**** BEGIN LOGGING AT Tue Nov 07 03:00:03 2017 Nov 07 04:56:56 WARNING: No bb files matched BBFILE_PATTERN_core '^/home/bodangly/devel/pitco750.new-test/sources/poky/meta/' Nov 07 04:57:03 but the directory exists, looks normal to me Nov 07 04:57:11 and does have BB files in the various subdirs Nov 07 04:57:31 I am getting this for all my layers pretty much, not sure what I changed because it was working before Nov 07 04:57:43 my local.conf, bblayer.conf between a previous build which works and this one, are identical Nov 07 04:57:53 any ideas on what could cause this? Nov 07 10:04:04 does yocto comes by default with gnome/kde ? Nov 07 10:04:23 How to check if my image contain any of it Nov 07 10:04:29 ? Thank you Nov 07 10:05:27 is it possible to get any help to make gdbserver work on a yocto rootfs Nov 07 10:05:28 ranran> yocto is highly customisable, so we need more information about your image, your board etc ... Nov 07 10:06:31 alkino: see the docs Nov 07 10:06:45 rburton_: I have read all the doc :p Nov 07 10:06:54 ranran: out of the box, oe-core doesn't come with gnome or kde but a gtk+3 based ui of its own that is intentionally small/minimal Nov 07 10:07:08 ranran: meta-gnome has bits of gnome2 and gnome3, but nobody has actually packaged all of gnome3 Nov 07 10:07:38 alkino: so what's the problem? Nov 07 10:07:49 I use imx6, smarc (adlink) Nov 07 10:08:18 gdb can't find debug library for ld-linux.so.3 Nov 07 10:08:43 bt gives me "#0 0x4a1f9f8c in ?? ()" Nov 07 10:08:45 0x4a1f0000 0x4a210000 0x20000 0x0 /lib/ld-2.24.so Nov 07 10:08:47 0x000007c0 0x0001bc54 Yes /home/alkino/build/sysroots/cortexa5-overkiz-linux-gnueabi/lib/ld-linux.so.3 Nov 07 10:09:34 rburton_, how should I add gnome ? Nov 07 10:10:19 ranran: see meta-gnome. i don't think its all of gnome2 though. the recipes stopped working with new gcc, and nobody cared enough to fix them. Nov 07 10:10:26 ranran: start giving a look here http://layers.openembedded.org/layerindex/branch/master/layer/meta-gnome/ Nov 07 10:10:30 I set sysroot to the sysroot of the toolchain Nov 07 10:10:49 rburton_: do you know if there is a way to opt out of the volatile-binds symlinking of tmpfiles.d/etc.conf to /dev/null ? or is there another way to ensure that resolv.conf gets symlinked to systemd-resolved ? Nov 07 10:11:05 Thanks, another more general question, how do I know if my image already contain gnome (or anything else I am interested in) ? Nov 07 10:11:11 ranran: and make sure that you're using X11/wayland Nov 07 10:11:19 ranran: because you wrote the image recipe? Nov 07 10:11:32 alkino did you look at : http://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html#debugging-with-the-gnu-project-debugger-gdb-on-the-target ? Nov 07 10:11:52 ranran: if you want to know what is in an image then either look at the recipe, or build it and look at the manifest in deploy/ Nov 07 10:12:52 ranran: besides the meta-gnome layer, you can start parsing the list of built packages, eg: oe-pkgdata-util list-pkgs | grep gnome Nov 07 10:13:01 nayfe: I can't debug on target, there is no more space available Nov 07 10:13:14 ranran: or look at your image manifest Nov 07 10:13:24 Thanks, zero_not, rburton. Nov 07 10:13:33 rburton_: sorry for directly asking you, but you reviewed the patch adding the symlink, so i guess you might also be able to give me a hint Nov 07 10:14:17 not off the top of my head. my guess is that the resolved stuff needs a good look at Nov 07 10:14:24 previous chapter is about debugging remotly Nov 07 10:15:05 rburton_, I've tried the list and I see gnome...gnome-common-dbg connman-gnome-dev gnome-desktop-testing connman-gnome gnome-common gnome-desktop-testing-dev gnome-common-dev gnome-desktop-testing-dbg connman-gnome-dbg at-spi2-atk-gnome Nov 07 10:15:53 rburton_: alright Nov 07 10:16:35 Does it mean that I should expect to have desktop environment with my image ? Nov 07 10:18:45 nayfe: Yes I have already do it Nov 07 10:19:29 except that instead or unarchive rootfs, I install toolchain in a directory Nov 07 10:19:35 there is ".debug" in there Nov 07 10:20:06 in the manifest, I also see gnome ges/lec-imx6/fsl-image-qt5-lec-imx6-20171106104512.rootfs.manifest | grep gno connman-gnome cortexa9hf-vfp-neon 0.7 gnome-desktop-testing cortexa9hf-vfp-neon 2014.1 gnome-desktop-testing-dev cortexa9hf-vfp-neon 2014.1 Nov 07 10:20:27 ranran: its your image, what did you tell it to build? Nov 07 10:21:58 the image is called fsl-image-qt5 which suggests that it ships qt5 not gnome Nov 07 10:23:51 right Nov 07 10:23:56 it is here: https://drive.google.com/file/d/1mbErYpYO8DyqohQ1iJ3YKWD-uowzBnD8/view?usp=sharing Nov 07 10:24:31 Can I add gnome with it ? Nov 07 10:24:32 that isn't an image recipe Nov 07 10:24:52 as we've said, gnome is in meta-gnome. its not complete though and probably won't work enough to be a gnome desktop, unless you want to fix it. Nov 07 10:24:57 oh... so where is the image recipe ? Nov 07 10:25:07 the image recipe is called fsl-image-qt5.bb Nov 07 10:27:06 doesn't local.conf should name the image recipe "fsl-image-qt5.bb" ? I can't find where it is mentioned Nov 07 10:28:13 it's in meta-freescale or meta-fsl-arm stuff Nov 07 10:33:56 I have found gnome, in the manifest, and lists, (and rootfs), doesn't it mean it is supported ? Nov 07 10:38:53 I have found gnome, in the manifest, and lists, (and rootfs), does it mean gnome is supported (in fsl-image-qt5 image) ? Nov 07 10:39:54 ranran: you found a few recipes with gnome in the name, that is not the same as a gnome desktop Nov 07 10:40:01 read the bit where we spoke about meta-gnome again Nov 07 10:40:28 ok, I now see what you wrote, I've skipped that. sorry. Nov 07 10:42:23 Does it mean that yocto image are not supported with desktop environment ? or is it a specific bug in my image recipe which tried to use gnome ? Nov 07 10:43:15 I think I once worked with intel yocto image, and it had desktop environment. Nov 07 10:45:06 the images can contain anything you want to put in them, as long as there is a recipe for it Nov 07 10:45:19 sato images ship with the sato environment, which is a basic gtk+3 based one Nov 07 10:45:35 it seems there is more support for QT envirment with Freescale socs Nov 07 10:45:47 fsl definitely like qt5, yes Nov 07 10:45:59 sato would work just fine though Nov 07 10:46:20 the question is do you want gnome2, gnome3, or just something you can click and start apps with Nov 07 10:47:14 alkino> if i can ask, is it a custom app you'r trying to debug ? Nov 07 10:47:55 We thought to use the board (with its imx6) as a development environment (compile qt application on imx6) Nov 07 10:49:07 maybe try to use a recent community release ? Nov 07 10:49:49 should I move to sato image ? I need qt5 too Nov 07 10:50:23 rule of thumb: image recipes are example and by definition you'll be customising or writing your own Nov 07 10:59:33 Hi, what does the "t" mean in tunes ? e.g tune-cortexa9 vs tune-cortexa9t ? Nov 07 11:05:17 thumb2 or trustzone maybe ? :) Nov 07 11:24:32 I get "Read-only file system" with the image. Is there a way to fix it ? Nov 07 12:22:47 New news from stackoverflow: Yocto error kernel-module-iptable-mangle [closed] Nov 07 12:46:02 @nayfe very probably thumb2 since it seems to do something like cortexa9t2-vfp and cortexa9t2-neon. Thanks. Nov 07 12:52:52 New news from stackoverflow: iptables does not recognize the --dport argument Nov 07 12:56:35 can you tell us result of this command cat /etc/default/rcS | grep READ_ONLY= Nov 07 13:10:21 mmh I'm having trouble with python-native on krogoth Nov 07 13:10:25 there's a local patch which adds Nov 07 13:10:27 PREFIX = os.path.normpath(sys.prefix).replace( os.getenv("BUILD_SYS"), os.getenv("HOST_SYS") ) Nov 07 13:10:38 but turns out, os.getenv("BUILD_SYS") and os.getenv("HOST_SYS") are both empty Nov 07 13:10:47 so python yells "TypeError: expected a string or other character buffer object" Nov 07 13:12:46 joshuagl, rburton_, halstead: I put tumbleweed on vanilla back in the pool Nov 07 13:14:04 RP: neat, thanks Nov 07 13:15:13 halstead: the tumbleweed failures were looking like being the fault of the host OS kernel so we switched from kernel-default (openSUSE patches) to kernel-vanilla (upstream kernel) and things seem to have stabilised Nov 07 13:17:32 joshuagl: I put a comment in the bug Nov 07 13:17:44 RP: thanks! was just heading to do so Nov 07 13:18:02 RP: hmm, which bug? :-) Nov 07 13:18:28 joshuagl: https://bugzilla.yoctoproject.org/show_bug.cgi?id=12301 Nov 07 13:18:29 Bug 12301: normal, Undecided, ---, bruce.ashfield, ACCEPTED , Kernel boot hanging on 4.12.14 with qemux86-64 Nov 07 13:18:37 RP: thanks Nov 07 13:23:37 maxin: ping ? :) Nov 07 13:34:48 Heyho, does anyone got opkg version 0.3.4 to work with omap and openwrt sources? Nov 07 13:35:14 opkg 0.3.3 with morty works Nov 07 13:54:18 rburton_: looks like we need to document setuptools as an OE-Core dep and add it to the AB worker Nov 07 13:54:20 s Nov 07 13:54:27 Can somebody has a clue about following issue? I had a working yocto (morty) system before. This was running fine on my 32-bit linux machine (mint 18.1). Due to malloc space insufficience, I recently upgraded it to 64bit (multiarch) Nov 07 13:54:42 joshuagl: maybe. currently wrapping that call in a try/except :) Nov 07 13:54:43 Now everything compiles fine, but in the final stage, I have an issue Nov 07 13:55:05 QA Issue: File '/etc/skel/.bashrc' from base-files was already stripped, this will prevent future debugging! [already-stripped] Nov 07 13:55:08 etc... Nov 07 13:55:14 many issues like that Nov 07 13:55:14 rburton_: we stopped polluting the build env with AB modules, otherwise that might have passed without incident Nov 07 13:55:32 joshuagl: yeah thats probably why it worked last time i tried those patches Nov 07 13:57:56 Hi, is there a way of emptying a _append in a .bbappend? The issue is that an image is adding an image class via a IMAGE_CLASSES_append, but the appended image class is not existing anymore. I can not run a anonymous since I get a parse error while it tries to inherit... So I did try IMAGE_CLASSES_append = "" but that did not help... Nov 07 13:59:10 ... and IMAGE_CLASSES_remove did not either Nov 07 14:01:49 proposal: move npm.bbclass out of oe-core to be alongside nodejs.bb Nov 07 14:04:09 Hi guys, is there any reccomended way to persist iptables rules with systemd in Yocto? Any save / load iptables.service or somesuch? Nov 07 14:06:45 oh, found this: https://patchwork.openembedded.org/patch/131285/ Nov 07 14:18:04 RP, joshuagl: re-spun mut Nov 07 14:19:27 rburton_: cool, thanks Nov 07 14:27:51 hi Nov 07 14:29:02 i have lots of files in /etc so i set FILES_${PN}="/etc/" but one file needs to go into different ipk, so i do FILES_${PN}-addon = "/etc/thefile" Nov 07 14:29:28 but the packages-split for -addon is empty Nov 07 14:29:42 PN probably is before PN-addon in PACKAGES Nov 07 14:29:50 oh Nov 07 14:29:51 change the order so PN-addon is first Nov 07 14:29:57 thats sounds super easy Nov 07 14:29:58 :) Nov 07 14:30:47 rburton: many thanks! thats it Nov 07 14:39:09 Is it possible to use a library RecipeA generates while building its primary target? And if so, whats the path within my own Recipe which depends on that helper lib? Nov 07 15:11:45 Hello! What's the right mailing list to post a question on EGL / X11? Nov 07 15:26:00 Guest47029: in general, or specifically to yocto? the latter, yocto@. otherwise your BSP/GL provider support list Nov 07 15:28:54 rburton_: I am having trouble with EGL and X11 configuration with my Yocto build. I see OE-Core maintains X11, perhaps Yocto@ is the right place? Nov 07 15:47:29 Is it possible to use 32-bit sdk (/opt/poky) in 64-bit machine ? Nov 07 15:47:48 Hi. Nov 07 15:53:29 Is it possible to use 32-bit sdk (/opt/poky) in 64-bit machine ? Nov 07 15:53:31 YPTM: Access Informaton: - dial into: 1-800-262-0778/404-397-1527 - enter the bridge number: 88748961# Nov 07 15:53:43 YPTM: Stephen Joined Nov 07 15:53:55 wow you are early today Nov 07 15:54:52 better than being late Nov 07 15:56:59 YPTM: Joshua joined Nov 07 15:57:17 YPTM: armin joined Nov 07 15:57:46 YPTM: Stephano joined Nov 07 15:59:20 ין Nov 07 15:59:21 hi Nov 07 16:00:09 YPTM: bavery joined Nov 07 16:00:49 YPTM: Richard joined Nov 07 16:01:18 YPTM: Joshua Watt here Nov 07 16:01:19 YPTM: Mark here Nov 07 16:01:43 YPTM Michael here Nov 07 16:02:09 * armpit Yocto Vista point Nov 07 16:03:05 YPTM: Patrick here Nov 07 16:03:17 YPTM: Pascal here Nov 07 16:03:56 YPTM: Randy joined. Nov 07 16:05:08 YPTM: Saul's on Nov 07 16:05:29 YPTM: Manju joined Nov 07 16:09:41 Is that freeze only on the master branch? Nov 07 16:14:14 armpit: http://git.yoctoproject.org/clean/cgit.cgi/poky/commit/?h=master-next&id=71250efee4e194856e8c9a8ad80b66f1edd05ab1 Nov 07 16:14:23 k Nov 07 16:18:56 2.5: RP's *likely* themes: build stability!, containers, YP-compatible v2 + new website, layer setup tools, test automation (testopia?) Nov 07 16:28:02 YPTM is over. Nov 07 16:28:58 WARNING: No bb files matched BBFILE_PATTERN_core '^/home/bodangly/devel/pitco750.new-test/sources/poky/meta/' Nov 07 16:29:11 the directory exists and has bb files in it, its really bizarre Nov 07 16:29:19 what could be causing this? Nov 07 16:29:35 My bblayer.conf is identical to another version of the build, where it does work Nov 07 16:30:34 is there a good way to debug this? Nov 07 16:48:21 is it the . in the path name? Nov 07 16:48:31 if BBFILE_PATTERN is a regex Nov 07 16:54:32 RP, the morty pull request is stuck in oe-core-owners list. I'll resend with diff email addr Nov 07 16:55:56 armpit: ok, I don't think I get those Nov 07 16:56:31 bodangly: which poky release series was that? There was a bug like that in older versions Nov 07 16:56:46 mailman blocks if not a member and my work email is not registered Nov 07 16:57:05 bodangly: or maybe we never merged the fix, there was a problem with some patches if I remember rightly Nov 07 16:59:22 RP: its jethro Nov 07 16:59:35 RP: but it isn't just poky, its all my layers Nov 07 16:59:39 tull ? Nov 07 16:59:40 RP: should have mentioned that Nov 07 16:59:50 in the shuffling madness Nov 07 16:59:57 :D Nov 07 17:00:17 bodangly: http://git.openembedded.org/bitbake/commit/?h=1.34&id=91c3b34625fac2a0f093a4b46a46e89f813e7972 ? Nov 07 17:00:41 RP: I should be making sure I have this patch? Nov 07 17:00:58 bodangly: I'm asking if that could be the problem? Nov 07 17:01:22 without the patch there's no way to slow down Nov 07 17:08:46 RP: my cooker.py does not have this patch Nov 07 17:11:26 RP: but I don;t think its the patch that is the issue, because my older version of my build also is missing that patch but works fine... Nov 07 17:13:11 Is it possible to use sdk (/opt/poky) of 32-bit in 64-bit machine ? Nov 07 17:18:22 RP: is it possible that when I migrated the build, the pyc files are "stale", and don't reflect the actual .py files? Nov 07 17:19:13 RP: because when I diff the two builds, the pyc files differ, but the .py files are the same Nov 07 17:20:05 bodangly: I find that unlikely. You could check by deleting them Nov 07 17:20:33 RP: yeah they are in .gitignore, so it shouldn't be that... Nov 07 17:20:56 RP: hmm maybe I just need to start reverting commits to find where I fucked up this migration Nov 07 17:21:18 bodangly: or you could try applying the patch. or check if you set the line mention in that commit in any of your layers? Nov 07 17:21:27 bodangly: is the order in the bblayers.conf the same? Nov 07 17:21:40 yeah my bblayers.conf between working and broken is identical Nov 07 17:30:50 halstead, ping Nov 07 17:31:55 halstead, the githook changes, do they work on master only or stable branches ? Nov 07 17:42:39 armpit, Which githook changes are you asking about? Nov 07 17:42:50 cve Nov 07 17:43:39 Hi, ive seen application where downloads job is made to fetch everything then everything in that directory is soft linked to 4 different directories to run 4 different builds in separate containers. this is done through jenkins in a parametrized build. I try the same thing and Im getting the 'only one bitbake should be run at a time' which is to be expected because the bitbake.lock is in the build directory. How is that working? Nov 07 17:43:54 and how important is running just one bitbake on one directory Nov 07 17:44:02 halstead, master got a cve fix http://git.yoctoproject.org/cgit/cgit.cgi/poky/commit/?id=1ffaaa2c482badb593ed12dcd9f9e43e19487c49 Nov 07 17:44:07 want to run two bitbakes for two different machines Nov 07 17:44:58 Thought an email was supposed to hit the sec list Nov 07 17:45:13 armpit, I don't think there are any CVE related githooks in poky. Checking now. Nov 07 17:45:36 was OE then? Nov 07 17:45:39 armpit, There were CVE changes added to openembedded-core though. Nov 07 17:46:07 armpit, Still working on those though. Nov 07 17:46:10 ok then http://cgit.openembedded.org/openembedded-core/commit/?id=08f8d5db06647b94f96d655100c358047682dd2f Nov 07 17:46:18 ok Nov 07 17:46:48 in any case what branch with the hooks monitor? Nov 07 17:48:32 armpit, There isn't any branch filtering at the moment. It's just grepping for 'CVE: ' in the log between oldrev and newrev every push. Nov 07 17:49:45 k Nov 07 17:49:52 thanks Nov 07 17:50:30 armpit, Should we limit to certain branches? And do we need to run this on poky? Nov 07 17:51:54 I think having it on poky will just cause dup emails being sent out Nov 07 17:53:40 regarding the branch, if the branch name is not in the subject, it may confuse folks on the list if the hooks look at all changes. I think limiting to master is fine for now Nov 07 17:58:04 did something hit the security list Nov 07 17:58:57 Crofton|work, No success yet. Still working on it in spare cycles. Nov 07 17:59:03 ok Nov 07 17:59:40 not sure many commits have hit the reops Nov 07 18:00:12 armpit, has a bin update Nov 07 18:00:20 * An error in TSIG handling could permit unauthorized zone transfers Nov 07 18:00:20 or zone updates. These flaws are disclosed in CVE-2017-3142 and Nov 07 18:00:20 CVE-2017-3143. [RT #45383] Nov 07 18:02:11 Crofton|work, do you mean bind? Nov 07 18:02:28 yes :) Nov 07 18:06:24 Crofton|work, not sure how the git hook is being implemented.. git headers or patch headers or both Nov 07 18:38:22 Hi, I am trying to run an abi-dump on the target sysroot libraries generated by do_populate_sdk. I see that the debug symbols are missing even after adding dbg-pkgs, and dev-pkgs in EXTRA_IMAGE_FEATURES Nov 07 18:39:46 I also tried INHIBIT_SYSROOT_STRIP = "1". But, none of my .so files in /usr/lib of the target sysroot in the generated SDK have debug symbols in them. Am I missing something? Nov 07 18:52:40 fucking bizarre. If I take the sources directory of my working build and put it in my new build, the new one works. But all that is different in the diff are a couple bb files. Nov 07 19:00:32 armpit, It really is a grep of the logs as I said. You can look at https://paste.fedoraproject.org/paste/Bsb5jzffdUg6uGgmb0Xr0Q/raw to see exactly. Nov 07 19:09:33 ok. thanks Nov 07 21:24:28 New news from stackoverflow: How do I add more python modules to my yocto/openembedded project? Nov 08 00:50:49 myep Nov 08 01:43:18 need some advice. I have both imx-gst1.0-plugin and gstreamer1.0-plugins-imx in a build because I need some gstreamer plugins from NXP and some from the community FSL. There is one file that conflicts doing so though. What is a way to prevent a recipe from installing a specific file? Nov 08 01:43:53 basically I need to remove /usr/lib/gstreamer1.0/libgstimxcompositor.so from my final image Nov 08 01:48:04 should I make a bbappend with a do_install() that does rm ${D}/usr/lib/gstreamer1.0/libgstimxcompositor.so? Nov 08 02:07:11 hello, is that yocotproject? Nov 08 02:20:32 bobo Nov 08 02:21:40 bodangly: yes thats a good approach, although, I would suggest to solve it amicably between the two recipes meaning one not providing it Nov 08 02:21:50 permanently Nov 08 02:29:10 khem: well I guess I could take that approach, but how would I really accomplish that? It isn't a mistake from upstream, its just that I am doing something a little weird in my build. Nov 08 02:30:12 khem: I suppose I could modify the actual source to not provide it in imx-gst1.0-plugin but that will make it tricky to merge any patches from upstream in the future Nov 08 02:30:52 khem: anyhow thanks, glad that my approach is at least viable. I will need to think about whether its worth doing something more permanent Nov 08 02:38:21 bodangly: I was hinting to fix it upstream Nov 08 02:38:31 so you dont need bbappends in long run Nov 08 02:39:25 khem: It may be worthwhile to point it out to them. What ends up happening, is if I have both recipes in the build, then gstreamer hangs when scanning for the plugins, spitting out warnings on duplicate symbols. Nov 08 02:39:55 khem: but if I remove that file, then the two plugins play nice with each other Nov 08 02:41:10 khem: so in theory it should be possible for them to make them compatible. but I doubt NXP will do so, so probably it would have to come from the community side. But, I am not sure if its a problem in later versions of the layers. I am building Jethro currently. Nov 08 02:47:39 unfortunately the product uses a technexion SOM which doesn't have a BSP release later than Jethro :x **** ENDING LOGGING AT Wed Nov 08 03:00:02 2017