**** BEGIN LOGGING AT Thu Jul 08 02:59:57 2010 Jul 08 03:09:29 khem: Ah Jul 08 03:09:45 Did manage to hit that in my next round of stuff, but didn't think relocatible.bbclass hit that normally Jul 08 03:09:57 I'll push something shortly, thanks Jul 08 03:11:46 That's pretty dang crazy Jul 08 03:11:53 0x0000000f (RPATH) Library rpath: [/home/trini/work/OE-upstream/tmp.glibc.angstrom/sysroots/i686-linux/usr/lib:.oRIGIN/../lib] Jul 08 03:12:38 on /home/trini/work/OE-upstream/tmp.glibc.angstrom/sysroots/i686-linux/usr/bin/unzipsfx Jul 08 03:16:22 03Tom Rini  07org.openembedded.dev * rf142648c46 10openembedded.git/classes/relocatable.bbclass: Jul 08 03:16:22 relocatible.bbclass: Call oe.path.relative in try/except Jul 08 03:16:22 Some programs such as libxml-parser-perl-native can have junky Jul 08 03:16:22 RPATH directories. Found by Khem Raj. Jul 08 03:16:22 Signed-off-by: Tom Rini Jul 08 03:31:11 Tartarus: thx lemme try it here too Jul 08 03:34:15 Tartarus: works well here now thx for quickfix Jul 08 05:13:39 there are some "oddities" within the gcc recipes... Jul 08 05:14:01 gcc-3.4.6 is most likely unused, as it has no cross recipe Jul 08 05:15:25 and some of the .inc files include the zecke-no-host-includes.patch - those versions are likely unused too, as the native recipes probably wont be able to build anything Jul 08 05:19:39 grg: Yeah, that was a special one for qemu 0.9.x or so Jul 08 05:54:57 I think we can punt most of 3.x recipes I believe Jul 08 05:56:47 ./distro/nylon.conf still prefers 3.3.4 Jul 08 05:57:24 and there are several references to 3.4.4 Jul 08 05:57:36 ah Jul 08 05:57:43 yes nylon I remember Jul 08 05:58:02 I wonder if it is still buildable on .dev Jul 08 05:59:13 hmmm... gmustix.conf.... hasn't been maintained since 2006... Jul 08 06:01:57 wrt54oe since 2005 Jul 08 06:08:08 :) Jul 08 06:08:19 gotta sleep now Jul 08 06:08:56 grg: try gcc 4.5 I know u have mips I am interested in perf numbers Jul 08 06:09:20 khem, i'll see what i can do Jul 08 06:09:21 if u have any becnhmarks Jul 08 06:09:49 to contrast Jul 08 06:11:30 u can use unixbench_5.1.2.bb Jul 08 06:11:51 * khem sleeps Jul 08 06:38:42 Hi! I want to add to recipe patch from http URI. The one problem is that it is packed in tar.gz with another patch. I've added SRC_URI for this archive and checksums for it, OE succesfully fetchs and unpacks it and I can see both patches from archive with other "normal" patches in sources dir, but patch task fails with "Patch patches.tar.gz does not apply (enforce with -f)". Haven't found nothing similar in other recipes, could anybody help w Jul 08 06:58:18 can i just import one function from a bbclass without inheriting the whole thing? Jul 08 06:58:35 like in python with the from gnome import verdir oder something? Jul 08 07:01:13 gim: have you tried writing your own do_patch python function? Jul 08 07:02:08 or actually gim, doesn'teven have to be python Jul 08 07:03:59 if you have a linux-efika_2.6.21+git.bb that file has an arbitrary do_patch function Jul 08 07:04:39 i think you will have to manually do all patches by yourself that way then, no only the one from the tar.gz Jul 08 07:10:32 fraxinath: yes, seems like redefining do_patch is the only way Jul 08 07:33:44 Hm, redefining do_patch is ok except one thing -- I don't understand what to do with 'file://' patches, OE doesn't copy them to WORKDIR and I can't get how I'd copy them manually from recipe Jul 08 07:40:05 morning Jul 08 07:40:34 file://${WORKDIR}/my.patch maybe? Jul 08 07:52:27 hrw: Hi. In recipe SRC_URI contain some http items and one local. If I use standard patch function everything is OK -- 'http' items are fetched to sources dir and then unpacked to WORKDIR, 'file' patches are applied directly from recipes dir. Anyway, I looked at the python_do_patch definition and seems like I should somehow parse SRC_URI by myself in my patch function Jul 08 07:56:48 gim: file://${WORKDIR}/my.patch maybe? Jul 08 07:58:58 can i just import one function from a bbclass without inheriting the whole thing? like in python with the from gnome import gnome_verdir oder something? Jul 08 08:01:08 hrw: Sorry, but I couldn't understand how it'll work. There is no my.patch in WORKDIR, my.patch is in recipes directory Jul 08 08:02:06 my.patch isn't copied to WORKDIR neither when I use standard patch nor when I use my own Jul 08 08:02:49 OE patch function apply 'file' patches directly from recipes dir and now I try to understand how to do similar thing Jul 08 08:02:58 SRC_URI="http://patches.tar.gz \ file://${WORKDIR}/patch-which-is-in-patches.tar.gz.patch" Jul 08 08:03:16 OE unpacks patches.tar.gz to WORKDIR and then use one of them Jul 08 08:03:23 fraxinath: no Jul 08 08:04:11 gim: ah.. it complains about 'patches.tar.gz' not being patch? Jul 08 08:04:49 yes Jul 08 08:05:04 "Patch patches.tar.gz does not apply (enforce with -f)" Jul 08 08:05:19 though maybe 'apply=no' will help, I'll try it Jul 08 08:05:36 hrw: how do i assign a bb varibale from within a py func? Jul 08 08:05:44 looks like patch.bbclass needs to be changed Jul 08 08:05:52 fraxinath: bb.data.setVar Jul 08 08:05:57 thx Jul 08 08:10:05 hrw, what's the syntax of that? it takes 3 vars... is there a doc anywhere? Jul 08 08:10:57 hrw: You were right, SRC_URI="http://patches.tar.gz \ file://${WORKDIR}/patch-which-is-in-patches.tar.gz.patch" worked, thank you very much Jul 08 08:11:14 *SRC_URI="http://patches.tar.gz;apply=no \ file://${WORKDIR}/patch-which-is-in-patches.tar.gz.patch" Jul 08 08:14:15 fraxinath: do not remember Jul 08 08:14:46 o_o Jul 08 08:15:06 i'll see if i can find any occurances Jul 08 08:15:49 hmm the third param seems to be some kind of a flag, it's always d Jul 08 08:17:12 fraxinath: 'd' is the datastore object Jul 08 08:17:50 which datastore object? hmm it throws ERROR: global name 'd' is not defined while parsing /dream/oe1.6/openembedded/recipes/gnome/libsoup-2.4_2.29.5.bb Jul 08 08:19:07 getVar needs it too Jul 08 08:20:13 should i mention that this is not inside a task but in a deperate def Jul 08 08:22:04 well, the datastore object that you want to operate on. oe usually passes that as 'd' to python methods. Jul 08 08:22:31 if this is a standalone def'd function (whether desperate or not) then you will have to pass it in explicitly Jul 08 08:23:39 morning Jul 08 08:45:06 I have a feeling that my ubuntu work eats disk space faster then OE did Jul 08 08:51:26 hehe.. Jul 08 08:52:15 * Jay7 should remove freebsd from notebook to build x11-image Jul 08 08:52:24 .. or buy new hdd.. Jul 08 08:55:04 if it is not very old then new hdd Jul 08 08:55:44 I'm working on x86_64 machine and want to build x86 meta-toolchain for ARM target. As I understand I should use canadian-sdk recipe, but I couldn't get what variables should I set. Could anybody help with it? Jul 08 09:04:29 03Koen Kooi  07org.openembedded.dev * rf867b84543 10openembedded.git/recipes/usb-gadget-mode/ (files/beagleboard/usb-gadget.conf usb-gadget-mode.bb): usb-gadget-mode: do not load any module by default for beagleboard Jul 08 09:24:08 03Martin Jansa  07org.openembedded.dev * r4fe49913ae 10openembedded.git/recipes/liboauth/liboauth_0.8.8.bb: Jul 08 09:24:08 liboauth: add recipe for OAuth Core RFC 5849 standard implementation (will be used in elmdentica) Jul 08 09:24:08 Signed-off-by: Martin Jansa Jul 08 09:39:06 Is there any documentation describing meta-toolchain canadian building? Jul 08 09:44:53 gim: do you want cross compile linux apps on windows ? Jul 08 09:45:39 ah for meta toolchain you don't need canadian build Jul 08 09:47:07 CruX|: I'm working on x86_64 machine and want to build x86 meta-toolchain for ARM target. As I understand I should use canadian-sdk recipe, but I couldn't get what variables should I set. Jul 08 09:49:10 gim: i have one configuration: on x86 linux i'm compiling win32 meta toolchain for arm-linux Jul 08 09:50:58 gim: here is my local.conf http://pastebin.com/6G55U31z Jul 08 09:51:05 maybee it helps you Jul 08 10:20:18 CruX|: Thanks, seems that it should help. You build canadian-sdk recipe to build toolchain, right? Jul 08 10:30:59 gim: yes Jul 08 10:36:11 03Roger Monk  07org.openembedded.dev * r4551424c80 10openembedded.git/recipes/x-load/x-load_git.bb: Jul 08 10:36:11 x-load: Update to latest commit - fix hynix memory support Jul 08 10:36:11 * Add new commit to fix SDRC params Jul 08 10:36:11 * Bump PR Jul 08 10:36:11 Signed-off-by: Roger Monk Jul 08 10:46:45 03Koen Kooi  07org.openembedded.dev * r73db290223 10openembedded.git/recipes/opencv/openscenegraph_2.9.6.bb: openscenegraph: fix GLES2 and ffmpeg build Jul 08 11:33:13 CruX|: I copied SDK sections from your config to mine and removed all mingw references. Now building canadian-sdk failes, se Jul 08 11:33:44 ... according to log, OE still tries to build mingw packages Jul 08 11:35:50 It runs tasks from recipes mingw-runtime_3.15.2.bb, mingw-w32api-headers_3.13.bb etc Jul 08 12:31:08 03Koen Kooi  07org.openembedded.dev * re3a95a050f 10openembedded.git/recipes/gstreamer/gst-musicxml2midi_git.bb: gst-musicxml2midi: add git version Jul 08 12:31:09 03Koen Kooi  07org.openembedded.dev * rd2514f5bcb 10openembedded.git/recipes/ti/ti-msp430-chronos_1.05.00.00.bb: ti-msp430-chronos: bump PR Jul 08 12:31:09 03Koen Kooi  07org.openembedded.dev * r0e24e6eb83 10openembedded.git/recipes/opencv/openscenegraph_2.9.7.bb: openscenegraph: add 2.9.7 Jul 08 12:50:19 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r54f7168696 10openembedded.git/ (conf/machine/htcdream.conf recipes/tasks/task-x11.bb): Jul 08 12:50:19 htcdream.conf: use xf86-video-fbdev and xf86-video-tslib Jul 08 12:50:19 Signed-off-by: Denis 'GNUtoo' Carikli Jul 08 12:50:19 Acked-by: Martin Jansa Jul 08 12:50:19 Acked-by: Michael 'Mickey' Lauer Jul 08 13:30:22 huhu Jul 08 13:53:12 03Henning Heinold  07org.openembedded.dev * r22afd95e74 10openembedded.git/recipes/llvm/ (llvm-common-native.bb llvm-common.bb): llvm-common: use new staging system and switch to BBCLASSEXTENDED Jul 08 13:53:21 03Henning Heinold  07org.openembedded.dev * recb3cf8e4c 10openembedded.git/recipes/ (52 files in 6 dirs): openjdk: move over the recipes from jalimo Jul 08 14:03:46 hi jconnolly Jul 08 14:03:54 hi woglinde Jul 08 14:04:03 brb need to resize a partition Jul 08 14:04:07 builds are large ;D Jul 08 14:04:11 *g* Jul 08 14:05:55 03Koen Kooi  07org.openembedded.dev * rebe46b88fa 10openembedded.git/conf/distro/include/angstrom.inc: angstrom: include jalimo bits now that openjdk is in OE Jul 08 14:07:20 ho Jul 08 14:07:23 hi Jul 08 14:07:38 JaMa: Managed to get my chroot working Jul 08 14:08:59 now I'm using schroot and mount proc and devpts at chroot to it, everything working fine so far Jul 08 14:10:03 geo_gcju: good Jul 08 14:12:55 JaMa: so thanks for four hint with proc, from the errors I got I coulnd't spot the error Jul 08 14:59:09 florian: ping Jul 08 14:59:36 mickeyl: yes? Jul 08 14:59:49 hi mickeyl, florian Jul 08 15:00:10 hi all Jul 08 15:00:13 hi woglinde Jul 08 15:00:50 florian: i just received an reminder from Finanzamt Jul 08 15:01:00 florian: we need to submit the tax stuff by 30.7. Jul 08 15:02:48 hi pb_ and woglinde Jul 08 15:04:02 mickeyl: oops... ok, well... i guess that's possible. maybe not this week but i'll try to do it next week Jul 08 15:04:08 its not that much... Jul 08 15:04:34 ok. i will make sure the Tätigkeitsbericht is ready next week as well Jul 08 15:04:38 will send it to you as PDF then Jul 08 15:05:39 * florian needs a new brain and a fork of himself Jul 08 15:06:46 heh Jul 08 15:12:51 have we travel-budget this year? Jul 08 15:15:07 florian is the treasurer, however i think we don't have much left. Apparantly until now we have failed to acquire more donations Jul 08 15:16:18 right... Jul 08 15:16:38 * florian would love to work on oe marketing but currently there is not a single minute left Jul 08 15:21:16 we need to push our director of PR Jul 08 15:21:57 *g* Jul 08 15:22:19 heh, indeed. poor chap. Jul 08 15:29:53 wow.. do we have travel-budget? :) Jul 08 15:30:13 we had Jul 08 15:30:14 last year Jul 08 15:30:41 who is sponsored us? Jul 08 15:30:51 I forgot the company Jul 08 15:30:57 florian and mickeyl knows Jul 08 15:31:52 imho we should try to ask some funding from TI and HP at least :) Jul 08 15:35:05 hihi or eads Jul 08 15:35:38 or mentor, their products are expensive enough :-} Jul 08 15:36:28 or intel. we could threaten to steal RP and sell him to raise cash. Jul 08 15:37:44 pb_: :)) Jul 08 16:08:22 if someone still run OE source mirror then I have bad news for him: cleaning my DL_DIR Jul 08 16:23:28 what does OE use to determine if TMPDIR has moved? Jul 08 16:23:36 mine hasn't, but bitbake is certainly convinced that it has Jul 08 16:23:51 your paths are relative? Jul 08 16:24:01 and you called bitbake from difrent dir Jul 08 16:24:02 nope... Jul 08 16:25:26 that is also a negative Jul 08 16:34:46 Hello Jul 08 16:35:09 In OE when using bitbake it cross compile something Jul 08 16:35:11 but Jul 08 16:35:44 if i want to chnage some pararmaeters that can be passed to the configure, where i should look? Jul 08 17:10:59 kristianpaul: did you mean what to write into recipe to pass some parameters to configure? Jul 08 17:11:02 03Koen Kooi  07org.openembedded.dev * r479ea460e4 10openembedded.git/recipes/pam/ (libpam-base-files.bb libpam-base-files/pam.d/cups): libpam-base-files: cups installs its own pam file, so remove it from here Jul 08 17:11:13 03Koen Kooi  07org.openembedded.dev * r81467a049f 10openembedded.git/recipes/meta/meta-toolchain-qte.bb: meta-toolchain-qte: fix sh syntax Jul 08 17:36:30 * kergoth_ spends his morning fixing his laptop's install so he actually has enough space to do big builds... /eyeroll Jul 08 17:37:01 btw Jul 08 17:37:23 on powerdeveloper.org you can propose your project to get Efika MX hardware Jul 08 17:37:37 OE on Efika is already proposed by likewise :) Jul 08 17:37:45 proposed and even accepted Jul 08 17:40:58 tzanger: the old tmpdir var is your tmp dir in the saved_tmpdir file, you can see what is in there and compare with the actual dir Jul 08 17:41:58 khem, looked at the gcc problem I had, the problem is when compiling the lib stuff, (and in my case linux_atomic.c), issue is that a -I to the x686 sysroot dir is there causing the system to pick up the wrong include file. Jul 08 17:43:05 khem, I've seen this is from GMPINC which points to the sysroots/i686-linux/usr/include dir and hence causes this Jul 08 17:43:48 khem, I've analysed it and the problem is a -I${GMPINC) in build*/gcc/Makefile, when I removed that manually and restarted compilation all works fine. Jul 08 17:43:56 but not sure how to fix this properly Jul 08 17:45:03 configure says GMPINC and GMPLIBS are influential env vars, should they be overruled in the bb file Jul 08 17:45:43 (trying to tweak with gmp related ac vars didn't lead to success for me, but maybe I misunderstood things) Jul 08 18:04:26 03Koen Kooi  07org.openembedded.dev * r6e7a6d5036 10openembedded.git/recipes/shadow/ (files/securetty shadow.inc): shadow: add more ARM serial ports Jul 08 18:29:55 03Koen Kooi  07org.openembedded.dev * r423aff2324 10openembedded.git/contrib/angstrom/sort.sh: angstrom feed sorter: add support for omap4430-panda Jul 08 18:31:36 03Frans Meulenbroeks  07org.openembedded.dev * r302dbb63be 10openembedded.git/recipes/linux/linux-nios2-2.6.34/defconfig: Jul 08 18:31:36 linux-nios2-2.6.34: moved defconfig to the neek subdir Jul 08 18:31:36 so other machines can have their defconfig adjacent to this Jul 08 18:31:36 Signed-off-by: Frans Meulenbroeks Jul 08 18:34:58 eFfeM: how is the include file wrong, anyhow? Jul 08 18:35:35 Or is gcc using the GMPINC incs when it doesn't need to? Jul 08 18:43:35 Tartarus: this is what I replied to one of earlier email GMPCINC is hostinclude Jul 08 18:43:44 and gcc passes it for target libs too Jul 08 18:43:58 exactly Jul 08 18:44:18 Anybody know why it is that when I bitbake omap3-console-image, it compiles 2.6.33? I'm reading on the internets that the kernel that should be compiled will be the one whose bitbake file specifies DEFAULT_PREFERENCE_overo? linux-omap_2.6.29.bb is the file that contains that flag though? Jul 08 18:44:23 but I dont know the reason yet Jul 08 18:44:30 and my target asm/unistd.h differs from the host one, causing compilation of the gcc lib to fail Jul 08 18:44:49 bkinman: whats your MACHINE Jul 08 18:45:28 "overo", as specified by auto.conf Jul 08 18:47:25 One more reason I wish it was easier to just build gmp/mpfr/etc into gcc, heh Jul 08 18:47:52 khem, the GMPINC comes from gcc/Makefile.in; I think configure gives it the wrong value, but it is overridable with configure env vars Jul 08 18:48:17 and ofc I can just patch out the -I${GMPINC} but not sure if that is good either Jul 08 18:49:58 re Jul 08 18:51:55 afk for a while Jul 08 18:51:56 eFfeM: gcc itself uses this include Jul 08 18:52:00 ah ok Jul 08 18:52:02 still here Jul 08 18:52:28 re Jul 08 18:52:50 should we have a gmp-intermediate or so that compiles and builds gmp with gcc-intermediate ? Jul 08 18:53:07 effem do we need it? Jul 08 18:53:23 bkinman: it seems overo machine uses linux-omap-psp Jul 08 18:53:30 bkinman: not linux-omap Jul 08 18:53:41 blindvt: around ? Jul 08 18:53:44 linux-omap-psp has more features Jul 08 18:53:46 eFfeM: The problem is why is gcc using host stuff when doing target stuff Jul 08 18:54:03 Tartarus: its a gcc bug/need I dont know Jul 08 18:54:09 khem: Interesting! Where did you find that? Jul 08 18:54:13 Tartarus: I think what we see is that configure finds a gmp and uses that Jul 08 18:54:14 Tartarus: there might be a reason why gcc does that Jul 08 18:54:23 khem: Right Jul 08 18:54:24 or may be just an oversight Jul 08 18:54:30 But that seems to be a potential for a problem Jul 08 18:54:32 I dont know yet Jul 08 18:54:36 I havent looked into it Jul 08 18:54:39 re pb Jul 08 18:54:50 ie 'normally' you would this this problem, for nios2 Jul 08 18:54:56 Since a normal gmpinc will be /usr/include Jul 08 18:55:00 yeah Jul 08 18:55:04 with host asm/unistd.h and so on Jul 08 18:55:12 or out host sysroot Jul 08 18:55:22 our Jul 08 18:55:30 Or now that our host sysroot has it, yes :) Jul 08 18:56:27 well, the issue is triggered by the introduction of linux-libc-headers-native which puts asm/unistd.h in x686/sysroot Jul 08 18:56:41 hm Jul 08 18:56:47 and the nios2 one differs from the native one as nios2 has introduced a new system call Jul 08 18:57:23 if I do ASSUME_PROVIDED_linux-libc-headers-native everything works fine Jul 08 18:57:50 because there is no faulty unistd.h in the GMPINC dir Jul 08 18:57:56 the old kernel-header problem Jul 08 18:58:46 yes, I already feared something nasty like this would happen with the introduction of linux-libc-headers-native; in this case I'd rather been proven wrong ... Jul 08 18:59:52 khem: I dont see linux-omap-psp bitbake file for kernel version 2.6.33? Jul 08 18:59:54 anyway my workaround is to have ASSUME_PROVIDED += "linux-libc-headers-native" Jul 08 18:59:54 in include/nios2.inc; despite people disliking this line in it Jul 08 19:00:36 psp is 2.6.32 Jul 08 19:00:42 yes Jul 08 19:00:50 open for suggestions, but temporarily moving out of this hot room... Jul 08 19:01:48 hi, how do I oewarn etc...in python? Jul 08 19:03:07 gnutoo grep? Jul 08 19:03:36 bkinman: whatever it is at is what you can use Jul 08 19:03:47 if you need 2.6.33 then you have to port Jul 08 19:04:10 bb.note Jul 08 19:04:13 found it Jul 08 19:05:01 khem uh thats not task for some one alone Jul 08 19:07:40 woglinde: heh Jul 08 19:16:44 eFfeM: But the question is, how does the nios2 stuff build anywhere at all? Jul 08 19:16:57 That might be an interesting gcc build log to look over Jul 08 19:17:48 hi JaMa Jul 08 19:17:53 I found the reason Jul 08 19:18:22 enabled = bb.data.getVar("ENABLE_BINARY_LOCALE_GENERATION", d, 1) Jul 08 19:18:29 khem, pong Jul 08 19:18:31 if not enabled: Jul 08 19:18:35 return Jul 08 19:18:39 enabled is a string Jul 08 19:18:44 so enabled is true Jul 08 19:19:08 even if it's 0 Jul 08 19:19:34 GNUtoo|laptop: Ick, silly python :( Jul 08 19:19:38 lol Jul 08 19:19:47 khem, i built qemuppc, qemumips, qemux86, qemuarm and my own qemux86_64 successfully after the nptl overhaul, yes Jul 08 19:19:48 Found a case like that in the strict hash stuff the other day Jul 08 19:19:53 khem, what ICE do you get? Jul 08 19:21:10 khem, i shall look into nobx now. What do you mean by "I think you do not have a function to disable already enabled feature." ? please explain? Jul 08 19:22:31 khem, the 'rem' translate into a sed /SOMETHING/d stanza, if you mean that? Jul 08 19:22:38 GNUtoo|laptop: i'm newbie in this oe stuff. what should i do to complete leviathan receipe ? Jul 08 19:22:59 zenob, go in the source dir Jul 08 19:23:06 go in the git dir Jul 08 19:23:29 note the name of leviathan recipe's source dir Jul 08 19:23:36 move or rm it Jul 08 19:23:40 and check it out manually Jul 08 19:23:42 khem, thinking about it, i'd much prefer to explicitly ask for BX, perhaps in some generic arm include? Jul 08 19:23:56 and rename it to the name you noted Jul 08 19:24:19 zenob, btw could you provide the exact error Jul 08 19:24:20 ? Jul 08 19:26:58 ping JaMa Jul 08 19:27:17 khem: http://patchwork.openembedded.org/patch/2235/ is that an acked-by? :) Jul 08 19:27:46 GNUtoo|laptop: http://pastebin.com/Mj0SVMQg Jul 08 19:28:05 hi mickeyl, about the git issue, did it apply to all gitorious repos? Jul 08 19:28:50 zenob, that doesn't seem a fetch error anymore... Jul 08 19:30:31 oh, sorry, i've tried to download manally *.tar.gz file, renamed it, and copied to dowload dir Jul 08 19:31:09 zenob, strange Jul 08 19:31:15 scripts/unifdef.c is present for me Jul 08 19:31:27 zenob, I meant the git Jul 08 19:31:31 clone it Jul 08 19:31:46 tar.gz is only the current source Jul 08 19:31:57 it will be changed at each source revision change Jul 08 19:32:23 ok, i'll try Jul 08 19:32:54 like in that dir Jul 08 19:32:59 /home/gnutoo/embedded/oe/sources/git/gitorious.org.htc-msm-2-6-32.leviathan-incoming.git Jul 08 19:33:05 maybe instead of cloning Jul 08 19:33:10 just run the command that failed Jul 08 19:51:21 i hate computers. Jul 08 19:51:38 kergoth me too Jul 08 19:52:04 And they hate us too Jul 08 20:03:22 Tartarus: just saw your q: don't really understand it "how does the nios2 stuff build anywhere at all?" Jul 08 20:03:24 so it seems Jul 08 20:03:53 Tartarus: it builds, see my tinderbox log of 2010-07-07 19:08:31 Jul 08 20:04:59 sorry wrong tinderbox ref Jul 08 20:05:55 somehow my X just exploded on my work laptop, stuck in a crappy resolution, etc now Jul 08 20:05:57 * kergoth_ scratches head Jul 08 20:06:00 oh Jul 08 20:06:07 maybe i'm so low on disk space i can't login :) Jul 08 20:06:41 system-rescue-cd Jul 08 20:07:17 hm, somehow I cannot find a proper tinderbox log, will try to make a new one, also tinderbox package search seem to have some issues Jul 08 20:08:41 Tartarus: but no idea why you think the nios2 stuff would not build (choose MACHINE neek and bitbake gcc-cross and you're on Jul 08 20:09:20 hi Jul 08 20:12:16 Tartarus: it is building now, http://tinderbox.openembedded.net/builds/78028/ will be back in a while Jul 08 20:19:19 eFfeM: I mean outside of OE Jul 08 20:19:23 I have no idea how it could build Jul 08 20:19:34 Since /usr/include will be used for GMPINC Jul 08 20:19:42 And that will have the bad asm/unistd.h Jul 08 20:24:24 Tartarus: originally I got a binary Jul 08 20:25:48 and this is the published way: http://www.nioswiki.com/Build_the_gcc4_toolchain Jul 08 20:26:03 never tried that it is based upon somw windriver/csl script or so Jul 08 20:26:32 the non-mmu version builds upon the uclinux build env Jul 08 20:27:43 don't have those sources and scripts handy Jul 08 20:32:42 is there some sort of emulator for nios2, btw? Jul 08 20:34:16 eFfeM, k Jul 08 20:34:21 we have nios2 support in uClibc but i never attempted to test it nor keep it alive, admittedly. Some sort of emulator would help tremendously to keep that port alive, in my pov Jul 08 20:35:26 khem: Actually, I AM using 2.6.33, it seems that bitbake omap3-console-image IS compiling a 2.6.33 kernel? Jul 08 20:40:02 khem: i really wish, for the life of me, that i could figure out which files were responsible for selecting this kernel. Jul 08 20:41:08 03Tom Rini  07org.openembedded.dev * rbcb57eab74 10openembedded.git/recipes/ipkg-utils/ipkg-utils-native_1.6+cvs20050404.bb: Jul 08 20:41:08 ipkg-utils-native: Switch to non-legacy staging Jul 08 20:41:08 Signed-off-by: Tom Rini Jul 08 20:48:06 03Tom Rini  07org.openembedded.dev * r51ee5b59b4 10openembedded.git/recipes/ipkg-utils/ipkg-utils-native_1.6+cvs20050404.bb: Jul 08 20:48:06 ipkg-utils-native: Switch for real... Jul 08 20:48:06 Signed-off-by: Tom Rini Jul 08 20:48:14 ~lart self Jul 08 20:48:15 * ibot takes out a cattle prod and gives self a good jolt Jul 08 20:48:32 ouch Jul 08 21:03:02 * Tartarus tests python-native switch, properly now Jul 08 21:05:21 hi Jul 08 21:05:23 I've Unknown fetch Error Jul 08 21:05:30 NOTE: Task failed: Unknown fetch Error: [Errno 2] No such file or directory: Jul 08 21:05:33 and the tarball Jul 08 21:05:38 I've messed up my source dir Jul 08 21:05:41 what can I do Jul 08 21:05:47 I don't want to rebuild from scratch Jul 08 21:05:52 I'm debugging a git issue Jul 08 21:06:16 turn on packaged-staging so ditching TMPDIR hurts much less, next time :) Jul 08 21:06:22 bitbake -c clean whatever is failing Jul 08 21:06:27 I've packaging staging Jul 08 21:06:29 mv $DL_DIR $DL_DIR.old Jul 08 21:06:33 clean doesn't seem to fix it Jul 08 21:06:33 ok Jul 08 21:06:42 Well, save off your pstaging dir Jul 08 21:06:45 and wipe out the rest Jul 08 21:06:50 or put pstaging outside of tmpdir :) Jul 08 21:07:10 basically I want to force it to refetch the git repo Jul 08 21:07:13 $ grep DISTRO conf/local.conf Jul 08 21:07:13 TMPDIR = "/home/trini/work/OE-upstream/tmp.${LIBC}.${DISTRO}" Jul 08 21:07:13 PSTAGE_DIR = "/home/trini/work/OE-upstream/pstage.${LIBC}.${DISTRO}" Jul 08 21:07:15 is what I do, heh Jul 08 21:07:18 ok Jul 08 21:07:32 so nuking TMPDIR is easy, don't have to think about it Jul 08 21:07:44 I also use autorev Jul 08 21:07:51 so I need to keep the local count Jul 08 21:08:13 moving source didn't solve it Jul 08 21:08:21 03Tom Rini  07org.openembedded.dev * ra753e68001 10openembedded.git/recipes/python/ (python-native_2.6.1.bb python-native_2.6.4.bb): Jul 08 21:08:21 python-native: Switch to non-legacy staging Jul 08 21:08:21 Signed-off-by: Tom Rini Jul 08 21:08:27 it doesn't seem the stamps either Jul 08 21:08:56 Well, I don't quite know what your problem is from that one line :) Jul 08 21:09:01 ok Jul 08 21:09:07 basically when you fetch Jul 08 21:09:14 from git you make a tarball of it Jul 08 21:09:20 Yeah Jul 08 21:09:20 it doesn't find that tarball Jul 08 21:09:31 So something is off in that part of bitbake Jul 08 21:09:42 I removed the tarball Jul 08 21:09:52 basically here's the issue Jul 08 21:10:04 I've an htcdream Jul 08 21:10:16 the android kernel is non-standard Jul 08 21:10:23 so we use a gitorious repo for our kenrel Jul 08 21:10:32 that gitorious repo fails to be fetched Jul 08 21:10:35 I want to debug that Jul 08 21:11:33 so I removed the tarball Jul 08 21:11:38 and the stamps Jul 08 21:11:44 for forcing it to refetch it Jul 08 21:11:52 and it wanted the tarball again Jul 08 21:14:22 ah Jul 08 21:14:27 maybe in stamps in staging Jul 08 21:14:48 hmmm no Jul 08 21:15:21 does someone knows? Jul 08 21:17:50 GNUtoo, i don't know if it helps, but think there's something with gitorius, i'm still getting fatal: early EOF during git clone git://gitorious.org/htc-msm-2-6-32/leviathan-incoming.git Jul 08 21:18:18 hmm Jul 08 21:18:26 zenob: its a bug on gitorious Jul 08 21:18:47 so what's the solution ? Jul 08 21:19:22 zenob, I'm trying to debug that now Jul 08 21:19:30 but I messed up my tmpdir Jul 08 21:22:17 When i do bitbake linux, it builds kernel 2.6.35, anybody know why? Jul 08 21:24:54 bkinman: no? Jul 08 21:25:09 bkinman: which device? Jul 08 21:25:16 I'll try -DDD Jul 08 21:25:20 leviathan: MACHINE is overo. Jul 08 21:25:25 bkinman: ohh Jul 08 21:25:26 hmm Jul 08 21:25:49 bkinman: perhaps they did not set a standart kernel Jul 08 21:25:51 or Jul 08 21:26:00 over is in official kernel Jul 08 21:26:01 :-) Jul 08 21:26:36 leviathan: this is killing me, i cant for the life of me figure out how it determines which kernel to build, i've been trying to figure this out for days. Jul 08 21:27:05 uhm Jul 08 21:27:09 try it in local.conf Jul 08 21:27:11 was bitbke's cache Jul 08 21:27:18 GNUtoo|laptop: ok Jul 08 21:27:36 thanks a lot for the help Jul 08 21:27:37 leviathan: local.conf does not specify a kernel as far as i can tell. Jul 08 21:27:56 bkinman: in local.conf we can set the definitions which virtual package should be used Jul 08 21:28:01 e.g. Jul 08 21:28:25 bkinman: PREFERRED_PROVIDER_virtual/kernel = "linux-leviathan" Jul 08 21:28:27 or so Jul 08 21:29:01 leviathan: yeah, there are no directives like that in my local.conf. Just DISTRO and DISTRO_TYPE Jul 08 21:29:31 bkinman: because its the last possible solution to fix your problem Jul 08 21:29:34 but it will do it Jul 08 21:29:35 ;-) Jul 08 21:31:12 leviathan: ah, you are making a suggestion, gotcha. Jul 08 21:31:37 yes Jul 08 21:31:39 :-) Jul 08 21:36:46 * blindvt curses Jul 08 21:47:02 Hrmm. bitbake linux seems to produce a kernel that doesnt work... However, bitbake omap3-console-image produced a kernel image that does seem to work, does that make sense to anybody? Jul 08 21:47:32 "linux" is one *specific* kernel recipe Jul 08 21:47:37 not necessarily the one your machine needs Jul 08 21:47:41 you want to bitbake virtual/kernel Jul 08 21:52:31 dammed. what was the http_proxy and ftp_proxy settings again that were whitelisted? Jul 08 21:53:04 what a pita. Looking for my notes Jul 08 21:55:09 oh wow Jul 08 21:55:24 -cclean linux blows away the entire source and a subsequent bitbake linux rebuilds it from source Jul 08 21:55:32 I guess -cclean isn't necessary if I change a defconfig :-) Jul 08 21:59:16 when will the 2.6.34 for akita be released? Jul 08 21:59:50 buZz: you can use it today :) Jul 08 21:59:51 we're at 2.6.34.1 by now Jul 08 21:59:58 woa Jul 08 22:00:47 can i just let bitbake pick a different kernel for my machine? Jul 08 22:00:55 ant have testing it sometimes Jul 08 22:01:25 buZz: set DEFAULT_PREFERENCE_akita="1" in recipe with appropriate kernel Jul 08 22:01:45 * Jay7 still not tested latest kernels on akita and tosa :( Jul 08 22:02:15 hmm Jul 08 22:02:48 none of the AT91 linux defconfigs have CONFIG_EARLY_AT91_USART set... I would have thought that they would have wanted that Jul 08 22:03:14 according to someone in my email, the 2.6.34 will allow me to run newer udev Jul 08 22:03:31 and thus, get gudev, and be able to continue my gnome-image based on xorg Jul 08 22:03:32 :P Jul 08 22:03:33 buZz: yes, udev-151 Jul 08 22:04:15 but then revert Stanislav Brabec's commit that have glued udev-14x for zauruses Jul 08 22:04:53 i still have to get used to dealing with stuff like that on git .. Jul 08 22:05:59 distro/include/angstrom-2008-preferred-versions.inc:PREFERRED_VERSION_udev_akita = "141" Jul 08 22:06:26 bkinman: which branch of oe are you using ? Jul 08 22:06:26 git log distro/include/angstrom-2008-preferred-versions.inc Jul 08 22:06:35 then git revert Jul 08 22:06:51 but git checkout -b your_own_branch before Jul 08 22:07:04 btw Jul 08 22:07:18 do we have renamed org.oe.dev to master already? Jul 08 22:07:25 khem: let me check. Jul 08 22:08:15 khem: the overo branch Jul 08 22:14:34 I've the error again Jul 08 22:14:41 how do I force to reparse everything? Jul 08 22:14:54 rm cache? Jul 08 22:15:15 ok Jul 08 22:15:28 apart that I shouldn't rm the local count Jul 08 22:15:35 thanks Jul 08 22:16:08 may be there is some bitbake option for this but I don't know.. Jul 08 22:16:22 ok Jul 08 22:17:17 sigh Jul 08 22:17:22 I will never fix the issue Jul 08 22:18:10 can't you just build single receipe ? Jul 08 22:18:27 I can't fetch Jul 08 22:18:37 NOTE: Task failed: Unknown fetch Error: [Errno 2] No such file or directory: '/home/gnutoo/embedded/oe/sources/git_gitorious.org.htc-msm-2-6-32.leviathan-incoming.git_3a5d683e5893ae75ace4bdfef4f5605555d40740.tar.gz Jul 08 22:18:50 hmm Jul 08 22:18:56 yes, but what may cause this ? Jul 08 22:18:57 got into the sources dir Jul 08 22:18:59 and without fetching I can't solve that gitorious issue Jul 08 22:19:08 leviathan, ? Jul 08 22:19:11 and clone manually Jul 08 22:19:14 :-( Jul 08 22:19:17 leviathan, no Jul 08 22:19:30 gitorious is broken somehow Jul 08 22:19:37 in combination with bitbake Jul 08 22:19:39 leviathan, I *WANT* to fix gitorious issue Jul 08 22:19:39 dunno why Jul 08 22:19:40 i've a foobar http fetcher that doesn't respect my http_proxy nor my ftp_proxy. sigh Jul 08 22:19:44 ahh Jul 08 22:19:47 GNUtoo|laptop: hmm Jul 08 22:20:00 and in order to do that I need to force it to fetch from gitorious Jul 08 22:20:03 instead of locally Jul 08 22:20:11 this is so extremely unproductive Jul 08 22:20:14 GNUtoo|laptop: it might be an issue with the commandline options Jul 08 22:20:17 * blindvt calls it a day Jul 08 22:20:26 g'night all Jul 08 22:20:31 so does the calender! Jul 08 22:20:32 leviathan, I don't think so Jul 08 22:20:33 blindvt: gn8 Jul 08 22:20:58 I think it's a variable or something like that Jul 08 22:21:02 but bitbake is complex Jul 08 22:21:22 I didin't even succeed at printing what git did in fetchcmd Jul 08 22:26:20 Any other kids in here doing kernel hacking for the overo? Jul 08 22:29:26 sigh Jul 08 22:30:24 oh.. Jul 08 22:30:53 * Jay7 is trying to sync with psplash mainstream and with patches from psplash recipe Jul 08 22:31:05 grrr Jul 08 22:32:49 GNUtoo|laptop: override git call with own script Jul 08 22:33:03 that will store command line somewhere Jul 08 22:33:14 ok but Jul 08 22:33:19 I can't even make git fetch Jul 08 22:33:30 I'm struggeling with that error: Jul 08 22:33:44 NOTE: Task failed: Unknown fetch Error: [Errno 2] No such file or directory: '/home/gnutoo/embedded/oe/sources/git_gitorious.org.htc-msm-2-6-32.leviathan-incoming.git_3a5d683e5893ae75ace4bdfef4f5605555d40740.tar.gz' Jul 08 22:34:00 I'm trying pass on a lot of bitbake function Jul 08 22:34:08 and I've still the same error Jul 08 22:34:30 hmm Jul 08 22:34:33 cbrake_away, you here? Jul 08 22:35:02 GNUtoo|laptop: I'm writing here presentation for my MEG presentation Jul 08 22:35:07 ok Jul 08 22:35:10 and why its not a perpetuum mobile Jul 08 22:35:12 so Jul 08 22:35:14 but manual clone also doesn't work and hangs at 15% Jul 08 22:35:14 and btw Jul 08 22:35:19 I should sleep now Jul 08 22:35:23 ok Jul 08 22:35:35 I'll look into it Jul 08 22:35:50 thanks Jul 08 22:36:02 btw, GNUtoo|laptop because a supra conductor is a perpetuum mobile Jul 08 22:36:09 and I'm using quantum mechanics Jul 08 22:36:27 my construction could really be the first real type 2 perpetuum mobile xD Jul 08 22:36:33 zenob, ok I'm really trying hard to fix it Jul 08 22:36:37 but anyway Jul 08 22:36:38 ^^ Jul 08 22:36:49 GNUtoo|laptop: I'll look into it Jul 08 22:36:53 promised Jul 08 22:36:56 leviathan, into what? Jul 08 22:36:58 bitbake? Jul 08 22:37:02 yes Jul 08 22:37:07 I'll try to fix it Jul 08 22:37:11 ok Jul 08 22:37:13 if you do not manage to do so Jul 08 22:37:14 but it seem hard Jul 08 22:37:17 the next days Jul 08 22:37:19 yes Jul 08 22:37:39 I've an idea Jul 08 22:37:43 rename the recipe Jul 08 22:37:45 locally Jul 08 22:38:40 GNUtoo|laptop: gn8 Jul 08 22:38:43 ok Jul 08 22:38:50 same error Jul 08 22:38:53 ok Jul 08 22:38:54 it was a bad idea Jul 08 22:38:58 because the git is the same Jul 08 22:39:03 yes Jul 08 22:39:07 its the git url Jul 08 22:39:14 maybe I could change that Jul 08 22:39:19 to http Jul 08 22:39:31 GNUtoo|laptop: look how other gitorious using recipes have done it Jul 08 22:39:33 :-) Jul 08 22:39:38 http is a good idea Jul 08 22:39:40 there are not much Jul 08 22:40:04 GNUtoo|laptop: I guess its because of broken git server Jul 08 22:40:06 try http Jul 08 22:40:15 no Jul 08 22:40:22 I guess it's a variable Jul 08 22:41:16 i tend to agree. using the very same command in a user shell always seems to work for me Jul 08 22:41:33 so it's probably something that OE blacklists Jul 08 22:41:39 but that git needs for this server Jul 08 22:42:53 I suspect errors in psplash svn.. Jul 08 22:43:05 that rgb ordering looking wrong Jul 08 22:43:22 ok Jul 08 22:44:04 LANG="en_US.utf8" ? Jul 08 22:44:08 kergoth, somehow f84f382f340d6db15b9e5afb8c7c93969249a958 from that dell guy completely broke proxy handling in the wget fetcher for me. I have both http_proxy and ftp_proxy set in my environment proper (plus the odd upper-case variants in my local.conf) and bb's wget fetcher hangs trying to contact any target directly instread of via the given proxy Jul 08 22:44:34 because git needs LC_ALL LANG or something like that Jul 08 22:44:35 s/instread/instread/;# even Jul 08 22:45:30 fingers playing tricks on me. neither insns nor instruction nor read. but instead, really Jul 08 22:46:27 sleep. now. really Jul 08 22:46:55 mickeyl, how can you fetch if it says: Jul 08 22:47:04 git and bugzilla offline for 30m to 1hr...for hardware replacement Jul 08 22:47:09 NOTE: Task failed: Unknown fetch Error: [Errno 2] No such file or directory: '.../git_gitorious.org.htc-msm-2-6-32.leviathan-incoming.git_3a5d683e5893ae75ace4bdfef4f5605555d40740.tar.gz' **** ENDING LOGGING AT Thu Jul 08 22:47:23 2010 **** BEGIN LOGGING AT Fri Jul 09 00:18:14 2010 Jul 09 00:18:18 oh nevermind it works Jul 09 00:18:47 that answers that Jul 09 00:19:01 can somebody check bugzilla please? Jul 09 00:19:59 ka6sox-work: seems to be working Jul 09 00:20:16 ka6sox-work: http://bugzilla.openembedded.org/ works for me Jul 09 00:20:31 ah git is working too Jul 09 00:20:33 thats nice Jul 09 00:20:49 thanks Jul 09 00:22:26 ka6sox-work: so all is well now ? Jul 09 00:22:28 it's back, yes :) Jul 09 00:22:40 bot is hung up perhaps Jul 09 00:22:47 03Tom Rini  07org.openembedded.dev * r8ea3456309 10openembedded.git/recipes/stage-manager/ (files/stage-manager-ipkg stagemanager-native_0.0.1.bb): (log message trimmed) Jul 09 00:22:47 stagemanager-native: Expand our gunzip/tar Jul 09 00:22:47 As we've gone over before, gunzip/tar can have a broken pipe, and Jul 09 00:22:47 this is allowed. In python code, we have things setup to ignore the Jul 09 00:22:47 problem. stage-manager-ipkg is written in shell however and we can Jul 09 00:22:51 ah there we go Jul 09 00:22:53 heh Jul 09 00:32:03 03Bernhard Reutner-Fischer  07org.openembedded.dev * ra0be67df8b 10openembedded.git/ (57 files in 49 dirs): Jul 09 00:32:03 uclibc: add 0.9.31 Jul 09 00:32:03 Signed-off-by: Bernhard Reutner-Fischer Jul 09 00:32:03 Signed-off-by: Khem Raj Jul 09 00:32:05 03Bernhard Reutner-Fischer  07org.openembedded.dev * r7bb8608aef 10openembedded.git/ (15 files in 2 dirs): Jul 09 00:32:05 uClibc: redo configuration Jul 09 00:32:05 Signed-off-by: Bernhard Reutner-Fischer Jul 09 00:32:22 attr_2.4.44.bb: Specify PLATFORM otherwise it uses uname. Jul 09 00:32:22 * Using uname to get OS name may not work on non linux build Jul 09 00:32:22 systems. Jul 09 00:32:22 * Cater for largefile feature which is knob'able in uclibc Jul 09 00:32:22 Signed-off-by: Khem Raj Jul 09 00:32:30 Signed-off-by: Khem Raj Jul 09 00:32:30 03Khem Raj  07org.openembedded.dev * rb568cfe4ae 10openembedded.git/recipes/uclibc/ (uclibc-config.inc uclibc.inc uclibc_git.bb): Jul 09 00:32:31 uclibc: Add thumb-interwork as local distro feature to enable bx Jul 09 00:32:31 * Bump SRC_REV for uclibc_git Jul 09 00:32:32 * Pass SSP_ALL_CFLAGS=${TARGET_LINK_HASH_STYLE} otherwise Jul 09 00:32:32 utils dont get the hash_style right and QA complains. Jul 09 00:32:33 Signed-off-by: Khem Raj Jul 09 00:32:59 03Khem Raj  07org.openembedded.dev * r4efe2e328a 10openembedded.git/conf/distro/minimal-uclibc.conf: Jul 09 00:32:59 minimal-uclibc.conf: Add ipv6, ipv4, largefile to DISTRO_FEATURES. Jul 09 00:32:59 Signed-off-by: Khem Raj Jul 09 00:43:21 Tartarus: hmmm Jul 09 00:43:31 Tartarus: see http://pastebin.com/HqBSNrKx Jul 09 00:43:50 Tartarus: bitbake -b /home/kraj/work/oe/openembedded/recipes/ipkg-utils/ipkg-utils-native_1.6+cvs20050404.bb Jul 09 00:43:57 and you get it Jul 09 00:44:10 its one of your patches i guess Jul 09 00:44:13 Add an install -d ${D}${bindir} Jul 09 00:44:16 and see if that does it Jul 09 00:44:18 if so, push :L) Jul 09 00:44:19 :) Jul 09 00:44:25 yeah thats the fix Jul 09 00:44:28 but need to run now Jul 09 00:44:33 pick up kids before 6 Jul 09 00:44:34 k Jul 09 00:45:46 03Tom Rini  07org.openembedded.dev * r2a6e03c72e 10openembedded.git/recipes/ipkg-utils/ipkg-utils-native_1.6+cvs20050404.bb: Jul 09 00:45:46 ipkg-utils-native: Ensure ${D}${bindir} exists Jul 09 00:45:46 Signed-off-by: Tom Rini Jul 09 00:56:51 googlecode is giveing 502 bad gateway again... Jul 09 02:32:23 is there a magic command to tell bitbake to unpack the source, apply the patches and then stop so I can hack on the code a bit? Jul 09 02:32:36 or do I simply make it, making sure that the rm_work isn't inherited? Jul 09 02:33:07 bitbake -c patch whatever Jul 09 02:33:32 you can stop at any point by specifying the task you want to be executed Jul 09 02:33:40 fetch, unpack, patch, configure, etc Jul 09 02:34:45 ooh Jul 09 02:34:51 bitbake -c patch foo ; bitbake -c devshell foo Jul 09 02:34:55 Might be quite helpful to you Jul 09 02:36:43 devshell already patches Jul 09 02:37:25 which is a sight annoyance, actually, since it means you can't use -c devshell directly to fix a patch application failure, but its more common to use it for other purposes, and the PATCHRESOLVE thing works for the patch fixing anyway.. Jul 09 02:38:10 Tartarus: fyi, my laptop basically decided to spontaneously shit itself. reinstalling ubuntu, cause I just can't be bothered with spending hours trying to track down the cause Jul 09 02:38:13 heh Jul 09 02:38:17 (work laptop) Jul 09 02:39:15 k Jul 09 02:39:36 spent today banging my head into the wall trying to get it to behave :( Jul 09 02:42:56 devshell eh? Jul 09 02:43:00 bitbake -k world sucked up the entire contents of my drive :) was going to kill the windows partition so i could continue, and then things just went all explodey Jul 09 02:43:11 tzanger: it spawns an xterm/screen/etc inside the unpacked, patched source tree Jul 09 02:43:12 hmm, it's not building anything now after I patched Jul 09 02:43:41 it's just saying xyz tasks, of which xyz didn't need to run Jul 09 02:43:59 so you built it fully previously Jul 09 02:44:04 it thinks it doesn't need to rerun the tasks Jul 09 02:44:12 yes this is true Jul 09 02:44:14 bitbake doens't know about the contents of the source tree, really Jul 09 02:44:15 -cclean would blow it all away Jul 09 02:44:21 it has "stamp" files that indicate a task is complete Jul 09 02:44:23 what you can do is use -f Jul 09 02:44:28 bitbake -c configure -f whatever Jul 09 02:44:32 will force it to run configure Jul 09 02:44:34 then bitbake whatver Jul 09 02:44:37 will rerun the rest Jul 09 02:44:41 since it'll see configure changed Jul 09 02:44:52 more steps than would be ideal, i know, but it should work Jul 09 02:44:59 so -ccompile -f foo? Jul 09 02:45:12 assuming you don't need it to re-run the configure step first, yep Jul 09 02:45:14 exactly Jul 09 02:45:27 but don't forget the bitbake foo afterward, cause that'll *only* run compile Jul 09 02:45:31 depends on what you need to do Jul 09 02:45:33 yeah I won't have to do that Jul 09 02:45:35 yeah I read that Jul 09 02:45:36 * kergoth nods Jul 09 02:45:44 okay then, have fun :) Jul 09 02:45:59 it'd really be nice to do a better job of handling the active development cases Jul 09 02:46:04 there's just too much overhead Jul 09 02:46:30 still, better than nothing Jul 09 02:46:43 I certainly appreciate the help :-) Jul 09 02:46:51 np Jul 09 02:46:59 hmm Jul 09 02:47:03 you know what'd be cool Jul 09 02:47:04 there's a lot ot keep in mind as you start out with oe Jul 09 02:47:08 make -c undersatnd multiple tasks Jul 09 02:47:16 bitbake -c configure,build -f foo Jul 09 02:47:27 * kergoth ponders Jul 09 02:47:57 odd Jul 09 02:48:03 -ccompile -f foo failed Jul 09 02:48:06 are you just trying to get the thing to crosscompile, or are you actually coding on the app? Jul 09 02:48:09 but not because of what I did Jul 09 02:48:14 hmm, how so? Jul 09 02:48:19 what's the error, i mean Jul 09 02:48:25 it failed looking for the cross compiler's libraries Jul 09 02:48:42 strange. Jul 09 02:48:46 very Jul 09 02:51:08 | arm-angstrom-linux-gnueabi-gcc: crt0_gnu.o: No such file or directory Jul 09 02:51:09 stuff like that Jul 09 02:53:54 Hmm, I just removed all the .o files and bitbake'd foo (at91bootstrap) and it seemed ok now **** ENDING LOGGING AT Fri Jul 09 02:59:57 2010