**** BEGIN LOGGING AT Sat Apr 04 03:00:20 2020 Apr 04 09:34:57 khem: thanks, the build did fail badly. Will requeue Apr 04 14:40:08 Hi! I'm using yocto warrior inside a docker image in order to create an image for my board. I want to create a recipe in order to compile a software of mine. The software is very big and it is based on qt5.8. Usually I compile it using a docker and a qmake command. Here, how can I compile my soft using a yocto recipe? :) Apr 04 15:02:43 alexb3600: http://layers.openembedded.org/layerindex/branch/master/classes/?q=qmake&search=1 Apr 04 15:05:43 Crofton|road: Hum so I just need to download the layer then I build qmake with bitbake and then in my recipe I will be able to do the command Apr 04 15:05:49 Am I right? Apr 04 15:07:17 Yeah basically, you'll need meta-qt5 layer, not sure what qt version is there, and mayb elook for some qmake based recipes to give you ideas Apr 04 15:08:04 But This should install qmake on the target no? Me I just need qmake for the yocto recipe. Ido not want qmake on the target Apr 04 15:14:39 there classes should help you use wmake in a recipe Apr 04 15:30:33 okay I will look at it, thanks :) Apr 04 16:22:43 I am fighting this the whole day, can somebody help please?, I am trying to remove this file : ${D}${sysconfdir}/systemd/system/getty.target.wants/getty@tty1.service Apr 04 16:23:54 tried with do_install_append() in sytemd_%.bbappend, but it says: cannot remove '...build/tmp/work/cortexa7t2hf-neon-vfpv4-poky-linux-gnueabi/systemd/1_243.2-r0/image/etc/systemd/system/getty.target.wants/getty@tty1.service': No such file or directory Apr 04 16:24:32 I couldn't figure out where this file being created :( Apr 04 16:39:17 timemaster: oe-pkgdata-util is your friend Apr 04 16:43:14 LetoThe2nd: haha, another reading..:) thanks for this clue, lets see what is it Apr 04 16:44:57 ok, this is probably not what I am looking for Apr 04 16:46:52 because this search for files, but this one is tricky because it is a symlink and presumably automatically created, somehow somewhere Apr 04 16:50:00 New news from stackoverflow: Python modules missing in compile-time - Yocto Apr 04 16:55:48 so just to confirm, oe-pkgdata-util didn't help, it showed only file getty.target.wants/serial-getty@ttyS0.service, which I would expect to be very similar issue, but is not :( that's why I am lost for so long Apr 04 16:57:35 ERROR: Unable to find any package producing path /etc/systemd/system/getty.target.wants/getty@tty1.service Apr 04 17:05:30 RP: sigh, another subtle issue on rebuilds where the filter is not effective, sending a v4 now Apr 04 17:05:34 * khem ducks Apr 04 17:05:42 timemaster: maybe related to this? https://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta/recipes-core/systemd/systemd_244.3.bb#n105 Apr 04 17:05:49 (can't really dig here at the moment, sorry) Apr 04 17:08:20 no problem, thanks for any help.. Apr 04 17:09:01 unfortunatelly not, if this is related, I didn't find how, this is for the other file called serial-getty@ttyS0.service, which is ok Apr 04 17:09:03 hi, I'm adding recipe for this repo: https://github.com/logimic/shape all seems to be find just *.so libraries are put to -dev package ant not to main package Apr 04 17:09:08 is there any reason for that? Apr 04 17:09:20 if they are not linked with binary then are moved to -dev package? Apr 04 17:10:15 but now I am pretty sure, this file is created by something in the distro, like systemd or agetty, because everything I tried so far ends up with rm: cannot remove, file not found... Apr 04 17:10:48 My last attemp was to create a class with a task and chain this task after do_rootfs Apr 04 17:11:09 and even there it failed with file not found error Apr 04 17:12:33 timemaster: i'd rather guess that this is a systemd default. Apr 04 17:12:53 no, I take it back.. if I open the filesystem image (tar.bz2) the file is there.. f*ck :( Apr 04 17:14:52 LetoThe2nd, yes, it seems like a agetty default or sth like that.. :( Apr 04 17:21:02 timemaster: have you outright grepped the work directories for it? Apr 04 17:21:20 (thats a simple idea i have without much manuel searching) Apr 04 17:22:45 yes, I tried, I was really desperate.. :) no info about this file Apr 04 17:23:15 the name is probably compiled during some script run Apr 04 17:37:27 but it should show up i a grp then, (not in a find, probably) Apr 04 17:37:40 whats the deeper reason behind removing it anyways? Apr 04 18:18:17 I know what you mean, I've tried both, find and recursive grep.. strange Apr 04 18:19:39 This service breaks my weston image, the image boots up, shows weston desktop and in a few sec it will switches to tty1 because of this service Apr 04 18:20:09 and then after some time, switches back to weston.. once I removed this file on the target manually, the behaviour is ok.. Apr 04 18:21:08 I think this service is not needed at all because systemd starts up agetty on demand anyway Apr 04 18:31:40 Finally moving a little, I was grepping badly :( DEBUG: Executing shell function systemd_preset_all, ln -s /lib/systemd/system/getty@.service ,,,,system/getty.target.wants/getty@tty1.service Apr 04 18:34:49 which is in IMAGE_PREPROCESS_COMMAND and runs systemctl of course.. Apr 04 18:44:57 RP: icu patch v4 is good, I was trying filters with it which are now working seemingly Apr 04 18:52:09 RP: I was seeing musl error, which I sent a separate patch on top Apr 04 18:57:56 LetoThe2nd, solved for now with a custom function written as bbclass and then included in the image definition, not ideal but at least it works :) thanks! Apr 04 18:58:08 :) Apr 04 18:58:27 timemaster: care to write a small rundown and send to the ml, so we can properly fix it for weson? Apr 04 18:59:42 sure, will do.. Apr 04 18:59:51 thanks! Apr 04 19:00:23 no problem Apr 04 19:00:25 :) Apr 04 19:02:50 Hey Apr 04 19:03:13 Did someone out there manage to get golang support into the SDK with poky master? Apr 04 19:05:12 As a first shot I added nativesdk-go nativesdk-go-dep nativesdk-go-runtime but like this default golang modules like fmt can not be found. Apr 04 19:05:15 LetoThe2nd: which ml would you prefer? Poky,oe-core? I am not sure.. Apr 04 19:05:27 poky Apr 04 19:05:34 ah crap, i mean yocto Apr 04 19:05:49 yocto, all right then Apr 04 19:05:53 thanks Apr 04 19:05:56 np Apr 04 19:30:01 khem: ok, I'll need another testing cycle I guess :/ Apr 04 20:10:05 RP: I will send v5, in few minutes, I think one issue is that I should have perhaps used BUILD_CC and tools Apr 04 20:10:13 its failing on one of ubuntu hosts Apr 04 20:10:20 old ones Apr 04 20:39:01 khem: ok, I'd not got to the v4 yet ;-) Apr 04 21:43:24 RP: v5 is already out :) Apr 04 21:43:51 I think I need to take deep breathe Apr 04 21:45:15 but this is very important since icudata is 27M and second largest shared library after browser Apr 04 21:46:15 so kind of trying to get the baseline into core before release Apr 04 21:57:16 ah mips bites again Apr 04 22:46:26 RP: is there some way where I can ask tar to overwrite existing files ? Apr 04 22:47:11 I have two tarballs, in src_uri second one has newer files which should be used instead of the ones in original tarball Apr 04 22:53:25 khem: extract to a different subdir, them move the files over yourself? Apr 04 22:53:49 khem: I'm getting worried about the risk from icu in the release now :/ Apr 04 22:54:37 We are patching a Makefile.in inside the data/ so I have to defer a patch after extract Apr 04 22:54:41 its a simple change Apr 04 22:55:19 khem: I mean that we're on a v6 of the patchset and seeing issues :/ Apr 04 22:56:17 I should wait until my slow mips builder finished before sending Apr 04 22:56:53 I am testing it with and without filter so combinations are a bit overwhelming Apr 04 22:57:05 khem: I've added v6 to the autobuilder, was that right? Apr 04 22:57:12 but I am seeing end of it now Apr 04 22:57:29 no wait for v7 which I am preparing as we speak Apr 04 22:58:01 khem: let me stop that build Apr 04 22:58:21 RP: sorry for being so crazy today Apr 04 22:58:53 khem: it matches the rest of the world atm! Apr 04 22:59:01 RP: http://sprunge.us/75yqy7 this is the diff over v6 Apr 04 23:01:23 khem: ok, I can fire a build with that Apr 04 23:03:11 RP: sent Apr 04 23:03:22 take v7 Apr 04 23:03:38 and I will pray nothing falls over this time Apr 04 23:03:42 khem: musl patch still needed? Apr 04 23:03:47 yes Apr 04 23:04:04 thats independent even if we dont take the other change its required Apr 04 23:06:16 RP: so its working ok on arm and mips as of now on musl and glibc Apr 04 23:06:34 rest will run through full autobuilder here and see tomorrow as well Apr 04 23:06:41 looks ok thus far Apr 04 23:07:32 khem: new build started on the AB too Apr 04 23:09:21 * RP -> sleep Apr 04 23:18:00 gn Apr 04 23:20:44 hello...I m PXE booting my yocto distro. However it asks for password at login. The same kernel rootfs combination when booted from ssd does not require password for "root" user. Where is the issue Apr 04 23:53:35 RP: icuinfo says ICU Initialization returned: U_ZERO_ERROR :) on qemumips so I can now sleep in peace Apr 04 23:54:56 kiwi_29: Are you sure its mounting exact samee rootfs in both cases ? Apr 04 23:56:55 khem ..yes. however I dug more and the kernel boot logs for pxe has numerous systems services failing ... in the end the login service fails too Apr 04 23:56:59 I wonder if that is the problem Apr 05 00:04:05 seems so Apr 05 00:04:17 you need pam and login working Apr 05 00:07:05 I m debugging that...will post here when I find more ...however I do get a login prompt btw... khem **** ENDING LOGGING AT Sun Apr 05 02:59:58 2020