**** BEGIN LOGGING AT Thu Jul 04 02:59:59 2013 Jul 04 08:22:54 good morning Jul 04 08:47:51 hello everybody Jul 04 08:48:09 I am seeing a problem with avahi recipe Jul 04 08:48:43 if we configure or system on systemd then we are suppose to not include sysvinit stuff Jul 04 08:49:15 but following line in avahi recipe pull sysvinit and friends Jul 04 08:49:17 RDEPENDS_avahi-daemon = "sysvinit-pidof" Jul 04 08:49:50 do you think that this is correct? Jul 04 08:53:32 Noor: you'd have to use "git blame" to find out why that was added Jul 04 08:54:08 bluelightning: I was doing that and it was added by 59a08907eafffde664079b9a2068f47131dd3f5d Jul 04 08:54:18 avahi: Refactor recipe to move UI to avahi-ui Jul 04 08:54:47 Noor: are you sure it wasn't added earlier? Jul 04 08:54:55 no specific reason was mentioned Jul 04 08:55:11 + Jul 04 08:55:12 +RDEPENDS_avahi-daemon = "sysvinit-pidof" Jul 04 08:55:12 + Jul 04 08:55:38 may be it was added in the combined recipe Jul 04 08:55:55 digging more Jul 04 08:57:16 Yeah it was there in combine recipe as well Jul 04 08:57:28 need to look into old recipe Jul 04 09:03:07 heh this is really ancient ..... I think above thing was migrated from old openembedded layer Jul 04 09:03:26 can see RDEPENDS_avahi-daemon += "sysvinit-pido in avahi.inc file present in old openembedded Jul 04 09:09:01 Noor: sysvinit-pidof is just one command Jul 04 09:09:10 Noor: you will not get sysvinit in image Jul 04 09:18:47 hrw: Yeah .... but I was thinking can we stop building sysvinit and its friends when we are building systemd Jul 04 09:19:09 coz this bring sysvinit recipe to build that full sysvinit-tabs Jul 04 09:19:16 and then chain goes on Jul 04 09:19:50 Noor: pidof is part of sysvinit so it got built Jul 04 09:20:07 Noor: many things are built during build. not every of them ends in an image Jul 04 09:20:33 hrw: yup ... that is true Jul 04 09:24:57 hrw: just noticed that pidof is provided by busybox as well Jul 04 09:25:07 can we use that one? Jul 04 09:26:08 Noor: no idea Jul 04 09:26:27 Noor: probably yes Jul 04 09:26:38 hrw: OK we will try on our side and it it works fine then send the pull request Jul 04 09:27:10 cool Jul 04 10:22:52 Hiya... Jul 04 10:22:55 Anyone around? Jul 04 10:22:56 Anyone around? Jul 04 10:23:17 I' Jul 04 10:24:01 I'm getting confused trying to work out the difference between what look like two OE GIT repos - one on github and one via openembedded.org - can anyone give me a very short summary of their relationships? Jul 04 10:31:38 * florian refuses to comment on any github fork Jul 04 10:32:44 Fair enough - I'm not botherered about forking politics but would love to know what the upstream version is...! Jul 04 10:33:00 so I can ignore all that sorta guff... Jul 04 10:58:01 pev: openembedded.org is overloaded Jul 04 10:58:20 just because of this we suggest to clone from github Jul 04 11:00:44 Hi ant: Thanks for that - if that' Jul 04 11:01:18 s the case, why does it look like the contents of both repositories are different though? It's not just a mirror? Jul 04 11:02:22 the repos are aligned Jul 04 11:03:17 meta-oe: 46354a697 oe-core: 3e1dbabbf33a Jul 04 11:09:52 I'm a little bit confused, what is xuser and why do we need it? Jul 04 11:10:11 I can't seem to find the patches where it was initially brought in to read the background Jul 04 11:10:27 Rootless X is owned by xuser Jul 04 11:10:53 ok, so why does Connman depend on xuser, when it's not an X application? Jul 04 11:11:08 and why do I get xuser in my image, when I don't have X11? Jul 04 11:11:11 introduced with: https://lists.yoctoproject.org/pipermail/poky/2011-February/003942.html Jul 04 11:11:21 I assume, it's because I use connman Jul 04 11:21:27 heh, yes, conman and xserver-nodm-init are needing xuser Jul 04 11:22:49 reading oe-core ML of todfay, it seems there will be another user for conman possibly Jul 04 11:24:05 yeah I've just had a quick chat with Ross about it and it seems something may be done in the future Jul 04 13:21:51 hi everyone Jul 04 13:23:46 SynchronE: hello Jul 04 13:24:18 i'm looking for a proper MCU to use in this case: a 1-key composite usb-device consisting of a keyboard emulator, that sends keystrokes, configured via (and that's the second part of composite) some interface Jul 04 13:25:15 oh ... i'm probably on a wrong channel Jul 04 13:27:47 yeah this channel is for the OpenEmbedded project specifically Jul 04 13:50:04 hrw: thank you for the hint. the result would be that the 'customer' build still ends up with all the ipks? Jul 04 13:50:15 zecke: yes Jul 04 13:50:50 zecke: but remember to make use of file:// mirror for sstate-cache in customer builds so you will not 'infect' other customers Jul 04 13:51:08 zecke: and I did not tried that yet Jul 04 13:54:36 hrw: ultimately I would only like to have ipk's i built for this specific customer. but this is more like a wish (e.g. like of everything from the 'base' build is ASSUME_PROVIDED.. without playing that game) Jul 04 13:55:21 zecke: would your customer do builds or do they rely on you? Jul 04 13:55:45 zecke: you will have deploy/ipk/ with base+customer set Jul 04 13:56:23 zecke: but no one stops you from PR += "customer1" or sth like that in customer layers Jul 04 14:05:28 hrw: I will do them for now. I just want to have a more clear separation between what is generic and what is customer specific Jul 04 14:05:48 hrw: I thought about being evil and use PACKAGE_ARCH = "customer1" as sort of bruteforce Jul 04 14:07:05 zecke: ugly Jul 04 14:07:36 zecke: if they would like to do builds then give them sstate-cache from customer build (which will be base+customer probably) Jul 04 22:41:47 JaMa: about TUNE_FEATURES Jul 04 22:41:49 http://pastebin.com/5v6Kzbp3 Jul 04 22:58:54 ant_home: and? Jul 04 22:59:07 TUNE_FEATURES wasn't incorrectly expanded before showing Jul 04 23:02:56 ant_home: see bdce39f22a0e8c8e1cf237322657220e4201077d Jul 04 23:03:38 ok, it seems strange Distro has *more* packages in the feeds Jul 04 23:03:52 I thought you aimed at reducing the nr Jul 04 23:05:02 I'll finish the build overnight and verify that Jul 04 23:05:05 gn Jul 04 23:18:16 * bluelightning will backport that patch to dylan Jul 05 00:03:50 someone using systemtap-uprobes? Jul 05 00:04:58 I have kernel which has CONFIG_UTRACE=y, so do_compile is trying to build uprobes but in directory ${S}/runtime/uprobes/ which doesn't exist **** BEGIN LOGGING AT Fri Jul 05 01:50:50 2013 **** ENDING LOGGING AT Fri Jul 05 02:59:58 2013