**** BEGIN LOGGING AT Tue Jun 23 02:59:57 2009 Jun 23 05:46:48 anyone have an idea of how to get a library to use GNU_HASH Jun 23 05:47:02 ive tried putting -hash-style=gnu in various places Jun 23 05:47:10 the package is 'slang' btw Jun 23 05:47:28 from objdump -p: HASH 0x000000b4 Jun 23 05:55:33 m4t: you need to pass it in LDFLAGS Jun 23 05:57:51 i added it to local.conf Jun 23 05:57:56 under ldflags Jun 23 05:58:19 let's see Jun 23 06:04:35 i think it already has --hash-style=gnu when configure tests the compiler Jun 23 06:04:44 maybe the Makefile is broken? Jun 23 06:06:44 morning Jun 23 06:08:14 good morning everybody Jun 23 06:08:32 i have a quesion about local overlay, anybody awake? :) Jun 23 06:09:27 Goddchen: I am yet to sleep Jun 23 06:09:34 ;) Jun 23 06:09:58 i need to overwrite INITSCRIPT_PARAMS in recipes/udev/udev_141.bb Jun 23 06:10:24 so i just copied that file to my local overlay local/recipes/udev/udev_141.bb and changed the setting Jun 23 06:10:33 it's getting parsed actually, but not used Jun 23 06:10:54 any idea what could be the problem? i should local overlay files are preferred to the original files Jun 23 06:11:48 according to this Jun 23 06:11:58 http://www.kernel-labs.org/files/openembedded-guide/openembedded-guide.html#toc6.4 Jun 23 06:12:41 03Koen Kooi  07org.openembedded.dev * r35a551aee9 10openembedded.git/ (conf/checksums.ini recipes/tango/tango-icon-theme_0.8.90.bb): tango-icon-theme: add 0.8.90 Jun 23 06:12:44 03Koen Kooi  07org.openembedded.dev * r955a41e2df 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev Jun 23 06:13:53 khem: configure fails when i have LDFLAGS="-hash-style=gnu" or "--hash-style=gnu" Jun 23 06:14:13 but i already saw --hash-style=gnu being appended to the cc line Jun 23 06:15:30 ...tmp/staging/ppc405-angstrom-linux/usr/lib -Wl,-O1 -Wl,--hash-style=gnu Jun 23 06:16:54 m4t: add TARGET_CC_ARCH += "${LDFLAGS}" in your .bb file Jun 23 06:18:23 Goddchen: add DEFAULT_PREFERENCE=1 to bb file in your overlay and see if that work Jun 23 06:20:52 could it be something with DEFAULT_PREFERRENCE="-1" in that bb file? Jun 23 06:24:24 Goddchen: dont think so unless you are using slugos distro Jun 23 06:27:29 hm should it recompile udev when i'm changing INITSCRIPT_PARAMS? Jun 23 06:28:59 03Koen Kooi  07org.openembedded.dev * r05386b0638 10openembedded.git/recipes/nbench-byte/nbench-byte_2.2.3.bb: nbnench-byte: fix QA errors Jun 23 06:35:05 khem do you feel like seeing if slang compiles under your toolchain? Jun 23 06:35:29 Goddchen: do you have your overlay added to BBFILES Jun 23 06:39:11 a native build shows gnu_hash Jun 23 06:39:20 as do all my other cross-compiled libraries Jun 23 06:40:33 m4t: Did you make the change I suggested ? Jun 23 06:41:17 yes Jun 23 06:41:25 did not help ? Jun 23 06:41:29 no Jun 23 06:41:43 i wasnt clear on if i should have ldflags defined in local.conf as welll Jun 23 06:41:51 i'd rather not to it globally to fix a single package Jun 23 06:42:19 i tried passing some ldflags= env variables to make as well Jun 23 06:43:37 oh i fixed it Jun 23 06:43:45 m4t: in slang_1.4.9.bb Jun 23 06:43:55 i just commented out all the bs they had defined in slang_1.4.9.bb Jun 23 06:44:00 now its gnu_hash Jun 23 06:44:18 * khem -> sleep () Jun 23 06:46:23 thanks Jun 23 06:59:09 damn it, it's still using openembedded/recipes/udev/udev_141.bb instead of local/recipes/udev/udev_141.bb Jun 23 06:59:36 my local overlay works since i use a package that i created in there which is now present in openembedded recipes Jun 23 07:03:55 good morning Jun 23 07:06:05 i now copied not only the file i need to my local overlay, but all files in recipes/udev, bitbake is still not using the local overlay Jun 23 07:06:14 cool i fixed it Jun 23 07:06:50 the elflib build rule was missing $(LDFLAGS) Jun 23 07:06:51 ie. d $(ELFDIR); $(ELF_LINK_CMD) -o $(ELFLIB_MAJOR_MINOR) $(LDFLAGS) $(OFILES) $(ELF_DEP_LIBS) Jun 23 07:06:57 *cd Jun 23 07:07:44 objdump shows gnu_hash now though Jun 23 07:07:53 the ldflags were already correct Jun 23 07:07:59 ah, shame on me, i had a typo in my local.conf :( hopefully that was the source of all that evil ;) Jun 23 07:08:03 i got the idea from a 'libgnuhash' fix Jun 23 07:08:25 from another pkg Jun 23 07:08:28 iwill make a patch Jun 23 07:10:57 Hello there Jun 23 07:12:20 I have some problem compiling oe. I am getting error from flex package. Is this a well known problem ? Jun 23 07:13:55 Nobody here? Jun 23 07:26:52 hi aykut Jun 23 07:27:02 what do you mean by "compiling oe"? Jun 23 07:31:05 hi Goddchen Jun 23 07:31:25 I am trying to have a minimal image for davinci dm355 targer Jun 23 07:31:27 t Jun 23 07:32:20 bitbake -D minimal-image command gives error on flex package Jun 23 07:32:54 Error is "checking lex output file root... configure: error: cannot find output from flex; giving up" Jun 23 07:32:59 morning Jun 23 07:35:24 sry have to leave... Jun 23 07:35:35 hrw: hi Jun 23 07:35:58 no problem :) Jun 23 07:37:44 03Michael Smith  07stable/2009 * r49c1d5d399 10openembedded.git/ (4 files in 2 dirs): Jun 23 07:37:44 syslog-ng: fix SRC_URIs Jun 23 07:37:44 Signed-off-by: Michael Smith Jun 23 07:37:44 Acked-by: Denys Dmytriyenko Jun 23 07:37:44 Signed-off-by: Marcin Juszkiewicz Jun 23 07:37:45 Acked-by: Koen Kooi Jun 23 07:37:45 hi Robert Jun 23 07:37:49 03Martin Dietze  07stable/2009 * ra220ec4eb4 10openembedded.git/recipes/mobilemesh/mobilemesh_1.2.bb: Jun 23 07:37:54 mobilemesh: fixed broken download URL Jun 23 07:37:56 Signed-off-by: Marcin Juszkiewicz Jun 23 07:37:58 Acked-by: Koen Kooi Jun 23 07:38:00 03Martin Dietze  07stable/2009 * rc4105d8c4c 10openembedded.git/recipes/openswan/ (openswan_2.2.0.bb openswan_2.4.7.bb): Jun 23 07:38:05 openswan: fixed broken download URL Jun 23 07:38:07 Signed-off-by: Marcin Juszkiewicz Jun 23 07:38:09 Acked-by: Koen Kooi Jun 23 07:38:11 03Rolf Leggewie  07stable/2009 * re80eab10ef 10openembedded.git/recipes/atd/ (atd_0.70.bb atd_0.80.bb): Jun 23 07:38:14 atd: fix spelling of LICENSE Jun 23 07:38:16 Signed-off-by: Marcin Juszkiewicz Jun 23 07:38:18 Acked-by: Koen Kooi Jun 23 07:38:20 Acked-by: Denys Dmytriyenko Jun 23 07:38:22 03Martin Dietze  07stable/2009 * r660606d724 10openembedded.git/recipes/elvis/elvis_2.2.0.bb: Jun 23 07:38:29 Fixed broken download URL for elvis Jun 23 07:38:31 Signed-off-by: Marcin Juszkiewicz Jun 23 07:38:33 Acked-by: Koen Kooi Jun 23 07:38:35 03Tom Rini  07stable/2009 * rd8d356ab01 10openembedded.git/recipes/mpfr/mpfr.inc: Jun 23 07:38:40 mpfr: Both GPL and LGPL components Jun 23 07:38:42 Signed-off-by: Marcin Juszkiewicz Jun 23 07:38:46 Acked-by: Koen Kooi Jun 23 07:38:48 Acked-by: Denys Dmytriyenko Jun 23 07:38:50 03Otavio Salvador  07stable/2009 * r1ff2b24a3c 10openembedded.git/ (4 files in 3 dirs): Jun 23 07:38:53 fixes wrong license value usage for GPLv2 Jun 23 07:38:55 Signed-off-by: Otavio Salvador Jun 23 07:38:57 Signed-off-by: Marcin Juszkiewicz Jun 23 07:38:59 (42 lines omitted) Jun 23 07:41:12 is INHERIT += "rm_work" still working? I've just found out that part of an x11-image filled 26G of tmp Jun 23 07:42:38 valhalla: is should still work Jun 23 07:42:56 valhalla: "bitbake -crm_work_all x11-image" should free it Jun 23 07:43:28 I'll try it Jun 23 07:44:08 so I should wonder why it didn't do the rm_work stages before Jun 23 07:44:42 morning Jun 23 07:52:22 hrw: it's doing builing stages, and no rm_work Jun 23 07:52:28 ping booxter Jun 23 07:59:03 hrw: hi, I have a question about license for you Jun 23 07:59:42 I am getting crazy with cross-compiling a qt4-embedded based package ! Jun 23 08:01:07 zaurus-updater has LICENSE = "zaurus-updater", I added zaurus-installer (just packing together updater.sh+kexecboot-kernel) and if I run recipe_sanity it says I should add a license. Jun 23 08:02:05 could someone point me to a recipe that actually works with qt4-embedded and uses the BUILD-SYS qmake (the one from qt4-native-tools) ? Jun 23 08:02:45 hrw: zaurus-updater is using the encdec_updater, which is GPL Jun 23 08:04:33 03Koen Kooi  07org.openembedded.dev * r1a2d098a0b 10openembedded.git/recipes/linux/ (2 files in 2 dirs): linux-omap-pm 2.6.29: fix beagleboard audio Jun 23 08:04:33 03Koen Kooi  07org.openembedded.dev * r9fe59c9b8a 10openembedded.git/contrib/angstrom/build-feeds.sh: Angstrom feed builder: add more packages Jun 23 09:07:24 3 next sets of patches on oe ml Jun 23 09:10:00 ant_work: zaurus-updater == updater.sh == unknown license Jun 23 09:11:42 hrw: I see...still needs GPL stuff to build Jun 23 09:12:05 but itself it is nongpl Jun 23 09:12:29 ant_work: if linux kernel would need BSD-install to build would you call linux kernel BSD licensed? Jun 23 09:12:29 should I set LICENSE = "zaurus-installer" ? :D Jun 23 09:12:33 yes Jun 23 09:14:17 hrw: grepping recipes for 'LICENSE' gave several flavours... Jun 23 09:14:36 so? Jun 23 09:14:40 one is even 'Public Domain' iirc Jun 23 09:14:47 checked its sources? Jun 23 09:14:58 sqlite(3) is PD for example Jun 23 09:15:12 yeah, that's fine, "public domain" is a perfectly respectable licence. Jun 23 09:15:30 iirc was just one in whole /recipe tree Jun 23 09:15:37 so? Jun 23 09:15:39 strange then Jun 23 09:15:55 statistically speaking Jun 23 09:16:09 yeah, there are probably some variant spellings. it'd be nice to unify them but I find it hard to get too excited about that. Jun 23 09:16:35 as hrw says, sqlite sets LICENSE = "PD", which is presumably supposed to mean public domain. there might be other forms as well. Jun 23 09:17:48 actually, grep reveals quite a lot of "PD", so I guess that'd be the one to keep. if you wanted to change that "Public Domain" one to match then that'd be fine. Jun 23 09:18:54 hm...instaed of creating a bogus license, couldn't I just set PD for zaurus-installer (no own sources) ? Jun 23 09:19:00 ant_work: if I will release umbaumba 24.1 and write LICENSE="absyntware" will you complain/ Jun 23 09:19:18 I couldn't care less :) Jun 23 09:19:30 mckoan: if you're the distro maintainer for KaeilOS then you don't really need to send patches for your own .conf files to the list. you might as well just go ahead and check them in. Jun 23 09:19:42 ant_work: we do not know history of updater.sh - it was taken from sharp as it was iirc Jun 23 09:19:47 pb_: he lacks rw Jun 23 09:19:58 hrw: so posibly is 'leaked' stuff ? Jun 23 09:20:02 ant_work: er, is zaurus-updater really public domain? that doesn't seem very likely. Jun 23 09:20:05 hrw: oh, right, I see Jun 23 09:20:23 ant_work: it did not leaked Jun 23 09:20:45 ant_work: it was (probably still) available from sharp zaurus website with sharp system image updates Jun 23 09:21:23 ant_work: but never had license header rather Jun 23 09:21:28 hrw: I never spent time on the licensing issue...and vendors GPL breakages. Jun 23 09:21:40 did Sharp release all what supposedly should? Jun 23 09:21:50 you mean? Jun 23 09:22:18 some drivers are still proprietary (for 2.4) iirc Jun 23 09:22:28 never disclosed the sources Jun 23 09:22:33 is it right? Jun 23 09:22:40 they never did Jun 23 09:22:58 for !collie we do not need SD driver sources as this is just pxa... Jun 23 09:23:06 for collie we reveng driver Jun 23 09:25:20 but even on collie we use in-tree driver: mmc-spi + board stuff iirc Jun 23 09:25:41 pb_: I haven't RW access yet Jun 23 09:28:24 mckoan: ah, sorry, I thought you did. fair enough then. Jun 23 09:28:51 ant_work: the only thing for which maybe having source would be good is sharp camera cf driver. but this hardware is terrible so no big miss Jun 23 09:29:00 :) Jun 23 09:29:18 ant_work: of course, would be nice to officially have LOCOMO documentation... Jun 23 09:30:04 kergoth has that, right? Jun 23 09:31:04 pb_: we got it from anonymous Jun 23 09:31:39 oh, I thought kergoth got it officially from sharp. isn't that why he can't share it with anybody else? Jun 23 09:31:43 pb_: and yes, kergoth has it - he signed NDA Jun 23 09:32:52 03Marco Cavallini  07org.openembedded.dev * r363ae30123 10openembedded.git/conf/distro/kaeilos.conf: Jun 23 09:32:52 Added PREFERRED_VERSION_classpath to support JVM in KaeilOS distro with contribution of Ronetix (www.ronetix.at) Jun 23 09:32:52 Signed-off-by: Marco Cavallini Jun 23 09:33:32 pb_: and whole datasheet is in japanese Jun 23 09:35:30 ok, I go to pack my bag for LinuxTag Jun 23 09:35:38 be back later Jun 23 09:35:48 thx for puching Jun 23 09:35:57 s/puching/pushing Jun 23 09:36:39 would be possible to push also this ? http://patchwork.openembedded.org/patch/771/ Jun 23 09:37:07 hrw|gone: yeah, that's probably unavoidable for a chip that was designed in japan. I doubt sharp would consider it reasonable to translate the entire datasheet just for oe's convenience. Jun 23 09:37:32 lots of vendors wouldn't even bother to create datasheets for internal parts so I guess we should consider ourselves lucky that there is one at all :-) Jun 23 09:37:33 mckoan: it is missing deps on java and other Jun 23 09:38:05 pb_: this guy seems to have don esame work as thesing... http://josejx.net/collie/ Jun 23 09:38:11 pb_: remember that Laibsch is fluent in Japanese ;D Jun 23 09:38:28 mckoan: it seems a bit weird that a kaeilos image is setting ANGSTROM_xx variables. is that right? Jun 23 09:38:35 ant_work: no, jose rev-eng driver and wrote infos. thesing used those info Jun 23 09:38:42 other than that, yeah, it looks fine. I can check it in later if you're happy with the angstrom thing. Jun 23 09:38:43 ant_work: clean reverse engineering Jun 23 09:38:45 ah, good to know Jun 23 09:38:48 bye Jun 23 09:43:05 pb_: yes is weird, but is a starting point ;-) Jun 23 09:43:27 heh, fair enough Jun 23 09:44:09 pb_: I already have more patches but I'm scared of submit them Jun 23 09:44:26 pb_: I'd like to have my own branch Jun 23 09:45:04 pb_: the main problem is that is too difficult get them accepted, like the other just rejected by hrw Jun 23 09:45:36 pb_: so I'm continuously stuck at the same point :-( Jun 23 09:46:37 * mckoan is debated whether create a local repo or try to submit his patches and ideas to OE mainline Jun 23 09:47:03 right, that is a bad situation. I don't see why you couldn't have your own branch, or even just direct commit access to the main repository for your own distro config files. Jun 23 09:47:29 as far as I am concerned, you are welcome to break the kaeilos config files any way you like so long as you don't damage anything else. Jun 23 09:48:47 pb_: I understand you feel safe in such case, but problems arise when I try to ass a recipe or an image Jun 23 09:48:59 /ass/add Jun 23 09:49:20 ~curse Dyslexia Jun 23 09:49:55 even then, adding new images shouldn't do any harm to people who aren't using them. the only situation where you might want to get review for your patches is if you are changing an existing recipe, or (particularly) a bbclass file or other bits of infrastructure. Jun 23 09:50:22 pb_: of course Jun 23 09:51:05 03Marco Cavallini  07org.openembedded.dev * r1b9b11cdb9 10openembedded.git/recipes/images/x11-jvm-image.bb: Jun 23 09:51:05 Added recipes/images/x11-jvm-image.bb for KaeilOS with JVM inclusion Jun 23 09:51:05 Signed-off-by: Marco Cavallini Jun 23 09:51:40 anyway, there you go, that other patch is in now. Jun 23 09:54:16 another issue would be whether is correct to add images into recipes/images or maybe would be better to use a subdirectory Jun 23 09:54:29 but in the second case there are 2 problems Jun 23 09:54:50 1. how to scan .bb in the subdir with bitbake Jun 23 09:55:16 2. how to organize subdirs (by name, role, distro) ? Jun 23 09:57:56 this is the scary distro x image matrix... Jun 23 10:02:46 yeah, creating subdirs there seems like it would be more trouble than it's worth. Jun 23 10:03:06 if you were going to do it, I guess the obvious thing would be to organise the images by DISTRO and arrange for images/${DISTRO}/*.bb to be included in BBFILES. Jun 23 10:03:43 but, there are a few images that are useful with multiple distros and it seems a shame to break the ability to share them without a compelling reason. Jun 23 10:04:54 yep, that's the main point against distro based organization, better to use role Jun 23 10:06:02 I'm not very convinced that roles provide any useful segregation. How would you classify them? Jun 23 10:07:03 I haven't idea yet, let's grow the /recipes/images dir and see what happens Jun 23 10:13:25 hello Jun 23 10:17:41 pb_: do we have COMPATIBLE_DISTRO? Jun 23 10:18:44 the distro-maintainer could add its distro in each image he tested Jun 23 10:19:34 well..here is the point... s/tested/built/ Jun 23 10:20:24 I don't think there's any such mechanism right now, but it'd be easy enough to add. It's certainly not a crazy idea. Jun 23 10:23:01 in therory it would be possible to create a 3d matrix distro-image-machine parsing tinderbox results Jun 23 11:43:06 Hi there. A quick question: Jun 23 11:43:38 Whenever I do a git fetch from stable/2009, it takes ages for the command to complete, and it ends up downloading a 100MB pack file. Jun 23 11:43:57 Is there some sort of local setting I need to use, or is this the nature of the beast? Jun 23 11:44:54 This is an incremental update, BTW. The last fetch was about two weeks ago. Jun 23 11:45:46 It certainly shouldn't be downloading the whole thing every time. Are you calling git fetch directly, or via git pull? Jun 23 11:46:26 git fetch. I am merging into my local branch later. I have a remote for openembedded (oe), and I use: Jun 23 11:46:34 git fetch oe stable/2009 Jun 23 11:47:22 What it tells me after a bit of blurb is: Jun 23 11:47:25 Getting pack 6431fad665c0fcd938f604945f0a3fc7e8d8df57 which contains 058b6ebc0e50370548fbae543b368c39394463ef Jun 23 11:47:29 that looks like it should be OK. I have no idea why it would be downloading so much data for you. Jun 23 11:48:02 100MB is not too bad, but with a 8-10K/sec download rate, it hurts. :) Jun 23 11:48:09 not that I use the stable branch myself, but I'm pretty sure even .dev doesn't accumulate 100MB of changes in two weeks. Jun 23 11:48:31 My thoughts exactly. Jun 23 11:54:21 I checked it. It was 11 patches in total that I needed. :( Jun 23 11:54:50 vhallac: try using git:// not http:// Jun 23 11:55:45 Yay! BET, it looks like you are spot on. It was http. Jun 23 11:55:57 I'll give it a try next week. ;-) Jun 23 12:04:26 Yup, it did the trick. I got a bunch of licensing changes in about 10 seconds instead of 3 hours. Jun 23 12:12:02 I am in Berlin Jun 23 12:14:38 Crofton: wunderbar! Jun 23 12:15:17 yes Jun 23 12:15:45 I am consuming my complimentry water Jun 23 12:17:17 don't forget to grab some döner kebap Jun 23 12:17:50 yeah Jun 23 12:17:54 I saw one of those places Jun 23 12:18:03 don't Germans eat German food? Jun 23 12:18:18 actually turkish fast food is the most popular here Jun 23 12:18:32 more popular than mcdonalds and such Jun 23 12:18:40 yeah, at least OEDEM we went to one of those stands Jun 23 12:18:57 eew that sounds pathologic Jun 23 12:18:58 XorA|gone, chose the Chinese looking stand next door due to the menu in English Jun 23 12:39:17 03Rolf Leggewie  07org.openembedded.dev * r007c2e2b94 10openembedded.git/recipes/bluez/ (bluez4_4.39.bb bluez4_4.41.bb): bluez4: rebase 4.39 and 4.41 on new .inc file Jun 23 12:39:18 03Rolf Leggewie  07org.openembedded.dev * r78ca80197a 10openembedded.git/recipes/bluez/ (bluez4.inc bluez4_4.42.bb): bluez4: poor man's unification Jun 23 13:09:53 How do I specify in a recipe that I need the opengl headers from mesa-full ? Jun 23 13:34:49 mickeyl: good morning Jun 23 13:35:33 morning pb_ Jun 23 13:45:04 How do I specify in a recipe that I need the opengl headers from mesa-full ? Jun 23 13:46:34 DEPENDS = "mesa-full"? Jun 23 13:46:48 <_rw_> hi, im having the problem "undefined reference: __sync_add_and_fetch_4" problem with boost 1.36. anyone knows how to apply the patches correctly? Even though I use them I end up with an error in a boost .so file Jun 23 13:47:29 pb_: sure but the headers are not copied into the HOST's staging dir ... Jun 23 13:58:43 _rw_: what arch are you on? __sync_add_and_fetch_4 have to be implemented manually if using gcc 4.3 or erlier on arm Jun 23 13:58:56 afaic Jun 23 13:59:06 k Jun 23 14:00:12 _rw_: i have not been using boost but i can help you implementing the missing gcc intrinsic Jun 23 14:00:20 why bitbake gcc-cross (for akita) is compilng qemu-naive ? Jun 23 14:00:30 what does it us qemu for ? Jun 23 14:01:49 <_rw_> xranby: an armv5te... I saw a fix but that was only for armv7 I think... Jun 23 14:02:34 _rw_: there area linux kernelhelper for arm that can be used to implement atomic operations in userland code . __kernel_cmpxchg can be used to implement the atomic __sync_add_and_fetch_4 for armv5 and armv4 Jun 23 14:03:14 i have been using it to make openjdk run on armv5 Jun 23 14:03:29 <_rw_> xranby: ok thanks... Im not familiar with kernel_cmpxcg, what is it/how can I get it? Jun 23 14:04:47 <_rw_> i guess its a kernel pathc? Jun 23 14:05:07 _rw_: it are part of the main arm linux kernel since 2.6.24 i think. Jun 23 14:07:30 <_rw_> xranby: ok so what do I need to do? Jun 23 14:08:57 _rw_: i will dig up some code for you Jun 23 14:09:45 <_rw_> xranby: thanks a lot Jun 23 14:11:50 _rw_: http://icedtea.classpath.org/hg/icedtea6/file/59b0e40b0d30/ports/hotspot/src/os_cpu/linux_zero/vm/atomic_linux_zero.inline.hpp Jun 23 14:14:00 <_rw_> xranby: thanks, sorry for my newbieness but how to apply this? Jun 23 14:14:06 _rw_: basically you will have to find the places in boost that uses __sync_add_and_fetch and make arm targets call arm_add_and_fetch Jun 23 14:15:05 <_rw_> Ok I see, well that seems pretty doable Jun 23 14:15:49 the lines 115 typedef int (__kernel_cmpxchg_t)(int oldval, int newval, volatile int *ptr); 116 #define __kernel_cmpxchg (*(__kernel_cmpxchg_t *) 0xffff0fc0) Jun 23 14:16:19 creates a method call that goes directly to the linux kernel and lets the kernel perform the atomic operation Jun 23 14:16:32 used by arm_add_and_fetch Jun 23 14:18:12 <_rw_> ok pretty nice Jun 23 14:19:38 <_rw_> xranby: I assume this means I cant use bitbake as I have to build the modified sources? Jun 23 14:21:14 _rw_: you can use bitbake.. go to tmp/work/armv5.../boost../boost.. Jun 23 14:21:26 use quilt to edit the files. Jun 23 14:21:59 with quilt you can quickly create a patch and then modify the boost recipy to use the new patch Jun 23 14:22:52 <_rw_> aha ok Jun 23 14:23:14 <_rw_> i need to study bitbake :) Jun 23 14:24:09 quilt are a independednt tool bitbake uses it internally to patch the source of various projects Jun 23 14:24:27 you can use quilt push or quilt pop to add or remove a patch Jun 23 14:25:00 quilt edit to att changes to an existing patch and quilt diff or quilt refresh to genereate a neew patch file Jun 23 14:25:32 <_rw_> ok I will check it out Jun 23 14:26:38 the patch itself are stored in the patch folder inside the project .. when you are done copy the patch to oe/recipes/boost/boost-versionnumber/mynew.patch and add SRC_URI += "file://arm-atomic.patch;patch=1" to the oe recipy Jun 23 14:26:58 im sure you will have a great time hacking :) Jun 23 14:54:20 hi, I don't suceed in adding -dbg packages to the image,I tried: Jun 23 14:54:32 RDEPEND += Jun 23 14:54:57 I'll try that: Jun 23 14:55:14 IMAGE_INSTALL += Jun 23 14:58:24 doesn't work either Jun 23 14:58:35 I have to add them into the image Jun 23 14:58:52 because It segfault at any program using shared libraries Jun 23 14:59:06 so even init or sh/ash from busybox segfaults Jun 23 15:01:08 or are there any script for gdb that looks for the .debug version automagically? Jun 23 15:01:28 the .debug beeing on my laptop...and gdbserver running on my target Jun 23 15:12:44 03Felix Domke  07org.openembedded.dreambox * r2ae8279099 10openembedded.git/packages/enigma2/ (enigma2-plugins.bb enigma2.bb): enigma2{,plugins}.bb: use same SRCDATE to prevent API conflict Jun 23 16:30:16 has someone already had a linking error with boost ? undefined reference to `__sync_add_and_fetch_4' Jun 23 16:31:38 it looks like it a optimizing assembly function that is not written/existing for arm Jun 23 16:35:50 Longfield, I've a patch for it Jun 23 16:36:01 Longfield, are you trying to cross compile wesnoth? Jun 23 16:36:16 Gnutoo: yeah I have just found your patch on the web ;) Jun 23 16:36:18 Longfield, which wesnoth version? Jun 23 16:36:25 Longfield, ok which one? Jun 23 16:36:37 Longfield, because there is one in C++ and one in assembly Jun 23 16:36:58 not wesnoth, but a program that uses boot thread library Jun 23 16:37:07 (written in my lab) Jun 23 16:37:22 the c++ one should be better because it let the compiler do the obtimizations Jun 23 16:37:26 ok Jun 23 16:37:51 Longfield, if you have commit access could you commit the c++ version of my patch? Jun 23 16:38:08 and test it (at runtime) of course Jun 23 16:38:13 I think I have found the asssembly version Jun 23 16:38:23 ok I'll look the c++ version Jun 23 16:38:47 but I don't have commit access .... but sure I'll give it a try Jun 23 16:40:14 Longfield, which version of boost do you use? Jun 23 16:40:24 I've also some recipes for 38 and 39 Jun 23 16:40:41 but they are not very tested and may need polishing Jun 23 16:41:09 36 Jun 23 16:41:38 Longfield, here's the patch but it's for 39...you may have to do very little adaptations: http://pastebin.com/m6bf559a6 Jun 23 16:41:51 (not in the code but only to make it apply) Jun 23 16:42:12 Longfield, by the way I did have a lot of help with this patches... Jun 23 16:42:19 s/did have/had Jun 23 16:43:53 by the way does someone have an idea on how to include -dbg packages in an image I tried different variable and none of them worked Jun 23 16:47:08 Gnutoo: do you still use ARM_INSTRUCTION_SET="arm" with your patch enabled ? Jun 23 16:48:52 Gnutoo, if you're feeling up to it, port the feature from poky to do that :) Jun 23 16:49:11 Tartarus, wow thanks I'll look Jun 23 16:49:27 Longfield, I'll look Jun 23 16:50:27 Longfield, no I think I've a patch for that... Jun 23 16:50:46 Longfield, do you want a tarball of my boost recipes? Jun 23 16:52:06 mmhhh, I think actual OE version already has an implementation for that, but the problem is that it has ARM_INSTRUCTION_SET="arm" and the assembly code is only set to ARMV7 (and I have ARMV6) Jun 23 16:52:15 Longfield, I've boost 33,34.1,36,38 and 39 with some patches Jun 23 16:52:29 Longfield, ok Jun 23 16:52:40 Longfield, i'll do a diff to look at the patches Jun 23 16:52:57 Gnutoo: well yes please give me at least version .36 Jun 23 16:53:04 ok Jun 23 16:53:38 Longfield, a tarball with all would be easier for me and you would get all the patches and everything Jun 23 16:53:51 Longfield, I'll post the tarball Jun 23 16:56:41 http://gnutoo.homelinux.org/open_embedded/boost.tar.bz2 Jun 23 16:57:02 here it is...feel free to ask questions about it but it's far in my memory Jun 23 16:57:15 some random infos: Jun 23 16:57:21 *I took some patches from debian Jun 23 16:58:43 *I didn't commit because It was for wesnoth and so because wesnoth segfault and that I don't know at all C++ I wasn't able to find the reason even with gdb and some help(it seems non trivial) from a wesnoth dev Jun 23 16:59:01 so I forgeten to commit it Jun 23 16:59:53 Gnutoo: thanks a lot. I will give it a try (at least version .36) and give you some feedback Jun 23 17:00:23 ok Jun 23 17:32:39 03Michael 'Mickey' Lauer  07fso/milestone5.5 * r83c3b73767 10openembedded.git/recipes/ (14 files in 6 dirs): Jun 23 17:32:39 libascenario-data: this no longer exists; move to fso-alsa-data Jun 23 17:32:39 libascenario: remove custom version which is deprecated; fso is using also directly now Jun 23 17:33:18 03Michael 'Mickey' Lauer  07fso/milestone5.5 * rabb3aff6b3 10openembedded.git/recipes/tasks/task-fso2-compliance.bb: task-fso2-compliance: ship fso-alsa-data Jun 23 17:38:08 see you next week - after LinuxTag Jun 23 17:38:16 or some - during LT :D Jun 23 17:38:45 hrw: don't drink too much beer :) Jun 23 17:38:49 haha Jun 23 17:39:16 nie Piwo Jun 23 18:04:10 flameman: I think qemu-native is used for locale stuff Jun 23 18:04:33 building locales AFAIR Jun 23 18:22:17 yes, that's right Jun 23 18:49:37 pb_: BTW, do you have force-defaults enabled locally for opkg-cl? Jun 23 18:49:40 do_rootfs Jun 23 18:52:28 03Klaus Kurzmann  07shr/import * rf9fafd8aea 10openembedded.git/recipes/tasks/task-shr-feed.bb: Jun 23 18:52:28 task-shr-feed: add opimd-utils Jun 23 18:52:28 Signed-off-by: Klaus Kurzmann Jun 23 19:14:33 kergoth: What is the status in Jun 23 19:14:37 !oebug 4452 Jun 23 19:14:39 * * Bug 4452, Status: NEEDSINFO, Created: 2008-07-23 22:25 Jun 23 19:14:40 * * kergoth(AT)gmail.com: Fetchers not run with a clean environment causes wget to obey user wgetrc. Jun 23 19:14:41 * * http://bugs.openembedded.org/show_bug.cgi?id=4452 Jun 23 19:15:47 CoreDump|cf-18: same question about Jun 23 19:15:50 !oebug 2493 Jun 23 19:15:51 * * Bug 2493, Status: NEEDSINFO, Created: 2007-06-15 18:39 Jun 23 19:15:52 * * treitmayr(AT)devbase.at: Subversion reports error when writing to repository (slugosbe) Jun 23 19:15:53 * * http://bugs.openembedded.org/show_bug.cgi?id=2493 Jun 23 19:17:55 * khem is happy to see Laibsch back Jun 23 19:18:02 * mickeyl as well Jun 23 19:20:35 thank you ;-) Jun 23 19:30:24 mario-goulart: hi there Jun 23 19:30:26 mario-goulart: welcome ;-) Jun 23 19:31:02 Hi otavio. Hi folks. Jun 23 19:31:22 Anyone experiencing problems with busybox on mips? Jun 23 19:31:56 mario-goulart: I think it is interesting to tell them that busybox-static does work Jun 23 19:31:58 mario-goulart: ;-) Jun 23 19:32:18 otavio: I was about to describe the scenario. :-) Jun 23 19:32:50 mario-goulart: ah; i forgot to ask, before asking you to join here, if it outputs any information on the serial console Jun 23 19:33:19 otavio: nothing much valuable, it seems. Jun 23 19:33:26 otavio: Just "Kernel panic - not syncing: Attempted to kill init!" Jun 23 19:33:34 mario-goulart: :( Jun 23 19:33:56 * otavio thinks how it could be debugged :( Jun 23 19:33:59 init is a shell script which echoes something and execs /bin/sh Jun 23 19:34:51 That init script works with busybox-static, but doesn't work with busybox(-dynamic) Jun 23 19:35:34 mario-goulart: one idea that just came to my mind is to install any stupid binary app, that uses libc, and test it using the busybox-static shell Jun 23 19:35:48 I've also tested using the helloword binary (from the helloworld recipe) as init and it works. Jun 23 19:35:48 mario-goulart: so we might find out if it is related to libc or not Jun 23 19:37:06 otavio: you mean a binary which uses dynamic loading? Jun 23 19:37:12 mario-goulart: it barely uses libc ... Jun 23 19:37:22 mario-goulart: yes ... maybe pciutils Jun 23 19:37:35 mario-goulart: lspci ought to be useful for us now anyway Jun 23 19:37:36 Laibsch: no, I don't think so Jun 23 19:37:50 otavio: right, I'm gonna try it. Jun 23 19:46:33 pb_: But you built gpe-image, right? How did you deal with bluez3 and bluez4 stepping on each others' toes during do_rootfs? Jun 23 19:53:54 mario-goulart: otavio you could try dynamically linked hello world program Jun 23 19:54:04 * * OE Bug 1831 has been RESOLVED (FIXED) by Jun 23 19:54:05 I suspect dynamic linker could be it Jun 23 19:54:06 * * gpe-dm hangs during shutdown Jun 23 19:54:08 * * http://bugs.openembedded.net/show_bug.cgi?id=1831 Jun 23 19:57:42 khem: just tried lspci here and it segfaults. Now trying a newer uclibc. Jun 23 19:58:30 khem: (using busybox-static as init.) Jun 23 19:58:32 mario-goulart: newer uclibc or newer binutils? Jun 23 19:58:46 otavio: sorry, binutils. Jun 23 19:58:51 question: what's the best way to use openembedded to build a minimal(/custom) rootfs? Jun 23 19:59:23 after sitting in on TI's tech day last week, one of the lectures was given on 2 second linux boot on the beagle board. Jun 23 20:00:01 I want to build a ramdisk w/ bb and some kernel modules, and could certainly script it, but i'd like to leverage oe where possible Jun 23 20:00:19 is there a wiki entry that helps bootstrap the python learning process? Jun 23 20:00:37 (by python, I mean the recipies and includes for oe) Jun 23 20:02:04 tomasg: what distro are you using on the host OS? Jun 23 20:02:12 host os is fedora Jun 23 20:02:18 does that really matter? Jun 23 20:02:32 something with deb packaging would have been better ;-) Jun 23 20:02:36 ah! Jun 23 20:02:37 But you're not out of luck Jun 23 20:02:44 take a look at the wiki Jun 23 20:02:53 the "Getting Started" page Jun 23 20:03:02 should be prominently linked Jun 23 20:03:10 yes.. i'm a little uncomfortable w/ the ipkg/opkg/oemake stuff too.. my target root fs will be RO and ~ 6-10MB Jun 23 20:03:12 or searchable even with sucky mediawiki search Jun 23 20:03:49 tomasg: try the micro distribution in OE and let OE create the rootfs for you Jun 23 20:04:22 i've so far customized the standard base and console-images, but now i'd like to play with running from a ramdisk and mounting my mtds after runtime Jun 23 20:04:43 my searches in google must not be turning up these pages... i'm looking at http://bec-systems.com/web/content/view/65/9/ right now Jun 23 20:04:54 wiki.openembedded.org Jun 23 20:05:04 or plain www.openembedded.org Jun 23 20:05:11 should also get you there IIRC Jun 23 20:09:21 Laibsch: That didn't seem to happen for me. I'm not quite sure why not. Jun 23 20:09:32 mario-goulart, I have problem with shared libs on mips Jun 23 20:10:01 mario-goulart, me too I've attempt to kill init Jun 23 20:10:46 pb_: Can you take a look at your local changes? Jun 23 20:11:12 otavio, hi what problem does mario-goulart have? Jun 23 20:11:15 on one of my boxes, I dropped bluez from the gpe task chain completely, because it was just such a mess to fix Jun 23 20:12:03 Gnutoo: dynamic linked binaries crash when executed on mips Jun 23 20:12:15 Laibsch: I've scanned the wiki, but it looks pretty void of documentation on this subject Jun 23 20:12:17 mario-goulart, wow I've exactly the same issue Jun 23 20:12:37 Gnutoo: so when using dynamic linked busybox as init, it crashes too. Jun 23 20:12:59 mario-goulart, I used uclibc+busybox+micro-base-image Jun 23 20:13:14 + 2.6.30 kernel from opwnwrt built with oe Jun 23 20:13:42 mario-goulart, I've not looked yet to the abi...maybe readelf could tells us something Jun 23 20:13:43 Gnutoo: same here. :-) Jun 23 20:13:44 what wiki page or earch criteria are you referring to? I've tried "root fs" "custom file system" Jun 23 20:14:00 mario-goulart, realy? openwrt??? Jun 23 20:14:29 Gnutoo: yes. And those bunch of patches. :-) Jun 23 20:15:01 mario-goulart, how a coincidence... Jun 23 20:15:07 s/how/what/ Jun 23 20:15:46 mario-goulart, how did you had the idea of porting openwrt to openembedded? Jun 23 20:17:01 Gnutoo: actually it was otavio's idea. Well, it seems easier, since they got it working already. Jun 23 20:17:16 mario-goulart, ah ok Jun 23 20:17:48 mario-goulart, did you have problem for tftp-booting the kernel? Jun 23 20:18:08 Gnutoo: yes, because of the kernel size, I suppose. Jun 23 20:18:23 yes but you needed: Jun 23 20:18:25 Gnutoo: After shrinking the rootfs it worked. Jun 23 20:18:36 ok I didn't add the rootfs yet Jun 23 20:18:55 I'll try readelf Jun 23 20:19:41 Gnutoo: here I'm using the vmlinux file generated in the OE's kernel workdir Jun 23 20:19:52 mario-goulart, ok me too Jun 23 20:20:19 mario-goulart, did you try to make a debug image? Jun 23 20:20:40 mario-goulart, there is something in poky that permit to do that Jun 23 20:20:55 Gnutoo: In OE too, but I'm using a release image. Jun 23 20:21:30 mario-goulart, ah how do you do that in OE ? Jun 23 20:21:47 mario-goulart, I've tried many variables inside the image recipe Jun 23 20:22:03 mario-goulart, is it the RELEASE or DEBUG variable in local.conf? Jun 23 20:22:23 Gnutoo: release Jun 23 20:22:37 ok Jun 23 20:22:39 I'll try Jun 23 20:23:31 Gnutoo: beware that, IIRC, release adds "quiet" to the kernel command line, so it may be hiding some things at boot time. Jun 23 20:23:45 I'll grep the manual Jun 23 20:24:14 Gnutoo: maybe it's worth checking the generated .config in the kernel workdir. Jun 23 20:24:32 mmm...DISTRO_TYPE may not be the right variable Jun 23 20:24:38 for me the config are nearly the same apart for: Jun 23 20:24:45 Gnutoo: .config's CONFIG_CMDLINE Jun 23 20:24:59 *jffs2 and mtd replaced by NFS Jun 23 20:25:05 *CMDLINE changed Jun 23 20:25:21 what router do you have exacrly? Jun 23 20:25:27 s/r/t/ Jun 23 20:25:29 Gnutoo: router board 450 Jun 23 20:25:35 what's that? Jun 23 20:25:42 I've wrt54gs v4 Jun 23 20:25:45 i don't see images/ micro-image Jun 23 20:25:56 Gnutoo: it's a mikrotik board Jun 23 20:25:56 under openembedded recipies for 2009.stable Jun 23 20:26:14 tomasg, maybe it's only in .dev ? Jun 23 20:26:26 grr Jun 23 20:26:28 gotcha Jun 23 20:26:40 mario-goulart, or I'll copy manualy the .debug files it would be faster... Jun 23 20:27:25 well... the other issue w/ using micro is that my question is really one of customization anyways... building just defers me for a little while, but i'll be back trying to understand how to extend the micro image to suit my device Jun 23 20:28:22 in a perfect work, ipkg/opkg on the host would allow me to take deployed packages (busybox, kernel mods, ubifs-tools) and "install them" to a loop-mounted ramdisk image Jun 23 20:28:27 s/work/world Jun 23 20:28:44 tomasg, ah ok,could the manual help you on this point,for me too I used it as a starting point... Jun 23 20:28:53 k00l Jun 23 20:29:11 it means the micro distro Jun 23 20:29:33 ok micro distro, not micro image Jun 23 20:29:56 looking at oe user manual on docs.openembedded.org Jun 23 20:30:00 I use wrt distro + micro-image Jun 23 20:30:23 umm? ok i can use angstrom distro + micro-image ? Jun 23 20:30:29 tomasg, basically you do "require micro-something.bb" and start customizing Jun 23 20:30:34 tomasg, yes Jun 23 20:31:48 Gnutoo: here we're using an image based on micro-uclibc Jun 23 20:31:52 the manual has no instance of "micro" but it does reference using "inherit image" Jun 23 20:33:12 tomasg, yes it says how to customize an image but say nothing about micro Jun 23 20:33:44 mario-goulart, ok me too Jun 23 20:33:50 mario-goulart, micro-base-image Jun 23 20:33:58 + wrt distro Jun 23 20:34:11 Gnutoo: is it OE? Jun 23 20:34:24 mario-goulart, what do you mean? Jun 23 20:34:30 I've made some small changes... Jun 23 20:34:41 Gnutoo: are you using Openembedded or poky (or both)? Jun 23 20:34:49 openembedded.dev Jun 23 20:34:54 Gnutoo: ah, ok. Jun 23 20:35:05 poky is just what I was told to look into in order to import the -dbg image Jun 23 20:35:16 right Jun 23 20:37:59 Gnutoo: thank you. i'm reading up on this. Sounds good. Is there a way to find out how to generate things like ramdisks, etc.? does oe do all the dd if=/dev/zero ... and filesystem creation? Jun 23 20:38:21 tomasg, yes IMAGE_FSTYPE or something like this Jun 23 20:38:29 great.. thx. Jun 23 20:38:37 mario-goulart: some other dude did mention problems with dynamic linked binaries on mips the other day; it does sound like that is a bit broken. are you using uclibc or glibc? Jun 23 20:38:54 pb__: uclibc Jun 23 20:38:57 pb__, I was this person Jun 23 20:39:02 :-) Jun 23 20:39:09 pb__, *I bet I was this person Jun 23 20:50:15 mmm...I was too impacient: Jun 23 20:50:17 2aaa8000-2aab2000 r-xp 00000000 00:0e 24128441 /mnt/NFS/lib/ld-uClibc.so.0 Jun 23 20:50:34 and the segfault is in 0x2aaa8a50 in ?? () Jun 23 20:50:52 oops Jun 23 20:51:10 s/oops// Jun 23 20:55:30 Hi! - is it possible to compile a native application depending on mesa even though mesa-native is a non existing recipe? Jun 23 21:02:14 custom image recipe is coming along... 2 issues: 1. there is apparently no ramdisk support (?) gumstix has a thread on this failing to work. 2. i started with the minimal image and customized it, but there is a lot of initscripts, udev, and other potentially time consuming assumptions. is micro-image better? Jun 23 21:05:56 strange...using readelf on openwrt's lib it gave : "There are no sections in this file." Jun 23 21:08:09 it's the same with openwrt's readelf Jun 23 21:08:37 use the cross-readelf tool Jun 23 21:08:38 i found DISTRO_INIT_MANAGER for controlling the init manager - thinking this should be "busybox" or "init" instead of "sysvinit" Jun 23 21:08:48 likewise, obviously that's what I used... Jun 23 21:08:54 in cross/bin dir Jun 23 21:08:55 Gnutoo: ok, just checking :-) Jun 23 21:09:34 tomasg, yes I wonder why it is set to sysvinit and not busybox... Jun 23 21:10:03 sgh, create the recipe Jun 23 21:11:24 Gnutoo: Just require the recipe and inherit nativ should do i for most I guess ? Jun 23 21:11:42 sgh, yes you could try that Jun 23 21:12:46 Gnutoo: thank you :) Jun 23 21:12:54 np Jun 23 21:17:50 Gnutoo: btw... do you happen to know how to manipulate ipk-files ? Jun 23 21:18:03 sgh, file package.ipk Jun 23 21:18:09 it will gives you the format Jun 23 21:18:17 if it's debian something use ar Jun 23 21:18:36 if it's gzip use tar z Jun 23 21:18:43 Gnutoo: ok Jun 23 21:23:36 otavio, hi how come did you have the idea of porting openwrt to oe? Jun 23 21:24:25 otavio, were you aware that I was also trying to do that(maybe if mario-goulart knew we could have shared the work) Jun 23 21:26:34 I'll go to sleep bye Jun 23 21:27:32 bye Jun 23 21:36:22 tomasg: as I said, the "Getting started" page is what I think you need first Jun 23 21:36:32 It's linked from the main page Jun 23 21:48:25 Laibsch: we are miscommunicating Jun 23 21:48:40 i'm hoping for some direction, but i've had an environment running for over a year Jun 23 21:48:56 getting started covers the BASIC nube howto use bitbake/oe Jun 23 21:49:05 I am looking to build a custom image Jun 23 21:49:25 Gnutoo directed me to the user manual which has got me to the point of building my custom image Jun 23 21:49:26 I understand your first to be that is what you needed Jun 23 21:49:33 good Jun 23 21:49:39 thx Jun 23 21:50:31 i now have a question w/ the image recipe: I want to include my kernel modules. Is there a way to derive the IMAGE_INSTALL arg for that? Jun 23 21:52:54 Anyone here having success with x11-native .... it does a "Floating point exception" on for me. :( Jun 23 21:55:22 oh; Gnutoo got out Jun 23 21:55:58 please tell him that I and Mario are going to travel tomorrow but we can get in touch at Monday and share work, for sure Jun 23 21:57:08 otavio: there is something called botmail, but I'm not sure how to operate it Jun 23 22:00:35 anybody pulled recently and got a fresh local tree? the HEAD got rewritten again... :( Jun 23 22:02:49 denix: how recent? Jun 23 22:02:55 I pulled a few hours ago Jun 23 22:03:00 What happened? Jun 23 22:03:43 Laibsch: someone overwritten the HEAD again, it's 8 days old now - http://cgit.openembedded.org/cgit.cgi/openembedded/log/?h=origin/org.openembedded.dev Jun 23 22:04:20 Laibsch: let me check if koen has anything latest... Jun 23 22:04:58 http://cgit.openembedded.net/cgit.cgi/openembedded/log/ is the last I see Jun 23 22:05:15 And I do have those (obviously, since I pushed them) Jun 23 22:05:30 never mind, false alarm Jun 23 22:05:35 OK Jun 23 22:05:57 when you say again, when did anything like this happen previously? Jun 23 22:06:03 was it ever discussed? Jun 23 22:06:59 my fault this time. actually, someone pushed an old origin/org.openembedded.dev and I clicked on it instead of just org.openembedde.dev Jun 23 22:07:01 so to answer my own question, it looks like images can cherry-pick modules with the IMAGE_INSTALL atom "kernel-module-MODULENAME" Jun 23 22:07:02 it's a bit strange to see that the link you gave only has changes 8 days and older Jun 23 22:07:37 it happened before though - there was no nook to prevent pushing non-fast-forwarded trees Jun 23 22:08:03 and no.. this wasn't in the user manual .. ;) Jun 23 22:09:11 Laibsch: if you take closer look at the branch name, you'll see the reason Jun 23 22:09:49 oh, yes Jun 23 22:09:57 I remember now ;-) Jun 23 22:10:08 Laibsch: it's origin/org.openembedded.dev instead of just org.openembedded.dev Jun 23 22:10:33 Laibsch: and I didn't see the difference at first and panicked :) Jun 23 22:10:42 mickey should delete that branch Jun 23 22:17:31 can't you? :) Jun 23 22:18:12 hi polyonymous_, haven't seen you around for a while Jun 23 22:18:20 Yeah, I haven't been there. Jun 23 22:18:31 Trying to get myself to build oe again. Jun 23 22:18:49 deleting branches is turned off in OE repo with a hook Jun 23 22:18:53 on the server Jun 23 22:18:55 Ah ok. Jun 23 22:18:59 But creating isn't? Jun 23 22:19:02 no Jun 23 22:19:12 funny. But well, it's safe. Jun 23 22:19:16 yes Jun 23 22:19:40 makes sense, me thinks Jun 23 22:20:18 Well, it was nice to see you, but I'm off to sleep now. Hope to see you again soon. Jun 23 23:41:35 i'm using a version of OE with busybox 1.2.1 on it. any ideas on how to disable syslog writing to the serial console? Jun 23 23:42:19 the device has a watchdog running that sends a message to the console every 20 or so seconds and is driving me crazy. Jun 23 23:56:57 berto-: busybox syslogd --help Jun 23 23:57:12 does it have any option that can help you Jun 23 23:59:07 berto-: is there -D option Jun 24 00:00:34 khem: no help. Jun 24 00:01:03 /sbin/syslogd: invalid option -- D Jun 24 00:02:00 berto-: does it read /etc/syslog.conf Jun 24 00:02:20 hmm, might not be syslog that's doing it. i killed syslog and the messages keep appearing. Jun 24 00:02:33 khem: yes, but it's not a standard syslog.conf. Jun 24 00:02:47 though, i may have been barking up the wrong tree this whole time. Jun 24 00:02:50 so find out who is generating messages Jun 24 00:12:50 hey khem i dunno if you saw Jun 24 00:12:56 i figured out the problem with slang though Jun 24 00:13:55 from src/makefile.in: Jun 24 00:14:02 +cd $(ELFDIR); $(ELF_LINK_CMD) -o $(ELFLIB_MAJOR_MINOR) $(LDFLAGS) $(OFILES) $(ELF_DEP_LIBS) Jun 24 00:14:11 its missing the $(LDFLAGS) there Jun 24 01:39:42 i fixed groff too Jun 24 01:40:16 config.log had bogus include paths for x_includes and x_libs Jun 24 01:42:55 i got a lynx.bb and lynx.inc going too if someone wants to try it Jun 24 01:43:10 it compiled and everything looks right i just dont have my board going yet **** ENDING LOGGING AT Wed Jun 24 02:59:57 2009