**** BEGIN LOGGING AT Tue Jun 12 03:00:05 2018 Jun 12 04:31:49 Is there a way to create wic images with rootfs > 4 GB automatically? I'm using meta-intel, targetting a NUC. Currently it's failing due to using hddimg as a middle step, which is limited to 4 GB due to using a vfat partition in an intermediate step in the image creation process. Currently I just generate an ext4 file instead and use the wic tools manually after bitbake finishes. Jun 12 06:30:38 erbo_: you can override which images are created by having IMAGE_FSTYPES = "wic" Jun 12 06:45:02 anujm: yes, but image type wic seems to use hddimg as a middle step and because of that it's limited to 4 GB rootfs Jun 12 06:49:05 but by using IMAGE_FSTYPES = "ext4" and running wic manually you don't need use the hddimg step, so then you can have larger rootfs. I figured that someone might already have seen the same issue and found a solution. Otherwise I guess I would have to either have to keep doing the last step manually, or automate it in either a new fstype or rework the original wic one. Jun 12 07:02:21 erbo_: i don't think that it should be doing that. x86 images build hddimg along with wic by default. so you will have to override image_fstypes. Jun 12 07:32:33 New news from stackoverflow: Wifi of Bluetooth for synchronizing devices [on hold] Jun 12 07:42:36 hello. i am wondering why my self-contained yocto sdk is 5GB huge.. why is yocto installaing tools shipped in the runtime image also into the sdk, and then stripped? And why is yocto installing stipped libs and debug libs? should it not be okay to ship only the unstipped libs with the sdk and stipped ones with the runtime image? Jun 12 08:06:00 anyone home? I've a problem with getting libgfortran built with rocko + a backport of a libbacktrace patch. Jun 12 08:06:42 I've this patch on the recipe: https://patchwork.openembedded.org/patch/146744/ Jun 12 08:07:17 and elements of this series to make it find libbacktrace at all: https://www.mail-archive.com/openembedded-core@lists.openembedded.org/msg98442.html Jun 12 08:09:51 Eventually, however, I get an issue when it comes time to link. This particular command (among others) yields an x86-64 ELF object file rather than an ARM compatible one: https://pastebin.com/J6wW3Tda Jun 12 08:11:45 output from bitbake/make complaining about libbacktrace and those native files: https://pastebin.com/80ETMhS2 Jun 12 08:12:21 Never had much to do with libtool before, any suggestions welcome. Jun 12 08:13:25 Oops, should have been more specific regarding versions: I'm trying to get it built for GCC 7.3, and I'm running on the head of rocko branch + the patches mentioned above. Jun 12 08:14:40 (gonna disappear here, if someone has anything useful to contribute, please pop it in here and I'll grab it via a log later). Jun 12 10:07:07 hello, I would like to remove a dependency for a package in a .bbappend: I'm doing RDEPENDS_${PN}_remove = "${PN}-python" but it doesn't work...while if I directly remove it from the original .bb file, it works fine Jun 12 10:07:30 does anyone know what I'm doing wrong ? Jun 12 11:25:27 it seems i'm hitting the same issue as http://lists.openembedded.org/pipermail/bitbake-devel/2018-January/009150.html Jun 12 11:33:22 New news from stackoverflow: Yocto Linux on a virtual machine Jun 12 11:48:42 hey guys! could you please hint me where to poke if I see " undefined reference to `__init_array_end'" when building a shared library recipe? Jun 12 11:49:30 emptying SECURITY_CFLAGS doesn't seem to help ;-( Jun 12 11:50:44 luneff: most obvious question: does it build cleanly outside of OE? most obvious tip: compare the build settings between outside and inside biuld Jun 12 11:54:55 I got similar error when building zeromq in AGL layer set. The solution was to manually edit the libtool script to omit "-lc" from the last step linker flags. Nothing similar in this Qt .pro lib, though ;-( It's building fine for all other platforms and Yocto flavors, though Jun 12 11:55:10 this one is AGL as well Jun 12 11:56:03 ok, no idea. never used that. Jun 12 11:57:40 me neither ;-D Jun 12 12:03:28 New news from stackoverflow: Machine specific layers in yocto Jun 12 12:03:31 hm. putting those SECURITY_CFLAGS = "" on both zeromq and my lib helped me out :-D Jun 12 12:05:09 What is the relationship between github meta-qt5 https://github.com/meta-qt5/meta-qt5, and Qt http://code.qt.io/cgit/yocto/meta-qt5.git/ ? Jun 12 14:03:53 New news from stackoverflow: How to enable/disable a particular bbappend for a specific MACHINE in Yocto Jun 12 16:00:46 armpit: did you just get kicked from the call? Jun 12 16:00:57 yes Jun 12 16:01:15 just as Mr. Halstead showed up Jun 12 16:01:33 rburton: same here Jun 12 16:02:04 i liked how it said 'goodbye' Jun 12 16:02:07 sounded quite terminal Jun 12 16:02:12 hm dialed in but on hold Jun 12 16:02:13 "Please wait for the host to start this meeting" Jun 12 16:02:19 yep Jun 12 16:02:20 halstead: did you break the meeting? Jun 12 16:02:53 i got it briefly Jun 12 16:03:25 * armpit the sync sunk Jun 12 16:03:28 you're in the meeting / first participant / goodbye Jun 12 16:03:34 rburton: I did as well, briefly Jun 12 16:03:42 zoom is great! Jun 12 16:03:50 never thought i'd say 'wish this was an intel bridge' Jun 12 16:04:05 hmm, heard halstead then it dropped me Jun 12 16:04:15 yes, me too Jun 12 16:04:35 halstead is mutally exclusive with everyone else? Jun 12 16:05:24 rburton, I'm not sure what is happening here. When I try to join it says waiting for host to start the meeting. Jun 12 16:05:47 do we blame stephen? :) Jun 12 16:05:57 (mainly as he isn't here to defend himself) Jun 12 16:06:14 rburton, When I switched to the host account and started the meeting it began but kicked me off after a moment. Jun 12 16:06:21 well, anyone got anything to discuss? :) Jun 12 16:07:07 nope. for me, I am one hold for QA Jun 12 16:07:21 I have 6 builders basically ready. Waiting on a long copy to new hard drives to move the storage over and free up rack space. Jun 12 16:07:43 the latest RCs of TPM things are working, so i'll be pushing recipes to meta-secure-core for released versions of TPM as well as _git recipes for the latest updates Jun 12 16:07:52 I'll need a downtime to finalize the storage move. Jun 12 16:08:19 halstead: just let us know when you're ready. 2.6 M1 build is the main thing looming Jun 12 16:09:14 we have a meeting Jun 12 16:09:14 RP: do you have an eta for m1? Jun 12 16:09:14 i'm still working through some additional layers (meta-security-isafw, meta-intel-iot-security, and meta-selinux) Jun 12 16:09:48 in Jun 12 16:11:46 armpit: we're back in Jun 12 16:11:59 armpit, We got it working if you'd like to rejoin. Jun 12 16:18:21 * rburton sprays coffee Jun 12 16:18:26 SIXTY SEVEN TERABYTES Jun 12 16:34:25 New news from stackoverflow: do_rootfs function failed in yocto project Jun 12 20:57:26 i can understand what this error is; https://paste.fedoraproject.org/paste/xuD9SNuqqR0z2ig9y7s0kQ Jun 12 21:03:51 s/can/can't/ **** ENDING LOGGING AT Wed Jun 13 03:00:01 2018