**** BEGIN LOGGING AT Mon Feb 25 02:59:57 2013 Feb 25 09:21:48 good morning Feb 25 09:21:57 hi mckoan Feb 25 09:22:34 morning mckoan, woglinde_, all Feb 25 11:38:24 Hi. Feb 25 11:38:57 I am desperately looking for some advice. Feb 25 11:39:24 pompomJuice: your always better just asking your question than asking to ask! Feb 25 11:39:46 Ok here we go. Feb 25 11:39:48 http://arago-project.org/git/projects/?p=boot-monitor.git;a=summary Feb 25 11:40:05 that git repo is failing when bitbake tries to do a git ls-remote Feb 25 11:40:19 pompomJuice: known problem, it's in the process of being solved Feb 25 11:40:24 aah Feb 25 11:40:27 ok thanks :D Feb 25 11:40:39 pompomJuice: issue with arago/meta-ti git servers most likely Feb 25 11:40:40 where can I go to see such things? Feb 25 11:40:47 meta-ti mailing list Feb 25 11:40:50 aah ok Feb 25 11:41:00 thanks so much for helping. Feb 25 11:41:17 I think the original convo started in the bitbake-devel list and then migrated over to the meta-ti list Feb 25 11:41:20 someone who has built meta-ti recently can probably furnish you with the tarball required! Feb 25 11:41:44 that would help Feb 25 11:41:47 XorA: meta-ti uses git tags, so it has to talk to the sever every time Feb 25 11:41:53 since it has detected the version change Feb 25 11:42:01 jackmitchell: I thought we had a variable to disable that these days? Feb 25 11:42:31 XorA: possibly, I haven't really looked into it, just been watching the discussion Feb 25 11:43:20 with the tarball could always setup on githuib with a simple push and change the URI ;-) Feb 25 11:44:10 that would work Feb 25 11:44:14 or even an old tarball Feb 25 13:41:44 Hi I add https://github.com/meta-qt5/meta-qt5.git and I try to build for raspeberrypi but build crash... What did I do wrong? Feb 25 13:54:17 nothing Feb 25 13:54:33 meta-qt5 is not stable yet Feb 25 13:56:41 need a new FAQ, if it only came out 5 minutes ago, its probably not working yet! Feb 25 13:58:39 woglinde: ok, thanks. Feb 25 14:02:54 jama wants to fix it this week Feb 25 14:06:13 manuxx81: but please describe "build crash" Feb 25 14:06:25 manuxx81: log on pastebin would be nice Feb 25 14:06:46 Jama|off: ok Feb 25 16:25:50 moin Feb 25 17:26:04 Question about the default udev scripts, are they enough to have udev populate /dev, or do I have to add more rules? I'm building a first-stage initramfs with OE and when I run udev on it, it loads modules but doesn't make device nodes. Feb 25 17:28:25 mewyn: recent udev does not create /dev nodes anymore; kernel is responsible for this now Feb 25 17:28:32 you should mount devtmpfs Feb 25 17:36:39 Hm Feb 25 17:36:41 Okay Feb 25 17:36:50 Now let's put it back in the kernel Feb 25 17:37:22 I'll have to check my kernel config. I thought I mounted devtmpfs on /dev, but it never showed up in mounts Feb 25 17:52:03 better check under Device Drivers -> Generic Driver Options and make sure... Feb 25 17:56:07 *nod* I remember where the option was, I didn't read fully on this one (it's new since I've last built a kernel), and thought it was just a version of tmpfs that's designed for /dev Feb 25 18:01:14 being a gentoo dev, i'm pretty damn familair with kernel config too, but sometimes these embedded kernels get a little whacked Feb 25 18:01:40 i managed to break the efikamx kernel build pretty quick Feb 25 18:01:55 their patches seem a little wonky Feb 25 18:02:32 i already harrased the upstream guy (since he is also a gentoo dev) Feb 25 18:13:38 Hm, okay, seems I must have fat-fingered the first try. Another question, is there a better way than udev for loading all applicable modules? Feb 25 18:26:11 busybox Feb 25 18:26:27 s/better/another/ Feb 25 18:27:23 The kicker is I don't know what modules I need to load, did busybox add something new to handle this (I read mdev still doesn't) Feb 25 18:28:44 not that i know of, so probably not... Feb 25 18:28:50 okay Feb 25 18:28:54 does anybody else have problems with meta-oe's libvpx recipe? Feb 25 18:29:03 "Libdir /usr/lib must be a subdirectory of /usr/local" Feb 25 18:29:06 I could probably write a script for it, but for now that's too much trouble. Feb 25 18:29:19 I though it was supposed to be fixed with this patch: http://www.mail-archive.com/openembedded-devel@lists.openembedded.org/msg10834.html but it still appears Feb 25 18:29:22 I'll just continue starting udev, triggering, killing udev Feb 25 18:31:35 i only use a custom initramfs on gentoo boxes with separate /usr mounts but normally you would specify which modules you want or default to all Feb 25 18:32:00 one would think this would be similar... Feb 25 18:32:57 not sure what the oe initramfs tools are like, so that's probably not very helpful... Feb 25 18:34:52 *shrug* for the most part I'm just having OE spit out the cpio.gz for the first phase booter Feb 25 18:36:31 mewyn: devtmpfs is not mounted automatically for initramfs Feb 25 18:36:47 dv_: are you saying you confirmed the patch was applied in your build? Feb 25 18:36:49 I'm making a linux build for arcade machines (hobbyist, not commercial; typically running MAME or other emulators) Feb 25 18:36:59 yes Feb 25 18:37:26 ensc|w: *nod* Yeah, I had a mount line that didn't work on an early build and ended up taking it out. Now I put it back and it's fine. Feb 25 18:38:48 dv_: then i punt Feb 25 18:46:43 And the reason I'm looking to load applicable modules in the initramfs is I want to include all storage drivers in it and load the ones I need, then find my / Feb 25 18:49:27 sounds fine, in principle... Feb 25 18:49:32 nerdboy: ? Feb 25 18:50:06 punt to kergoth or... Feb 25 18:50:24 ah Feb 25 18:51:16 so I have a nativesdk package for python-cheetah Feb 25 18:52:03 and it fails to run because it needs the cgi module. What is the correct way to tell cheetah it depends on another nativesdk python package? Feb 25 18:53:55 I probably have this issue for the non sdk version as well Feb 25 19:07:07 python rdepends are madness Feb 25 19:11:41 i remember back in the oe-classic days seeing numpy as a dependency, which in turn pulled in BLAS and lapack etc. Feb 25 19:11:47 ... for a beagleboard. Feb 25 19:12:50 numpy will do that Feb 25 19:13:03 because that is the point of numpy Feb 25 19:13:09 oe-core can't change that Feb 25 19:13:45 the python module stuff looks like it gets done by hand, using trial and error Feb 25 19:24:42 Crofton: mickeyl does it as a script then uploads the result Feb 25 19:25:20 hmmm Feb 25 19:25:43 but the python recipes that are not using the manifest stuff are done by hand? **** ENDING LOGGING AT Tue Feb 26 02:59:58 2013