**** BEGIN LOGGING AT Mon Jun 01 02:59:59 2015 Jun 01 06:07:36 hi guys, good morning Jun 01 06:13:04 may I know, in yocto build engine, the rootfs is formed? Jun 01 06:13:38 I need to make some changes into one /etc/securetty during build Jun 01 06:16:02 kanupatar, at least if you're using shadow, securetty comes from the shadow-securetty recipe Jun 01 06:17:17 AndersD: I am using yocto first time man, May I know which file is responsible for this? Jun 01 06:34:37 hi guys Jun 01 06:34:52 may I know, how rootfs is generated after yocto build? Jun 01 06:53:17 kanupatar: you may find this section of the manual useful in explaining how image construction works: http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#image-generation-dev-environment Jun 01 06:57:44 bluelightning: thanks..but my doubt was to see where I can confgure the rootfs properties Jun 01 06:58:09 kanupatar: depends on what "properties" you mean Jun 01 06:58:49 bluelightning: means, configuring the users, login serial access permission etc Jun 01 06:59:18 bluelightning: means, need to configure contents of /etc/securetty during build Jun 01 08:11:54 good morning Jun 01 08:16:17 (UMT) Jun 01 08:18:44 lo Jun 01 08:18:54 is there a plan to include Jun 01 08:19:15 http://lists.openembedded.org/pipermail/openembedded-core/2015-May/105030.html in the fido branch? Jun 01 08:25:09 there is, I just wait for them to land in master first and then backport them Jun 01 08:25:46 it's in master now so will get picked up for the proposed updates I make this week Jun 01 08:27:01 cool Jun 01 08:27:02 thx Jun 01 08:44:02 morning all Jun 01 09:02:33 hello, is there a way to list all packages that are getting installed into the final target root filesystem? Jun 01 09:05:48 ericbutters: you can see it after do_rootfs by looking in the manifest file next to the image Jun 01 09:06:16 thanks Jun 01 09:08:28 bluelightning: in my build there is no manifest file right now, but i know that i got such a file with other builds.. Jun 01 09:08:52 in tmp/deploy/images// Jun 01 09:09:01 if your image successfully built, one should be written out Jun 01 09:10:33 hm.. yes. the image tar.gz is there and working. is there an different way to force the manifest file? Jun 01 09:11:15 this is on dora btw Jun 01 09:17:02 ericbutters: ah right, manifest creation next to the image didn't happen until daisy Jun 01 09:17:10 ericbutters: you could enable buildhistory though Jun 01 09:18:15 how? and do i need to do some clean build then or only "bitbake -f -c rootfs "? Jun 01 09:20:43 Hi, I'm running into a problem with a BSP that I'm working on. Jun 01 09:21:44 doing bitbake core-image-sato it breaks in the end with: error: Can't install packagegroup-core-boot-1.0-r17@vab820_quad: no package provides busybox-hwclock Jun 01 09:22:36 doesn't busybox automatically provide the busybox-hwclock package? Jun 01 09:24:08 ericbutters: http://www.yoctoproject.org/docs/current/ref-manual/ref-manual.html#maintaining-build-output-quality Jun 01 09:24:35 ericbutters: just enabling it then building the image should give you results at least for the image Jun 01 09:25:18 imrehg: depends if your busybox config has that enabled Jun 01 09:25:51 imrehg: the dependency on busybox-hwclock is dependent on whether "rtc" is in MACHINE_FEATURES, FWIW Jun 01 09:26:21 so if your machine doesn't actually have an rtc you could just remove that from MACHINE_FEATURES Jun 01 09:27:00 otherwise you'd need to look at how busybox is configured in your case and enable the appropriate option Jun 01 09:27:31 bluelightning: it's in the machien features: MACHINE_FEATURES += " pci rtc" Jun 01 09:28:07 bluelightning: though if I take it out, the build fails the same way depending on busybox-hwclock Jun 01 09:28:36 imrehg: which version of the build system are you using? Jun 01 09:29:02 1.8, and based on the meta-fsl-arm stuff (my board is another i.MX6) Jun 01 09:30:01 ok, well, if you look at meta/recipes-core/packagegroups/packagegroup-core-boot.bb you should see that that dependency is definitely conditional on rtc being in MACHINE_FEATURES Jun 01 09:34:08 if rtc is in there and hwclock is a dependency, what do I need to change in the machine files to get it actually built? Jun 01 09:39:13 btw if I remove rtc from MACHINE_FEATURES, it still breaks on that dependency Jun 01 09:42:57 I don't know what to suggest, other than maybe that packagegroup somehow isn't being rebuilt? Jun 01 09:44:06 I've rebuilt things a few times (removing tmp/), will poke around a bit more, thanks! Jun 01 09:44:14 if you want to get busybox-hwclock built, you'll need to have a look at how busybox is being configured in your situation Jun 01 09:44:33 perhaps one of the layers you are using is specifying its own configuration Jun 01 09:45:56 the default busybox defconfig has CONFIG_HWCLOCK=y in it Jun 01 09:53:59 bluelighting: interesting, indeed CONFIG_HWCLOCK is not set in my case. Couldn't see any layer that messes with that config, will have another look Jun 01 09:56:36 imrehg: look for busybox in the output of "bitbake-layers show-appends" and "bitbake-layers show-overlayed" Jun 01 10:00:28 not in overlayed, and appended only in meta-yocto... Jun 01 10:03:51 If I do a "bitbake -c menuconfig busybox", hwclock is actually enabled. Then a "bitbake busybox", then a "bitbake core-image-minimal" and it breaks with the same "not provided...." Jun 01 10:05:39 I'll keep looking, thanks for the pointers! Jun 01 10:07:44 that is strange... Jun 01 10:09:30 i got to go at the moment, but will follow up if I find something. cheers! Jun 01 10:30:56 bluelightning: did you see my question on bash of last week? Jun 01 10:32:56 https://www.yoctoproject.org/irc/%23yocto.2015-05-29.log.html#t2015-05-29T09:42:10 Jun 01 13:28:00 denix: did you still have that init-ifupdown rebuild problem? I think I might understand it Jun 01 13:41:10 RP: no, don't see init-ifupdown rebuilding. I had other generic packages rebuilding, adding them to safe deps resolved that Jun 01 13:50:23 denix: I guess I'm misremembering the issue then. I did just find an issue where init-ifupdown does rebuild :/ Jun 01 13:53:48 RP: it could have been JaMa as he had a whole bunch of them... Jun 01 15:12:16 hello. there seems to be a problem with "meta-toolchain-qt5", it builds qttools with CONFIG+=linguistonly but then aborts at some point because qttools-plugins is missing. Jun 01 15:17:28 JaMa: does that ring any bells? ^ Jun 01 15:24:53 hmm Jun 01 15:25:07 seems like bitbake is clearing http_proxy related envvars Jun 01 15:29:41 that shouldn't be the case, it's explicitly kept in the fetch env whitelist Jun 01 16:09:25 kergoth: both http_proxy and https_proxy? Jun 01 16:09:39 is it possible that this recipe is clearing those envvars? Jun 01 16:10:49 kergoth: wait, is it only whitelisted in do_fetch? Perhaps it's cleared during do_compile? Jun 01 16:11:13 if you're fetching in do_compile, you're doing it wrong Jun 01 16:11:46 kergoth: agreed. It is wrong. I'm trying to get upstream to fix it Jun 01 16:11:55 in the meantime, I'm investigating work arounds **** ENDING LOGGING AT Tue Jun 02 02:59:58 2015