**** BEGIN LOGGING AT Sat May 31 02:59:57 2008 May 31 03:20:14 hello all May 31 03:20:59 i am a complete newbie to bitbake, and my configure step is failing with the error "cross compiler not found" - any ideas? May 31 03:48:36 alternate question: is there a way to set all the environment variables and such in the terminal so that i can try running the configure step manually? May 31 03:58:19 hello May 31 03:58:32 I amtrying to build navit on angstrom (2007.12 and 2008.1) May 31 03:58:39 on stable I get bug http://bugs.openembedded.net/show_bug.cgi?id=4334 May 31 03:59:06 on dev I get bug http://bugs.openembedded.net/show_bug.cgi?id=4344 May 31 03:59:19 (after working arond a few other issues, that is) May 31 03:59:37 in short it means I cannot build my program. May 31 03:59:45 stable == dev May 31 04:00:07 how can I help fix the bugs? May 31 04:00:50 the dbus bug is weird: May 31 04:00:51 bus-sysdeps-unix.c:996: error: storage size of 'cr' isn't known May 31 04:00:51 dbus-sysdeps-unix.c:999: warning: pointer targets in passing argument 5 of 'getsockopt' differ in signedness May 31 04:00:51 dbus-sysdeps-unix.c:996: warning: unused variable 'cr' May 31 04:01:08 cr is not used and the storage size *must* be known? May 31 04:09:53 The other bog is also strange as it is part of a kernel compile. May 31 04:10:13 aside from the reasons of *building* a kernel as dependency of an app May 31 04:10:25 (extract the code and you have what you need) May 31 04:10:38 it's strange that it fails in the way reported May 31 09:12:06 * * OE Bug 4347 has been created by udovdh(AT)xs4all.nl May 31 09:12:08 * * fakeroot missing May 31 09:12:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4347 May 31 09:23:26 yay May 31 09:23:33 I found fakeroo elsewhere May 31 09:23:38 s/roo/root/ May 31 09:23:50 and finished building navit 0.0.3 with the GENERIC distro May 31 09:23:52 hi May 31 09:23:56 hello woglinde May 31 09:24:10 how can I get the generic distro to build the latest cvs for navit? May 31 09:24:27 (same for angstrom if we can get the bugs fixed...) May 31 09:36:28 udovdh edit your local.conf and set PREFERRED_VERSION_navit = navit_cvs May 31 09:36:56 thats override the generic distro May 31 09:37:20 than look at the navit_cvs.bb if contains a revision and edit them May 31 09:50:43 thanks woglinde May 31 09:51:04 saw that navit is gotten from cvs May 31 09:51:13 just version is added to the ipkg name May 31 09:51:56 morning all May 31 09:53:23 hi cyril May 31 10:03:00 slapin_nb: That question was asked on linux-arm-kernel recently May 31 10:09:37 hi rp May 31 10:10:41 hi woglinde, RP May 31 10:11:01 Missing or unbuildable dependency chain was: ['console-image', 'task-base-extended', 'kernel-module-zd1211rw', '${INITRAMFS_IMAGE}'] May 31 10:11:08 What's the fix for this? May 31 10:11:30 hm May 31 10:11:41 INITRAMFS_IMAGE isn't set May 31 10:11:42 I guess kernel-module-zd1211rw isnt anywhere May 31 10:11:46 hm okay May 31 10:12:53 I don't have one :) May 31 10:13:10 Do machines that do not init ram fs' need to set it to ""? May 31 10:14:11 Crofton|work: I guess there should be an INITRAMFS_IMAGE ?= "" somewhere May 31 10:14:31 yeah, I'm not sure where May 31 10:14:41 Where is it used? May 31 10:14:55 task-base I think May 31 10:14:59 grepping now May 31 10:16:10 packages/tasks/task-base.bb: ${@base_contains('COMBINED_FEATURES', 'usbhost', 'kernel-module-zd1211rw', '',d)} \ May 31 10:16:30 Thats not where INITRAMFS_IMAGE is used May 31 10:17:08 well, thats the only hi from grep that would go into console-image ... May 31 10:17:32 urg May 31 10:17:37 thats the kernel module May 31 10:17:42 that worries me also May 31 10:18:19 INITRAMFS is in kernel.bbclass May 31 10:18:35 it's early May 31 10:18:45 Thats where it should be set to a sane default then May 31 10:21:21 new problem May 31 10:22:02 http://rafb.net/p/1OszR377.html May 31 10:22:32 oeh May 31 10:22:45 ah, I know why this happens :/ May 31 10:24:38 Crofton|work: In that anonymous function its setting INITRAMFS_TASK to ":do_rootfs" May 31 10:24:50 Crofton|work: That function needs fixing May 31 10:25:10 this is "greek" to me May 31 10:25:21 with apologies to stelios May 31 10:26:21 Crofton|work: Remove the ?= and change the code to something like if bb.data.getVar(...) is not None and is not "": May 31 10:26:49 To python I suspect None != '' May 31 10:30:47 can you be alittle clearer what I need to type in to kernel.bbclass? May 31 10:31:47 03koen 07org.oe.dev * r870d6ff5... 10/ (8 files in 5 dirs): May 31 10:31:47 opkg: update to 0.1.4 May 31 10:31:47 packagekit: update to 0.2.1 to match opkg API (with addional patch from openmoko) May 31 10:31:57 Crofton|work: http://rafb.net/p/1Lafz371.html May 31 10:32:08 anybody here who can help me with bug #4334? May 31 10:32:11 Crofton|work: and remove the ?= "" you added whereever May 31 10:33:06 i.e.: http://bugs.openembedded.net/show_bug.cgi?id=4334 May 31 10:33:12 ah, you were perfactly clear :) May 31 10:34:02 ERROR: 85:SyntaxError:' if bb.data.getVar('INITRAMFS_IMAGE', d, True) is not None and is not '': May 31 10:34:24 gah May 31 10:34:37 my thoughts exactly May 31 10:35:19 http://rafb.net/p/goP7IA39.html May 31 10:35:56 what are the main differences between the generic distro and angstrom-2007.1? May 31 10:36:09 i.e.: why does building for generic work and fail for angstrom-2007.1? May 31 10:36:18 I am trying to build navit May 31 10:37:30 udovdh: It looks like something on your host system is confusing the dbus build May 31 10:37:46 what could that be? May 31 10:37:54 I am on a more or less standard Fedora 9 box May 31 10:37:58 something header related perhaps... May 31 10:38:02 and should the host confuse the bitbake environment? May 31 10:38:12 hm maybee amd64 problem May 31 10:38:19 I am on amd64 yes May 31 10:38:27 It shouldn't but the native packages are influenced by the system May 31 10:38:40 We do try and minimise the influence May 31 10:38:42 hmm.. but how to fix or circumvent? May 31 10:38:55 that is the question May 31 10:38:59 hm I have to look at the code May 31 10:39:07 maybe you must define cr yourself May 31 10:39:36 how could I put that into the bitbake machinery? May 31 10:39:42 RP, now I hvae an indentation error May 31 10:39:53 I can apply and make a patch (if I know what/how to code) May 31 10:39:59 right May 31 10:40:15 i.e. I have experience with vi, diff, patch and make/gcc May 31 10:40:29 but this is a step further May 31 10:40:59 use quilt May 31 10:41:32 oe uses quilt for patching the code too May 31 10:44:46 ah. hmmm. May 31 10:44:53 hm oh May 31 10:45:00 in dev is dbus 1.2.1 May 31 10:45:09 maybee this fixed smoething May 31 10:45:23 I tried to build navit from dev May 31 10:45:28 i.e. angstrom 2008-1 May 31 10:45:35 and that bombed on a different bug May 31 10:45:48 http://bugs.openembedded.net/show_bug.cgi?id=4344 May 31 10:46:46 Crofton|work: make sure the lines are four space indented May 31 10:47:20 udovdh: I seem to remember off errors like that with some versions of dbus May 31 10:47:38 You could see which if a different version of dbus works May 31 10:47:41 RP could I force stable to use a different version of dbus? May 31 10:47:46 ah. May 31 10:47:50 how to force? May 31 10:48:01 I am now running a stable build again, it will stop on dbus May 31 10:48:10 then I can retry with a different version May 31 10:48:16 Add a recipe for the version you want, remove the ones you don't want May 31 10:48:25 There are more compliacted ways of course :) May 31 10:48:31 which version should work? May 31 10:49:04 I'm using 1.1.20 May 31 10:49:15 before that 1.0.3 worked for me May 31 10:49:25 but this was with poky, not OE May 31 10:49:34 stable is at 1.0.1 so try 1.0.3 for now? May 31 10:49:36 not that it probably makes much difference in this case May 31 10:49:50 I go for 1.0.3 May 31 10:49:51 you could, yes May 31 10:50:15 hm 1.2.1 for i386 build fines May 31 10:51:12 so i copy dbus-native_1.0.1.bb to dbus-native_1.0.3.bb May 31 10:51:15 remove dbus-native_1.0.1.bb May 31 10:51:26 and change references from 1.0.1 to 1.0.3 in dbus-native_1.0.3.bb May 31 10:52:13 it's still dying May 31 10:52:29 I'm just randomly guessing at indentation at this point May 31 10:53:03 corfton mtd diff May 31 10:53:05 aergs May 31 10:53:07 mtn May 31 10:53:25 mtd? got issues with mtd on h2200 May 31 10:53:31 flashing don't work May 31 10:53:44 http://bugs.openembedded.net/show_bug.cgi?id=4324 May 31 10:53:44 http://rafb.net/p/SAIzVu12.html May 31 10:53:58 morniing May 31 10:54:55 Crofton|work: four spaces before "image =" and the "if ...", 8 before the bb.data.setVar May 31 10:55:02 hi methril May 31 10:55:42 RP i tried some of your poky receipes for OSX but, finally i reboot in my Linux system :D May 31 10:56:10 i have a UMTS connection for the moment and it didn't work, but now i could work on it May 31 10:56:28 methril: OSX is a work in progress... May 31 10:56:39 wes, i could see May 31 10:56:45 indentation error May 31 10:56:47 i'd like to help May 31 10:57:07 the ':' at the end of the forst if looks funny May 31 10:57:41 methril: I don't have an OSX system but I'll take any patches and review them May 31 10:57:48 ok May 31 10:57:48 Crofton|work: Remove the first if May 31 10:57:55 i'll try next reboot May 31 10:57:56 :D May 31 10:58:05 methril: :) May 31 10:58:49 Crofton|work: You're misapplied that patch somehow.... May 31 10:58:59 I am doing it by hand May 31 10:59:07 becuase I am an idiot May 31 10:59:15 and I thought it would be "easier" May 31 11:00:08 Crofton|work: I think this problem is coming from somewhere else, I'm also seeing a build failure May 31 11:00:15 well, backtrace from bitbake May 31 11:00:27 good :) May 31 11:00:36 I'm going to revert May 31 11:00:58 I need to rest the arm anyway May 31 11:01:07 and get more coffee May 31 11:03:56 someone mannaging the minimal images? May 31 11:04:19 ./packages/linux/linux-rp.inc:INITRAMFS_IMAGE=initramfs-kexec-image May 31 11:04:22 Grrrr May 31 11:04:44 Perhaps someone could have mentioned that little gem to me before committing it? May 31 11:08:26 Laibsch: you were looking for me yesterday? May 31 11:12:29 Crofton|work: Fix tested and pushed May 31 11:14:12 03rpurdie 07org.oe.dev * r0380aa91... 10/ (1 classes/kernel.bbclass): classes/kernen.bbclass: Fix INITRAMFS logic to stop breaking builds for people not using it May 31 11:17:20 danke schoen for the fix May 31 11:23:38 thanks RP , build is going now May 31 11:42:18 crofotn did you read that? http://kernelslacker.livejournal.com/124350.html *g* May 31 11:47:39 rwhitby: yes May 31 11:47:51 I believe you are the best person to ask this May 31 11:49:28 I am wondering what NAS device you'd recommend. slug won't do it because it does not house the HD. Is there anything below $50-75 that runs a full Linux? May 31 11:52:44 interesting q May 31 11:57:11 * cyrilRomain didn't noticed the http://www.gumstix.com/ main page displays angstrom ;-) May 31 11:57:18 you see, I got myself a NAS recently for considerably less than that. It runs some proprietary OS and that of course is annoying at times. But it has also limited my willingness to spend. It seems that this should be doable and I am not really aware of any low-cost device as successor to the slug which was released a quite some time ago, wasn't it? May 31 11:57:39 http://blog.leggewie.org/?p=42 May 31 12:30:01 another dbus followup May 31 12:30:07 I now get: May 31 12:30:07 IOError: [Errno 2] No such file or directory: '/usr/src/ipaq/org.openembedded.stable/packages/dbus/dbus-native_1.0.1.bb' May 31 12:30:08 ERROR: Task 689 (/usr/src/ipaq/org.openembedded.stable/packages/dbus/dbus-native_1.0.1.bb, do_fetch) failed May 31 12:30:57 so it still expects 1.0.1 to be used May 31 12:32:18 workaround: rename the bb for 1.0.3 to the 1.0.1 name? May 31 12:32:42 udovdh: no, you probably need to change PREFERRED_VERSION_dbus to 1.0.3 May 31 12:32:43 it then continues, but does it find the sources OK? May 31 12:32:50 in my build.conf? May 31 12:33:10 err local.conf May 31 12:34:59 udovdh: yes May 31 12:35:12 ok, trying that May 31 12:37:27 udovdh: why don't you build against angstrom-2007.1 ? Angstrom uses dbus 1.0.2 in stable May 31 12:37:43 I am using angstrom 2007.1 May 31 12:37:48 and it uses 1.0.1 May 31 12:37:56 so what could be wrong? May 31 12:38:24 or at least we assume it uses 1.0.1 as that is the dbus-native recipe May 31 12:38:43 udovdh: angstrom uses dbus-native 1.0.1, but dbus 1.0.2 for the target device May 31 12:39:05 and it fails at 1.0.1 native May 31 12:39:14 so I want to tackle that hurdle first May 31 12:39:26 udovdh: ok so then it's 'PREFERRED_VERSION_dbus-native = "1.0.3"' that you need May 31 12:39:27 1.0.3 is supposed to work May 31 12:39:34 ah, ok May 31 12:39:36 thanks for the tip May 31 12:40:05 udovdh: you may want to stick with 1.0.2 for your device May 31 12:40:26 udovdh: if not, also define 'PREFERRED_VERSION_dbus = "1.0.3"' May 31 12:40:38 thanks. May 31 12:40:48 if 1.0.2 fails I will try 1.0.3 May 31 12:40:53 thanks again May 31 12:53:03 back again May 31 12:53:10 dbus-1.0.3 is used May 31 12:53:13 but patching fails: May 31 12:53:14 NOTE: Applying patch 'cross.patch' (/usr/src/ipaq/org.openembedded.stable/packages/dbus/dbus-1.0.3/cross.patch) May 31 12:53:14 NOTE: Task failed: Error: /usr/src/ipaq/org.openembedded.stable/packages/dbus/dbus-1.0.3/cross.patch not found. May 31 12:53:14 NOTE: package dbus-native-1.0.3-r0: task do_patch: failed May 31 12:53:39 comment out the patch? May 31 12:54:40 trying... May 31 12:55:53 udovdh: you may need to create a packages/dbus/dbus-1.0.3/ folder with updated patches from dbus-1.0.2 May 31 12:56:42 dbus 1.0.1 just had that one patch May 31 12:56:49 will look into it May 31 13:02:09 dbus 1.0.3 gives the same error May 31 13:03:06 refresh http://bugs.openembedded.net/show_bug.cgi?id=4334 for details May 31 13:03:17 what dbus version is next candidate for dbus-native? May 31 13:04:51 trying 1.1.20... May 31 13:07:06 as I said May 31 13:07:35 try 1.2.1 from dev May 31 13:11:10 dbus 1.1.20 gives different error. May 31 13:11:14 will try 1.1.20; May 31 13:11:19 what about necessary patches? May 31 13:16:14 copied whole recipe from .dev May 31 13:24:20 udovdh: about http://bugs.openembedded.net/show_bug.cgi?id=4344 it seems scripts/mod/sumversion.c needs a '#include ' May 31 13:24:45 s/limit.h/limits.h/ May 31 13:24:59 ah, thanks May 31 13:25:09 could you put that in dev? May 31 13:25:19 I am now trying the stable repo May 31 13:25:33 patching dbus-1.2.1 doesn't work May 31 13:25:36 yet May 31 13:25:45 ???? May 31 13:26:19 maybee you forgot to copy the patches too May 31 13:26:37 nope May 31 13:26:42 inc file for dev is needed as well May 31 13:28:18 stable inc doesn't work apparently; compile runs now May 31 13:28:38 appears to be ok, dbus-glib is busy now May 31 13:28:42 hmmm May 31 13:39:12 dbus compile bug appears to be gone with 1.2.1 May 31 14:31:15 he florian May 31 14:34:37 hi woglinde May 31 14:34:52 woglinde: sounds like you made it back home ;) May 31 14:35:00 yes May 31 14:35:02 *g* May 31 14:35:16 was at home befor 10 p.m. May 31 14:35:34 so I heard the aertze konzerte May 31 14:35:44 so... May 31 14:35:57 the navit build finally succeeds more or less for angstrom 2007.1 May 31 14:35:57 because the stage is approx 1 km from my home May 31 14:36:05 I now get, as finals steps: May 31 14:36:07 NOTE: QA checking staging May 31 14:36:07 ERROR: QA Issue: libgmodule-2.0.la failed sanity test (workdir) in path /usr/src/ipaq/angstrom-tmp//staging/arm-angstrom-linux-gnueabi/lib May 31 14:36:07 NOTE: a QA error occured but will not be logged because QA_LOG is not set May 31 14:36:27 how serious are these? (I got multiple of these lines) May 31 14:36:34 can they be avoided? May 31 14:37:12 udovdh: yes they can, and they should be avoided even May 31 14:37:32 how? May 31 14:39:56 shall I make it a bug? or? May 31 14:41:59 sed can be used to fix reference to working path, but not sure this is the best way to fix May 31 14:42:11 udovdh: yes you can report it as a bug May 31 14:42:26 ok May 31 14:44:09 does this mean the binaries I just buult are bad? May 31 14:44:16 s/buult/built/ May 31 14:44:50 udovdh: QA is here to prevent not always fatal but often weird bugs May 31 14:45:05 * * OE Bug 4348 has been created by udovdh(AT)xs4all.nl May 31 14:45:08 * * QA errors after bitbake navit May 31 14:45:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4348 May 31 14:45:20 hmmm. so every QA is one too many May 31 14:45:33 I used the stable tree May 31 14:45:41 but had to do some fixes to make stuff work May 31 14:45:55 biggest is to dbus-native (change to 1.2.1) May 31 14:46:40 udovdh: feel free to report bugs (and possibly with patches) to the bugtracker May 31 14:46:55 udovdh: you can also send mails to the stable branch mailing list May 31 14:46:56 patches are harder May 31 14:47:05 well May 31 14:47:18 I am doing a small wiki page now for the navit project with my changes May 31 14:47:26 then I can use that page to show what I did May 31 14:47:35 and tehn ask people to look into issues May 31 14:47:38 udovdh: good May 31 14:47:44 s/tehn/then/ May 31 14:53:37 cyrilRomain, openembedded-devel@lists.openembedded.org ? May 31 14:53:41 udovdh: BTW, you might want to test navit 0.0.4 from .dev May 31 14:53:54 I just built cvs May 31 14:54:13 but I will give it a go May 31 14:54:21 anyway, problems with .stable shoudl be reported :) May 31 14:54:28 listname is OK? May 31 14:55:00 udovdh: openembedded-stablebranch@openembedded.org May 31 14:55:15 udovdh: openembedded-devel@lists.openembedded.org for .dev May 31 14:55:25 and other oe devel discussions May 31 14:57:44 stablebranch is not mentioned on the website? May 31 14:57:54 also the confirmation email is slower than for devel May 31 14:58:00 (still not in..) May 31 14:58:17 udovdh: it is May 31 14:58:30 ah it does wore :-) May 31 14:58:32 work.. May 31 14:58:33 udovdh: the second post :) May 31 15:06:42 cyrilRomain|bbl, email sent May 31 15:29:10 where can I find what tests are performed in QA testing and which ones fail? May 31 15:31:16 udovdh: classes/insane.bbclass May 31 15:31:26 danke schoen! May 31 15:32:42 can I replay the bitbake to make the QA tests more verbose? May 31 15:32:49 to see exactly what is found? May 31 15:33:30 udovdh: add 'QA_LOG = "1"' in your local.conf May 31 15:33:40 danke May 31 15:33:45 and then rerun the bitbake? May 31 15:34:02 so that qa error are logged in temp directory with other log files (log.qa.*) May 31 15:34:47 udovdh: yeah, but you will need to clean the package if it built successfully since QA errors does not make the build fail May 31 15:35:00 aha May 31 15:35:57 udovdh: 'bitbake -c clean thepackage && bibake thepackage', or 'bitbake -c rebuild thepackage' May 31 15:36:18 ah... even better! May 31 15:36:41 udovdh: prefer the former if you have a old bitbake version May 31 15:37:04 udovdh: there was a problem with the rebuild, fixed some weeks ago May 31 15:52:23 ping rwhitby May 31 16:34:52 Hi guys, one of the oe wiki pages is spammed. can someone with write access restore last good version please? "OEandYourDistro" May 31 17:04:23 will "bitbake nano" build a "special" version of tools like libtool, m4, etc? is that normal? (it is doing at the moment) May 31 17:04:55 bison, gettext too. well, nevermind, maybe I'll understand this a few hours later... May 31 17:05:36 it is building its own tool-chain, yes May 31 17:06:14 hm, ok thanks, I hope it will succeed :) May 31 17:15:28 03koen 07org.oe.dev * r82171c9e... 10/ (3 files in 3 dirs): dbus 1.2.1: fix pidfile location May 31 18:29:45 finally managed to get angstrom console-image built for my ts7260 and it's working like a charm. keep up the goog work! May 31 18:30:32 kario fine May 31 18:30:37 stable or dev? May 31 19:12:17 dev May 31 19:17:15 kario fine May 31 19:59:17 03trini 07org.oe.stable * r29c205b8... 10/ (26 files in 3 dirs): May 31 19:59:17 glibc, linux-libc-headers: sort out scsi headers May 31 19:59:17 http://lists.linuxtogo.org/pipermail/openembedded-stablebranch/2008-May/000052.html May 31 19:59:17 Acked-by: Koen Kooi May 31 19:59:17 Acked-by: Leon Woestenberg May 31 21:12:34 if i build build invidual packages, how do i update the corresponding Packages.gz file under deploy/glibc/ipk/. trying to setup my own ipkg feed May 31 21:13:18 bitbake package-index May 31 21:20:32 \o/ works **** ENDING LOGGING AT Sun Jun 01 02:59:56 2008