**** BEGIN LOGGING AT Mon Jul 29 02:59:58 2013 Jul 29 06:09:38 morning #oe Jul 29 07:12:40 good morning Jul 29 07:33:45 hi Jul 29 07:34:22 i have a doubt in oe core picking up the bb files for module.. i have two meta-layers in whihc i have bb files with same name Jul 29 07:34:36 so which one will oe takes to build the module? Jul 29 07:35:06 can i specify anywhere to select the specified bb file? Jul 29 07:38:48 johny_: it will pick the one with the 'higher' version or the 'highest' priority. Jul 29 07:39:46 you can use a PREFERED_VERSION to indicate which one to use. Jul 29 07:39:54 ohh i'm not giving any versions as of now.. Jul 29 07:40:03 and u mean layer priority? Jul 29 07:40:29 you can use bitbake-layers show-recipes to list all recipes and their 'preferred' version. Jul 29 07:40:47 well, each recipe has a 'version' Jul 29 07:40:58 _.bb Jul 29 07:41:40 oops, there was a typo above, it's PREFERRED_VERSION Jul 29 07:41:47 yes but actually i'm downloding 2 layers from some source so one of the module is been provided by two layers.. and there is versions for that module.. Jul 29 07:41:51 you can grep for that string for examples in tree. Jul 29 07:42:00 so wha t can i do to take my cchoice of bb file Jul 29 07:42:29 i think it's best to use PREFERRED_VERSION_ then. Jul 29 07:42:47 if you give a layer a higher priority then it will impact all recipes from that layer. Jul 29 07:43:00 bb files name is just .bb in both the layers..!!! Jul 29 07:43:35 then the version must be set in the .bb file. Jul 29 07:43:49 oh sory , both the layers has same versions. Jul 29 07:43:53 it's taken automatically from the .bb file when there is one. Jul 29 07:43:59 they are including one more file in side bb Jul 29 07:44:12 versions are same in both Jul 29 07:44:58 hmm... are these 2 public layers ? can you tell more about them? Jul 29 07:45:04 but configurations and patch files are different Jul 29 07:45:11 no no.. these are not public layers Jul 29 07:45:19 created by some one Jul 29 07:45:24 ok. Jul 29 07:45:50 then i think you might get around with priority, but i have never used that too much. i don't know what other 'problems' you might get. Jul 29 07:46:08 maybe be better to wait for others to chime in. Jul 29 07:46:11 i want to use some of the modules form one layer and some from another layer but some common modules are there for both Jul 29 07:46:41 as far as i know, we wil give priority to bb layers.. it should take accordinly. Jul 29 07:46:49 but it is not happening so!! Jul 29 07:56:36 i have a doubt in oe core picking up the bb files for module.. i have two meta-layers in whihc i have bb files with same name and version, so which one will oe takes to build the module? Jul 29 07:56:56 is it based on PRIORITY of layers ? Jul 29 08:09:26 johny_: http://lists.linuxtogo.org/pipermail/openembedded-core/2011-August/007387.html Jul 29 08:21:05 morning all Jul 29 08:22:04 ant_work: johny_: FWIW, the bug described in that email has been fixed for quite a while now Jul 29 08:24:39 morning bluelightning, all Jul 29 08:26:21 oh ok Jul 29 08:26:47 so what comes first? priority of layer or latest version of bb? Jul 29 08:38:44 bluelightning: sweet, I can remove one more link from my archive Jul 29 08:40:02 bluelightning: just for fun, what happens with xserver-nodm-init now with default prefs? Which one is chosen if you hav emeta-oe? Jul 29 08:42:33 ant_work: the meta-oe one Jul 29 08:42:59 I see, thx Jul 29 08:43:13 ant_work: but that's about to be merged into OE-Core, and then we can send a patch to remove the meta-oe one Jul 29 08:43:20 (can't check from here) Jul 29 08:43:24 great news Jul 29 08:44:13 bluelightning: the xinput/pointercal is still in queue? Jul 29 08:44:45 ant_work: it's part of the same patch Jul 29 08:45:11 yumm Jul 29 08:47:32 * bluelightning just send dylan & bitbake 1.18 changes incl a bunch of changes from JaMa|Off Jul 29 08:57:41 morning all, bluelightning , ant_work Jul 29 09:03:36 It is possible made an append of an .inc file? Jul 29 09:10:43 silvio_: no, .inc and .classes can not be appended Jul 29 09:12:11 ant_work, thanks, if I what check that i use last revisio of a version of a certain version of .bb, for example gvfs-1.8.2.bb? Jul 29 09:14:53 I'm looking here http://layers.openembedded.org/layerindex/recipes/?q=gvfs, but I see last version , and I would last revision of gvfs-1.8.2.bb Jul 29 09:15:45 silvio_: ah, you mean the previous revision Jul 29 09:17:29 in this case you should just set the PREFERRED_VERSION-pn or mask the more recent packages Jul 29 09:18:23 ah, ofc you have to provide the recipe in some layer Jul 29 09:25:03 silvio_: typo...note that it is actually an underscore like in: PREFERRED_VERSION_u-boot = "2012.10" Jul 29 09:25:50 iirc PREFERRED_VERSION_pn-u-boot = "2012.10" if you put it in your local.conf instead Jul 29 09:29:05 ant_work, sorry I was not clear I had a trouble with gvfs-1.8.2-r7, and I found some patches online that to r9, but how I can sure that r9 is last revision of that version of recipe, and there is not an r10 revision or more? Jul 29 09:30:26 silvio_: the r number is just an incrementing revision number, it doesn't have any relevance outside of OE / distros derived from it Jul 29 09:31:17 it is only a nightmare for package management ;) Jul 29 09:32:36 silvio_: if I understand it, you have some extra patches in your layer so you want create a .bbappend and add those extra patches Jul 29 09:34:11 ant_work, yes , I want add this extra patches to gvfs-1.8.2-r7, but I m not sure that last patch that solve my trouble is r9 Jul 29 09:34:38 I try to loooking in git repo, but seam last is r8 ?!!? Jul 29 09:35:47 but i found a r9 with google so I cant understand with is the last patch made to gvfs-1.8.2 to make it finish compilation in right way Jul 29 09:36:21 http://lists.linuxtogo.org/pipermail/openembedded-devel/2012-December/043203.html Jul 29 09:37:00 do now i go in incremental steps... Jul 29 09:43:48 morning all Jul 29 09:48:15 hi pb_ Jul 29 09:49:41 silvio_: you re using danny it seems Jul 29 09:49:53 ant_work, yes Jul 29 09:53:44 silvio_: seems you need to backport some patches http://cgit.openembedded.org/meta-openembedded/log/meta-gnome/recipes-gnome/gvfs Jul 29 09:56:52 ant_work, seams I m becoming crazy... :) Jul 29 09:57:03 ant_work, seems* I m becoming crazy... :) Jul 29 10:01:28 hi bluelightning Jul 29 11:10:06 the tfkill thread is somewhat epic Jul 29 11:14:08 Crofton: not yet escalated like the kernel's one about DT bindings as ABI ;) Jul 29 11:15:46 :) Jul 29 12:31:57 ant_work, bluelightning , thanks for help I applier r8 patch and seems that all work like a charm Jul 29 12:32:35 silvio_: great :) Jul 29 12:33:38 Can I mix the sstate-cache directory for different build platforms? like fedora 17 and a debian-7.0? Jul 29 12:41:54 silviof: you can, but the native parts won't be shared by default (since as you can probably see, there is a subdirectory named with the value of NATIVELSBSTRING for each distro) Jul 29 12:42:36 silviof: the reason being, executables built against a newer glibc will break if run with an older version Jul 29 12:43:56 bluelightning: oh ok - thanks. I just wanted to make sure. Jul 29 12:53:13 otavio: I have seen you have done some work in meta-browser. You know Is there some work done for firefox22? Jul 29 12:53:31 silviof: it is not Jul 29 12:53:45 silviof: if you're willing to work on this, please do Jul 29 12:54:43 yeah - i work on it :-P (now) Jul 29 14:23:31 Hello! Jul 29 14:24:30 my questions might be off topic, but i could need a hint here. I have this Device: http://www.bandrich.com/p530_over.html Jul 29 14:24:58 it runs open embedded on it i think. It has a Web GUI and a telnet Login. Jul 29 14:25:41 Any idea how i can "root" the device? I want to pimp it a little bit and do some custom stuff on it... Jul 29 14:26:13 the Backup file seems to be some DOS Binary?! Jul 29 14:28:29 you're asking about backdoors in oe? or for some 0day? Jul 29 14:35:12 spiekey: OE is a build system, rather than a specific operating system, the manufacturer of that device could have made any number of custom changes Jul 29 14:35:56 spiekey: if it runs OE then it runs Linux and as such you're probably better looking for general Linux exploits Jul 29 14:36:39 *probably runs Linux, OE could in theory be used to build other operating systems Jul 29 14:39:25 jackmitchell: even if that theory is very far fetched :) Jul 29 14:41:13 jackmitchell, hm. okay. Last time i downloaded the Backup file, added my user in /etc/passwd /etc/secrets and uploaded the backup file and i was able to telnet into it. Then i was able to set up my openvpn tunnel. which was nice Jul 29 14:42:01 spiekey: thats is not specific to OE at all. Sounds like they wrote something on their own their for handling backups Jul 29 14:42:04 stefan_schmidt_w: indeed, I realised I had said something that could possibly be wrong; I didn't want the internet police coming to get me ;) Jul 29 14:42:26 stefan_schmidt: we use OE to build binaries (though admittedly not entire filesystems) for windows. Jul 29 14:42:32 okay, thanks a lot Jul 29 14:42:34 jackmitchell: I cannot hear them. Seems like they are out of black vans right now. Lucky you :) Jul 29 14:43:18 pb_: I buy that without doubt. I have yet to see OE build a fully system for something else than Linux Jul 29 14:43:48 pb_: Still a surprise to me that you do that :) Jul 29 14:44:05 pb_: mingw integrated as compiler or how? Jul 29 14:44:21 stefan_schmidt_w: right, exactly. Jul 29 14:45:01 pb_: cool, did not know people are doing that Jul 29 14:52:36 stefan_schmidt_w: you may be interested to hear that poky was used to build apps + dependencies for rooted iOS devices some years ago :) Jul 29 14:53:08 bluelightning: brr, now I have goose bumps. :) Jul 29 14:53:29 :D Jul 29 14:53:30 bluelightning: Still its binaries but not full images Jul 29 14:53:57 indeed... would be interesting to see someone build something else, but probably quite a lot of work Jul 29 14:54:04 Actually I remember these crazy folks at OHand did such things Jul 29 14:54:13 yeah, that was an OpenedHand project Jul 29 15:08:45 hi. how does bitbake know which PACKAGE_ARCH to consider for the image? My (A/V) firmwares are only SoC specific, not machine (or board) - so I built them using PACKAGE_ARCH = ${SOC_FAMILY} - but it seems I need to add this somewhere else as the image creation eventually fails complaining that it can't find the required packages Jul 29 15:15:05 andre_d: it uses everything in ${PACKAGE_ARCHS} Jul 29 15:21:05 pb_: thanks. So would I just add ${SOC_FAMILY} to PACKAGE_EXTRA_ARCHS_append? Jul 29 15:22:02 yeah, something like that. Jul 29 15:25:27 curious, I can't find anyone else doing that e.g. in meta-ti Jul 29 15:26:08 bluelightning: what do you mean? Jul 29 15:27:05 denix: I meant, I would have expected similar necessity in other layers such as meta-ti and yet it doesn't appear to be done there Jul 29 15:27:38 bluelightning: yeah, I meant to ask if you refer to SOC_FAMILY topic or something else Jul 29 15:28:09 denix: that's what I was referring to yes Jul 29 15:28:42 bluelightning: there's soc-family.inc in oe-core now and we include it Jul 29 15:29:17 erm does anyone know how to tun it off: "bootconsole [earlycon0] enabled" when linux is booting ? Jul 29 15:30:15 denix: sure, it's just that what andre_d is asking about seems semi-logical so it seemed a bit odd that we're not doing it automatically, but maybe there is a good reason for that Jul 29 15:31:05 adding ${SOC_FAMILY} to PACKAGE_EXTRA_ARCHS in my machine.conf worked nicely for me Jul 29 15:31:12 i don't think there are any firmware for a/v in meta-ti, yet. no? Jul 29 15:31:32 andre_d: adding means 'append'. right? Jul 29 15:31:40 yes Jul 29 15:32:07 well, bluelightning i agree that is SOC is not a 'core' feature, we might want to add SOC to PACKAGES_ARCHS. Jul 29 15:32:09 bluelightning: ah, I see. we don't do SOC on the package level on purpose Jul 29 15:32:36 not sure if ideally this should be added to the ipkg configuration on the target, so packages can be update, too. But it's not an issue for me right now Jul 29 15:33:27 doing it on a package level would require updating package manager (opkg) to match run-time experience Jul 29 15:34:03 hence we only use SOC_FAMILY for overrides at build time Jul 29 15:34:55 denix: but there is no firmware in meta-ti, iirc. Jul 29 15:34:57 right? Jul 29 15:35:01 yeah, similar to what andre_d said :) Jul 29 15:35:16 ndec: no a/v, correct. there's pm though Jul 29 15:35:36 ok, then the same would apply. Jul 29 16:02:34 bluelightning: hi. not sure if that was intentional, but you didn't signoff my patch that you merged in dylan... Jul 29 16:03:25 ndec: my policy is not to unless I've actually made a change to the patch Jul 29 16:03:53 ok. Jul 29 16:03:59 i didn't know that. Jul 29 16:04:11 I know Richard and others sign everything off... I just figure there's not a lot of point adding my name to the list :) Jul 29 16:04:26 ndec: thanks for checking though Jul 29 16:05:04 it typically indicates the 'upstream' path, but you are the committer anyway... Jul 29 16:31:17 i just noticed that pulseaudio was in DISTRO_FEATURES_BACKFILL. however by default, pulseaudio does not seem to be pulled into my image. does that mean that pulseaudio is now the de-facto support solution, and that bad things could happen if i disable pulseaudio completely (BACKFILL_CONSIDERED)? Jul 29 16:32:58 ndec: no, it's only there since it was floating / on by default in older configurations, so when adding the feature we had to back-fill it into people's existing distro configs Jul 29 16:33:09 ndec: you're free to disable it Jul 29 16:33:53 bluelightning: ok... we will see , i restarted a build, but because it touches DISTRO_FEATURE, it will take me a while.. Jul 29 16:34:08 indeed Jul 29 16:34:37 bluelightning: i noticed that gst-meta-audio does not depend on plugins-good-pulse. that's on purpose too? Jul 29 16:35:09 with the current default, we seem to enable pulse at build time, but it doesn't have any effect on image (hence i didn't notice before) Jul 29 16:56:20 Can somebody smell this fix I made to make sure I'm doing this right? … so I set DISTRO_FEATURES_append+="pam", and the target build of shadow broke, because shadow thinks that if I --enable-pam, it doesn't need to -lcrypt. I "fixed" this by adding BUILD_LDFLAGS="-lcrypt" in a .bbappend. Jul 29 17:21:54 ndec: it depends on what you build I guess... the main drivers for adding the feature were gstreamer and qt; I know qt does build differently depending on the feature, I'm not as close to gstreamer as others perhaps Jul 29 17:23:40 rtollert_: that sounds rather strange, since I believe shadow should indeed not need to -lcrypt when it's using PAM (since the crypto is abstracted behind the PAM api). What exactly is the undefined reference you get without it? Jul 29 17:24:38 there are cases where crypt is needed for other reasons, even with PAM enabled.. shadow uses crypt for it's 'passwd', and other commands in order to set and manipulate the passwd/group files.. Jul 29 17:37:22 pb_: shadow-4.1.4.3/lib/encrypt.c:48: undefined reference to `crypt' when linking chsh Jul 29 17:37:28 nerdboy, I wrote the image (rootfs.rpi-sdimg) to the SD card using dd, I can mount both partitions and everything, but it won't boot for some reason Jul 29 17:38:06 when I plug the thing in, PWR and ACT LEDs come on, then after a few seconds the activity LED goes off and that's that Jul 29 17:38:14 no output on composite Jul 29 17:53:43 pb_ fray: is this something I should submit as a patch? Jul 29 19:46:01 Hi. I have a problem with my kernel. I have: UBOOT_LOADADDRESS := 0x80008000, but when booting u-boot reports: Load Address: 00008000 Jul 29 19:46:58 I have no idea why it is relocated to the wrong address, works fine if I use the built toolchain to build the kernel, with the same bootloader Jul 29 19:55:20 looking at the logs, I see that during the bitbake build the kernel reports Load Address: 00008000, while buiding it separately reports 80008000 Jul 29 19:55:29 any ideas where to look? I'm blank Jul 30 00:31:32 nerdboy, I built yocto instead, thought that might work, but nope, when trying to boot rpi-basic-image, ACT lights up for a few seconds and that's it, no display output or anything. it's probably not booting given that the ethernet-related LEDs remain off even when powering up the RPi with a network cable attached Jul 30 01:46:16 hi Jul 30 01:50:09 i have build aminimal image and ogt rootfs, when i execute ldd it is showing that it is linking to some of system libraries... is it correct? for example Jul 30 01:50:41 ldd libsqlite3.so giving me this result Jul 30 01:50:41 linux-gate.so.1 => (0xf76f1000) libdl.so.2 => /lib/i386-linux-gnu/libdl.so.2 (0xf76ca000) libpthread.so.0 => /lib/i386-linux-gnu/libpthread.so.0 (0xf76af000) libc.so.6 => /lib/i386-linux-gnu/libc.so.6 (0xf7505000) /lib/ld-linux.so.2 (0xf76f2000) Jul 30 01:51:03 why it is using my system libs? it should not happen as of my understanding.. Jul 30 01:51:11 any body has any idea on this.? Jul 30 02:38:39 i have build aminimal image and ogt rootfs, when i execute ldd it is showing that it is linking to some of system libraries... is it correct? for example Jul 30 02:38:45 ldd libsqlite3.so giving me this result Jul 30 02:38:51 linux-gate.so.1 => (0xf76f1000) libdl.so.2 => /lib/i386-linux-gnu/libdl.so.2 (0xf76ca000) libpthread.so.0 => /lib/i386-linux-gnu/libpthread.so.0 (0xf76af000) libc.so.6 => /lib/i386-linux-gnu/libc.so.6 (0xf7505000) /lib/ld-linux.so.2 (0xf76f2000) Jul 30 02:38:56 any body has any idea on this.? **** ENDING LOGGING AT Tue Jul 30 02:59:58 2013