**** BEGIN LOGGING AT Mon Mar 20 03:00:01 2017 Mar 20 03:46:54 Hello there, I would like to ask how I can use my library at other component. Mar 20 03:47:44 I am trying to link my library but I cannot find a way. It just shows NO LIB Mar 20 03:50:17 bkim: how is your package organized ? Mar 20 03:50:19 what exactly are you trying to do? can you pastebin an actual error? Mar 20 03:50:24 are they in two different recipes ? Mar 20 03:50:45 Yes two recipies Mar 20 03:51:08 * nerdboy had to sed his own autotools-based lib recipe... Mar 20 03:51:19 I made a library that is simply containint a function to print out on console Mar 20 03:52:09 I built it as shared library and checked that libhello.so Mar 20 03:53:12 In another recipies and Cmakelist, I tried to link and find the library with target_link_libraries and find_package Mar 20 03:53:18 can you pastebin the failing log and the rcipe? Mar 20 03:53:43 Ok I will share in 15 min Mar 20 03:53:53 oh yay, cmake... Mar 20 03:54:26 Thanks in advance. I couldnt find any clue in a number of google search TT Mar 20 03:54:38 are you inheriting cmake? Mar 20 03:54:54 yes Mar 20 03:54:58 grep the recipes in your build layers Mar 20 03:55:15 often better than google... Mar 20 03:56:07 mr khem knows more about the internals but i've hacked a recipe or two Mar 20 03:59:05 i normally only touch cmake when i have to fix it Mar 20 03:59:29 and pretty much everyone does it different Mar 20 04:01:04 I am quite new at cmake and bitbake. From last week I ve stuyied alone with google and official manual but the thing is quite now easy to understand at once. Mar 20 04:02:02 essentially you need to tell your find_package stuff to look in the sysroot Mar 20 04:02:50 I need to write PACKAGE += ${PN} on bb? Mar 20 04:03:15 so library seems to have ${PN}-dev Mar 20 04:03:30 So I need to make that as well Mar 20 04:05:52 there's some auto-lib stuff in there Mar 20 04:06:18 depending on package naming and inherits i think Mar 20 04:06:51 * nerdboy waits for khem to jump back in Mar 20 04:07:29 did you name it libhello by any chance? Mar 20 04:08:05 I am preparing my recipies and cmake Mar 20 04:10:13 maybe set LEAD_SONAME = "lbhello.so" ? Mar 20 04:10:47 only if there's more than one lib, otherwise that's not going to accomplish anything AFAIK Mar 20 04:24:49 My company blocks irc :( Mar 20 04:27:38 * nerdboy has a fairly shiny autotools example Mar 20 04:27:44 just not cmake... Mar 20 04:28:25 Thanks guys anyway. Mar 20 04:28:34 I ll share if i success Mar 20 04:58:53 bluelightning: bitbake -p 299.50s user 2.92s system 662% cpu 45.648 total Mar 20 06:52:11 I have a problem with using gdb remote Mar 20 06:52:39 I have run the bitbake meta-ide-support Mar 20 06:53:52 but the sysroot it gives me still strip all the symbols https://paste.fedoraproject.org/paste/HE5tWoxbqP93jzUEPCJjLl5M1UNdIGYhyRLivL9gydE= Mar 20 08:09:57 Is it possible to make a bbappend that is applied to a package when building core-image but not when building with -c populate_sdk ? Mar 20 08:10:38 Something like: if !sdk SRC_URI += ... endif Mar 20 08:11:41 or perhaps just if !native SRC_URI += ... endif Mar 20 08:12:04 (If it is even possible to use "if" at all in the bb files...) Mar 20 08:28:39 deva: you can try with _append_class-nativesdk Mar 20 08:29:54 mborzecki, Does that work with _remove as well? Mar 20 08:30:38 that would be my guess :) Mar 20 08:30:49 Because then I could do a SRC_URI += "foo"; SRC_URI_remove_class-nativesdk = "foo" Mar 20 08:31:15 I'll give it a shot Mar 20 08:45:53 mborzecki, It appears to work. Is there a class for the "regular" image as well? Mar 20 08:46:17 I have an install_append which I need to remove (or not append) to the nativesdk Mar 20 08:54:32 i suppose you can have a function that does common part of install eg. do_common_install() and then do_install() { do_common_install; extra stuff} and do_install_class-nativesdk() { do_common_install } Mar 20 08:55:55 I am currently trying '["$CLASSOVERRIDE" == "class-nativesdk"] || foo' for each of the lines in my do_install_append Mar 20 08:56:02 now, whether overriding do_install for sdk is a good thing, I don't know, did a quick grep, only dbus recipe defines do_install_class-nativesdk Mar 20 08:56:11 That should execute foo only if the class is not nativesdk Mar 20 08:57:12 I can't set do_install directly, because I am in a bbappend and need the original do_install to still be executed Mar 20 08:57:36 I suppose there is no way of calling the "super class" do_install from a bbappend recipe? Mar 20 08:57:39 hm maybe you can fix the upstream recipe then? Mar 20 08:57:56 or just copy the recipe to your layer Mar 20 08:59:09 It's not a fix, but a customization that is only needed on the target which breaks tyhe nativesdk Mar 20 08:59:42 Copying it would work, but Id' rather stick as close to upstream as possible Mar 20 09:01:15 well if install is not that complicated you can define do_install_class-nativesdk and perform all the steps yourself Mar 20 10:27:59 hm mingw broke again Mar 20 10:29:09 rburton: there appears to be something in -next/mut doing this :/ Mar 20 10:29:53 theres very little in next, going to fire a master on that builder to verify Mar 20 10:32:18 kanavin: https://autobuilder.yocto.io/builders/nightly-oe-selftest/builds/218/steps/Running%20oe-selftest/logs/stdio looks like your fault Mar 20 11:33:14 khem: your ' binutils: Enable threading when gold is enabled and is not default linker' patch broke builds for windows msd Mar 20 11:33:19 erm, windows sdk Mar 20 11:48:07 rburton: did you test different target on purpose? (qemux86 vs qemux86-64) Mar 20 11:49:13 jku: the selftest? tbh i just saw the last error was gpg signing,and alex was changing that Mar 20 11:50:00 rburton: oh I thought you were testing the binutils failure Mar 20 11:50:20 oh right Mar 20 11:50:23 yes, i was Mar 20 12:31:02 rburton: just remembered: kanavin isn't working today Mar 20 12:31:35 ah Mar 20 12:34:40 I'll take a look, maybe it's not complex Mar 20 12:35:40 thanks Mar 20 13:59:03 marquiz: https://bugzilla.yoctoproject.org/show_bug.cgi?id=11191 -- I guess your rpm signing commit isn't related but can you still have a look? Mar 20 13:59:04 Bug 11191: normal, Undecided, ---, alexander.kanavin, NEW , test_signing_packages failed on AB Mar 20 14:09:41 Something in lvm2 is currently broken, some RDEPENDS_append seems to lack a leading space: Mar 20 14:09:41 ERROR: Nothing RPROVIDES 'lvm2-native-scripts-native' (but virtual:native:/fast/work/intel-iot-refkit//meta-openembedded/meta-oe/recipes-support/lvm2/lvm2_2.02.166.bb RDEPENDS on or otherwise requires it) Mar 20 14:09:41 NOTE: Runtime target 'lvm2-native-scripts-native' is unbuildable, removing... Mar 20 14:10:07 The problem is I can't even examine that problem with "bitbake -e lvm2-native", because that just ends up printing the exact same error :-/ Mar 20 14:12:54 Or is it the native.bbclass which fails to handle RRECOMMENDS_lvm2="lvm2-scripts (= 2.02.166-r0)" (from the target recipe)? Mar 20 14:17:03 Yes, that's it. Is the right solution to set RDEPENDS and RRECOMMENDS for extra packages only for the target? Mar 20 14:17:20 Because those extra packages do not exist in the non-target cases? Mar 20 14:36:59 Hey guys, I just discovered that if an image with SysV init is so minimal that "modutils-initscripts" is not included, it won't be automatically included if KERNEL_MODULE_AUTOLOAD is not empty.Should I report a bug, or is this expected? Mar 20 15:59:14 RP: renaming to -initial, now its gone from all dependencies even from packages where its saying DEPENDS += "go-native-initial" Mar 20 15:59:24 RP: there seems to be more to magic Mar 20 16:30:21 khem: https://autobuilder.yocto.io/builders/nightly-musl/builds/197 well it happened again Mar 20 16:34:48 khem: hmm :/ Mar 20 16:35:00 khem: it may be that -initial is only used by -cross recipes :/ Mar 20 16:52:39 RP: I sent a different fix, found that there were two scripts in conflict renamed them for bootstrap version that should do it Mar 20 16:53:59 RP: quick test shows it holds, I can still build go for rpi3 and run it on target to build packages after this fix Mar 20 16:54:34 but I never had that race happen here for my images since I do not use go-native for packages Mar 20 16:55:00 lets see if this works for mender folks Mar 20 16:55:06 khem: ok Mar 20 17:20:56 Hmm. Yocto bug #9859 is a pseudo thing, and I know the change is merged, but I don't remember whether it's in the branch currently in Yocto. Mar 20 17:20:57 Bug https://bugzilla.yoctoproject.org/show_bug.cgi?id=9859 normal, Medium, 2.3 M3, seebs, IN PROGRESS IMPLEMENTATION , setcap doesn't work Mar 20 17:21:00 If it is, the bug can be marked out. Mar 20 17:22:21 joshuagl: ^^^ do you happen to know? Mar 20 17:35:49 devtool is slowww.. devtool modify -x opencdmi 17.21s user 8.48s system 14% cpu 2:58.04 total Mar 20 17:36:02 it took 3mins Mar 20 17:36:18 for a tiny package whose sources are fetched in dl_dir already Mar 20 17:36:22 I wonder what it it doing Mar 20 17:54:42 khem: would be interesting to know... Mar 20 17:56:21 paul just filed a bug which basically said "it does too much work" Mar 20 17:56:26 so he appears to know its a bit slow Mar 20 17:56:46 khem: did it refetch? Mar 20 17:57:40 -ETOOMANYPAUL Mar 20 17:57:48 paule :) Mar 20 17:57:58 too many blue* too Mar 20 17:57:59 figured as much :) Mar 20 18:26:40 RP: yeah I discussed briefly with him yesterday Mar 20 18:27:03 RP: it was sitting in do_fetch for long Mar 20 18:27:10 its a small package from gh Mar 20 18:27:21 even a refetch should have been quick Mar 20 18:37:44 -ETOOMANYPAUL seems to be a common error Mar 20 18:46:47 we should just limit to using nicks to avoid it may be Mar 20 19:08:11 seebs: RP: pretty sure that setcap was in pseudo 1.8.2 which we have in OE-Core Mar 20 19:11:42 okay, then that bug is probably in the wrong state. I don't have a live yocto setup right now, though. Mar 20 19:15:54 seebs: you can mark as resolved → fixed and assign to me, I'll verify it tomorrow Mar 20 21:53:44 hi there Mar 20 21:54:00 I know there is smart logic in creating dsk.xz files that Yocto currently produces Mar 20 21:54:22 I have 16GB USB stick with live Ubuntu on it, I would love to dump that stick into .dsk.xz file (compressed one) Mar 20 21:54:33 would anybody here know how to do that? Mar 20 21:55:17 I played with clonezilla for some time, but I figured it doesn't do iso files Mar 20 21:55:29 iso/dsk Mar 20 22:42:20 RP: i sent the first batch of fixes in yp compatible script, this is my working branch http://git.yoctoproject.org/cgit/cgit.cgi/poky-contrib/log/?h=alimon/yp_compatible Mar 20 22:42:42 RP: remains the isolate environment and able to display signature detail diffs **** ENDING LOGGING AT Tue Mar 21 03:00:01 2017