**** BEGIN LOGGING AT Tue Oct 15 02:59:57 2019
Oct 15 04:28:23 <jeranio> Hello
Oct 15 04:28:55 <jeranio> Can anyone help here with cross compiling issues?
Oct 15 04:31:02 <jeranio> I am using eclipse with yocto toolchain
Oct 15 05:16:36 <yocti> New news from stackoverflow: How to blacklist the galcore kernel module from yocto rootfs <https://stackoverflow.com/questions/58387637/how-to-blacklist-the-galcore-kernel-module-from-yocto-rootfs>
Oct 15 06:16:44 <yocti> New news from stackoverflow: Rauc and Yocto on Jetson Nano - Unable to find primary boot slot <https://stackoverflow.com/questions/58265118/rauc-and-yocto-on-jetson-nano-unable-to-find-primary-boot-slot>
Oct 15 07:06:16 <alessioigor> good morning!
Oct 15 07:12:32 <mckoan> good morning :-)
Oct 15 07:25:27 <guerinoni> hi to everyone, I need to set hostname with dynamic IP to my device, I already installed avahi in the system but when I change hostname other device can't ping through new hostname... Can someone help me?
Oct 15 07:46:58 <yocti> New news from stackoverflow: How to use OpenCV with camera on Jetson Nano with Yocto/poky <https://stackoverflow.com/questions/57989992/how-to-use-opencv-with-camera-on-jetson-nano-with-yocto-poky>
Oct 15 07:49:54 <fredrigu> Hi, I'm trying to add support for a new package format, more than deb, rpm and ipk. This works fine so far, but building python3-native the task do_package is run when I've my own package_apk defined. However if I've package_rpm defined do_package is not run for python3-native. How can I prevent do_package to be run for native packages when I've package_apk defined?
Oct 15 07:53:58 <ruru4143> LetoThe2nd: to make my problem reproduceable i wrote a script to clone/config yocto and then run bitbake. (https://termbin.com/wxkw) with that script, it works, i have no idea why it didn't work before
Oct 15 07:54:39 <ruru4143> LetoThe2nd: but i'm very happy now that it's working
Oct 15 07:56:00 <qschulz> armpit: thanks for taking the patch (maybe speaking too fast since it's not in the tree yet..) for thud for arm64 MACHINEOVERRIDES ordering.
Oct 15 08:17:03 <yocti> New news from stackoverflow: How to add libgif as a package in yocto build <https://stackoverflow.com/questions/58389887/how-to-add-libgif-as-a-package-in-yocto-build>
Oct 15 09:17:13 <yocti> New news from stackoverflow: IMX7D Sabre Yocto Project with OP-TEE examples <https://stackoverflow.com/questions/58391185/imx7d-sabre-yocto-project-with-op-tee-examples>
Oct 15 09:57:22 <iceaway> LetoThe2nd: about what I asked yesterday about "wifi" distro feature requiring x11. If I used DISTRO_FEATURES_append instead of DISTRO_FEATURES +=, the X11 requirement disappeared.
Oct 15 09:58:55 <qschulz> iceaway: time to dig into bitbake -e to see why it's different
Oct 15 09:58:59 <rburton> two reasons
Oct 15 09:59:10 <rburton> 1) DISTRO_FEATURES is assigned with ?= in the guts so you can easily wipe it out
Oct 15 09:59:11 <litb> hello folks!
Oct 15 09:59:17 <rburton> 2) remember to use leading whitespace with _append
Oct 15 09:59:49 <litb> what happens if I define and export cmake-mingw32_do_generate_toolchain_file() in my cmake-mingw32.bbclass, and cmake.bbclass does the  same? what function will be exported to the recipes?
Oct 15 10:00:01 <litb> (if those inherit both classes). will there be an error message?
Oct 15 10:00:45 <litb> i did see code that did  _append +=
Oct 15 10:01:04 <litb> looked like it does the whitespace appending for you if you use +=, but not sure.
Oct 15 10:02:26 <litb> i've no idea what "_append =+" would do, though xD
Oct 15 10:02:56 <rburton> append += is stupid
Oct 15 10:02:59 <qschulz> rburton: for 1) I'm torn, not having something more than when using _append (x11 is there when using += not with _append)
Oct 15 10:03:22 <qschulz> seems fishy to me
Oct 15 10:04:00 <rburton> could easily be += was wiping out the weak assignment, or a lack of whitespace with _append meaning the added distro feature was "systemdwifi" or similar
Oct 15 10:05:02 <qschulz> rburton: yes but in that case you have nothing more than what you have in your +=, so x11 shouldn't be picked up. The missing space in _append would explain it though, breaking the thing
Oct 15 10:17:23 <yocti> New news from stackoverflow: ONVIF implementation in Yocto <https://stackoverflow.com/questions/58392229/onvif-implementation-in-yocto>
Oct 15 10:26:45 <litb> rburton, hm the mega manual has it, SRC_URI_append += "file://patch-file-one"
Oct 15 10:29:08 <rburton> litb: not any more :)
Oct 15 10:46:28 <ricardocrudo> how to set the default locale on yocto.
Oct 15 10:46:31 <ricardocrudo> ?
Oct 15 10:58:28 <ricardocrudo> actually, I'm getting this error `character map file `utf8' not found: No such file or directory` when I add `GLIBC_GENERATE_LOCALES += "en_GB.utf8"`
Oct 15 11:00:09 <rburton> en_GB is utf-8 out of the box
Oct 15 11:00:39 <rburton> for "reasons" en_GB is utf8 and you need to specify encoding to get other variations
Oct 15 11:29:24 <litb> if a task has the attribute [noexec] assigned. can I still manually execute it and it actually executes, within another task?
Oct 15 12:14:46 <tgamblin> khem: saw your email but looks like the rsyslog update is already on zeus and master. Did you still need me to do something there?
Oct 15 12:21:33 <bojones> Hello, anyone using the barebox bootloader on x86 systems?
Oct 15 12:44:26 <OnkelUlla> bojones: ask for it in #barebox
Oct 15 12:44:43 <bojones> Doh.. Thanks :)
Oct 15 12:47:03 <ricardocrudo> After a fresh image install when I run `localectl status` I get "System Locale: n/a". How can I set System Locale = en by default? I have it installed in my image as I see in `localectl list-locales`
Oct 15 13:04:36 <mcfrisk> is there a way to run target/cross gdb in bitbake devshell? can't run it on target and feel too lazy to extend SDK to debug a crash..
Oct 15 14:01:21 <litb> gcc-crosssdk includes  gcc-cross_${PV}.bb.    when building gcc-crosssdk, will gcc-cross_%.bbappend also apply then? i.e for include or requires lines?
Oct 15 14:01:32 <litb> or only when bitbake processes a .bb at the "toplevel" ?
Oct 15 14:04:17 <qschulz> litb: bitbake -e and see for yourself :) ? (and let me know because I find the question interesting)
Oct 15 14:05:34 <litb> thanks, I will try when all the files should work in theory
Oct 15 14:19:34 <fury> "The metadata is not deterministic and this needs to be fixed." oof. do i have to fix it to a specific revision and commit the metadata every time i make a change to the app? the commit hash sometimes goes backwards and i think this is causing it to not unpack the git repo into the temp directory, making other steps fail O_o
Oct 15 14:22:29 <LetoThe2nd> fury: you can set SRCREV to autorev. but generally, bitbake tells you when it thinks it might be unable to reproduce what it is just doing
Oct 15 14:23:36 <fury> i think it is already set to autorev, the tmp directory gets _1.0+gitAUTOINC+lastcommithash-r0 appended to it
Oct 15 14:23:49 <LetoThe2nd> fury: for development purposes, if you set the app source to your local check with devtool modify -n then all is well, for example
Oct 15 14:24:21 <fury> i had a problem the first couple times i tried to build it, committed a couple fixes, but the third time i tried to run it it had no git folder in the tmp directory for that revision
Oct 15 14:24:59 <fury> deleted tmp and trying again from scratch to make sure i'm not just crazy
Oct 15 14:30:34 <armpit> qschulz, you are welcome. its up to RP to do the merge..
Oct 15 14:43:10 <apteryx> just stumbled on bitbake unable to checkout one repo, because it failed parsing my ~/.ssh/config due to the use of the 'ProxyJump' directive.
Oct 15 14:54:14 <RP> armpit: think I just did that
Oct 15 14:54:52 <armpit> yep. thanks
Oct 15 15:04:22 <tlwoerner> could someone drop the link to the planning doc here?
Oct 15 15:07:08 <paulbarker> tlwoerner: This one from the status email? https://docs.google.com/document/d/1UKZIGe88-eq3-pOPtkAvFAegbQDzhy_f4ye64yjnABc/edit?usp=sharing
Oct 15 15:08:31 <tlwoerner> paulbarker: ah thanks!!
Oct 15 15:15:16 <milloni> where do i send patches for poky?
Oct 15 15:16:04 <milloni> bitbake-devel@lists.openembedded.org i assume?
Oct 15 15:16:17 <milloni> sorry, poky@yoctoproject.org
Oct 15 15:16:17 <milloni>  rather
Oct 15 15:16:17 <tlwoerner> milloni: that would be for bitbake patches
Oct 15 15:16:28 <tlwoerner> milloni: are your patches poky-specific?
Oct 15 15:17:01 <milloni> it's a patch for poky/meta/recipes-devtools/python
Oct 15 15:18:02 <milloni> so it's the "meta" layer - i think that lives in two different places, or im confused
Oct 15 15:19:02 <tlwoerner> milloni: send it to openembeded-core
Oct 15 15:19:08 <milloni> ok, thanks!
Oct 15 15:33:02 <milloni> "You need to subscribe to the appropriate mailing-list in order to be able to send your patch(es) there" do i *have to* be subscribed to send a patch?
Oct 15 15:34:27 <JPEW> milloni: On some lists they will be sent to a moderator for review before posting if your not signed up.... but I'm not sure if that applies to the lists you are referring to.
Oct 15 15:36:00 <milloni> sure, i thought this was an anti spam feature, i'm just being wary of the traffic to my mail address
Oct 15 15:59:49 <rburton> milloni: feel free to turn off delivery of list mails when you've subscribed
Oct 15 16:00:05 <rburton> milloni: meta is entirely in oe-core
Oct 15 16:37:54 <litb> WTF, what could cause this error when executing "bitbake gcc-cross" ?
Oct 15 16:37:55 <litb> https://coliru.stacked-crooked.com/a/09af2585aa89c979
Oct 15 16:38:30 <litb> I looked into the offending recipe, it looks fine. bitbake '-e' doesn't output anything
Oct 15 16:39:00 <rburton> pastebin your log4cpp recipe?
Oct 15 16:39:29 <rburton> or at least, SRC_URI
Oct 15 16:40:22 <rburton> and  -e should output *something
Oct 15 16:40:43 <JaMa> at least the same error when it cannot parse anything
Oct 15 16:41:34 <rburton> SRC_URI = "http://downloads.sourceforge.net/${BPN}/${BP}${PRC}.tar.gz \
Oct 15 16:41:35 <rburton>            file://fix-pc.patch;striplevel=2 \
Oct 15 16:41:47 <rburton> litb: £10 says your mingw class sets PRC
Oct 15 16:43:36 <litb> hold on, I'll paste log4cpp and my cmake class (although the log4cpp is plain vanilla meta-oe, and I'm not using it. the error happens at parse time, apparently)
Oct 15 16:44:04 <rburton> right
Oct 15 16:44:34 <rburton> the error message suggests that you're setting PRC to includedir or something
Oct 15 16:45:13 <litb> rburton, this is my cmake class: https://coliru.stacked-crooked.com/a/ee4cdf78e23e2619
Oct 15 16:45:30 <litb> possibly I'm doing something illegal there and mess up with some internals. not sure
Oct 15 16:45:44 <rburton> question: what does EXPORT_FUNCTIONS do
Oct 15 16:47:09 <rburton> you can prove its that class by not inheriting it and seeing if it works again
Oct 15 16:47:12 <rburton> i suspect its not that class
Oct 15 16:47:20 <rburton> and that you've something else setting PRC
Oct 15 16:48:03 <rburton> my question is a bit evil and i'm expecting the answer 'i don't know other classes do it so i did'.  remove that line.
Oct 15 16:48:06 <litb> ah I said "after do_path" instead of "after do_patch". maybe it's that
Oct 15 16:48:51 <litb> rburton, I ran into an error like "huh do_modify_toolchain_file is empty in <foobar.bb>" for all recipes of all layers (because I INHERIT that cmake-mingw32 class)
Oct 15 16:48:53 <rburton> you can do addtask entirely in code, so i'd just add the task if the conditions are right, instead of trying to remove it again if not
Oct 15 16:48:59 <litb> when I added the export, the error went away
Oct 15 16:49:22 <rburton> i did wonder why your function and the task don't match
Oct 15 16:49:22 <litb> rburton, ah I had no idea. I'll look into how to make the addtask using code
Oct 15 16:49:35 <rburton> bb.build.addtask :)
Oct 15 17:08:45 <litb> rburton, ah damnit, it's the RC variable
Oct 15 17:09:12 <litb> because meta-mingw does  WINDRES_mingw32 = "${HOST_PREFIX}windres --include-dir=${STAGING_INCDIR}" and then  RC_mingw32 = "${WINDRES}"
Oct 15 17:17:48 <litb> IMO they should use PR instead of RC for that in log4cpp.
Oct 15 17:43:04 <litb> rburton_, many recipes definitely need to be patched to make use of ${EXEEXT}, ${SOLIBS} and ${SOLIBSDEV}. often that's the only reason why one would need a bbappend in the mingw layer, I suspect
Oct 15 17:43:28 <litb> so I thought I can do this without patching oe-core, but I think I need to
Oct 15 17:45:29 <SeanSill> Hi all, I'm having trouble getting yocto to regenerate a fitImage when I change something
Oct 15 17:46:07 <SeanSill> Currently trying to get it to identify a xilinx generated device tree that works with an SD card boot
Oct 15 18:01:23 <SeanSill> None of the clean args seem to help
Oct 15 18:01:23 <rburton_> litb: solibs etc should just be sent to oe
Oct 15 18:01:23 <rburton_> litb: also PR is *package* revision. that string should be in PV.
Oct 15 18:01:23 <JPEW> litb: I don't see any SOLIBS or SOLIBSDEV changes in meta-mingw FWIW
Oct 15 18:01:23 <litb> JPEW, SOLIBS is for the target and used in FILES_<...>. Not sure what one should use for the SDK machine OS, but meta-mingw is an SDK machine layer
**** BEGIN LOGGING AT Tue Oct 15 18:02:35 2019
Oct 15 18:28:51 <rburton_> oh right, the old recipe used PR
Oct 15 18:29:07 <rburton_> in master it uses PRC, which is why when i pasted hours ago it had PRC :)
Oct 15 18:29:15 <rburton_> litb need to pick meta-oe bfe144b038ee49290bed7b7d851bca52d340af2e
Oct 15 18:29:21 <rburton_>     This recipe cannot be parsed when using meta-mingw because ${RC} expands
Oct 15 18:29:21 <rburton_>     to the Windows resource compiler. Let's use PRC instead to avoid that
Oct 15 18:29:23 <rburton_>     problem.
Oct 15 19:28:05 <mischief> does anyone have advice on integrating dm-verity with the yocto image classes?
Oct 15 19:48:11 * armpit hmm, earthquake 
Oct 15 19:56:01 <rburton_> mischief: pretty sure i saw a slide deck about that recently, tried google?
Oct 15 19:56:40 <mischief> a bit yea, not much has appeared. i saw one slide deck that only complained about it being 'difficult' -_-
Oct 15 19:57:02 <mischief> right now i'm seeing if i can just add another image type
Oct 15 21:07:03 <neverpanic> mischief: I've seen a verity image type
Oct 15 21:07:15 <neverpanic> not sure whether we implemented that ourselves, though.
Oct 15 21:35:45 <mischief> hm. i have a class now that does dm verity as a 'image type', but i am not sure how i can use an artifact in IMGDEPLOYDIR when making my initramfs
Oct 15 21:49:21 <yocti> New news from stackoverflow: Yocto Conflict between attempted installs <https://stackoverflow.com/questions/50664736/yocto-conflict-between-attempted-installs>
Oct 15 22:49:30 <yocti> New news from stackoverflow: How to conditionally install and ship files in yocto bb recipe? <https://stackoverflow.com/questions/58114055/how-to-conditionally-install-and-ship-files-in-yocto-bb-recipe>
Oct 15 23:44:31 <mischief> uh, how do i fix a dep loop between systemd and cryptsetup :(
Oct 15 23:52:11 <mischief> oh. meta-openembedded master fixed it i guess
**** ENDING LOGGING AT Wed Oct 16 02:59:57 2019