**** BEGIN LOGGING AT Thu Jan 31 02:59:58 2013 Jan 31 08:18:48 good morning Jan 31 08:45:03 morning Jan 31 09:25:40 morning all Jan 31 09:25:53 gm too Jan 31 12:23:28 JaMa: about systemd vs openrc http://solpeth.wordpress.com/2012/12/19/systemd-vs-openrc-on-gentoo-a-completely-non-technical-users-perspective/ Jan 31 12:24:12 openrc shell based? Jan 31 12:25:19 it's C Jan 31 12:25:26 woglinde: only some parts (and init.d scripts) but core is in C Jan 31 12:28:41 its missing a feature you really want Jan 31 12:28:44 om embedded Jan 31 12:28:48 Ability to auto-restart services when they fail Jan 31 12:29:02 according to the wiki page Jan 31 12:29:28 so what does openrc arrange the shell skripts starting better? Jan 31 12:32:01 http://dev.gentoo.org/~vapier/openrc/projects/openrc/ticket/120.html Jan 31 12:32:07 woglinde: ^^ Jan 31 12:32:19 woglinde: it has declarative dependencies between services, so it can prepare right order of services and start some of them in parallel Jan 31 12:33:11 yes but it has to open subshells Jan 31 13:15:14 hey hello everybody Jan 31 13:21:09 woglinde: and no dbus... Jan 31 13:37:15 Does anybody has some feedback about oe an QT 5 ? Jan 31 14:40:22 ascor check the mailinglist it is worked on Jan 31 14:41:33 ok Jan 31 14:58:30 Hi all, I'm using the openembeded-core/meta layer and I'm trying to add python to my image. I have an IMAGE_INSTALL_append += "python", but building the image fails saying it cannot find the python package. Any idea what I've done wrong? Jan 31 14:58:46 (folk in #yocto suggested I might ask here) Jan 31 14:59:24 use e.g. "python-core" Jan 31 14:59:37 adalton: Hi - does IMAGE_INSTALL end up in the DEPENDS as well? Jan 31 15:00:23 aloril: or python-modules if you want to install also all python modules Jan 31 15:00:26 hum, how do I know? bitbake -e? Jan 31 15:00:32 adalton: ^ Jan 31 15:00:35 Does it fail while assembling the filesystem or right on start? Jan 31 15:00:51 assembling the fs Jan 31 15:01:00 florian: there isn't PN in python PACKAGES.. Jan 31 15:01:33 ah, python-core is the proper name? Jan 31 15:01:40 florian: how long is your patch queue today? :) Jan 31 15:01:47 adalton: depends on what you want to install Jan 31 15:01:56 adalton: python recipe produces a lot of smaller packages Jan 31 15:02:02 but not "python" package Jan 31 15:02:14 and IMAGE_INSTALL is about packages not recipes Jan 31 15:02:37 ah, that was actually a question I had, and I guess that's the answer. The question was, does a recipe automatically provide a package with the same name? Jan 31 15:02:44 JaMa: good point... your patches are on top of the stack but somehow life refuses to give me a little bit of free time. Jan 31 15:02:47 I'm guessing "no" Jan 31 15:03:09 but in theory it should work.. RPROVIDES_${PN}-core = "${PN}" Jan 31 15:03:21 adalton: depends on PACKAGES value Jan 31 15:03:25 adalton: usually it does Jan 31 15:03:40 florian: top of the stack is good :) Jan 31 15:03:57 JaMa: so I noted the the python recipe never did PACKAGES = "..." only += and =+, so I assumed there would be a python package Jan 31 15:04:23 adalton: see openembedded-core/meta/recipes-devtools/python/python-2.7-manifest.inc Jan 31 15:05:28 ah, nice -- I didn't know that was there :-) Jan 31 15:06:09 it was included from python recipe Jan 31 15:06:10 :) Jan 31 15:06:47 okay, so I don't see just a ${PN} in PACKAGES, I guess that explains why it couldn't find it :) Jan 31 15:07:25 adalton if you are not sure look in the deploy dir at the ipk's name Jan 31 15:07:34 you asked what I'm after. I'm looking to install a basic python runtime. I may end up needing additional modules later, but for now I'm shooting for vanilla python. That's python-core? Jan 31 15:08:36 woglinde: thanks, that's helpful too Jan 31 15:10:45 I'll give python-core a shot Jan 31 15:15:12 argh.. die modemmanager die Jan 31 15:15:38 stealing any serial port connected to the machine is evil :( Jan 31 15:16:36 bluelightning hm modemmanager was quit usefull for me Jan 31 15:16:50 well, when it's managing modems I'm sure it's fine Jan 31 15:17:08 ah now I get it Jan 31 15:17:12 but managing my dev boards when I'm trying to connect to them with a terminal is very suboptimal... Jan 31 15:17:16 tell him not to do it Jan 31 15:17:18 thanks folks -- that helped a lot Jan 31 15:17:37 woglinde: I took the radical approach of just removing it for the moment Jan 31 15:24:24 bl hehe Jan 31 16:22:19 bluelightning: same here Jan 31 16:23:25 the fun is when device restarts, appears on usb bus. modemmanager sends AT command. device does not boot because u-boot took AT command as 'lets break boot cause user wants my shell' Jan 31 16:23:35 heh Jan 31 16:23:46 and then you run picocom/etc to see what is going on Jan 31 16:24:18 and you get blocked... Jan 31 16:24:22 must be fun with devices that are less discriminating about receiving such data Jan 31 16:32:53 bluelightning: to not have modemmanager interacting with the uarts you can : edit /usr/share/dbus-1/system-services/org.freedesktop.ModemManager.service and comment out the line #Exec=/usr/sbin/modem-manager Jan 31 16:33:03 this way no more AT commands sent on the uart Jan 31 16:33:33 ericben: supposedly there's a blacklist somewhere as well but it doesn't seem like the right approach to me... Jan 31 16:35:17 bluelightning: that's the only way I found for the moment. On a dev machine that's easy and save some headache with some boards Jan 31 18:22:31 interesting... apparently monitors are prohibited on eurostar Jan 31 18:24:42 glad I checked before I tried to take one to the station... Jan 31 18:55:56 hi, is uclibc/armv7a on oe-core supposed to work on the current master? :) Jan 31 18:57:39 mnemoc yes but not with systemd Jan 31 18:59:50 woglinde: it doesn't even compile for me :< http://sprunge.us/CDcR (building core-image-minimal) Jan 31 19:06:19 hm intressting Jan 31 19:06:32 looks like khem tested only qemuarm Jan 31 19:07:11 what is your machine? Jan 31 19:07:19 and what gcc and what kernel headers? Jan 31 19:08:59 hey I just stumbled on a patch bc4da2573dfb59ea2fc4af359701818df20f7663 Jan 31 19:09:21 it introduces "MULTILIB_GLOBAL_VARIANTS".. this looks like a -really- bad idea.. hard coding the variant values Jan 31 19:09:37 I'm using variants that are not in the list on some of my systems Jan 31 19:10:18 mnemoc you could try uclibc 0.33 version not the git Jan 31 19:16:27 woglinde: machine still dummy. just importing tune-cortexa8.inc Jan 31 19:17:04 woglinde: everything else is whatever comes by default with core-image-minimal on oe-core's master Jan 31 19:17:55 woglinde: will try uclibc 0.33 after I find how :) Jan 31 19:18:25 my distro is still dummy too Jan 31 19:26:55 woglinde: 0.9.33 failed too. might it be because I don't have a kernel configured yet? Jan 31 19:38:31 mnemoc that could be Jan 31 19:38:38 try to build for beagleboard Jan 31 19:56:59 woglinde: meh :( .... ERROR: Unable to parse /srv/build/amery/x/oe-beagleboard/../oe-core/meta-ti/recipes-misc/images/cloud9-image.bb: Exited with "1" Jan 31 20:00:18 can these "parse errors" actually dependencies? ie. missing layers Jan 31 20:00:48 no Jan 31 20:00:55 bitbake checks stricter Jan 31 20:00:58 use minimal-image Jan 31 20:01:04 or dont use a image at all Jan 31 20:01:28 for quick and dirty Jan 31 20:01:43 just do a rm for all .bb which gives errors Jan 31 20:03:42 rm-ed 5 files :p Jan 31 20:58:56 woglinde: I got the same error :< Jan 31 21:05:29 mnemoc hm okay please send it to the oe-core ml Jan 31 21:05:35 so khme can take a look Jan 31 21:08:54 woglinde: ok. thanks. I'll subscribe and compose a mail now Jan 31 21:20:11 Favorite bit bake moment of the day: FATAL: Invalid user interface 'depexp' specified. Jan 31 21:20:12 Valid interfaces are 'ncurses', 'depexp' or the default, 'knotty'. Jan 31 21:20:51 It's an old bitbake. Jan 31 21:26:06 evening all Jan 31 21:26:13 * pb_ completes his tax return, phew Jan 31 21:39:29 * sr105 stills needs to do his own Jan 31 21:52:56 * mwester has decided not to pay taxes anymore. Jan 31 22:42:20 Does someone know how I can 'prepend' to a recipe in an overlay Jan 31 22:42:31 ie, '*.prepend' or something? **** ENDING LOGGING AT Thu Jan 31 22:46:41 2013 **** BEGIN LOGGING AT Thu Jan 31 22:48:22 2013 Feb 01 00:03:21 Can I move a tmp dir? Feb 01 00:03:36 other than changing TMP_DIR, anything I would need to do? Feb 01 00:21:28 hi Feb 01 00:22:09 when doing bitbake --show-versions and shows the pattern N: what does the integer N mean? Feb 01 02:35:38 epoch Feb 01 02:35:45 ah, he left **** ENDING LOGGING AT Fri Feb 01 02:59:58 2013