**** BEGIN LOGGING AT Wed Oct 14 02:59:59 2015 Oct 14 03:07:39 WARNING: core-image-minimal-initramfs: usermod command did not succeed. Retrying... Oct 14 03:07:41 ERROR: core-image-minimal-initramfs: Tried running usermod command 1 times without success, giving up Oct 14 03:07:43 :/ Oct 14 04:13:36 damn, I'm really not sure how to get the shallow git support to work well for linux-yocto.. Oct 14 04:14:09 hmm Oct 14 05:19:58 hi Oct 14 05:20:30 i woul like to include openjdk-8 with my yocto build Oct 14 05:21:48 so can anybody suggest me how to include this version. I am already trying but i am getting an error. Please see the following error-- Oct 14 05:23:12 build/tmp/work/x86_64-linux/icedtea7-native/2.1.3-r1.0/icedtea-2.1.3/build/openjdk.build-boot/hotspot/outputdir' | >&2 echo "*** This OS is not supported:" `uname -a`; exit 1; Oct 14 06:07:47 Good Morning! Someone here who can help me regarding poky 1.8 and nostdlib behaviour for building of package with includes c++ stuff? Oct 14 06:09:42 hok: whats the problem you see Oct 14 06:11:33 i'm trying to include libsocket with yocto 1.8. C part is runing find. C++ part needs to add includes for c++ headers which is already done. But linking is not working. Needed libs are not found. Oct 14 06:12:38 in special i tried to set LINK_DIRECTORY and CMAKE_MODULE_LINKER_FLAGS in CMakeLists.txt. Linker gets the options but with same situation Oct 14 06:12:52 echo "SET(CMAKE_MODULE_LINKER_FLAGS \" -Wl,-rpath-link,${STAGING_LIBDIR}\")" >> ${S}/C++/CMakeLists.txt Oct 14 06:12:57 echo "LINK_DIRECTORIES(${STAGING_LIBDIR})" >> ${S}/C++/CMakeLists.txt Oct 14 06:20:54 sources are https://github.com/dermesser/libsocket Oct 14 07:22:31 morning all Oct 14 07:41:22 good morning internet Oct 14 07:43:05 morning Ox4 Oct 14 07:56:35 guys, is it possible to create .img image by yocto? Oct 14 08:09:36 Ox4: please define ".img" Oct 14 08:10:28 Ox4: and while you're at it, please also define "yocto" *SCNR* Oct 14 08:17:47 LetoThe2nd: .img image instead of jffs2 or tar.bz2 Oct 14 08:18:07 LetoThe2nd: whad does mean *SCNR*? Oct 14 08:19:34 Ox4: for the first point: jffs2 and tar.bz2 are properly properly defined file formats. whereas .img is just a random filename extension with many, many meanings. Oct 14 08:20:06 Ox4: for the second - the search engine of your least distrust will certainly provide you with the meaning of "SCNR" Oct 14 08:22:04 Hey, I'm trying to build a x86_64 bit SDK on a 64 bit machine but gcc-crosssdk-initial fails with xgcc not found. Does someone had that issue or has an idea of what could be the problem? Oct 14 08:24:00 did someone have Oct 14 08:25:23 LetoThe2nd: understood. I will clarify from my devs what img format they want Oct 14 08:25:52 LetoThe2nd: Also I have one more question: is it possible to build image which will include kernel in it? Oct 14 08:26:00 Ox4: you can also read that as: feel free to add whatever image creation mechanism you want in your own layers. Oct 14 08:27:10 Ox4: sure, just include the kernel package Oct 14 08:28:05 domidimi: the question rings a bell, though i'm not sure where i've heard it. maybe fray or RP know, but they won't wake up until in 5 or 6 hours. Oct 14 08:29:32 LetoThe2nd: to my machine conf, right? Oct 14 08:32:22 Ox4: erm... no Oct 14 08:32:47 Ox4: bascially, to the recipe of the image you want to create. Oct 14 08:34:08 LetoThe2nd: thanks Oct 14 08:35:19 Ox4: see https://www.yoctoproject.org/sites/default/files/ypdd-mar2015-intro-lab.pdf Oct 14 08:35:29 Ox4: exercises 6 through 8 Oct 14 08:36:49 LetoThe2nd: thank you again :) Oct 14 08:37:52 Ox4: I'd suggest you take some time and work properly through the yocto QS and these slides, they really cover a lot of the common stuff. Oct 14 08:41:41 LetoThe2nd: what about ubifs? Oct 14 08:42:06 Ox4: ubifs is supported and AFAIK properly documented. Oct 14 08:44:39 The official doc of ubifs: http://www.linux-mtd.infradead.org/doc/ubifs.html Oct 14 08:45:53 Ox4: see https://github.com/linux4sam/meta-atmel/blob/master/conf/machine/sama5d3xek.conf for a rather simple example of using ubi Oct 14 08:46:36 to build a ubifs image, I think you should to add UBI or UBIFS to IMAGE_TYPES into you machine configuration Oct 14 08:47:02 LetoThe2nd: thank you master Yoda :) Oct 14 08:47:12 ^ it's a good example Oct 14 08:47:17 t0mmy_: plus a bunch of parameters essential for ubi creation, hence the example given :) Oct 14 08:47:42 yes, MKUBIFS_ARGS and UBINIZE_ARGS Oct 14 08:50:18 ^^ Oct 14 08:50:30 hups - wrong window, sorry Oct 14 08:51:05 silviof: you're also welcome here :) Oct 14 08:51:50 I also have a problem with building of watchdog: http://dpaste.com/32HVX23 Oct 14 08:52:20 WDIOC_PAUSE is defined in the linux/include in the kernel source Oct 14 08:53:04 I don't understand why it cannot be found, I pass the kernel sources through -I Oct 14 08:53:51 Ox4: that sounds a lot like you are trying to work around proper dependency handling Oct 14 08:56:28 and it sounds even more like some internal piece of software that nobody can comment on anyways. Oct 14 08:57:25 LetoThe2nd: thx for the welcoming :-) Oct 14 09:04:38 Morning folks Oct 14 12:30:43 guys, could somebody take a look: http://dpaste.com/25GN1DJ ? Oct 14 12:32:09 can I avoid to set SECTION? Oct 14 12:36:30 LetoThe2nd: According to ubifs, I have ubinize.cfg file: http://dpaste.com/1T9Q0NS How can I configure yocto to set such parameters from the config? Oct 14 12:39:22 Ox4: SECTION defaults to "base"... are you unsetting it? Oct 14 12:41:44 Ox4: did you have a look at the example given? Oct 14 12:43:37 LetoThe2nd: I am looking at there right now Oct 14 12:43:50 bluelightning: I didn't set it at all Oct 14 12:44:09 Ox4: and the parameters given there are not enough? Oct 14 12:44:40 Ox4: well that's odd because it's set by default... try bitbake -e yourrecipename | less then search for SECTION with / Oct 14 12:49:57 LetoThe2nd: I am not sure about vol_size=52080KiB, vol_type=dynamic, vol_name=rootfs and vol_flags=autoresize Oct 14 12:50:48 ah I see UBI_VOLNAME = "rootfs" Oct 14 12:51:30 Ox4: yes, I was just about to say... why not look again, then try it just as is, and after that have a look at what the parameters given really meant, through the logs? Oct 14 12:52:18 Ox4: i don't know them all by heart too, an reading the flags up in the documentation of ubinize and mkubifs is possible for you just as it is for me. Oct 14 12:56:04 http://dpaste.com/1M4X1ZK Oct 14 12:56:05 hm Oct 14 12:58:23 well then how about trying to increase the leb value as suggested by the message? Oct 14 13:18:53 hey folks Oct 14 13:19:35 hi raykinsella78 Oct 14 13:30:15 bluelightning: anythoughts on the best way to override KERNEL_MODULES_AUTOLOAD Oct 14 13:30:45 bluelightning: at the end of the build, there is a whole bunch of stuff in their I would rather wasn't autoloaded. Oct 14 13:30:58 bluelightning: btusb for example :-) Oct 14 13:31:41 it needs to be set at the configuration level, so basically local.conf, machine config or distro config Oct 14 13:34:42 bluelightning: tried KERNEL_MODULES_AUTOLOAD = "" is my local.conf, no effect. Oct 14 13:35:34 raykinsella78: it's probably being overridden, use bitbake -e | less to find out where/how Oct 14 13:35:57 raykinsella78: also it's KERNEL_MODULE_AUTOLOAD not KERNEL_MODULES_AUTOLOAD Oct 14 13:37:12 bluelightning: ok, will do Oct 14 13:40:51 any ideas why libtool might not be included in SDK? Oct 14 13:42:54 auto(-re-)conf is there, automake is there, somehow libtool is missing, hence autoconf complains about undefined AM_PROG_LIBTOOL in configure.ac Oct 14 13:44:08 hm libtoolize is there as well, it's just libtool that is missing Oct 14 13:52:36 bluelightning: how do I figure out where in the build KERNEL_MODULE_AUTOLOAD is being eval'ed? Oct 14 13:52:44 bluelightning: which package/action. Oct 14 13:53:35 raykinsella78: I can tell you, it's the kernel recipe (and any external module recipes) Oct 14 13:53:57 bluelightning: virtual/kernel Oct 14 13:54:05 bluelightning: ya - thanks Oct 14 13:58:32 RP is not around :/ seems like he might have fixed it in http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/meta/recipes-devtools/libtool?id=4b9fc479b6a68e6a3fb7cd9da6bdc9c4a786e4d3 Oct 14 14:01:39 bluelightning: KERNEL_MODULE_AUTOLOAD_remove works fine Oct 14 14:35:34 wow, I found nice wiki page according to UBIFS: http://www.crashcourse.ca/wiki/index.php/Poky_Image_Types Oct 14 14:53:55 Ox4: some parts of this wiki page are in image_types.bbclass, it seems to me you Oct 14 14:54:25 and in rpi layer Oct 14 14:58:56 t0mmy_: yes, I am curious about ubinize.cfg file Oct 14 15:51:06 I have one question,hope someone answer, tks Oct 14 15:51:39 guys, how the $PATH is set during the boot in yocto? Oct 14 15:52:16 I can use bitbake, but cant use bitbake-layers,yocto-layer, no the command when i use it, so what do i miss for installation? Oct 14 16:01:45 bluelightning: is there a wix tutotal somewhere? Oct 14 16:02:01 bluelightning: tutorial ... I mean Oct 14 16:02:26 bluelightning: ah its called wic, never mind Oct 14 16:02:35 raykinsella781: right.. it should be covered briefly in the manual Oct 14 16:02:59 Ox4: most likely the same way it is on nearly all linux machines, /etc/profile, and for the user, the bits copied from /etc/skel Oct 14 16:04:42 kergoth: the PATH should be set without user login in my case Oct 14 16:05:10 kergoth: and as far as I know /etc/profile is read when a user logs in Oct 14 16:05:49 for startup scripts and whatnot i think they generally have a hardcoded set default, i.e. just /usr/bin and /bin, /usr/sbin and /sbin, but i'm not sure where that's set offhand Oct 14 16:05:51 sorry Oct 14 16:07:42 as I see the PATH is set on each runlevel script Oct 14 16:07:55 also I patched rcS script, but it didn't help :-( Oct 14 16:08:27 I don't see my path after booting Oct 14 16:08:42 kergoth: btw I'm looking into the virtual/kernel thing, unless I'm doing something wrong it seems like the bb.providers functions don't expect to accept such targets Oct 14 16:16:07 https://github.com/MentorEmbedded/meta-mentor/blob/master/meta-mel/lib/recipetool/kernel.py#L58-L61 is what i used for the kernel recipetool commands Oct 14 16:16:10 not sure if its ideal, but it works Oct 14 16:16:20 afk Oct 14 16:17:13 there might be useful code in bb too, since i had to resolve providers there -- https://github.com/kergoth/bb Oct 14 16:17:30 k, actually afk now, walking up to the park with my wife and son, will likley get back on and work some from there Oct 14 16:20:48 kergoth: thanks Oct 14 17:16:13 Any folks that know linux-yocto stuff well around, by chance? Oct 14 17:21:55 kergoth, I'm otp, but might be able to help Oct 14 17:23:10 dvhart_: I'm working on shallow git repo support. Currently it keeps any branches listed in the uris, trims them to the SRCREV (that is, removes any commits on the branch after SRCREV), and loses any other branches. Do you know if the linux-yocto build process requires any dependent branches stick around for validation purposes, or if there are any other obvious complications that come to mind? Oct 14 17:25:45 in the past i was getting branch validation failures when building from shallow, but for some reason that's no longer the case, now i'm getting failures to apply patches, i'm guessing that the checkout is different between the two, working on isolating that now Oct 14 17:27:12 * kergoth digs Oct 14 17:50:41 hmmm, http://beagleboard.org/blog/2015-10-14-beagleboard-x15/ Oct 14 18:23:49 oh! i think i see the problem Oct 14 18:27:21 I'm guessing that the patch process is normally able to recognize that patches have already been applied in the form of commits, and when I remove that history, it's not able to recognize that they've already been applied, so fails to apply them Oct 14 18:27:25 just need to retain more of the history Oct 14 18:27:42 I might need to rework the usability of this slightly Oct 14 18:28:53 * kergoth tests Oct 14 18:29:55 Has anyone ever seen python land in a finish build missing modules? Oct 14 18:30:05 I am installing python 2, and it is weirdly missing the subprocess module Oct 14 18:32:14 ryanstur1er: if you need all of the standard modules you need also to install python-modules Oct 14 18:32:37 ryanstur1er: we have pretty fine granularity on our python packaging so you can be selective in order to save space on the target Oct 14 18:35:13 And there it is! Oct 14 18:35:15 Thank you. Oct 14 18:37:59 Do I DEPENDS or RDEPENDS on python-modules? Oct 14 18:39:41 ryanstur1er: RDEPENDS Oct 14 18:43:04 Got it! Thanks! Oct 14 20:29:19 Hmm, I think I need to separate the git shallow depth from enable/disable Oct 14 20:47:11 damnit Oct 14 20:49:50 yeah it's a damn it day Oct 14 20:49:52 asklasklakslzaksl;ask Oct 14 20:57:30 Where can I configure the bitbake worker NFS sstate filepath character limitation? Oct 14 20:57:37 It seems to be set to 144, which isn't always enough Oct 14 21:16:49 Hmm, if I do a depth 10 shallow git clone of a repo where the two of those commits merged in a ton of other commits, will i keep all those commits, since the depth included the merges? Oct 14 21:17:05 I'm so sick of this Oct 14 21:19:09 * kergoth grumbles Oct 14 21:20:04 if i make v4.1.8 the shallow graft, it does very little good, since the branch merged in v4.1.7 and others previously, so all grafting v4.1.8 accomplished was removing the commits between v4.1.7 and v4.1.8 from history, nothing else, since it came in via previous commits.. Oct 14 21:20:26 either I need to graft at the merge commit, graft a whole bunch of different commits, or try to graft every commit in a git rev-list v4.1.8 Oct 14 23:15:09 wtf, I'm getting 'no bb files matched bbfile_pattern' for every layer in bblayers Oct 14 23:15:25 anyone run into this? Oct 14 23:29:53 hmm **** ENDING LOGGING AT Thu Oct 15 02:59:59 2015