**** BEGIN LOGGING AT Fri Jun 13 02:59:59 2014 Jun 13 07:46:25 good morning Jun 13 08:11:30 hi, I'm trying to add an extra init script in a bbappend for a package that already inherits update-rc.d, but I don't seem to get it to work Jun 13 08:12:39 I have set INITSCRIPT_PACKAGES += "${PN}-ifplugd" and defined INITSCRIPT_NAME_${PN}-ifplugd and INITSCRIPT_PARAMS_${PN}-ifplugd Jun 13 08:12:57 do I need to inherit update-rc.d again or is there some other thing to do ? Jun 13 08:34:35 morning all Jun 13 09:17:55 PulkoMandy: inheriting again certainly won't help Jun 13 09:18:21 yes, I guessed so, but I'm not sure what else I can try Jun 13 09:18:57 PulkoMandy: is the package also in PACKAGES ? Jun 13 09:19:16 PulkoMandy: does the package get generated at all (missing the initscript)? Jun 13 09:21:10 mh... no, I think not Jun 13 09:21:15 so I should probably add that Jun 13 09:22:39 PulkoMandy: we could add a sanity check for that, you could file an ehancement bug for it (or even send a patch) Jun 13 09:22:43 :) Jun 13 09:23:16 there is no magic bullet for usability but we can improve things incrementally... Jun 13 10:48:34 mh... I still don't get this working Jun 13 10:48:35 http://paste.debian.net/104774/ Jun 13 10:48:57 here is my bbapend file, I think I added the new package properly, but it is not getting built at all Jun 13 10:49:22 however the rc file is installed on my image, I think from the main busybox package Jun 13 10:49:32 but the update-rc is not happening Jun 13 13:32:33 i want to install both gst-0.10 a nd gst-1.0 on my target. to enable gst-1.0 I have added lines like the following to conf/local.conf Jun 13 13:32:33 CORE_IMAGE_EXTRA_INSTALL += " gstreamer1.0-plugins-good-meta" Jun 13 13:33:42 but wehen building the image i get message like Jun 13 13:33:49 WARNING: gst-plugins-good-alaw-0.10.31 was registered as shlib provider for libgstalaw.so, changing it to gstreamer1.0-plugins-good-alaw-1.2.4 because it was built later Jun 13 13:34:21 and on the target i only have gst-1.0 libs, even though gst-0.10 libs have been built Jun 13 13:34:59 so I wonder how to tell yocto to copy both libs to the target? Jun 13 13:35:42 tobiash: the warning can be ignored Jun 13 13:35:52 and presumably you didnt actually install the 0.10 plugins Jun 13 13:36:17 rburton: autocomplete... Jun 13 13:36:22 mh... is it not possible to use ssh:// in an SRC_URI to fetch a folder ? do_fetch is working well (I get the folder in the download/ dir) but then do_unpack seems to be confused and copy the whole download dir, including sources for other packages Jun 13 13:36:32 or maybe my URI must not end with a / ? Jun 13 13:36:54 PulkoMandy: that sounds like a bug to me Jun 13 13:37:28 rburton: some 0.10 libs git installed though (those that are only available for 0.10 but not for 1.0) Jun 13 13:37:31 PulkoMandy: it's likely that the ssh fetcher isn't widely used Jun 13 13:39:45 bluelightning: I am finding a ".debug" directory populated under /bin , by different packages. LSB standard prohibits any sub directory under /bin directory. I dont think there are any explicit packaging commands in those recipes to create this. Jun 13 13:40:01 So any way to prevent creation of debug rpms ? Jun 13 13:40:16 milan__: those are meant to go into the -dbg package, which you wouldn't install unless you want those files Jun 13 13:40:51 bluelightning: damnit adium! Jun 13 13:41:23 TobSnyder: they should parallel install fine Jun 13 13:45:20 bluelightning: ok, the problem was having a / after the final folder name in the URI, removing that works Jun 13 13:46:49 Hi all, I've got trouble where I've defined IMAGE_ROOTFS_EXTRA_SPACE but now bitbake seems to freeze up on do_bootimg, after verbosing it more last message I'm getting is File "rootfs.img" too big Jun 13 13:47:19 am I just being impatient or is there a bug and it is truely freezing? Jun 13 13:48:52 PulkoMandy: could you file a bug for that? it shouldn't make a difference Jun 13 13:49:46 stuartw_: that does sound like a bug; I wonder which tool is printing that Jun 13 13:50:11 stuartw_: you should be able to increase that variable's value without any adverse behaviour Jun 13 13:50:42 It's just the output I'm getting after killing the python processes then rebuilding with -vvv option Jun 13 13:52:25 it'll be from one of the dommands that do_bootimg is calling Jun 13 13:52:27 commands Jun 13 13:54:26 hmm okay, what do you reckon it's best to do? file a bug then? Jun 13 13:54:37 that would be a good first step yes Jun 13 13:56:24 I have a recipe with pulls from a GIT repository. There is a variable setup in the recipe with selects the GIT branch, i.e. BRANCH ?= "master". Can i override this variable in the local.conf like BRANCH_myrecipe = "newbranch" Jun 13 13:57:36 bluelightning: https://bugzilla.yoctoproject.org/show_bug.cgi?id=6448 Jun 13 13:57:38 Bug 6448: normal, Undecided, ---, richard.purdie, NEW , ssh fetcher problem with url ending in / Jun 13 13:58:40 PulkoMandy: which branch / version are you using? Jun 13 14:00:04 I'm using dora, not sure where to look for the bitbake version Jun 13 14:01:06 ok, I'll put 1.5.2 in the version field - thanks for filing it Jun 13 14:06:35 PulkoMandy: check BB_MIN_VERSION variable in the meta/conf/sanity.conf of the oecore layer Jun 13 14:08:00 well, technically the actual bitbake version is reported every time a build starts Jun 13 14:09:25 oh, didn't even see that Jun 13 14:09:26 BB_VERSION = "1.20.0" Jun 13 14:09:36 bluelightning: https://bugzilla.yoctoproject.org/show_bug.cgi?id=6449 Jun 13 14:09:38 Bug 6449: normal, Undecided, ---, richard.purdie, NEW , do_bootimg in bitbake locks up Jun 13 14:17:19 bluelightning: I am suspecting my packagegroup-core-basic (modified) is the cause behind introducing the debug packages. Would you be able to give some idea seeing at the changes:http://pastebin.com/jCwcN4Ev Jun 13 14:17:56 I did not see it before with packagegroup-core-full-cmdline Jun 13 14:18:32 if there was a dependency from a normal package on a -dbg package you would get a QA warning Jun 13 14:18:49 at least I think you would Jun 13 14:20:04 Ok...but basically what does this block of code does ( given link ) ? Jun 13 14:21:00 it's intended to create a runtime package migration path from the old recipe to the new one Jun 13 14:21:23 ok Jun 13 14:21:57 which is one reason why having both in the same build would not be recommended Jun 13 14:22:06 Are the qa warnings logged anywhere? Jun 13 14:22:34 hey there Jun 13 14:22:54 why would libtool require /bin/bash in the image? I'm unable to even trace the dependency Jun 13 14:23:53 something is shipping a script that starts / Jun 13 14:23:56 whoops Jun 13 14:24:01 starts #! /bin/bash Jun 13 14:24:20 so just grep for #! /bin/bash and look what is shipped? Jun 13 14:24:27 or any smarter way to do it? Jun 13 14:26:50 imx-test Jun 13 14:26:55 sorry, wrong window Jun 13 14:32:12 if you know what package has the dependency, then you can just grep the contents of that package Jun 13 14:32:24 then either see if you can convert the script to be posix sh not bash sh, or admit that you need to build bash Jun 13 14:33:53 rburton: great thanks, I would be fine with shipping bash if I knew how to set /bin/sh to ash Jun 13 14:34:06 or to busybox to be more specific Jun 13 14:34:07 jaskij: update-alternatives Jun 13 14:34:54 the class? Jun 13 14:38:25 rburton: when removing all those CORE_IMAGE_EXTRA_INSTALL += " gstreamer1.0".. lines from local.conf (and therefore going with default gstreamer-0.10) all required gst-0.10 libs are installed (installed for me means in the deploy../tar.gz file containing rootfs) Jun 13 14:43:52 ok wrong Jun 13 14:44:25 libgstudp.so was built but again not copied into tar.bz2 file containing RFS Jun 13 14:44:35 so some INSTALL is missing I guess Jun 13 14:51:33 I have a recipe with pulls from a GIT repository. There is a variable setup in the recipe with selects the GIT branch, i.e. BRANCH ?= "master". Can i override this variable in the local.conf like BRANCH_myrecipe = "newbranch". This may sound like a basic question but unforunately it is not working for me. Jun 13 14:51:52 milan__: yes, tmp/qa.log Jun 13 14:52:38 dguthrie: from outside the recipe it would need to be BRANCH_pn-myrecipe = "newbranch" Jun 13 14:54:55 Thanks bluelightning Jun 13 14:57:21 bluelightning: Thanks Jun 13 15:04:33 How can I see the contents of a specific package? I'm a complete noob and I can't really find anything useful in google. Jun 13 15:04:58 I changed my companies license file. md5sum doesn't match Jun 13 15:05:09 yet, no automatic build update Jun 13 15:43:31 jaskij: one way is to look in the packages-split subdirectory of the workdir for the recipe (tmp/work////packages-split/) Jun 13 15:44:21 volker-: can you be a bit more specific, i.e. how is that license file is associated with the build? Jun 13 15:52:26 bluelightning: I created a ticket Jun 13 15:52:47 like so often ;-) Jun 13 15:54:25 bluelightning: but to come back to your question 'LICENSE="mycompany"\nLIC_FILES_CHKSUM = "file://${COREBASE}/meta-/LICENSE.mycompany;md5=my-after-update-wrong-md5sum"' Jun 13 15:54:53 volker-: that's not the ideal way to use LIC_FILES_CHKSUM though Jun 13 15:55:32 LIC_FILES_CHKSUM is supposed to detect when changes occur in the upstream source licensing (usually when you do an upgrade) Jun 13 15:55:45 there are lots of instances like that in oe-core though :/ Jun 13 15:55:46 so the license file you point to should be part of the source that the recipe fetches Jun 13 15:56:04 rburton: yes, there are, and we should change that where possible Jun 13 15:56:04 yes, I took the style from yocto, so don't blame me about the style :p Jun 13 15:56:25 the instances in yocto are where the upstream source doesn't have a license of its own Jun 13 15:56:44 (not an explicit one anyway) Jun 13 15:56:59 volker-: I'm just trying to explain why what you think should happen does not happen Jun 13 15:57:14 ok, wrong style. Still would have expected that it checks the md5sum on every run to point out changes here Jun 13 15:59:13 volker-: it's checked at the end of do_configure Jun 13 15:59:28 I believe the reason is that in some cases the license that applies is selected as part of the configure process Jun 13 15:59:32 which does not seem to be executed if the license file changes but not the recipe Jun 13 16:00:01 no, but if you think about how LIC_FILES_CHKSUM is meant to be used, it's not meant to point to a file we would normally monitor Jun 13 16:01:23 Anyone else run into problems with gdb loading debugging symbols when using our seprated out debug info with image-prelink enabled? Jun 13 16:08:02 kergoth: I didn't have problems on ARM/edison. Jun 13 16:08:12 hmm, k Jun 13 16:33:39 kergoth: Am I ok to merge the meta-mentor git patches into master? Jun 13 16:35:50 You're referring to the two commits at https://github.com/MentorEmbedded/meta-mentor/commits/master/meta-mentor-staging/recipes-devtools/git/git_1.9.0.bbappend, I'm assuming? Jun 13 16:36:00 * kergoth needs to find some time to push more pending/staging stuff Jun 13 16:36:32 kergoth: yes Jun 13 16:36:49 afaik they should be good to go, we've had no issues with it anymore Jun 13 16:37:12 searching for yocto on monster.com only shows 6 results Jun 13 16:38:05 kergoth: cool, just trying to clear some of the bugzilla entries... Jun 13 16:38:13 ah, cool Jun 13 16:39:32 kergoth: also keeps annoying me in the autobuilder logs ;-) Jun 13 17:16:36 kergoth: yes, lots of problems for multi-threaded apps Jun 13 17:30:04 hmm: Operating Systems used in ATS: OpenEmbedded Linux (Yocto Project), Ubuntu, Windows (XP and 7), and Android (Gingerbread and JellyBean). Jun 13 17:30:33 Experience with GIT/Yocto Jun 13 17:31:46 >Strong development experience with modern Linux development environment and tools such as Yocto. Jun 13 17:31:55 > You will be helping assist with the transition of Embedded Linux SW stack to OpenEmbedded. Jun 13 17:32:33 WR needs a Yocto guy, fancy that Jun 13 17:32:43 heh Jun 13 17:33:03 Understanding of and experience with Yocto embedded distribution architecture Jun 13 17:33:03 Familiar with distrubution systems like open embedded or system builder Jun 13 17:33:39 :) Jun 13 17:34:03 also, searching for openembedded gets liek 2-3 additional positions Jun 13 17:34:19 rofl Jun 13 17:34:28 Experience defining/building and maintaining custom Linux configurations (Yacto, OpenEmbedded, Gentoo, OpenWRT). Jun 13 17:34:45 Crofton: only 6 hits for yocto was a little bit less then I expected Jun 13 17:34:50 but the intel page seems to have much more Jun 13 17:34:51 ;-) Jun 13 17:35:07 maybe you should search for Yacto Jun 13 17:35:52 only BAE systems comes up when searching for yacto ;-) Jun 13 17:36:04 we should study search keywords vs openwrt/buildroot etc Jun 13 17:36:06 rofl Jun 13 17:36:15 typical defence contractor Jun 13 17:36:26 I few months ago I had to do the technical interview of a ex-fed guy. Jun 13 17:36:46 Umbunto and such typos all over his 5 page technical expertice Jun 13 17:39:35 heh Jun 13 19:52:36 hello, I've got a question: how do you validate a build? Jun 13 20:40:18 where will be an appropriate place to but PREFERRED_VERSION_u-boot = "2013.07" inside my layer? Jun 13 20:44:16 For the moment the only place that I manage to do it was inside my machine conf ... But I was interested in keeping this change only on my layer? is it possible to append the machine/.conf file ? Jun 13 20:50:28 I’m equally curious about doing appends to existing machine files Jun 13 20:50:34 equivalent to a .bbappend, but for a .conf Jun 13 20:50:42 not sure if such a thing is possible Jun 13 20:53:39 mkeeter, my test were not successful :( Jun 13 20:53:56 alright Jun 13 20:54:04 I ended up duplicating the machine with a new name Jun 13 20:54:22 then adding a kernel recipe .bbappend that adds the appropriate KBRANCH, KMACHINE, etc Jun 13 20:54:43 you might be able to include the existing machine file in your new one, save some duplication Jun 13 21:02:14 mkeeter, that's what I thought but I ask before doing it since it require a bit more of work... If I recall reading on mailing list that if we need to change bootloader kernel version machine are perfect for that... Jun 13 21:02:19 mkeeter, thx Jun 13 21:10:22 which packages do I need to install to be able to compile kernel modules on a device? Jun 13 21:20:10 smurray, reading the ref man about the include/require in that case require will be better in my case. But the looking in the poky source I see require being used only with .inc file in machine folder I didn't see it with conf. Do you have any example to point to? Jun 13 21:21:02 it's used just like include. The difference is if file is not found one continues and one fails. Jun 13 21:21:59 blloyd, that's what I understand, thx Jun 13 21:22:24 yeah, that's the case Jun 13 22:39:50 one of your projects matchbox-window-manager, i am trying to fullscreen the chromium-browser with its --kiosk --start-fullscreen but it only "maximizes" it when using matchbox-window-manager, other WMs will truely "fullscreen" (f11) it, is there any flag or setting for this? Jun 14 00:29:25 Hello. Can someone point me to documentation regarding cleaning the images directory? I would like to remove the files, but the readme says "NO!" and isn't clear on the proper procedure. Jun 14 00:36:57 (I haven't been around that long but I think that's the documentation I've been charged with creating zman97211) Jun 14 00:42:16 LCyrin, can you help me? **** ENDING LOGGING AT Sat Jun 14 02:59:58 2014