**** BEGIN LOGGING AT Mon Jan 14 02:59:59 2013 Jan 14 08:12:05 good morning Jan 14 09:13:28 morning all Jan 14 09:16:03 morning bluelightning Jan 14 09:16:14 hi rburton Jan 14 09:16:59 bluelightning: welcome back with snow Jan 14 09:17:18 rburton: heh Jan 14 09:17:30 thankfully at the moment it's just snow and not snow + ice Jan 14 10:40:05 hello evry one Jan 14 10:41:11 i'v tryed to generate a dev-image using IMAGE_FEATURES+="dev-pkgs" however this is not working i'v checked with many packages not being automatically installed, Jan 14 10:41:16 is this a yocto bug ? Jan 14 10:41:55 b737800: what version of yocto? Jan 14 10:42:01 b737800: there was a bug exactly like that, fixed in danny Jan 14 10:42:08 version 8.0 danny Jan 14 10:42:18 hm, file a bug with exact replication steps then Jan 14 10:42:32 ok i will do that, Jan 14 10:42:37 and the lists of what you get installed/what should be installed Jan 14 10:43:36 However in my project i plan to write extra glu code (in a bbclass) that do that i.e 1- list all packages that will be installed 2.check if a dev pkg exist for each package 3.install it in the dev image any hint on doring that ? Jan 14 12:03:07 I am having a hard time to find a problem with git fetcher Jan 14 12:03:50 it fails when I try to fetch inside bitbake; it works when I fetch in same machine from my shell ... Jan 14 12:04:15 so it seems something related to OE but I don't know what. Do someone has any clue? Jan 14 12:11:06 otavio: not related to anything in your ~/.gitconfig by any chance? Jan 14 12:15:19 otavio, what is the error? Jan 14 12:15:52 bluelightning: no; it just work when clonning locally Jan 14 12:15:59 bluelightning: but fails when in bitbake Jan 14 12:16:11 agherzan: error -3 Jan 14 12:16:14 hold Jan 14 12:16:34 otavio, what recipe? Jan 14 12:16:51 http://paste.debian.net/224536/ Jan 14 12:17:10 (-3 is another bug sorry, not the git one) Jan 14 12:17:19 it's the same as the one from rpi Jan 14 12:17:29 try cloning it on your machine with --quite Jan 14 12:17:46 agherzan: with --quiet? Jan 14 12:18:04 --quiet Jan 14 12:18:12 typo* Jan 14 12:18:41 doing Jan 14 12:19:05 agherzan: the crazy thing it works when I run it byhand Jan 14 12:19:23 otavio, it works with or without that flag? Jan 14 12:19:38 agherzan: I have open an issue ticket in github and they said it worked for them; so I started to try it locally and it worked Jan 14 12:19:44 (running with --quiet) Jan 14 12:20:15 hmmm.... strange. I have the same issue in rpi. They fixed it once (after a ticket) and now it's here again. Jan 14 12:20:40 let me check if i can clone it locally with and without --quiet (brb) Jan 14 12:23:04 agherzan: I am cloning with --quiet now Jan 14 12:26:56 agherzan: oh; it fails with --quiet Jan 14 12:28:20 agherzan: I am testing it once again to check Jan 14 12:30:32 otavio, well in this case we have the same issues which github fixed in the past Jan 14 12:31:43 they just mailed me (github). Jan 14 12:33:03 otavio, "Does it work with that parameter if you use https instead of the git protocol?" lol Jan 14 12:33:49 lol Jan 14 12:34:32 the idea goes like this - bitbake adds --quiet - and with this clone fails Jan 14 12:35:18 this is reproducible (at least for me) on host as well. so it's not from bitbake but from github Jan 14 12:36:32 agherzan: I agree Jan 14 12:39:25 I mailed them again Jan 14 12:41:28 otavio, very good - will spam them again too :) Jan 14 12:42:02 it's a common issue at large github repos - they should fix it once forever Jan 14 12:54:57 why is it so "lol" to ask about https:// access instead of git:// ? sure that's likely to go a rather different path inside github Jan 14 13:50:48 i'm trying to figure out why my useradd recipe fails. though calling "bitbake -D -vvv", the actual cmd being executed does not show up in the log, all i see is "ERROR: tried running groupadd command 10 times without success, giving up" Jan 14 13:55:26 schramae: have you looked in the log file for the failed task? Jan 14 13:56:57 bluelightning: yes, log.do_rootfs only says ERROR Jan 14 13:58:29 when looking at the /etc/group file in the sysroot, the groups i specified within the recipe are sucessfully created though Jan 14 14:00:14 B4gder, i wasn't looking for a workaround but for a fix. that's why "lol" Jan 14 14:01:32 agherzan: in my eyes it could've been a question to also help them narrow down the cause of the problem in their end Jan 14 14:01:46 or just understand it more Jan 14 14:02:47 B4gder, i know - but sounded funny - or as a workaround. You're right anyway Jan 14 15:24:03 is there a listing of kernel features somewhere? Jan 14 15:28:21 Crofton|work. it's on the 1.4 TODO list, I have a script that dumps them and then we can put them somewhere useful. (assuming you are talking about the linux-yocto style in tree features). Jan 14 15:28:56 yeah Jan 14 15:29:11 I am looking at adapating my custom zynq recipe to use them Jan 14 15:32:58 if you drop me an email with a few details, I can fire you off some ways to do it while waiting for the 1.4 features (if you need something ASAP). details would just be source kernel tree, if you want to re-use the meta-data from the existing released kernels, that sort of thing. I'm about to be hauled off for a few minutes so won't be at my keyboard. Jan 14 15:33:12 ok Jan 14 15:33:34 basically, I want to use the current xilinx repo, and use kernel features to config it Jan 14 15:33:38 let me thrash some more Jan 14 15:33:50 their git repo is killing qgit :( Jan 14 16:42:12 i'm desperately trying to get xtscal to work, but it always comes up with "XCALIBRATE extension missing: Resource temporarily unavailable". i added "--enable-xcalibrate --enable-tslib" to xserver-xorg and enabled and installed calibrateproto, libxcalibrate, tslib* like sb. suggested on the oe-mailinglist, but that does not help so far Jan 14 16:59:37 guys, im seeing errors like this: https://gist.github.com/4513124 only on centos 5.x Jan 14 16:59:46 it seems weird to be only on this older distro Jan 14 16:59:50 this is on danny btw Jan 14 17:13:27 Hello walters Jan 14 17:14:27 msm: looking Jan 14 17:15:07 walters: I am trying an upgrade to polkit 0.110, but there is a hard dependency on mozjs185, and in poky there is only mozjs180 Jan 14 17:15:25 walters: do you have any insights into this? Jan 14 17:15:46 sgw1: im doing more tests now too Jan 14 17:15:52 sgw1: it seems related to sstate Jan 14 17:16:03 sgw1: as in the first time it works, deplying from sstate it fails Jan 14 17:16:09 im double checking that statement now Jan 14 17:16:23 msm: ok, I am just looking at the recipe depends now Jan 14 17:17:04 sgw1: the depends seem ok Jan 14 17:17:28 msm, looks like it's not getting the lttng2 version Jan 14 17:21:21 msm: must be a runtime dependency, I am not even finding a liblttng-ust-ctl package on the autobuilder Jan 14 17:21:51 yea, rpm is being nice and finding it's missing Jan 14 17:21:58 but it's odd that it's only on my centos 5.x box Jan 14 17:22:10 what does your lttng-tools pkgdata look look like? Jan 14 17:23:00 msm: and the first build was OK? Is that package being generated in the tmp/deploy/rpm's dir? Jan 14 17:23:21 second Jan 14 17:23:33 you want want lttng2-ust pkgdata… or? Jan 14 17:23:48 msm: lttng-tools Jan 14 17:24:59 sgw1: https://gist.github.com/4531678 Jan 14 17:25:06 what do you want to see in it? Jan 14 17:25:38 sgw1: im rebuilding now with a fresh lttng2-ust (cleanall'ed) and it's going slow as lots of builds are running Jan 14 17:26:00 msm: how about the runtime/lttng-tools contents, also was a RPM package generated for liblttng-ust-ctl? Jan 14 17:28:53 sgw1: i've got lttng-ust and lttng-ust2 Jan 14 17:28:54 https://gist.github.com/4531711 Jan 14 17:29:08 2 has the correct library Jan 14 17:29:16 although this is a "fresh" non-sstate cache build Jan 14 17:29:21 but.. should not really matter Jan 14 17:29:44 sgw1: here is the pkgdata Jan 14 17:29:45 https://gist.github.com/4531717 Jan 14 17:37:24 msm: lttng2-ust should be providing the liblttng-ust-ctl library that lttng-tools is FILERDEPENDS on based on line 18 of the pkgdata file you posted. Somehow it's getting translated to a package on it's own, can you confirm that lttng2-ust has the right FILESRPROVIDES in it's pkgdata/runtime/lttng2-ust. Jan 14 17:39:20 sgw1: sec my machine is running slow… it's running 4 big builds atm Jan 14 17:40:19 https://gist.github.com/4531795 Jan 14 17:41:11 sgw1: it looks correct... Jan 14 17:41:44 sgw1: let me get some of these builds finished Jan 14 17:41:52 it could have been corrupt sstate-cache on a mirror Jan 14 17:42:16 msm: OK, all this looks good, so something else is going on, might need fray to chime in on RPM solver stuff Jan 14 17:42:31 ok, does that use any host rpm stuff? Jan 14 17:42:40 or is it all built -native? Jan 14 17:46:22 msm: should be the -native versions Jan 14 17:47:40 k Jan 14 19:21:33 is INCOMPATIBLE_LICENSE = "GPLv3" officially supported by yocto? Jan 14 19:23:58 an Jan 14 19:24:25 kergoth: I think so, there is something in these lines Jan 14 19:24:29 butnot sure Jan 14 19:24:55 i'm pretty sure the intent is that it should work, but i'm wondering about the offiical stance Jan 14 19:36:58 kergoth: i think it is semi official Jan 14 19:37:14 they do non-GPLv3 builds on the autobuilder normally, don't they? Jan 14 19:38:24 good question, didn't think to check that Jan 14 19:38:35 will check into that Jan 14 19:38:39 kergoth: http://autobuilder.yoctoproject.org:8010/builders/nightly-non-gpl3 Jan 14 19:47:59 thanks Jan 14 19:58:32 hello everybody Jan 14 20:01:51 could someone give me a hint about the following situation. I have a core-image-base built on pandaboard. Trying to bring the wifi up so I've got the drivers activated and compiled in to the linux-omap4-3.4 kernel. Also placed the necessary ti-connectiviy firmware to the /lib/firmware dir. so far so good. For some bloody reason, the driver is complaining about not finding the firmware. My last hope is this IRC so please, if someone could Jan 14 20:03:44 my idea is that the udev do not have a firmware rule or it is simply not running.. or the version of the udev falling into one of the well known disastrous release that the community were complaining about Jan 14 20:04:42 long story short, the wl12xx.ko and wlcore_sdio.ko drivers can not reach the filesystem for some reason. Jan 14 21:31:32 hmm, xz's LICENSE isn't quite right. while yes, there's a gplv3 file in the source tree, it's an m4 macro that we don't package.. afaict the packages we emit are GPLv2 and PD Jan 14 21:33:45 also, license.bbclass's handling of per-package licenses is wrong.. Jan 14 21:33:49 * kergoth makes a todo Jan 14 21:34:01 license = d.getVar("LICENSE_%s-%s" % (pn, package), True) if package else None Jan 14 21:34:14 LICENSE_ should work, not LICENSE_- Jan 14 21:34:39 and if we fix that, turns out there are a number of invalid license strings set for individual binary packages.. Jan 14 21:34:43 * kergoth mutters Jan 14 21:40:39 er, think there's some incorrect logic in the dont_skip_recipe handling in base.bbclass, too Jan 14 23:01:10 i have a recipe that has some libraries Jan 14 23:02:39 that installs a .so->.so.1->so.1.4 for instance Jan 14 23:03:06 how do I install a .so in -dev package? Jan 14 23:03:16 that's default. i don't understand the question Jan 14 23:03:17 because in this case the .so will be a link Jan 14 23:03:26 .so symlinks always go into -dev packages by default Jan 14 23:03:42 but its a link to nothing, right? Jan 14 23:03:52 no, the -dev package depends on the main package Jan 14 23:04:01 use the package manager properly and you'll be fine Jan 14 23:04:02 oh, ok Jan 14 23:04:18 ok, then its fine Jan 15 00:42:41 Hi , I got a message "WARNING: There were 3 hardware options requested that do not have a corresponding value present in the final ".config" file. This probably means you aren't getting the config you wanted. The full list can be found in your kernel src dir at: meta/cfg/standard/jasperforest/mismatch.cfg" Jan 15 00:43:02 what I did is "bitbake -c menuconfig linux-yocto" Jan 15 00:43:34 and change the ethernet driver "Intel 82576" from "M" to "*" Jan 15 00:46:34 should I add any recipe if want to modify kernel module to internal driver? Jan 15 02:07:46 guys, in my sysroot I have a library that has .so->.so.1->.so.1.2 for instance Jan 15 02:08:00 and another library is been compiled and linked against that first library Jan 15 02:08:28 but its linking against the .so instead of the .so.1 Jan 15 02:08:41 so its braking un runtime Jan 15 02:31:09 that sounds particularly unlikely, but if that's actually the case, then the library was linked with the wrong soname. make sure it's being linked correctly **** ENDING LOGGING AT Tue Jan 15 02:59:58 2013