**** BEGIN LOGGING AT Tue Feb 05 02:59:57 2008 Feb 05 09:01:27 Howdy folks! Is bugs.openembedded.org down? I can't connect Feb 05 09:01:41 jwilliams: look up Feb 05 09:03:28 XorA: Thanks. I was following a link on the Angstrom page. I will ask the maintainer to update the links Feb 05 09:03:52 jwilliams: it will be fixed when new server is in place, which hopefully is soon Feb 05 09:04:25 OK, cheers Feb 05 09:11:37 Bonjour! Feb 05 09:26:46 does gnash support alsa Feb 05 09:27:52 i don't remember Feb 05 09:55:25 hi, I have a small question, why is it, that oe is allways downloading the package after every build? Feb 05 09:55:36 cant it store it locally if it installed it once? Feb 05 10:09:26 moin Feb 05 10:09:57 morning Feb 05 10:11:28 morning! Feb 05 10:12:16 hey guys Feb 05 10:12:29 morning all Feb 05 10:12:43 Azrae1: it only downloads once unless you delete the downloads directory Feb 05 10:13:56 ah ok, so it only need some more packages, i was a little bit confused because i compiled task-base and then the bootstrap-image, and it said that it is a image of task-base Feb 05 10:47:30 03rwhitby 07org.oe.dev * rddd7d8fb... 10/ (5 files in 4 dirs): linux-ixp4xx: Update defconfigs and sane-srcrevs for latest patches that have been accepted upstream Feb 05 10:47:34 03rwhitby 07org.oe.dev * r93048632... 10/ (1 conf/machine/include/ixp4xx.inc): ixp4xx.inc: Move default ixp4xx kernel to 2.6.24 Feb 05 10:47:40 03rwhitby 07org.oe.dev * r7ed2521d... 10/ (14 files in 9 dirs): ixp4xx-kernel: Replaced by linux-ixp4xx since Nov 2007 Feb 05 11:13:24 good morning Feb 05 11:14:14 Ello Feb 05 11:15:21 lo Feb 05 11:24:05 ERROR: Error in executing: Feb 05 11:24:06 ERROR: Exception: Message:[Errno 2] No such file or directory: '/tmp/work/em-x270-angstrom-linux-gnueabi/angstrom-version-1_2008.1-test-20080205-r1/install/angstrom-version.lock' Feb 05 11:24:15 03xora 07org.oe.dev * r77ca7603... 10/ (1 packages/opkg packages/opkg/opkg_svn.bb): Feb 05 11:24:15 opkg_svn.bb : The Open Package Manager, a fork of ipkg actively developed Feb 05 11:24:15 with OE patches already incorporated and new features being add like Feb 05 11:24:15 signed feeds. Feb 05 11:24:35 sry, but im a little bit frustrates of the compelation, allways is something not worging Feb 05 11:24:57 Azrae1: Ive only ever seen that message when my machine ooopsed Feb 05 11:25:35 nop, machine is working fine, maybe it is because i lost my internet connectoin, so what should i do? Feb 05 11:25:57 Azrae1: Im not sure as I dont 100% know what that means Feb 05 11:26:21 so do i XD Feb 05 11:26:51 can i make a clean, somehow? Feb 05 11:26:59 without deleting the downloaded packages Feb 05 11:27:24 Azrae1: rm -rf tmp/ should do it; iirc, the downloaded packages are stored in sources/ Feb 05 11:27:34 er, sorry Feb 05 11:28:11 Someone else will need to confirm that Feb 05 11:28:16 my build dir is clean ATM Feb 05 11:28:31 XorA: I think mickey|zzZZzz already committed opkg Feb 05 11:28:39 Do we have it twice now? Feb 05 11:28:39 hm, i think i could try this on Feb 05 11:29:01 Azrae1: Just check that the downloaded packages aren't in tmp/, and if so, it should be safe to remove that Feb 05 11:29:23 Laibsch: I cant see it in my repo Feb 05 11:29:28 i simply moved the tmp-dir, so i have a backup Feb 05 11:29:49 * Laibsch searches for opkg Feb 05 11:31:22 03xora 07org.oe.dev * r4a6c8f44... 10/ (1 packages/opkg/opkg_svn.bb): opkg_svn.bb : fix the DEPENDS Feb 05 11:32:05 XorA: I can't either Feb 05 11:32:16 It probably was just me recollecting him talking about it Feb 05 11:32:20 Thanks for committing Feb 05 11:32:29 all talk no action from mickey|zzZZzz :-D Feb 05 11:32:34 What needs to be done to have opkg instead of ipkg for a distro? Feb 05 11:32:48 XorA: yeah, he is notorious for that ;-) Feb 05 11:32:55 NOT Feb 05 11:33:18 Laibsch: opkg should be pretty much a work alike ipkg, but everything is called opkg Feb 05 11:33:25 I know Feb 05 11:33:40 Laibsch: so thats all you need to do to replace it in distro Feb 05 11:33:41 So all I need is change the image files? Feb 05 11:34:05 ok, only moving the tmp-director is enouth, i don't have to redownload the sources Feb 05 11:34:10 and fixup so there is opkg-collateral stuff, and sign you Packages files Feb 05 11:34:35 thanks a lot for you help, I will now go and drink some coffee(for 3-4h damn slow computer) Feb 05 11:35:43 XorA: ipkg and opkg can coexist? Feb 05 11:36:13 Laibsch: they conflict on update-alternatives Feb 05 11:36:16 Laibsch: but other than that they will sit side by side Feb 05 11:36:19 * Laibsch will need to take a closer look at this Feb 05 11:37:25 re Feb 05 11:38:00 Laibsch: sonkei could be the first opkg distro in OE :-) Feb 05 11:38:27 * Laibsch rushes ;-) Feb 05 11:38:37 But what about openmoko? Feb 05 11:38:43 Do they still use ipkg? Feb 05 11:38:59 Sonkei is experimental enough at this point that I am willing to break it Feb 05 11:41:51 Laibsch: openmoko will take a while, opkg is being working in in parallel with GTA02 release, so not likely to switch until after that Feb 05 11:42:05 OK Feb 05 11:44:43 XorA: will opkg-utils-native follow? Feb 05 11:45:37 and opkg-native Feb 05 11:45:50 hrw: you know the answer to that as you saw it over ---? Feb 05 11:45:52 > Feb 05 11:47:02 XorA: to tell the true I did not had a time to follow om-dev channel Feb 05 11:47:21 hrw: well Im waiting on new release of opkg for native Feb 05 11:47:33 k Feb 05 11:47:42 ipkg-utils-native will work for opkg I think Feb 05 11:52:22 XorA: Fetching opkg fails Feb 05 11:53:06 Laibsch: dont forget to choose a SRCREV :-) Feb 05 11:53:29 Oh, I need to do that? Feb 05 11:53:36 local.conf or sonkei.conf, right? Feb 05 11:53:50 But that is tricky to make such a commit Feb 05 11:53:52 ;-) Feb 05 11:53:54 wait a bit as I forgot to set version in PV anyway Feb 05 11:54:17 Which SRCREV is the most recent? Feb 05 11:56:36 Laibsch: 4017 Feb 05 11:56:51 pushing updates to OE now Feb 05 11:57:45 Can' t that be committed to sane-srcrevs.inc? Feb 05 11:59:48 4018, even Feb 05 12:00:38 What is ipkg-collateral doing after all? Feb 05 12:00:49 I don't think there is an opkg-collateral, yet Feb 05 12:05:34 03xora 07org.oe.dev * rb84b8bcb... 10/ (4 files in 3 dirs): opkg_svn.bb : remember to set PV and update (in)sane-srrevs.inc Feb 05 12:21:59 Thanks, XorA Feb 05 12:22:26 hrw: I am reworking what xjqian|zzzzzzzzz has done to fix Feb 05 12:22:30 !oebug 3720 Feb 05 12:22:31 * * Bug 3720, Status: NEW, Created: 2008-01-29 21:48 Feb 05 12:22:32 * * GNUtoo(AT)no-log.org: opie 1.2.3 fails to start (black screen) under qemu Feb 05 12:22:34 * * http://bugs.openembedded.net/show_bug.cgi?id=3720 Feb 05 12:22:58 Does opie-ttf-support need to call update-fonts-common in postinst? Feb 05 12:23:02 I don't think so Feb 05 12:23:31 Laibsch: might be an idea to, just to get the ttf fonts that already on system before you install ttf support Feb 05 12:23:50 otherwise you'll need to install a font to get them to show Feb 05 12:24:12 XorA: then it would need to RDEPENDS on font-update-common, too Feb 05 12:24:33 * Laibsch is loosing oversight over the tangled dependencies for this thing Feb 05 12:25:00 My take is that if somebody has non-packaged fonts on the system, they need to call update-font-common by hand Feb 05 12:25:17 or was the script called update-fonts? Feb 05 12:25:24 I think so Feb 05 12:27:49 well, currently, 02update-$blah is called, so that works Feb 05 12:28:10 I'll just leave that in Feb 05 12:33:15 i have just compiled gnash, i get the error "Could not find any -gnash executable under directory /usr/bin, with in" Feb 05 12:33:34 the gnash package didnt gave me only an executable called gnash Feb 05 12:35:17 gnash is a shell script Feb 05 12:39:32 XorA: still interested in failure reports? http://rafb.net/p/P5oBnZ41.html Feb 05 12:48:00 03pfalcon 07org.oe.dev * r647847c7... 10/ (1 conf/machine/h5000.conf packages/tasks/task-base.bb): Feb 05 12:48:00 h5000.conf: Ship BT module. Feb 05 12:48:00 * Fixes #3549 Feb 05 12:48:00 * Based on the patch by Michal Panczyk Feb 05 12:48:05 03xora 07org.oe.dev * r8dd7d29f... 10/ (4 files in 3 dirs): Feb 05 12:48:05 opkg-native_svn.bb : opkg native version for brave people who want Feb 05 12:48:05 to try and build images with opkg. Feb 05 12:48:12 03pfalcon 07org.oe.angstrom-2007.12-stable * r7d0b1888... 10/ (1 BACKPORTS.txt): BACKPORTS.txt: Note pending h5000 fixes. Feb 05 12:59:05 * * OE Bug 3720 has been RESOLVED (FIXED) by Feb 05 12:59:07 * * opie 1.2.3 fails to start (black screen) under qemu Feb 05 12:59:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3720 Feb 05 12:59:28 Damn. There are already 3 or 4 classes for representing display devices. And none of them is sufficient for my case. Feb 05 13:03:26 03Laibsch 07org.oe.dev * r47b75b67... 10/ (3 files in 3 dirs): Feb 05 13:03:26 opie-ttf-support: add 02qtttffont-update and move everything possible from the bb file there Feb 05 13:03:26 * improves readability and robustness Feb 05 13:03:26 * closes 3720 together with e25947a6e6d2be708fd836b7d5209a3d074a28a5 from xjqian Feb 05 13:37:24 XorA: opkg-native bitbakes fine, opkg does not Feb 05 13:39:09 opkg works here Feb 05 13:41:20 \o/ Feb 05 13:41:26 i should try it too Feb 05 13:41:49 i'm trying to get nand flash working on my x86 device Feb 05 13:42:30 the bios doesn't support it so i try to modify the kernel module of the controller to set the register Feb 05 13:44:37 i don't know really if nand flash need to be hard mapped or not Feb 05 14:11:37 XorA: missing dependency? Feb 05 14:12:05 * * OE Bug 3742 has been created by Feb 05 14:12:06 * * Latest opie images fail calibration Feb 05 14:12:09 * * http://bugs.openembedded.net/show_bug.cgi?id=3742 Feb 05 14:12:48 Laibsch: whats the error? Feb 05 14:13:01 http://rafb.net/p/P5oBnZ41.html Feb 05 14:13:09 I thought you read that Feb 05 14:13:48 Laibsch: must have got lost in the void, that looks like a too old gpgme/libgcrypt Feb 05 14:14:36 XorA: libgcrypt is 1.2.3-r1 Feb 05 14:15:16 XorA: gpgme is 1.1.4-r1 Feb 05 14:15:50 hmm, think it might need libgcrypt 1.4.0, I remeber the author asking me to upgrade that packge Feb 05 14:15:59 only libgcrypt has a later version in oe Feb 05 14:16:11 OK, I'll bake it and retry Feb 05 14:18:48 XorA: it still fails /home/laibsch/work/tmp/sonkei-0.0.1/work/armv5te-sonkei-linux-gnueabi/opkg-0.0 Feb 05 14:19:02 http://rafb.net/p/4Q6FZK34.html Feb 05 14:19:06 there you go Feb 05 14:21:48 Laibsch: libgpg-error? Feb 05 14:22:01 I wouldn't know Feb 05 14:22:09 I have 1.4 of that here Feb 05 14:22:19 * Laibsch looks Feb 05 14:26:01 XorA: Looks like I have 1.4, too Feb 05 14:26:08 :-( Feb 05 14:26:15 really no idea on that error then Feb 05 14:26:50 :-( Feb 05 14:31:31 XorA: http://rafb.net/p/U5VrjP64.html for now? Feb 05 14:32:01 I think opkg uses libgcrypt via gpgme Feb 05 14:32:07 so it doesnt directly use it Feb 05 14:32:26 yes, but opkg depends on that specific version Feb 05 14:33:45 does it? I thought that didnt fix the error? Feb 05 14:36:11 Oh, still the same error? Feb 05 14:36:22 I thought we had moved on to the next error ;-) Feb 05 14:41:04 * * OE Bug 3743 has been created by Feb 05 14:41:06 * * opkg fails do_compile Feb 05 14:41:08 * * http://bugs.openembedded.net/show_bug.cgi?id=3743 Feb 05 14:44:01 morning Feb 05 14:50:08 is there a maintainer for ipaq-sleep? I'd like to make some updates, but the author's contact information is invalid Feb 05 14:56:02 03xora 07org.oe.dev * rbbb0c0e6... 10/ (1 packages/gpgme/gpgme_1.1.4.bb): gpgme_1.1.4.bb : improve packaging, gpgme-config moved to -dev package Feb 05 14:56:09 apaulsen: oh... that might be outdated right. I can take care about it... Feb 05 14:56:58 * florian wishes he had a little bit of time to replace this and a lot of other stuff with a good middleware Feb 05 14:58:01 There is a package called gxine in packages directory. when i try to do bitbake gxine i get an error saying : ERROR: Required build target 'gxine' has no buildable providers. Feb 05 14:58:56 florian: do you mean you will fix the contact information or that I should send updates to you? Feb 05 15:01:36 apaulsen: send updates to me Feb 05 15:02:55 apaulsen: iirc I was the last who worked on it and colin and pb_ are not that active any more Feb 05 15:05:33 florian: ok, if I get around to making the updates and they actually work, I will contact you Feb 05 15:05:49 apaulsen: ok great - many thanks Feb 05 15:05:54 * florian loves updates Feb 05 15:12:29 rschuster: did you tried to use glibc directly from maemo sdk instead of building it? Feb 05 15:13:12 hrw: no. just took the one that was in OE Feb 05 15:13:18 ok Feb 05 15:14:05 rschuster: I am thinking about creating own feed for n810 and wondering how many OE generated pacakges should I remove before 'apt-get update' Feb 05 15:16:25 Should every bb file contain a PROVIDES statement? Feb 05 15:17:08 hrw: the glibc wont upgrade,libstdc++ does, but doesnt seem to break anything Feb 05 15:17:16 hrw: at least the important ones should not be taken from OE (glibc, glib, gtk and such) Feb 05 15:17:23 anyone here using a graphics tablet? Anything special to consider for Linux? Feb 05 15:17:40 hrw: nokia provides a list of officially distributed libraries Feb 05 15:18:25 hrw: http://tablets-dev.nokia.com/4.0/4.0b_vs_4.0_content_comparison.html Feb 05 15:18:38 hrw: perhaps this helps to find out what to leave out for the feeds Feb 05 15:20:41 thx Feb 05 15:21:13 ~lart nokia for two Shift keys working as LShift Feb 05 15:21:14 * ibot drops a baby grand on nokia for two Shift keys working as LShift Feb 05 15:22:55 hrw: is it hardware problem or software? Feb 05 15:24:55 according to Nokia guy it is hardware Feb 05 15:25:01 http://article.gmane.org/gmane.comp.handhelds.maemo.devel/12585 Feb 05 15:25:06 The keyboard isn't wired to Feb 05 15:25:06 distinguish left and right shift so you only get . Feb 05 15:26:29 lumag: remember - it is Nokia.. Finland.. cold.. alcohol is expensive.. Feb 05 15:27:03 :) Feb 05 15:27:08 :))) Feb 05 15:27:24 hrw: canada is cold too and alcohol can be expensive too Feb 05 15:35:25 03mickeyl 07org.oe.dev * rbbd1c5e5... 10/ (1 packages/gstreamer/gst-plugin-pulse_0.9.7.bb): add gst-plugin-pulse 0.9.7 (to go w/ PA 0.9.8 and later), DP = -1 Feb 05 15:35:29 03xora 07org.oe.dev * receb5780... 10/ (4 files in 2 dirs): Feb 05 15:35:29 angstrom-2007-preferred-versions.inc : renamed for openmoko as its only really Feb 05 15:35:29 been left in OE for that use. angstrom-2007-for-openmoko-versions.inc Feb 05 15:36:16 Damn, damn lcd, backlight, etc. interface. Feb 05 15:36:35 !lart class_backlight, info->bl_dev and company Feb 05 15:36:35 * cdbot2 pulls out his louisville slugger and uses class_backlight, info->bl_dev and company's head to break the homerun record Feb 05 15:41:07 * lumag is trying to properly hook up the lcd & bl driver on tosa. Feb 05 16:06:24 03pH5 07org.oe.dev * rd835703f... 10/ (1 packages/xorg-proto/glproto_1.4.9.bb): glproto: add 1.4.9 Feb 05 16:06:28 03pH5 07org.oe.dev * r6369a7be... 10/ (1 packages/xorg-proto/xproto_7.0.11.bb): xproto: add 7.0.11 Feb 05 16:06:32 03pH5 07org.oe.dev * r11c8bbda... 10/ (1 packages/xorg-app/bitmap_1.0.3.bb): bitmap: add 1.0.3 Feb 05 16:06:37 03pH5 07org.oe.dev * r8a29ca42... 10/ (1 packages/xorg-app/luit_1.0.3.bb): luit: add 1.0.3 Feb 05 16:06:41 03pH5 07org.oe.dev * r4c220b92... 10/ (1 packages/xorg-app/xcalc_1.0.2.bb): xcalc: add 1.0.2 Feb 05 16:06:46 03pH5 07org.oe.dev * rddbd12bb... 10/ (1 packages/xorg-app/xclock_1.0.3.bb): xclock: add 1.0.3 Feb 05 16:06:50 03pH5 07org.oe.dev * rcee217f0... 10/ (1 packages/xorg-app/xconsole_1.0.3.bb): xconsole: add 1.0.3 Feb 05 16:06:55 03pH5 07org.oe.dev * r576b3d7a... 10/ (1 packages/xorg-app/xcursorgen_1.0.2.bb): xcursorgen: add 1.0.2 Feb 05 16:07:00 03pH5 07org.oe.dev * rd35887c6... 10/ (1 packages/xorg-app/xdriinfo_1.0.2.bb): xdriinfo: add 1.0.2 Feb 05 16:07:04 03pH5 07org.oe.dev * r84305d74... 10/ (1 packages/xorg-app/xinit_1.0.7.bb): xinit: add 1.0.7 Feb 05 16:07:09 03pH5 07org.oe.dev * rae923f15... 10/ (1 packages/xorg-app/xload_1.0.2.bb): xload: add 1.0.2 Feb 05 16:07:13 03pH5 07org.oe.dev * r094c0f06... 10/ (1 packages/xorg-app/xlsfonts_1.0.2.bb): xlsfonts: add 1.0.2 Feb 05 16:07:17 03pH5 07org.oe.dev * r5ca1d96c... 10/ (1 packages/xorg-app/xmag_1.0.2.bb): xmag: add 1.0.2 Feb 05 16:07:22 03pH5 07org.oe.dev * r32459eac... 10/ (1 packages/xorg-app/xman_1.0.3.bb): xman: add 1.0.3 Feb 05 16:07:26 03pH5 07org.oe.dev * r92b1f63c... 10/ (1 packages/xorg-app/xmessage_1.0.2.bb): xmessage: add 1.0.2 Feb 05 16:07:31 03pH5 07org.oe.dev * r2304edb6... 10/ (1 packages/xorg-app/xsetpointer_1.0.1.bb): xsetpointer: add 1.0.1 Feb 05 16:07:36 03pH5 07org.oe.dev * r74619a24... 10/ (1 packages/xorg-app/xwininfo_1.0.3.bb): xwininfo: add 1.0.3 Feb 05 16:07:40 03pH5 07org.oe.dev * rd7826d9c... 10/ (1 contrib/oe-stylize.py): oe-stylize.py: add support for the PE (package epoch) variable Feb 05 16:26:23 * lumag is going now Feb 05 16:26:24 bye! Feb 05 16:29:42 when i try to build gxine, i get the errror: ERROR: Required build target 'gxine' has no buildable providers. Feb 05 16:30:58 i added a PROVIDES="virtual/gxine" in gxine.bb file, but without sucess Feb 05 16:46:40 ashu: sounds like it's looking for "gxine", not "virtual/gxine" Feb 05 17:14:45 bye Feb 05 18:10:21 XorA: BTW, Sonkei uses some of the Angstrom stuff in OE for its overlay Feb 05 18:25:36 Hey. Newbie here with a few questions. Anyone up for stupid questions? Feb 05 18:25:53 * teamnoir waves Feb 05 18:26:07 just ask Feb 05 18:26:16 If we ave answers, we speak Feb 05 18:26:23 otherwise we keep quiet :) Feb 05 18:26:50 * Crofton knows irc can be really frustrating for "new users" Feb 05 18:26:55 What's the difference between "def foo():" and "python foo () {" Feb 05 18:27:03 um Feb 05 18:27:09 * teamnoir was an irc person over a decade ago, but times change. Feb 05 18:27:19 that seems to be a python question? Feb 05 18:27:31 I think it's a bitbake question. Feb 05 18:27:40 hmm Feb 05 18:27:48 * Crofton is not a python person Feb 05 18:27:55 although that may have to change ... Feb 05 18:28:01 Neither of those forms are python. This looked like the best cannel for bitbake questions. Feb 05 18:28:14 yeah Feb 05 18:28:15 it is Feb 05 18:28:23 the bitbake guys are not many though Feb 05 18:28:25 All of my questions are about the python/bitbake interface. Feb 05 18:28:28 and they stay very busy Feb 05 18:28:36 * teamnoir chuckles Feb 05 18:28:44 It looks like the sort of project that encourages that. Feb 05 18:28:51 teamnoir: the first is a python function definition in python, the latter is a python function definition in a bitbake recipe. Feb 05 18:29:00 stick around, they can be pretty good about checking the scroll back when they have time Feb 05 18:29:16 ph5: what's the difference in usage? Feb 05 18:29:23 teamnoir: we have tasks, e.g. do_compile() {} normally they are shell Feb 05 18:29:36 * teamnoir nods Feb 05 18:29:44 teamnoir: but with saying python do_compile you set the python flag on this task Feb 05 18:29:59 teamnoir: and when bb.build.exec_task gets called the python interpreter gets used Feb 05 18:30:09 * teamnoir nods Feb 05 18:30:19 teamnoir: the other is pure python, and in the global 'namespace/objectspace' of the bitbake runtime Feb 05 18:30:42 teamnoir: so they at least have two difference scopes Feb 05 18:30:48 zecke: so it's a question of scope? Of the visibility of the function? Feb 05 18:31:02 teamnoir: def global(), python a task with no scope/local one Feb 05 18:31:12 zecke: L=2.5e-3; Feb 05 18:31:12 T=1024.0/25e6; Feb 05 18:31:21 zecke woops Feb 05 18:31:21 zecke: ty. Feb 05 18:31:45 zecke: http://marketplace.publicradio.org/display/web/2008/02/04/currywurst/ Feb 05 18:32:22 next question - the "d" values to setVar and getVar... these are some form of bitbake scoping? I'm setting a variable, seeing it set in the d.dict, but subsequent uses, (like SRC_URI), aren't seeing those values. So I'm confused. Feb 05 18:33:07 teamnoir: 'd' is a global variable and valid in the executed task (with anonfuncs as exception) Feb 05 18:33:50 Prior to executing a task... d represents something else it seems... Feb 05 18:33:59 teamnoir: e.g. when you execute a task, the current bb.data dict gets put into globals() Feb 05 18:34:05 teamnoir: current is the current bitbake file. Feb 05 18:34:27 zecke: ok, so I'm trying to set a variable prior to the bitbake file being read. Feb 05 18:34:40 teamnoir: it has always a scope. When parsing the configuration files, d is the configuration object, inherited/used as a base for all bb files Feb 05 18:34:46 zecke: as I want to inform the value of SRC_URI. Feb 05 18:35:04 zecke: ok. Feb 05 18:35:33 teamnoir: why would you want that? the only way would be in the global configuration (and using OVERRIDES SRC_URI_pn-mypackage) Feb 05 18:37:05 It's a long story, involving a distribution that isn't a common application of OE. But what I'm trying to implement is a single file which both describes what packages are included in an image as well as their versions. (complicated with a local convention for "submissions" that are stored in conventional subversion locations). Feb 05 18:38:25 teamnoir: the closest you can get is global configuration + overrides Feb 05 18:38:28 So far I have a class, specified in INHERIT, which defines a python structure of these values. And I'm trying to set variables from the structure by traversing and setVar. Feb 05 18:38:36 teamnoir: What about storing the versioning info in a file like sane-srcdates.inc? Feb 05 18:39:16 hvontreslwork: I don't know sane-srcdates, but I'm looking for a structure that includes both image inclusion as well as versioning info concurrently. Feb 05 18:40:04 The problem I'm seeing in my current approach is that base_set_filenames is apparently broken because it can't expand ${SUBISSION_busybox}. Feb 05 18:40:24 So I'm presuming that I'm not setVar'ing in the right context. Feb 05 18:40:37 hi everybody Feb 05 18:45:20 Ok, thanks all. I guess I'm off to puzzle through some more python. Feb 05 18:45:34 * teamnoir waves Feb 05 18:46:20 good luck Feb 05 18:46:28 ty Feb 05 18:49:48 zecke: could you look at bug 3739 and commit it if ok or comment if not ok? Feb 05 19:04:46 !oebug 3739 Feb 05 19:04:48 * * Bug 3739, Status: NEW, Created: 2008-02-03 17:01 Feb 05 19:04:49 * * thommycheck(AT)gmx.de: add an initramfs that mounts root and kexecs a kernel from there. Feb 05 19:04:50 * * http://bugs.openembedded.net/show_bug.cgi?id=3739 Feb 05 19:06:23 thesing: are you going to be here in 30 minutes? Feb 05 19:06:34 zecke: yes. Feb 05 19:13:45 hello, does someone know well busybox...why chroot throw me an illegal instruction? Feb 05 19:21:59 GNUtoo: last time it happend to me was yesterday Feb 05 19:22:19 it was because my qemu did not support eabi Feb 05 19:22:26 keesj, i'm compiling a static chroot Feb 05 19:22:46 keesj, it's on a real machine and i think that the abi is good (i'll verify it) Feb 05 19:22:51 but is it the chroot of the bash/init after the chroot that throws the exepction Feb 05 19:23:58 keesj, no i want to update the linux distro of my device so i want to try mine with chroot Feb 05 19:25:08 so you need a "native" chroot right? so does chroot / work? Feb 05 19:25:29 GNUtoo: that is all I know sorry :p Feb 05 19:25:36 keesj, yes it work ...mabe that's because of the 2 different libc Feb 05 19:27:08 I don't thikn that could be it. the busybox inside your new distro is staticaly linked right? Feb 05 19:27:45 re Feb 05 19:33:24 keesj, yes it is Feb 05 19:39:17 thesing: okay, what is this kexec patch? is that a copy of another kexec patch? Feb 05 19:40:05 zecke: yes. I wanted to make a softlink to the normal kexec patchdir. But monotone doesn't support links. Feb 05 19:40:40 zecke: so I copied the patches. Feb 05 19:42:21 thesing: why do you need the other kexec tools copy again? Feb 05 19:42:38 (it is not the area where I would have super cow knowledge) Feb 05 19:43:30 zecke: because I don't want to include glibc in the initramfs. Feb 05 19:45:07 thesing: let me take another look Feb 05 19:45:21 thesing: so -static is the only difference? Feb 05 19:46:37 zecke: yes. kexecs configure hardcodes cflags. The recipe adds -static to them. (kexec-static.patch) Feb 05 19:47:12 thesing: grep for FILESDIR in the packages/ dir, you want to use that :) Feb 05 19:48:59 thesing: Laibsch is pretty good at this, I always try to push as much as possible into a .inc file Feb 05 19:51:13 zecke: so I could set FILESDIR to the dir of normal kexec and add the kexec-static patch there. Feb 05 19:52:05 thesing: you can add the normal kexec dir and use the common patch, and have the -static patch in the other directory :) Feb 05 19:52:35 thesing: this would be the minimum before I commit and push, and then I encourage you to try to minimize the differences between the two versions :) Feb 05 19:52:52 thesing: and then you should get push access as well, and we should get git and reviewboard going Feb 05 19:53:24 zecke: is the other stuff ok? Feb 05 19:54:41 yeah, looks sane Feb 05 19:56:00 zecke: I will change this. Feb 05 19:59:53 zecke: There is not much to unify between the two kexec recipes apart from the patches. Feb 05 20:05:57 is there a quality check for a package (to see if some files are missing from the packages) Feb 05 20:06:42 keesj: missing in regard to a previous version? no, installed but not packages, yes Feb 05 20:09:35 yes , I guess installed but not packaged Feb 05 20:25:35 zecke: was "DEPENDS = "task-boot task-qtopia-core-gui task-qtopia-core-gui" " in qtopia-core-gui-image.bb right? Feb 05 20:26:05 task-boot probably yes, the others not Feb 05 20:30:57 keesj, i just saw that chroot is not staticaly linked...i'll compile coreutils staticaly Feb 05 20:31:56 zecke: I attached the a new patch to the bug. Feb 05 20:33:25 thesing: can you remove the garbage at the beginning of the file? :) Feb 05 20:33:49 that is oelint? ugly :) Feb 05 20:34:57 zecke: this was oe-stylize. it should use stderr for warnings.. Feb 05 20:35:35 I don't feel like removing these lines myself Feb 05 20:35:39 *lazy* Feb 05 20:41:46 zecke: updated the patch again. I had do remove these line locally anyway. Feb 05 20:46:02 thesing: oh, sorry to bother you. Normally we don't require other bb files :) Feb 05 20:47:26 thesing: could you move the FILESDIR into the -static.bb as well? Feb 05 20:47:35 zecke: so I should remove the revision. rename it to .inc and create a new .bb with just the revision inside? Feb 05 20:48:05 thesing: yes, most of the times I mv the .bb to .inc and then make it 'sharable' Feb 05 20:51:34 * Laibsch looks at bug 3739 Feb 05 20:51:37 sounds interesting Feb 05 20:51:57 * Laibsch votes in favor of thesing RW access Feb 05 20:52:33 Laibsch: it works great. My kernel even has a logo ;) Feb 05 20:52:54 * Laibsch wonders what a kernel with a logo looks like Feb 05 20:54:19 thesing: I think bb files should at least have a DESCRIPTION Feb 05 20:54:33 plus LICENSE Feb 05 20:55:29 AUTHOR would be nice Feb 05 20:55:40 In this case, I guess it would be you Feb 05 20:56:10 aren't all bb-files supposed to be bsd licensed anyway? Feb 05 20:58:07 Laibsch: not many bbs have AUTHOR set. Feb 05 20:58:41 We have the strange policy to keep the author/maintainer info outside... Feb 05 20:59:08 Not quite right Feb 05 20:59:25 LICENSE and AUTHOR of course refer to the outside code Feb 05 20:59:29 not the bb file itself Feb 05 20:59:33 The bb is MIT Feb 05 20:59:49 In this case AUTHOR might not be applicable Feb 05 21:04:01 ok. patch updated again. Feb 05 21:04:41 but I forgot to test if it still builds. Feb 05 21:05:05 which it doesn't. Feb 05 21:15:59 Laibsch: do you know how this filesdir stuff works? the kexec-static.patch isn't applied somehow. Feb 05 21:16:32 I am not sure Feb 05 21:16:48 But I think FILESDIR specifies where bitbake should look for files Feb 05 21:17:00 specified in the SRC_URI Feb 05 21:17:06 What are you trying to do? Feb 05 21:17:17 Can we break this to individual pieces? Feb 05 21:17:28 thesing: What is the first recipe you need ot add? Feb 05 21:17:36 to Feb 05 21:18:27 thesing: your filesdir looked wrong :) Feb 05 21:18:40 the klibc-utils-static file Feb 05 21:19:46 thesing: and sorry you want FILESPATH and append to it :) Feb 05 21:20:14 I will post separate patches. Feb 05 21:20:22 zecke: I don't like mega-patches Feb 05 21:20:28 thesing: That would be great Feb 05 21:20:38 divide and conquer ;-) Feb 05 21:20:54 zecke: Any objection to the klibc-utils-static bb file? Feb 05 21:21:21 not really, it should go in Feb 05 21:21:48 thesing: Can't you use the .inc file for klibc-utils-static? Feb 05 21:22:18 Laibsch: not without some work. Feb 05 21:22:20 If there is any setting that is not appropriate for your version, you can either override by specifying a different value in the bb file Feb 05 21:22:38 Can you paste just the bb file? Feb 05 21:22:39 I will try. Feb 05 21:22:42 One of the things were OE is lacking is variants/USE flags, recursive rebuilding, debian packages, easily changing flags (e.g. just adding -static to a couple of packages) Feb 05 21:22:44 I'll do it for you Feb 05 21:22:57 Laibsch: thanks Feb 05 21:23:14 anyone in Berlin searching for a kernel hacker job? Feb 05 21:23:24 (fulltime though) Feb 05 21:23:39 zecke: You want to hack on the kernel? Feb 05 21:23:52 Or you have an offer? Feb 05 21:24:18 zecke: variants are not a strong point of OE Feb 05 21:24:40 But I guess one can make do with .inc files and specifying just what is different in a bunch of bb files Feb 05 21:24:50 not the most elegant, but I think it should work Feb 05 21:25:32 Laibsch: can offer one Feb 05 21:25:46 zecke, thesing: xxdiff and "diff -y" are great tools for unification Feb 05 21:26:21 Laibsch: I like macports variants, but this does not work for packaged distributions Feb 05 21:27:00