dec 18 20:52:02 indeed.... Dec 18 22:39:56 I'm working on a Angstrom/OE Gumstix board. Are all files belonging to an opkg package? How can I find which package provides a specific file? Dec 18 22:40:32 opkg --help Dec 18 22:40:58 E.g. opkg search bash turns up empty... Dec 18 22:42:34 opkg update? Dec 18 22:43:01 Already run Dec 18 22:45:11 Specifically I'm trying to find which package, if any, provides /usr/bin/SGX/demos/ Dec 18 22:45:29 but I'm lost in the opkg world Dec 19 00:54:20 hmmm Dec 19 02:21:10 hey, can some point me to a simple interrupt handler that i can look at as an example? i have the LDD 3rd edition, but i am having a disconnect on how to write the handler and the return methods etc Dec 19 02:26:39 not on topic for this channel. this is specific to the OE project. #edev or #elinux would be a better bet, or even #kernelnewbies on oftc Dec 19 02:41:04 sorry Dec 19 02:54:58 anybody work with the realease shogo? it ships with angstrom and they provide kernel source and an angstrom-2.0b1-blah-armv6-blah sdk Dec 19 02:55:04 but no recipes / machine conf Dec 19 02:55:18 i figure they must have them kicking around somewhere **** ENDING LOGGING AT Sun Dec 19 02:59:57 2010