**** BEGIN LOGGING AT Tue Feb 22 02:59:58 2011 Feb 22 04:13:50 k, i'll poke a few things in then, thanks Feb 22 07:31:48 gm Feb 22 07:40:43 good morning Feb 22 08:31:06 how to over ride the do_install_prepend( ) in my recipe Feb 22 08:47:23 gi florian Feb 22 08:47:32 good morning Feb 22 08:47:34 sreddy? Feb 22 08:47:52 do_install_prepend you cannt override Feb 22 08:48:50 morning woglinde Feb 22 08:49:03 sreedy what will you achieve? Feb 22 08:49:05 hi ant Feb 22 08:55:11 bluelightning: morning Feb 22 08:55:23 hi Jay7, all Feb 22 08:55:52 bluelightning: did you tried to build (recent) fbreader with opie frontend? Feb 22 08:55:56 just asking Feb 22 08:56:16 Jay7: no, I haven't... does it not work now? Feb 22 08:56:35 I don't know.. I see that we have no fbreader-opie recipes.. Feb 22 08:56:38 only for x11 Feb 22 08:56:46 but I'll prefer to have one Feb 22 09:43:26 03Denis 'GNUtoo' Carikli  07master * r8e4bd2fd4c 10openembedded.git/recipes/linux/linux-2.6.34/om-gta01/defconfig: (log message trimmed) Feb 22 09:43:26 linux-2.6.34: add unwinding to om-gta01's defconfig Feb 22 09:43:26 * taken from om-gta02 change 1516588acd3c4b4dd4add71d06ab8ce0d1bafa02 Feb 22 09:43:26 Adding unwinding remove CONFIG_FRAME_POINTER while preserving its Feb 22 09:43:26 debug functionalities at the cost of a bigger image Feb 22 09:43:26 (that we mitigate by using LZMA compression instead of gzip) Feb 22 09:43:27 Pratically speaking that should result in speed improvements: Feb 22 10:23:11 gm hrw Feb 22 10:23:26 gm hrw Feb 22 10:24:00 hi djwillis Feb 22 10:24:11 Hi woglinde_ Feb 22 10:29:24 morning Feb 22 10:31:26 03Koen Kooi  07org.openembedded.dev * r0c1366f3bb 10openembedded.git/recipes/opencv/ (2 files in 2 dirs): Feb 22 10:31:26 opencv 2.2: unbreak SIFT for ARM Feb 22 10:31:26 https://code.ros.org/trac/opencv/ticket/910 Feb 22 10:31:26 Signed-off-by: Koen Kooi Feb 22 11:49:23 hello all iam new to embedded environment can any body tell me what does bsp's means and Feb 22 11:49:28 what will be the work on that Feb 22 11:49:29 ..? Feb 22 11:51:47 board support package Feb 22 11:52:34 hrw: that is abbrevation for that.. Feb 22 11:53:00 but what would be the work on that.. Feb 22 11:53:57 vlrk: http://en.wikipedia.org/wiki/Board_support_package Feb 22 11:54:06 vlrk: machine definition, kernel config (and possible patches), etc - see the Yocto documentation http://www.yoctoproject.org/sites/default/files/bsp-guide_7.pdf Feb 22 11:55:36 incandescant:thx Feb 22 12:39:39 * mwester decides he hates autoconf. Feb 22 13:39:55 I'm working on a user config to switch between dhcp and static ip. Are there any utils that can help me with that, or must I build my own /etc/network/interfaces files? Feb 22 13:53:26 anr78, it's easier to build your own than to try to get NetworkManager (aka NetworkMangler) working with all its dependencies and assumptions. Feb 22 13:54:22 mwester: yeah. just remembered how much I have struggled with that one on Ubuntu, so it will be a nice little python script :) Feb 22 13:55:14 But, that's my opinion -- there are others who strongly believe that embedded devices should be able to handle desktop tools; as evidenced by the increasing number of base packages that now pull in dbus, and therefore xlib, and a plethora of similar stuff. Feb 22 13:57:11 :q Feb 22 13:57:18 oopsie Feb 22 13:57:45 I was wondering if you were makeing some sort of cryptic comment on network manager, eFfeM_work :D Feb 22 13:57:56 nope Feb 22 13:58:22 i've not really played with it, I'm a cmd line person :-) Feb 22 13:59:01 As am i, which is probably the root cause of my distaste for the "new" way to manage services on Linux (dbus) Feb 22 14:03:08 there's also connman; it's supposed to be simpler than nm Feb 22 14:08:47 Doesn't connman also require dbus? Feb 22 14:08:48 re Feb 22 14:14:16 mwester: if you are opposed to having dbus installed then it won't be suitable, no... Feb 22 14:19:11 I run without X, so command line will be fine :) Feb 22 14:32:42 re Feb 22 15:25:31 If I'm using sane-toolchain.inc which defines the prefered gcc as 4.5 why is it: 4.5-r31.2+svnr170123? Feb 22 15:26:27 because the 4.5 recipe is at svn Feb 22 15:26:59 and thats a sane default? Feb 22 15:27:23 Yes Feb 22 15:27:29 Since 4.5 is a released branch of gcc Feb 22 15:27:38 And they have clear guidelines Feb 22 15:27:48 ahh so the branch is considered stable Feb 22 15:27:53 * ajb_oe sees Feb 22 15:29:35 ajb_oe: OE gcc-4.5 takes code from fsf svn repo and adds lot of linaro patches (which speedup arm) Feb 22 15:30:10 * DJWillis wonders how 'stable' bbappend and recipe priority is these days (with BB 1.12) Feb 22 15:48:49 03Mike Westerhof  07org.openembedded.dev * rb162231b35 10openembedded.git/recipes/squid/squid_3.1.9.bb: Feb 22 15:48:49 This patch changes hard-coded references to /usr and /opt in the various Feb 22 15:48:49 squid config.test scripts to point to the correct sysroot paths. Feb 22 15:48:49 This version accommodates the unusual staging dir layout used by micro. Feb 22 15:48:49 Signed-off-by: Mike Westerhof Feb 22 15:48:49 Acked-by: Khem Raj Feb 22 15:48:54 03Mike Westerhof  07org.openembedded.dev * r1f71b10332 10openembedded.git/recipes/meta/slugos-packages.bb: SlugOS: slugos-packages - reinstate squid, add task-sdk-native Feb 22 15:58:23 * ajb_oe frowns at a failing groff build Feb 22 15:58:45 ajb_oe: apropos your libnet thing, have you tried just setting RDEPENDS_${PN}-dev = "" to stop libnet-dev getting the dependency on libnet in the first place? Feb 22 16:07:22 pb_: no, no-one suggested it at the time. I'll give it a spin Feb 22 16:22:32 woglinde : thanks i got it Feb 22 16:27:13 sreddy he okay Feb 22 16:31:08 woglinde : Apart from adding to DEPENDS +=" liboil " and adding -I${LIBOILPATH} what modification i need to do in my recipe to make it compile with a gicen xyz.so file Feb 22 16:49:28 Anyone around that can delete branches? Feb 22 16:50:12 03Tom Rini  07master * r9c5b9db297 10openembedded.git/conf/bitbake.conf: Feb 22 16:50:12 bitbake.conf: Simplify TARGET_CPPFLAGS/LDFLAGS Feb 22 16:50:12 This is based on poky's ba2e1f4d933c37b372d6749d64614f2510ee9d7b, which Feb 22 16:50:12 simplifies TARGET_CPPFLAGS (and thus CFLAGS) and TARGET_LDFLAGS Feb 22 16:50:12 based on gcc any company having --sysroot Feb 22 16:50:13 Acked-by:Khem Raj Feb 22 16:50:13 Signed-off-by: Tom Rini Feb 22 16:51:47 Tartarus: I dont know how to work git to check if this made it in, but there was another obstacle I had found during that process Feb 22 16:52:05 Tartarus, I could probably do it but cbrake would be better. Feb 22 16:52:32 k, thanks Feb 22 16:52:33 conf/distro/include/sane-toolchain.inc contained a TARGET_CC_ARCH_pn-cairo_armv5te = "-march=armv4t" which was a problem for me, may not be necessary Feb 22 16:52:58 demigod2k, yeah, there's a few things like that which do need an audit again Feb 22 16:53:01 cbrake, can you help Tartarus? Feb 22 16:53:51 Tartarus: sure, which ones? Feb 22 16:54:08 trini/simplify-target-flags Feb 22 16:55:46 Tartarus: done Feb 22 16:56:16 thanks Feb 22 17:08:03 Is anyone building groff in their tree, I have no idea why this build is failing? Feb 22 17:08:13 that BB_THREADS made a huge difference.... load average: 10.08, 7.49, 6.60 Feb 22 17:08:22 in thats case BB_THREADS = 4 Feb 22 17:08:27 wow, bitbake 1.12 has progressbar and ETA Feb 22 17:13:16 hrw you didnt use master for a while? Feb 22 17:13:49 woglinde: I did not used OE for quite long time Feb 22 17:13:59 oh Feb 22 17:14:10 it even has parallel parsing Feb 22 17:14:15 thanks to kergoth Feb 22 17:37:09 Good evening everybody Feb 22 17:41:25 hi, regarding http://patches.openembedded.org/patch/589/ Feb 22 17:41:34 when is the review going to be completed? Feb 22 17:42:57 filip when some is doing it? Feb 22 17:43:05 try to ping on the ml Feb 22 17:44:45 woglinde: I'm just curious about the procedure Feb 22 17:45:41 ? Feb 22 17:45:53 there is nothing special Feb 22 17:45:58 if some one looks at it Feb 22 17:46:01 mostly khem Feb 22 17:46:05 and give his ack Feb 22 17:46:12 he also is commiting it Feb 22 17:46:22 if nobody looks Feb 22 17:46:25 nothing happens Feb 22 17:46:38 okay Feb 22 17:48:29 and if nothing happens for long time then you ping again Feb 22 18:00:19 Hey guys, short question: The "micro" distribution (conf/distro/micro.conf) is unmaintained and not guaranteed to build, right? Feb 22 18:01:00 hrw or give up Feb 22 18:04:21 And what about conf/distro/minimal.conf? Is it still valuable as a starting point to build a distro? Feb 22 18:04:39 minimal should work Feb 22 18:04:55 but what is the problem you are going to solve? Feb 22 18:05:27 jo jay7 Feb 22 18:05:30 lackS: minimal builds a TON of stuff that I wouldn't really describe as minimal, which has all been obstacles for me Feb 22 18:05:58 lackS: and because it's so hierarchical it can be a real challenge to comment out things which are blocking the build Feb 22 18:06:46 I think its a better starting point than anything else but beware the name is deceiving Feb 22 18:07:07 libmpc refused to build because it did not recognise a proper version of gmp (one of its dependencies) Feb 22 18:07:28 minimal is supposed to be the minimal things defined in a distro file Feb 22 18:07:30 Stop thinking about it, I'm currently doing a clean bitbake task-base with my config Feb 22 18:07:34 I think it is out of hand again Feb 22 18:07:37 demigod2k for you too building != what ends uo in image Feb 22 18:08:11 I was a bad guy and interrupted the building process to change my config file several times, so I guess a clean start is the best way to see if the problem really exists or was just caused by my messing Feb 22 18:08:12 there are some deps mostly bluetooth which drags in the gnome because fo gstreamer Feb 22 18:08:26 bette bluetooth-sound drags it in Feb 22 18:08:30 woglinde: that could be. I'm retrying with a -k right now hoping it'll skip ahead and be able to create an image anyway. Two weeks in and I'm still working on that :( Feb 22 18:08:35 search the ml for diskussion Feb 22 18:08:39 about it Feb 22 18:08:46 ? Feb 22 18:08:46 woglinde: other thing too is I'm starting with "bitbake console-image" which might not be the best starting point? Feb 22 18:08:47 demigod2k, what image are you trying to build? Feb 22 18:08:56 woglinde, to whom are you talking? Feb 22 18:08:57 console-image. is there a better more minimal choice? Feb 22 18:09:08 probably Feb 22 18:09:11 demigod2k Feb 22 18:09:14 thx Feb 22 18:09:21 pretty much all I want is busybox in any sort of filesystem in order to prove-out that I can build something. console seemed small Feb 22 18:09:23 what do yuo need in an image? Feb 22 18:10:19 Crofton|work: almost nothing yet. I'm still trying to prove-out that we can switch-off of buildroot. later I'll worry about cherry-picking packages that I want Feb 22 18:10:34 lackS you might get distro where the false versions are pinned Feb 22 18:10:47 lacks try angstroem 2010.x Feb 22 18:10:51 and console-image Feb 22 18:10:56 or shr Feb 22 18:11:46 angstrom seems way to heavy for my purpose... or is there a quick way to build an angstrom image with a minimal set like busybox + some more (<20) packages? Feb 22 18:11:52 Crofton|work: the challenge for example has been that console-image pulled in this libfso (smartphone?) which doesn't built with older glibc and I'm using an external toolchain, etc, etc, etc. Feb 22 18:12:11 lacks *sigh* Feb 22 18:12:26 woglinde, sorry if I appear to be dumb Feb 22 18:12:26 angstroem is distro it not defines whats in an image Feb 22 18:12:39 hmm Feb 22 18:12:47 I wonder why console-image gets libfso Feb 22 18:12:53 it defines versions and some other strategies Feb 22 18:12:56 I know that angstrom is a distro. Question: How can you define what packages will be included in an image? Feb 22 18:12:58 console-imag is not designed to be super small Feb 22 18:13:02 crofton -g Feb 22 18:13:03 but that sounds funny Feb 22 18:13:25 and examine the pn- file Feb 22 18:13:25 woglinde: I've got the same misunderstanding I suppose. If I did a "bitbake -k console-image" is it possible that tasks will fail to build but it'll still spit out a jffs2 image? Feb 22 18:13:40 demigod2k it should fail Feb 22 18:13:44 not Feb 22 18:13:51 why is it failure at your side? Feb 22 18:14:22 urg, I need to debug this fpga though Feb 22 18:14:44 hehe Feb 22 18:14:49 woglinde: for example I "bitbake console-image" tries to build several pacakges which have an unexpressed dependency of glibc > 2.8. those packages fail to build and apparantly they're needed for hte image Feb 22 18:14:53 was there a release today? Feb 22 18:14:53 verilog verilog verilog Feb 22 18:15:04 wrong channel..sorry Feb 22 18:15:05 glibc 2.8? Feb 22 18:15:16 what distro are you using? Feb 22 18:15:23 try angstroem 2010.x Feb 22 18:15:25 woglinde: I'm working on resolving those things but the catch is that it's including all this stuff that I may or may not care to have in the image, but because the packages won't compile it ultimately blocks me from progressing to building an entire image Feb 22 18:15:28 choose uclibc Feb 22 18:15:32 and uclibc git version Feb 22 18:15:39 woglinde: the other trick is that I must use an external toolchain. it's not possible to build one for my chip Feb 22 18:15:45 ah right Feb 22 18:15:49 he Feb 22 18:15:54 its external toolchain Feb 22 18:16:16 blame your vendor Feb 22 18:16:18 not us Feb 22 18:16:19 libfso ends up in the image, or it just tried to build it? Feb 22 18:16:32 crofotn mostly only build Feb 22 18:17:06 woglinde: could be. the ultimate difficulty is that it's pretty difficult for me to track down which *.bb contains some line that decides to add each thing, and whether or not I wanted that library which failed to build it's difficult for me to remove it Feb 22 18:17:19 woglinde: I don't blame you. it's my problem in the end to solve it's just not trivial to solve Feb 22 18:17:47 demigod2k remove bt support Feb 22 18:18:00 so the set should be small Feb 22 18:18:10 Crofton: would appear to be in the image since it was listed in the *.dot file, tried to build, and with a -k it won't spit out a jffs2 image but honestly I'm too new at this to track down much more detail Feb 22 18:19:10 is there a micro-image? Feb 22 18:20:15 yes there is an I'm trying it right now! Feb 22 18:20:40 my last build choked because it couldn't build tcpdump Feb 22 18:22:07 I haven't been able to run graphviz yet on *.dot but I think it would also be more obvious if for a given image I could see "ok task-base includes... task-cli-image includes..." etc so it's more obvious why a package gets included and how it's disabled Feb 22 18:23:20 it's always a tradeoff. of course it becomes unmanageable if every target had a list 1 mile long that named every image include individually but right now you have to pick through several layers of hierarchy to figure out why a certain package got picked up. it can be quite difficult Feb 22 18:26:00 Crofton: both micro-image and minimal-image built and spit out an image for me. I'll try flashing them onto my board in a sec. THANKS for the tip, that really was simpler than console-image and I may be recommending that as a better starting point for the wiki Feb 22 18:26:54 demigod2k, it really depends on what yuo are trying to do Feb 22 18:27:05 console-image does dar more useful things for many users Feb 22 18:28:05 true. perhaps I mean more of a split that if you're a board vendor looking to bring-up, start simple, and do something equivalent to buildroot then that's a better example Feb 22 18:28:29 if you're a hobbyist or somebody that we'd consider in the aftermarket, then they want a different path because OE may already be supporting the piece of hardware they have in hand Feb 22 18:31:01 jo kergoth Feb 22 18:31:53 demigod2k, which hardware are you trying to run? Feb 22 18:32:37 demigod2k its a wiki you can edit it too Feb 22 18:33:15 demigod2k and you cannt run the produced .dot files with graphiz really Feb 22 18:33:19 they are to dense Feb 22 18:33:27 I look at them with vi or less Feb 22 18:33:29 lackS: we make a few atmel sam9261 based products using buildroot (240mhz / 16mb nor / 64mb sdram). switching to Marvell 78200 (1 ghz / sata / ddr2 socket) for the upcoming stuff Feb 22 18:34:30 woglinde: ya I will glady work on docs as soon as I'm actually building code that executes :) the *.dot I noticed that too. I was thinking if it was possible to do like the image to the package name level it might work. they are super dense though Feb 22 18:35:04 demigod2k the pn- file is okay Feb 22 18:35:13 demigod2k: Sounds fine, thanks for the info. I'm playing around with a Marvell PXA320 (806 MHz / 128 MB RAM/ 1 GB internal flash HD) Feb 22 18:35:21 I am mostly find now the faulty deps I am searching for Feb 22 18:36:22 hm good point the pn- looks smarter. what does pn- stand for though? Feb 22 18:36:47 packagename Feb 22 18:36:54 or dont know Feb 22 18:37:19 pxa is really outdated Feb 22 18:37:19 hi woglinde Feb 22 18:37:20 *g* Feb 22 18:37:34 seems my ISP connection is bad today :( Feb 22 18:38:21 lackS: At first I was going to run away screaming from the chips but I've grown to accept them. Poor low-power performance has been my main complaint so far Feb 22 18:38:59 well plus the inadequate documentation and how far things have forked from any mainline linux support Feb 22 18:40:25 demigod2k, also follow the yocto integration discussions Feb 22 18:40:39 your observations should be useful to the people doing that work Feb 22 18:41:40 I should comment that the work here IS good, which is why we're looking at it. I think the whole ipkg thing to build a root filesystem is a serious, serious improvement over how buildroot works Feb 22 18:42:58 woglinde: It's a Toradex Colibri and lives on a SODIMM. We have pretty strict requirements concerning device size, that's why we're choosing it. Feb 22 18:46:08 Crofton: Just another general comment, I think the distro is a very foreign concept to somebody in my shoes. Our usual approach is to select a bunch of packages in buildroot and create a squashfs filesystem that gets flashed onto a device Feb 22 18:47:01 lacks we already hat the discussion Feb 22 18:47:03 *g* Feb 22 18:47:10 Yep... ;-) Feb 22 18:47:13 I pointed you to the tegra moduls Feb 22 18:47:35 I know. Unfortunately, choice of hardware is not my decision. Feb 22 18:49:40 Crofton: so where the wiki suggests not to start with certain things, in my shoes the very first thing you want to do is clone and modify a couple files (distribution conf, image conf, external toolchain conf, etc) to specify what you're building Feb 22 18:50:38 woglinde, the Toradex Tegra2 are not production state btw... engineering samples only Feb 22 18:54:27 yes Feb 22 18:56:11 re Feb 22 18:58:37 03Tom Rini  07master * r52d750320e 10openembedded.git/recipes/strongswan/strongswan_4.5.1.bb: Feb 22 18:58:37 strongswan: Add version 4.5.1 Feb 22 18:58:37 strongSwan is an alternative to openswan. Feb 22 18:58:37 Signed-off-by: Tom Rini Feb 22 19:12:53 03Steffen Sledz  07org.openembedded.dev * r33912e223d 10openembedded.git/recipes/jansson/jansson_1.3.bb: Feb 22 19:12:53 jansson: new version 1.3 Feb 22 19:12:53 Signed-off-by: Steffen Sledz Feb 22 19:14:12 general question, OE built an /sbin/init which depends on libc.so.6. is it typical for unix that /sbin still depends on the libc? Feb 22 19:14:54 its ok it resulted in a failure because I forgot to specify that libc needed to go into the image but I was surprised the binary wasn't completely static Feb 22 19:19:41 Anyone here do linux power regulator device drivers? Feb 22 19:20:08 GatorBoz: lrg maintains whole subsystem iirc Feb 22 19:20:18 is that Liam? Feb 22 19:20:57 i've got a regulator that is always on Feb 22 19:21:05 but the PMIC that is compiled in doesn't exist Feb 22 19:21:30 so I think that is why the driver for a device is not working Feb 22 19:24:22 hrw: has lrg been on today? Feb 22 19:29:30 ~seen lrg Feb 22 19:29:35 lrg <~lrg@slimlogic.co.uk> was last seen on IRC in channel #oe, 71d 3h 42m 7s ago, saying: 'RP: congrats - excellent news !'. Feb 22 19:34:39 :( bitbake 1.12.0 hangs when starting the parse on my buildhost... Feb 22 19:35:13 Looks like it creates 4 threads, which promptly enter a [defunct] state, and then the main thread needs to be killed to get it back. Feb 22 19:35:30 no helpful info from -D -D -D -D debug flags. Feb 22 19:39:40 mwester: do u have assume provided in your local.cong Feb 22 19:39:47 local.conf Feb 22 19:42:01 mwester, what host and what python version? Feb 22 19:43:41 khem, yes I do. Tartarus, Fedora 12 (64-bit), python (host) 2.6.2 Feb 22 19:44:34 * ajb_oe wonders why bitbake can't install udhcp into his image when the package is there Feb 22 19:44:36 64bit + 2.6.5 is OK, so that's strange Feb 22 19:46:01 Ok, removing the assume provided doesn't change it. Feb 22 19:46:19 can you pastebin what's left? Feb 22 19:46:25 But I'm being called out to a meeting -- so this will have to wait until I get back in an hour or so. Feb 22 19:46:29 k Feb 22 19:47:40 http://pastebin.com/R292AM7C Feb 22 19:47:52 :) Now I must dash off! Feb 22 19:52:50 hi guys. i'm currently struggling with some xorg modules. i can't find the pointer where i should look. So here is my problem using bitbake 0.10.2 and the corresponding xorg-1.9.3 stuff. when starting xorg it complains about missing int10 lib and other libs. they are definatelly in the image. if i change the recipie not to make a module of it it works as expected. Feb 22 19:53:03 mwester, good news / bad news, it's not a local.conf thing. That + BBFILES is parsing away on my u10.10 64bit box Feb 22 19:53:06 er, 10.04 Feb 22 19:57:01 ok. i re-enabled hal. thought i can get rid of it. need more coffee. sorry for the noise Feb 22 19:57:25 hal should be replacable, heh Feb 22 19:58:01 I'll leave for today guys Feb 22 19:58:20 Have a nice day/eve/night (whatever timezone you are in ;-) and see you! Feb 22 19:58:22 --off Feb 22 19:59:22 Tartarus, i thought i had it already running without hal. Feb 22 20:00:22 Tartarus, it seems i still have problems with that int10. Feb 22 20:01:20 Tartarus, do you get enough information about the builds from hudson to be useful to us for solving OE problems? Feb 22 20:01:20 \ Feb 22 20:04:48 paepke, you need other stuff instead, iirc Feb 22 20:04:57 ka6sox, well, it depends on disk space, really Feb 22 20:05:16 internally, for product stuff, we archive $TMPDIR/work/*/*/temp Feb 22 20:05:22 along with console output and so on Feb 22 20:05:36 But that takes a lot of space Feb 22 20:05:38 Tartarus, i'm currently digging through the xorg recipe. Feb 22 20:05:48 Tartarus, ya, that does... Feb 22 20:05:57 For community builds, I just capture console output and then sometimes it does and sometimes I need to reproduce the problem Feb 22 20:06:32 I guess I look at it like this Feb 22 20:06:39 Hudson/Jenksin can do jobs written in python Feb 22 20:07:35 So, ${insert dev effort here} and we can know what failed and then archive just the logs for that recipe Feb 22 20:07:56 (since sometimes you want log.do_configure to see what went off for log.do_compiles strange looking failure, and so on) Feb 22 20:09:09 okay, can we get that by "package" or only for the entire build? Feb 22 20:10:39 without some sort of mangling or post-processing, we can only get it all, or just console Feb 22 20:11:41 I"m trying to figure out how to write a plugin that "manages" the offsite builds by "watching" them but they don't have to stay connected (ala JNLP) Feb 22 20:11:52 ok Feb 22 20:12:02 community builds are easy Feb 22 20:12:22 so our server would be the slave for the master but would report that things are fine out in build land. Feb 22 20:12:22 just use remote API and post every package status + logs Feb 22 20:12:47 but batched up...so by package. Feb 22 20:12:58 but I'm sure Hudson/Jenkins can't produce reports I've noted on wiki page Feb 22 20:14:10 * ka6sox dunnos...I'd rather keep the "custom" un-maintainium stuff down to a minimum. Feb 22 20:15:14 I own one plugin so far and it's not too bad on the plugin maintainers Feb 22 20:18:23 hm.. jenkins + jasper-repors Feb 22 20:18:30 or even jasper BI suite Feb 22 20:24:18 okay so this is a "work in progress" Feb 22 20:24:34 after SCALE I'll have a bit more time to setup something with jenkins on it. Feb 22 20:29:04 * Jay7 is looking on ka6sox's word 'un-maintainium' Feb 22 20:29:15 it's pretty descriptive :) Feb 22 20:31:34 Jay7, it does describe "systems" that are architected by 1 person, are not documented well, and which, over time, fall into disrepair and then either break or go out of control. Feb 22 20:33:12 ka6sox: IMHO we should collaborate with yocto here Feb 22 20:33:27 they are using? Feb 22 20:33:50 now they are using buildbot (master + slaves) Feb 22 20:34:13 afaik, they have no client-master mode Feb 22 20:34:43 but we may discuss it together Feb 22 20:35:06 okay I'll ask RP when I see him. Feb 22 20:35:32 but hopefully its more forgiving of non-fulltime connected slaves. Feb 22 20:35:53 ka6sox: I'll ping him on #yocto Feb 22 20:36:17 I suppose I could go there unless its a closed thing. Feb 22 20:37:39 channel is open Feb 22 20:37:51 I am here? Feb 22 20:38:47 RP__: we are trying to choose right software for QA/build-testing Feb 22 20:39:03 I've posted link to wiki with our req's on #yocto Feb 22 20:39:43 ok, I'll reply on #yocto Feb 22 20:40:06 ka6sox: join it please :) Feb 22 20:40:54 hi rp Feb 22 20:44:21 is there an explanation anywhere for what INSANE_SKIP_* does? I'm trying to fixup my external-toolchain-* recipe to actually put the glibc on the target and can't find that in the user manual Feb 22 21:09:44 Tartarus, khem : any news on http://patches.openembedded.org/patch/589/ ? Feb 22 21:31:46 filip, ask khem, he's the one that's listed as the reviewer Feb 22 21:32:16 Tartarus: I've updated BuildStats page with log from #yocto Feb 22 21:32:25 please look and join #yocto :) Feb 22 21:39:35 03Tom Rini  07master * rc120ad4cab 10openembedded.git/site/ (13 files): Feb 22 21:39:35 site: Move rpppoe_cv_pack_bitfields to common-linux Feb 22 21:39:35 Signed-off-by: Tom Rini Feb 22 21:53:38 minimal-image got me to a command prompt -- there's progress! Feb 22 21:55:39 03Tom Rini  07master * r1fca8dc50a 10openembedded.git/conf/bitbake.conf: Feb 22 21:55:39 bitbake.conf: Define tar.xz image types. Feb 22 21:55:39 Make it depend on both tar-native and xz-native to make sure Feb 22 21:55:39 that we get a tar that groks --xz. Feb 22 21:55:39 Signed-off-by: Tom Rini Feb 22 21:56:21 I've also decided the *.dot files are a heck of a lot more useful if you strip everything with a -native in the name, so you're only looking at things which go onto the target Feb 22 21:57:42 hehe Feb 22 21:57:53 hi ant Feb 22 21:58:21 if you leave the -native in there you get the not-so-useful clutter that darn near everything in the entire project depends on autotools-native, etc Feb 22 21:58:26 hey woglinde Feb 22 22:04:41 Tartarus: ok, thanks Feb 22 22:14:30 every recipes/images/*.bb has IMAGE_INSTALL. How does that relate to what's in the user manual (PACKAGE_INSTALL, etc)? Couldn't find that variable name anywhere in the manual Feb 22 22:14:34 did it get changed at some point? Feb 23 02:55:18 Can someone try to build libcap2 on current OE? **** ENDING LOGGING AT Wed Feb 23 02:59:56 2011