**** BEGIN LOGGING AT Wed Jan 17 03:00:02 2018 Jan 17 06:47:25 hello, is there already someone working on integrating systemd 236? Jan 17 06:49:08 Hi, guys! Has somebody worked with linux-dummy? I can't find any instructions... Jan 17 07:53:48 Hello, how one can build for a particular machine defined in meta-fsl-arm from some other custom layer ? Jan 17 08:40:33 how do i disalbe gcc optimisation for one package .. seems -O1 pr -O2 or inroduced somewhere Jan 17 10:12:39 hello how can I add headers/libs from some packages which are not in rootfs ? (bitbake -c populate_sdk) Jan 17 10:12:49 packages into sdk Jan 17 10:40:56 jozo: http://www.yoctoproject.org/docs/current/sdk-manual/sdk-manual.html#sdk-adding-individual-packages Jan 17 10:56:56 New news from stackoverflow: Bitbake the latest Gstramer in Yocto Morty Jan 17 11:00:25 rburton: Hello, I am working with Rocko and using systemd 234. Although systemd provides a --disable-hwdb option via autotools-generated configure script, it is not among the PACKAGECONFIG options in the bitbake recipe. How do I properly disable hwdb functionality in systemd and avoid including the udev-hwdb package? Jan 17 11:01:35 rburton: it became apparent to my colleague that hwdb has a significant negative impact on boot time for our system and does not look to be necessary for our needs Jan 17 11:03:25 eduardas_m: its a recommends so use BAD_RECOMMENDS Jan 17 11:03:33 erm, BAD_RECOMMENDATIONS even Jan 17 11:03:52 (or add the packageconfig and send a patch) Jan 17 11:05:16 rburton: are you the one actually maintaining the systemd recipe? You seem to be listed as the OpenEmbedded Core/User Space maintainer Jan 17 11:05:28 no, i'm not Jan 17 11:05:37 RECIPE_MAINTAINER_pn-systemd = "Chen Qi " Jan 17 11:09:56 rburton: thank you. was not really aware of the maintainers.inc file Jan 17 11:10:29 Seems like packages using 'inherit pkg-config' have 'bash' as a dependency, which includes bash in the final target images. Is this expected? I think this behaviour is new in Yocto 2.3/2.4 Jan 17 11:10:55 ttllkk: easy to prove, can you do so? Jan 17 11:11:39 what do you mean 'do so' ? Jan 17 11:11:45 prove it Jan 17 11:12:03 your assertion is vague Jan 17 11:12:05 No, not suer Jan 17 11:12:18 That's why I said 'Seems like', I am not suer.. Jan 17 11:12:23 "sure" Jan 17 11:12:35 I was checking packages depending on bash by: 'cat recipe-depends.dot | grep "\-> \"bash" | grep -v "\-completion"' Jan 17 11:12:38 and I get the list Jan 17 11:12:53 build depends != runtime depends Jan 17 11:13:32 I have a package 'p' that is listed with that command, as in ' "p" -> "bash" ' Jan 17 11:13:43 is there a way I can get the root cause of that dependency? Jan 17 11:13:51 I tried with 'bitbake -e p | less ' and having a look Jan 17 11:14:01 that would tell you if it were in DEPENDS, yes Jan 17 11:14:13 I check that for example, but it is not there Jan 17 11:14:20 if you can replicate this with a package in oe-core then others would be able to help Jan 17 11:14:30 what is the actual problem though Jan 17 11:14:35 as i said, build depends != runtime depends Jan 17 11:14:58 The final problem is that 'bash' is ending up in the target, and I would like to get rid of it. Jan 17 11:15:14 So I assume the reason for that is that there is a package that *runtime depends* on bash, not *depends* on bash? Jan 17 11:15:26 so run rpm -e bash or whatever on the target to see what it says is stopping it being removed Jan 17 11:15:39 (easiest way to find out what the dependency was) Jan 17 11:15:53 Ouch, I didn't think of that Jan 17 11:15:55 thanks! Jan 17 11:15:58 I will try that Jan 17 11:16:23 there are tools to inspect dependency trees but that is so much easier Jan 17 11:16:30 Seems obvious now. I was trying to get the informatino just from Yocto. I am sure there is a way. Jan 17 11:16:35 But checking on the target is good enough for me Jan 17 11:16:50 error: Failed dependencies: /bin/bash is needed by (installed) rpm-1:4.13.90+git0+a8e51b3bb0-r0.0.cortexa9hf_neon bash is needed by (installed) rpm-1:4.13.90+git0+a8e51b3bb0-r0.0.cortexa9hf_neon Jan 17 11:16:58 there you go Jan 17 11:17:04 that would mean that rpm runtime depends on bash? Jan 17 11:17:06 yes Jan 17 11:17:35 so either 1) switch to opkg 2) dont ship package management tools 3) look at why rpm needs bash, and see if the scripts can be ported to pure sh Jan 17 11:18:24 However, should bash appear here, then? Jan 17 11:18:39 it appears there for a reason: rpm is shipping a bsh script Jan 17 11:18:40 https://pastebin.com/Dka8aTMz Jan 17 11:18:49 or could it be that it is an indirect dependency? Jan 17 11:18:57 (shouldn't rpm report the direct dependency in that case?) Jan 17 11:19:24 no, as some dependencies are only know at build time (another example is every library dependency) Jan 17 11:20:08 I see Jan 17 11:20:11 thank so much for all the help! Jan 17 12:10:33 fontforge-native is now failing with "/bin/bash: msgfmt: command not found" even when it inherits gettext, any idea what might be wrong? Jan 17 12:15:34 JaMa: damnit! Jan 17 12:16:02 JaMa: needs explicit dependency on gettext-native Jan 17 12:16:29 native disables NLS so just has gettext-minimal-native in the depends Jan 17 12:19:55 Marex: hello Jan 17 12:20:07 Marex: please provide the u-boot upgrade for 2018.01 Jan 17 12:28:15 rburton: thanks, will try that Jan 17 12:31:35 otavio: heh, git mv no longer works ? :) Jan 17 12:31:46 I saw the autoupdater report ... Jan 17 12:33:00 Did the auh emails make it to the mailing list? Jan 17 12:55:17 RP: it was cc'd but i suspect they're all in moderation? Jan 17 12:55:50 i got one for my recipe that was attempted which cc'd core but none of the others Jan 17 13:02:55 Hello everyone Jan 17 13:03:20 I want to use the linux uio driver with yocto Jan 17 13:03:41 do i have to add it to the build? Jan 17 13:04:05 if so, how? Jan 17 13:08:47 RP: rburton: it's bizarre, they do have oe-core list in CC, but they aren't actually sent there. Only the overall report is. Jan 17 13:09:02 I'm not convinced we want them on the list in the first place though Jan 17 13:09:59 kanavin: tend to agree Jan 17 13:10:00 kanavin: Iwonder if its a moderation or subscription issue Jan 17 13:10:07 summary yes, patches no Jan 17 13:10:23 it would be good to see them *somewhere* but probably not oe-core Jan 17 13:11:52 RP: I tend to think it's some kind of bug in AUH, as both the updates and the overall report are sent from the same email address Jan 17 13:12:24 kanavin: yes, I wondered why some made it and some didn't Jan 17 13:14:41 RP: I did ask halstead to whitelist auh@auh.yoctoproject.org for oe-core, and at least for the overall reports it works as intended Jan 17 13:22:41 kanavin: right, that makes sense Jan 17 13:57:24 New news from stackoverflow: CMake cannot generate a safe linker search path - yocto 2.4 Jan 17 14:16:30 I have built an x86_64 image using the meta-intel layer and am trying to run it in a VMWare virtual machine but I am getting stuck. I have both a .hddimg and a .wic output file (I'm not sure what the difference is in disk layout) and for both I have tried directly converting them to vmdk images. This only gave a bootable result for the hddimg file, but it got stuck during boot for whatever reason. Jan 17 14:17:14 I have also tried using those files as virtual removable media but the VM would not boot from either of them Jan 17 14:57:35 New news from stackoverflow: ttyAMA0 serial port RaspberryPI3 Jan 17 15:12:30 * kergoth yawns Jan 17 15:27:22 * nayfe yawns too Jan 17 16:08:00 rburton: I did find the make error and sent a v4 for it Jan 17 16:08:10 yeah saw, thanks Jan 17 16:19:06 rburton: I tried your ninja patch, needs rebase Jan 17 16:19:17 rburton: I agree with the patch, for sure. Jan 17 16:22:13 I'm checking with the gnuradio ninja enthusiasts Jan 17 16:27:48 New news from stackoverflow: Linux UIO driver include Jan 17 17:02:21 rburton, for ninja, is it easy to tell cmake to use make if you have a project with some custom make? Jan 17 17:02:45 Crofton|work: can you do that? Jan 17 17:03:32 dunno :) Jan 17 17:03:47 mainly relaying a reamrk from one of the gnuradio guys Jan 17 17:04:12 apparently one of them ran into a case where they had to use make on some "library" Jan 17 17:04:45 I can look at it, but fighting banks today Jan 17 17:04:54 Crofton|work: can you ask them what library? Jan 17 17:07:48 rburton, he is in #oe Jan 17 17:07:54 ETOOMANYDUPE CHANNELS Jan 17 17:13:16 when i pull yocto poky, can do i know what version of the gnu toolchain will be used as part of poky? i.e. is it consistent, set in stone, or can it vary based on major version? Jan 17 17:15:03 zarzar: its defined in the metadata. A given version of poky uses a specific version of the toolchain Jan 17 17:15:28 RP: great! Jan 17 17:15:48 RP: can you point me to said metadata for reference? Jan 17 17:16:35 zarzar: http://git.yoctoproject.org/cgit.cgi/poky/tree/meta/recipes-devtools/gcc - gcc recipes used by the master branch ? Jan 17 17:19:17 RP: ok thanks, to be sure, that is for building yocto, not the toolchain deployed in image? Jan 17 17:20:10 zarzar: cross bits are building yocto, "plain gcc" bits are target Jan 17 17:20:23 RP: thanks! Jan 17 17:47:45 JaMa: for the host-contamination error you just do a cleansstate for each affected package, as workaround? Jan 17 18:10:34 Hello, I've created a custom Yocto image and my drive is found to be 0 bytes in size when attached to the image but has the correct size in my popular linux image machine. I'm probably missing an important module. Any ideas on how to discover the missing module? Jan 17 19:21:14 Hi, how to apply a patch after my do_configure_prepend() ? Is this even possible? Jan 17 19:22:07 why would you want to? Jan 18 01:26:07 Hello, How do I inherit a class based on some conditions Jan 18 02:29:18 New news from stackoverflow: How to recover Intel Edison board **** ENDING LOGGING AT Thu Jan 18 03:00:01 2018