**** BEGIN LOGGING AT Wed Jul 14 02:59:56 2010 Jul 14 03:20:16 Tartarus: I am back Jul 14 03:25:45 03Graham Gower  07org.openembedded.dev * r5ce3d5e328 10openembedded.git/recipes/perl/perl_5.8.8.bb: (log message trimmed) Jul 14 03:25:45 perl_5.8.8.bb: Don't use hostperl's $Config{cppflags} with cross compiler. Jul 14 03:25:45 Fixes /usr/local/include Badness. Jul 14 03:25:45 make[2]: Entering directory `/home/grg/oe2/tmp/work/mipsel-oe-linux/perl-5.8.8-r34/perl-5.8.8/ext/Errno' Jul 14 03:25:45 hostperl -I/home/grg/oe2/tmp/work/mipsel-oe-linux/perl-5.8.8-r34/perl-5.8.8/Cross/../fake_config_library -MConfig Errno_pm.PL Errno.pm Jul 14 03:25:53 03Khem Raj  07org.openembedded.dev * re78de4e1cd 10openembedded.git/recipes/gtk+/ (4 files in 3 dirs): Jul 14 03:25:53 gtk+-2.20.0,gtk+-2.20.1: Fix case where /usr/bin/nm was being generated into libtool. Jul 14 03:25:53 Signed-off-by: Khem Raj Jul 14 03:28:02 * Tartarus is still not sure if it's compressed patches that're broken or what Jul 14 03:29:44 Tartarus: compressed patch should now be not recognised as a patch Jul 14 03:29:52 unless apply=yes Jul 14 03:30:17 Tartarus: I am trying to patch linux_2.6.28.bb as example Jul 14 03:31:31 Tartarus: http://pastebin.com/y8nWuuHk Jul 14 03:31:47 Hey JaMa up so early :) Jul 14 03:35:37 Tartarus: the symlink points to DL_DIR/.bz2 Jul 14 03:35:52 and the patch is applied Jul 14 03:36:01 so seems ok Jul 14 03:36:22 there is a copy of unzipped patch in WORKDIR may be thats not needed Jul 14 06:08:13 03Robert Schuster  07org.openembedded.dev * r5dfb4130a5 10openembedded.git/recipes/llvm/llvm-common.bb: llvm-common: Fixed installation path of helper script (led to issues in OpenJDK build), bumped PR. Jul 14 06:08:18 03Robert Schuster  07org.openembedded.dev * r1fc243e252 10openembedded.git/recipes/cacao/ (cacao.inc cacao_0.99.3.bb cacao_0.99.4.bb): Jul 14 06:08:18 cacao 0.99.3: Fixed packaging of libjvm.so. Jul 14 06:08:18 cacao 0.99.4: Dito. Jul 14 06:08:18 03Robert Schuster  07org.openembedded.dev * r1bd457c582 10openembedded.git/recipes/jamvm/ (4 files in 2 dirs): Jul 14 06:08:18 jamvm-initial 1.4.5: Face random crashes with a silent retry. Jul 14 06:09:58 khem, Tartarus, just backread the your conversation, i seem to have problems with the symlinking of patches and using quilt; if a patch fails (e.g. when moving to a new version) I use quilt to update the patch, but quilt push -f told me that the patch was empty or not valid or something like that) Jul 14 06:10:20 (happened two days ago @ home, so don't know exact details any more) Jul 14 06:14:24 03Robert Schuster  07org.openembedded.dev * r440b956c99 10openembedded.git/recipes/jamvm/ (4 files in 2 dirs): jamvm-initial 1.4.5,1.5.0: Fixed binary to be called in script. Jul 14 06:32:12 03Graham Gower  07org.openembedded.dev * rb4cdc30178 10openembedded.git/docs/usermanual/chapters/recipes.xml: Jul 14 06:32:12 usermanual/chapters/recipes.xml: Executables should be executable. Jul 14 06:32:13 Signed-off-by: Graham Gower Jul 14 06:32:13 Signed-off-by: Khem Raj Jul 14 06:45:37 eFfeM_work: I have patched quilt-native to follow symlinks Jul 14 06:46:07 eFfeM_work: so if you refresh a patch in your workdir it will get updated whereever the symlink points to Jul 14 06:46:33 but you need to use quilt-native Jul 14 06:58:01 khem: ah ok, will try that (actually I used quilt from the host, not quilt-native) Jul 14 06:58:30 khem would it be an idea to have a var in local.conf to specify if you want symlinks or not ? Jul 14 07:27:34 morning Jul 14 07:28:17 guys, question... Jul 14 07:28:41 what do i have to do to come a change in a tasks file into effect Jul 14 07:28:59 if it's included in the machine conf Jul 14 07:57:44 khem: hehe, no, pushing before going to bed.. 2:20AM Jul 14 08:10:03 moin woglinde Jul 14 08:10:15 hi ant Jul 14 08:10:44 still tropical temps there? Jul 14 08:12:17 hm not yert Jul 14 08:12:22 but later I think so Jul 14 08:20:29 he Jul 14 08:26:13 morgen Jul 14 08:27:07 I'm editing a recipe. do_compile takes ages and I only changed FILES_${PN}something and I'd like to see if it's correct now Jul 14 08:27:35 is there a task that I can run to just test this? Jul 14 08:27:48 I mean the selection of what file goes to what package Jul 14 08:31:42 zub bitbake -f -c stage foobarr Jul 14 08:31:47 thanks Jul 14 08:32:04 and than lookinto workdir/packages-split Jul 14 08:32:15 yup, found that already :) Jul 14 08:32:18 okay Jul 14 08:32:28 just didn't know which task is responsible Jul 14 08:32:33 hm or -c package Jul 14 08:32:35 sorry Jul 14 08:32:38 ah, ok Jul 14 08:36:35 package worked :) Jul 14 08:36:46 what does distribute_sources do? Jul 14 08:37:07 for distros-maintainers Jul 14 08:37:24 you can put the source for the packages besides the packages Jul 14 08:37:31 like debian does Jul 14 08:37:39 and palm didnt understand Jul 14 08:38:05 hm, ok Jul 14 08:45:29 if i add something in a task file which is included in the machine conf used for an image... what do i have to bitbake in order for those changes to take effect? Jul 14 08:48:04 bitbake -f foo-image Jul 14 08:48:19 -f! ok thanks Jul 14 08:48:56 oh uhm Jul 14 08:49:00 i don't think that's gonna work Jul 14 08:49:27 it doesn't even work if i -c clean the image package and rebake Jul 14 08:49:35 fraxinas: rm tmp && bitbake whateveryouneed Jul 14 08:49:48 eFfeM_work: that would take 3 hours Jul 14 08:49:50 takes some time but then you have something :-) Jul 14 08:50:01 :) Jul 14 08:50:02 effem *g* Jul 14 08:50:02 i want to do experiments Jul 14 08:50:16 clean image and bitbake -f image Jul 14 08:50:19 should work Jul 14 08:50:27 or the task Jul 14 08:50:41 i can specifically run the task? Jul 14 08:50:44 cool didn't know that Jul 14 08:50:53 well we don't know about your change (what task/recipe are affected) but you could clean those Jul 14 08:51:06 fraxinas: bitbake -ctask -f recipe Jul 14 08:51:20 e.g. bitbake -cclean console-image Jul 14 08:51:38 -f is handy in case you want to force things, e.g. in combination with -ccompile Jul 14 08:51:45 -c cleaning the image doesn't work Jul 14 08:52:02 -f is pretty much the same as deleting the stamp which i had also already tried Jul 14 08:52:28 okay but running the task which i had changes by itself might work Jul 14 08:52:29 -c clean will remove the image, but if your task affects the underlying components/recipes: those are not cleaned Jul 14 08:53:02 * eFfeM_work takes a wild guess and suggests your task might not be added properly and just not run Jul 14 08:53:37 yeah it was originally run Jul 14 08:53:48 just not since i had done changes to it Jul 14 08:54:18 pastebin your changes ? Jul 14 08:56:57 hm, aren't you confusing -c task vs. recipes/tasks/* ? Jul 14 08:57:06 I mean in the discussion :) Jul 14 08:57:33 oh shoot i broke youtube Jul 14 08:57:56 fraxinas what? Jul 14 08:58:06 or does it work for you? Jul 14 08:58:11 zub hm? Jul 14 08:58:23 you can use -b recipes/tasks/fii.bb Jul 14 08:58:27 yes Jul 14 08:58:32 so bitbake dont does the parasing Jul 14 08:58:37 uops parsing Jul 14 08:58:47 and it seemed to me fraxinas meant that task, not -c task Jul 14 08:58:48 but have to know what you are doing Jul 14 08:58:50 anyway, what do I know Jul 14 08:59:14 zub I indeed was talking about bitbake tasks Jul 14 09:00:31 okay cool Jul 14 09:00:35 so it worked now Jul 14 09:00:35 if it is a task recipe just clean and bake it (as well as the recipes that depend on it Jul 14 09:01:15 what i did was bitbake -c clean task-opendreambox-enigma2 && bitbake task-opendreambox-enigma2 && bitbake -f dreambox-image Jul 14 09:01:16 et voilà Jul 14 09:01:30 sure Jul 14 09:01:35 thanks for the hint about running the task directly Jul 14 09:01:37 we are using it for a long time Jul 14 09:01:45 hehe ;) Jul 14 09:01:52 so am i... three years actually Jul 14 09:02:14 just never touched those task files before Jul 14 09:29:16 time for me to complain ;D Jul 14 09:29:22 too many native packages Jul 14 09:29:35 hm? Jul 14 09:30:30 "bitbake gcc-cross" takes openssl, python Jul 14 09:32:10 hi jama Jul 14 09:32:34 hm does it builds git-native? Jul 14 09:32:40 probably Jul 14 09:33:03 khem let quilt depends on perl package Jul 14 09:33:06 but I found it Jul 14 09:33:11 and he reverted it Jul 14 09:33:13 no.. I have it in ASSUM_PR Jul 14 10:31:04 hm anyone knows how to send the 0/x mail with git-sendmail? Jul 14 10:37:24 Anybody here know where the register addresses are defined for a particular platform? Particularly, i'm wondering about for the omap? Jul 14 10:40:04 ?????? Jul 14 10:40:08 wrong channel Jul 14 10:41:45 bkinman: #linux-omap? Jul 14 10:45:11 I had no idea such a channel existed! Awesome, thanks! Jul 14 10:47:43 jo crofton Jul 14 10:47:44 leviathan: what is the fix for lsof problem ? Jul 14 10:49:50 gm Jul 14 11:04:51 woglinde: --cover-letter Jul 14 11:05:18 doing format-patch Jul 14 11:09:32 mm, I wonder why git feels the need to send all of those as separate emails. it would be rather less annoying if it combined them into a single mail with the individual patches as mime attachments. Jul 14 11:10:34 ant cool Jul 14 11:10:53 pb but than patchwork dont works proably Jul 14 11:11:12 is already broken :/ Jul 14 11:11:24 hm what we have to do to get native debug binaries? Jul 14 11:12:03 dcordes: you need to unpack manually Jul 14 11:12:20 somehow unpacking done by the recipe itself is broken Jul 14 11:12:47 leviathan: how can we fix it ? Jul 14 11:13:17 doing the unpack routine differently Jul 14 11:13:28 should I do a patch? Jul 14 11:30:29 woglinde: true, but patchwork could be fixed. Jul 14 11:30:35 and, well, who cares about patchwork anyway :-} Jul 14 11:30:56 hm I have 22 patches for oe-stable Jul 14 11:31:25 so should I spam the list? Jul 14 11:31:38 it seems Jul 14 11:32:21 woglinde, send them to the list Jul 14 11:32:30 okay Jul 14 11:32:30 stop wasting time pissing and moaning about how it works Jul 14 11:32:36 hm? Jul 14 11:32:55 I dont moan Jul 14 11:33:17 well, the conversation seemed to be headed that way :) Jul 14 11:33:43 cover-letter was what I searched Jul 14 11:34:19 I think I used an english phrase that translates poorly Jul 14 11:43:22 hi, i'm getting error gee/Makefile.am:95: HAVE_INTROSPECTION does not appear in AM_CONDITIONAL while executing libgee_git.bb Jul 14 11:44:53 zenob yes Jul 14 11:45:03 its known Jul 14 11:45:10 gnutoo is working on it Jul 14 11:45:46 ok, thx Jul 14 11:47:17 hm Jul 14 11:47:29 why he left Jul 14 11:47:47 only stops by to drop problems Jul 14 11:47:49 I am was about to explain him how to solve it for the moment Jul 14 12:16:09 03Dmitry Eremin-Solenikov  07org.openembedded.dev * r10fabfa670 10openembedded.git/recipes/gdb/ (16 files): Jul 14 12:16:10 gdb: stop conflicting with binutils Jul 14 12:16:10 To stop gdb file conflicts with binutils (in different ways) move Jul 14 12:16:10 removing of bfd/opcodes/libiberty stuff to gdb-common.inc. Also bump PRs Jul 14 12:16:10 to force rebuild. Jul 14 12:47:40 problem executing bitbake omap3-console-image Jul 14 12:48:06 ERROR: Error in executing: /home/team/overo-oe/org.openembedded.dev/recipes/coreutils/coreutils-native_8.5.bb Jul 14 12:48:11 has anyone seen this? Jul 14 12:50:33 here's the error: ERROR: Exception:exceptions.AttributeError Message:'str' object has no attribute 'partition' Jul 14 12:56:01 which bitbake you are using? Jul 14 12:56:58 BitBake Build Tool Core version 1.8.18, bitbake version 1.8.18 Jul 14 12:57:36 try using bitbake master version Jul 14 12:57:53 its faster anyway for dev Jul 14 12:58:51 this is how I got bitbake:git clone git://git.openembedded.net/bitbake bitbake Jul 14 12:59:01 is it a different directory for the master? Jul 14 12:59:31 hm shoud be latest Jul 14 12:59:35 try bitbake --version Jul 14 13:00:09 that's how I got the version line listed above... hmm... Jul 14 13:01:51 hm... I had also executed: git checkout 1.8.18 Jul 14 13:02:22 that was per instructions here: http://www.gumstix.net/Setup-and-Programming/view/Overo-Setup-and-Programming/Setting-up-a-build-environment/111.html Jul 14 13:08:44 thats not master Jul 14 13:08:53 thats the 1.8.18 version Jul 14 13:12:45 should I just recheckout bitbake and not get the 1.8.18 version? Jul 14 13:15:37 hm get familiar with git and branches Jul 14 13:15:52 git checkout master Jul 14 13:17:04 whats gmx.de? Jul 14 13:18:03 I did: git checkout master and now when I try bitbake --version, I get:File "/home/team/overo-oe/bitbake/bin/bitbake", line 203 except Exception as e: ^ SyntaxError: invalid syntax Jul 14 13:19:03 Crofton german email provider Jul 14 13:19:32 lstotz whats your python version? Jul 14 13:20:24 Python 2.4.4 (#1, Aug 8 2007, 13:47:44)... hmm... looks like I should update that... Jul 14 13:22:04 thought so Jul 14 13:22:13 woglinde, whanks Jul 14 13:22:39 lstotz iehks Jul 14 13:22:45 you need 2.6 at least Jul 14 13:22:48 for master Jul 14 13:23:17 crofton no prob Jul 14 13:24:18 hm I can push my stable stuff as user-branch Jul 14 13:24:20 good idea Jul 14 13:25:24 yep Jul 14 13:27:02 hm okay letsee how I can tweak git config to use my key from usbstick Jul 14 13:30:06 hm seems easy Jul 14 13:32:26 hm could someone please show me his gitconfig with ssh access Jul 14 13:40:59 woglinde: see pm Jul 14 13:41:13 yes Jul 14 13:41:17 let me try this Jul 14 13:46:10 yes, updating python got me further! Thanks woglinde... now I will go onto the next problem... funny that the website says that I need python, but does not indicate a version, my system must just be old Jul 14 13:49:24 lstotz: http://wiki.openembedded.net/index.php/Getting_Started Jul 14 13:49:25 One note for those who want to play with development versions of BitBake - Python 2.6 may be required by newer versions. This can be a problem for some Linux distributions. Jul 14 13:50:04 hm seems I took the wrong key from home Jul 14 13:51:41 http://wiki.openembedded.net/index.php/RequiredSoftware says that I just need python 2.4 or higher, which is what I had Jul 14 13:52:15 this is what I get now: NOTE: Running task 35 of 4383 (ID: 21, /home/team/overo-oe/org.openembedded.dev/recipes/quilt/quilt-native_0.48.bb, do_install) ERROR: function do_install failed Jul 14 13:52:46 lstotz is a wiki Jul 14 13:52:49 fix it Jul 14 13:52:55 +it is Jul 14 13:53:06 woglinde, you were just ahead of me, wanted to give the same reply Jul 14 13:54:00 but I will do it for now Jul 14 14:00:39 point taken, it is changed Jul 14 14:01:01 hm oh Jul 14 14:01:02 thanks Jul 14 14:07:16 Hi everybody. Jul 14 14:08:38 woglinde: fyi: icedtea6-native_1.7.3.bb seems to be missing fastjar-native in DEPENDS Jul 14 14:09:08 I have a issues by compiling with bitbake on Ubuntu 9.10. Jul 14 14:09:30 uh really? Jul 14 14:09:35 let me check Jul 14 14:09:41 I worked through http://wiki.openembedded.net/index.php/Getting_started and getting following compile error: Jul 14 14:09:50 http://pastebin.com/LmDwh8z4 Jul 14 14:10:00 jama hm but you needclasspath Jul 14 14:10:03 and classpath has it Jul 14 14:11:24 woglinde: building "bitbake openjdk-6" didn't build fastjar-native before icedtea6-native Jul 14 14:12:02 woglinde: hmm and still /bin/sh: jar: command not found Jul 14 14:12:20 jama hm olay Jul 14 14:12:26 put it in back Jul 14 14:12:27 thanks Jul 14 14:12:42 seems I took the wrong key from home Jul 14 14:12:42 gcc was already downgraded to 4.1.x and is now 4.2.4 Jul 14 14:13:19 taalas are you using dev and which distribution? Jul 14 14:13:35 woglinde: what whould provide "jar" when fastjar doesn't put it in sysroots? Jul 14 14:13:51 taalas, what filesystem are you using Jul 14 14:14:02 jama gjar from classpath Jul 14 14:14:48 jama problem is there is know alternatives for native stuff Jul 14 14:14:50 in local.conf I had at first Angstroem I also tryied also with minimal. Jul 14 14:14:54 woglinde: ah, than why is icedtea6-native_1.7.3.bb calling just "jar"? missing jar -> gjar link? Jul 14 14:14:54 ups know/no Jul 14 14:15:15 woglinde: now trying with sane-toolchain-java + angstrom-jalimo Jul 14 14:15:22 eFfem_work if you are talking about my dev pc then I use ext3 Jul 14 14:15:27 taalas, what host distro Jul 14 14:15:36 woglinde: before I had only alternatives from compatibility-providers.. Jul 14 14:15:52 taalas, yeah wanted to make sure you didn't put in on a fat fs or so, sometimes people do try that Jul 14 14:15:54 eFfeM_work Ubuntu 9.10 Jul 14 14:16:00 jama but fastjar should work Jul 14 14:16:09 and fastjar works in all situations Jul 14 14:16:10 taalas, can you compile a helloworld proggie Jul 14 14:16:56 I only tried to compile using "bitbake nano" Jul 14 14:17:06 woglinde: http://paste.pocoo.org/show/237452/ Jul 14 14:17:57 hm I thought I patched this for native too Jul 14 14:19:54 taalas this is no good: Jul 14 14:20:04 | configure: error: C compiler cannot create executables Jul 14 14:20:28 morning Jul 14 14:20:30 suggest to try verify first that host gcc is working Jul 14 14:20:34 hi kergoth__ Jul 14 14:20:56 hi kergoth Jul 14 14:21:18 jama could you please pastebin the config.log Jul 14 14:21:48 eFfeM_work Output for gcc -v is correct. Jul 14 14:22:03 gcc version 4.2.4 (Ubuntu 4.2.4-5ubuntu1) Jul 14 14:22:16 and you can compile and run a simple prog ? Jul 14 14:22:27 eFfeM_work and the helloworld program is also failing Jul 14 14:23:03 gcc -o hello hello.c (after writing hello.c) ? Jul 14 14:23:48 eFfeM_work japp worked Jul 14 14:23:56 hm, very strange Jul 14 14:24:12 out of ideas atm Jul 14 14:25:11 eFfeM_work I installed also Ubuntu 9.04 to verify if there is any error. Jul 14 14:25:46 i don't think this is ubuntu related, I've build oe head many times on many ubuntu versions Jul 14 14:27:05 you mightwant to check if there is somethingodd in the config.log Jul 14 14:29:13 g'day kergoth__ Jul 14 14:29:15 eFfem_work only error there is: configure:4490: error: in `/stuff/build/tmp/work/i686-linux/coreutils-native-8.5-r0/coreutils-8.5': Jul 14 14:29:25 configure:4492: error: C compiler cannot create executables Jul 14 14:29:40 se config.log for more details Jul 14 14:29:54 tehre is a config.log somehwere in that dir you gave Jul 14 14:29:59 woglinde: http://paste.pocoo.org/show/237456/ Jul 14 14:30:41 I grepped for 'error' thru the config.log and that was the only thing I got Jul 14 14:30:51 Everything else looks fine for me Jul 14 14:31:28 i guess configure will try to run some test program and it fails because something is missing Jul 14 14:31:59 ok I found some permission problems with ccache: ccache: failed to create /home/hctm/.ccache/CACHEDIR.TAG (Permission denied) Jul 14 14:32:16 hm, no idea if that is related Jul 14 14:33:01 woglinde: .pc/build-hacks-native.patch/Makefile.am calls just "jar" in WITH_RHINO section Jul 14 14:33:40 jama hm ah Jul 14 14:35:44 Oh damn I think I got it... Jul 14 14:35:50 jama hm I forgot to update the patch Jul 14 14:36:06 andrew made a better patch Jul 14 14:36:43 Taalas, enlighten us Jul 14 14:36:48 (or entertain us :-) ) Jul 14 14:37:02 I think it will entertain you... Jul 14 14:37:08 :-) Jul 14 14:37:11 jama http://icedtea.classpath.org/hg/icedtea6/rev/7674917fa451 Jul 14 14:37:36 .ccache was owned by root because I created something with sudo ltib previously... Jul 14 14:37:54 So he .ccache folder was only writable for root... Jul 14 14:37:57 jama fell free to check it in Jul 14 14:37:58 ah ok, those things happen Jul 14 14:38:05 otherwise I will do it this evening Jul 14 14:38:11 next time I'll ask to ccache gcc hello.c :-) Jul 14 14:38:52 :-) Thanks for your help eFfeM_work Jul 14 14:39:04 yw Jul 14 14:46:47 woglinde: thanks, trying Jul 14 14:52:52 JaMa: do you happen to remember the thread where someone (koen) analyzed the build time? Jul 14 14:53:47 or you did the analysis? can't remember Jul 14 14:54:19 build time of image? Jul 14 14:54:30 I don't remember any such thread.. Jul 14 14:55:58 ant_work: maybe you're talking about thread where I said that IO is bottleneck and koen analyzed how much time he saved by converting to new staging? IIRC this was in "minimalistic chroot" thread Jul 14 14:55:59 np Jul 14 14:56:04 oh, yes Jul 14 14:56:20 io is bottleneck Jul 14 14:56:43 jkridner in #beagle is disappointed: he has tmpfs anfd 68GB ram..8 cores Jul 14 14:56:50 no big improements Jul 14 14:57:05 what disks? Jul 14 14:57:12 iirc bttleneck are legacy staging and disabled parallel build Jul 14 14:57:18 for some recipes Jul 14 14:57:37 woglinde: all in ram Jul 14 14:58:05 ccache too? Jul 14 15:00:43 disabled iirc Jul 14 15:00:51 he, I know what you think Jul 14 15:01:10 fwiw I tried with ccache on tempfs and without ccache Jul 14 15:01:15 no big diff Jul 14 15:01:24 so I radicated ccache forever ;) Jul 14 15:01:29 *eradicated Jul 14 15:01:38 hm right Jul 14 15:01:39 sorry Jul 14 15:02:20 I think that 30-40% of build time is for toolchain (for little images) Jul 14 15:02:46 JaMa: have you a link? Jul 14 15:03:22 np, found Jul 14 15:03:40 ant_work: http://www.mail-archive.com/openembedded-devel@lists.openembedded.org/msg03803.html Jul 14 15:29:35 args Jul 14 15:29:40 one h missed Jul 14 15:29:47 for the key Jul 14 15:35:31 JaMa: thx Jul 14 15:35:33 bl Jul 14 15:38:26 GNUtoo|laptop: have you found how to fix libgee problem ? Jul 14 15:38:37 zenob, no Jul 14 15:38:49 zenob, I sent a mail in the dev ml Jul 14 15:46:55 leviathan: yes a patch would be nice Jul 14 15:47:24 hm, just hit the gtk+-native race-or-what again. bitbaking again resolves it. Jul 14 15:47:47 > Cannot load module /mnt/data/zoom2/OE/tmp/work/x86_64-linux/gtk+-native-2.20.1-r8.1/gtk+-2.20.1/modules/input/im-xim.la: /mnt/data/zoom2/OE/tmp/work/x86_64-linux/gtk+-native-2.20.1-r8.1/gtk+-2.20.1/modules/input/.libs/im-xim.so: undefined symbol: g_realloc_n Jul 14 15:48:09 has anybody got an idea of what's happenign there? Jul 14 15:52:15 jama still working on it? Jul 14 15:53:55 woglinde: patch backported to 1.7*, waiting for do_compile to finish Jul 14 15:54:05 oh okay Jul 14 15:54:23 woglinde: seems like hanging somewhere :/ Jul 14 15:54:56 jama there is a bug somewhere in jamvm Jul 14 15:55:11 which random let it segfaults Jul 14 15:55:15 we are working on it Jul 14 15:55:50 yup, I had 2 segfaults already (restarting build worked in both cases), but now it just hangs Jul 14 15:55:58 I'll kill and restart Jul 14 15:56:13 last line in log.do_compile is [copy] Copying 1 file to /OE/tmpdir-dev-shr/work/x86_64-linux/icedtea6-native-1.7.3-r0.0/icedtea6-1.7.3/openjdk-ecj/build/linux-amd64/langtools/build/bootstrap/bin Jul 14 15:56:28 hm Jul 14 15:57:05 hm how shoud I push my user/branch Jul 14 15:57:27 is git push origin yourname/yourfeature still safe? Jul 14 15:59:07 * kergoth prefers github, personally, just because he doesn't have to pester a git.openembedded.org admin to remove / rebase the user branches later on Jul 14 15:59:23 kergoth hm Jul 14 15:59:39 yeah, me too Jul 14 15:59:46 hms Jul 14 16:01:17 * JaMa too Jul 14 16:02:52 03Martin Jansa  07org.openembedded.dev * rbe6ff5541d 10openembedded.git/recipes/icedtea/ (2 files in 2 dirs): Jul 14 16:02:52 icedtea6-native_1.7.3: backport patch for building without jar in system Jul 14 16:02:52 Signed-off-by: Martin Jansa Jul 14 16:03:00 03Martin Jansa  07org.openembedded.dev * r2b80712f64 10openembedded.git/conf/distro/ (include/preferred-shr-versions.inc shr.conf): Jul 14 16:03:00 shr: include sane-toolchain-java.inc and angstrom-jalimo.conf, add largefile to DISTRO_FEATURES Jul 14 16:03:00 Signed-off-by: Martin Jansa Jul 14 16:03:50 goddamn, like 75% of the commits between 1.8 and 1.10 are unusable, makes a git bisect a pain in the ass Jul 14 16:04:21 thanks jama Jul 14 16:04:25 woglinde: yw Jul 14 16:08:01 hm I branched from oe stable Jul 14 16:08:09 how can I that push to github now Jul 14 16:08:18 the same way you push to anything else. Jul 14 16:08:25 git push origin localbranch:remotebranch Jul 14 16:08:29 or whatever the remote is Jul 14 16:08:53 hm I made origin2 Jul 14 16:09:13 so git push origin2 localbranch:remotebranch, or just git push origin2 localbranch Jul 14 16:09:45 see man git-push :-} Jul 14 16:10:11 woglinde: not sure if github has same feature as gitorious, but on gitorious it's faster if you clone git repo from web interface Jul 14 16:10:29 error: src refspec jalimo-stable does not match any. Jul 14 16:10:30 woglinde: and then only push your changes while sharing common git objects Jul 14 16:11:03 hm maybee I have to push something other first Jul 14 16:11:35 that sounds like you just got the name of your local branch wrong. it's hard to say what exact command you need without knowing all the context. Jul 14 16:12:11 git branch shows me * stable-jalimo Jul 14 16:12:21 ah, that would be the problem then Jul 14 16:12:27 jalimo-stable != stable-jalimo Jul 14 16:12:34 ups Jul 14 16:12:36 try again with that name and see if the situation improves Jul 14 16:12:37 damn Jul 14 16:12:38 thanks Jul 14 16:17:01 woglinde: one more issue: ( cd /OE/tmpdir-dev-shr/work/x86_64-linux/icedtea6-native-1.7.3-r0.0/icedtea6-1.7.3/openjdk-ecj/build/linux-amd64 && /usr/bin/unzip -o /OE/tmpdir-dev-shr/work/x86_64-linux/icedtea6-native-1.7.3-r0.0/icedtea6-1.7.3/openjdk-ecj/build/linux-amd64/corba/dist/lib/bin.zip ) Jul 14 16:17:06 /bin/sh: /usr/bin/unzip: No such file or directory Jul 14 16:17:27 * JaMa gtg Jul 14 16:17:40 yeah already noticed this Jul 14 16:18:42 are you able to fix it yourself? Jul 14 16:27:47 bye for now Jul 14 16:41:38 khem, did the disable-multilib go in? Jul 14 16:51:26 03Xerxes Rånby  07org.openembedded.dev * r1ea15a2cd7 10openembedded.git/recipes/openjdk/ (3 files in 2 dirs): openjdk-6-6b18: Add ARM Linux Shark CPU features detection code. Jul 14 16:55:40 03Xerxes Rånby  07org.openembedded.dev * r74d4f4b860 10openembedded.git/recipes/openjdk/ (3 files in 2 dirs): openjdk-6-6b18: Remove vestigial thumbee.patch. Jul 14 17:08:50 Tartarus: no I have the patch locally not pushed yet Jul 14 17:09:04 Tartarus: has been trying combinations Jul 14 17:09:54 combinations of? Jul 14 17:17:52 somebody able to ack large mail (w/patch) on mobile-linux-discuss waiting for moderator approval ? Jul 14 17:18:07 sorry for the OT Jul 14 17:19:20 mickeyl: ping ? Jul 14 17:50:23 Tartarus: different arches and distro Jul 14 17:50:36 Tartarus: one of mips,ppc,arm and uclibc eglibc etc. Jul 14 17:50:55 Tartarus: I can post the patch for some acks Jul 14 18:13:08 of making multilib work? Jul 14 18:13:18 based on romans work i assume Jul 14 18:14:30 Tartarus: no time to do that Jul 14 18:14:31 here Jul 14 18:15:06 Then what are you testing, I'm confused Jul 14 18:15:13 Change enable to disable, commit :) Jul 14 18:15:18 you already got acks Jul 14 18:46:10 > Jul 14 18:46:10 > Jul 14 18:46:10 > ------------------------------------------------------------------------ Jul 14 18:46:10 > Jul 14 18:46:10 > _______________________________________________ Jul 14 18:46:11 > Openembedded-devel mailing list Jul 14 18:46:11 > Openembedded-devel@lists.openembedded.org Jul 14 18:46:12 > http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel Jul 14 18:46:28 Bah, naughty gio... Jul 14 18:46:35 and bad me for leaving the irc window open Jul 14 19:09:13 rofl Jul 14 19:19:57 03Khem Raj  07org.openembedded.dev * r847c0dadb9 10openembedded.git/recipes/gcc/gcc-configure-common.inc: Jul 14 19:19:57 gcc-configure-common.inc: Disable multilib generation. Jul 14 19:19:57 Signed-off-by: Khem Raj Jul 14 19:19:57 Acked-by: Roman I Khimov Jul 14 19:19:57 Acked-by: Tom Rini Jul 14 19:34:40 03Frans Meulenbroeks  07org.openembedded.dev * r8a68f0d1c4 10openembedded.git/classes/base.bbclass: (log message trimmed) Jul 14 19:34:40 base.bbclass: introduce COMPATIBLE_TARGET_SYS Jul 14 19:34:40 This patch introduces COMPATIBLE_TARGET_SYS Jul 14 19:34:40 It is similar to COMPATIBLE_MACHINE but where COMPATIBLE_MACHINE Jul 14 19:34:40 is used to specify that a certain recipe is for a certain machine Jul 14 19:43:02 anyone have a recommended way to deal with installing configuration files to an OE distro? Jul 14 19:43:21 you're going to have to elaborate Jul 14 19:45:41 stupid one- is there any way to kick out myself being still logged in on freenode from other machine? I forgot to log out when I was lieaving the office ... ;] Jul 14 19:45:47 ? Jul 14 19:46:29 nazgee_, yes, it's called ghosting, as long as you're registered iwth nickserv using your main nick Jul 14 19:46:35 /msg nickserv help Jul 14 19:46:40 thx ;] Jul 14 19:50:11 khem, bb master was somehow broken WRT proxy handling in the fetcher last week so any testing would need me fixing this first Jul 14 19:50:21 s/me/somebody/ Jul 14 19:51:04 WRT to configuration files, I'm trying to configure update the ssh service on my distro to enable PAM. Jul 14 19:51:26 that's not a configuration file question Jul 14 19:51:32 Essentially two files need to be modified, /etc/pam.d/sshd and /etc/ssh/sshd_config. Jul 14 19:51:59 khem, bb or oe, didn't look closely yet Jul 14 19:52:00 What's the best way to inject these into the system ? Jul 14 19:52:07 most services need to be buit without --without-pam, too Jul 14 19:52:18 and are you sure it's only ssh, not the /bin/login also? Jul 14 19:52:44 you can leverage FILESPATHBASE to provide a different config file to the recipe, if th erecipe pulls those files using file:// uris in SRC_URI Jul 14 19:52:46 kergoth, actually no, the login will be replaced as well... Jul 14 19:52:48 re Jul 14 19:52:54 "replaced"? Jul 14 19:53:05 its likely you're using busybox login, which likely doesn't support pam at all Jul 14 19:53:22 * kergoth is guessing there, though, would want to check Jul 14 19:53:50 there's a bit of background. Jul 14 19:54:28 how does the FILESPATHBASE work? Jul 14 19:55:58 03Tom Rini  07org.openembedded.dev * r350abdb99e 10openembedded.git/recipes/ipsec-tools/ (files/ipsec-tools-install.patch ipsec-tools_0.7.2.bb): Jul 14 19:55:58 ipsec-tools: Add a patch to not 'install' vmbuf.h twice Jul 14 19:55:58 Patch taken form pld-linux.org Jul 14 19:55:58 Signed-off-by: Tom Rini Jul 14 19:56:34 darknighte grep some recipes you will see Jul 14 19:57:19 woglinde, that's strange, when I grep for FILESPATHBASE, I only get a hit from the bitbake.conf file. Jul 14 19:57:55 Looks like it's used in FILESPATH, which is widely reference in recipes... Jul 14 20:00:52 can anyone confirm that using FILESPATH requires an entire recipe to be copied into a local overlay? Jul 14 20:02:40 03Denys Dmytriyenko  07org.openembedded.dev * rc40eea2a6c 10openembedded.git/recipes/libc-client/libc-client_2007b.bb: (log message trimmed) Jul 14 20:02:40 libc-client: remove extra "openssl" from the include path Jul 14 20:02:40 Since normally OpenSSL headers are already included from "openssl" Jul 14 20:02:40 directory, don't add it to the path again, which breaks the build Jul 14 20:02:40 with external toolchains. Thanks to Pierluigi Passaro for suggesting Jul 14 20:24:11 any autotools experts here? :) Jul 14 20:25:08 hm, if I plan to send a patch for a recipe, should I increment PR? Jul 14 20:25:16 denix depends Jul 14 20:25:18 ask Jul 14 20:25:27 zub yes Jul 14 20:25:40 configure.ac:23: option `silent-rules' not recognized Jul 14 20:25:53 hm Jul 14 20:26:02 thx Jul 14 20:26:18 denix this a relative new option Jul 14 20:26:24 maybee autoconf is to old Jul 14 20:26:58 crap. my build is stuck with: "ERROR: QA Issue with python-pygps: non debug package contains .debug directory:" message. my oe is 2099 commits behind dev branch. does it make sense to pull, wishing this error would go away? Jul 14 20:27:03 here's full log - http://pastebin.com/9ZNDBxDr Jul 14 20:28:36 gettext-0.17 worked fine, new 0.18 fails like that... Jul 14 20:28:37 nazgee gues Jul 14 20:28:40 re ob Jul 14 20:28:43 ups pb Jul 14 20:28:48 denix hm? Jul 14 20:29:12 automake (GNU automake) 1.10.3 Jul 14 20:29:16 you need 1.11 Jul 14 20:29:47 denix what distro are you using? Jul 14 20:30:22 slightly outdated Gentoo... Jul 14 20:30:31 1.10.3 isn't THAT old! it's 2009 :) Jul 14 20:31:09 that is so last year Jul 14 20:31:24 Crofton|work: :) Jul 14 20:33:06 woglinde: i'll give it a try ;] i was hoping to find some 'magic' here, that could make it go away this way or another. Jul 14 20:33:33 but it seems ain't harry potter's world Jul 14 20:36:33 requiring the very latest automake doesn't seem like a good thing... Jul 14 20:37:53 hi woglinde Jul 14 20:44:52 03Tom Rini  07org.openembedded.dev * r098bbaa744 10openembedded.git/recipes/clamav/ (clamav-0.96.1/fix-libclamav.pc.in.patch clamav_0.96.1.bb): Jul 14 20:44:52 clamav 0.96.1: Add a patch to drop @CFLAGS@ from libclamav.pc.in, bump PR Jul 14 20:44:52 @CFLAGS@ doesn't belong in pkgconfig files Jul 14 20:44:52 Signed-off-by: Tom Rini Jul 14 20:45:05 03Tom Rini  07org.openembedded.dev * r9ea5318120 10openembedded.git/classes/pkgconfig.bbclass: Jul 14 20:45:05 pkconfig.bbclass: Add -R... to the regex, and reindent. Jul 14 20:45:05 clamav 0.96.1 makes use of this flag Jul 14 20:45:05 Signed-off-by: Tom Rini Jul 14 20:46:41 I am dumping something with readelf using os.Popen("readelf") Jul 14 20:47:06 I want to put the output into some data structure that I can use later Jul 14 20:47:13 whats the best way to do it in python Jul 14 20:47:18 Um, sec Jul 14 20:47:22 There's examples Jul 14 20:47:30 Tartarus: pointer ? Jul 14 20:47:49 relocatable.bbclass Jul 14 20:47:51 Tartarus: btw if you know if any example that plays around with dum[s Jul 14 20:47:54 dumps Jul 14 20:47:55 jo khem Jul 14 20:48:06 uses chrpath not readelf Jul 14 20:48:10 But same basic idea Jul 14 20:48:38 err, out = p.communicate() Jul 14 20:48:43 throws stderr to err, stdout to out Jul 14 20:48:44 khem is there no readelf python-modul? Jul 14 20:50:20 what's the best way to install config files for a service? Jul 14 20:51:42 bbl Jul 14 20:54:42 03Frans Meulenbroeks  07org.openembedded.dev * r6bd90b5b7e 10openembedded.git/recipes/gcc/ (6 files in 4 dirs): (log message trimmed) Jul 14 20:54:43 gcc 4.1.x: patch Makefile.in for cross compile badness Jul 14 20:54:43 When compiling gcc target code make passes -I(GMPINC) Jul 14 20:54:43 but this one points to the host dir (it is set to HOST_GMPINC Jul 14 20:54:43 which in oe points to tmp/sysroots/i686-linux/usr/include) Jul 14 21:18:46 when using u-boot... whats the use of "boot.ini"? sometimes my environment is not readed from that file, someone has said me that u-boot has his own boot.ini and under some behaviours it loads it and refuses the one u put in the boot partition... is it true? Jul 14 21:37:02 ~tell kergoth please let me know if those two mini bb patches are ok to install -- TIA Jul 14 21:38:18 kms for vodoo-chipset Jul 14 21:38:21 oh my god Jul 14 21:38:23 khem, hm. is there a msg-bot in #here? I forgot to say that in a coverletter Jul 14 21:38:38 voodoo1 i hope ;) Jul 14 21:38:51 http://www.phoronix.com/scan.php?page=news_item&px=ODQxMw Jul 14 21:42:02 cool stuff. More than one decade after the initial impl Jul 14 21:45:45 bitbaking my new image on dev branch failes with "libtool-native_2.2.6b.bb do_install failed", crying that it cant Jul 14 21:46:04 is tinderbox flaky? NOTE: oestats: error sending task, disabling stats Jul 14 21:46:06 locate x86_64-linux-libtool Jul 14 21:46:33 hm cve's for freetype Jul 14 21:46:37 someone should update Jul 14 21:46:58 denix could be Jul 14 21:47:16 nazgee hm Jul 14 21:47:27 whats your target machine? Jul 14 21:47:37 igep0020b Jul 14 21:47:44 beagle clone Jul 14 21:47:52 hm Jul 14 21:48:10 are you sure there is cruft left over? Jul 14 21:48:19 libtool-native worked Jul 14 21:48:37 okay I have to now Jul 14 21:48:40 good nite Jul 14 21:48:46 bye Jul 14 21:49:14 03Denys Dmytriyenko  07org.openembedded.dev * rfc29167bbe 10openembedded.git/recipes/php/php_5.2.13.bb: Jul 14 21:49:14 php: make sed avoid unaccessible .pc files left by quilt Jul 14 21:49:14 Signed-off-by: Denys Dmytriyenko Jul 14 22:07:22 Tartarus: around ? Jul 14 22:08:05 Tartarus: thx for relocatable hint Jul 14 22:39:00 gdb-cross-sdk_7.1 fails to install in a meta-toolchain-* for me because of libiberty.a Jul 14 22:39:36 I guess that's the issue what was fixed by 10fabfa670e4e49a9c9a025d94a87ef0d9c86cc0 gdb: stop conflicting with binutils Jul 14 22:40:26 but on my host (x86_64) I see the troubling libiberty.a is in usr/local/angstrom/arm/lib64 Jul 14 22:41:39 the patch ^ added "rm -f ${D}${libdir}/libiberty.a", but doesn't seem to be enough on my box/on x86_64 Jul 14 22:42:03 addind somehing as ugly as: rm ${D}${libdir}64/libiberty.a fices it shough Jul 14 22:42:07 though Jul 14 22:42:27 but how should I correctly refer to lib64? ${libdir}64 looks ugly Jul 14 22:55:46 khem, there's something wrong with gcc-cross_svn for me, fyi: http://paste.debian.net/80478/ Ideas? I configured like this for qemuppc: http://paste.debian.net/80480/ Jul 14 22:58:08 I'm seeing some more problems in do_populate_sdk Jul 14 22:58:22 * check_data_file_clashes: Package glproto-dev wants to install file .../armv7a-angstrom-linux-gnueabi/meta-toolchain-eccam-1.0-r1/sdk/image Jul 14 22:58:26 //usr/local/angstrom/arm/arm-angstrom-linux-gnueabi/usr/include/GL/internal/glcore.h Jul 14 22:58:28 But that file is already provided by package * libgl-dev Jul 14 22:58:40 is this caused by some recent change, that made latent issues visible? Jul 14 22:58:41 khem, btw.. is that ppc module compilation bug fixed already by now? http://bugzilla.kernel.org/show_bug.cgi?id=11143 Jul 14 23:08:18 blindvt: I think yes Jul 14 23:08:21 its fixed Jul 14 23:12:40 too tired to see why it doesn't like gcc-cross_svn Jul 14 23:12:48 g'night & Jul 14 23:36:50 khem, of course it was gcc-cross_svn.bb ruining PV, thus: http://paste.debian.net/80483/ Jul 14 23:51:15 is gettext on the host OS a requirement for building OE? Jul 15 00:20:10 khem: eh, compressed patches should still be seen as patches without apply=yes Jul 15 00:20:27 zenlinuxPDX, shouldn't be. OE will build a native version. Jul 15 00:22:55 grg, thanks, that's what I thought. I'm working on a new recipe that seems to be insisting on finding the host version of gettext tools like msgmerge Jul 15 00:24:12 zenlinuxPDX, if you make it DEPEND upon gettext-native then those tools should be in the path for the build Jul 15 00:24:26 (i'm guessing) Jul 15 00:34:54 kergoth: a python question Jul 15 00:35:02 I have a list of strings Jul 15 00:35:20 actually I want to compare two lists of strings Jul 15 00:50:00 wtf? g++ is getting oom killed during webkit do_compile. I have 6gb of ram! Jul 15 00:50:16 gcc-4.5 Jul 15 00:59:17 PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ P COMMAND Jul 15 00:59:17 5070 grg 20 0 6477m 5.6g 2016 D 1 95.5 0:11.62 4 cc1plus Jul 15 00:59:30 !!! Jul 15 01:01:31 grg: nice one ;] i wonder what could have happened if it it was java, not cpp ;] Jul 15 01:02:21 nazgee, i'm not going to build openoffice to find out... Jul 15 01:07:43 haha... xchat oom killed Jul 15 01:08:05 added 4g more swap. still no good Jul 15 01:30:14 khem: compare in what way? Jul 15 01:38:48 my journey into trying to build Angstrom fast continues without much promise.... bitbake minimal-image took a real time of 62m54.608s, despite Angstrom, downloads, and my tmp directory all being entirely in a 'ramfs'. :( Jul 15 01:39:55 I can't figure out how I'm still I/O bound when everything is in a 'ramfs'. Jul 15 01:53:01 kergoth: I knocked it somehow Jul 15 01:53:15 kergoth: I am writing something to compare 2 elf files Jul 15 01:53:28 kergoth: for differences Jul 15 01:53:39 kergoth: but not all differences Jul 15 01:53:49 something that could be wrong Jul 15 01:53:54 like function sizes Jul 15 01:54:17 if you compile same file with same compiler twice it should produce identical function size Jul 15 01:54:27 if its different that means something is wrong Jul 15 01:54:51 kergoth: can you think of more things to thats would make sense Jul 15 01:54:57 to compare in ELF Jul 15 01:55:06 hmm Jul 15 01:55:06 this is statically linked ELF file Jul 15 01:55:31 I have got the function size comparision working Jul 15 01:55:33 :) Jul 15 02:47:56 kergoth: when are ${@ evaluated Jul 15 02:48:09 same time ${} are evaluated Jul 15 02:48:14 when used, or immediate if := Jul 15 02:48:21 kergoth: hmmm Jul 15 02:48:29 I am trying to use something in distro.conf Jul 15 02:48:35 which is set in machine.conf Jul 15 02:48:48 and machine.conf is included prior to distro.conf Jul 15 02:49:34 I do this DISTRO_FEATURES_append_arm =' ${@["thumb-interwork",""][bb.data.getVar('THUMB_INTERWORK', d, 1) == "yes"]}' Jul 15 02:49:44 seems not to evaluate Jul 15 02:50:24 in a python function in uclibc-config.inc Jul 15 02:50:30 def features_to_uclibc_settings(d): cnf, rem = ([], []) distro_features = bb.data.getVar('DISTRO_FEATURES', d).split() Jul 15 02:50:46 here I am expecting the above to be expanded Jul 15 02:51:15 but if I print distro_features here Jul 15 02:51:47 ['nfs', 'smbfs', 'wifi', 'ppp', 'alsa', 'ext2', 'vfat', 'pcmcia', 'usbgadget', 'usbhost', 'pci', 'eabi', 'pam', 'largefile', 'ipv4', 'ipv6', '${@["thumb-interwork",""][bb.data.getVar(\'THUMB_INTERWORK\',', 'd,', '1)', '==', '"yes"]}'] Jul 15 02:51:52 thats what I get Jul 15 02:52:18 so it substituted the str Jul 15 02:52:28 but did not evaluate it Jul 15 02:52:39 any help Jul 15 02:57:45 ~ugt Jul 15 02:57:46 i guess ugt is Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html **** ENDING LOGGING AT Thu Jul 15 02:59:56 2010