**** BEGIN LOGGING AT Mon Jan 03 02:59:58 2011 Jan 03 07:29:18 good morning and a good 2011 everyone ! Jan 03 08:35:25 morning, HNY'11 Jan 03 08:43:22 03Sebastian Krzyszkowiak  07master * r761cb1c799 10openembedded.git/recipes/shr/frameworkd-config-shr_git.bb: Jan 03 08:43:22 frameworkd-config-shr: bump SRCREV for nokia900 updates Jan 03 08:43:22 Signed-off-by: Martin Jansa Jan 03 08:43:33 03?ukasz Pankowski  07master * rfa82a13faa 10openembedded.git/recipes/ffalarms/ffalarms_git.bb: Jan 03 08:43:34 ffalarms: bump SRCREV to fix compilation with valac 0.11.2 Jan 03 08:43:34 Signed-off-by: Martin Jansa Jan 03 09:12:33 good morning and Happy new year Jan 03 09:23:24 morning, HNY'11 Jan 03 09:32:58 hi hrw Jan 03 09:45:02 moin zecke, frohes Neues Jan 03 09:48:12 mickey|office: guten jahre^Wmorgen Jan 03 09:49:44 hrw: :)) hi. happy new year Jan 03 09:49:53 how's the family? Jan 03 09:59:49 mickey|office: hey! Jan 03 10:03:55 mickey|office: fine, Mira got some flu but will get it Jan 03 10:04:18 mickey|office: how's yours? Jan 03 10:05:20 hrw: good. in case you didn't know yet... there's a baby coming :) Jan 03 10:05:42 there are always more... Jan 03 10:05:48 hehe Jan 03 10:06:03 no one respects old grumpy people Jan 03 10:06:07 indeed Jan 03 10:06:10 mickey|office: CONGRATULATIONS!!! Jan 03 10:07:20 hrw: :)) thanks a lot Jan 03 10:07:29 mickey|office: boy or girl? Jan 03 10:08:30 hrw: not clear yet. we'll know probably next month Jan 03 10:09:04 ah, fresh news Jan 03 10:09:44 :) Jan 03 10:10:07 well, she's in 5th month yet and last time the doctor said it's 50% for a girl ;) Jan 03 10:10:43 ;D Jan 03 10:10:58 mickey|office: so greet Sabine from us Jan 03 10:11:36 hrw: will do, thanks Jan 03 10:12:30 mickey|office: Ania asked does everything goes fine Jan 03 10:14:24 hrw: ah greet her from me, please. yes, everything ok until now. first three months she had a bit of problems with sleeping and nausea, but last two months have been just fine. Jan 03 10:15:13 good to hear Jan 03 10:17:48 kergoth: fyi "fetch: resurrect branch inclusion in git revision key" fall back doesn't work correctly.. resets every localcount to 1 here Jan 03 10:18:15 kergoth: ie: http://paste.pocoo.org/show/313608/ Jan 03 10:23:03 kergoth: ah, sorry.. maybe it was because my cache was messed after last branch inclusion in key so key with branch was found but with wrong _count, here is cache before bitbake upgrade, which explains it.. http://paste.pocoo.org/show/313610/ Jan 03 10:55:21 03Koen Kooi  07org.openembedded.dev * r83134fa031 10openembedded.git/recipes/qt4/ (qt4-embedded.inc qt4-x11-free.inc qt4.inc): Jan 03 10:55:22 qt4: switch to mysql5 to avoid header clashes in staging Jan 03 10:55:22 Signed-off-by: Koen Kooi Jan 03 10:55:27 03Koen Kooi  07org.openembedded.dev * reaf0fc03ac 10openembedded.git/recipes/mysql/mysql5_5.1.40.inc: Jan 03 10:55:27 mysql5: remove legacy staging Jan 03 10:55:27 Signed-off-by: Koen Kooi Jan 03 10:55:28 03Koen Kooi  07org.openembedded.dev * r68ec0723d8 10openembedded.git/recipes/postgresql/postgresql.inc: Jan 03 10:55:28 postgresql: remove legacy staging Jan 03 10:55:28 Signed-off-by: Koen Kooi Jan 03 12:42:03 03Koen Kooi  07org.openembedded.dev * r070cbda8e9 10openembedded.git/recipes/postgresql/postgresql.inc: Jan 03 12:42:03 postgres: partially fix dependency on host zic Jan 03 12:42:03 Signed-off-by: Koen Kooi Jan 03 13:01:46 for your entertainment. The (as usually highly nuanced) view of Koen on yocto: http://lists.linuxtogo.org/pipermail/angstrom-distro-devel/2011-January/004292.html Jan 03 13:05:16 eFfeM_work: looks like (as expected) yocto is giving no benefits to OE Jan 03 13:06:58 mckoan: I'm not too sure that it does not give benefits to OE, I think a better build system would be of use to us too Jan 03 13:07:54 and although yocto has much less recipes than we have (and I feel quite some are dearly missed like e.g. mysql), it is my impression that the overall quality of the recipes is better Jan 03 13:09:02 As such for deeply embedded systems (the ones without a UI so no need for QT or X11 or so) yocto might actually be a better choice Jan 03 13:09:41 but if you want to do QT or lots of media playback or so yocto is probably not the best option Jan 03 13:10:37 I am still in favour of yocto as OE base Jan 03 13:10:39 Hi all Jan 03 13:11:08 angelox_123: hi Jan 03 13:16:14 hrw: so am I, (and sad to see so little feedback on your proposal) Jan 03 13:16:23 angelox_123: hi Jan 03 13:17:15 hrw: i feel that using yocto as oe base would give us the best of both worlds Jan 03 13:17:49 * hrw -> lunch Jan 03 13:17:56 enjoy! Jan 03 13:38:04 Ok Jan 03 13:38:10 i'm seeing the example helloworld Jan 03 13:38:26 it create the source file with a printf command Jan 03 13:38:47 i can use a local file to compile? without use printf ? how? Jan 03 13:41:27 angelox_123: what do you mean? "i can use a local file to compile?" Jan 03 13:41:39 yes see the example helloworld Jan 03 13:41:46 and? Jan 03 13:41:49 it create the helloworld.c with printf Jan 03 13:41:54 sure Jan 03 13:42:07 then? Jan 03 13:42:18 i can use a local file (like openembedded/recipes/helloworld/files/helloworld.c) to compile? Jan 03 13:42:31 to compile what? Jan 03 13:43:10 the helloworld.c... Jan 03 13:43:52 i want to remove the printf command Jan 03 13:44:08 angelox_123: do you mean you want to modify the auto-generated helloworld.c and rebuild it? Jan 03 13:44:20 i think so... Jan 03 13:44:56 you can but it's complicate Jan 03 13:45:01 you can Jan 03 13:45:21 how? Jan 03 13:45:31 I suggest you to build any custom package outside OE with its toolchain Jan 03 13:46:07 if you want to work inside OE you have 2 options (AFAIK) Jan 03 13:46:43 1. modify the recipe helloworld_1.0.0.bb creating a new custom .c file Jan 03 13:46:57 angelox_123, yes you can Jan 03 13:47:15 put the file in files Jan 03 13:47:18 2. play with devshell Jan 03 13:47:21 and copy it in destdir Jan 03 13:47:33 ah yes devshell is easier, but les automatic Jan 03 13:47:38 *less Jan 03 13:47:53 GNUtoo|laptop,exact but i don't know the command to copy... Jan 03 13:47:55 by copy it in desdir I mean tell the recipe to do that Jan 03 13:47:55 :( Jan 03 13:47:59 install or cp Jan 03 13:48:10 cp work? Jan 03 13:48:13 good Jan 03 13:48:16 then override do_configure with : Jan 03 13:48:27 and add gcc command in do_compile Jan 03 13:48:32 like that Jan 03 13:48:40 ${CC} ... Jan 03 13:48:59 angelox_123, always use ${} Jan 03 13:49:05 ok Jan 03 13:49:07 checking here Jan 03 13:49:08 the manual references theses Jan 03 13:49:20 for instance ${D} for destdir Jan 03 13:49:39 i use ${CP} for cp command ? Jan 03 13:49:43 no Jan 03 13:49:53 just install or cp Jan 03 13:49:53 only cp? Jan 03 13:49:56 yes Jan 03 13:49:57 ah ok Jan 03 13:51:20 angelox_123, there is an example in the manual Jan 03 13:51:38 ${CC} ${CFLAGS} ${LDFLAGS} ${WORKDIR}/myhelloworld.c -o myhelloworld Jan 03 13:51:52 knowing that files:// are copied in workdir Jan 03 13:52:00 you just have to do that: Jan 03 13:52:42 cp "${WORKDIR}/myhelloworld.c" ${S}/ Jan 03 13:53:05 then override do_configure like that: Jan 03 13:53:14 do_configure(){ : } Jan 03 13:53:25 and add the do_compile that is in the manuel Jan 03 13:53:31 don't forget LDFLAGS Jan 03 13:53:39 else you have gnu hash issue Jan 03 13:54:00 hum.. ok i'll try..thanks by all help Jan 03 13:54:14 i found now the manual :) Jan 03 13:54:17 then you need a manual do_install Jan 03 13:54:23 that install with the install command Jan 03 13:54:29 to ${D} Jan 03 13:55:50 follow the do_install example in the manual Jan 03 13:56:10 ah there is even a full example Jan 03 13:57:11 angelox_123, copy the recipe of the manual then Jan 03 13:57:18 yes :) Jan 03 13:57:22 i'm trying that :) Jan 03 14:07:12 GNUtoo|laptop : Worked..but my new source code for tests have some header files dependencies and it don't compile...i tried add DEPENDS but don't worked... any idea ? Jan 03 14:07:35 show your source Jan 03 14:08:00 else you could try the usual flags Jan 03 14:08:08 -I for include Jan 03 14:08:16 -L for libraries search path Jan 03 14:08:28 -lfoo for linking to libfoo Jan 03 14:09:08 my code needs elementary.h (fast crazy change LOL) :) Jan 03 14:09:19 here the clock.c : http://pastebin.com/KNWP0YuY Jan 03 14:09:37 here the clock_1.0.0.bb : http://angelox_123.pastebin.com/YGLbqAYv Jan 03 14:10:01 ops Jan 03 14:10:15 do_stage..i tried but don't worked Jan 03 14:14:32 And... ? Jan 03 14:19:48 in do_stage you're supposed to stage your own libraries' header, not other ones. i recommend you're adding something like `pkg-config --cflags --libs elementary`to your CFLAGS instead Jan 03 14:20:38 mickey|office Sorry...how to do that? Jan 03 14:20:48 * angelox_123 is sincerity newbie Jan 03 14:21:05 well, since you're not using a buildsystem, but just compiling in do_compile directly, just add it there Jan 03 14:21:44 so remove ${CFLAGS} and add pkg-config --cflags --libs elementary ? Jan 03 14:21:57 probably not removing, but just adding Jan 03 14:22:04 and don't forget the ` Jan 03 14:22:18 since you want to add the result of calling the command, not the name of the command Jan 03 14:29:24 eFfeM_work, pong. Re your uClibc++ version remark: sounds plausible. I'll have a look ASAP Jan 03 14:34:12 does anybody know something about the status of the "nylon" distro? Is it still used or built regularly? It pins ancient versions.. perhaps we could remove it? Jan 03 15:05:19 blindvt write a proposal to the ml Jan 03 15:21:33 I think that nylon can be dropped Jan 03 15:21:41 but RFC would be nice Jan 03 15:28:57 Thanks mickey|office it worked :) and sorry by that newbie questions :( Jan 03 15:34:43 hmmm Jan 03 15:34:47 intone is down? Jan 03 15:35:33 google code seem down Jan 03 15:35:56 sure: Jan 03 15:35:58 ? Jan 03 15:36:04 maybee dns problems on your provider side? Jan 03 15:36:21 http://intone.googlecode.com/svn/trunk/ Jan 03 15:36:22 here ok Jan 03 15:36:35 I autorev that in local.conf Jan 03 15:36:48 no..likes down yes! Jan 03 15:39:04 http://code.google.com/p/intone/ Jan 03 15:39:05 hmmm Jan 03 15:40:31 but the svn is still at http://intone.googlecode.com/svn/trunk/ Jan 03 15:40:39 maybe only their svn is down Jan 03 15:59:12 blindvt` thanks (was afk) Jan 03 16:01:27 03Otavio Salvador  07master * rc9d0fef176 10openembedded.git/recipes/hotplug2/hotplug2_svn.bb: Jan 03 16:01:27 hotplug2 (svn): add Jan 03 16:01:27 Signed-off-by: Otavio Salvador Jan 03 16:01:29 03Otavio Salvador  07master * rf18bbc75bd 10openembedded.git/recipes/freerdp/freerdp_git.bb: Jan 03 16:01:29 freerdp (git): use gitpkgv and bump SRCREV Jan 03 16:01:29 Signed-off-by: Otavio Salvador Jan 03 16:01:39 03Otavio Salvador  07master * r6a0d197215 10openembedded.git/classes/gitpkgv.bbclass: Jan 03 16:01:40 classes/gitpkgv.bbclass: add GITPKGVTAG that uses 'git describe' Jan 03 16:01:40 Using ${GITPKGVTAG} allows for automatic versioning based on the Jan 03 16:01:40 repository tags. For those that doesn't want to use it, ${GITPKGV} is Jan 03 16:01:40 still available. Jan 03 16:01:40 Signed-off-by: Otavio Salvador Jan 03 16:01:41 Acked-by: Martin Jansa Jan 03 16:38:31 quick question on PN: When i try to bake busybox-static-mdev, i get NoProvider. Now, the busybox-static recipe (which inherits the normal busybox) does list busybox-static-mdev as a package. Can i build just a package directly, off a recipe, rather than the containing recipe so that my dumb attempt to use IMAGE_INSTALL="busybox-static-mdev" would work? If so, how? Jan 03 16:44:42 blindvt: you need to specify the recipe, not the package with the busybox command -- the recipe will create all the packages associated with the recipe in question. Jan 03 16:45:42 But you will reference the package, not the recipe, in your DEPENDS or RDEPENDS lines in other recipes. Jan 03 16:49:58 mwester, yea. So, to summarize, i cannot _bake_ packages, just whole recipes. I see. Thanks! Jan 03 17:04:14 have a nice rest of day Jan 03 17:04:58 mwester: DEPENDS references recipes, RDEPENDS/IMAGE_INSTALL/PACKAGE_INSTALL reference packages Jan 03 17:25:05 hrw, thanks -- that's a nuance I have missed all this time. It also leaves me even more confused than I was regarding Koen's responses to my proposed e2fsprogs patch. Jan 03 17:28:59 * Tartarus tries to enjoy one more day of vacation and not think about what craziness he's missed on MLs Jan 03 17:34:38 * Crofton|work kicks Tartarus from the channel Jan 03 17:34:57 lol Jan 03 17:35:09 Crofton|work: ! Jan 03 17:35:40 florian, he is trying to vacation Jan 03 17:35:44 I am helping him Jan 03 17:35:48 anyway, no ops Jan 03 17:35:52 hehe Jan 03 17:39:50 03Koen Kooi  07org.openembedded.dev * r725c1e52a0 10openembedded.git/recipes/linux/linux-omap4/omap4430-panda/defconfig: Jan 03 17:39:50 linux-omap4: tweak pandaboard defconfig Jan 03 17:39:50 Signed-off-by: Koen Kooi Jan 03 18:16:10 how can i make it so i don't have to update my bitbake? Jan 03 18:16:28 whta? Jan 03 18:16:36 right now it seems bitbake refs against something to notify itself it its out of date or something Jan 03 18:17:04 or maybe everytime i do a bitbake, OE gets updated, all i know is its annoying Jan 03 18:17:23 and its breaking by build environment that was good till the end of dec Jan 03 18:17:31 now its busted again Jan 03 18:18:03 git pull for oe and bitbake Jan 03 18:18:10 dont know whats the difficult for it Jan 03 18:18:19 yeah, i want to not have to do that Jan 03 18:18:30 dont use oe if you dont want Jan 03 18:18:35 i had a perfectly stable env to build my stuff in with oe Jan 03 18:18:41 dinner now Jan 03 18:18:50 umm, why does OE want to update itself when i do my bitbake Jan 03 18:18:51 with oe-dev? Jan 03 18:18:57 use stable release Jan 03 18:18:59 i didn't that i had oe-dev Jan 03 18:19:06 ? Jan 03 18:19:15 i thought i had pulled down oe-stable Jan 03 18:19:35 dont know if oe-stable is the oe 12/2010 release now Jan 03 18:19:43 please check it Jan 03 18:21:17 obi Jan 03 18:29:00 gm Jan 03 18:36:29 khem, paulepanter, you updated DP of glib-2.0, thanks! 2.26.1 needs to be taught to handle !IPv6. Until this is applied upstream (current seems to be 2.27.5, btw), may i ask somebody who actually uses that thing to test and eventually apply https://bugzilla.gnome.org/show_bug.cgi?id=638593 ? Jan 03 18:49:52 03Frans Meulenbroeks  07org.openembedded.dev * r013ab198dc 10openembedded.git/recipes/tgt/tgt_1.0.10.bb: Jan 03 18:49:53 tgt: updated to 1.0.12 Jan 03 18:49:53 Signed-off-by: Frans Meulenbroeks Jan 03 19:19:17 who maintains the gstreamer recipes? Jan 03 19:23:19 seems like do_split_packages has a flaw when called from multiple packages that store their files in the same dir, as such is done in gst-plugins-package.inc - all the gst-plugins are marked as being provided by each base/good/bad/ugly package Jan 03 19:52:51 how to compile a qt4 project ? Jan 03 19:53:00 i mean: make a bb file of it Jan 03 19:53:08 i tried: http://www.developmentboard.net/index.php/article/Creating+Qt4+project+with+bitbake+for+Mini2440V2 Jan 03 19:53:34 but i don't have the string "NAMES ${QT_MODULE} ${QT_MODULE}4" and i get a error Jan 03 19:53:47 like | CMake Error: The source directory "/home/angelo/openembedded/angstrom-setup-scripts/build/tmp-metano/work/armv5te-oe-linux-gnueabi/lcdslider-1.0-r0/lcdslider-1.0" does not appear to contain CMakeLists.txt. Jan 03 19:53:47 | Specify --help for usage, or press the help button on the CMake GUI. Jan 03 19:53:47 | ERROR: Function do_configure failed Jan 03 19:54:10 look at other qt-apps recipes Jan 03 19:56:10 good nite Jan 03 19:56:42 any body knows any qt-app recipe name? Jan 03 20:00:29 hey. i updated my bitbake, now i get this error about a conffile in the nw package is different from the older one and the new one will be used Jan 03 20:00:33 then the build fails Jan 03 20:01:20 angelox_123, grep Jan 03 20:01:25 use grep Jan 03 20:01:41 to findout what recipe depends on qt4 Jan 03 20:06:29 Hello Jan 03 20:06:38 I am having somethings strange in my "layer" Jan 03 20:06:53 probably something quite stupid from my side, but anyhow I'd like to ask here for it Jan 03 20:07:31 So say to we check :) Jan 03 20:07:31 define strange Jan 03 20:07:32 it looks like I am failing to generate image recipes in the layer Jan 03 20:08:02 it might be because I've created a: foo/recipes/images/foo.bb Jan 03 20:09:19 otavio, i had did the same thing that you. and i didn't get any error.. can you show to us? Jan 03 20:10:24 angelox_123: sure; look at http://pastebin.com/t0csHFUQ Jan 03 20:11:09 hum Jan 03 20:11:35 which directory are: "osnetboot-video-geode.bb" ? Jan 03 20:11:58 s/are/is/ Jan 03 20:15:48 ops Jan 03 20:16:06 what directory the file "osnetboot-video-geode.bb" is? Jan 03 20:17:40 angelox_123: ossystems/recipes/images/ Jan 03 20:17:48 angelox_123: it is inside of oe/ Jan 03 20:18:01 hum... Jan 03 20:18:48 and you are bitbaking what? i mean,when you type: bitbake "something",what is that "something"? Jan 03 20:19:08 angelox_123: it Jan 03 20:19:13 angelox_123: osnetboot-video-geode Jan 03 20:19:27 hum...having a conflict in somewhere Jan 03 20:20:00 check osnetboot-video-geode.bb and see if SRC_URI parameter have osnetboot-video-geode.bb Jan 03 20:20:23 angelox_123: sorry; it is inside ossystems/recipes/osnt/ Jan 03 20:20:33 ah ok Jan 03 20:20:37 angelox_123: and I only have IMAGE_INSTALL on it Jan 03 20:20:45 let me think a bit Jan 03 20:20:47 :) Jan 03 20:21:01 so you are: Jan 03 20:21:11 I am at oe/ Jan 03 20:21:13 and do: Jan 03 20:21:18 bitbake osnetboot-video-geode Jan 03 20:21:31 ok Jan 03 20:21:42 try: Jan 03 20:21:45 ops Jan 03 20:21:49 so you are doing this: Jan 03 20:21:54 bitbaking a image named osnetboot-video-geode Jan 03 20:22:02 that it includes the osnetboot-video-geode.bb Jan 03 20:22:03 ok Jan 03 20:22:25 can you give the image named osnetboot-video-geode ? Jan 03 20:22:35 or no? if no we can find another solution Jan 03 20:22:50 angelox_123: without .bb? Jan 03 20:23:12 the image,on ossystems/recipes/images/osnetboot-video-geode Jan 03 20:23:17 .bb i forget Jan 03 20:23:33 angelox_123: I can move it to images/ Jan 03 20:23:36 hold Jan 03 20:26:14 angelox_123: does not work Jan 03 20:26:47 can you give a pastebin from ossystems/recipes/images/osnetboot-video-geode.bb Jan 03 20:27:05 angelox_123: yes Jan 03 20:27:10 but i think you are having filenames conflicts Jan 03 20:27:34 angelox_123: http://pastebin.com/iQKwhDe5 Jan 03 20:28:03 angelox_123: the class is just a bunch of variables and an inherit image Jan 03 20:28:15 try this: Jan 03 20:28:27 edit line inherit ossystems-image-module to inherit image Jan 03 20:28:33 or give your ossystems-image-module Jan 03 20:28:56 http://pastebin.com/ahqRh8rS Jan 03 20:29:51 very strange Jan 03 20:30:05 try what i said... Jan 03 20:30:18 and if don't work,try renaming the image name Jan 03 20:31:35 but it makes no sense to it break it Jan 03 20:32:56 -inherit ossystems-image-module Jan 03 20:32:56 +inherit image Jan 03 20:33:01 keeps failing Jan 03 20:33:06 strange error Jan 03 20:33:10 yes Jan 03 20:33:23 it looks like it is passing the wrong URL for do_unpack Jan 03 20:33:43 yes..but i'm so confused :| Jan 03 20:34:19 http://pastebin.com/ciWNCEGJ Jan 03 20:34:26 the layer is at the paste Jan 03 20:34:54 i don't know very things about layers..sorry Jan 03 20:35:10 i don't understand what are the problem... :( Jan 03 20:36:32 kergoth: any idea about the issue above? Jan 03 20:36:41 kergoth: it looks like the do_unpack path is wrong Jan 03 20:37:59 * otavio needs to go away; will be back later. If someone knows about the issue please leave a message. Jan 03 21:01:28 hi, I found an issue in iw Jan 03 21:01:36 I bet I will have to write a mail Jan 03 21:01:46 basically with libnl1 biabtken: Jan 03 21:01:58 /home/gnutoo/embedded/oe/oetmps/htcdream/sysroots/x86_64-linux/usr/armv6-novfp/lib/gcc/arm-oe-linux-gnueabi/4.5.3/../../../../arm-oe-linux-gnueabi/bin/ld: warning: libnl.so.2, needed by /home/gnutoo/embedded/oe/oetmps/htcdream/sysroots/armv6-novfp-oe-linux-gnueabi/usr/lib/libnl-genl.so, may conflict with libnl.so.1 Jan 03 21:02:01 and at runtime: Jan 03 21:02:27 http://pastebin.com/EaDch4vw Jan 03 21:02:35 but if you bitbake -c clean libnl2 Jan 03 21:02:40 it builds fine and runs fine Jan 03 22:24:49 has bitbake ever been tested on python 2.7? Jan 03 22:25:17 I just tried using bitbake again after not trying it for a month and this is the traceback: http://pastebin.com/LtSmKRLr Jan 03 23:10:45 mwester: the web interface for nslu2 does not appear anymore :( do u know how to fix it Jan 03 23:11:06 Reflash the original Linksys firmware? Jan 03 23:11:09 mwester: its using the static IP by default Jan 03 23:11:19 mwester: hmmm Jan 03 23:11:23 its working well otherwise Jan 03 23:11:25 It uses whatever the network settings were for the originl firmware. Jan 03 23:11:39 There is no GUI with SlugOS. Jan 03 23:11:48 o i c Jan 03 23:12:05 I put a 4Gb USB stick Jan 03 23:12:07 http://www.nslu2-linux.org/wiki/SlugOS/InstallandTurnupABasicSlugOSSystem Jan 03 23:12:12 yep Jan 03 23:12:15 thats what I have used Jan 03 23:12:31 now its booting from the USB stick Jan 03 23:12:37 Ok, then that's pretty much what SlugOS gets you - a pretty basic Linux box to build from... Jan 03 23:12:47 ok Jan 03 23:12:58 now I wanted to upgrade it with latest stuff Jan 03 23:13:06 do u have feeds for new stuff ? Jan 03 23:13:28 I can screw the rfs on USB right ? Jan 03 23:13:47 and if I mess it up then I plug is out and it boots from flash ? Jan 03 23:13:55 and redo the stick again Jan 03 23:14:23 The feeds are defined, and may have packages in them, but I don't know how correct they are -- I need to clean them out and rebuild them from scratch. It really depends on how the autobuilder has been; i've not checked on that one for many months. Jan 03 23:14:45 mwester: ok. Jan 03 23:14:51 I did opkg update Jan 03 23:15:06 and it synced the feed data Jan 03 23:15:12 but no updates happened Jan 03 23:15:18 it seemed the system is uptodate Jan 03 23:15:30 but that means no updates were pushed Jan 03 23:15:41 may be I will set my own private feeds Jan 03 23:15:47 Heh - my mistake - I told you wrongly. Jan 03 23:15:48 and build it using gcc 4.5 Jan 03 23:16:28 The SlugOS 5.3 feeds are static and are not being updated -- the "unstable" feeds are being updated, but are not defined for the standard SlguOS 5.3 image. Jan 03 23:16:42 i c Jan 03 23:16:59 That won't change, because I don't think that opkg upgrade can do it, considering the change to libc Jan 03 23:17:16 hmmm Jan 03 23:17:33 So users will ahve to reflash with SlugOS 5.6 (or more probably 6.0 depending on what we call it), no upgrade available. Jan 03 23:17:39 libc change should be upgradable Jan 03 23:17:46 If we can switch to GCC 4.5 by then, that would be fine. Jan 03 23:18:22 we can but I need to kind of try the images geerated by gcc 4.5 Jan 03 23:18:34 I'd also like to upgrade the kernel, and that's not upgradable for SlugOS... yet. Jan 03 23:18:38 mwester, should I update my slugs? Jan 03 23:18:45 ka6sox, nope. Jan 03 23:18:56 mwester, kk Jan 03 23:18:58 very unstable right now. :) But better than they have been for a long time. Jan 03 23:19:45 mwester: can I just untar the file system on the USB stick and will it boot ? Jan 03 23:19:49 khem and eFfeM_work have been building SlugOS, so when I had some time around Christmas, I was able to work on the image itself rather than troubleshooting build problems. :) Nice. Jan 03 23:20:25 right. thats the purpose of testing branch Jan 03 23:20:35 khem, for testing images, I use nfs. But I think you can untar onto the USB stick as well. Jan 03 23:20:38 to keep distro/machines going which devs care Jan 03 23:20:53 mwester: ah nfs would be better Jan 03 23:21:04 mwester: how to set that up ? Jan 03 23:21:14 I remember seeing sme howto on that Jan 03 23:21:21 but cant remember where it was Jan 03 23:23:02 export a "slugos_rootfs" or suchlike directory on your NFS server -- then on the NSLU2, you do "turnup nfs -i :" -- and it should work, assuming that busybox mount works ok in the latest images. Jan 03 23:23:35 mwester: ok Jan 03 23:23:37 cool Jan 03 23:23:56 I have nfs server working for another boards Jan 03 23:23:57 That will copy the rootfs from flash to the nfs directory. From there you can edit, copy, at will. You can even untar the new rootfs overtop of the export on the host, which is how I do it. Jan 03 23:24:06 so it should be pretty simple to add nslu Jan 03 23:24:10 Yep. Jan 03 23:24:54 let me build an image with gcc 4.5 Jan 03 23:25:28 If you have any trouble, it is likely to involve rpc services which don't exist on the rootfs on the NSLU2 -- one has to use the -o nolock option to mount, IIRC. And I seem to recall that busybox had some issues with nfs mounts as well, but I don't recall the details and haven't tested that yet. Jan 03 23:26:42 well then I will shove it onto usb stick :) Jan 03 23:27:39 hopefully it is all fixed with busybox 1.18, and it will all just work. Jan 03 23:29:04 mwester: yes Jan 03 23:29:17 mwester: you should think of using sane-toolchain.inc Jan 03 23:29:28 to get toolchain versions Jan 03 23:31:01 but /me is insane! Jan 03 23:31:22 mwester: then it will make you insaner :) Jan 03 23:31:38 * khem hacks preferred-slugos-versions.inc Jan 03 23:32:37 mwester: MACHINE=nslu2be DISTRO=slugos bitbake native-sdk-image Jan 03 23:32:41 will that work > Jan 03 23:34:44 dunno. bitbake -k meta-toolchain-slugos used to work, but I had library conflicts that I was unable to solve, so I left it for another day. You can try it and see if it works better with gcc 4.5 Jan 03 23:35:18 * mwester wonders what native-sdk-image would do, so goes off to try to bitbake it... Jan 03 23:37:02 What version of python is most commonly used for bitbake users? Or what distro and version? Jan 03 23:37:29 if 2.6 is known to work, ill compile that Jan 03 23:38:15 2.6 AFAIK is the most common. Not sure if 2.4 is still supported. Jan 03 23:43:26 mwester: native-sdk-image will have all bits that you need to compile helloworld on the nslu2 itself Jan 03 23:45:08 Ah. Of course, with only 32MB of ram, helloworld is about all that one can actually compile on the nslu2! :D Just ask ka6sox -- IIRC he was involved in a project to compile Debian on an NSLU2... Jan 03 23:49:00 mwester, ow ow owwwwwwwww Jan 03 23:49:04 hurt me baby Jan 03 23:50:17 What's that old phrase? Something like "the wonder isn't in how _well_ the bear dances, rather that it dances at all!" :) Jan 03 23:52:46 32Mb might be ok for gcc Jan 03 23:53:03 I do native compiles on osk5912 which has 32M too Jan 03 23:53:48 but with gcc 4.5 is might be interesting Jan 03 23:53:50 khem, autotools sucks on a slug Jan 03 23:54:02 sorry but it just spends forever configuring Jan 03 23:54:07 dont they suck everywhere Jan 03 23:54:18 true, but worse on a slug Jan 03 23:54:22 like HOURS worse. Jan 03 23:54:32 I dont compile real software locally Jan 03 23:54:39 but run gcc testsuites Jan 03 23:54:52 or build bench marks Jan 03 23:54:54 kk Jan 04 01:01:52 hmm python 2.6 wont work either Jan 04 01:03:35 Using Python 2.7 from Arch Linux here Jan 04 01:03:51 i tried that and i thought that 2.7 was the problem Jan 04 01:03:58 so i tried 2.6 Jan 04 01:04:19 what was the problem with 2.7? Jan 04 01:05:00 2.7 - http://pastebin.com/LtSmKRLr 2.6 - http://pastebin.com/h1MmxhJt Jan 04 01:06:03 im running out of tricks to pull. Jan 04 01:08:01 can't say I've every had that problem, sorry :P Jan 04 01:08:48 maybe something more to do with building semaphore support in pythong or something? Jan 04 01:08:52 vadmium1, looks like a system problem to me. Jan 04 01:09:09 im in a chroot environment Jan 04 01:09:33 You aren't using one of those "build it from source code yourself" distros, are you? if so, you need to check your "use flags" or whatever, to make sure you have support for semaphores and other locks... Jan 04 01:09:57 no im on fedora with arch mounted inside it Jan 04 01:10:06 this also doesnt work with arch inside rch Jan 04 01:10:07 *arch Jan 04 01:10:25 once you chroot, you're in a differnt environment... only the kernel remains. Jan 04 01:10:47 Why not just use fedora? (I do) Jan 04 01:11:49 because im trying to get fancy, allowing the user to build OE inside a different environment Jan 04 01:12:13 to avoid problems to the actual running system Jan 04 01:12:26 But if you want to use arch, then you'll have to find a binary that supports whatevers missing in yours, or if you built it yourself, then you'll have to rebuild it with the right support (again, I don't know precisely what's missing - just guessing that it's in libc or other low-level app support) Jan 04 01:13:11 * mwester uses nothing but Fedora hosts, and has never had troubles with OE causing problems for hosts -- just make sure you never run bitbake as root. Jan 04 01:14:17 mwester, due you bbuild binary locales? Jan 04 01:14:24 or you turn of selinux/ Jan 04 01:14:54 i just turned it off Jan 04 01:15:10 i kept getting these AVC denials just now Jan 04 01:15:31 brb rebooting Jan 04 01:15:46 I used to build binary locales, but disable selinux on all my systems. Habit, mostly -- I'm sure that the early bugs which rendered hosts almost useless for dev work with selinux enabled have been long ago solved. Jan 04 01:17:15 yeah Jan 04 01:17:24 I just disable binary locales Jan 04 01:17:44 They take waaay too long to build anyway. :) Jan 04 01:22:20 hmm turns out that bitbake or python or whatever needs /dev to be fully mounted Jan 04 01:22:30 works fine now Jan 04 01:24:13 yay my chroot project is done **** ENDING LOGGING AT Tue Jan 04 02:59:58 2011