**** BEGIN LOGGING AT Thu Sep 24 02:59:58 2015 Sep 24 04:05:55 is there really no chat in here for almost 6 hrs?? Sep 24 04:07:17 not familiar with timezones? Sep 24 04:07:22 it can't be busy 24/7 Sep 24 04:07:39 just wanted to see if my client died or something Sep 24 04:07:40 thx Sep 24 05:56:16 hi guys Sep 24 06:10:43 does anyone know if it's possible to have a recipe that is machine specific ? for example i have a PKGCONFIG variable that I want to be different upon the selected machine Sep 24 06:32:09 drou: you can use machine overrides like PKGCONFIG_machineName = ..., see http://www.yoctoproject.org/docs/latest/dev-manual/dev-manual.html#understanding-recipe-syntax Sep 24 06:42:20 nice, thanks, didn't know about that ! Sep 24 10:05:56 is there a difference between oe.utils.conditional() and base_conditional() ? Sep 24 10:08:35 jku: base)conditional just calls oe.utis.conditional Sep 24 10:16:38 rburton: so it does... what's the point? I assume not just saving 4 characters Sep 24 10:18:04 jku: i suspect base_conditional predates the move of code to oe.utils, but that's a guess Sep 24 10:29:45 that's my memory of it Sep 24 10:39:40 bluelightning: am I correct that devtool can't really handle modifying a patch in the middle of a patch series without making metadata changes in other patches as well? Sep 24 10:40:48 jku: right, it's a little too naive for that at the moment - but I suspect it's something we could fix (and I'd see it as a legitimate bug, so we could fix it in M4) Sep 24 10:41:15 jku: as a workaround you can just git checkout the unnecessarily modified files after update-recipe Sep 24 10:44:05 bluelightning: I think the root cause for disappearing "Upstream-Status" lines here has the same cause: http://pastebin.com/3gy68T50 Sep 24 10:44:35 It's technically not part of the git message body so it gets dropped Sep 24 10:45:27 that probably happens even for legitimately modified patches Sep 24 10:45:45 Arguably the patch is wrong in the first place Sep 24 10:55:22 bluelightning: I guess there's no bug for the "modify patch in middle of patch series" case and I should file it? Sep 24 11:06:21 jku: there isn't, if you could file one that would be great Sep 24 11:11:15 sure Sep 24 12:04:55 hi Sep 24 12:05:27 is there a way to put an instruction into a recipe to source an env file for example, before running make Sep 24 12:25:25 blotunga, yes, but you will need to create a custom do_compile step Sep 24 12:25:41 grep do_compile -r oe-core/meta Sep 24 12:26:29 Crofton: thanks, I'll check Sep 24 12:27:10 blotunga: or do_compile_prepend() if you don't need to do anything special in the compile command itself Sep 24 12:28:03 jku: I just need to set some environment variables. Ah and I think I'll have one more problem. I need something that replaces buildroot's rootfs_path.env Sep 24 12:28:09 is there something similar? Sep 24 12:28:25 I'm just making the switch and am kind of overwhelmed :D Sep 24 12:33:30 sorry not familiar with buildroot. But do_compile_prepend() should work for setting env variables required during make Sep 24 12:35:59 jku: the rootfs_path.env in buildroot basically added the toolchain into the path and sets the path to the libraries etc. Sep 24 12:36:12 but probably it's not needed Sep 24 14:21:30 I am seeing that bitbake is adamant on looking at /mnt/home today Sep 24 14:21:42 even though I do not have that directory Sep 24 14:21:45 any ideas ? Sep 24 14:22:16 http://pastebin.com/tZhuReh0 Sep 24 14:22:37 That seems "incorrect" Sep 24 14:23:25 khem: consolelogfile gets set from BB_CONSOLELOG Sep 24 14:23:26 that almost looks like it's in the UI itself.. (consolelogfile) that is busted.. Sep 24 14:23:35 Ahh there ya go Sep 24 14:23:57 whose default value is "${LOG_DIR}/cooker/${MACHINE}/${DATETIME}.log" Sep 24 14:24:15 well a bitbake -e should hopefully show you waht it's set to Sep 24 14:24:20 true Sep 24 14:24:22 my guess is somehow LOG_DIR got mangled Sep 24 14:24:28 LOG_DIR defaults to ${TMPDIR}/log Sep 24 14:26:28 bluelightning yes no changes to LOG_DIR Sep 24 14:27:57 this workspace worked so something must be borked in env Sep 24 14:27:59 lets see Sep 24 14:28:13 deleting conf/ dir helped Sep 24 14:42:04 khem: that doesn't really help figure out what caused it... Sep 24 15:00:19 yo Jefro! Sep 24 19:02:17 halstead: IM Sep 24 19:02:27 Hi lamego Sep 24 19:17:13 Hi all... I'm still learning the basics of yocto, but I'm trying to add a recipe that is added to the image I'm generating, and it seems to build and install it, but it's not in the image when I install it on the device. Sep 24 19:18:14 The package is being installed to tmp/work/at91sam9x5ek-poky-linux-gnueabi/{imgname}/1.0-r0/rootfs/etc/init.d/{filename} Sep 24 19:19:17 But when I put the final UBIFS onto the board, it's missing..? Am I doing something wrong? I just re-ran bitbake after adding this package. Do I need to bitbake from scratch or something? Sep 24 19:28:51 Hmm, no systemd support in weston-image? would be nice to have a weston demo image that worked with both sysvinit and systemd.. Sep 24 19:31:42 er, wrong image Sep 24 19:34:15 core-image-weston, that is Sep 24 20:10:23 what package provides gi.repository? Sep 24 20:10:39 i'm having a helavatime finding it Sep 24 20:13:02 maybe a more generic question: if i have a file, how do I find the recipe/package that provides it? Sep 24 20:17:06 if it's already been built, that's easy, with oe-pkgdata-util. if not, it's rather more difficult, since bitbake doesn't knwo what files will be available until the recipe is built, other than the patterns in the FILES varaibles Sep 24 20:20:27 tricky. Sep 24 20:20:42 looks like there is a meta-gir that's outside of poky and meta-oe **** ENDING LOGGING AT Fri Sep 25 02:59:58 2015