**** BEGIN LOGGING AT Mon Jun 18 02:59:58 2012 Jun 18 08:17:05 which is latest DISTRO for angstrom? i think "angstrom-2008.1" is too old, right? Jun 18 09:04:22 pb_: Are you around? I'm wondering how opkg should behave with dropbear and openssh Jun 18 09:05:51 pb_: Specifically, if openssh RCONFLICTS with dropbear (but not RREPLACES) what should happen if you install dropbear or openssh. The behaviour I'm looking for is for dropbear to get replaced in the image by openssh, iff openssh was specifically installed either individually or along with dropbear Jun 18 09:12:47 What do I need to write to a .bb file to make it add a crontab line when the package is installed? Jun 18 09:15:56 Cron from the cronie package doesn't seem to run the stuff in /etc/cron.daily, weekly, etc. I guess that's because /etc/crontab doesn't do it. Jun 18 09:24:17 pb_: according to the debian manuals there is no way to do this as far as I can tell :/ Jun 18 09:32:27 RP__: Replaces: might do what you want. Jun 18 09:32:43 RP__: Though that sounds like you just want a virtual package for ssh with dropbear as the default. Jun 18 09:35:00 broonie: replaces is too strong, it is more the virtual package with a default which I need, yes Jun 18 09:35:23 broonie: trouble is I'm not sure opkg can support that, or how we'd configure that from oe :/ Jun 18 09:36:10 Dunno about the opkg side but from OE I'd expect something like just make sure that openssh is installed first, then the virtual package. Jun 18 09:36:33 If OpenSSH wasn't installed then the virtual package will pull in the default, otherwise installing it is a noop. Jun 18 09:38:22 broonie: Its certainly worth exploring, thanks Jun 18 09:49:12 I'm having difficulties figuring out the boundaries between the yocto, poky and oe-core mailing lists Jun 18 10:31:55 Zagor: oe-core covers the meta directory, poky meta-yocto directory and the yocto list, all the other yocto projects Jun 18 10:32:06 (see git.yoctoproject.org) Jun 18 10:33:54 morning all Jun 18 10:34:29 hi Jun 18 10:34:50 hi pb_, hi hrw Jun 18 10:35:22 RP__: I don't think opkg can quite do what you want, no. Obviously you can get the approximate effect using update-alternatives, but I can't think of any way to actually get dropbear uninstalled. Jun 18 10:35:46 long time since last time I booted OE generated system Jun 18 10:39:54 hmm. yocto core-image-lsb-sdk lacks opkg but has rpm. fun Jun 18 10:45:59 hi all Jun 18 10:47:42 hrw: change the package type if you want something else, its configurable as always :) Jun 18 10:47:46 hi bluelightning Jun 18 10:47:51 question to old OE devs: which big distros were bootstrapped using OE? Jun 18 10:48:03 RP__: I am generating ipk packages Jun 18 10:48:27 hrw: and rpm as well? Jun 18 10:48:34 RP__: no, only ipk Jun 18 10:48:49 hrw: then it shouldn't have installed rpm :/ Jun 18 10:49:16 RP__: Package zip version 3.0-r1 has no valid architecture, ignoring. Jun 18 10:49:28 ufly output from 'opkg-cl list_installed' Jun 18 10:49:58 but know probably why Jun 18 10:51:12 'installing' opkg-config-base helps Jun 18 10:51:42 ar x + tar xf roxx Jun 18 10:52:00 update-alternatives: Error: not linking //etc/init.d/functions to functions.lsbinitscripts since //etc/init.d/functions exists and is not a link Jun 18 10:52:08 I know, 'patches welcome' ;D Jun 18 10:55:21 how well tested is target toolchain in Yocto/OE? Jun 18 11:09:17 RP__: thanks. I guess my testing mail to yocto should rather be in oe-core then? Jun 18 11:31:45 Zagor: I think that one is ok there, you could have gone for either in that case Jun 18 11:32:09 hrw: We have self-hosted images now, so OE can build OE Jun 18 11:32:19 hrw: That means it should be fairly well tested :) Jun 18 11:33:47 RP__: OE was self hosted 5-6 years ago ;) http://marcin.juszkiewicz.com.pl/2006/12/18/self-hosted-angstrom-build/ Jun 18 11:34:05 RP__: I am doing other build now to check some failures Jun 18 11:34:37 hrw: well, yes. I mean its being promoted and more regularly used/tested now though with a vmdk image type and so on Jun 18 11:34:55 hrw: You can download an image that will launch into hob etc Jun 18 11:35:49 cool Jun 18 11:36:16 RP__: there are plans to bootstrap ARMv8 (64bit) with OE Jun 18 11:43:27 hrw: great :) Jun 18 11:44:22 RP__: have to blog about it one day Jun 18 11:45:12 and again OE reminds me that there is no such thing as 'fast enough build machine' ;D Jun 18 12:00:21 hrw, Intel backs the Yocto Project to drive sales of high end build machines Jun 18 12:00:39 hehe Jun 18 12:00:39 Crofton|work: I know ;) Jun 18 12:00:50 hrw: We've been working on trying to a) use more CPU power effectively and b) optimise the build process but its still takes too long Jun 18 12:01:01 Crofton|work: you did not got discount on new intel hardware? Jun 18 12:01:12 Crofton|work: gah, the secret is out! ;-) Jun 18 12:01:44 Remeber when Micron stopped sponsoring and bitbake's memory consumption was cut to a fraction of what it was? Jun 18 12:01:45 :) Jun 18 12:02:06 Crofton|work: that was when Intel paid more to Zecke than Micron did? Jun 18 12:02:14 ah Jun 18 12:02:36 this will probably end up reported in Phoronix Jun 18 12:02:41 ;D Jun 18 12:02:58 Crofton|work: LWN would be better but it may be harder to get there Jun 18 13:37:43 Hello everybody !! I have one problem, i hope you could help me. Jun 18 13:38:36 I want to install OE on a board AT91SAM9263EK and i have an error when i use bitbake and google can't help me Jun 18 13:39:09 the error is ERROR: Function 'Fetch failed: Unable to fetch URL http://download.savannah.gnu.org/releases/quilt/quilt-0.48.tar.gz from any source.' failed ERROR: Task 24 (/home/asilhol/prog/stuff/openembedded/recipes/quilt/quilt-native_0.48.bb, do_fetch) failed with exit code '1' ERROR: '/home/asilhol/prog/stuff/openembedded/recipes/quilt/quilt-native_0.48.bb' failed asilhol@viveris:~/prog/stuff$ Jun 18 13:39:48 did you have this error ? can you help me ? Jun 18 13:39:51 thanks Jun 18 14:56:30 Hello everybody !! I have one problem, i hope you could help me. I want to install OE on a board AT91SAM9263EK and i have an error when i use bitbake and google can't help me. the error is ERROR: Function 'Fetch failed: Unable to fetch URL http://download.savannah.gnu.org/releases/quilt/quilt-0.48.tar.gz from any source.' failed ERROR: Task 24 (/home/asilhol/prog/stuff/openembedded/recipes/quilt/quilt-native_0.48.bb, do_fetch) failed Jun 18 14:56:56 could you help me ? Jun 18 15:18:49 hi, all! Jun 18 15:19:19 have anybody tried to make mixed OE-based rootfs with android graphics? Jun 18 15:20:19 and also - is there some official place with allwinner a10 sources, or maybe already-made overlay? Jun 18 17:30:50 i've been trying to get the interfaces file from netbase overridden with my own, but it seems there's a dozen ways to do so, and i haven't found one that i could get to work... anyone managed to do this? Jun 18 18:37:28 khem, task-core-tools-profile uses lttng-0.x instead of lttng-2.x. Why so? Nobody got time to switch over? Jun 18 18:38:56 sgw, looks like you were the last to touch that task Jun 18 18:41:21 kergoth, about bitbake trying to md5sum a socket (see my previous mail to the list).. Would you prefer to catch that in md5_file() ? Jun 18 18:42:49 blindvt: your probably right, when lttng-2 was added the tools-profile task did not get updated, a patch for that would be most welcome! Jun 18 18:43:32 kergoth, RP_, or should we classify that as a pathological case that we just ignore ? :) Jun 18 18:44:06 sgw, ok, i will keep it in mind Jun 18 21:15:36 mmh... the gold linker strips away sections which were kept with the old bfd one :( Jun 18 21:16:39 ensc: yes gold is more accurate in dependencies Jun 18 21:16:58 so if you have unused sections consider them gone Jun 18 21:17:04 khem: but these sections are marked with KEEP() in the linker srcipts Jun 18 21:17:40 oh Jun 18 21:18:44 hmm which version of binutils are you at ? Jun 18 21:19:22 khem: http://pastebin.com/RgrRcXQf Jun 18 21:20:02 recent oe-core + oe-meta Jun 18 21:20:49 in the pastebin, x1 is the oe gold linker, x2 is from the host system linker (Fedora 17, seems to be bfd) Jun 18 21:24:04 ensc: it seems a bug to me Jun 18 21:32:28 khem: really odd things happen when the _start + _end pointers are placed inside the section: http://pastebin.com/RLAEkxWw **** ENDING LOGGING AT Tue Jun 19 02:59:58 2012