**** BEGIN LOGGING AT Mon Mar 06 03:00:00 2017 Mar 06 06:50:23 Marex: send a patch with rc Mar 06 06:50:36 and that will help to stabilize it this week Mar 06 07:05:10 hello Mar 06 07:05:19 i have problems with rppmv5 Mar 06 07:05:35 rpm -qf $(which dc) says 'bc' Mar 06 07:06:10 however, `rpm -ql bc` does not list dc Mar 06 07:06:21 which dc Mar 06 07:07:40 anybody else sees similar problems with rpm v5 ? Mar 06 07:44:14 I am trying to set up toaster using this tutorial: https://wiki.yoctoproject.org/wiki/Setting_up_a_production_instance_of_Toaster However, I get stuck at setting SECRET_KEY and STATIC_ROOT. How exactly do I set these variables? Must this happen within the virtual environment? Mar 06 07:48:10 Great, solved it. has also to be set in /var/www/toaster/poky/bitbake/lib/toaster/toastermain/settings.py Mar 06 08:20:44 morning Mar 06 09:28:50 guys, i am having some small recipes where is place their actual files via SRC_URI="file://XYZ.sh" into hte meta-layer directly .. still i get the "Unable to get checksum for XYZ SRC_URI entry" warning .. Mar 06 09:29:43 is there a better way to ship files via the meta layer ( small scripts etc) or will i need to point to a actual tarball and use the checksum stuff as intended Mar 06 09:48:25 rob_w: what is at the eend of the warning? " file could not be found" ? Mar 06 09:49:13 file not found except in DL_DIR Mar 06 09:50:10 i wonder if files ending in .patch are treated different ( obviously they do ) so i wonder how i could reach this with other files Mar 06 09:51:07 the issue seems to be SRC_URI: where did you put the file (uder which dir?) Mar 06 09:51:18 is just this Mar 06 09:52:53 i will have a recipe where i state them under SRC_URI="" inside that folder where the recipe is , i will have a "files" directory where those files reside Mar 06 09:53:36 the SRC_URI then looks like SRC_URI = "file://xyz.sh \ etc Mar 06 09:59:19 rob_w: sorry, phone call Mar 06 10:00:12 it should just work but please try to rename 'files' dir to 'xyz' Mar 06 10:01:09 better to be specific, it comes first in the search path Mar 06 10:12:02 ant_work: hmm looks like it works, thx alot Mar 06 15:20:08 nrossi: your qemu packageconfig series needs rebasing, can you do that? part 2 is just because there's a —with-python to EXTRA_OECONF but part 4 also fails Mar 06 15:21:00 rburton: Sure, i've already rebased locally. Should I send out a updated set of patches or would you like a remote to pull? Mar 06 15:22:30 the list should see them, and i'm happy with branch or mails Mar 06 15:27:31 rburton: Sent Mar 06 15:28:14 cheers Mar 06 15:30:00 rburton: also, there is still one other patch that was part of that group of patch sets. https://patchwork.openembedded.org/patch/136585/ is that in your queue? it still applies fine to master Mar 06 15:37:02 nrossi: barfs with current master Mar 06 15:37:06 (the qemu stuff) Mar 06 15:44:24 nrossi: doesn't the gettext-libintl package get renamed if it's got a library in? Mar 06 15:48:26 rburton: hmmm thats odd, i just rebased it on top of oe-core 784a837926, are you applying to a different master? Mar 06 15:50:01 poky master but should be the same as far as qemu is concerned Mar 06 15:50:02 rburton: with the gettext-libintl what do you mean by renamed if its got a library in it? Mar 06 15:50:24 if you have a package called gettext-libintl which contains libintl.so.1, it gets renamed to libintl Mar 06 15:51:09 i did not observe that during testing. Which class does that renaming? Mar 06 15:51:11 i'll try picking from your branch Mar 06 15:51:14 debian.bbclass Mar 06 15:51:40 nrossi: ok picking from your branch worked. whatever :) Mar 06 15:52:37 nrossi: what is this .git-cover-letter stuff from? Mar 06 15:52:58 rburton: ah just drop the top commit, i store my cover letter in git history Mar 06 15:53:11 custom tooling? Mar 06 15:53:15 yer Mar 06 15:53:28 have you tried git-series Mar 06 15:54:00 i have, but ended going with a custom tool. Mar 06 15:55:00 is there a simple way to trace in how variables are set during a yocto build? e.g. which recipes modify a certain PACKAGECONFIG variable? Mar 06 15:55:29 CoLa|work: bitbake -e? Mar 06 15:56:02 yeah, bitbake -e [recipe] Mar 06 15:56:16 pipe to less, search for ^PACKAGECONFIG Mar 06 15:56:46 thanks Mar 06 15:57:24 NOOOO why does bitbake want to fetch a kernel all of a sudden Mar 06 15:59:11 oh yeah, bsp update. coffee time i guess Mar 06 16:10:58 Hello guys Mar 06 16:11:20 Is it possible to have a recipe depends on a package dynamic generated ? Mar 06 17:23:22 rburton: regarding my patch. I think it made that function to fail when it was a last statement in the function. I moved it up in a hope it will help. Why do you think it will not? Mar 06 17:24:37 ed2: generally we run shell scripts with set -e, which makes any non-zero exit status fail the script Mar 06 17:24:52 * kergoth yawns Mar 06 17:28:26 kergoth: $ cat test.sh Mar 06 17:28:26 foo() { Mar 06 17:28:26 [ -h /etc/fstab ] && echo 'fstab is a symlink' Mar 06 17:28:26 echo 'end of foo' Mar 06 17:28:26 } Mar 06 17:28:26 foo Mar 06 17:28:27 echo $? Mar 06 17:28:27 ~ $ Mar 06 17:28:28 ~ $ Mar 06 17:28:28 ~ $ sh -e test.sh Mar 06 17:28:29 end of foo Mar 06 17:28:29 0 Mar 06 17:28:46 kergoth: am I missing something here? Mar 06 18:11:00 rburton: so figured out why i didn't see the debian class renaming. It only applies to "lib/*.so"... which mingw doesn't have. Do you think the package naming still makes sense? should I update the change to add DEBIAN_NOAUTONAME? or should this sort of lib be handled in a different way? Mar 06 19:06:02 I think debian autorenaming provides a good solution for coexistence, I think non-linux targets would probably not need it as much and given container/crops solutions I think mingw should disappear in future Mar 06 20:09:52 hey! are there resources for first time contributors? Mar 06 20:10:44 i'm interested in yocto's outreachy internship, so i'm trying to find a way to contribute as part of that :) Mar 06 20:31:20 hey garbados, you want to contact jefro: jeffrey.osier-mixon@intel.com Mar 06 20:31:54 thanks stephano :D Mar 06 23:46:52 sgw_: around? Mar 06 23:51:09 RP: hi, I tried to debug kernel menuconfig. Looks like there is a problem because I have to abuse of cml1.bbclass Mar 06 23:52:05 there is do_menuconfig[depends] += "ncurses-native:do_populate_sysroot" Mar 06 23:52:50 but somehow ncurses is needed instead of ncurses-native...doh Mar 06 23:53:42 ant_home: why would it need a target ncurses? the menus aren't running on the target? Mar 06 23:54:27 logging now Mar 06 23:54:49 first is: Mar 06 23:54:51 ggi HOSTLD scripts/kconfig/mconf Mar 06 23:54:51 scripts/kconfig/mconf.o: In function `show_help': Mar 06 23:54:51 mconf.c:(.text+0x914): undefined reference to `stdscr' Mar 06 23:55:23 then lxdialog stuff Mar 06 23:55:28 scripts/kconfig/lxdialog/checklist.o: In function `print_arrows': Mar 06 23:55:28 checklist.c:(.text+0x2c): undefined reference to `wmove' Mar 06 23:55:34 etc Mar 06 23:56:16 RP: here Mar 06 23:56:57 RP: is an old issue, in the past I did build ncurses Mar 06 23:57:42 I guess RSS exposed it, now you can reproduce it: -c cleansstate ncurses and try to -c menuconfig virtual/kernel Mar 06 23:58:03 sgw_: just wondering about a kernel fix http://git.yoctoproject.org/cgit.cgi/poky-contrib/commit/?h=rpurdie/wip-rss2&id=f1b71802a8b908c3606e5bfe9b779608a5cfbc96 Mar 06 23:58:50 sgw_: of course fixing this doesn't really help me much as it will take an age to make it into production systems but it would make me happier to know the problem would eventually go away Mar 06 23:59:24 ant_home: Whilst it may "fix" a problem, that problem still doesn't make sense. It shouldn't need a target ncurses Mar 06 23:59:50 RP: heh, sure... Mar 07 00:00:46 but look at the comments of [YOCTO #2513], one cml1 patches Mar 07 00:01:45 RP: are you submitting this to mainline? Mar 07 00:04:57 sgw_: I'm wondering about it Mar 07 00:06:17 ant_home: you mean http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=7e62e24b6cfd1683cd4b7ba565537b25d1b604f8 ? Mar 07 00:06:34 ant_home: or https://bugzilla.yoctoproject.org/show_bug.cgi?id=2513 ? Mar 07 00:06:35 Bug 2513: minor, Low, 1.3 M1, richard.purdie, VERIFIED FIXED, bitbake -c menuconfig silently fails if ncurses is not installed Mar 07 00:06:55 I don't understand which comments I'm looking at? Mar 07 00:08:02 RP: I coud not find anything strange in the old commits Mar 07 00:08:22 but a bell rings reading cml1: fix tasks after default [dirs] changed Mar 07 00:10:33 ant_home: I'm really not seeing the connections between these things... Mar 07 00:11:00 heh, too many changes and an old issue reappearing :/ Mar 07 00:12:26 maybe linux-yocto does some trick to fix that. standard kernels do not Mar 07 00:12:59 I'll verify that, otherwise there is a new bug Mar 07 00:20:09 RP: ah, the logs shows the path, indeed Mar 07 00:20:43 /tmp/build/tmp-glibc/work-shared/h3600/kernel-source/Makefile:546: recipe for target 'menuconfig' failed Mar 07 00:24:04 full story: http://pastebin.com/qTyJThw2 Mar 07 00:36:10 people actually use menuconfig? Mar 07 00:36:13 * paulg runs Mar 07 00:37:49 next folks will be saying people use busybox interactively.... Mar 07 00:44:47 gn **** ENDING LOGGING AT Tue Mar 07 03:00:02 2017