**** BEGIN LOGGING AT Sat Aug 01 02:59:58 2009 Aug 01 08:15:38 03Khem Raj  07org.openembedded.dev * r66d4bf24f0 10openembedded.git/conf/distro/minimal.conf: Aug 01 08:15:38 minimal.conf: Add PREFERRED_PROVIDER for libusb0. Aug 01 08:15:38 Signed-off-by: Khem Raj Aug 01 08:15:48 03Khem Raj  07org.openembedded.dev * r794d00d4bf 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev Aug 01 08:15:49 03Khem Raj  07org.openembedded.dev * r68b24e04cd 10openembedded.git/ (34 files in 17 dirs): Aug 01 08:15:49 uclibc: Convert eligible uclibc recipes to use git instead of svn. Aug 01 08:15:49 uclibc has changed its SCM system to git. We need Aug 01 08:15:51 to adjust the recipes to fetch from uclibc git Aug 01 08:15:53 Signed-off-by: Khem Raj Aug 01 08:50:07 03Khem Raj  07org.openembedded.dev * r3617ddf640 10openembedded.git/recipes/uclibc/uclibc_git.bb: Aug 01 08:50:07 uclibc_git: Remove DEFAULT_PREFERENCE Aug 01 08:50:07 Signed-off-by: Khem Raj Aug 01 11:24:27 03Holger Hans Peter Freyther  07org.openembedded.dev * r7d34a1965b 10openembedded.git/recipes/qt4/ (3 files in 2 dirs): Aug 01 11:24:27 qt4: Add patch to fix the visibility and bsymbolic detection Aug 01 11:24:27 Do not use "$COMPILER" to execute the test. This fails when Aug 01 11:24:27 the passed compiler is something like "ccache gcc". Aug 01 11:24:28 03Holger Hans Peter Freyther  07org.openembedded.dev * r5a63763189 10openembedded.git/recipes/qt4/qt4.inc: Aug 01 11:24:32 qt4: Do not recreate the link if it exists otherwise do_install will fail Aug 01 11:24:34 When running configure more than once the do_install stage Aug 01 11:24:36 will fail due excessive symlinks... Stop doing that. Aug 01 11:24:38 03Holger Hans Peter Freyther  07org.openembedded.dev * ra969fd16fd 10openembedded.git/recipes/qt4/ (5 files): qt4: Use require instead of include to fail on missing includes Aug 01 12:55:32 woglinde: RP: ready to review a base.bbclass change? Aug 01 13:14:56 anyone? Aug 01 13:16:55 morning Aug 01 13:28:16 zecke_: sorry, being called away :( Aug 01 13:28:29 RP: no problem Aug 01 13:32:35 zecke send it to the mailinglist? Aug 01 13:33:02 woglinde: lalala Aug 01 13:34:22 *g* Aug 01 13:48:57 jo ant Aug 01 13:49:37 woglinde: hey Aug 01 15:52:05 re Aug 01 15:54:45 woglinde: sometime ago robert added BPN to strip -native, -cross... away from the PN Aug 01 15:55:01 woglinde: he added a function to base.bbclass and I have added a case for -sdk Aug 01 15:55:09 woglinde: this unbreaks libtool build Aug 01 15:57:59 hm Aug 01 15:58:09 libtool breaks with sdk? Aug 01 15:59:40 woglinde: yes, S points to the wrong directory Aug 01 16:00:45 hm and you have a fix for review? Aug 01 16:04:35 woglinde: my 2nd ego reviewed it already... the fix is elif pn.endswith("-sdk"): return pn[:-6] Aug 01 16:05:22 zecke pasted "review?" at http://paste.lisp.org/display/84573 Aug 01 16:05:32 woglinde: ^^ that is the patch Aug 01 16:06:43 micro + meta-toolchain does not mix well :( Aug 01 16:08:52 gcc does not like to search its' headers in /include Aug 01 16:11:33 heh Aug 01 16:11:42 gcc-cross-sdk to be specific Aug 01 16:12:05 ok Aug 01 16:12:08 must leave house Aug 01 16:12:10 gn Aug 01 16:13:45 zecke zecke hm seems okay Aug 01 16:13:47 push it Aug 01 16:16:48 and gcc-cross and gcc-cross-sdk configure's are almost identical Aug 01 16:17:06 woglinde: thanks, BPN is only used by libtool, so the possible fuckup is quite low Aug 01 16:17:48 has anyone used package_tar lately? Aug 01 16:21:35 me not Aug 01 16:22:07 03Holger Hans Peter Freyther  07org.openembedded.dev * r775c746f69 10openembedded.git/classes/base.bbclass: Aug 01 16:22:07 base.bbclass: Strip -sdk from the PN for BPN as well Aug 01 16:22:07 This is fixing the compilation of libtool-sdk_2.6.6a Aug 01 16:22:07 which is currently the only user of BPN. Aug 01 17:48:37 zecke_: That sounds sane. I'm pretty sure poky does things like that too Aug 01 17:51:23 hi rp Aug 01 18:04:41 if I have to pick a few patches from one branch into other( not complete merge ) is cherry-pick the right way ? Aug 01 18:05:33 yes Aug 01 18:05:55 cherry-pick rockz Aug 01 18:54:16 I gather that Ubuntu is not a good distro choice for an OE development machine, given the problem with /bin/sh pointing to dash instead of bash. Am I right? Aug 01 18:57:06 nope Aug 01 18:57:10 just type bash Aug 01 18:57:20 or set a link from /bin/sh to bash Aug 01 18:58:31 s/or/and Aug 01 19:04:47 dpkg-reconfigure dash Aug 01 19:06:22 yeah "sudo dpkg-reconfigure dash" and say "no", and if your running 8.10 or 9.04 you'll have to do the following as well probably: Aug 01 19:06:22 sudo -s Aug 01 19:06:22 echo -e "vm.vdso_enabled = 0\nvm.mmap_min_addr = 0" > /etc/sysctl.d/60-oe-settings.conf Aug 01 19:06:22 exit Aug 01 19:06:22 sudo invoke-rc.d procps start Aug 01 19:07:09 NvrBst, due to qemu? Aug 01 19:07:17 Tartarus yes Aug 01 19:07:27 glibc locales generating Aug 01 19:07:32 Ah, that's why I don't have that in my script, qemu never ever works for me Aug 01 19:07:44 tartarus hm? Aug 01 19:08:17 For my ppc stuff, I have the segfault on running any program problem Aug 01 19:08:43 I forget if armv6 or v7 works now, but last time I tried it didn't Aug 01 19:10:27 (A quick google showed that the ppc "fix" was to upgrade to 0.10.4, which didn't help, nor did 0.10.5, both of which are a real easy update for someone that can test qemu stuff) Aug 01 19:10:29 hm arm should work flawless Aug 01 19:10:37 ppc is another topic Aug 01 19:10:51 0.10.6 is out Aug 01 19:11:24 not on qemu.org yet :) Aug 01 19:11:57 they forgot update the website Aug 01 19:11:58 i see the ml post tho Aug 01 19:12:03 debian unstable has it Aug 01 19:12:05 and post on ml Aug 01 19:12:55 http://lists.gnu.org/archive/html/qemu-devel/2009-07/msg02520.html Aug 01 19:12:58 probably also an easy update, for someone that can test qemu anyhow :) Aug 01 20:20:21 re Aug 01 20:22:46 jo florian Aug 01 20:59:40 Yay for big company payrolls, i see my check was postmarked the 29th :) Aug 01 20:59:42 ga Aug 01 21:00:18 check postmark? Aug 01 21:00:38 god thanks I live in germany why the money flows directly to my bank account Aug 01 21:00:44 heh Aug 01 21:00:57 Takes up to 2 pay periods for that to happen here :| Aug 01 21:01:37 jesus Aug 01 21:17:00 jo ant Aug 01 21:20:57 hi woglinde Aug 01 21:21:35 woglinde: I stll have the doubt: how to properly stage mtd-user.h and mtd-abi.h (klibc)? Aug 01 21:22:04 ant Aug 01 21:22:19 do the fowllowing Aug 01 21:22:28 copy them over in the source per patch Aug 01 21:23:08 woglinde: http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=8361ab6537f2335efa18fde8ef369167250127b3 Aug 01 21:23:09 thats a good solution and nobdy gets hurd Aug 01 21:24:19 yeah Aug 01 21:25:16 is it decided in kernel.bbclass which headers to stage? Aug 01 21:25:28 ?????? Aug 01 21:25:46 you mean from kernel-dir? Aug 01 21:25:59 I'm a bit confused..somebody said now the kernel headers are sanitized enough to be staged instead of the libc-headers... Aug 01 21:26:15 ?????? Aug 01 21:26:33 kernel.bbclass ->kernel_do_stage() { Aug 01 21:26:49 nope Aug 01 21:26:57 the kernelheaders comes from /devel/arm/git/openembedded/recipes/linux-libc-headers/ Aug 01 21:29:58 the mtd headers I have are in /oe/build/tmp/staging/armv5te-angstrom-linux-gnueabi/usr/include/mt Aug 01 21:29:59 d Aug 01 21:30:28 (mtd-user.h and mtd-abi.h) Aug 01 21:31:25 otherwise I have mtd.h in /oe/build/tmp/staging/armv5te-angstrom-linux-gnueabi/lib/klibc/include/linux/mtd Aug 01 21:31:40 but I avoided it Aug 01 21:32:40 Maybe you need a newer version of linux-libc-headers? Aug 01 21:32:47 hey Tartarus Aug 01 21:32:56 I just followed mtd-utils path.. Aug 01 21:33:04 staging own headers Aug 01 22:13:01 hello xjqian Aug 01 22:13:54 hi ant Aug 01 22:19:58 hi bluelightning Aug 01 22:50:48 03Leon Woestenberg  07org.openembedded.dev * rc68980af2f 10openembedded.git/ (3 files in 2 dirs): libnxml: Added 0.18.3. Changed upstream URL www2.* to www.*. Aug 01 22:50:52 03Leon Woestenberg  07org.openembedded.dev * r455a67ee98 10openembedded.git/recipes/libnxml/libnxml_0.16.bb: libnxml: Use .inc and INC_PR. Aug 01 22:52:17 03Leon Woestenberg  07org.openembedded.dev * r9c637acc4d 10openembedded.git/recipes/libnxml/libnxml.inc: libnxml.inc: Fix INC_PR from "0" to "r0". Aug 01 22:56:44 03Leon Woestenberg  07org.openembedded.dev * r309ba5df8f 10openembedded.git/ (3 files in 2 dirs): libmrss: Use .inc and INC_PR. Aug 01 23:04:40 03Leon Woestenberg  07org.openembedded.dev * rd5b3bbde9c 10openembedded.git/ (4 files in 2 dirs): libmrss: Added 0.19.2. Aug 01 23:04:41 03Leon Woestenberg  07org.openembedded.dev * r421bc0e6fc 10openembedded.git/recipes/libnxml/libnxml_0.18.3.bb: libnxml: Added 0.18.3. Aug 01 23:28:59 03Leon Woestenberg  07org.openembedded.dev * r5473be7cf9 10openembedded.git/ (conf/checksums.ini recipes/rsstail/rsstail_1.5.bb): rsstail: New recipe for 'tail' like RSS feed reader. (Needs QA love.) Aug 02 02:18:12 03Holger Hans Peter Freyther  07org.openembedded.dev * rd7931d9aff 10openembedded.git/recipes/gdb/ (3 files in 2 dirs): Aug 02 02:18:12 gdb: Add a patch to fix compilation of gdb 6.8 Aug 02 02:18:12 gcc warns about the subscript_array being uninitialized, Aug 02 02:18:12 just memset this area... luckily I can not break much as Aug 02 02:18:12 this is for fortran code. **** ENDING LOGGING AT Sun Aug 02 02:59:57 2009