**** BEGIN LOGGING AT Fri Mar 04 02:59:58 2016 Mar 04 08:24:30 Hi there o/ Mar 04 08:25:09 I have gcc set in PATH and LD_LIBRARY for this gcc and building cmake-native fails Mar 04 08:25:40 as my default system gcc older than gcc set by env vars Mar 04 08:26:26 and cmkae bootstrap fails to run as he can't find libc in newer version from gcc Mar 04 08:27:57 is there any other option that passing rpath during cmake compilation or installing newer gcc as a default with newer libc ? Mar 04 08:43:06 Hi all. How can i configure smart in such way, that it uses a local folder as package source? Mar 04 08:44:15 I've found this [ https://wiki.yoctoproject.org/wiki/Smart ] but i don't get the --data-dir option Mar 04 09:44:41 cmake-native/3.3.1-r0/build/Bootstrap.cmk/cmake uses RPATH from x86_64 linux sysroot /usr/lib and /lib directories, while he was compiled from host gcc and using LD_LIBRARY_PATH Mar 04 09:50:25 why bootstap cmake has this rpath to sysroot ? Mar 04 10:05:18 How can i generate the neccesarry metadata for my rpm packages, so i can serve the deploy/rpm/<...>/ directory as repository? Mar 04 10:13:57 Anticom: images are generated using those directories are repositories, so its all already there Mar 04 10:16:08 ah cool Mar 04 10:16:36 Is there any documentation on how to configure smart channel using recipies? Mar 04 10:16:46 channels* Mar 04 10:19:24 set PACKAGE_FEED_URIS Mar 04 10:19:30 to the uri of where the feed will be Mar 04 10:20:08 i run a local httpd on my build machine so have PACKAGE_FEED_URIS="http://flashheart.local/poky-master/" in my local.conf Mar 04 10:58:22 rburton: and i don't need to specify the other smart attributes like name of the channel or type=rpm-md ? Mar 04 11:01:12 smart is so kind... 'smart upgrade --check' just tells me 'There are new upgrades available!' but not which packages it would like to update Mar 04 11:02:48 i guess --dump is what i wanted Mar 04 11:23:03 Anticom: that variable will write a basic smart config for you Mar 04 12:12:24 does PR_ = ... make any sense? guessing not, but i want to verify my understanding. Mar 04 12:13:12 the recipe will only be "run" once to generate all the packages, so you shouldn't be able to specify per-package PRs like that (unless there's specific support beyond overrides) Mar 04 12:34:03 Ulfalizer: no, PR is recipe wide Mar 04 12:34:20 for the reason you gave, basically Mar 04 12:39:24 ok, thanks Mar 04 12:41:21 would using fakeroot for perform_packagecopy cause more issues or just possibly worse performance Mar 04 12:41:24 ? Mar 04 12:41:28 same thing for LICENSE i guess Mar 04 12:41:35 * Ulfalizer is cleaning up a bit Mar 04 12:45:45 ah, apparently specifying licenses on a per-package basis is supported Mar 04 12:50:33 rburton: but how do i specify e.g. whether it should be of type rpm-md or rpm-dir? Mar 04 12:51:58 Anticom: if you use PACKAGE_FEED_URIS it uses rpm-md. literally no idea what the difference is, if you want more control then write your own repo config Mar 04 12:53:15 rburton: afaik rpm-md uses repomd.xml, filelists.xml.gz and primary.xml.gz while rpm-dir just scans a directory for *.rpm Mar 04 12:53:47 rburton: and does PACKAGE_FEED_URIS work with opkg / dpkg? Mar 04 12:53:50 or is it rpm only? Mar 04 12:53:53 opkg, yes Mar 04 12:53:57 no idea about dpkg Mar 04 12:54:03 okay thank's Mar 04 12:54:12 (i use it with opkg and had to check it exists for rpm) Mar 04 12:54:12 well i guess if it doesn't atm it will be in the future :) Mar 04 12:54:49 rburton: btw what's the advantage of using opkg over rpm besides saving some disk space? Mar 04 12:55:03 everyone seems really exicted about using opkg Mar 04 12:56:00 * rburton shrugs Mar 04 12:56:03 both package managers Mar 04 12:56:09 rpm is bigger - no need to ship python Mar 04 12:56:14 erm, need to ship python Mar 04 12:56:23 opkg is just enough package manager to work Mar 04 12:56:32 so it depends on what you want from a package manager really Mar 04 13:00:50 opkg is more flexbile and not so binded to the distro Mar 04 13:00:55 as rpm is Mar 04 14:01:22 does recent openssl cve fix breaks fido for anyone here? Mar 04 14:13:14 <_questor> hi. I've tried compiling a rasp-image and got a compilation error from m2crypto.so: undefined symbol SSLv2_method (ubuntu15.10, jethro-branch, package crda). is that an known error? in the crda-recipe depends shows "pyhton-m2crypto-native", but it seems there's still something missing.. Mar 04 14:20:11 hi Mar 04 14:35:46 rburton: oh FYI you can do `smart channel --help-type ` to find out more about how the available types differ from each other :> Mar 04 14:36:06 Anticom: my desktops are debian and i use opkg in oe :) Mar 04 14:36:19 Just saying, if you should ever need smart Mar 04 15:36:41 anyone seeing NOTE: recipe meta-ide-support-1.0-r3: task do_package_write_ipk: Failed ? Mar 04 15:37:25 meta-ide-support/1.0-r3/temp/run.sstate_create_package.5091: 124: cd: can't cd to ..workdir..meta-ide-support/1.0-r3/sstate-build-package_write_ipk/ Mar 04 15:45:36 denix, in what way? Mar 04 18:07:04 halstead: fedora23 just disappeared :( Mar 04 18:07:49 RP1, A hard lock-up. I'm investigating now. I may have to switch back to the 4.2 series kernel. Mar 04 18:08:23 RP1, It looks like the builds assigned to it were successfully rescheduled. Mar 04 18:09:03 halstead: right, just thought I'd mention it in case you didn't know... Mar 04 18:09:39 Thank you. I did just notice it 5 minutes ago. Mar 04 18:30:30 I couldn't find anything online, does anyone know if yocto supports multiple partitions at the moment? Mar 04 18:30:56 Only hits I saw were something in January of last year suggesting it wasn't supported at that time. Mar 04 18:31:43 what do you mean by multiple partitions? Mar 04 18:33:11 generating a separate image for /var Mar 04 18:33:29 Since /var is populated by the build system Mar 04 18:35:37 I'm not against (trying) to contribute support if needed. Mar 04 18:36:29 I don't know enough of the guts of bitbake/yocto yet to know where the tool that creates images is located, however. Mar 04 18:37:04 Hmm. Interesting question. I don't know of a thing to do that, but it shouldn't be too hard to manually do it after the fact. Mar 04 18:37:25 We actually already have python scripts to kludge some data into our rootfs Mar 04 18:38:42 I imagine as part of that step, I could move /var to a separate image. Mar 04 18:39:06 I'd still rather integrate it into yocto if possible though. Mar 04 18:39:18 ideally, it would even use it to generate the fstab. Mar 04 18:40:07 You might take a look at the 'wic' tool as well Mar 04 18:40:11 http://www.yoctoproject.org/docs/2.0/mega-manual/mega-manual.html#creating-partitioned-images Mar 04 18:41:02 stephano: thank you, I will investigate. Mar 04 18:41:40 I've read a good chunk of of the 1.8 manual, but in bits and pieces, and not straight-through. Mar 04 18:42:17 As well as the bitbake manual. Sadly, I still feel somewhat lost at times without just digging into the python itself. Mar 04 18:42:27 Only so much I can retain, heh. Mar 04 20:13:30 <_nico_> Hi, I'm looking for some info on how to create a yocto based image for the RPi3. Mar 04 20:18:22 _nico_, I'd start here http://git.yoctoproject.org/cgit/cgit.cgi/meta-raspberrypi Mar 04 20:18:38 may need to meial the maintainers and ask if they have a pi 3 to work with Mar 04 21:37:52 zeddii: sob...a new issue arose **** ENDING LOGGING AT Sat Mar 05 02:59:58 2016