**** BEGIN LOGGING AT Tue Sep 20 02:59:56 2011 Sep 20 06:41:10 good morning Sep 20 06:44:43 hi mckoan, all Sep 20 08:05:38 I want to implement my own do_fetch function. Is there a variable that points to the recipy folder itself? Sep 20 08:11:27 I'm still basing my local repo and work on org.openembedded.dev. Am I old-fashioned? is oe-core the place I should follow for Angstrom/Beagle development? Sep 20 08:16:38 tasslehoff, hasn't koen moved the latest Angstrom to oe-core? Sep 20 08:16:49 thats what it looks like in narcissus Sep 20 08:27:08 hi all Sep 20 08:27:20 just have a minor question Sep 20 08:28:25 if we are using different layers and selection of particular layer for recipe only depends on BBLAYER_PRIORITY or there is some thing else as well Sep 20 08:29:30 I am using my own layer meta-oe and oe-core .... uboot recipe is present in mylayer and oe-core but oe-core u-boot recipe is being picked all the time Sep 20 08:29:36 there is other to know: http://lists.linuxtogo.org/pipermail/bitbake-devel/2011-August/001169.html Sep 20 08:33:12 thanks ant_work ..... I am looking innto it and get back to you if I have more questions Sep 20 08:34:34 priority knowledge is split between Sep 20 08:34:34 layer.conf and bblayers.conf today Sep 20 08:34:57 I agree with others: that seems bad to me Sep 20 08:40:43 ka6sox-away: ah, then I should probably follow him :) Sep 20 08:47:21 JaMa: ping Sep 20 08:47:26 pong Sep 20 08:47:50 hi, I've seen you're experimenting with X Sep 20 08:48:11 so, which server for meta-hh ? Sep 20 08:48:37 I'm using 1.11 but that's only in jansa/x11 branch in oe-core-contrib Sep 20 08:48:55 and you would need also 3 patches from jansa/spitz for meta-hh Sep 20 08:49:16 yes, that's the kdrive/lite iirc Sep 20 08:49:28 not full xorg-xserver Sep 20 08:49:28 https://gitorious.org/shr/meta-handheld/commits/jansa/spitz Sep 20 08:50:08 indeed, back to xserver-xf86-lite Sep 20 08:50:34 at this point we are stuck with calibration, with both servers Sep 20 08:50:46 xinput-pointercal is working Sep 20 08:50:51 jitter? Sep 20 08:51:06 that's kernel driver issue not xorg Sep 20 08:51:21 yes but is still unusable... Sep 20 08:51:49 weel, it was, I did not retest it lately Sep 20 08:55:57 ant_work: and -lite is not kdrive.. it's just withoug glx and dri Sep 20 09:00:30 morning all Sep 20 09:08:12 JaMa: what about the xorg-xserver in meta-oe? Sep 20 09:08:22 ant_work: it's removed in jansa/x11 Sep 20 09:08:26 :) Sep 20 09:08:36 bluelightning: taking note? Sep 20 09:09:00 ant_work: about? Sep 20 09:09:46 many machines in meta-hh are using PREFERRED_PROVIDER_virtual/xserver = "xserver-xorg" Sep 20 09:10:09 this is the meta-oe xserver Sep 20 09:10:45 it looks like we can do https://gitorious.org/shr/meta-handheld/commit/9fad2c5a2ac10365688a582f457d2224f1866245 Sep 20 09:11:06 for many if not all devices to have a common testbase Sep 20 09:11:46 ant_work: ah no big difference http://paste.pocoo.org/show/479031/ Sep 20 09:13:47 ant_work: ah no big difference http://paste.pocoo.org/show/479031/ Sep 20 09:13:55 ups Sep 20 09:16:12 ant_work: right Sep 20 09:57:39 if I have set priority of my-collection oe-core meta-oe like 3,2 and 1 ..... my linux recipe is present in my-collection and oe-core and and both have same version ...... in my opinion recipe from my collection should be picked as it has higher priority but bitbake is picking from oe-core Sep 20 09:58:22 what other things are involved in it? Sep 20 09:58:47 ant_work: I have gone through the thread but I have different problem Sep 20 10:00:31 hey guies Has anyone worked on Pollux processors ,board NC600 or NC650 series ,I am trying to find tht in open emmbeded repo ,whether we hwve it? Sep 20 10:15:20 morning! Sep 20 10:15:40 any wiki/doc I should read when moving to oe-core? Sep 20 10:31:10 found the problem in priority of layers Sep 20 10:32:13 the BBFILES variable was not correctly set so that is why it was not picking recipes from my layer if the same version is present in lower priority layer as well Sep 20 10:32:49 after settign the correct value of BBFILES recipes from correct layer is being picked Sep 20 10:33:40 just wondering if BBFILES is not correctly set then no recipes should be picked from that layer .... isn't it a bug that recipes are still being picked? Sep 20 11:07:59 tasslehoff: http://www.openembedded.org/wiki/OpenEmbedded-Core if you haven't seen it already Sep 20 11:08:29 Noor: if you set the priority of your layer higher it should override other layers Sep 20 11:09:19 Noor: what were the incorrect/correct values of BBFILES in your case? Sep 20 11:10:48 bluelightning: in my case BBFILES was being set to $LAYERDIR .... it should be be $LAYERDIR/recipes/*/*.bb Sep 20 11:14:33 bluelightning: had actually not seen that one. thanks. Sep 20 11:15:15 tasslehoff: inexplicably is not on Main_Page Sep 20 11:15:15 bluelightning: the interesting thing is that I had set PREFERRED_PROVIDER_virtual/kernel = linux ... that recuipe was present in my layer and it was being picked and my layer had highest priority Sep 20 11:15:57 bluelightning: i copied the same recipe to oe-core and bitbake started picking the recipe from oe-core instead of my layer Sep 20 11:19:23 so it is looking strange to me that if we set preferred_provider and that recipe is only present in my layer and BBFILES variable of that layer is not pointing to correct files bitbake still picks the recipe Sep 20 12:33:16 ant_work: odd :) Sep 20 12:34:18 should my layers of choice be cloned inside or alongside oe-core? Sep 20 12:36:55 morning all Sep 20 12:41:31 hi pb_ Sep 20 12:41:55 tasslehoff: alongside, keeps things tidy Sep 20 13:02:57 03Steffen Sledz  07org.openembedded.dev * rd4c60de1b5 10openembedded.git/recipes/rsyslog/rsyslog.inc: Sep 20 13:02:57 rsyslog: change path for config includes to /etc/rsyslog.d Sep 20 13:02:57 Signed-off-by: Steffen Sledz Sep 20 13:35:17 lumag, thanks for taking care of the proxy statement! Sep 20 13:35:35 I was just thinking about looking into it when your email popped up Sep 20 13:36:15 Crofton|work, :) Sep 20 13:36:38 now all I need to find is someone to run a TSC election Sep 20 13:36:59 denix0, you awake? Sep 20 14:07:41 bye! Sep 20 14:41:53 Hi Sep 20 14:44:13 I want to use external toolchain to prevent unnecessary compilation and to reduce build time Sep 20 14:44:47 I have added these lines to my local conf up to an old email from oe mail list Sep 20 14:44:57 morning Sep 20 14:45:02 gm Sep 20 14:45:22 bah why cant I find a generic pcb spacer kit Sep 20 14:45:35 TARGET_VENDOR = "-none" Sep 20 14:45:44 TOOLCHAIN_TYPE = "external" Sep 20 14:45:54 TOOLCHAIN_BRAND = "csl" Sep 20 14:46:20 is there anything I have missed? Sep 20 14:46:32 you're using classic oe, not oe-core? Sep 20 14:46:33 mgundes: which setup u r using oe-core or openembedded Sep 20 14:46:39 :) Sep 20 14:46:41 my cross compiler is in my PATH Sep 20 14:46:51 openembedded Sep 20 14:48:22 morning kergoth Sep 20 14:48:40 hey pb_ Sep 20 14:50:59 mgundes: u need to set TOOLCHAIN_PATH and TOOLCHAIN_SYSPATH as well Sep 20 14:52:00 ok I am trying Sep 20 14:52:38 TOOLCHAIN_PATH = "/opt/CodeSourcery/Sourcery_G++_Lite/" Sep 20 14:52:48 TOOLCHAIN_SYSPATH = "${TOOLCHAIN_PATH}/arm-none-linux-gnueabi" Sep 20 14:53:12 should I set anything to or unset LIBC? Sep 20 15:06:43 mgundes: I think that is it Sep 20 15:09:38 Noor: I think I dont need LIBC, since my toolchain have libraries and other binaries Sep 20 15:09:42 right? Sep 20 15:10:21 if you are using openembedded external-toolchain-csl recipe then LIBC is not required Sep 20 15:13:03 but in many configuration files there are checks about LIBC like that: LIBC ?= "eglibc" Sep 20 15:13:26 then one type of LIBC will be compiled, right? Sep 20 15:13:27 03Philip Balister  07master * r4addbe94be 10openembedded.git/recipes/uhd/uhd_git.bb: Sep 20 15:13:27 uhd : Update SRCREV to head of master branch. Sep 20 15:13:27 This revision update is needed to pick up support for current USRP E100 Sep 20 15:13:27 fpga interface. Sep 20 15:13:27 Signed-off-by: Philip Balister Sep 20 15:13:41 how can I prevent it, have any idea? Sep 20 15:13:59 a pox on people that won't make releases :) Sep 20 15:22:03 a pox on power failures Sep 20 15:22:09 goo dthing I pushed Sep 20 15:24:10 mgundes: no, LIBC controls the behavior of things that care what libc you're using. it has nothing to do with whether it'll be built from source or used from external Sep 20 15:28:56 hm ok thanks kergoth, I hope any glibc would not be compiled. I will try it. Sep 20 15:30:58 by the way, I checked environment after parsing and I see that in environments: Sep 20 15:31:01 PREFERRED_PROVIDER_virtual/libc="eglibc" Sep 20 15:32:17 TARGET_SYS="arm-oe-linux-gnueabi" Sep 20 15:32:33 export CC="arm-oe-linux-gnueabi-gcc -march=armv6k -mtune=arm1176jz-s -mthumb-interwork -mno-thumb" Sep 20 15:32:49 hm I think something is wrong in my config files Sep 20 15:39:11 can we update BB_HASHBASE_WHITELIST in any class .... I did it in native.bbclass but that value that I added was not going in cachefile Sep 20 16:04:01 Crofton: pong Sep 20 16:16:21 Crofton|work: pong Sep 20 16:24:14 did you find an answer to building linux-3.0 with the release branch? Sep 20 16:25:41 Crofton|work: sifting through the patches in oe.dev/meta-oe Sep 20 16:26:48 morning Sep 20 16:26:56 I think I'll just have to backport all the fixes from meta-oe to maintenance Sep 20 16:27:10 urg Sep 20 16:42:43 denix0, do you know when you will be able to submit the fixes? Sep 20 16:43:14 I also think the compiler in dev is making faster code than the one in maint Sep 20 16:43:18 I need to look at this some Sep 20 16:52:20 bluelightning: ping? Sep 20 16:52:32 Jay7: pong Sep 20 16:52:38 bluelightning: WRT konsole bug Sep 20 16:52:48 yep Sep 20 16:52:49 konsole --nofork does the trik Sep 20 16:53:08 no, unfortunately that only works when running from within another konsole session :/ Sep 20 16:53:13 I tried that already Sep 20 16:53:18 ah, hm.. Sep 20 16:53:29 well, I'll look deeper :) Sep 20 16:53:49 AFAICT it requires a number of fixes within konsole itself Sep 20 16:55:37 well.. then I'll install some other *term :) Sep 20 17:50:55 evening! Sep 20 17:58:46 I wander: if I fix sdk problems by changing crossdk .inc files in oe-core, should I also submit PR bump patches to meta-oe? Sep 20 19:26:17 I have a working u-boot recipe, but I want to have two u-boot images with different default environments. What's the recommended approach for creating the recipe for this other image? Sep 20 19:26:43 I could just duplicate the recipe, add one patch to the copy and be done with it, but that's not pretty, right? Sep 20 19:30:05 koobe_: there are some recipes about for UBOOT_MACHINES to build u-boot for multiple cofigurations Sep 20 19:30:16 not the plural Sep 20 19:30:18 note* Sep 20 19:31:34 UBOOT_MACHINE is the name of u-boot configuration that is to be selected for a given OE machine? Sep 20 19:31:52 but I'd like to have two slightly different u-boot images for the same machine Sep 20 19:33:02 oh, I see. uboot_machineS would invoke multiple u-boot configurations for a single machine? Sep 20 19:36:36 Can anyone tell me how to tell openembedded to use a particular kernel version (besides the default)? Sep 20 19:37:03 e.g. if I wanted to build 2.6.34 instead of 2.6.39 Sep 20 19:39:09 pack3754: did you try PREFERRED_VERSION_linux in your machine.conf? Sep 20 19:39:52 pack3754, usually it's done in the kernel recipes itself: Sep 20 19:40:23 you usually have DEFAULT_PREFERENCE_machine = "1" Sep 20 19:40:38 and DEFAULT_PREFERENCE = "-1" #applies to all other machines than "machine" Sep 20 19:41:11 since he asked about version I told the version Sep 20 19:41:11 then there is still what kind of kernel you want Sep 20 19:41:12 like linux-foo Sep 20 19:41:15 or linux-bar Sep 20 19:41:22 that is done in machines usually Sep 20 19:41:43 a third variable named COMPATIBLE_MACHINE also exist Sep 20 19:42:32 s/machines usually/machines config usually/ Sep 20 19:43:03 msm: thanks, I think that UBOOT_MACHINES is the right way to go Sep 20 19:43:04 COMPATIBLE_MACHINE prevent bitbake to parse the recipe if your machine is incompatible Sep 20 19:43:04 it's also added in the kenrel recipe in question Sep 20 19:43:24 Ok, so I'm looking at conf/machine/overo.conf Sep 20 19:43:41 and I see Sep 20 19:43:42 PREFERRED_PROVIDER_virtual/kernel = "linux-omap3" Sep 20 19:44:17 pack3754, yes that is for the kind of kernel, not its version Sep 20 19:44:30 version is like 2.6.32 Sep 20 19:44:50 koobe_: yes, you got it UBOOT_MACHINES = "P1020RDB P1020RDB_SPIFLASH etc" Sep 20 19:45:29 ok, I don't see a PREFERRED_VERSION_linux defined Sep 20 19:55:15 GNUtoo: I feel PREFERRED_VERSION_linux should be defined in the machine.conf file.. is that correct? Sep 20 19:55:55 no idea on what "should be", I can only tell what's done Sep 20 20:00:00 so my default version is 2.6.39... but I don't see this referenced anywhere in conf/* Sep 20 20:05:10 GNUtoo: before I start mucking around with things.. I've seen that it's not recommended to make any changes inside org.oe.dev/ , but instead to create a directory user.collection. Sep 20 20:05:50 what should go inside the user.collection dir? Sep 20 20:06:00 I'll be back later but learn about oe-core Sep 20 20:06:18 ok, thanks GNU Sep 20 20:11:42 pack3754, learn about openembedded-core since openembedded.dev is deprecated and nearly everybody switched to openembedded-core Sep 20 20:12:47 if I search for oe-core, it seems like Yocto is heavily involved with this Sep 20 20:12:57 yocto and oe both use oe-core as their core Sep 20 20:13:12 hoi kergoth Sep 20 20:13:30 yocto then uses meta-yocto and other bits on top of that, and oe has meta-oe Sep 20 20:13:32 hey woglinde Sep 20 20:13:35 how's it going? Sep 20 20:14:17 hm still have my meta-tegra layer laying arround here Sep 20 20:14:38 kergoth: thanks! this means plain old OE is not going anywhere, just that Yocto is also building on top of the same oe-core? Sep 20 20:14:46 pack3754: no Sep 20 20:14:57 the old openembedded repository isn't being developed in at all. Sep 20 20:15:08 new work is done based on oe-core + meta-openembedded Sep 20 20:15:20 recipes are transitioned from oe to whatever layer is appropriate for them Sep 20 20:15:37 meta-oe being the catch-all for stuff that doesn't belong anywhere else Sep 20 20:16:04 ok, this is good to know.. I'm trying to follow instructions from gumstix. Feels like some things may be out of date Sep 20 20:16:16 hi woglinde Sep 20 20:16:47 woglinde, meta-tegra, does the tegra you have have a neon unit? Sep 20 20:16:55 if not how do you handle that? Sep 20 20:18:23 tegra has none neon Sep 20 20:18:25 thats the problem Sep 20 20:19:27 ah even tegra1? Sep 20 20:19:31 I tought that only tegra2 lacked neon Sep 20 20:19:47 I'm interested because I know that some machines have broken neon: Sep 20 20:19:49 for instance: Sep 20 20:20:18 *palmpre* machines have neon issues I think(not sure but what could make mplayer,enlightenment etc...freeze the phone otherwise...) Sep 20 20:20:35 *some imx51 of rev < 3 have broken neon Sep 20 20:20:41 gnutoo hm? Sep 20 20:20:51 and strangely mplayer can freeze the machine too Sep 20 20:20:55 tegra-meta will contain more tegra Sep 20 20:21:03 maybee tegra3 for instance Sep 20 20:21:09 ok Sep 20 20:29:02 hey. i'm trying to configure a dapm device, and i was wondering if anyone here has experience with this Sep 20 20:37:46 hi bluelightning_ Sep 20 20:38:34 CMoH, maybe try in the alsa soc channel Sep 20 20:38:47 alsa soc? Sep 20 20:39:01 let me search for it Sep 20 20:39:28 ty Sep 20 21:09:15 is there a way to do the reverse of an overrides? Sep 20 21:09:25 like IMAGE_INSTALL_notdef(var) Sep 20 21:12:20 what do you mean? Sep 20 21:12:38 I want to remove a package from an image if's some OVERRIDES is defined Sep 20 21:12:44 if you want a value when no override is applied, that's what the main variable name is Sep 20 21:13:09 that's the reverse of an override append, not the reverse of an override :) Sep 20 21:13:09 right, i want to be included for all images except powerpc64 Sep 20 21:13:25 FOO = "base value" Sep 20 21:13:30 FOO_powerpc64 = "other value" Sep 20 21:13:36 IMAGE_INSTALL_append = "${FOO}" Sep 20 21:13:41 or whatever Sep 20 21:13:49 but in my case FOO is a package list Sep 20 21:13:54 hmm ok wait.. i see it thanks Sep 20 21:13:56 and? :) Sep 20 21:14:14 the key is separating out the bit that's affected into its own variable, so you can juse use overrides as usual there Sep 20 21:14:17 s/juse/just/ Sep 20 21:14:43 right, the own variable part is key =) Sep 20 21:14:46 * kergoth nods Sep 20 21:15:15 lumag: hi Sep 20 21:15:18 msm the real way would fix the packages for ppc64 Sep 20 21:15:25 hi bluelightning Sep 20 21:15:31 hi woglinde Sep 20 21:15:48 woglinde: trust me that's my intention Sep 20 21:16:14 woglinde: problem right now is the recipe (interal crap mind you) only builds on 32bit Sep 20 21:16:43 i've got a lot of fixes for ppc64 forthcoming Sep 20 21:17:28 03Joshua Lock  07master * rab6c63f357 10bitbake.git/lib/bb/ui/ (crumbs/hobeventhandler.py hob.py): (log message trimmed) Sep 20 21:17:28 hob: fix build again when building packages only Sep 20 21:17:28 The different code paths for package vs image builds have lead to some bit Sep 20 21:17:28 rot in the package only build and an incompatability between package only Sep 20 21:17:28 builds and the build again mechanism. Sep 20 21:17:28 This patch unifies the code paths and fixes build again for package only Sep 20 21:17:29 builds. Sep 20 21:17:36 03Joshua Lock  07master * rbdbd724e80 10bitbake.git/lib/bb/ui/crumbs/hobprefs.py: (log message trimmed) Sep 20 21:17:36 ui/crumbs/hobprefs: re-enable reloading of data after prefs changes Sep 20 21:17:36 When disabling the toolchain w/ headers checkbox in an early commit I Sep 20 21:17:36 accidentally disabled too much code. This re-enables adding the custom Sep 20 21:17:36 response handler to the Preferences dialog which triggeres a reparse of Sep 20 21:17:37 the metadata when appropriate preference changes are made. Sep 20 21:17:38 Fixes [YOCTO #1479] Sep 20 21:17:38 03Joshua Lock  07master * r705156e281 10bitbake.git/lib/bb/ui/hob.py: Sep 20 21:17:39 hob: enable package only builds even if an image has been built Sep 20 21:17:39 Once an image build has been performed the selected_image variable would Sep 20 21:17:40 prevent the package only build choice being triggered - by setting the Sep 20 21:17:40 variable to None on reset we are able to perform package only builds after Sep 20 21:17:41 an image build. Sep 20 21:17:41 Signed-off-by: Joshua Lock Sep 20 21:19:11 woglinde: You mentioned earlier I could resolve a problem with Kernel.org being down by using the github mirror instead.. Where would I specify which mirror I would like to use? Sep 20 21:19:41 pack I dont know which tree you want from kernel.org Sep 20 21:21:25 ha =P neither do i! I'll try to track sown some details for you Sep 20 21:22:00 or "details for me" rather, that might be useful for you when I ask for help Sep 20 21:22:33 hm maybee I will be at sleep then Sep 20 21:23:33 well, I won't complain too loudly Sep 20 21:23:41 as long as you're back some time in the nenxt days ;) Sep 20 21:24:53 uhm Sep 20 21:25:14 I think you then would figure it out on your own Sep 20 21:25:49 pack3754: there are a couple of ways I can think of Sep 20 21:25:59 pack3754: one would be to add a regex to PREMIRRORS Sep 20 21:26:29 the other would be to set SRC_URI_pn-linux within your local.conf but that's a bit messy Sep 20 21:26:37 particularly if you have patches :/ Sep 20 21:26:45 woglinde: I hope you're right about that Sep 20 21:26:51 i'd think you'd want to define KERNELORG_MIRROR Sep 20 21:26:57 which defaults to http://kernel.org/pub Sep 20 21:27:25 kergoth: ah yes that would probably be more useful :) Sep 20 21:27:29 :) Sep 20 21:27:37 kergoth hm not for git stuff Sep 20 21:27:37 MIRRORS isn't the most friendly thing around, sadly Sep 20 21:27:41 woglinde: true Sep 20 21:28:23 bluelightening & kergoth: thanks for the input! I'll try it out Sep 20 21:29:08 KERNELORG_MIRROR, ... should that be defined in local.conf? Sep 20 21:29:43 that would be fine, yes Sep 20 21:46:31 bye! Sep 20 21:52:50 cuuurious... the package OE was failing to find this morning (i think due to kernel.org being down...) was found! Sep 21 00:28:05 hey kergoth, what was that thing you found to build a series of git commits to help ensure bisectability? Sep 21 00:34:26 Tartarus: http://dustin.github.com/2010/03/28/git-test-sequence.html **** ENDING LOGGING AT Wed Sep 21 02:59:57 2011