**** BEGIN LOGGING AT Mon Aug 18 02:59:57 2008 Aug 18 05:00:10 03  07org.openembedded.dreambox * rcfbcfb0a26 10OE.dev/packages/python/python-pybluez_0.15.bb: python-pybluez: add package Aug 18 05:00:10 03  07org.openembedded.dreambox * r3c97876a0c 10OE.dev/packages/dreambox/dreambox-secondstage.bb: dreambox-secondstage: update Aug 18 05:00:11 03  07org.openembedded.dreambox * r8b24fb2616 10OE.dev/packages/gdb/ (gdb-cross_6.8.bb gdb_6.8.bb): gdb: update to gdb-6.8 Aug 18 05:00:12 03  07org.openembedded.dreambox * r0ce6b313d0 10OE.dev/packages/madwifi/ (4 files in 2 dirs): madwifi: update to new hal Aug 18 05:00:15 03  07org.openembedded.dreambox * rdee5e327ea 10OE.dev/packages/enigma2/enigma2.bb: enigma2: depend on gst-plugin-ivorbis, not --vorbis Aug 18 05:00:18 03  07org.openembedded.dreambox * r62b39dc566 10OE.dev/packages/zd1211/ (4 files in 2 dirs): zd1211b: update Aug 18 05:00:23 03  07org.openembedded.dreambox * r27d9079feb 10OE.dev/packages/wlan-rt73/ (4 files in 2 dirs): wlan-rt73: change interface name to wlan, fix build Aug 18 05:00:26 03  07org.openembedded.dreambox * ra64ca06604 10OE.dev/packages/libdreamdvd/libdreamdvd.bb: libdreamdvd: update Aug 18 05:00:29 03  07org.openembedded.dreambox * r808a8d44b1 10OE.dev/packages/initscripts/ (3 files in 3 dirs): initscripts-opendreambox: update Aug 18 05:00:34 03  07org.openembedded.dreambox * r034f638874 10OE.dev/packages/wpa-supplicant/ (9 files in 3 dirs): wpa_supplicant: update to 0.5.8 Aug 18 07:09:31 morning Aug 18 07:55:09 monotone.oe.org working for anyone? Aug 18 07:55:46 mtn: updated to base revision c3b0961cee4c8d8129f4365b8bd96da9e04bdffa Aug 18 07:55:50 was working 30 mins ago Aug 18 07:56:06 but does appear to be dead now Aug 18 07:56:07 :-( Aug 18 07:56:47 oh no, just really really slow Aug 18 07:58:09 yep, just slow. Aug 18 07:58:52 and http://git.openembedded.net/?p=org.openembedded.dev.git;a=summary is stuck at 15 hours ago Aug 18 08:04:35 03  07org.openembedded.dev * r4632894274 10OE.dev/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: Added libnotify and notification-daemon for gestures dependencies Aug 18 08:05:22 03  07master * r4632894274 10OE.dev/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: Added libnotify and notification-daemon for gestures dependencies Aug 18 08:06:24 Laibsch: seems you're getting the same classpath-native failures that I am - any ideas? Aug 18 08:06:29 http://tinderbox.openembedded.net/public/logs/763036.txt Aug 18 08:06:58 It seems that ecj-initial is not creating the directory hierarchy for output files Aug 18 08:07:08 rwhitby: I get those errors as well Aug 18 08:07:09 this works on Ubuntu 7.10, but not on Debian Lenny. Aug 18 08:07:22 rwhitby: on lenny as well :-) Aug 18 08:07:42 XorA: what does OM buildhost run? Can I remove these packages from task-openmoko-feed since they don't build for me ;-) Aug 18 08:07:58 rwhitby: it runs debian stable Aug 18 08:08:09 rwhitby: they are already commented out in OM.dev Aug 18 08:09:02 * rwhitby looks, to see if he can get a clean task-openmoko-feed build by commenting out the same Aug 18 08:09:58 hmm - looks like task-openmoko-feed.bb hasn't had any upstream pushing love for a while ... Aug 18 08:11:37 XorA: does commenting out like that actually work - do the following packages get built? Aug 18 08:13:03 rwhitby: yes, its a handy bitbake parser bug :-) Aug 18 08:20:01 bonjour Aug 18 08:21:20 XorA: what's the plan for org.openmoko.dev pushes back to org.openembedded.dev? Aug 18 08:21:37 it seems to be quite divergent ... Aug 18 08:22:12 waiting for OE git to go live Aug 18 08:24:41 No objections if I change task-openmoko-feed in OE to match what can actually be built there? Aug 18 08:24:45 XorA: ^^ Aug 18 08:24:58 rwhitby: go for it Aug 18 08:25:46 XorA: you may want to add git, navit, libnotify and notification-daemon in OM's version Aug 18 08:29:44 XorA: I'll keep the order the same to make it an easier merge Aug 18 08:30:09 rwhitby: cheers Aug 18 08:30:31 * XorA is not looking forward to merges, checksums.ini is insane Aug 18 08:30:37 I think we need a new way to do that Aug 18 09:24:38 morning all Aug 18 09:25:03 hey RP Aug 18 09:28:08 g'mornin' Aug 18 09:43:48 XorA: I see the mkimage task only applies '-C gzip' when 'if test -e arch/${ARCH}/boot/compressed/vmlinux' fails... Aug 18 09:46:17 but I see I have better compression (but only a few kb) by gzipping zImage.bin by hand and launching mkimage with '-C gzip' Aug 18 09:56:09 well, it can only be the decompressor code get compressed... Aug 18 09:57:43 and the introduced boot-delay is negligible Aug 18 10:01:09 ant__work: we planned it to use kernel decompressor becuase it takes 2s where as u-boot was taking 10s to decompress Aug 18 10:01:26 ant__work: on most devices the couple of K extra is irrelevant Aug 18 10:01:43 10 secs? wich machine ? Aug 18 10:05:05 anyone else have a problem with objdump cross (arm) dissassembling? (segfaults here) Aug 18 10:56:59 03  07org.openembedded.dev * ra4ced5b433 10OE.dev/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: synced with org.openmoko.dev branch Aug 18 10:57:06 03  07master * ra4ced5b433 10OE.dev/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: synced with org.openmoko.dev branch Aug 18 11:08:53 save Aug 18 12:13:39 help: when doing: "bitbake minimal-image" I'm getting: Aug 18 12:13:59 * ERROR: Cannot satisfy the following dependencies for task-boot: Aug 18 12:14:06 * base-passwd * busybox * initscripts * modutils-initscripts * netbase * udev * sysvinit * tinylogin * Aug 18 12:14:50 even after a "bitbake base-passwd busybox initscripts modutils-initscripts netbase udev sysvinit tinylogin" Aug 18 12:45:37 any ideas? Aug 18 12:55:59 dicklebeck: Sounds like configuration or whatever is horribly wrong. This should never happen. Aug 18 13:03:24 yes, this could be it, it feels horrible Aug 18 13:10:17 hi piroko Aug 18 13:15:46 hi woglinde Aug 18 13:16:27 i'll be considered rude, if i ask you if you had any chance to look for the bug 4429? Aug 18 13:22:53 03  07org.openembedded.dev * r028c4535ce 10OE.dev/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: Added u-boot-openmoko Aug 18 13:23:02 03  07master * r028c4535ce 10OE.dev/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: Added u-boot-openmoko Aug 18 13:24:14 diego ah Aug 18 13:24:22 es I worked a little bit further Aug 18 13:24:29 but apr is piece of shit Aug 18 13:24:41 depending on build-directory Aug 18 13:59:07 woglinde, ok thanks Aug 18 14:01:04 03  07org.openembedded.dev * rd30fa16c5c 10OE.dev/packages/freesmartphone/illume-theme-freesmartphone_git.bb: illume-theme-fso: Add config elements to make FSO a complete profile which can be switched Aug 18 14:01:05 03  07org.openembedded.dev * re3ebf3fdc0 10OE.dev/: Aug 18 14:01:05 merge of '2a8702db4281b5dc1be59182ca755ae0bac6b5fa' Aug 18 14:01:05 and 'c7d0711a18c512a5aa6f7085793a18d0849cf9d2' Aug 18 14:01:09 03  07org.openembedded.dev * r94ea981881 10OE.dev/conf/distro/include/sane-srcrevs.inc: sane-srcrevs.inc: Bump illume-theme-freesmartphone Aug 18 14:01:14 03  07master * rd30fa16c5c 10OE.dev/packages/freesmartphone/illume-theme-freesmartphone_git.bb: illume-theme-fso: Add config elements to make FSO a complete profile which can be switched Aug 18 14:01:15 03  07master * re3ebf3fdc0 10OE.dev/: Aug 18 14:01:19 merge of '2a8702db4281b5dc1be59182ca755ae0bac6b5fa' Aug 18 14:01:21 and 'c7d0711a18c512a5aa6f7085793a18d0849cf9d2' Aug 18 14:01:23 03  07master * r94ea981881 10OE.dev/conf/distro/include/sane-srcrevs.inc: sane-srcrevs.inc: Bump illume-theme-freesmartphone Aug 18 14:18:51 hi all Aug 18 14:19:24 hi Unclejack Aug 18 14:19:35 I've got issues with bitbake while using a local overlay, who could help me plz ? Aug 18 14:19:50 Unclejack: what is the issue? Aug 18 14:21:23 simple, I've created an overlay with some stuff inside... today I've created a new package to include in my images recipes, but during the rootfs creation BB says that it cannot find my package Aug 18 14:22:28 Unclejack: did you add the overlay directory to your local.conf and to BBPATH? Aug 18 15:49:58 good morning! Aug 18 15:50:29 I'm running into a strange problem -- hoping someone else has seen this and knows the cause :-) Aug 18 15:51:19 I'm building on 2 machines, kernel builds on one machine work, on the other they fail Aug 18 15:51:57 they fail to build or run ? is the machine that fails 64bit ? Aug 18 15:52:02 The failure mode seems to be lib/firmware files missing from the kernel install directory Aug 18 15:52:46 they are present in the work git/firmware directory, but missing from the work install directory Aug 18 15:53:22 in this case the failing machine is 32 bit, the working is 64 bit! Aug 18 15:54:02 example failure log: http://www.sakoman.net:8000/public/logs/253527.txt Aug 18 15:55:07 The failure seems to have suspiciously started occuring when I updated my kernel builds to 2.6.27-rc3 Aug 18 15:57:06 In comparing the work directories of both machines, the only difference I can see is that the lib/firmware files are missing from the image directory. Aug 18 15:57:11 the /home/sakoman/source/omap3-oe/tmp/staging/i686-linux/usr/bin/install-sh: ./-m0644 does not exist. Aug 18 15:57:14 is suspicious Aug 18 15:57:29 like an empty env var for a command Aug 18 15:57:36 yes, fix that to -m 0644 Aug 18 15:57:49 install.sh doesnt handle args stuck together Aug 18 15:58:37 where do I look to fix that? Aug 18 15:59:36 kernel Makefile somewhere Aug 18 15:59:44 should be able to grep for m0644 :-) Aug 18 16:00:12 what is ttusb-budget Aug 18 16:00:16 maybe look there Aug 18 16:00:33 Crofton|work: I think it is a usb sound driver Aug 18 16:00:58 * mwester encountered that some time ago as well; was a poorly-written package that had its own install.sh that took arguments in that non-standard form. Aug 18 16:00:59 why would it work on one machine and not another? very strange! Aug 18 16:03:31 interesting - 3 places in the kernel tree: http://pastebin.com/m18d157d0 Aug 18 16:03:39 heh, I'm sure something will be obvious once the problem is sorted out Aug 18 16:03:47 the last one looks like the culprit Aug 18 16:04:05 since it is in the firmware instal command line Aug 18 16:04:51 I understand why it fails now, I don't understand why it works on AMD64 Ubuntu! Aug 18 16:05:00 :) Aug 18 16:08:49 hello, anybody knows how to modify rootfs after an installation ? I've tried to use pkg_postinst() but it doesn't work apparently Aug 18 16:09:13 sakoman : It's one of these rare cases where it works on 64bit :) enjoy it as long as it lasts :) Aug 18 16:09:58 seliosk: I've added the space and am now testing on both machines, so the enjoyment will be hopefully short-lived :-) Aug 18 16:12:22 sakoman : wild guess. On the 32 bit do you run the python accelerator ? It won't run on 64bit Aug 18 16:13:01 but on the other hand it should not make any difference on a Makefie Aug 18 16:13:27 steliosk: yes I do use the accelerator on 32 bit Aug 18 16:13:37 hmmm what are the make versions on the machines ? Aug 18 16:15:04 GNU Make 3.81 on both Aug 18 16:15:17 can't be make then Aug 18 16:15:58 on the machine that runs does the file has a 644 attribute ? Aug 18 16:16:21 after the installation that is Aug 18 16:18:01 BTW if anyone is planning to visit Astricon visit and register to www.digital-opsis.com for discount codes and free passes (or late me know and i'll email them to you) :) Aug 18 16:18:16 www/digital-opsis.com/astricon08 Aug 18 16:18:32 * steliosk looks at his fat fingers Aug 18 16:18:56 steliosk: yes, it is 0644 Aug 18 16:19:15 sakoman : I am running out of ideas.... Aug 18 16:19:26 me too :-) Aug 18 16:19:58 builds should complete in a couple of minutes, so we'll know whether adding the space fixes the issue Aug 18 16:22:25 any idea about chowning files to a non-root user ? Aug 18 16:23:31 UncleJack : what do you mean ? after building the rootfs or during the build ? Aug 18 16:24:03 I guess that during the build is feasible, after not Aug 18 16:24:52 what do you think ? Aug 18 16:25:53 still can't understand what you want to do ? chown some files during the rootfs/image is bulding ? Aug 18 16:27:20 simple, I want a /home/user beloging to user, not to root Aug 18 16:30:01 is it feasible ? Aug 18 16:34:07 Unclejack: does it need to be done during the build process or aftwords? Aug 18 16:34:37 Unclejack: either way it should be easy. You can make a package that runs an ipkg postinst script that does adduser. Aug 18 16:36:08 yeah, I could add a script for doing that at first boot Aug 18 16:36:33 I was wondering whether there are other ways Aug 18 16:37:10 steliosk: as expected, adding the space resulted in a successful build on both 32 and 64 bit machines Aug 18 16:39:04 Unclejack: Look at the Asterisk package to see an example of doing adduser postinst via ipk. Aug 18 16:39:45 Unclejack: ppp-dialin_0.1.bb is another simpler example. Aug 18 16:41:07 neat It should work Aug 18 16:41:09 Unclejack: the pkg_postinst function will be run the first time the newly flashed board is booted, or right after the package is installed. Aug 18 16:41:16 ah ok Aug 18 16:41:23 Unclejack : Yes adding a package is one way (maybe the easiest), the other would be to modify rootfs bbclass and do it there Aug 18 16:41:48 do you know the difference Aug 18 16:42:03 between post_inst and post_inst prepend() ? Aug 18 16:43:37 with post_inst you set the post installation routine to be executes with post_inst prepend your post_install is executed prior to the existing one Aug 18 16:44:42 sakoman : well here goes the one thing 64bit had to brag over 32 bit incompatibility :) I wonder why the make does not fail in both cases.... Aug 18 16:45:29 time to go. Later all ! Aug 18 16:46:56 Unclejack: sometimes you have a require statement in a recipe that includes another file (somefile.inc) that has the post_inst already in it, you want to add your own post_inst stuff with out overwritting the existing post_inst so you can use post_inst_prepend or append. Aug 18 16:47:25 understood Aug 18 16:47:27 :) Aug 18 16:56:48 Who do we contact/email to get gittrial access? Aug 18 17:08:01 03  07master * rf49b1d1c0e 10OE.dev/packages/ (4 files in 3 dirs): ffmpeg git : bump SRCREV Aug 18 17:08:02 03  07org.openembedded.dev * rf49b1d1c0e 10OE.dev/packages/ (4 files in 3 dirs): ffmpeg git : bump SRCREV Aug 18 18:19:17 <_hc> hey all, I am trying to build freetype for armv5te and it dies on a libtool error: http://rafb.net/p/misgNW21.html Aug 18 18:19:31 <_hc> it seems that libtool needs something like "--tag=CC" added, Aug 18 18:19:54 <_hc> I'm going to try that know, but I was wondering whether anyone knows anything about this problem Aug 18 18:23:07 * * OE Bug 4504 has been created by hans(AT)eds.org Aug 18 18:23:09 * * freetype building for armv5te dies on libtool Aug 18 18:23:11 * * http://bugs.openembedded.net/show_bug.cgi?id=4504 Aug 18 19:19:04 good day all Aug 18 19:19:57 likewise, likewise Aug 18 19:48:06 re Aug 18 19:48:59 hi ant Aug 18 19:51:55 woglinde: hey Aug 18 19:58:05 * * OE Bug 4505 has been created by lrg(AT)kernel.org Aug 18 19:58:07 * * gtk+-2.12.11-r3 fails to link Aug 18 19:58:09 * * http://bugs.openembedded.net/show_bug.cgi?id=4505 Aug 18 20:23:34 re Aug 18 20:31:05 re florian Aug 18 20:32:00 woglinde: are you lazy on the couch or can assist me with shell/c scripts... Aug 18 20:32:12 *_* ? Aug 18 20:36:46 ant ask Aug 18 20:36:57 aeh show me what you want Aug 18 20:37:02 then I will decide Aug 18 20:37:02 we,, I'm doing pastebin now.. Aug 18 20:37:11 tia Aug 18 20:41:39 woglinde: here is the .bb http://gitweb.openembedded.net/?p=org.openembedded.dev.git;a=tree;f=packages/kexecboot;h=91a78e9f1bf8cd82749229a452741105dfe98437;hb=HEAD and here the sources: http://git.linuxtogo.org/?p=thesing/kexecboot.git;a=tree Aug 18 20:42:44 woglinde: here is my first hack: http://www.pastebin.ca/1177975 Aug 18 20:43:22 woglinde: the question is: how to pass that filtered cmdline to kexecboot ? Aug 18 20:43:41 me I would add a -c --cmdline option Aug 18 20:44:14 bah Aug 18 20:44:30 thats the problem thesing and I didnt found the solution Aug 18 20:44:31 I got it work hacking with getenv Aug 18 20:44:45 in kexecboot.c (pastebin follows) Aug 18 20:45:40 this is the first ugly hack http://www.pastebin.ca/1177979 Aug 18 20:45:48 now the 'better' one: Aug 18 20:46:40 http://www.pastebin.ca/1177981 Aug 18 20:48:19 uups.. I did add strlen but did not correct the line...cannot work Aug 18 20:50:43 hm but thats not a shell script Aug 18 20:53:28 well I parse /proc/cmdline during the init Aug 18 20:53:51 thesing suggested to write it in C, but it's overwork Aug 18 20:55:03 it's just about the best way to pass the filtered string (could be empty) Aug 18 20:55:30 c is fine Aug 18 20:56:00 well, reinvent the parser Aug 18 20:56:09 *_* Aug 18 20:57:48 I think one thirg arg (optional) could be easier to implement Aug 18 20:58:01 but I'll ask thesing Aug 18 20:58:30 (finally it's his work :-) Aug 18 21:32:20 Who do I contact/email to get gittrial access? Aug 18 21:35:18 laibsch I think Aug 18 21:40:29 woglinde: thanks. Aug 18 21:44:05 nite Aug 18 22:03:17 svolpe: RP is the one Aug 18 22:03:30 But I'd just wait until the real git rollout Aug 18 22:06:12 svolpe: you can pull the git repo without a key, so that at least allows you to use git repos for your work Aug 18 22:08:28 03  07org.openembedded.dev * rb6adcdc1cd 10OE.dev/packages/phoneme/phoneme-advanced.inc: phoneme-advanced.inc: Allow overriding the PhoneME profile. Aug 18 22:08:36 03  07master * rb6adcdc1cd 10OE.dev/packages/phoneme/phoneme-advanced.inc: phoneme-advanced.inc: Allow overriding the PhoneME profile. Aug 18 23:08:33 * kergoth`work plays around with topgit & pristine-tar Aug 18 23:15:19 Laibsch: I emailed RP last week, I figured he did not reply because he was not in charge of it :-) I will wait a couple more days and ping him on it. Aug 18 23:15:46 Please don't Aug 18 23:15:53 Testing is basically over Aug 18 23:15:57 Laibsch: ok. Aug 18 23:16:02 IOW, this will create unnecessary work Aug 18 23:16:11 which is probably why RP ignored it Aug 18 23:16:31 Laibsch: thanks for the info, I will just wait then. Aug 18 23:16:53 Laibsch: I don't want to cause extra work. Aug 19 00:24:06 hmmm **** ENDING LOGGING AT Tue Aug 19 02:59:56 2008