**** BEGIN LOGGING AT Sat Jan 10 02:59:57 2009 Jan 10 03:41:30 Can anyone recommend to me a reference for the best way to customize a RFS Jan 10 03:44:57 archae0pteryx: rfs == root file system? Jan 10 03:46:01 yes Jan 10 03:46:22 archae0pteryx: pick the one that the closest to you, check packages/images/* for rootfs's Jan 10 03:46:35 In addition to adding packages. I am fairly certain I can customize a package by editing the recipe but if I want to stuff on tiop of that Jan 10 03:46:44 like replacing default config files Jan 10 03:47:11 archae0pteryx: it really depends, rfs == your own image file Jan 10 03:47:24 yes Jan 10 03:47:29 archae0pteryx: package defaults, it depends again, sometimes the defaults make sense for everyone then just patch it Jan 10 03:47:52 archae0pteryx: sometimes only for your device/distro (e.g. /etc/network/interfaces) then use something like overrides... Jan 10 03:49:25 I have a recipe to make my own image simply to customize what packages i have installed. But I'll eventually want to have different config files based on image being built I believe Jan 10 03:51:57 in most cases the config is not depending on the image, e.g. a package might be installed later.. Jan 10 03:52:54 That's true. But maybe if I've added a package to an image I want to override it's default config files, not necessarily add or remove files Jan 10 03:53:17 I guess I could include a different recipe Jan 10 03:53:17 archae0pteryx: then put the config into separate packages? Jan 10 03:53:25 and on rootfs pick the packages you want? Jan 10 03:53:34 that's what I just came up with ;) thanks Jan 10 03:53:59 great minds think alike and all that Jan 10 03:54:39 * archae0pteryx is doing too much manually right now Jan 10 03:55:07 hehe Jan 10 03:58:11 I have to do it anyway I just wasn't planning on doing it soon Jan 10 04:11:40 03Tick Chen  07org.openembedded.dev * r35b33d640a 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Jan 10 04:11:40 opkg: Bump up the version to R196 Jan 10 04:11:40 fixing issues http://code.google.com/p/opkg/issues/detail?id=2 Jan 10 07:44:10 03Khem Raj  07org.openembedded.dev * r7cc6eda906 10openembedded.git/packages/uclibc/uclibc-svn/ (arm/uClibc.config uClibc.config uClibc.distro): Jan 10 07:44:10 uClibc.config: New Jan 10 07:44:10 uClibc.distro: New Jan 10 07:44:10 arm/uClibc.config: Set UCLIBC_SUSV4_LEGACY Jan 10 09:06:31 03Koen Kooi  07org.openembedded.dev * r06af67f2e6 10openembedded.git/packages/e17/e-wm_svn.bb: e-wm: ship unpackaged utils in ${PN}-utils Jan 10 09:06:40 03Koen Kooi  07org.openembedded.dev * r06af67f2e6 10openembedded.git/packages/e17/e-wm_svn.bb: e-wm: ship unpackaged utils in ${PN}-utils Jan 10 09:45:09 03Koen Kooi  07org.openembedded.dev * rd076ad3dbe 10openembedded.git/packages/ (45 files in 15 dirs): gumstix utils: add char-driver, gpio-event, i2c and microwindows Jan 10 09:49:30 !oelogs Jan 10 09:49:35 !logs Jan 10 09:49:35 Channel logs for #oe are archived at: Jan 10 09:49:36 http://hentges.net/tmp/logs/irc/%23oe Jan 10 09:49:37 Live-logs are available at Jan 10 09:49:39 http://hentges.net/tmp/logs/irc/livelogs/%23oe.livelog Jan 10 09:49:41 See ?? help-logs for usage instructions Jan 10 10:07:39 hello, i' Jan 10 10:07:40 oops Jan 10 10:08:09 hello, i've a problem with boost on angstrom...i've thumb enabled and it does that: /tmp/ccMefFub.s:639: Error: selected processor does not support `swp r2,r3,[r6]' Jan 10 10:08:25 how do I solve the problem without disabling all libs? Jan 10 10:09:37 ant_: hey :) Jan 10 10:10:48 i'll grep the source Jan 10 10:11:02 morning all Jan 10 10:16:10 I think I have to remplace swp %0, %1, [%2] Jan 10 10:16:14 morning Jan 10 10:28:22 mabe i've found the right post Jan 10 10:29:47 the strange thing is that it works with the opnemoko toolchain Jan 10 10:29:55 s/toolchain/distro as config file/ Jan 10 10:32:04 seems that openmoko doesn't use thumb Jan 10 10:32:38 that's why it worked...but it's strange that it worked on woglinde's machine(he use angstrom) Jan 10 10:46:28 mickey|zzZZzz, you awake man? Jan 10 11:09:39 hi, what would be the best way to set up an environment for two different distros (openwrt and angstrom) for two different boards (nslu2 and beagle). the boards are both arm based. I'd rather not have 4 copies of all tools, recipes etc Jan 10 11:24:15 03Jan Lübbe  07org.openembedded.dev * r1b6e9b10a1 10openembedded.git/packages/linux/ (4 files in 3 dirs): linux-openmoko-2.6.28/devel: reduce kernel size by configuring some stuff as modules Jan 10 11:39:44 I've a very problem...some python part of a recipe that work on woglinde are never executed: Jan 10 11:39:49 I tried: Jan 10 11:40:04 *re-cloning the org.openembedded.dev git Jan 10 11:40:16 *removing part of my config(inherit ipk) Jan 10 11:40:35 *changing distro from openmoko to angstrom... Jan 10 11:40:45 that required to recompile all from scratch Jan 10 11:41:46 I realy need help Jan 10 11:42:00 ah yes...and I also tried an ubuntu livecd Jan 10 11:42:06 my distro is gentoo Jan 10 11:42:25 and I also migrated to python 2.5 and removed all old python Jan 10 11:45:18 eFfeM, i'm not an expert but mabe just switch distro,oetmp,and machine when you compile...keep the sources and the config/package dir common Jan 10 11:46:43 eFfeM, i bet you can't use the same tmpdir for openwrt and angstrom because i bet that you don't have thumb on openwrt ...and angstrom uses thumb Jan 10 11:53:09 hello, is there someone there? Jan 10 12:05:19 now I try to remove bitbake and re-do a svn checkout Jan 10 12:05:30 s/I try/I'm trying Jan 10 12:07:58 GNUtoo, thanks for your info (was afk 4 lunch, I don't mind using multiple tmp dirs, but I would like to use a single git image Jan 10 12:08:51 although i had the impression that the thumb compiler would be named differently or be controlled by different cflags Jan 10 12:09:27 I don't know if the thumb compiler is named differently... Jan 10 12:09:45 eFfeM, by the way what are you compiling? Jan 10 12:11:15 GNUtoo i want to be able to create an image for my NSLU2 of both angstrom and openwrt or openslug, but I am in the process of moving to a beagle board and would like to be able to build beagle images in the same tree Jan 10 12:11:34 ok Jan 10 12:11:44 so you do not have evrything setup? Jan 10 12:11:53 because I'd like someone to test a recipe Jan 10 12:12:10 because woglinde is not there at the moment Jan 10 12:12:26 (there is a recipe that works on his machine but not on mine) Jan 10 12:12:29 i have a slugos set up, but i'm at the moment moving my oe tree (reinstalled linux on a new hard disk) Jan 10 12:12:41 which recipe would you want to have tested Jan 10 12:13:04 a modified boost_1.36 Jan 10 12:13:26 I want to know if it split the packages correctly Jan 10 12:13:33 give me a few minutes then I can tell whether I can test it Jan 10 12:14:21 ok thanks a lot Jan 10 12:14:27 i'll put it on my server Jan 10 12:16:20 I've a last thing to test before making you testing it Jan 10 12:16:50 ok, np, still moving the dir and need to see if it works either Jan 10 12:22:59 anyway the bb files and patches are here: http://gnutoo.homelinux.org/open_embedded/boost.tar.bz2 Jan 10 12:23:47 will give a try later, still moving around data ... Jan 10 12:23:55 eFfeM, ah I have forgetten...be carefull when moving the tmpdir...it has to be at the same location Jan 10 12:24:26 otherwise you will have an error and won't be able to continue compiling Jan 10 12:26:09 GNUtoo, the user name and path remains the same, it is only the partition data that is moved from one partition to another Jan 10 12:26:22 so that's ok Jan 10 12:26:31 but as you will have 2 tmpdir... Jan 10 12:26:34 be carefull Jan 10 12:27:31 i've noticed once that there is an option to have somethign like tmpdir_$distro or so; that was also used for autobuild.,need to find that msg again, but thought someone here could tell me right away Jan 10 12:28:42 kristoffer: good morning Jan 10 12:28:45 ok Jan 10 12:28:57 that's interresting as I build openmoko and angstrom Jan 10 12:29:15 otherwise i could use the python base conditional Jan 10 12:29:49 but I don't know if that works in config files... Jan 10 12:30:28 issue is also that you need to override local.conf Jan 10 12:31:21 one simple solution i've been thinking of is to have several local.conf variants and issue a command upfront which will creat a symlink from tmpdir to the tmpdir variant and a link of local.conf Jan 10 12:32:13 but thought that perhaps there is a better way Jan 10 12:41:09 mickeyl, I'm sorry to bother you but I've a problem that me and woglinde couldn't resolve: I've a recipe at: http://gnutoo.homelinux.org/open_embedded/boost.tar.bz2 (recipies and patches) that works on woglinde's machine but not mine...the problem is that it doesn't split boost in multiples packages and packages evrything in ${PN}-dev...mabe because of the defaults inside bitbake.conf...here's what I've tried: http://rafb.n Jan 10 12:41:09 et/p/i4Anfv43.html what should I do now? Jan 10 12:42:15 attach it your tarball to a bug, describing the problem and I'll take a look Jan 10 12:42:20 s/it// Jan 10 12:42:37 ok thanks a lot!!! Jan 10 12:42:43 np Jan 10 12:44:46 shoudl I put the bug in bitbake or openembedded? Jan 10 12:47:15 openembedded first Jan 10 12:47:36 since it's more likely to be a problem in oe Jan 10 12:47:45 but we'll see Jan 10 12:47:56 ok thanks a lot Jan 10 12:53:50 mickeyl, had some issues with ssh but think I have it solved, going to try shortly Jan 10 12:56:52 oki Jan 10 12:57:06 * * OE Bug 4975 has been created by GNUtoo(AT)no-log.org Jan 10 12:57:08 * * I've a recipe that works on woglinde's machine but not mine Jan 10 12:57:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4975 Jan 10 13:06:56 bother source.mvista.com is down I think Jan 10 13:07:08 GNUtoo: try using python populate_packages_prepend() Jan 10 13:07:12 not python split_pkg() Jan 10 13:07:25 I also tried it but i'll retry Jan 10 13:14:59 I saw qwerty!!!!!!!!!!!!!!!!!!!!!!!!!! Jan 10 13:15:35 but still not splited Jan 10 13:15:56 LEAD_SONAME not defined... Jan 10 13:16:00 i'll check that out Jan 10 13:18:48 good Jan 10 13:18:52 progress :) Jan 10 13:19:31 03Philip Balister  07org.openembedded.dev * r6291d154ec 10openembedded.git/packages/linux/linux-davinci_2.6.x+git-davinci.bb: linux-davinci : Change SRC_URI from source.mvista.com to git.kernel.org. Jan 10 13:19:40 * Crofton|work crosses his fingers that his email is still set in git :) Jan 10 13:19:44 yay Jan 10 13:19:54 :) Jan 10 13:20:06 Crofton|work: will you send an ADMIN email wrt. the domains? Jan 10 13:20:13 yes Jan 10 13:20:17 RP: are we set on friday before FOSDEM? Jan 10 13:20:18 I'll get that done today Jan 10 13:20:22 cool, thanks Jan 10 13:20:38 I need to switch dns from eay dns also Jan 10 13:44:21 Crofton|work, you seems to know well boost...what SONAME should I chose? Jan 10 13:44:38 because it appears that there is no main library Jan 10 13:46:08 s/SONAME/LEAD_SONAME Jan 10 13:47:54 heh Jan 10 13:48:01 I hate boost, but have to use it Jan 10 13:48:12 GNUtoo, I don't know Jan 10 13:48:22 mainly I got the bits I needed to work Jan 10 13:48:27 ok i'll pick one at random...thanks anyway Jan 10 13:48:32 ok Jan 10 13:48:43 you are working toward being able to use multiple version, right? Jan 10 13:49:09 Crofton|work, no i'll do that later...I first need to make it package fine Jan 10 13:49:15 ok Jan 10 13:49:48 hopefully I have time to work on stuff that needs boost in a few weeks Jan 10 14:12:36 what is the current binutils version built by OE for arm? Jan 10 14:15:29 http://pastebin.com/m7890916f Jan 10 14:15:36 anyone know a solution for this error when running bitbake? Jan 10 14:17:34 GNUtoo, i've build your boost, what did you want to know ? Jan 10 14:18:03 btw also found out about the tmp dir, trick is to have one src tree and different tmpdirs e.g. by using an env var Jan 10 14:19:51 hmm, not getting past fakeroot at all. seems like all versions are dead in the water (non-existant). Jan 10 14:22:07 GNUtoo, afk for an hour or so, but feel free to leave a msg Jan 10 14:32:44 (codesourcery objdump works) Jan 10 14:44:10 eFfeM, ok I wanted to know if it was splited in multiple packages Jan 10 14:54:30 I understand a part of the problem...here's the package order: boost-dbg boost boost-doc boost-dev boost-locale boost-dev boost Jan 10 14:54:43 so boot-dbg is the main package according to: Jan 10 14:55:13 mainpkg = packages[0] Jan 10 14:55:17 the comment is: Jan 10 14:55:24 # This is *really* broken Jan 10 14:55:25 lol Jan 10 15:00:02 ah no after it .remplace 'dbg' and 'dev' by nothing Jan 10 15:00:10 so it should find the package Jan 10 15:10:06 * * OE Bug 4976 has been created by sander(AT)3v8.net Jan 10 15:10:08 * * o add support for command line url parameter Jan 10 15:10:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4976 Jan 10 15:19:20 eFfeM, the problem is that on my system it packages evrything in -dev Jan 10 15:28:47 eFfeM, and I want it splited like libboost-filesystem Jan 10 15:29:29 i'll check (just back) Jan 10 15:30:50 eFfeM, ok thanks a lot Jan 10 15:32:48 GNUtoo where exactly do you want me to look Jan 10 15:33:06 i did a bitbake -b so it did not end up on deploy Jan 10 15:33:41 guess you are boost-1.36.0-r5 Jan 10 15:33:45 yes Jan 10 15:33:48 has there been a change in how AUTOREV is handled? Jan 10 15:33:59 I have some recipes in a collection that are always rebuilding now Jan 10 15:34:19 eFfeM, what gives an ls inside the dir that corresponds to ~/oetmp_angstrom/work/armv4t-angstrom-linux-gnueabi/boost-1.36.0-r5/install on your box? Jan 10 15:35:05 hm there is no install Jan 10 15:35:54 sry did not check the log properly compile failed Jan 10 15:36:07 ah ok Jan 10 15:36:19 bjam command not found Jan 10 15:36:40 wrong bb file Jan 10 15:36:44 no Jan 10 15:36:57 yes, my bad Jan 10 15:37:07 i build 1.33.1 Jan 10 15:37:11 it's only that you didn't bitbakel the boost-jam-native Jan 10 15:38:02 ah that's because you don't have: PREFERRED_VERSION_boost = "1.36.0" and PREFERRED_VERSION_boost-jam-native = "3.1.16" Jan 10 15:38:12 ah, ok building that one now, i also did build .33.1 not .36.0 i was too much in a rush to start things Jan 10 15:38:26 ok np Jan 10 15:38:50 GNUtoo, even worse, I was lazy, copied boost then applied your changes then did a bitbake -b packages/boost-/*.bb Jan 10 15:38:57 so that did build the wrong one Jan 10 15:39:06 lol Jan 10 15:39:41 so mabe we'll see that later as i've made some change in the recipe Jan 10 15:39:42 thought there was only one, didn't know the exact name, so thought *.bb would help me to avoid having to look it up Jan 10 15:40:29 btw anyone know why I get: Jan 10 15:40:35 /home/frans/oe/bitbake/lib/bb/COW.py:29: DeprecationWarning: the sets module isdeprecated Jan 10 15:40:35 import types, sets Jan 10 15:41:04 when I call bitbake? this is on opensuse, when I build under ubuntu I do not get that warning Jan 10 15:41:04 mmm...don't know and I must go Jan 10 15:41:19 ok will you stay in the channel Jan 10 15:41:30 if so I'll msg you the results Jan 10 15:41:38 yes of course...the computer will stay not me Jan 10 15:41:50 ok thanks a lot Jan 10 15:42:45 eFfeM, can I pastebin a new version of boost-36.inc? Jan 10 15:42:53 sure Jan 10 15:43:42 http://rafb.net/p/GdYhpv74.html Jan 10 15:43:44 thanks a lot Jan 10 15:46:34 GNUtoo my pidigin became nonresponsive, i missed the url if you pasted it Jan 10 15:46:55 eFfeM, http://rafb.net/p/GdYhpv74.html Jan 10 15:47:03 I must go now Jan 10 15:47:17 ok cya Jan 10 15:53:53 Gnutoo, it does not patch: Jan 10 15:53:55 NOTE: Applying patch 'no-long-double-math.patch' (/home/frans/oe/openembedded/packages/boost-/files/./no-long-double-math.patch) Jan 10 15:53:55 NOTE: Task failed: Error: /home/frans/oe/openembedded/packages/boost-/files/./no-long-double-math.patch not found. Jan 10 15:54:16 frans@linux-cdv2:~/oe/openembedded/packages> find boost | grep math Jan 10 15:54:16 frans@linux-cdv2:~/oe/openembedded/packages> find boost- | grep math Jan 10 15:54:24 eFfeM1, strange but I must go Jan 10 15:54:25 so no patch Jan 10 15:54:33 sorry i realy must go Jan 10 15:54:33 okie Jan 10 15:55:53 mabe it's the package name Jan 10 15:56:27 the bb's name is boost_1.36.0.bb Jan 10 15:56:36 otherwise it means that you do not have the patch Jan 10 15:56:44 eFfeM, ^^^ Jan 10 15:56:46 i go Jan 10 15:56:53 hi Jan 10 15:57:05 GNUtoo i do not have the patch file Jan 10 15:57:09 cya Jan 10 15:57:24 eFfeM1, did I gave you the tarball? Jan 10 15:57:37 i got an reeror message with org.opendreambox branch Jan 10 15:58:00 http://gnutoo.homelinux.org/open_embedded/boost.tar.bz2 Jan 10 15:58:09 Exception: Message:local variable 'bb' referenced before assignment Jan 10 15:58:14 eFfeM ^^^ Jan 10 15:58:32 seem to be a known problem but I can't fix it Jan 10 15:59:13 http://bugs.openembedded.net/show_bug.cgi?id=2427 <- related to my problem Jan 10 16:00:40 got it Jan 10 16:02:29 GNUtoo if you are still there, is it ok that there are no files in the files/ dir in your tar ? Jan 10 16:03:11 GNUtoo and if you read this later, try to catch me tomorrow if you still need someone to test Jan 10 16:27:26 03Philip Balister  07org.openembedded.dev * re17f080c93 10openembedded.git/packages/powertop/powertop_1.11.bb: powertop : Add bb file for version 1.11. Jan 10 16:50:15 mickeyl: Do we have a list of people who can attend? Jan 10 16:50:35 mickeyl: I'm starting to wonder if an online meeting might have more of the key people we need : Jan 10 16:50:36 :/ Jan 10 16:51:28 right now it looks like crofton, pb_, and me Jan 10 16:51:33 (and you) Jan 10 16:51:42 no response from zecke yet Jan 10 16:51:58 i don't know, it looks i will probably skip it if we don't meet as well Jan 10 16:59:36 03Michael 'Mickey' Lauer  07org.openembedded.dev * rf4daa9b92f 10openembedded.git/ (21 files in 5 dirs): Jan 10 16:59:36 linux-openmoko: split into linux-openmoko-2.6.24 (old stable), linux-openmoko-2.6.28 (new stable), land Jan 10 16:59:36 linux-openmoko-devel (new testing) Jan 10 16:59:37 03Michael 'Mickey' Lauer  07org.openembedded.dev * r96bed68c2d 10openembedded.git/conf/machine/ (om-gta01.conf om-gta02.conf): Jan 10 16:59:39 om-gta01.conf: prefer kernel 2.6.24 until 2.6.28 has been tested Jan 10 16:59:41 om-gta02.conf: prefer kernel 2.6.28 Jan 10 17:00:36 mickeyl: I'm afraid we don't look to have critical mass :/ Jan 10 17:00:50 ok Jan 10 17:10:57 hi, I'm trying to build "helloworld-image" for a PXA270 target, but get stuck on fakeroot-native_1.12.bb becuase of a patch failing to apply - any suggestions? (Am I the only one experiencing this error?) Jan 10 17:18:52 RP, mickeyl we should at least try to organize a dinner or something for OE affiliated attendees Jan 10 17:23:14 03Michael 'Mickey' Lauer  07org.openembedded.dev * rd5a683e1fc 10openembedded.git/ (3 files in 2 dirs): vala[-native]: update to 0.5.5 Jan 10 17:24:26 Crofton|work: That sounds like a good idea Jan 10 17:26:53 * eFfeM wonders if he is affiliated (haven't decided yet if I go either) Jan 10 17:29:09 eFfeM, you are here and doing OE work ... Jan 10 17:29:14 what package do I need to be able to compile modules on a target ? I'm missing lib/modules/2.6.26-1-ixp4xx/build, which pkg creates that Jan 10 17:30:07 eFfeM, ok i'm back Jan 10 17:30:10 Crofton|work: I even still have an OE key although it has been a while since I committed anything (got sidetracked into something else, but trying to pick things up with beagle) Jan 10 17:30:17 wb GNUtoo Jan 10 17:30:36 GNUtoo your tarball had an empty files subdir guess that is where the patch had to be Jan 10 17:31:06 eFfeM, strange on my server it hasn't Jan 10 17:31:22 eFfeM, for instace: -rw-r--r-- gnutoo/gnutoo 1430 2009-01-04 01:04 boost/files/unit_test_log10f.patch Jan 10 17:32:48 Crofton|work: i think it would be fun to meet some oe people in real life Jan 10 17:33:14 gnutoo not in the tar you sent me afaik can see, but will check again Jan 10 17:33:29 eFfeM, i doubble checked...the tar is ok Jan 10 17:33:41 eFfeM, do you go to osdem? Jan 10 17:33:51 s/osdem/Fosdem/ Jan 10 17:34:01 GNUtoo haven't decided yet Jan 10 17:34:18 but the Mrs liked the idea to spend a weekend in brussels so maybe yes Jan 10 17:34:27 eFfeM,I went to it twice Jan 10 17:34:40 I live about 1.5-2hrs driving from brussels Jan 10 17:34:43 but now i'm in italy Jan 10 17:35:08 eFfeM, so why not going there...it's not too far... Jan 10 17:35:38 got the patches but ark has beaten me, I moved from ubuntu to suse today and thought that if I clicked on files folder then extract, i would get the files extracted, apparently not Jan 10 17:35:53 eFfeM, once I even assisted to an oe conference about bitbake(I had only a pma430 at the time) Jan 10 17:35:57 GNUtoo it is indeed not far and the weekend is not booked yet Jan 10 17:36:22 actually if the mrs decides not to join me it is possibel that i only get there for a day Jan 10 17:36:26 * eFfeM only has an NSLU2 Jan 10 17:36:38 but I'll get a beagle as soon as C1 is out Jan 10 17:36:49 eFfeM, ok Jan 10 17:39:35 it is compiling Jan 10 17:41:19 eFfeM, i have: pma430->nightmare(not in oe,gcc-2.95 can't change it,kernel with proprietary modules etc...),openmoko(great),openwrt(not running oe yet...running openwrt),and mabe i'll buy a pandorabox(i won't use the 3d) (i've also a ti89 not yet swiched to pedrom and a chdk camera)...i'm a little bit like the dog here: http://www.freesoftwaremagazine.com/columns/bizarre_cathedral_31 Jan 10 17:42:29 LOL Jan 10 17:42:49 (btw I still have a ti57 but the battery is broken) Jan 10 17:44:42 it is still compiling, afk for a few minutes Jan 10 17:45:01 ok Jan 10 17:51:37 Is everyone using that standard source package places? fakeroot doesnt even exist in the "old" location anymore Jan 10 17:57:41 I wouldn't know, I'm kinda new to using OE. What should I do to get past that fakeroot build problem? Jan 10 17:59:05 well, getting the source package is one way (1.9.6) or adding the new package source (1.11.4) Jan 10 17:59:17 Ive been banging my head against that a couple of hours now Jan 10 17:59:31 gr8 :-) Jan 10 18:14:36 kristoffer: is there a convenient way of getting the MD5 and SHA256 sums of a new package into the checksums.ini file? (instead of doing it by hand) Jan 10 18:14:50 cat tmp/checksum.ini Jan 10 18:15:01 then copy'n'paste Jan 10 18:15:41 thx! Jan 10 18:20:08 03Kristoffer Ericson  07org.openembedded.dev * r75e88aff01 10openembedded.git/conf/distro/jlime-mongo.conf: conf/distro/jlime-mongo.conf : add tar.gz package output (aswell as ipk) Jan 10 18:20:19 03Kristoffer Ericson  07org.openembedded.dev * rd75f60d524 10openembedded.git/conf/checksums.ini: conf/checksums.ini : Add checksums for fakeroot 1.11.4 Jan 10 18:21:24 mickeyl, looks like it works. Jan 10 18:21:32 good Jan 10 18:34:00 xorcist68: There have been many host-distro-sensitive troubles with fakeroot, which is why once a version that works is found, OE tends to stick with it... Jan 10 18:51:04 ouch...mabe it will work now...if it works it was because of a typo: libbboost instead of libboost Jan 10 18:51:35 hi gnutoo Jan 10 18:52:13 woglinde, hi i've made a typo in boost...i don't know why it worked on your machine...i even made it try by eFfeM and he has the same results than me Jan 10 18:52:42 hm Jan 10 18:53:00 boost 1.36.0 right? Jan 10 18:53:10 woglinde, but i'm compiling it right now didn't look if it works yet Jan 10 18:53:15 woglinde, yes Jan 10 18:53:36 in dev I have only the headers and the .a libs Jan 10 18:54:02 I am compiling for armv machine and uclibc Jan 10 18:54:07 but this shouldnt matter Jan 10 18:54:13 whats your bitbake version? Jan 10 18:54:19 I have latest svn Jan 10 18:54:24 maybee thats the problem Jan 10 18:55:02 me too Jan 10 18:55:08 latest svn Jan 10 18:55:11 1.8 branch Jan 10 18:56:22 NOTE: boost-dbg boost boost-doc boost-dev boost-locale boost-filesystem boost-program-options boost-iostreams boost-math-tr1f boost-date-time boost-system boost-math-c99 boost-regex boost-unit-test-framework boost-thread boost-math-c99f boost-prg-exec-monitor boost-signals boost-math-tr1 Jan 10 18:56:30 it works!!!! Jan 10 18:56:36 thanks a lot for all your help Jan 10 18:56:59 *g* Jan 10 18:57:03 ah no Jan 10 18:57:19 seems that i have them but it still package evrything in dev but i'll fix that soon Jan 10 18:59:56 by the way match_path seems used only in the regex matching...better leaving it at true Jan 10 19:05:58 i'll remove it at the match is used later Jan 10 19:06:39 kristoffer: did u get past that fakeroot build problem yet? There's a patch that doesn't apply cleanly to 1.11.4... Jan 10 19:07:17 woglinde, by the way is there something wrong with frozen-bubble beside the PR(mabe i haven't changed it)? or mabe you didn't have time to commit it Jan 10 19:08:44 03Tom Rini  07rpurdie/canadian-sofar * r00f5fabc4b 10openembedded.git/classes/insane.bbclass: insane.bbclass: Fix thinko in package_qa_check_staged logic for iscrossnative Jan 10 19:09:01 gnutoo I am at testing it Jan 10 19:09:10 woglinde, ah ok thanks a lot Jan 10 19:09:12 but with uclibc and linuxthread perl isnt compiling Jan 10 19:09:18 woglinde, ah ok Jan 10 19:09:20 so I could test if it builds Jan 10 19:09:26 couldnt test Jan 10 19:09:39 and I got stuned by the mesa tls stuff Jan 10 19:09:46 which I fixed yesterday Jan 10 19:09:47 hi ant Jan 10 19:10:25 hi woglinde Jan 10 19:13:29 ant_: evening :) Jan 10 19:14:01 hey Jan 10 19:14:12 hi jay7 Jan 10 19:14:21 woglinde: hey :) Jan 10 19:14:41 ant_: did you checked kexecboot in init-mode? :) Jan 10 19:14:47 Jay7: without other hacks I'm still using init I suppose... Jan 10 19:15:21 I've almost done reboot and rescan options Jan 10 19:15:29 hm.. Jan 10 19:15:35 ant_: what's wrong? Jan 10 19:16:00 nothing :-) Jan 10 19:16:06 ah Jan 10 19:16:18 but I think I'm still using the init provided by initramfs-kexecboot Jan 10 19:16:19 yes, initramfs should be fixed.. Jan 10 19:16:25 exactly Jan 10 19:17:05 look into initramfs-kexecboot_1.0.bb Jan 10 19:17:07 have you sen the log btw? Jan 10 19:17:13 do_compile and do_install Jan 10 19:18:03 03Michael 'Mickey' Lauer  07org.openembedded.dev * rb3f8a99fbe 10openembedded.git/ (2 files in 2 dirs): libgdbus: bump to 0.2+git Jan 10 19:18:03 do_compile is not needed, in do_install kexecboot should be linked as init Jan 10 19:18:07 and about klibc-utils-static-sh ? Jan 10 19:18:09 03Michael 'Mickey' Lauer  07org.openembedded.dev * r9492ade502 10openembedded.git/packages/connman/ (5 files): connman: remove 0.1 and 0.2; fix metadata Jan 10 19:18:36 but I don't know where is kexecboot placed in initramfs (/bin, /sbin, etc.) Jan 10 19:18:49 looking Jan 10 19:19:00 ah Jan 10 19:19:05 /usr/bin/kexecboot Jan 10 19:19:09 from init.sh Jan 10 19:19:37 confirmed (inspected the cpio.gz) Jan 10 19:19:58 Jay7: I van hack mt cpio for the moment... Jan 10 19:20:10 this is ok for now Jan 10 19:20:50 my point is getting rid of the shell..possible? Jan 10 19:21:27 try to remove it Jan 10 19:21:40 it should not be called from anywhere now Jan 10 19:23:23 hmm..return -1 Jan 10 19:23:49 from where? :) Jan 10 19:37:56 03Michael 'Mickey' Lauer  07org.openembedded.dev * r81b581a29c 10openembedded.git/packages/linux/ (12 files in 2 dirs): linux-openmoko: remove obsolete files Jan 10 19:43:28 03Michael 'Mickey' Lauer  07mickey/python26 * r7fb704a0d1 10openembedded.git/packages/python/ (33 files in 6 dirs): python: remove 2.4 and 2.5 Jan 10 20:13:55 bother, I forgot to add the checksumds for powertop-1.11 Jan 10 20:14:47 crofton hehe Jan 10 20:26:42 hi florian Jan 10 20:27:41 hi all Jan 10 20:36:19 03Philip Balister  07org.openembedded.dev * r9917f3a840 10openembedded.git/conf/checksums.ini: checksums.ini : Add checksums for powertop-1.11 and sort. Jan 10 20:45:55 there, I have corrected that error Jan 10 21:06:54 woglinde, it continues to ship everything in ${PN}-dev even with the python function from the old recipe...but at least it creates the directories Jan 10 21:07:15 woglinde, what's your setup? why it worked on your computer? Jan 10 21:09:10 mabe it's that: ${libdir}/libboost_*.a" \ Jan 10 21:12:30 anyone familiar with what is on oe/oz domains dns? Jan 10 21:16:30 gnutoo hm Jan 10 21:17:34 removing ${libdir}/libboost_*.a" \ didn't change anything...i'll pastebin the recipe Jan 10 21:18:27 http://rafb.net/p/d3QlHS46.html Jan 10 21:22:04 03Koen Kooi  07org.openembedded.dev * r3a6e211cc0 10openembedded.git/packages/i2c-tools/ (17 files in 2 dirs): Jan 10 21:22:04 pico-dlpcontrol: use a modified version of the gumstix 'i2c' program and wrap it with a script Jan 10 21:22:04 * 'picodlp-control hflip 1' will turn on horizontil flipping Jan 10 21:39:28 FILE_{PN}-something is defined before the PACKAGES+={PN}-something mabe it's that? Jan 10 21:40:12 mmm Jan 10 21:40:13 NOTE: FILES_boost-date-time is /usr/lib/libboost_date_time*.so* Jan 10 21:40:48 the later is bb.data.getVar("FILES_%s" % pkg, d, 1) Jan 10 21:41:13 gnutoo you could run bitbake -i Jan 10 21:41:20 i'll re-read the part on including files Jan 10 21:41:22 ok i'll try Jan 10 21:41:35 and then peek boost-1.36.0 FILE_{PN}-something Jan 10 21:41:44 or every other var Jan 10 21:43:25 tanks a lot Jan 10 21:45:25 woglinde, i've a question on this shell...how do I clean the result of packaging...because clean clean all...i want to keep the result of the compilation and the installation Jan 10 21:45:38 build Jan 10 21:45:42 sorry Jan 10 21:45:52 didnt know you didnt know Jan 10 21:45:54 *g* Jan 10 21:45:59 bitbake -c build moo Jan 10 21:46:11 ok thanks a lot Jan 10 21:47:05 I didn't know because i didn't have the time to look at it Jan 10 21:50:56 doesn't work(the build thing) Jan 10 21:51:05 ???? Jan 10 21:51:15 mabe i wasn't very clear Jan 10 21:51:16 bitbake -c build boost Jan 10 21:51:25 yes that's what I did Jan 10 21:51:31 but let me explain better Jan 10 21:51:31 then you habve image install pkg-staging Jan 10 21:51:33 all in work Jan 10 21:51:52 yes I know but build will build the package Jan 10 21:51:59 and I don't want that Jan 10 21:52:01 sh: rm: command not found Jan 10 21:52:04 hehe.. Jan 10 21:52:17 what I want is to clean what does do_package Jan 10 21:52:21 while doing bitbake -c patch linux-rp :) Jan 10 21:52:27 without cleaning do_install Jan 10 21:52:44 so I do not have to recompile it each time i change something in the package Jan 10 21:53:06 there is no way Jan 10 21:53:15 with command Jan 10 21:53:16 ok thanks anyway Jan 10 21:53:20 delete the stamp files Jan 10 21:53:28 lol Jan 10 21:53:38 and try Jan 10 21:53:50 i'll move it Jan 10 21:53:53 thanks a lot Jan 10 22:10:14 woglinde, peek doesn't work for python code Jan 10 22:10:46 woglinde, or code that needs to look into the installation result first Jan 10 22:11:08 for instance: peek boost-1.36.0 BOOST_PACKAGE => None Jan 10 22:11:16 oops Jan 10 22:11:22 s/BOOST_PACKAGE/BOOST_PACKAGES Jan 10 22:11:39 but I have written BOOST_PACKAGES on the shell Jan 10 22:13:14 peek boost FILES_boost-serialization works Jan 10 22:13:59 ah yes Jan 10 22:14:52 you want peek boost PACKAGES Jan 10 22:15:06 woglinde, by the way i see fnmatch only in bitbake(not in bbclass) and at lib/bb/shell.py: return fnmatch.filter( names, pattern ) Jan 10 22:15:26 that is in the function globfilter Jan 10 22:15:39 and globefilter is only in shell.py Jan 10 22:15:45 strange isn't it Jan 10 22:17:07 woglinde, no i have PACKAGES but it returns boost-dbg boost boost-doc boost-dev boost-locale ${BOOST_PACKAGES} Jan 10 22:17:39 hm Jan 10 22:17:40 okay Jan 10 22:18:02 1. ERROR in org/thenesis/microbackend/ui/awt/AbstractAWTBackend.java (at line 20) import java.awt.Dimension; Jan 10 22:18:13 wrong channel Jan 10 22:19:33 woglinde, and peek boost-1.36.0 FILES_boost-filesystem => None Jan 10 22:19:44 that's because serialization is overridden Jan 10 22:19:57 FILES_boost-serialization = "${libdir}/libboost_serialization*.so* \ Jan 10 22:20:15 so what should I do now? Jan 10 22:20:43 becaue i must find the equivalent of the fname function Jan 10 22:20:58 s/fnmae/fnmatch Jan 10 22:29:31 woglinde, who sould I ask for fnmatch? Jan 10 22:30:01 ah org.openembedded.dev/site/sh-common Jan 10 22:32:32 ah no it's for autoconf Jan 10 22:41:41 woglinde, are you there? Jan 10 22:46:57 woglinde, i'll go to sleep...bye Jan 10 22:47:01 yes Jan 10 22:47:12 but I am working on an diffrent problem Jan 10 22:47:58 woglinde, ok thanks a lot anyway...i'll ask someone else for fnmatch tomorrow Jan 10 22:51:33 anybody interesting in teaching linux and programming c on linux Jan 10 22:54:26 hm Jan 10 22:54:40 there are enugh examples in the www Jan 10 22:54:56 and programmin c is n't so much diffrent Jan 10 22:55:09 or are you a visual studio user? Jan 10 22:55:35 yes I am vS2005 user Jan 10 22:56:01 I am a newbie on linux but somehow attracted towards linux Jan 10 22:56:09 specially embedded linux Jan 10 22:56:28 but need somebody expert in linux to teach me Jan 10 22:56:36 I am ready to pay for the time Jan 11 00:12:36 woglinde: stll active? Jan 11 00:13:06 hm yes Jan 11 00:13:09 args Jan 11 00:13:16 to late already Jan 11 00:13:37 just for a quick consulting...can be tomorrow, np Jan 11 00:13:43 no ask Jan 11 00:13:47 5 minutes Jan 11 00:15:16 we try to slip in cpio.gz one files coming from another package Jan 11 00:15:21 one file Jan 11 00:15:42 I had to hardcode the path Jan 11 00:15:57 ant why? Jan 11 00:16:05 PACKAGE_ARCH differs so I'm a bit lost with directories Jan 11 00:17:18 we try to use kexecboot instead of init Jan 11 00:18:42 I'm thinking that is task for initramfs-kexecboot-image.bb Jan 11 00:23:35 I have somehow the idea that 2 packages could be enough Jan 11 00:24:17 hm Jan 11 00:24:18 hehe Jan 11 00:24:26 so whats the problem now? Jan 11 00:24:37 hardcode path Jan 11 00:24:42 now Jan 11 00:25:11 as I told you Jan 11 00:25:15 this will not work Jan 11 00:26:21 I'm already happy it works..now we should improve the .bb files Jan 11 00:26:43 woglinde: it's more a proof of concept, for the moment Jan 11 00:35:08 03Mike Westerhof  07org.openembedded.dev * r1ca0d99090 10openembedded.git/packages/opkg/ (2 files in 2 dirs): opkg_nogpg-nocurl: update fix_tmpdirs patch, bump PR Jan 11 00:47:23 * Jay7 is looking for ThinkPad x60/x61 tablet Jan 11 00:51:33 03woglinde  07org.openembedded.dev * r095aeec3dd 10openembedded.git/packages/uclibc/ (uclibc-0.9.30/pthread_atfork.patch uclibc_0.9.30.bb): Jan 11 00:51:33 uclibc: correct pthread_atfork function for linuxthread Jan 11 00:51:33 * bump PR Jan 11 00:56:12 03woglinde  07org.openembedded.dev * rd4e4f05c5e 10openembedded.git/conf/checksums.ini: checksums.ini: add sdlpango Jan 11 00:57:57 * Jay7 -> sleep Jan 11 01:10:55 'nite Jan 11 02:44:15 Have anybody run Doom on your oe image? **** ENDING LOGGING AT Sun Jan 11 02:59:57 2009