**** BEGIN LOGGING AT Fri Jan 14 02:59:58 2011 Jan 14 07:20:19 03Steffen Sledz  07org.openembedded.dev * rc74b7712d8 10openembedded.git/recipes/gnome/epiphany_2.30.2.bb: (log message trimmed) Jan 14 07:20:19 epiphany-2.30.2: unbreak build by disabling CA files Jan 14 07:20:19 This is a workaround to unbreak the build of epiphany for various host Jan 14 07:20:19 operating systems, since nobody was able/willing to fix it in the right Jan 14 07:20:19 way. Jan 14 07:20:20 see also: Jan 14 07:20:20 * Jan 14 07:51:07 good morning Jan 14 09:24:48 03Martin Jansa  07master * r8818455b7f 10openembedded.git/recipes/python/ (33 files in 4 dirs): Jan 14 09:24:48 python-native: drop old versions 2.6.1 and 2.6.4, remove unneeded python-native-* directories Jan 14 09:24:48 * python-native-2.7.1 is the same as python-2.7.1 and thanks to FILESPATHPKG .= ':python-:python' it's not needed to keep it twice Jan 14 09:24:48 Acked-by: Khem Raj Jan 14 09:24:48 Signed-off-by: Martin Jansa Jan 14 09:24:58 03Martin Jansa  07master * r2dba8fb680 10openembedded.git/recipes/linux/ (3 files in 3 dirs): Jan 14 09:24:59 linux-2.6.34: drop DRM patches from shr.patch Jan 14 09:24:59 Signed-off-by: Martin Jansa Jan 14 09:42:52 stefan_schmidt, when is the next patch weekend? Jan 14 09:43:52 ka6sox: not determined yet Jan 14 09:45:02 kk, will you post it to the ML? Jan 14 09:46:30 ka6sox: yes Jan 14 09:46:56 stefan_schmidt, thanks. I just want to have things on the Top Line when you do. Jan 14 09:47:42 ka6sox: ok Jan 14 09:47:56 ka6sox: let me know if anything would block it Jan 14 09:48:05 I'm flexibale with this Jan 14 09:48:07 stefan_schmidt, so far I don't think so Jan 14 09:48:28 I'd just ask that it isn't the 25-27th of February Jan 14 09:48:55 ok, will keep this in mind Jan 14 09:50:19 I will be @ www.socallinuxexpo.org Jan 14 09:50:32 we have a stand there as well. Jan 14 10:55:37 khem: building ... **** BEGIN LOGGING AT Fri Jan 14 11:39:04 2011 Jan 14 11:48:04 03Martin Jansa  07master * r24b76af7bd 10openembedded.git/recipes/freesmartphone/ (3 files): Jan 14 11:48:04 fso-specs,libfso-glib,libfreesmartphone-glib: bump SRCREV Jan 14 11:48:04 Signed-off-by: Martin Jansa Jan 14 12:04:31 gm Jan 14 12:12:03 03Martin Jansa  07master * r5e4ece4a93 10openembedded.git/recipes/freesmartphone/libfreesmartphone-glib_git.bb: Jan 14 12:12:03 libfreesmartphone-glib: fix SRCREV Jan 14 12:12:03 Signed-off-by: Martin Jansa Jan 14 14:43:27 03Tom Rini  07org.openembedded.dev * r6e8dd072b0 10openembedded.git/recipes/e2fsprogs/e2fsprogs.inc: Jan 14 14:43:27 e2fsprogs: Fix chattr update-alternatives location Jan 14 14:43:27 Reported by Jaap de Jong. Jan 14 14:43:27 Signed-off-by: Tom Rini Jan 14 15:11:53 hmm.... very funny Jan 14 15:12:49 while we are discussing whether develop a relationship with the Yocto Projec, I was'nt able to find any reference to OE in Yocto website Jan 14 15:14:18 is compatible with the OpenEmbedded Build System Jan 14 15:25:09 03Koen Kooi  07org.openembedded.dev * rd5e10f2ec8 10openembedded.git/recipes/tasks/angstrom-task-gnome.bb: Jan 14 15:25:09 angstrom-task-gnome: add cron and ntpdate Jan 14 15:25:09 Signed-off-by: Koen Kooi Jan 14 15:38:27 khem: it worked fine here. Jan 14 16:57:40 03Tom Rini  07org.openembedded.dev * r005d673172 10openembedded.git/recipes/linux/linux-jlime-jornada6xx_2.6.25.bb: Jan 14 16:57:40 linux-jlime-jornada6xx 2.6.25: Move to nonworking Jan 14 16:57:40 The git tree in question no longer seems to exist and wasn't trivial Jan 14 16:57:40 to find where it's moved to (if it was). Jan 14 16:57:40 Signed-off-by: Tom Rini Jan 14 16:59:40 Anyone seeing problems w/ gcc* and fetch failing? Jan 14 16:59:48 http://pastebin.com/SEdtsDKw Jan 14 16:59:54 in the middle of an svn checkout going already Jan 14 17:04:09 sgw, time to use a pastebin? :) Jan 14 17:05:13 right! Jan 14 17:15:29 words of wisdom - don't use webkit-gtk_svn at this point, leaks like a sieve on ajax requests. The 1.2.3 stable does not exhibit this behavior. Jan 14 17:21:43 OK, I will be at FOSDEM! Jan 14 17:33:43  Jan 14 17:34:03 oops Jan 14 17:34:04 :) Jan 14 17:52:19 * otavio still fighting with sem_open support Jan 14 17:52:27 no glue why it is failing Jan 14 17:53:54 * kergoth_ has no glue either Jan 14 17:55:11 kergoth_: did you see my reply about the OLDEST_KERNEL setting? Jan 14 17:56:57 no, i was just being a smartass. i think you meant clue, not glue ;) Jan 14 17:58:11 clue :P Jan 14 17:58:13 right Jan 14 18:06:32 is there a limit to the number of child processes a build can spawn? Jan 14 18:12:33 gm Jan 14 18:12:39 03Khem Raj  07master * r1eb54c1b78 10openembedded.git/contrib/qemu/run-qemu.sh: Jan 14 18:12:39 run-qemu.sh: Add new commandline arg to specify image Jan 14 18:12:39 * This was earlier hardcoded in the script so you needed to Jan 14 18:12:39 modify the script everytime you wanted a different image to boot Jan 14 18:12:39 Signed-off-by: Khem Raj Jan 14 18:12:39 hi Jan 14 18:12:44 otavio: thanks I tested it here too Jan 14 18:12:46 03Khem Raj  07master * r42fcc11e94 10openembedded.git/recipes/linux/ (linux-2.6.37/qemuppc/defconfig linux_2.6.37.bb): Jan 14 18:12:46 linux_2.6.37.bb: Add qemuppc to list of supported machines Jan 14 18:12:46 Signed-off-by: Khem Raj Jan 14 18:12:48 and its ok Jan 14 18:13:10 Tartarus: must be some little hitch on gcc svn server Jan 14 18:13:21 blindvt: around ? Jan 14 18:13:44 blindvt: go ahead and apply Carmelo's protected symbol patches and spin out rc2 Jan 14 18:14:09 err well ack them and he will push Jan 14 18:14:16 he is waiting on you Jan 14 18:16:25 khem: did you see my mail about sem_open? Jan 14 18:16:48 khem: OLDEST_KERNEL didn't help, I also checked if sem_open was "available" on libpthread and it seems to be Jan 14 18:16:59 hmmm ok Jan 14 18:17:15 otavio: Does it happen only on arm builds Jan 14 18:17:18 khem: so I have no clue (see kergoth_ .. learned it :P) about it. Do you? Jan 14 18:17:20 or also on i586 ? Jan 14 18:17:26 khem: I am using i586 only Jan 14 18:17:33 khem: didn't test arm Jan 14 18:17:46 otavio: hmm someone else was reporting it on arm Jan 14 18:17:47 too Jan 14 18:17:52 do u have /dev/shm Jan 14 18:17:57 on target Jan 14 18:18:03 khem: I guess so; checking it Jan 14 18:19:54 khem: it seems the mode was wrong Jan 14 18:19:59 khem: doing a new test Jan 14 18:25:29 khem: well but it wouldn't matter since the app was running as root Jan 14 18:25:38 khem: so the mode wasn't important Jan 14 18:25:48 khem: except if glibc checks the mode, which I doubt Jan 14 18:29:40 otavio: Do u have a small testcase ? Jan 14 18:29:46 or can you generate one ? Jan 14 18:29:52 I have try to debug it here Jan 14 18:30:10 I think the problem is probably misconfiguration on OE part Jan 14 18:30:16 particularly eglibc Jan 14 18:30:24 but am not sure unless I see thru it Jan 14 18:30:29 khem: I guess so. Jan 14 18:30:29 ok time to drive Jan 14 18:30:37 khem: freerdp is a good test case Jan 14 18:30:42 otavio: ok write a small testcase Jan 14 18:30:52 khem: it fails to start if it lacks sem_open Jan 14 18:30:59 should be simple enough to just make a call to sem_open ? Jan 14 18:31:08 in a main() Jan 14 18:31:18 khem: yes and check if it "works or not" Jan 14 18:31:24 khem: just that Jan 14 18:32:06 alright Jan 14 18:32:06 hoi khem Jan 14 18:32:10 hi otavio Jan 14 18:32:13 hey woglinde Jan 14 18:32:18 need to drive now Jan 14 18:32:19 ttyl Jan 14 18:32:25 hm seems I will make lightsoark recipe the next time Jan 14 18:32:30 khem: in my app it checks it as: http://gitorious.org/+ossystems-hackers/freerdp/ossystems-hackers-tree/blobs/master/X11/xfreerdp.c#line674 Jan 14 18:32:32 ups lightspark Jan 14 18:32:43 xfreerdp? Jan 14 18:32:47 woglinde: yep Jan 14 18:32:56 whats that good fro? Jan 14 18:33:03 we already have nx Jan 14 18:33:15 woglinde: rdesktop alternative and with a better team and codebase Jan 14 18:33:28 woglinde: so rdp client that is going to rock soon Jan 14 18:34:03 hm Jan 14 18:34:23 oh nomachine has nx4 preview out Jan 14 18:41:05 03Otavio Salvador  07master * r3e299ddd4f 10openembedded.git/recipes/hal/ (5 files): Jan 14 18:41:05 hal: use INC_PR Jan 14 18:41:05 Signed-off-by: Otavio Salvador Jan 14 18:41:09 03Otavio Salvador  07master * r819e1c13dc 10openembedded.git/conf/machine/i586-generic.conf: Jan 14 18:41:09 i586-generic: define udevdir and OLDEST_KERNEL Jan 14 18:41:09 Signed-off-by: Otavio Salvador Jan 14 18:41:10 03Otavio Salvador  07master * reb888ab886 10openembedded.git/recipes/dbus/dbus.inc: Jan 14 18:41:10 dbus: fix init script if $prefix is empty Jan 14 18:41:10 Signed-off-by: Otavio Salvador Jan 14 18:41:12 03Otavio Salvador  07master * r097294f756 10openembedded.git/recipes/hal/hal.inc: Jan 14 18:41:12 hal: fix init script if $prefix is empty Jan 14 18:41:12 Signed-off-by: Otavio Salvador Jan 14 18:43:04 That reminds me Jan 14 18:43:15 Do we have stuff using hal that we can't migrate to devicekit? Jan 14 18:43:51 Tartarus: no idea Jan 14 18:43:59 Tartarus: I never migrated an app to devicekit Jan 14 18:44:03 Tartarus: is it hard? Jan 14 18:45:32 not sure either, honestly Jan 14 18:45:44 It's hopefully just a matter of using newer versions when needed Jan 14 18:46:31 Tartarus: I mean porting in regarding to code Jan 14 18:46:44 yeah Jan 14 18:46:47 Never done it Jan 14 18:46:58 Tartarus: I understand that most FS project has already done it, but some internal code needs porting Jan 14 18:47:08 Tartarus: I see Jan 14 19:18:16 otavio: I wrote a small example and linked it with -pthread Jan 14 19:18:22 otavio: worked ok on arm Jan 14 19:18:29 khem: grr Jan 14 19:18:32 are you linking with pthread or librt Jan 14 19:18:39 khem: -pthread Jan 14 19:18:49 ok Jan 14 19:19:10 otavio: http://pastebin.com/vdf07bdP Jan 14 19:19:18 try to compile this and run it on your target Jan 14 19:19:26 use option -pthread Jan 14 19:19:29 to compiler Jan 14 19:20:20 khem: doing Jan 14 19:24:09 otavio: what was the error message you got from your real program Jan 14 19:24:23 I am trying to associate that to errno Jan 14 19:25:37 Not implemented Jan 14 19:25:47 hmmm Jan 14 19:32:01 khem, otavio: can i have your ack on this patch or is there anything left to change before pushing it? http://patches.openembedded.org/patch/280/ Jan 14 19:35:09 obi: you have my ack for that one Jan 14 19:35:12 thanks for redoing it Jan 14 19:35:36 otavio: can you print the name of the semaphore in the failing case please ? Jan 14 19:36:16 thank you Jan 14 19:36:53 otavio: nm it will be xfreerdp_**** Jan 14 19:39:52 otavio: it seems its failing because it can not find mount point of shmfs Jan 14 19:40:51 otavio: this is what I have tmpfs on /dev/shm type tmpfs (rw,mode=777) Jan 14 19:41:05 on the system where it is working Jan 14 19:41:39 khem: but does it need to be 777? Jan 14 19:41:42 khem: testing it now Jan 14 19:41:56 03Eric BENARD  07org.openembedded.dev * r18fd668a4d 10openembedded.git/recipes/strace/strace_4.5.20.bb: Jan 14 19:41:56 strace 4.5.20: stable since 04/14/2010 so remove DP=-1 Jan 14 19:41:56 Signed-off-by: Eric Bénard Jan 14 19:41:56 Acked-by: Tom Rini Jan 14 19:41:56 Acked-by: Khem Raj Jan 14 19:42:13 03Eric BENARD  07org.openembedded.dev * r09e47006b8 10openembedded.git/recipes/mtd/ (10 files in 2 dirs): (log message trimmed) Jan 14 19:42:13 mtd-utils: add 1.4.1 Jan 14 19:42:13 * latest stable version + a fix to mkubifs. Jan 14 19:42:13 * switch to lzo2 Jan 14 19:42:14 * tested to generate ubi rootfs on the host. Jan 14 19:42:14 Signed-off-by: Eric Bénard Jan 14 19:42:15 Acked-by: Tom Rini Jan 14 19:42:38 otavio: rw,nosuid,nodev should be enough infact Jan 14 19:45:28 haha today I stumbled over the one second boot qt app on sh4 again Jan 14 19:48:28 woglinde: are u using sh4 Jan 14 19:48:49 no Jan 14 19:49:11 http://www.embedded-bits.co.uk/2011/1-second-linux-boot-to-qt/ Jan 14 19:50:11 03Daniele Ricci  07master * r5b560a0eee 10openembedded.git/recipes/mokosuite2/mokopanel_git.bb: Jan 14 19:50:11 mokopanel: SRCREV bump Jan 14 19:50:11 Added GPS status icon in panel Jan 14 19:50:11 Signed-off-by: Daniele Ricci Jan 14 19:50:11 Signed-off-by: Martin Jansa Jan 14 19:50:14 woglinde: ah cool hmmm SH7724 which is sh4a Jan 14 19:50:48 it runs at half a gig Jan 14 19:51:01 I wonder why cortex socs could not boot in 1 sec Jan 14 19:54:46 khem: I saw a beagle article a while back with high 2s for a beagle Jan 14 19:55:06 the sh is using NOR instead of NAND so less bootloader overhead maybe Jan 14 19:56:00 martinmeba, khem guys you should look at the slides Jan 14 19:56:18 woglinde: I was looking at them this morning Jan 14 19:56:24 woglinde: http://www.youtube.com/watch?v=wUWBkIquQaI this one is not bad also Jan 14 19:56:32 I don't want to remove the crc check on the kernel either Jan 14 19:57:03 ericben yes the linked it from the arcticle Jan 14 19:57:33 definitely food for thought though Jan 14 20:01:37 khem: same error Jan 14 20:01:43 khem: function not implemented Jan 14 20:01:53 khem: /dev/shm is mounted and with 1777 Jan 14 20:08:59 jo mickeyl Jan 14 20:09:19 hi woglinde Jan 14 20:35:03 hi all Jan 14 21:02:27 khem, testing now but uclibc+ppc needs to be -O2 not -Os, too I'm pretty sure Jan 14 21:02:42 (bison is an example of something that fails for the Os vs O2 reason) Jan 14 21:25:32 hi hrw Jan 14 21:25:42 morgen woglinde Jan 14 21:25:50 or rather guten aben Jan 14 21:28:07 hi ant Jan 14 21:28:12 hey woglinde Jan 14 21:32:20 03Klaus Kurzmann  07org.openembedded.dev * r09ad80f201 10openembedded.git/recipes/freesmartphone/fsodeviced_git.bb: Jan 14 21:32:20 fsodeviced_git.bb: remove do_configure_prepend with the now unneeded i2c header magic Jan 14 21:32:20 Signed-off-by: Klaus Kurzmann Jan 14 21:33:57 03Michael 'Mickey' Lauer  07org.openembedded.dev * r62c30a60c8 10openembedded.git/recipes/dbus/ (dbus-cxx/cppcompliance.patch dbus-cxx_0.7.0.bb): dbus-cxx: new recipe; c++ bindings for dbus Jan 14 21:34:01 03Michael 'Mickey' Lauer  07org.openembedded.dev * r1be4ec8b0e 10openembedded.git/recipes/gnome/mm-common_0.9.1.bb: mm-common: add BBCLASSEXTEND Jan 14 21:34:02 03Michael 'Mickey' Lauer  07org.openembedded.dev * r0e50dd7fb9 10openembedded.git/recipes/libsigc++-2.0/libsigc++-2.0_2.2.8.bb: libsigc++-2.0: add new version 2.2.8; this one features BBCLASSEXTENDS Jan 14 21:34:13 03Michael 'Mickey' Lauer  07org.openembedded.dev * ref4fa654cd 10openembedded.git/recipes/dbus/dbus-cxx_0.7.0.bb: dbus-cxx: fix typo in BBCLASSEXTEND Jan 14 21:34:15 03Michael 'Mickey' Lauer  07org.openembedded.dev * re3f3e11d65 10openembedded.git/recipes/commoncpp/ (commoncpp2/cppcompliance.patch commoncpp2_1.8.1.bb): commoncpp: add version 1.8.1 Jan 14 21:45:43 03Thomas Zimmermann  07org.openembedded.dev * refa2cae6a3 10openembedded.git/recipes/libzrtpcpp/ (3 files in 2 dirs): Jan 14 21:45:43 libzrtpcpp: update to 1.5.3 as 1.4.6 doesn't compile anymore Jan 14 21:45:43 Signed-off-by: Thomas Zimmermann Jan 14 21:47:14 03Thomas Zimmermann  07org.openembedded.dev * r7f34f127e5 10openembedded.git/recipes/ccrtp/ccrtp_1.7.2.bb: Jan 14 21:47:14 ccrtp: update to version 1.7.2 Jan 14 21:47:14 Signed-off-by: Thomas Zimmermann Jan 14 21:50:01 Heinervdm: thanks Jan 14 22:17:12 03Thomas Zimmermann  07org.openembedded.dev * rd7fb492bd6 10openembedded.git/recipes/libzrtpcpp/ (3 files in 2 dirs): Jan 14 22:17:12 libzrtpcpp: update to 1.6.0 as sources of 1.5.3 were removed today Jan 14 22:17:12 Signed-off-by: Thomas Zimmermann Jan 14 22:17:17 03Thomas Zimmermann  07org.openembedded.dev * r5a2fcbb9ad 10openembedded.git/recipes/ccrtp/ (ccrtp_1.7.2.bb ccrtp_1.7.0.bb): Jan 14 22:17:17 ccrtp: update to 1.8.0 as sources of 1.7.2 were removed today Jan 14 22:17:17 Signed-off-by: Thomas Zimmermann Jan 14 22:19:25 otavio: strange is it mounted 1777 or 777 Jan 14 22:19:44 Tartarus: uclibc+ppc I am close to find the problem Jan 14 22:20:12 Tartarus: the problem seems to be in busybox or init.sysvinit Jan 14 22:20:37 the only issue with -Os and -O2 was how it was bringing in libgcc.a Jan 14 22:20:48 but that must be solved (I hope) Jan 14 22:21:19 03Thomas Zimmermann  07org.openembedded.dev * rea9751831e 10openembedded.git/recipes/dbus/ (3 files in 2 dirs): Jan 14 22:21:20 dbus-c++: update to latest SRCREV Jan 14 22:21:20 * convert to BBCLASSEXTEND and remove -native as it has Jan 14 22:21:20 staged the wrapper script instead of the binary Jan 14 22:21:20 Signed-off-by: Thomas Zimmermann Jan 14 22:25:38 khem: 1777 Jan 14 22:29:55 khem, i'm referring to stuff that doesn't build Jan 14 22:30:04 and it's not today Jan 14 22:35:42 khem: I am building it using uclibc and will give it a try Jan 14 22:35:50 khem: so we will know if it is a eglibc issue or not Jan 14 22:43:46 khem, def fixed a few items Jan 14 22:45:30 heh, bison has another issue Jan 14 22:45:35 *g* Jan 14 22:45:38 probably a glibc vs uclibc thing, will dig next Jan 14 22:50:24 oh right, heh Jan 14 22:56:09 03Thomas Zimmermann  07org.openembedded.dev * rc6daf0c9c3 10openembedded.git/recipes/dbus/ (dbus-c++/fix-Makefile.patch dbus-c++_git.bb): Jan 14 22:56:09 dbus-c++: add patch to fix the Makefile Jan 14 22:56:09 * dbus-c++ stages the header files in dbus-c++-1/dbus-c++ but Jan 14 22:56:09 includes in the autogenerated files just dbus-c++ Jan 14 22:56:09 Signed-off-by: Thomas Zimmermann Jan 14 22:57:10 Tartarus: what is not building for ppc/uclibc ? Jan 14 22:57:26 Tartarus: I know that it does not boot :) Jan 14 22:57:31 but thats a different aspect Jan 14 22:58:08 No images Jan 14 22:58:17 perl is pulled in for minimal and that fails Jan 14 22:58:36 tgt needs blacklisting on uclibc until it adds posix_fallocate Jan 14 22:58:44 (do we have an easy way to do that today? heh) Jan 14 23:00:21 woglinde, according to tinderbox you've been the only one to successfully build pyqt in ages - but your building a ver thats not in oe? Jan 14 23:00:34 4.7.7 Jan 14 23:00:40 tharvey I tried an upgrade Jan 14 23:00:48 but it didnt worked out Jan 14 23:00:55 http://tinderbox.openembedded.net/packages/python-pyqt/ Jan 14 23:00:57 and than I was pissed Jan 14 23:01:01 and give it up Jan 14 23:01:01 ah... lol Jan 14 23:01:08 ok, thats very good info Jan 14 23:01:28 ya, I'm thrashing around with it now and not having much fun Jan 14 23:01:34 yes Jan 14 23:01:40 its a shitty mess Jan 14 23:01:43 started looking at pyside as well but their build system is funky as well Jan 14 23:01:44 maybee pyside is better Jan 14 23:02:00 i think it makes more sense Jan 14 23:02:10 at building Jan 14 23:02:16 on debian pyqt works fine Jan 14 23:02:19 trying to disect their build system to figure out how to use cross-compile env Jan 14 23:03:04 I've got a beagleboard over here and I notice that angstrom feeds has a pyqt built and packaged as well, but haven't been able to test it yet - need to rebuild fs on a larger flash Jan 14 23:03:25 if it works was going to investigate it more to see how/when it built Jan 14 23:05:34 khem, I'm getting out of time / over my uclibc head, but none of minimal-image meta-toolchain x11-image console-image native-sdk-image nas-server-image build for uclibc + ppc Jan 14 23:05:44 And doing -O2 not -Os got them a lot closer Jan 14 23:06:00 (and cut the fail list in half at least) Jan 14 23:12:10 anyone familiar with using the cmake.bbclass to build stuff using OE but with an external toolchain (toolchain built by OE if that helps)? Jan 14 23:28:20 or even just with getting OE to use an external toolchain in general? I'm aware of the existence of the external-toolchain-xxx.bb meta recipes but not sure how to go about using them Jan 14 23:32:28 Have you googled? Jan 14 23:32:41 There's at least a few threads on them and the example conf file changes needed Jan 14 23:33:21 yup, found a couple references to external-toolchain-csl but haven't found those as useful. Found this one: http://lists.linuxtogo.org/pipermail/openembedded-devel/2009-November/014699.html Jan 14 23:35:19 hbeck, what machine are you building for? Jan 14 23:35:43 And what's your local.conf look like at? Jan 14 23:35:45 atm Jan 14 23:35:53 arago dist is based on OE and has a goal of a very fast build time, which it seems to do a good job of, by using external toolchain Jan 14 23:37:19 our distro and local are set up to use the internal toolchain, but we have a specific set of recipes that we want to build with an external toolchain Jan 14 23:37:38 machine is x86 or 'smart' (arm architecture) Jan 14 23:39:01 Tartarus: On the contrary its building fine for me Jan 14 23:39:10 Tartarus: whats your DISTRO when doing uclibc Jan 14 23:39:20 Tartarus: I hope its DISTRO=minimal-uclibc Jan 14 23:41:37 yeap Jan 14 23:42:00 khem, all of those targets? Jan 14 23:42:11 btw, there are logs this time in tinderbox Jan 14 23:42:14 i gotta go now tho Jan 14 23:43:10 Tartarus: I have not tried nas-server-image Jan 14 23:43:16 other seems file Jan 14 23:43:18 fine Jan 14 23:43:26 Tartarus: show me your local.conf Jan 14 23:43:42 and the error messages too Jan 14 23:53:05 btw, can someone fix: Jan 14 23:53:07 03Thomas Zimmermann  07org.openembedded.dev * r9d8c3d675e 10openembedded.git/recipes/dbus/ (dbus-c++/fix-Makefile.patch dbus-c++_git.bb): Jan 14 23:53:07 Revert "dbus-c++: add patch to fix the Makefile" Jan 14 23:53:07 This reverts commit c6daf0c9c3f5050bd31daf30e7eb2d9eecc8cf50. Jan 14 23:53:07 WARNING: Deprecated usage of 'patch' url param for '/oe/openembedded/recipes/ltp/ltp-20100831/runalltests.patch', please use 'apply={yes,no}' Jan 14 23:55:12 ant__: correct it Jan 14 23:55:55 well..I'll'have to stash other stuff ;) too many terminals on screen ... Jan 14 23:57:02 ok fixedc Jan 15 00:01:07 03Khem Raj  07master * re4b9119190 10openembedded.git/recipes/ltp/ltp_20100831.bb: Jan 15 00:01:08 ltp_20100831.bb: Fix deprecated usage to specify patch in SRC_URI Jan 15 00:01:08 Signed-off-by: Khem Raj Jan 15 00:28:59 03Andrea Adami  07org.openembedded.dev * r023a77e2af 10openembedded.git/recipes/linux/linux-kexecboot.inc: Jan 15 00:28:59 linux-kexecboot: bump PR after kexecboot changes. Jan 15 00:28:59 Signed-off-by: Andrea Adami Jan 15 00:29:01 03Andrea Adami  07org.openembedded.dev * r018c23fd98 10openembedded.git/recipes/kexecboot/ (kexecboot-klibc_git.bb kexecboot.inc kexecboot_git.bb): Jan 15 00:29:01 kexecboot: increase delay before device detection for Zaurus. Jan 15 00:29:01 * Kernels 2.6.3x boot more quickly so we have to compensate the slow Jan 15 00:29:01 * init of CF (or even slower internal HDD on spitz) adding more delay. Jan 15 00:29:02 * Bump PR. Jan 15 00:29:02 Signed-off-by: Andrea Adami Jan 15 00:29:02 03Andrea Adami  07org.openembedded.dev * r9c72d93981 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Jan 15 00:29:03 linux-kexecboot: add c7x0 defconfig for 2.6.37. Jan 15 00:29:03 Signed-off-by: Andrea Adami Jan 15 00:29:04 03Andrea Adami  07org.openembedded.dev * r276ed84e0c 10openembedded.git/recipes/kexecboot/kexecboot-cfg_0.1.bb: Jan 15 00:29:04 kexecboot-cfg: show DISTRO and MACHINE in the label. Jan 15 00:29:05 * Bump PR. Jan 15 00:29:05 Signed-off-by: Andrea Adami Jan 15 00:41:16 03Andrea Adami  07org.openembedded.dev * rf6a804b234 10openembedded.git/recipes/linux/linux-kexecboot-2.6.37/akita/defconfig: Jan 15 00:41:16 linux-kexecboot: add akita defconfig for 2.6.37. Jan 15 00:41:16 Signed-off-by: Andrea Adami Jan 15 00:41:44 Having some problems compiling SHR for the iPhone3G. Anyone mind taking a look at my pastebin? http://pastebin.com/csQFVd81 Jan 15 00:48:37 Lopi: patch devmem2 works ok here Jan 15 00:48:50 Lopi: can you do bitbake -c clean devmem2 Jan 15 00:48:55 and bitbake devmem2 Jan 15 00:54:27 khem: seems to be compiling fine Jan 15 00:54:44 k Jan 15 00:56:43 khem: so when I run into errors like these it's best to clean the package and try again? Jan 15 00:56:58 I dont know why it happened Jan 15 00:57:12 but this should not happen commonly Jan 15 00:57:26 hm okay Jan 15 01:19:45 khem, fyi, these failures happen on u10.04 Jan 15 01:19:47 crazy long path Jan 15 01:19:55 i'll try in a normal length one Jan 15 01:21:21 khem, also, note it's p2020ds (e500v2) not qemuppc Jan 15 01:22:06 hmmm Jan 15 01:22:12 e500 ok Jan 15 01:22:15 donig qemuppc + sane path now Jan 15 01:22:18 I was using qemuppc Jan 15 01:22:21 will check in, in the morning Jan 15 01:22:26 k Jan 15 01:22:27 e300c2 or so also chokes Jan 15 01:22:30 (mpc8315e-rdb) Jan 15 01:22:36 Tartarus: ok Jan 15 01:22:48 I will see if I can build something on mpc8315e-rdb Jan 15 01:22:54 I dont have boards Jan 15 01:22:58 k Jan 15 01:23:02 these are all build time failures Jan 15 01:23:02 but it will be a build test Jan 15 01:23:08 ok Jan 15 01:23:11 i'm not boot testing these either :) Jan 15 01:23:14 (no time) Jan 15 01:23:28 do u have the logs in tinderbox? Jan 15 01:23:31 yes Jan 15 01:23:35 mentor-graphics-autobuilder Jan 15 01:23:40 I boot test qemu on uclibc/eglibc fairly often Jan 15 01:23:48 I want to automate it Jan 15 01:23:51 but no time Jan 15 01:23:57 yeah Jan 15 01:24:12 i had some semi automated testing for real HW, but need to kick one of the lab folks Jan 15 01:24:22 do u have some idea on how I can make sure it booted ? Jan 15 01:24:26 (and also need to make it way smarter about HW reservation, people odn't like a board always checked out) Jan 15 01:24:37 khem, if you can make it even build, I'll find some way to boot test something Jan 15 01:25:06 so u say minimal-image will fail on Jan 15 01:25:06 17:17 < khem> Tartarus: ok Jan 15 01:25:06 17:18 < khem> I will see if I can build something on mpc8315e-rdb Jan 15 01:25:06 17:18 < khem> I dont have boards Jan 15 01:25:11 hrrrmm Jan 15 01:25:14 mpc8315e-rdb Jan 15 01:25:39 ah, no Jan 15 01:25:43 p2020ds will for minimal Jan 15 01:25:48 but 8315 is better, sorry Jan 15 01:26:13 native-sdk, nas-server both fail for 8315 Jan 15 01:26:17 afk Jan 15 01:27:24 ok native-sdk-image I will try on mpc8315e-rdb Jan 15 01:27:41 newbie question: if I 'require' some .inc or .bb file from a recipe in order to change conf settings like PREFERRED_VERSION_xxx it will only apply to the context of that recipe's build and not be sticky outside of that context correct? Jan 15 01:32:55 hbeck? Jan 15 01:33:08 PREFERRED_VERSION you should set in your local.conf Jan 15 01:34:31 woglinde: yeah, I may be trying to do something weird. I want to set up for cross-compiling with a prebuilt external toolchain, but only for specific recipes Jan 15 01:35:03 o.O Jan 15 01:35:04 everything else I'm building 'normally' with the internal Jan 15 01:36:05 hm I think you cannt do that Jan 15 01:39:28 the idea was to build the toolchain with OE, install it, and then use OE to automate building/packaging some of our own libraries with the toolchain at that point Jan 15 01:40:17 as both a sanity check and user-friendliness test, if that makes sense. Maybe I'm taking the wrong approach and need a different overlay area and local.conf? Jan 15 01:41:05 khem: Now pidgin won't compile. I tried cleaning the package then compiling it by itself, but no luck. http://pastebin.com/6ttFpF9Y Jan 15 01:47:50 hbeck: not sure what problem you are trying to solve by doing that extra work. But if you insist upon extra work, then you'll need to do that by using a separate OE build area to build the toolchain, copy the toolchain elsewhere where you will use it as an external toolchain in a separate OE instance to build the rest of your stuff. Jan 15 01:48:44 Of course that answer is provided in the context of an assumption I am making about what problem you are, in fact, trying to solve by doing this. :) Jan 15 01:54:59 mwester-laptop: well, it may not be worth the work. Other than a sanity check on our constructed toolchain I don't think there's that much of a benefit to forcing the external toolchain ... maybe a slightly different version of gcc? Right now I'm writing recipes to build a bunch of small utilities we have (which already build via CMake), OE providing some nice automation and packaging convenience there. Jan 15 01:57:00 sounds like a headache to get it working nicely, a shell script to test builds against our toolchain may be a better solution here in terms of pain Jan 15 02:23:20 Lopi: thats a real problem I think in build Jan 15 02:23:29 Lopi: take a look at pidgin Jan 15 02:23:46 and see if disabling parallel make helps Jan 15 02:23:53 or it could be related to gettext Jan 15 02:23:55 who knows Jan 15 02:24:11 mwester-laptop: did u get a chance to play with latest slugos-image ? Jan 15 02:24:20 mwester-laptop: its been kind to me :) Jan 15 02:24:49 I'm on the road again... NSLU2 is back at home... so, no. Jan 15 02:24:56 heh ok Jan 15 02:26:15 Home tomorrow afternoon, then leaving again on Sunday, back late on Tuesday... so from a practical point-of-view, I doubt I'll get to testing the image on real hardware for a week or so. Jan 15 02:26:40 hmm ok Jan 15 02:27:33 I took my sheevaplug along, though. Jan 15 02:28:21 It's making me think we need a completely different image to be created; not just larger but perhaps ubifs. Jan 15 02:28:51 heh! After the NSLU2, I'm not even sure how to use 512MB of flash! Jan 15 02:58:14 Tartarus: you were right about Os/O2 thing for ppc/uclibc it boots well with -O2 Jan 15 02:58:19 but fails with Os Jan 15 02:58:21 hmmmm Jan 15 02:58:30 bad on one hand **** ENDING LOGGING AT Sat Jan 15 03:00:01 2011