**** BEGIN LOGGING AT Fri Dec 16 03:00:00 2016 Dec 16 08:17:17 Good morning. How do I assure it's used a variable I set on local.conf? I'm doing VAR = "foo", but another layer does VAR = "foo foo2", I want to prevent that other layer to override my VAR Dec 16 08:23:07 good morning Dec 16 08:23:26 aV_V: another layer should do VAR ?= "foo foo2" Dec 16 08:23:47 aV_V: no, nevermind sorry Dec 16 08:24:09 * mckoan yawns Dec 16 08:35:28 mckoan: yeah, that's how it should be but it isn't Dec 16 08:37:17 I guess there is no other option than doing a _remove Dec 16 08:38:09 aV_V: actually I can't figure out why you should need that Dec 16 09:18:11 I try to add my own layer, but bitbake does not like it. "WARNING: No bb files matched BBFILE_PATTERN_my-software-layer '^/local/home/user/oe-core/build/../layers/meta-my-software" Dec 16 09:18:39 this is the only message I get from bitbake concerning the layer Dec 16 09:19:07 is there any way to find out why it does not match? Dec 16 09:25:08 brrm: seems like the configuration in your layers cont/layer.conf is incorrect, or there just is no recipe and no bbappend in there (yet) Dec 16 09:25:25 brrm: see http://git.yoctoproject.org/cgit/cgit.cgi/poky/tree/meta-yocto-bsp/conf/layer.conf as an example Dec 16 11:04:00 hello everyone. I am using arago meta-layer ( http://arago-project.org/git/meta-arago.git?p=meta-arago.git;a=blob;f=meta-arago-distro/recipes-devtools/php/php_%25.bbappend;h=ab402f7e46569a03d63efe9851f215f447eb76b8;hb=HEAD ) and would like my build contains php-cgi module for lighttpd server with my php.ini configuration file. I created additional php bbappend file with the following lines: http://ix.io/1MIt . But after all I neither Dec 16 11:04:00 see /etc/php/ directory nor php.ini in it. Could somebody help me? Thank you. Dec 16 11:32:36 Ox4: have you considered to call yourself 0x4D? ;-) Dec 16 11:35:59 viengelm: nope :) Dec 16 11:36:07 viengelm: why 4D? Dec 16 11:36:26 because Oxford ;) Dec 16 11:38:57 :) Dec 16 12:13:29 nobody? Dec 16 12:16:50 Ox4: where u have put the append file? Dec 16 12:32:04 to my meta-layer Dec 16 12:32:14 aV_V: ^^ Dec 16 12:33:38 Ox4: and do u respect the original directory structure? Dec 16 12:39:00 oh, so there is already another php_%.bbappend, I don't know how exactly it's managed when there are multiple append files for the same recipe. If the 2 append files are not in conflict then it should not be problems I think but if they are then u must check priority layers Dec 16 12:39:01 aV_V: yes, sure. Other bbappend files work as I expected Dec 16 12:39:51 yes, but this files have no any conflicts between each other Dec 16 12:40:00 the other bbappend is yours? Dec 16 12:40:24 yep Dec 16 12:40:53 this ? http://arago-project.org/git/meta-arago.git?p=meta-arago.git;a=blob;f=meta-arago-distro/recipes-devtools/php/php_%25.bbappend;h=ab402f7e46569a03d63efe9851f215f447eb76b8;hb=HEAD Dec 16 12:41:46 when u create ur own bbappends, u must place them in your layer and create the needed directories Dec 16 12:42:03 aV_V: no, mine is http://ix.io/1MIt Dec 16 12:42:21 yes, I have created all needed stuff Dec 16 12:43:45 make sure it's applied with bitbake-layers show-appends Dec 16 12:43:57 otherwise I dont know what could be Dec 16 12:46:22 aV_V: just a second Dec 16 12:47:37 aV_V: http://ix.io/1MJn Dec 16 12:47:37 Dec 16 12:48:00 php_5.5.21.bb: Dec 16 12:48:00 /home/yocto/yocto_build/tisdk/sources/meta-arago/meta-arago-distro/recipes-devtools/php/php_5.5.21.bbappend Dec 16 12:48:00 /home/yocto/yocto_build/tisdk/sources/meta-sirin/recipes-extended/php/php_5.5.21.bbappend Dec 16 12:48:12 meta-sirin is mine Dec 16 16:01:28 offtopic: someone knows why u-boot doesn't detect ctrl+c? I've set bootdelay to 0 Dec 16 18:13:18 gah, keep getting a gtk3 do_install failure due to an issue with gobject-introspection Dec 16 18:15:58 halstead: when you're done with the AB work, can you fire ross/mut on the *new* AB cluster please? Dec 16 18:16:08 the old cluster is just finishing M1 RC1 \o/ Dec 16 18:16:21 it says to set GIR_EXTRA_LIBS_PATH, but it's already set? Dec 16 18:16:22 * kergoth sighs Dec 16 18:16:25 rburton1, Sure. Dec 16 18:16:35 thanks Dec 16 18:16:38 on that note, good weekend all Dec 16 18:16:48 ugh new cluster Dec 16 18:17:36 what do folks thing of hardened toolchain to be default meaning we make security flags defaults Dec 16 18:17:44 s/thing/think Dec 16 18:19:03 currently what we do is suboptimal a bit. Since we have to babysit the toolchain sequence on options and sometimes gcc does not agree with ld and sometimes with code and vice versa Dec 16 18:19:19 we keep patching for these but I think its probably wrong way Dec 16 18:19:38 we should make the toolchain do PIE by default Dec 16 18:22:02 seems reasonable, but defaults don't avoid the need to pass the flags, since we need to be able to support external toolchains Dec 16 18:22:15 if making them defaults hides those, then it's a net loss imo Dec 16 18:24:27 I think Ubuntu 16.10 is now enabling pie on their packages for most architectures Dec 16 18:36:12 kergoth: thats fine then the patching is passed on to external toolchain Dec 16 18:36:22 internal toolchain should work fine with or without patching Dec 16 18:36:55 security folks really want to enable pie Dec 16 18:37:20 reproducible builds guys probably dont like it so much Dec 16 18:51:55 oh! gtk+3:do_install failed because bitbake re-ran tasks, which ended up resulting in do-install compiling something, not just do_compile, and the GIR_EXTRA_LIBS_PATH was only exported in do_copmile, not do_install Dec 16 18:51:57 that explains it Dec 16 18:55:50 yeah this hidden nature of partial compile happening during make install broke some assumption here too Dec 16 18:56:14 where a tool was assuming do_compile completes all compilation work Dec 16 18:56:27 which is not true for many apps Dec 16 18:56:45 although I think its a bug in these apps make systems Dec 16 19:15:43 khem, can we do 64 bit builds for rpi3 in oe? Dec 16 19:19:42 Crofton: Currently not in master branch of meta-raspberrypi, there were some patches for this posted recently though Dec 16 19:51:27 I recall the patches Dec 16 21:38:21 Crofton: ys we can Dec 16 21:38:33 with few patches on ml Dec 16 21:38:39 which are not upstreamed yet Dec 16 21:39:02 I have staged all of the patches except kernel Dec 16 21:39:14 kernel we want to movet to 4.9 mainline Dec 16 21:39:29 also drop linux-raspberrypi as well Dec 16 21:39:45 Well, it sounds like I should look at getting one Dec 16 21:39:58 nooooooo webkitgtk is rebuilding Dec 16 21:39:59 * kergoth sighs Dec 16 21:40:21 Take the rest of the day off Dec 16 21:41:12 indeed Dec 16 22:28:43 this is odd, getting files in the rootfs owned by the build user, and it's certs. but there's no reason the certs shuld have wrong ownership, they're installed with 'install',w hich defaults to the current uid/gid Dec 16 22:35:31 kergoth: I feel your pain Dec 16 22:35:42 I deal with webkit everyday Dec 16 22:53:49 kergoth: we see that every couple of days once Dec 16 22:54:05 kergoth: haven't figured out what causes that, guess would be some problem in pseudo, but we have no idea Dec 16 23:50:30 neverpanic: good call, pseudo.log on the image shows symlink & inode mismatches for the symlinks in /etc/ssl/certs/ created by update-ca-certificates, even though the latter runs under pseudo as a rootfs-time postinst. thoughts, seebs? Dec 16 23:51:42 this is morty Dec 17 02:03:26 Does someone use systemd/netword's resolv.conf feature. I am especially interested how `ln -s /usr/lib/systemd/resolv.conf /etc/resolv.conf` is handled Dec 17 02:37:41 zzeroo: "[OE-core] [PATCH] systemd: point to correct resolv.conf when resolved is enabled" on the oe-core mailing list may be of interest. Dec 17 02:38:10 zzeroo: https://patchwork.openembedded.org/patch/135223/ Dec 17 02:40:21 kergoth: thank you. damn bleeding edge ^^. Dec 17 02:40:30 np **** ENDING LOGGING AT Sat Dec 17 03:00:00 2016