**** BEGIN LOGGING AT Fri Oct 16 02:59:58 2015 Oct 16 07:26:48 morning all Oct 16 08:23:22 Hello everybody! Oct 16 08:24:04 Could you tell me how I can determine how was set PACKAGECONFIG variable for specific recipe? Oct 16 08:24:39 I mean, i know what i have written in my image .conf file. But How it was really set by bitbake ? Oct 16 08:33:10 spaszPL: i guess inspecting the output of bitbake -e for that recipe should tell. Oct 16 08:52:25 hmm Oct 16 08:52:32 i will check it, thx Oct 16 09:26:27 LetoThe2nd, yes it works :) Oct 16 10:45:11 hi all Oct 16 10:49:07 bluelightning: I have small question related to opkg if you can answer Oct 16 10:50:01 Noor: I can try... Oct 16 10:50:55 bluelightning: do you know while creating rootfs opkg changes the ipk file names Oct 16 10:51:06 what is happening on our side is Oct 16 10:51:43 it is replacing / with underscore and thenn filename become so long that it says file name too long cant stat Oct 16 10:52:26 er... how do you come to have / in the file name? Oct 16 10:52:48 bluelightning: http://paste.ubuntu.com/12797728/ Oct 16 10:53:02 we have having this error Oct 16 10:53:17 at time of rootfs creation Oct 16 10:53:41 that's very strange indeed Oct 16 10:53:57 I couldn't tell you how that is even ending up in the ipk filename Oct 16 10:54:12 bluelightning: Ok no issues Oct 16 10:54:23 I'll wait for kergoth then :) Oct 16 11:12:30 shouldn't everything in 'image' in a recipes workdir make it into my image? Oct 16 11:13:40 tasslehoff: don't think so, everything that is matched by the FILES variable makes it into the package. Oct 16 11:15:29 LetoThe2nd: there is no specification of FILES_${PN}-dev in the recipe, but the symlink I want in the rootfs ends up in the dev package Oct 16 11:16:16 tasslehoff: some stuff for FILES is implicitly set by default. Oct 16 11:16:26 but you might need to expand it Oct 16 11:19:00 LetoThe2nd: it specifies ${libdir}/*, but the libcuda.so symlink in /usr/lib does not "make it" :) Oct 16 11:19:54 tasslehoff: hm. sorry, but at that point i've got to pass. Oct 16 11:20:26 LetoThe2nd: np. Oct 16 12:11:46 khem: ping Oct 16 12:44:49 khem: I have figured out how to make cuda run on my Jetson, and need your advice on a path I'll submit to meta-jetson-tk1 Oct 16 12:45:04 s/path/patch Oct 16 15:15:11 Noor: sounds like opkg encodes the url into its cache, not just the filename, and in this case the url is a file: url with a long path Oct 16 15:15:22 that's what it looks like in the pastebin, anyway Oct 16 17:22:52 I'm having issues with gitr0 not incrementing in a kernel recipe Oct 16 17:22:57 PV = "3.14.43+gitr${SRCPV}" Oct 16 17:22:57 SRCREV = "${AUTOREV}" Oct 16 17:22:57 SRC_URI = "git://gitlab.bec-systems.com/xyz/linux.git;user=git;branch=github-beagle-3.14.43;protocol=ssh Oct 16 17:23:26 version is "3.14.43+gitr0" on every build, even though I'm checking in changes on the branch Oct 16 17:23:39 do you use PR service? Oct 16 17:23:51 good question -- so that must need turned on Oct 16 17:24:10 yes for any recent OE release Oct 16 17:27:17 pr service is required for autorev to work properly? oh, it's handling AUTOINC? Oct 16 17:58:45 kergoth: only for AUTOINC, AUTOREV itself should work Oct 16 17:59:02 * kergoth nods Oct 16 17:59:06 kergoth: but with PV going sometimes backwards it's not very useful without working AUTOINC Oct 16 17:59:28 We should really think about just defaulting to the local autospawned pr server by default at some point Oct 16 17:59:37 * kergoth does that in mel.conf Oct 16 18:00:09 PRSERV_HOST = "localhost:0" fixed it Oct 16 18:00:22 our internal builds still do manual PR bumps Oct 16 18:00:49 as we don't care about upgrade paths in binary feed Oct 16 18:01:30 we even thinker about adding: Oct 16 18:01:32 EXTENDPRAUTO_append = "${@ '.%s' % str(d.getVar('BB_TASKHASH', True))[:10]}" Oct 16 18:02:09 instead of for us meaning less numbers from PR service which aren't synced correctly across multiple builders **** ENDING LOGGING AT Sat Oct 17 02:59:58 2015