**** BEGIN LOGGING AT Mon Nov 04 02:59:58 2013 Nov 04 08:40:10 morning all Nov 04 09:02:11 good morning Nov 04 09:36:57 hi, how to request a backport from master to dora? Nov 04 09:37:41 morning all Nov 04 09:38:04 morning Nov 04 09:38:21 hi JaMa Nov 04 09:39:45 beuh: just post a request on the mailing list; mark it as [dora] in the subject and CC the dora branch maintainer (Robert Yang ) Nov 04 09:41:59 bluelightning, ok thanks Nov 04 09:43:14 beuh: (assuming it's an OE-Core change you're referring to) Nov 04 09:44:00 bluelightning, it's a change in meta-openembedded Nov 04 09:44:36 beuh: ah ok, in that case send to the openembedded-devel mailing list instead, without CC Nov 04 09:44:45 ok Nov 04 09:45:01 sorry, should have asked before Nov 04 09:50:07 hi bluelightning, JaMa, all Nov 04 09:50:21 hi mckoan Nov 04 09:55:28 bluelightning: are layerindex inter-layer dependencies maintained from README files or only in web UI? Nov 04 09:55:43 JaMa: web UI Nov 04 09:56:02 they are per-branch as well, FWIW Nov 04 09:57:04 I'm trying to find how to change it in web UI Nov 04 09:57:41 edit layers allows me to change list of layer, but not layer-branch combinations Nov 04 10:04:04 JaMa: the assumption is that branches match up across layers Nov 04 10:04:21 JaMa: is that not the case here? Nov 04 10:05:10 it isn't, for meta-webos-ports/danny there wasn't meta-qt5 dependency at all Nov 04 10:05:26 dylan branch depends on older revision from meta-qt5/master Nov 04 10:05:39 dora and master are now compatible with latest meta-qt5/master Nov 04 10:08:47 updated README files pushed Nov 04 10:09:23 ok, I think the problem is it's trying to auto-create a danny branch record for meta-webos-ports and assuming the dependencies are the same as master Nov 04 10:09:46 I can solve that by creating a layer branch record manually without that dependency Nov 04 10:10:52 er... or not Nov 04 10:11:01 damned form doesn't let me edit the fields :/ Nov 04 10:11:07 it would be nice to be able to edit it from UI, because I think that different layer requirements are quite common for some project in different branches Nov 04 10:11:38 well, after the branch record exists you can definitely change those from the standard form Nov 04 10:11:40 e.g. now with separate meta-filesystems I need to edit multiple projects to include it since dora Nov 04 10:13:27 could be we need to look at an explicit "layer split" tool Nov 04 16:55:48 Are there any existing images/methods to build a distribution that runs entirely from RAM? I'm looking for something with similar behavior to a LiveCD or LiveUSB image. Nov 04 17:29:51 oneQubit: sure - if you include "live" in IMAGE_FSTYPES, it will build a LiveUSB image (.hddimg) Nov 04 17:49:56 moin Nov 04 17:50:31 bluelightning: erlang and gnat-gcc got in the way of nginx again... Nov 04 17:50:48 mr_science: np Nov 04 17:59:05 i'll try and get back to it tonight... Nov 04 18:00:34 it might also be partly that the erlang guy was a lot happier than upstream was with my patch... Nov 04 18:25:38 bluelightning, any major caveats to the "live" images running on an embedded device (available RAM permitting)? Nov 04 19:31:44 damn it, I am getting pull request for meta-sdr, before it is in meta-oe :) Nov 04 19:34:31 cbrake: Hmm, is your commit summary email script laying around somewhere, by chancE? Nov 04 20:05:21 hi all, i'm new in openembedded Nov 04 20:06:19 I would like to understand how it works and what is meant by meta layer Nov 04 20:06:47 i have a BeagleBoneBlack and i want to customize image for my board Nov 04 20:08:38 what are layer? Nov 04 20:11:11 and metadata? Nov 04 20:19:27 there is someone? Nov 04 20:20:20 metadata is the recipes and other informatoin in a layer Nov 04 20:22:42 recipes what is? Nov 04 20:23:35 mtkl_m: http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#understanding-and-creating-layers Nov 04 20:23:56 tlwoerner: fantastic Nov 04 20:24:50 http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#yocto-project-terms Nov 04 20:32:22 So there are this layer that are configuration file of my Image. After i write PATH of layer in bblayers.conf and with bitbake build image Nov 04 20:32:24 is right? Nov 04 20:34:41 kergoth: sure is: http://cgit.openembedded.org/openembedded-admin/ Nov 04 20:34:50 Crofton: thanks, noted Nov 04 20:34:54 cool, thanks Nov 04 20:36:09 is right? Nov 04 20:37:46 tlwoerner: Nov 04 20:37:46 ? Nov 04 20:43:03 but in my Beagle Board which Layer i can insert? Nov 04 21:59:00 anyone remember the patch to boot older arm kernels built with recent oe-core toolchain? Nov 04 22:00:49 crofton use gcc-4.6 Nov 04 22:01:01 grrr Nov 04 22:01:18 you might have to patch a lot stuff in the kernel Nov 04 22:01:24 anybody seen high CPU usage by connmand in the sato image? Nov 04 22:01:39 crofton but when you find it again let me know Nov 04 22:01:51 I have the same situation kernel 3.1 for tegra Nov 04 22:01:59 I have 3.5 on omap Nov 04 22:02:01 after a while, connmand suddenly needs a lot of CPU% . I have seen this on several machines. Nov 04 22:02:02 used to boot Nov 04 22:02:16 hm shouldnt 3.5 work? Nov 04 22:02:44 still fairly ancient Nov 04 22:04:48 dv_: sounds familiar when i built sato for rpi, but i didn't like conmand much anyway Nov 04 22:05:24 basic networking seems much more stable without conman Nov 04 22:05:28 Crofton: iirc mckoan sent a patch long ago and 2xiirc could even be committed Nov 04 22:05:30 perhaps its actually caused by the network icon in sato Nov 04 22:05:43 do you ahve a reference Nov 04 22:05:51 maybe...lemme see Nov 04 22:05:55 i tried conman in my non-sato image too Nov 04 22:06:12 pretty much the same wonky POS Nov 04 22:06:44 i didnt have stability issues so far. just this CPU spike. Nov 04 22:06:53 it refused to honor it's own static config and would "revert" to dhcp on its own some number of hours later Nov 04 22:07:18 completely ignores network/interfaces configs Nov 04 22:07:32 connman's purpose is to be a leaner networkmanager alternative for embedded, right? Nov 04 22:07:38 i gave it the boot from my xorg image Nov 04 22:07:49 dv_: in theory, yes Nov 04 22:08:14 and you know "in theory" is greek for "not really"... Nov 04 22:08:18 Crofton: http://patches.openembedded.org/patch/42279/ Nov 04 22:09:27 :) Nov 04 22:09:42 not that one Nov 04 22:09:51 I ahve a kernel that built but will not boot Nov 04 22:10:14 something gets miscompiled Nov 04 22:10:43 argh... then is gcc Nov 04 22:10:48 yeah Nov 04 22:10:54 but there is a way to fix he kernel Nov 04 22:11:09 dv_: i stuck with pulse for a while until it got better Nov 04 22:11:20 not sure i can do that with conman Nov 04 22:11:32 mr_science: and networkmanager is too big for embedded? Nov 04 22:11:39 networking being a bit more critical than audio... Nov 04 22:12:05 no, NetworkManager is in the gnome layer i think Nov 04 22:12:20 systemd being the other ingredient Nov 04 22:13:20 i'm still going for "lite weight" in my xorg image, so no conman, no systemd, no gnome Nov 04 22:14:06 the lightest alternative for a wifi gui would probably be wifi-radar Nov 04 22:15:03 not sure if there's a wicd recipe or not, but that would be my choice of the three (conman, wicd, NetworkManager) Nov 04 22:16:29 the memset stuff I think Nov 04 22:16:49 i did resurrect a wifi-radar recipe but it's hard-coded to an old dhcp client Nov 04 22:17:00 needs to be updated Nov 04 22:21:11 dv_: i would say "bloaty and unstable" rather than "too big" Nov 04 22:22:13 networkmanager itself can exist without gnome though Nov 04 22:22:17 I use it with KDE on my laptop Nov 04 22:22:59 true Nov 04 22:23:57 gnome is the bloatiest desktop in the OE world, and i think it's the only one that depends on NetworkManager Nov 04 22:25:24 i use it with my old gnome desktop, but not with any of my current xfce/other installs Nov 04 22:35:48 anarsoul: hi Nov 04 22:44:09 crofton https://github.com/torvalds/linux/commit/455bd4c430b0c0a361f38e8658a0d6cb469942b5 Nov 04 22:45:19 thanks Nov 04 22:45:21 I foiund it Nov 04 22:45:27 and another commit to fix that one Nov 04 22:45:31 testing now Nov 04 22:47:40 crofton which one? Nov 04 22:48:02 heh.. stubborn guy ;) Nov 04 22:48:51 418df63adac5684 Nov 04 22:49:28 *sigh* link Nov 04 22:49:31 *g* Nov 04 22:51:32 crofton do not let me down Nov 04 22:51:38 sorry, I was using a local repo to make patches Nov 04 22:52:55 https://github.com/torvalds/linux/commit/418df63adac56841ef6b0f1fcf435bc64d4ed177 Nov 04 22:54:00 crofton does it work now for you? Nov 04 22:54:39 still building Nov 04 22:54:47 well packaging now Nov 04 22:57:13 hrms how do I get out the patch now from github Nov 04 22:57:21 damn new interface Nov 04 22:58:54 woglinde, yep Nov 04 23:00:13 heh...nasty, yes Nov 04 23:04:53 funny all the pull request for kernel on github Nov 04 23:07:43 okay downloaded it from oe patchwork Nov 04 23:15:17 woglinde: view -> raw Nov 04 23:18:45 ant hm Nov 04 23:19:15 ant thats only the file Nov 04 23:19:18 not the patch Nov 04 23:21:41 okay good nite Nov 04 23:22:08 will testing the kernel tomorrow and maybe switch back to gcc-4.6 Nov 04 23:22:27 * woglinde wonders why the toradex guys did not pull in both patches Nov 04 23:25:16 might use an old gcc **** BEGIN LOGGING AT Tue Nov 05 00:58:43 2013 **** ENDING LOGGING AT Tue Nov 05 02:59:58 2013