**** BEGIN LOGGING AT Tue Feb 01 02:59:57 2011 Feb 01 03:28:19 khem you around? Feb 01 07:57:04 03Stefan Schmidt  07master * r644881e89c 10openembedded.git/recipes/connman/connman_0.68.bb: connman: Add recipe for 0.68 release Feb 01 08:11:14 hi all Feb 01 08:47:29 good morning Feb 01 08:49:02 hi mckoan Feb 01 09:05:09 03Koen Kooi  07org.openembedded.dev * re2cdad70c4 10openembedded.git/recipes/angstrom/angstrom-uboot-scripts/ (17 files): Feb 01 09:05:09 angstrom-uboot-scripts: add some more scripts Feb 01 09:05:09 Signed-off-by: Koen Kooi Feb 01 09:50:31 huhu Feb 01 09:55:33 good morning Feb 01 09:56:24 florian: I wish to remind about kexecboot-devel password :) Feb 01 09:56:46 florian: thesing lost his password long time ago, can you reset it? Feb 01 09:57:20 in guess so... at least i hope Feb 01 09:57:56 thanks Feb 01 11:43:00 hi , I cloned the deveopment kernel git clone git://arago-project.org/git/people/martin/linux-da830.git and built it with the configuration for Video VPIF Driver ... when I load it on the hawkboard with tftp server it doesnt proceed beyond the Starting kernel ... I compiled the kernel wiht hte mkimage of the u-boot-omapl1_v1 which i procurred from the Feb 01 11:43:05 http://elinux.org/Hawkboard#Sources Feb 01 11:43:17 what went wrong any idea Feb 01 13:40:43 where is the best place to manage an /etc custom population, which package is usually responsible to that? Feb 01 13:41:25 initscripts ? Feb 01 13:41:37 mwester: thank you Feb 01 13:42:53 You can also take a look at slugos-init for an example of how SlugOS installs a lot of very distro-specific utils (although not specific to /etc). Feb 01 13:45:38 mwester: very interesting thx Feb 01 15:19:18 03Tom Rini  07org.openembedded.dev * r61813317fc 10openembedded.git/site/ (14 files): Feb 01 15:19:18 site infos: Move ac_cv_sys_restartable_syscalls to common-linux Feb 01 15:19:18 Signed-off-by: Tom Rini Feb 01 15:19:19 03Tom Rini  07org.openembedded.dev * red3c8fff4f 10openembedded.git/site/sh-common: Feb 01 15:19:19 sh-common: Add postgresql info Feb 01 15:19:19 Signed-off-by: Tom Rini Feb 01 15:37:19 03Marek Sujak  07org.openembedded.dev * r66f62eb7fc 10openembedded.git/recipes/barebox/barebox-2010.10/vmx25/ (barebox_2010.10-vmx25-20110112.patch defconfig): Feb 01 15:37:19 barebox: add patch for vmx25 Feb 01 15:37:19 * memory timing, custom init scripts, mach id's Feb 01 15:37:19 * default config file Feb 01 15:37:19 * tested with angstrom 2010.x eglibc compiler Feb 01 15:37:19 Signed-off-by: Marek Sujak Feb 01 15:37:19 Signed-off-by: Eric BĂ©nard Feb 01 15:40:58 morning Feb 01 15:51:08 ka6sox-away: hello Feb 01 15:53:31 mornin mckoan hoping we might find the source of the bug in tinderbox today. Feb 01 16:00:25 ok guys, I'll be leaving for FOSDEM woon Feb 01 16:00:26 soon Feb 01 16:00:37 will even have time for some sight seeing, weather permitting Feb 01 16:01:00 hopefully I do not get caught in the storm headed to the east coast Feb 01 16:45:06 03Tom Rini  07master * r2e1e7af403 10openembedded.git/site/ (10 files): Feb 01 16:45:06 site info: Move ac_cv_conv_longlong_to_float to common-linux Feb 01 16:45:06 Signed-off-by: Tom Rini Feb 01 16:45:52 eFfeM_work, ping? Feb 01 16:46:16 eFfeM_work, you ack'd all of my perl stuff except 2/5, intentional or oversight? :) Feb 01 16:46:20 Getting ready to push now Feb 01 16:58:55 Jay7, is tinderbox working right? Feb 01 17:53:28 03Tom Rini  07master * r49362da572 10openembedded.git/recipes/perl/perl-native_5.10.1.bb: (log message trimmed) Feb 01 17:53:28 perl-native: Switch to userelocatableinc and not shared libperl Feb 01 17:53:28 While in here, we didn't use db nor gdbm so remove those as deps. Feb 01 17:53:28 This recipe is safe with no other deps so add INHIBIT_DEFAULT_DEPS Feb 01 17:53:28 and switch to patch as PATCHTOOL. Being relocatable means we don't need Feb 01 17:53:29 the config.sh mangling we had been doing and we want to be more clear Feb 01 17:53:30 about our mangling of scripts to use /usr/bin/env perl Feb 01 17:53:30 03Tom Rini  07master * r01996cb112 10openembedded.git/conf/bitbake.conf: Feb 01 17:53:31 bitbake.conf: Drop unused perl-native-runtime ASSUME_PROVIDED Feb 01 17:53:31 Acked-by: Koen Kooi Feb 01 17:53:32 Acked-by: Martin Jansa Feb 01 17:53:32 Acked-by: Frans Meulenbroeks Feb 01 17:53:33 Signed-off-by: Tom Rini Feb 01 17:53:34 03Tom Rini  07master * r55815b5ef0 10openembedded.git/recipes/autoconf/autoconf.inc: Feb 01 17:53:35 autoconf: Add perl-native as a dep Feb 01 17:53:35 Acked-by: Koen Kooi Feb 01 17:53:35 Acked-by: Martin Jansa Feb 01 19:33:43 ka6sox-away: hey Feb 01 19:34:00 tinderbox.oe.org showing me bugzilla page :) Feb 01 19:34:32 Jay7, dunno who did that... Feb 01 19:36:10 ka6sox-away: I think it is just falling into previous or default virtual host Feb 01 19:36:16 or next Feb 01 19:37:09 it shouldnt' its been SET that way...wonder why? Feb 01 19:37:17 wierd. Feb 01 19:37:26 okay...turning off and then trying again. Feb 01 19:53:04 khem ping Feb 01 20:21:42 Hello can anybody tell me where terminal messages are logged on Arago instead of /var/log/messages please? Feb 01 20:23:42 bertszoghy: cat /etc/syslog.conf ? Feb 01 20:25:40 bertszoghy: I replied to you 15 minutes ago... Feb 01 20:27:00 actually, more like 30 minutes ago... anyway - Hello Denix! Can you tell me where is /var/log/messages on Arago please? hey! do you have syslog running? and by default, it logs into memory buffer. you need to change it in /etc/syslog.conf change DESTINATION="buffer" into DESTINATION="file" - that should do it Feb 01 20:30:46 thanks denix, syslog was already running but the first line in /etc/syslog.conf was not set to file. Workss now, thanks! Feb 01 20:34:35 bertszoghy: any time... :) Feb 01 20:36:44 I want kernel include files to be available outside the bb process, ideally already in the root fs, but inside the modules tgz is good, or anything else that works. Feb 01 20:36:57 ideas? Feb 01 22:16:58 jo Feb 01 22:17:00 lol Feb 01 22:17:08 now clutter 1.6 is released Feb 01 22:28:10 03Tom Rini  07master * r01ef753e95 10openembedded.git/recipes/postgresql/ (5 files): Feb 01 22:28:10 postgresql: Pass --disable-spinlocks on SH4, use INC_PR Feb 01 22:28:10 Signed-off-by: Tom Rini Feb 01 22:31:29 zecke? Feb 01 22:31:33 no sleep? Feb 01 22:31:51 So.. Feb 01 22:31:59 hi tom Feb 01 22:32:08 Anyone around that cares for qt 4.6.x ? Feb 01 22:32:10 hi woglinde Feb 01 22:32:25 tartarus better write it to the ml Feb 01 22:32:28 Yeah Feb 01 22:32:35 there maybee some companys Feb 01 22:32:40 Just wondering if anyone off hand could answer my next Q Feb 01 22:32:47 4.7.x we use -xmlpatterns Feb 01 22:32:50 But not 4.6.x Feb 01 22:33:04 So qt4-x11-demo-image / qt4e-demo-image fail Feb 01 22:33:11 hm Feb 01 22:33:18 (They include task-whatever that spells out libqt4xmlpatterns) Feb 01 22:33:26 maybee it was forgotten or qt 4.6 dont hav it Feb 01 22:33:27 Haven't even tried passing -xmlpatterns to 4.6.x yet Feb 01 22:33:34 But a quick google shows it may work Feb 01 22:33:42 Will try and RFC the change on the ML, if it does Feb 01 22:33:57 (this is why in my autobuilder angstrom 2010.x builds the qt demo images, but not 2008.1) Feb 01 22:34:12 * Tartarus has a qemumips64 build going right now tho Feb 01 22:39:12 Tartarus: I care for Qt 4.6.x :) Feb 01 22:40:05 denix why? Feb 01 22:40:09 mostly from a "do not remove" perspective, though :) Feb 01 22:40:13 it doesnt have quick Feb 01 22:40:16 denix, know the anwswer to my Q off hand? Feb 01 22:40:46 Tartarus: not off hand, sorry. need to go through the configs, but we don't use xmlpatterns flag Feb 01 22:40:57 woglinde_: some of our products use qt-4.6.x Feb 01 22:40:58 yeah Feb 01 22:41:02 i'll test adding it soon Feb 01 22:41:23 * Tartarus made the change and will queue the build Feb 01 22:41:52 denix they dont work under 4,7? Feb 01 22:43:24 good nite Feb 01 22:43:28 some of them will, but some still need to be validated... people are afraid of any regressions with 4.7 - there might be some small performance degradation, which is not good. Feb 01 22:43:40 ? Feb 01 22:43:50 I think it will be the other way round Feb 01 22:43:57 and you get more performance Feb 01 22:44:21 but I am going now Feb 01 22:44:30 if it was up to me, I'd make a switch, but people are generally inert. so, now I have to wait for them to validate their apps against 4.7, that's it Feb 01 23:07:29 angstrom is somewhat unstable beeing based on a lot of git versions and not actual releases is there a distro containing more stable packages in oe ? Feb 01 23:12:42 unsolo, git versions are generally chosen because they are known to be more stable or more useful that release versions Feb 01 23:13:06 and there are many programs that dont have recent releases Feb 01 23:13:27 or don't have releases at all Feb 01 23:13:57 unsolo, is there a particular package/recipe taht you are having problems with? Feb 01 23:27:14 grg: i've created a patch to fix an issue with opkg and depmod. I've been told you're the right person to ask you to have a look at it: http://paste.ubuntu.com/561174/ Feb 01 23:27:46 the problem was, that opkg intercepts all calls to "depmod -a *" and replaces them with a single "depmod -A" Feb 01 23:28:07 from pkg_postinst etc. Feb 01 23:28:56 in oe, the depmod command line usually is depmod -a ${KERNEL_VERSION} Feb 01 23:29:32 which updates modules.dep only for the given kernel version Feb 01 23:30:43 without the kernel version argument, depmod updates modules.dep for the running kernel Feb 01 23:31:05 so stripping off $KERNEL_VERSION is wrong Feb 01 23:31:57 the other (imho worse) problem is, that -A doesn't do anything if the installed module's timestamp is older than modules.dep's timestamp Feb 01 23:32:31 grg: For reasons i cannot go into i would like to be able to build a distro based on specific releases (and patched up) .. on a omap3530 using kernel 2.6.34 Feb 01 23:33:04 which usually is the case, because the modules timestamp equals the build time, not the install time Feb 01 23:35:26 grg: ofc that means making a bb for my machine for that specific kernel .. and well thats sort of my issue.. but regarding the rest of the packages.. arent they mostly released Feb 01 23:36:47 * unsolo hopes he is making some kind of sense Feb 01 23:40:43 obi, thanks. I'll take a look. Feb 01 23:40:53 * grg is confused by unsolo Feb 01 23:41:59 grg let me rephrase i want a oe distro based on mostl "stable / released" packages although patched up.. that is quite up to date (not including kernel) Feb 01 23:42:40 i suppose you have to choose one of the available distros which you consider to be close to that. Or modify the closest to create your own distro Feb 01 23:43:01 yea so my question was rather which one would you guys consider to be the closest to that Feb 01 23:43:38 although i could build all of them ;) Feb 01 23:44:47 gumstics perhaps Feb 01 23:45:21 no.. Feb 01 23:46:50 well guess maybe modifying an angstrom is the way to go Feb 01 23:48:21 * unsolo needs sleep.. later Feb 02 00:09:10 03Andrea Adami  07org.openembedded.dev * re7c683253d 10openembedded.git/recipes/linux/linux-kexecboot/ (6 files in 6 dirs): Feb 02 00:09:10 linux-kexecboot: refresh Zaurus defconfigs for new SRCREV (2.6.38-rc3) Feb 02 00:09:10 * try to unbreak collie mtd Feb 02 00:09:10 Signed-off-by: Andrea Adami Feb 02 00:09:20 03Andrea Adami  07org.openembedded.dev * rb2a7257d4c 10openembedded.git/recipes/linux/linux-kexecboot_git.bb: Feb 02 00:09:20 linux-kexecboot: move SRCREV of git recipe to 2.6.38-rc3 Feb 02 00:09:20 Signed-off-by: Andrea Adami Feb 02 00:21:37 obi, can you point out a recipe which triggers the problem(s) you are fixing? (just so i can trace through what is happening) Feb 02 00:22:44 install a kernel module, modprobe that module, install a second module, try to modprobe that module Feb 02 00:22:50 the second modprobe will fail Feb 02 00:23:13 because modules.dep won't get updated once its timestamp is newer than the timestamp of the modules Feb 02 00:24:20 any kernel-module-* will do Feb 02 00:25:38 i'm using jffs2, but i suppose that the timestamp of a file installed by opkg will be the build time on any fs Feb 02 00:25:55 kernel.bbclass:pkg_postinst_modules() is what i was looking for :) Feb 02 00:26:19 ah, ok :) Feb 02 00:27:53 Is there a way to install the kernel-headers into a build instead of a stagging package? kernel-dev installs only the map , symbols, and config. Feb 02 00:28:56 grg: btw., the symptom was very odd. putting /sbin/depmod into pkg_postinst would fix the problem, too... Feb 02 00:29:55 it's quite puzzling if you don't know that opkg has this interception mechanism Feb 02 00:30:19 there is little about opkg that is not puzzling Feb 02 00:30:36 :) Feb 02 00:31:17 as i understand it, the intercept mechanism is there so that e.g. depmod is not called for each and every postinst script during an opkg install. So if you installed 16 modules, depmod would only be called once. Feb 02 00:31:50 with your patch, that is no longer the case Feb 02 00:31:59 so it needs to be tweaked a bit i think Feb 02 00:32:21 * grg is trying to remember how this all fits together Feb 02 00:32:58 no, i had that in mind and it should still work that way Feb 02 00:33:06 intercept/depmod creates a shell script Feb 02 00:33:14 ah. sort -u :) Feb 02 00:33:16 before it would put depmod -A into it Feb 02 00:33:18 my mistake Feb 02 00:33:22 yes :) Feb 02 00:33:59 the indirection is horribly complex Feb 02 00:35:08 it's unlikely that modules for different kernel versions get installed at the same time, but it's certainly possible Feb 02 00:35:31 that's why i used sort instead of just replacing -A with $@ Feb 02 00:37:39 did you check that the other intercept scripts don't have similar problems? Feb 02 00:38:56 i don't know whether ldconfig and update-modules are ever called with parameters Feb 02 00:41:29 i don't see any in my current /usr/lib/opkg/info Feb 02 00:41:44 interestingly, it looks like they all call ldconfig with /sbin/ldconfig Feb 02 00:41:57 so the intercept wont be called for any of those Feb 02 00:56:47 obi, i've committed your fix to the opkg repository. Feel free to bump the opkg SRCREV in opkg.inc. Feb 02 00:57:13 thanks! Feb 02 02:34:41 Tartarus, are you here? **** ENDING LOGGING AT Wed Feb 02 02:59:58 2011