**** BEGIN LOGGING AT Mon May 16 02:59:57 2011 May 16 06:54:29 good morning May 16 07:03:05 hi mckoan, all May 16 07:43:29 Hi May 16 07:43:37 I am getting an issue while making oe images May 16 07:44:00 simple utilities like ping, treaceroot also require root password May 16 07:44:05 how can i control this May 16 07:44:06 ?? May 16 08:09:18 03Stefan Schmidt  07master * r73772d69ba 10openembedded.git/recipes/llvm/ (llvm2.7/include-fixes.patch llvm2.7_2.7.bb): (log message trimmed) May 16 08:09:18 llvm2.7_2.7: Add patch for build errors on newer gcc / glibc combinations May 16 08:09:18 Its seems there is a stricter header checking again which results in errors like May 16 08:09:23 task-bug-feed-extra: Add gdbserver May 16 08:09:23 Signed-off-by: Stefan Schmidt May 16 08:09:24 03Andrew Turley  07master * r3de75ac543 10openembedded.git/recipes/buglabs-apps/bug-app.inc: May 16 08:09:30 com.buglabs.app.bugdash2: Bump to newer version May 16 08:09:30 Signed-off-by: Stefan Schmidt May 16 08:09:30 03John E. Connolly  07master * r115a388621 10openembedded.git/recipes/alsa/alsa-state/bug20/asound.state: May 16 08:09:37 bug-osgi.inc: Allow javac command to be specified by recipe. May 16 08:09:37 Signed-off-by: Stefan Schmidt May 16 08:09:37 03Stefan Schmidt  07master * r8b52ada6f9 10openembedded.git/recipes/linux/linux-omap-hah_2.6.31.bb: May 16 08:10:50 03Stefan Schmidt  07master * r24bb53f14c 10openembedded.git/recipes/tasks/task-bug.bb: May 16 08:10:51 task-bug: Add connman-plugin-ethernet back into task-bug. May 16 08:10:51 Signed-off-by: Stefan Schmidt May 16 09:15:13 03Stefan Schmidt  07master * rd0f1dbb8fe 10openembedded.git/recipes/llvm/ (llvm2.7/include-fixes.patch llvm2.7_2.7.bb): May 16 09:15:13 llvm2.7_2.7.bb: Update patch to use c++ include style. May 16 09:15:13 As henning suggested on we should use a c++ style include here instead. Thus May 16 09:20:21 hi all May 16 09:21:39 hi pb_ May 16 09:22:43 hello pb_, florian May 16 09:27:51 hm, this is a bit weird, guilt-native doesn't build for me in a fresh oe-core checkout May 16 09:28:02 and, looking at the code, I don't really understand how it can have been working for anybody else either. funky. May 16 09:32:19 oh, no, I see, it's just another one of these hardcoded /usrs May 16 09:36:15 heh, still poking around micro? May 16 09:41:57 yup May 16 10:30:01 03Eric Bénard  07master * r50c64f5170 10openembedded.git/conf/distro/include/angstrom-2010-preferred-versions.inc: May 16 10:30:01 angstrom-2010: prefer qt 4.7.3 May 16 10:30:01 Signed-off-by: Eric Bénard May 16 10:36:35 morning mickey May 16 10:44:43 hi woglinde May 16 10:44:49 hi pb May 16 11:16:51 stupid netqwork manager May 16 11:18:02 hehe May 16 11:18:08 thats why I avoid it May 16 11:18:50 and I still how it works if you dont log in May 16 11:18:56 +wonder May 16 11:42:15 crofton ping May 16 11:43:14 pong May 16 11:43:33 crofton your vote proposal is only to ack the list? May 16 11:43:55 right now I just want to klnow if people mind if we vote yes/no on the entire list May 16 11:44:06 then I will submit the actual proposal May 16 11:45:15 uhm May 16 11:45:16 is that clear? May 16 11:45:17 okay May 16 11:45:36 os its only about block vote or not May 16 11:45:56 there are more people than usual on the list, and just want to make sure people are ok with voting on the list, as opposed to voting for each individual May 16 11:45:59 rigth May 16 11:46:07 just a question May 16 11:46:45 send May 16 11:46:49 hi risca May 16 11:51:53 http://bicycles.stackexchange.com/ May 16 11:56:21 crofton aha May 16 12:03:45 Hi :-) May 16 12:08:32 hi woglinde May 16 12:08:42 Hi all May 16 12:28:24 03Koen Kooi  07org.openembedded.dev * rb4166d5421 10openembedded.git/conf/distro/include/angstrom-2010-preferred-versions.inc: May 16 12:28:24 angstrom: prefer udev 168 May 16 12:28:24 Signed-off-by: Koen Kooi May 16 12:36:20 bitbake nano returns " TypeError: warn() takes at least 2 arguments (1 given) " May 16 12:38:06 which bitbake version May 16 12:38:10 which oe? May 16 12:41:28 oka have to go May 16 12:43:57 1.8.18 May 16 12:44:02 sorry for the delay May 16 12:55:49 hi all May 16 12:57:19 is it possible to write a recipe for a svn package that doesn't use automake (so it doesn't have a configure) ? May 16 12:58:14 I have some projects in a local svn server that I need to include in my own layer but none of them have a configure...just bare Makefiles May 16 12:58:55 mrAlmond: there's no problem with that May 16 12:59:23 by default (from base.bbclass), do_configure is empty May 16 13:00:06 mrAlmond: just don't inherit autotools (since you're not using autotools) May 16 13:00:34 foerster: ok...but I suppose I've to change the Makefiles May 16 13:01:08 that at the moment require the path to the cross compiler bin directory May 16 13:02:00 must I simply use $CC or $STRIP (for using strip) ? May 16 13:02:24 in order to make yocto use the right cross compiler May 16 13:02:55 03Koen Kooi  07org.openembedded.dev * r37af029d0a 10openembedded.git/recipes/udev/ (3 files in 2 dirs): (log message trimmed) May 16 13:02:55 udev: add recipe to build from git (currently 168 + handfull of commits) May 16 13:02:55 This recipe removes all but one of the custom rules OE installs. The local.rules will need some looking at, since it assumes ifupdown support while the image might be use connman/networkmanager/etc instead. May 16 13:04:09 mrAlmond: not certain, I haven't written any non-autotools recipes May 16 13:04:41 mrAlmond: http://docs.openembedded.org/usermanual/html/recipes_examples.html May 16 13:05:11 you can see how they're using ${CC}, etc May 16 13:05:41 thank you! May 16 13:06:32 np May 16 13:08:07 I've found that also the zip recipe in recipes-extended/zip is for a non-autotool project May 16 13:08:14 so it seems to be a good start point May 16 13:17:20 hello.... I need to change the IP addres on a given network interface in C.... someone can help me? May 16 13:22:25 Is there a way to reverse a SRC_URI_append? I'm trying to avoid applying a certain set of patches in a recipe. May 16 13:22:37 Such as SRC_URI_remove? May 16 13:24:53 mkneissl: edit the recipe (or copy it and roll your own) May 16 13:25:40 eFfeM_work: OK, thanks. May 16 13:25:50 yw May 16 13:26:49 mkneissl: you could write a python function to filter out the bits you want, but there is no direct inverse to _append. May 16 13:29:03 pb_: of course you can do that, but if it is for private use only it is probably better to edit or edit a copy, and if it is for generic use probably the recipe should be reworked to avoid applying the patch if not needed May 16 13:30:57 pb_: I got your reply to the libpcsclite recipe May 16 13:31:14 pb_: how should I do it using sbin and like? May 16 13:33:06 one way would be to embed the script into a shell function within the .bb file as a here document. May 16 13:34:05 a better solution for the long term would be to find a way for recipes to simply provide the parameters and have the actual script be generated by some backend mechanism. May 16 13:34:26 pb_: yes but: May 16 13:34:32 pb_: here document is ugly May 16 13:34:40 pb_: and there's no backend for it atm May 16 13:34:42 pb_: so? heh May 16 13:34:59 the latter would avoid the proliferation of tiny scripts which are 90% identical and remove the potential for unnoticed errors. May 16 13:35:11 otavio: well, you could write the backend. May 16 13:35:21 and, as for ugly, given a choice between ugly and broken I think we would be better off with ugly. May 16 13:43:23 pb_: sent May 16 13:43:27 pb_: please check it May 16 13:43:54 RP__: do you mind to apply the cmake fix? or does it still has problems? May 16 13:46:22 hello.... I need to change the IP addres on a given network interface in C.... someone can help me? May 16 13:53:44 pb_: sent v3 now May 16 14:00:12 otavio: fwiw: I'm also in favor of calling it pcsc-lite which has ie ${PN}-lib in PACKAGES and debian.bbclass makes libpcsclite1 from it May 16 14:00:50 JaMa: really? it looks so ugly :-( May 16 14:01:11 JaMa: I won't object this change but I am not in favor of it May 16 14:01:45 otavio: I'm not using pcsc from OE personally but I wouldn't guess that recipe named libpcsclite has pcscd package in it May 16 14:02:11 JaMa: yes, you'd since you have it inside pcsc-lite dir May 16 14:02:23 JaMa: and also a package called pcscd too May 16 14:02:24 otavio: but if it's vice-versa I don't need to know that pcsc-lite has separate package with lib, because shlibs will RDEPEND on right name for me May 16 14:02:59 (devel)...dded/meta-oe/recipes-support/pcsc-lite% apt-cache search pcsc-lite May 16 14:03:00 libpcsc-perl - Perl interface to the PC/SC smart card library May 16 14:03:02 only this May 16 14:03:19 (devel)...dded/meta-oe/recipes-support/pcsc-lite% apt-cache show pcscd | grep Version May 16 14:03:22 Version: 1.5.5-4 May 16 14:03:26 same. May 16 14:03:35 I wouldn't look up for pcsc-lite but pcsc May 16 14:03:40 eix pcsc http://paste.pocoo.org/show/389909/ May 16 14:05:14 JaMa: heh, what eix is? May 16 14:06:04 utility to search gentoo metadata May 16 14:06:09 ah May 16 14:06:10 cool May 16 14:06:27 real 0m0.167s May 16 14:06:28 anyway there you're looking for source packages not binaries May 16 14:06:48 an in our case I am worried about the binaries May 16 14:07:05 but as I said I won't object to rename it May 16 14:08:38 and I won't NAK it if it's called libpcsclite.. just saying that pcsc-lite is IMHO better :) May 16 14:09:07 as debian.bbclass is handling renaming binaries including version May 16 14:10:19 * JaMa doesn't also like glib-2.0 as package name May 16 14:13:22 I'm attempting to do a 'git submodule update --init --recursive' on a repository to perform the checkout. any idea how to add that to a recipe? May 16 14:14:25 I was advised I could add it as a process step after do_fetch and before do_configure, but the downloads folder doesn't have a checkout (so I wouldn't be able to do it there even if I knew how to change directory to it) and the work folder already has a copy, but that copy doesn't have any git metadata. May 16 14:18:31 JaMa: sent an updated recipe with it renamed May 16 14:25:11 How can I handle svn repositories that need authentication in a recipe ? May 16 14:26:53 mrAlmond: I'd use it into .subversion of user May 16 14:28:19 otavio : I will try that...thank you. May 16 14:28:41 Is someone using linux kernel (instead of linux-yocto) on oe-core? May 16 15:05:10 hi, what would May 16 15:05:13 err May 16 15:06:02 what would I need to do to have OE produce a 32-bit SDK (for the host) intended to target 32-bit, but do this from a 64-bit build environment May 16 15:09:05 hmm, may either need to do something canadian cross, or run the whole thing under linux32 and add -m32 to the native/build cflags May 16 15:09:26 http://www.linux.com/news/featured-blogs/171-jonathan-corbet/445344-whats-up-with-arm May 16 15:10:02 hbeck: I had a similar issue and setup a chroot May 16 15:10:06 kergoth_: we been doing it from chroot type environment but would prefer to not have to mess with that. May 16 15:10:19 martinmeba: right, was hoping there was some handy config option to avoid chroot May 16 15:10:32 i'd try the linux32/-m32 route, see how it goes May 16 15:10:33 yeah - I wish that there was a simpler way as well... May 16 15:13:53 morning kergoth May 16 15:14:44 hey pb_ May 16 15:16:12 sigh, this stupid freescale bootloader seems designed for maximum infuriation potential May 16 15:16:50 freescale redboot? May 16 15:17:16 it relocates the application vector table, which is fine... but the relocated one is in a completely different format to the original. not only does it have 3 bytes per vector rather than 2, the ordering is reversed in memory so the reset vector comes first. May 16 15:17:32 florian: no, some crazed 9s08 bootloader May 16 15:17:58 uh May 16 15:18:15 I was hoping not to need a bootloader on this system: we were planning to just program the chips over the debug interface. but, freescale have somehow managed to create a debug interface that doesn't work if the chip isn't already programmed. May 16 15:18:52 (as well as requiring the host to bit-bang a serial protocol at 460000 baud) May 16 15:18:52 argh May 16 15:19:26 so, it turned out that using a serial bootloader was in fact the easiest approach. May 16 15:19:41 oh, and, of course: all the tools are windows only so I have to run them in virtualbox May 16 15:20:01 but enough of that, I am starting to get depressed again just talking about it May 16 15:20:19 otavio: thanks :) May 16 15:20:51 pb_: think positive... it could be worse: there are tools out there that fail to work with virtualbox May 16 15:20:57 heh, true May 16 15:21:54 * florian wonders why there are still companies out there stopping communication once you mention linux kernel sources May 16 15:22:35 hbeck: FWIW, openembedded-core can do what you describe (set SDKMACHINE accordingly) May 16 15:25:44 RP__: huh. Where could I peek at that? May 16 15:26:48 hm, drat, my oe-core build with uclibc has crashed and burned again May 16 15:27:14 I guess most of the oe-core folks must be using eglibc. May 16 15:27:35 ugh, I'm starting to suspect we have some lurking issues with reuse of sstates in oe-core. Multiple times now I've hit build failures that go away if I wipe my sstate-cache and rebuild :( May 16 15:28:08 someone has recipes for regular linux kernel that i can use? May 16 15:28:13 for oe-core I mean May 16 15:28:25 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r8500885ce1 10openembedded.git/ (2 files in 2 dirs): (log message trimmed) May 16 15:28:25 linux-nokia900-meego: remove oss and old alsa api May 16 15:28:25 The removal of the oss api is needed: May 16 15:28:45 do the ones from regular oe not work? May 16 15:34:49 otavio, not yet :) May 16 15:39:30 pb_: probably not; I am looking for someone using them so I can copy it May 16 15:39:31 hehe May 16 15:42:20 I dnot see why the regualr ones wouldn't work May 16 15:44:27 I guess you'd have to add the LIC_FILES_CHKSUM stuff at least, and I certainly wouldn't want to bet on the regular ones working out of the box. But I doubt it would be hard to fix them. May 16 15:57:18 hbeck: http://git.openembedded.net/cgit.cgi/openembedded-core/ May 16 15:57:42 kergoth_: ouch :( May 16 15:57:53 kergoth_: Any idea where the problem is roughly? May 16 15:58:16 haven't nailed down the specific circumstances yet :\ May 16 16:06:26 How to refresh the modules in the image.tar after reconfiguring and rebuilding the kernel? May 16 16:18:49 phew, uclibc working now apparently May 16 16:19:19 mkneissl: it should be automatic if you rerun the image target May 16 16:20:13 pb_: Yeah, I just figured out, that there have never been modules in the root fs. How are those supposed to be referenced? May 16 16:20:34 oh, spoke to soon, can't find -lgcc_s May 16 16:20:35 drat May 16 16:21:08 mkneissl: they'd need to be called in by something in ${PACKAGE_INSTALL}. May 16 16:24:39 pb_: I see that there are lots of kernel-module-*.ipk produced. But they don't make it into the image. May 16 16:27:42 yeah. so, like I said, you need to either mention them in ${PACKAGE_INSTALL}, or have them be depended on by something else that is in there. May 16 18:35:23 re May 16 18:40:58 hi florian, getting any sleep? May 16 18:41:58 ka6sox-sick: howdy ? sick ? May 16 18:42:05 ka6sox-sick: not enough... returned from travelling for a week... still need to catch up May 16 18:42:18 ka6sox-sick: but you do not look that good too May 16 18:42:43 khem, yep...trying to kick this.... May 16 18:42:57 florian: never have...this can't improve it! May 16 18:43:29 khem, are you going to kick off a full build on garnet after re-mounting /scratch where you want? May 16 18:43:37 florian, how was linuxtag? May 16 18:44:54 ka6sox-sick: pretty good - much better then last year. more oe people, better booth equipment and more interesting people there. May 16 18:45:25 florian, excellent... May 16 18:45:41 I hope next year @ SCaLE khem will have more help too... May 16 18:48:23 pb_, ping? May 16 19:02:45 hi May 16 19:03:21 it seems that a boss at intel told peoples : go and become a member of openembedded :-D May 16 19:03:26 ka6sox-sick: yes I will May 16 19:03:36 ka6sox-sick: right now there was a little breakage May 16 19:03:37 khem, thanks May 16 19:03:46 khem, understood. May 16 19:03:51 ka6sox-sick: btw. melo needs reboot May 16 19:03:58 I'd like to see that it still builds in < 1 day May 16 19:04:10 khem, whats wrong with melo? May 16 19:04:13 so u added some memory May 16 19:04:27 ka6sox-sick: security updates to lnx kernel :( May 16 19:04:32 I added memory and scratch May 16 19:04:35 nothing is wrong per say May 16 19:04:52 khem, I'll do that when europe goes to bed...about 0100UTC May 16 19:05:34 about 6pm here. May 16 19:21:07 hi ka6sox May 16 19:22:41 ka6sox-sick write an email in case it dont comes up May 16 19:29:53 woglinde, done May 16 19:38:50 03Eric Bénard  07master * r716bbddcf0 10openembedded.git/recipes/sip/sip-native_4.12.1.bb: May 16 19:38:50 sip-native: update to 4.12.2 May 16 19:38:50 4.12.1 is no more fetchable from the original source May 16 19:51:16 03Joshua Lock  07master * r2bc8f405ec 10bitbake.git/lib/bb/ui/ (crumbs/hobeventhandler.py crumbs/runningbuild.py depexp.py): May 16 19:51:16 bitbake/ui: Fix Gtk+ GUI's after recent cache changes May 16 19:51:16 Signed-off-by: Joshua Lock May 16 20:10:35 03Tom Rini  07master * r1e255fbd29 10openembedded.git/recipes/perl/perl-native_5.10.1.bb: May 16 20:10:35 perl-native: create_wrapper on perl${PV} too May 16 20:10:35 perl${PV} becomes hostperl when building for the target so we need a May 16 20:26:32 03Simon Busch  07master * r798faf0cbf 10openembedded.git/conf/distro/shr.conf: May 16 20:26:32 shr: use fsoaudiod as provider for alsa-state May 16 20:26:32 Signed-off-by: Simon Busch May 16 20:26:35 03Simon Busch  07master * rf4d615ceef 10openembedded.git/recipes/tasks/task-base.bb: (log message trimmed) May 16 20:26:35 task-base: make it possible for a distro to set it's alsa-state provider May 16 20:26:35 There different recipes in OE which provides alsa-state. In my case it's alsa-state and May 16 20:26:58 fsoaudiod: package even the alsa-plugin and bump PR May 16 20:26:58 Signed-off-by: Simon Busch May 16 20:30:53 esbenh: ping May 16 20:31:12 03Ulf Samuelsson  07master * rf0560cb5ef 10openembedded.git/recipes/linux/ (6 files in 3 dirs): (log message trimmed) May 16 20:31:12 Update AT91 linux-2.6.30 support May 16 20:31:12 Support use of exp.4 (the latest) patch for SAM9M10EKES May 16 20:33:51 Add new configs May 16 20:33:51 at91sam9m10g45ek_sd_config May 16 20:33:51 03Ulf Samuelsson  07master * r0b65da01d4 10openembedded.git/recipes/owl-wifi/ (4 files in 2 dirs): May 16 20:34:04 Set defaults to at91bootstrap-3.0 and u-boot-2009.11 May 16 20:34:04 Set the default bootstrap for at91 machines to bootstrap-3.0 May 16 20:34:04 Set the default u-boot to 2009.11 May 16 20:35:45 03Ulf Samuelsson  07master * r63e1602354 10openembedded.git/recipes/images/ (3 files): May 16 20:35:45 Update at91sam9 images May 16 20:35:45 Signed-off-by: Ulf Samuelsson May 16 22:02:44 03Koen Kooi  07org.openembedded.dev * r599426ea8f 10openembedded.git/recipes/systemd/ (4 files in 2 dirs): May 16 22:02:44 systemd: bump SRCREV May 16 22:02:44 The angstrom and svg patches were merged upstream, so remove them May 17 01:02:17 03Tom Rini  07master * rdc57d5696f 10openembedded.git/conf/machine/p2020rdb.conf: May 17 01:02:17 p2020rdb: Switch to linux-powerpc-fsl for kernel May 17 01:02:17 The 'linux' recipe has never been an actual good provider of a kernel **** ENDING LOGGING AT Tue May 17 02:59:58 2011