**** BEGIN LOGGING AT Thu May 22 02:59:58 2014 May 22 07:10:40 morning all May 22 07:10:58 RP_: what's your opinion on using cp --reflink for install/package ? May 22 07:11:04 hey bluelightning May 22 07:12:02 hi all May 22 07:13:53 hey woglinde May 22 07:15:30 hi koen, woglinde May 22 07:19:03 koen: we already extensively use hardlinks so it probably wouldn't buy us much May 22 07:20:53 right May 22 07:21:05 and I fear on non-COW systems it will fallback to regular cp :( May 22 07:21:50 koen: yes :( May 22 07:24:49 I pointed /work to the wrong mountpoint last week and wondered why the builds were taking so long May 22 07:25:05 turns out I pointed it to the btrfs partition instead of the ext4 one May 22 07:25:37 btrfs is great for DEPLOY_DIR and DL_DIR, but /work, no May 22 07:45:16 good morning May 22 08:14:16 hi mckoan May 22 08:33:29 hi woglinde May 22 12:49:13 what is the suggested way to trace the execution of bitbake ? Can I just inject 'print' statements into poky ? Or is there a canonical logging mechanism I should use ? May 22 12:50:10 (I'm trying to debug a failure the results in a python stack trace in meta/lib/oe/sdk.py) May 22 12:51:54 stefan__: you can do bb.warn('your message here') from python code and it will print a warning May 22 12:52:00 that's what I usually do May 22 12:52:16 bluelightning: very nice, thanks. May 22 12:56:00 bluelightning: what package defines 'pp' ? May 22 12:56:21 stefan__: you mean bb? May 22 12:56:24 (I'm trying to call that from sdk.py, and thus need to import it first). May 22 12:56:26 yes May 22 12:56:32 you shouldn't need to import it May 22 12:56:44 it's provided as part of bitbake and should be imported automatically May 22 12:57:00 not into python modules. sdk.py complains about it. May 22 12:57:08 ok, onse sec May 22 12:57:14 one sec May 22 12:57:55 that's odd, because I can see references to the bb module in that file already... May 22 12:58:31 we are talking about meta/lib/oe/sdk.py, yes ? There is no other reference to 'bb' in it. At least not in the 'daisy' branch I'm using. May 22 12:59:19 oooh, sorry. Pilot error. :-( May 22 12:59:53 right, I was about to say "yes there is - bb.utils.mkdirhier(self.sdk_output)" May 22 13:01:25 ok, good, I can debug this now, and will follow up with my report on the list. Thanks ! May 22 13:02:04 np May 22 14:30:30 howdy smart people May 22 14:32:59 any times on using bitbake's include/require function to files located somewhere else... somehow relative path does not work! May 22 14:33:00 ? May 22 14:33:18 uuh wait May 22 14:33:45 oops nm May 22 14:36:45 relative paths are found relative to the locations listed in BBPATH. BBPATH does include the directory where the recipe resides, temporarily May 22 15:56:40 Does https://gist.github.com/kergoth/9d7d31999db83ab045db seem reasonably sane? I think I'll open a yocto bugzilla feature request about it May 22 16:13:08 moin May 22 16:40:54 JaMa: I'm wondering if those RDEPEND lines in https://github.com/meta-qt5/meta-qt5/blob/master/recipes-qt/qt5/qtwebkit-examples.inc are correct. qtwebkit-examples.inc is required by qtwebkit-examples_5.2.1.bb and qtwebkit_git (for both, ${PN} == qtwebkit-examples). So, shouldn't RDEPENDS_${PN}-examples be expanded to RDEPENDS_qtwebkit-examples-examples? May 22 16:46:51 Surprisingly (to me, at least), the way it is now (RDEPENDS_${PN}-examples) works. If I remove -examples, the generated package doesn't depend on qtwebkit-qmlplugins ca-certificates: http://paste.debian.net/101231/ May 22 17:30:39 is there a good way to track a patch I submitted to oe-core's progress through the machine? May 22 17:33:01 i think that's what patchwork was for, originally, but i don't think it's used by saul & co May 22 17:33:41 oe-core doesn't use pw May 22 17:34:54 patchwork does track patches sent to the oe-core *list*, its just folks aren't using it during the merges May 22 17:35:56 ah May 22 17:36:14 I knew it wouldn't answer my question at least :) May 22 17:36:30 hehe May 22 17:40:10 JaMa: otavio helped me to understand that the qt5 stuff is actually correct. It's just a big gotcha to me that "bitbake qtwebkit-examples" actually generates a "qkwebkit-examples-examples" package. May 22 17:41:14 Specially because it's hard to see the second -examples when you tab-complete the package name. May 22 17:43:25 And it just generates an error at rootfs-generation time. May 22 17:43:52 (When you add qtwebkit-examples to the image) May 22 17:46:25 that does sound iffy May 22 17:51:19 A big gotcha, to say the least. I looked several times at "qtwebkit-examples-examples" in the middle of all those logs and huge pathnames and couldn't figure out there was two -examples there. May 22 18:29:26 htop fials to build, missing ncurses.h file May 22 18:30:48 | /home/balister/src/oe-core/build/tmp-eglibc/work/armv7at2hf-vfp-neon-oe-linux-gnueabi/htop/1.0.2-r0/htop-1.0.2/RichString.h:19:29: fatal error: ncursesw/curses.h: No such file or directory May 22 18:30:49 | #include May 22 18:31:04 missing dep? May 22 18:33:05 yeah, waiting on the build to finally fail completely May 22 18:34:14 htop depends on ncurses May 22 18:34:38 weird May 22 19:06:22 kergoth, ncurses.h exists but not in the ncursesw directory May 22 19:14:08 wtf May 22 19:35:36 hopefully jsut layer confusion due to switching across branches May 22 21:50:06 Crofton|work: set S = B and see if that helps May 22 21:50:17 I think its a failed configure test May 22 21:50:29 due to wrong m4 macro in htop most probably May 22 22:44:20 khem: people say, mips32r2 gained with gcc4.9 **** ENDING LOGGING AT Fri May 23 03:00:01 2014