**** BEGIN LOGGING AT Mon May 04 02:59:59 2015 May 04 07:02:30 good morning May 04 13:19:31 I need help May 04 13:19:47 What is the correct procedure to add kernel May 04 13:19:56 What is the correct procedure to add kernel "defconfig" options May 04 13:20:26 I thought it was=> bitbake savedefconfig; copy defconfig linux/machine May 04 13:20:56 but the defconfig generated is almost empty and does not contain all entries that was there before May 04 13:21:41 What I want to do is: bitbake virtual/kernel -c menuconfig ----> set some options ---> save it into the bitbake system May 04 13:23:17 and I don't have access to kernel config fragments so the documentation is useless to me May 04 13:23:49 pompomJuice: basically the latter way is ok, you just need to pick the .config file that menuconfig creates and push it to whereever you kernel recipe wants it May 04 13:25:43 how does defconfig get applied? May 04 13:25:47 or when? May 04 13:26:08 if I manually put my .config there it might jam with the current defconfig system May 04 13:26:24 um, why? May 04 13:26:44 your kernel recipe usually contains where it wants to pull its defconfig from May 04 13:26:53 place it there. May 04 13:27:34 There is a machanism/convention inside bitbake/oe that looks for a ${MACHINE}/defconfig folder inside your kernel folder and mashes it into your kernel config during build stage May 04 13:28:18 what you are saying is... if I replace my defconfig with the .config file all will be fine? May 04 13:28:57 pompomJuice: 'make savedefconfig' only saves the values that aren't set to their default value May 04 13:29:06 aaah May 04 13:29:37 so can I take the defconfig made with savedefconfig and use that as a defconfig and it will work even though it looks very empty. May 04 13:29:49 Let me try that May 04 13:29:52 i personally favor a fixed config included in my layer, but of course, YMMV May 04 13:29:53 right May 04 13:30:02 yes '' | ARCH=something make oldconfig May 04 13:30:05 replace 'something' May 04 13:30:16 that will populate .config will *all* config variables May 04 13:30:58 finally... that makes sense to me. May 04 13:31:09 dang May 04 13:31:13 we did it wrong then. May 04 13:34:42 pompomJuice: keep in mind that vendor kernel might have broken Kconfig dependencies May 04 13:35:08 so 'make menuconfig ; make savedefconfig ; cp defconfig . config ; yes '' | make oldconfig' can be lossy May 04 13:35:14 *cough* TI kernels *cough* May 04 13:35:39 took me a week to figure out why USB stopped working May 04 13:37:28 hehe May 04 13:37:34 luckaly we are moving away from TI kernels May 04 13:37:40 Ti is a swear word here now May 04 13:43:42 our next gen device freescale wandboard-quad... what a pleasure. May 04 14:19:04 koen, I need help again. I need to upgraede connman to 1.29. oe-core contains a connman.inc with an do_configure_append that contains outdated sed calls. What would be the way to overwrite/remove that append and add a new one without making upstream changes? May 04 14:19:17 there isn't May 04 14:19:23 ok May 04 14:19:24 unless you overlay the .inc May 04 14:19:26 I thought so May 04 14:19:33 overlay? May 04 14:19:41 copy the file completely? May 04 14:21:23 aah I see... overlay the meta layer May 04 14:21:24 ok thanks May 04 14:56:31 pompomJuice, koen: anonymous python function which gets the value of do_configure, removes the sed lines, and re-sets it May 04 14:56:41 pompomJuice, koen: not pretty, but I've done it successfully before May 04 14:57:50 aah May 04 14:57:53 that would work May 04 15:36:17 pompomJuice, are you using oe-core master? You could send in the update to that? May 04 15:38:00 I will do that... was in crisis mode I had to get it done quickly May 04 15:39:53 ok May 04 15:39:55 just checking May 04 15:40:10 I saw it is on the upgrade list May 04 15:56:32 pompomJuice, also, I see the connman update patches on the oe-core list. Looks like about 1.5 hours old :) May 04 15:58:17 oh my May 04 15:58:31 I want it May 04 16:13:43 RP, you have annoyed mranostay May 04 16:13:50 you set the bar pretty high already May 04 16:46:22 Crofton|work: heh. To be honest I doubt I'm going to do that every day May 04 16:47:10 yeah, but with a handful of short tides, you will stay on the weeks leaderboard :) May 04 16:51:13 Crofton|work: there is that :) May 04 16:53:44 there wasn't any cheating, right? May 04 16:59:52 Crofton|work: no, honestly no cheating on that. Just got a bit carried away and forgot to turn around May 04 17:00:34 Crofton|work: I have never been on 80% of that route so navigation was "fun", it is mostly cycleways though May 04 17:02:07 it look slike you alsmot made it halfway across the country May 04 17:02:40 Crofton|work: more like a third and I turned back before it got properly hilly May 04 17:03:03 we have different definitions of hilly :) May 04 17:03:14 lol.. all depends on where you live.. May 04 17:03:21 my definition of 'hilly' is from the US Midwest.. ;) May 04 17:03:54 * Crofton|work mocks fray May 04 17:04:03 (if you've never been to the midwest.. our hills are at the most 20-100 foot, rarely with more then a 10 degree slope.. ;) May 04 17:04:06 I had a friend that lived in Florida a while May 04 17:04:27 overpasses were climbs May 04 17:04:39 yup May 04 17:04:42 fray: If I went further it would mean the lake district and the UK's highest market town May 04 17:05:15 Much as I love Alston, I'm not up to cycling there May 04 17:10:18 hmm, someone is racy in cmake recipes now May 04 19:50:58 Hmm, we should think about having some general variable naming guidelines May 04 19:51:27 e.g. things like should multiword variable names be ordered from general to more speicifc, or vice versa, or the order of subject/object/verb May 04 19:51:36 USE_NLS vs *_ENABLED or *_DISABLEd May 04 19:51:50 when adding a new variable today, would you want DISTRO_FEATURES or FEATURES_DISTRO? May 04 19:52:07 we're not very consistent today.. May 04 19:57:25 i should say, USE_NLS vs *_ENABLED vs ENABLE_* vs *_DISABLED vs DISABLE_*, etc. ordering and meaning, typing relative to naming as well, when to use prefixes vs suffixes vs flags.. **** ENDING LOGGING AT Tue May 05 02:59:59 2015