**** BEGIN LOGGING AT Fri Aug 05 02:59:56 2011 Aug 05 07:35:42 hi all , I have seen that whenever i build opencv recipe standalone which will generate libopencv-core and so many libs but i have included opencv and opencv-dev packages to my recipe it is working but it is not installing any libs Aug 05 07:36:27 how can i install libopencv-core package with using recipe Aug 05 07:38:45 ? Aug 05 07:39:08 I thought koen fixed the deps Aug 05 07:50:00 i have included libopencv-core2.2 with IMAGE_INSTALL+ = in my recipe but it fails and complaining about package not found Aug 05 07:58:16 than it is called diffrent Aug 05 09:30:51 hi all Aug 05 09:32:52 hello there Aug 05 09:36:41 ant_work, got a patch which is required for pcmcia on collie. Where should I send it. To JaMa himself, or to some ML? Aug 05 09:37:20 I'd send to Jama now, as long as the kernels are in meta-zaurus Aug 05 09:37:30 I have a patch queue for him too Aug 05 09:38:09 then eventually someone will merge down to meta-handelds Aug 05 09:38:54 does CF card works fine now? Aug 05 09:41:52 yep Aug 05 09:42:09 great Aug 05 09:42:40 Now I see systemd errors from zaurusd and some other stuff :) Aug 05 09:43:24 Missing zaurus/mach-collie-vars :( Aug 05 09:45:02 someone says on the ML the oe-core rootfs created by Angstrom has wrong permissions, because of pseudo misbehavior Aug 05 09:45:55 I'm still working in earlu userspace atm, there I have wrong locomo keymap Aug 05 09:46:13 on poodle Aug 05 09:46:18 collie should be the same Aug 05 09:46:24 ant_work, I saw that behaviour before. Now it seems fixed. Aug 05 09:46:30 ah, great Aug 05 09:46:38 linux 2.6.39.3 or 3.x ? Aug 05 09:47:06 i mean in klibc userspace Aug 05 09:47:27 i.e. kexecboot keys Aug 05 09:48:21 ant_work, I meant rootfs permissions. Aug 05 09:48:27 ah, ok Aug 05 09:48:34 Don't have keyboard emulation yet. Aug 05 09:48:55 neither but kernel 2.6.2x had different keymaps Aug 05 09:50:03 maybe was a kernel patch Aug 05 09:51:38 ant_work, That's strange. I see no relevant patches. Aug 05 09:51:46 in mainline I mean Aug 05 10:55:11 hi ant_work, all Aug 05 11:21:08 hi pb_ Aug 05 11:21:17 hi pb Aug 05 11:31:00 lumag: is your patch for OE or for kernel? Aug 05 11:31:12 if it is kernel patch better to CC it to zaurus-devel@ ML Aug 05 11:31:18 for kernel in OE. not yet ready for mainline. Aug 05 11:31:21 Jay7, ok Aug 05 12:26:00 hi jinmasters Aug 05 12:26:03 ups jon Aug 05 13:36:40 florian: denix Crofton: 25 minutes until the meeting Aug 05 13:37:58 * florian is here :) Aug 05 13:38:01 hi all Aug 05 13:42:34 thanks Aug 05 13:42:39 hmm Aug 05 13:42:44 my calendar has noon :) Aug 05 13:42:54 denix, you awake Aug 05 13:58:23 ##oe-ev-board Aug 05 13:58:32 3 minutes :) Aug 05 14:05:00 florian: can we reschedule to 18? apparantly at least crofton got the UTC->CEST mixed up Aug 05 14:05:13 so denix might only appear in two hours as well Aug 05 14:05:40 yes that's fine for me Aug 05 14:05:52 ok, cool. then lets do that Aug 05 14:05:57 Crofton: ok with that? Aug 05 14:07:39 yes Aug 05 14:07:43 I'll be around Aug 05 15:02:05 Is mirrors.openembedded.org temporary offline or gone forever? Aug 05 15:06:19 * Tartarus kicks linuxtogo, didn't take my change bitbake-devel from digest to not change, yet Aug 05 15:08:01 You do usually get one final digest when you switch from digested to non-digested. If you think ltg is really stuck, let me know and I'll have a look. Aug 05 15:09:20 k, thanks Aug 05 15:10:36 hi mickeyl Aug 05 15:10:42 hi crfton Aug 05 15:10:48 hi tartarus Aug 05 15:11:07 hi woglinde Aug 05 16:15:47 dangit, getting bit by the perl-native/libxml-parser-perl-native/intltool-native thing again Aug 05 16:15:50 hrmph Aug 05 16:49:08 03Joshua Lock  07master * r3d66d4e64a 10bitbake.git/lib/bb/ui/hob.py: Aug 05 16:49:08 bb/ui/hob: reset event handler once build complete Aug 05 16:49:08 The HobEventHandler passes events on to RunningBuild when a build is in Aug 05 16:49:08 progress. This patch resets the HobEventHandler to handle events when the Aug 05 16:49:08 build is complete. Aug 05 16:49:09 Signed-off-by: Joshua Lock Aug 05 16:49:19 03Joshua Lock  07master * r5cacdc4f16 10bitbake.git/lib/bb/cooker.py: (log message trimmed) Aug 05 16:49:19 bb/cooker: mark parsed conf files as potential cache invalidators Aug 05 16:49:19 The include parameter to _parse() tells the parser whether to include the Aug 05 16:49:19 file in __base_depends, the contents of which have their mtime tested to Aug 05 16:49:20 invalidate the cache when they have changed. Aug 05 16:49:20 As I understand it we would want all of the configuration files that are Aug 05 16:49:21 currently handled with the _parse() method to invalidate the cache to set the Aug 05 17:14:17 tartarus ping? Aug 05 17:15:47 khem, ping? Aug 05 17:31:23 ka6sox: pong Aug 05 17:32:59 hi Tartarus when you are using your buildpower to regress builds, how much disk space and how much RAM were you using for those builds? Aug 05 17:33:18 Lets see Aug 05 17:33:29 At the time, 4 CPUs, 4GB memory, 50GB disk, using rm_work Aug 05 17:33:47 We got a lot done because there were ~30-40 I think going at a time Aug 05 17:34:49 so 4CPUs 24GB of RAM + 40GB of tmpfs would be enough? Aug 05 17:35:57 and was DL_DIR in that figure or outside? Aug 05 17:42:10 ka6sox: yes Aug 05 17:43:17 Tartarus: Do u have many machines doing it in parallel ? Aug 05 17:52:09 ka6sox: back Aug 05 17:52:17 DL_DIR was w/in that 50GB, fetching from a local mirror Aug 05 17:52:42 If I was to spec out a VM, it would depend on backend storage Aug 05 17:52:49 and you were regressing mimimal for a particular arch? Aug 05 17:52:51 This setup as super fast backend storage (a BlueArc) Aug 05 17:53:07 We were just building dang near everything, for a lot of machiens Aug 05 17:53:11 and seeing if something blew up, or not Aug 05 17:53:28 If you're looking for a single machine builder instead, I woudl go differently Aug 05 17:53:41 Since in this setup we had 30+ machines doing a different combo at the same time Aug 05 17:54:20 For a single, I'm partial to as many CPUs as you can shove at it and enough memory for that + SSD Aug 05 17:54:20 ji ka6sox Aug 05 17:54:33 ka6sox do you still own a nook? Aug 05 17:55:02 woglinde, yup...2 of em. Aug 05 17:56:58 ka6sox -> https://github.com/Kernelhacker/2.6.38.3-Adam Aug 05 17:57:12 I will include it in my tegra layer Aug 05 18:06:37 ka6sox: is opal setup Aug 05 18:07:24 ka6sox: or do u have new buildserver from LF Aug 05 18:08:32 I think I can turn opal over to you sometime today Aug 05 18:08:59 new build servers are coming, working with yocto to make sure we arent' buying the wrong thing. Aug 05 18:09:20 ok Aug 05 18:09:32 have you spec'ed them out ? Aug 05 18:40:25 khem, thats what I'm working on now. Aug 05 18:40:35 working with LF and Yocto Aug 05 19:36:31 hi Aug 05 20:34:14 I am having a tough time applying a debian patch to a recipe for netcat, it says it succeeded, but the files in the work dir show otherwise Aug 05 21:17:23 I'm trying to implement my own do_patch method. Is there a variable I can use that points to the patch filenames? Aug 05 21:20:42 mikec: do_unpack will copy all the patches to workdir Aug 05 21:20:49 then you are on your own after that Aug 05 21:43:16 khem: I finally sent out the klibc for meta-oe. Pls have a look, I can do corrections on weekend Aug 05 21:48:13 ant__: will do Aug 05 21:50:26 should I send some klibc-recipes to test-compile ? Aug 05 21:51:04 send to ML I mean (I'm thinking about kexec-tools-klibc) Aug 05 21:53:45 ant__: yes it would be nice if there was an image that could use it Aug 05 22:00:48 included are klibc-static-utils Aug 05 22:01:34 the images and the couple of klibc static binaries live in meta-zaurus atm Aug 05 22:02:06 you have the layer through Angstrom scripts Aug 05 22:02:36 (beware the layer preference/masking issue :/) Aug 05 22:10:11 thanks khem. As I also had to implement my own do_unpack method, I don't get the patch files in the workdir. Is there a variable that points at the recipy/files directory that I can use to copy the patch myself? Aug 05 22:11:15 yes ${S} Aug 05 22:15:06 mmm. doesn't ${S} point at the directory where the sources are copied? As I had to implement my own do_unpack, the patch files aren't copied to ${S}. So I'd like to adapt my do_unpack, but I don't find a way to point at the folder that contains my patch... Aug 05 22:19:08 no thats ${B} Aug 05 22:22:43 khem: btw you have klibc in meta-nslu2. You could test that Aug 05 22:29:10 ok thanks. got it now. **** ENDING LOGGING AT Sat Aug 06 02:59:57 2011