**** BEGIN LOGGING AT Thu Nov 08 02:59:57 2007 **** BEGIN LOGGING AT Thu Nov 08 04:01:32 2007 **** BEGIN LOGGING AT Thu Nov 08 04:20:19 2007 **** BEGIN LOGGING AT Thu Nov 08 04:35:58 2007 Nov 08 07:42:34 hello Nov 08 07:43:38 anyone taking care of bug 2954? Nov 08 07:51:39 morning Nov 08 07:52:14 XorA: morining Nov 08 08:05:02 !oebug 2954 Nov 08 08:05:04 * * Bug 2954, Status: NEW, Created: 2007-09-08 09:56 Nov 08 08:05:05 * * alessandro.iurlano(AT)gmail.com: New program and recipe: ipkgorphan Nov 08 08:05:06 * * http://bugs.openembedded.org/show_bug.cgi?id=2954 Nov 08 08:07:24 ~curse gta01 for battery non-management Nov 08 08:07:25 May the fleas of a thousand camels infest your most sensitive regions, gta01 for battery non-management ! Nov 08 08:07:43 morning Nov 08 08:31:04 * * OE Bug 3266 has been RESOLVED (INVALID) by joaquinduran(AT)adtelecom.es Nov 08 08:31:06 * *  new version of library libsigc++-2.0 Nov 08 08:31:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3266 Nov 08 09:02:18 03hrw 07org.oe.dev * rccec70ab... 10/ (1 conf/machine/em-x270.conf): em-x270: do not assume where mtd-utils are installed Nov 08 09:19:21 re Nov 08 09:23:34 florian: ping Nov 08 09:26:12 good moring Nov 08 09:26:20 Jin^eLD: here :-) Nov 08 09:26:45 hey :) Nov 08 09:26:55 http://bugs.openembedded.org/show_bug.cgi?id=2482 Nov 08 09:27:03 morning all Nov 08 09:27:05 it's a one line fix to binutils.inc Nov 08 09:27:19 florian: after that the meta-toolchain worked out fine, the problem was in binutils-cross-sdk Nov 08 09:27:24 hey RP Nov 08 09:28:32 florian: and then I had one more little weirdness regarding meta-toolchain, I changed the SDK_PREFIX in my configuration to make it put everything in /opt, however an empty /usr directory was put into the tar.bz2 Nov 08 09:28:36 Jin^eLD: very good, i'll try it too... Nov 08 09:29:05 I modified the .bb file to simply remove that dire since it is empty anyway, not sure if that's the right way, could not find who creates it Nov 08 09:29:36 but apart from that the meta-toolchain worked out fine Nov 08 09:29:38 Jin^eLD: hmm... yes, that sounds like a bug - but unluckily that could be caused by any component Nov 08 09:31:12 Scanning through the sdk toolchain packages might narrow down the problem... Nov 08 09:33:40 hi, im trying to build a system with a 2.4 series kernel Nov 08 09:33:51 but when compiling uclibc it exists with the following error Nov 08 09:33:51 | + ./extra/scripts/fix_includes.sh -k /OE/OpenEmbedded/toolchain/mipsel-kernel_2.4/mipsel-linux-uclibc -t mips Nov 08 09:33:54 | Unable to determine version for kernel headers Nov 08 09:34:13 its gcc 3.4.4 Nov 08 09:34:19 uclibc 0.9.28 Nov 08 09:45:06 nik0n_: and which libc-linux-headers? Nov 08 09:45:30 * hrw needs uicmoc3 mastah Nov 08 09:45:36 mickey|zzZZzz: wake up! Nov 08 09:46:23 hrw: well i didnt find a libc-linux-header for 2.4 Nov 08 09:46:29 there are only 2.6 recipies Nov 08 09:46:37 or doesn't that matter? Nov 08 09:47:04 does not Nov 08 09:47:30 ~wake mickey|zzZZzz Nov 08 09:47:31 * ibot throws a barrel-full of ice water on mickey|zzZZzz and shouts "GOOD MORNING!!!!" Nov 08 09:47:40 nik0n_: btw - which mips machine? Nov 08 09:47:46 03rpurdie * r982 10/ (4 files in 4 dirs): Nov 08 09:47:46 git.py: Use git-fetch instead of git-pull removing any need for merges when Nov 08 09:47:46 fetching (we don't care about the index). Fixes fetch errors. Add Nov 08 09:47:46 BB_GENERATE_MIRROR_TARBALLS option, set to 0 to make git fetches faster at the Nov 08 09:47:46 expense of not creating mirror tarballs. (from poky) Nov 08 09:47:48 wrt54 Nov 08 09:49:05 I have one - with old openwrt Nov 08 09:49:31 well, i wann try it with OE, cause i dont like that buildroot of kamikatze Nov 08 09:50:06 nik0n_: look at wrtoe DISTRO config then Nov 08 09:50:25 nik0n_: it generate packages compatible with openwrt (atleast whiterussian) Nov 08 09:55:35 good morning all Nov 08 09:55:39 hey koen Nov 08 09:55:49 XorA: do we want to enable networkmanager support in epiphany? Nov 08 09:55:53 hey Jin^eLD Nov 08 09:56:24 koen|away: no idea Nov 08 09:56:37 I have no idea either :) Nov 08 09:56:49 I did turn on distro branding in the recipe :D Nov 08 09:57:04 * * OE Bug 3270 has been created by autobuild(AT)openembedded.org Nov 08 09:57:05 networkmanager stinks Nov 08 09:57:07 * * gnash-0.8.1-r0-do_compile Nov 08 09:57:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3270 Nov 08 09:57:17 * * OE Bug 3261 has been RESOLVED (FIXED) by Nov 08 09:57:18 03oliver.fels 07org.oe.dev * r74f70c34... 10/ (6 files in 3 dirs): notecase: update to 1.6.9 Nov 08 09:57:18 * *  madwifi snapshot needs updating Nov 08 09:57:20 * * http://bugs.openembedded.org/show_bug.cgi?id=3261 Nov 08 09:57:22 03hrw 07org.oe.dev * r22f47091... 10/ (6 files in 3 dirs): notecase: unify recipes Nov 08 10:00:08 hello! Nov 08 10:00:45 anyone here familiar with the packaging of hostap-daemon? Nov 08 10:01:38 hrw: i enabled PREFERRED_VERSION_linux-libc-headers ?= "2.6.20" Nov 08 10:01:46 but still have the same error Nov 08 10:03:14 03koen 07org.oe.dev * rf4307e16... 10/ (3 files in 2 dirs): epiphany: add svn version to get the webkit backend to compile Nov 08 10:03:19 03koen 07org.oe.dev * r9cd0ed95... 10/ (1 conf/distro/include/sane-srcrevs.inc): sane-srcrevs: add epihpany, fix elf2flt Nov 08 10:06:40 XorA: http://www.angstrom-distribution.org/feeds/2008/ipk/glibc/unsorted/armv5te/epiphany_2.20.1+svnr7650-r0_armv5te.ipk Nov 08 10:07:28 ah..finally split is over :) Nov 08 10:07:35 hi everyone Nov 08 10:10:14 :> Nov 08 10:11:00 someone know good tool to merge addressbooks? Nov 08 10:11:08 meld? :)) Nov 08 10:11:16 hrw: which sort? Nov 08 10:11:29 hrw: export to vcard, concatenate, import? Nov 08 10:12:00 NAbyss: I have few files with vcards and want to merge data in them - phones/emails/addresses etc Nov 08 10:12:15 multisync *might* be able to help. Nov 08 10:12:19 hrw: Oh, I see =/ Nov 08 10:12:35 hrw: pen&paper and a lot of time? :) Nov 08 10:12:37 install OSX, import into Adressbook.app, export as vcard :) Nov 08 10:13:22 ;D Nov 08 10:13:25 if the answer to a question is "install OSX" then you need to rethink what you are doing :) Nov 08 10:13:42 hrw: i enabled PREFERRED_VERSION_linux-libc-headers ?= "2.6.20" in my distro.conf for the mips build, but still have the same error when compiling uclibc "Unable to determine version for kernel headers" Nov 08 10:14:02 aaargh Nov 08 10:14:12 koen/hrw: http://bugs.openembedded.org/show_bug.cgi?id=3178 <-- can you have a look at that.. I provided a patch, but it's probably not a clean solution Nov 08 10:14:15 mv bv4 discharged its battery while being attached to usb Nov 08 10:14:20 koen|away: same here Nov 08 10:14:36 my Z discharges itself whilst plugged into DC Nov 08 10:14:42 Do any body use icu (International Components for Unicode )? I found that strip didn't run on the package generation. Nov 08 10:15:40 gilligan_: gas/configure isn't regenerated in do_configure? Nov 08 10:15:59 benlau2: webkit uses it Nov 08 10:16:19 yes Nov 08 10:16:42 I am building a image with webkit Nov 08 10:16:48 ~curse tc6393-ohci for crushing in the irq handler Nov 08 10:16:49 May you be reincarnated as a Windows XP administrator, tc6393-ohci for crushing in the irq handler ! Nov 08 10:17:02 However, the icu library without strip is too large for my board... Nov 08 10:17:24 koen: I did a `-crebuild bitbake` (which certainly calls do_configure?) and it works Nov 08 10:17:43 koen|away: Any idea why strip didn't run ? Nov 08 10:17:48 benlau2: package.bbclass should run strip, so I wonder why it doesn't do that Nov 08 10:18:12 gilligan_: ok, I shot myself in the foot many times by patching configure instead of configure.ac :) Nov 08 10:18:26 koen: hehehe. ok Nov 08 10:18:28 keo: yes, I am also quit confused... the rest of packages are fine. Nov 08 10:19:04 + export BUILD_STRIP=strip Nov 08 10:19:04 + BUILD_STRIP=strip Nov 08 10:19:04 + cd /home/benlau/Desktop/OE/tmp/work/armv6-angstrom-linux-gnueabi/icu-3.6-r1/icu/source Nov 08 10:19:04 + do_package_write Nov 08 10:19:55 Can't see any line to export STRIP for host Nov 08 10:20:15 usr/lib/libicuio.so.36.0: ELF 32-bit LSB shared object, ARM, version 1 (SYSV), not stripped Nov 08 10:20:51 good.. it is not my own problem... Nov 08 10:24:13 koen|away: well this could probably also be applied in the configure.ac of the binutils root I guess.. but I am not so very tempted to wade through the endless autotools swamp... ;] Nov 08 10:26:03 benlau2: I see where he problem is Nov 08 10:26:18 if not os.path.islink(file) and not os.path.isdir(file) and isexec(file): runstrip(file, d) Nov 08 10:26:32 -rw-r--r-- 1 koen koen 1220129 Nov 7 17:37 libicuuc.so.36.0 Nov 08 10:26:44 the libs aren't executable Nov 08 10:27:04 koen|away: gas/configure.in also checks for ${need_libm} and if it is set does AC_CHECK_LIBM AC_SUBST(LIBM) so maybe need_libm should be set to yes there.. but I dunno Nov 08 10:28:06 RP: what do you think about removing isexec(file) from that check in package.bbclass? Nov 08 10:29:44 koen|away: oh..and i've got two more questions.. (1) how can I install all packages with debugging information and (2) libthread_db.so.1 should be part of uclibc I suppose? gdb tries to open it for debugging pthreads Nov 08 10:30:57 benlau2: http://rafb.net/p/dpjayA99.html Nov 08 10:31:22 gilligan_: ipkg install -dbg, that should pull in foo and all deps Nov 08 10:31:38 gilligan_: or ipkg install -dbg Nov 08 10:31:54 no idea on the thread-db Nov 08 10:32:12 koen|away: well I thought i could set some variable so that bitbake would install the dbg versions to the image Nov 08 10:38:25 keon|away: !!!! Nov 08 10:39:16 keon|away: So you changed the installation phase to add +x flag to those library? Nov 08 10:42:04 * * OE Bug 3271 has been created by autobuild(AT)openembedded.org Nov 08 10:42:06 * * epiphany-2.19.6-r1-do_configure Nov 08 10:42:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3271 Nov 08 10:44:05 * * OE Bug 3249 has been RESOLVED (FIXED) by Nov 08 10:44:07 * *  Notecase 1.6.9 update Nov 08 10:44:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3249 Nov 08 10:44:17 * * OE Bug 3272 has been created by jeremy.laine(AT)bolloretelecom.eu Nov 08 10:44:19 * * python-logging is missing some dependencies Nov 08 10:44:22 * * http://bugs.openembedded.org/show_bug.cgi?id=3272 Nov 08 10:53:03 benlau2: yes Nov 08 10:53:12 benlau2: mtn pull :) Nov 08 10:53:19 benlau2: thanks for spotting it Nov 08 10:53:20 thx!!! Nov 08 10:54:13 meh.. uclibc gdb segfaults! ;/ Nov 08 10:54:38 run it under gdb ;) Nov 08 10:55:00 * koen|away crossed fingers catchsegv works Nov 08 10:56:52 * gilligan_ slaps koen Nov 08 10:56:53 ;] Nov 08 10:57:26 03koen 07org.oe.dev * rc2e1d9a6... 10/ (1 packages/icu/icu_3.6.bb): icu: make libs executable so they get stripped Nov 08 10:59:36 grrr.. i need bloody gdb ;] Nov 08 11:03:41 koen|away: hm.. the package busybox-dbg is empty.. i guess that shouldn't be the case? Nov 08 11:05:25 gilligan_: I suspect busybox strips itself during compile :/ Nov 08 11:05:33 grrrr Nov 08 11:05:46 indeed Nov 08 11:06:02 that's where all the oe_runmake -e STRIP=echo stuff stems from Nov 08 11:06:35 i'm running from one dead end into the next Nov 08 11:07:10 welcome to embedded linux Nov 08 11:08:03 hehe Nov 08 11:10:00 nick jeremy|away Nov 08 11:18:04 gilligan_: How about the busybox package? Today I found that I must assign FILES_busybox manually. Nov 08 11:19:04 benlau2: huh? you have to do what? busybox per se works fine.. but I wish it wouldn't strip itself automatically Nov 08 11:20:32 Many files was installed , but not packed to another package. Nov 08 11:20:58 BB>> peek busybox FILES_busybox Nov 08 11:21:04 /usr/share/udhcpc Nov 08 11:21:46 The busybox package contains only the /usr/share/udhcpc. Nov 08 11:22:09 Therefore I set FILES_busybox to include other files. Nov 08 11:22:10 bitbake -i is broken Nov 08 11:22:33 it doesn't set FILES correctly Nov 08 11:22:38 99$ efika is back (for a limited time as i understood) -> http://www.directron.com/efika.html Nov 08 11:22:45 !!!!!! Nov 08 11:23:48 Which version should I rollback to ? Nov 08 11:25:55 benlau2: just do not use bitbake shell for building Nov 08 11:26:09 ok..... Nov 08 11:39:31 how the hell am I going to fix an alignment trap when the address of the trap reported lies in the middle of total nothingness..nggggrrh Nov 08 11:39:53 40005000-40006000 rw-p 40005000 00:00 0 Nov 08 11:40:14 right in that section.. which is some area reserved for the linker or so Nov 08 11:40:27 of course addr2line knows nothing about that Nov 08 11:40:49 and i can't debug and stop at the PC because... gdb segfaults ;] Nov 08 11:41:01 * gilligan_ punches his head against the table Nov 08 12:02:12 03florian 07org.oe.dev * r689e57ba... 10/ (4 files in 3 dirs): libchenabler: Add patch to fix i18n breakage. Nov 08 12:02:18 03hrw 07org.oe.dev * r1e16b8d9... 10/ (1 packages/uicmoc/uicmoc3-native_3.3.5.bb): uicmoc3-native: depend on qmake-native Nov 08 12:02:23 03hrw 07org.oe.dev * rd1a492ae... 10/ (5 files in 4 dirs): quasar: added 0.9beta - close #3224 Nov 08 12:02:37 ah.. guess i know why it crashes.. some library is missing Nov 08 12:06:14 any1 ever saw this on compiling for wrt54 machine with wrt54oe dist Nov 08 12:06:15 | mipsel-linux-uclibc-ld: cannot open linker script file ldscripts/elf32btsmip.xsw: No such file or directory Nov 08 12:18:00 ok, i have the file tmp/work/binutils-cross-2.17.50.0.12-r0/binutils-2.17.50.0.12/build.i686-linux.mipsel-linux-uclibc/ld/ldscripts/elf32btsmip.xsw Nov 08 12:18:07 why isnt it in the staging area? Nov 08 12:18:48 ~spell favorite Nov 08 12:18:50 'favorite' may be spelled correctly Nov 08 12:19:20 yay.. gdb fixed Nov 08 12:19:52 have to file a bug.. gdb is missing a dependency.. it need uclibc-dev Nov 08 12:25:56 koen: seems like uclibc also strips everything automatically Nov 08 12:33:04 * * OE Bug 3224 has been RESOLVED (FIXED) by Nov 08 12:33:06 * *  New Quasar recipes available Nov 08 12:33:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3224 Nov 08 12:53:38 do bugs in sanity.bbclass count an insane ? :) Nov 08 13:01:22 03eha 07org.oe.dev * ra5ebc445... 10/ (1 packages/apache2/apache2_2.2.3.bb): apache2: Use WORKDIR path to access apache2 init script - close #3226 Nov 08 13:18:04 * * OE Bug 3226 has been RESOLVED (FIXED) by Nov 08 13:18:06 * *  Use WORKDIR path to access apache2 init script Nov 08 13:18:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3226 Nov 08 13:27:55 hey mickeyl Nov 08 13:28:10 hey mickey Nov 08 13:28:23 hi guys Nov 08 13:28:31 hey mickeyl Nov 08 13:28:38 hi mickeyl Nov 08 13:29:51 mickeyl: you see your arrival seems to be quite a celebration ;] Nov 08 13:32:10 hi mickeyl Nov 08 13:32:32 mickeyl: I looked at that SRCREV stuff again. Did you see my message here last night? Nov 08 13:34:04 gm Nov 08 13:34:46 mickeyl: r3104 has a bugfix for the one issue with it found so far... Nov 08 13:34:53 hi Crofton|laptop Nov 08 13:35:18 morning Nov 08 13:35:44 morning.. which i just got up as well ;] Nov 08 13:38:16 koen, these GNU radios guys drink a lot :) Nov 08 13:38:51 Crofton|laptop: they have to do that to tune out all the waves in their heads ;) Nov 08 13:42:15 RP: no, i must have missed that. let me scroll up Nov 08 13:42:50 hmm, so much for the buffer Nov 08 13:42:51 heh Nov 08 13:43:05 RP: thanks anyways, i'll check the diff Nov 08 13:43:47 mickeyl: Poky revision r3101 Nov 08 13:44:01 mickeyl: http://svn.o-hand.com/view/poky?rev=3101&view=rev Nov 08 13:44:44 heh Nov 08 13:44:47 ISHOULDNEVEREXIST Nov 08 13:45:36 mickeyl: I will add some kind of cacheClear function or copy the dict instead of that before it goes into bitbake proper ;-) Nov 08 13:46:23 ISHOULDNEVEREXISTANDCERTAINLYWONTNOW :) Nov 08 13:46:25 RP: yeah, that'd be good. otherwise it looks nice. nicer than AUTOREV Nov 08 13:46:45 mickeyl: I still don't like it but it is more generic Nov 08 13:46:48 yep Nov 08 13:47:54 It also keeps the hack in __init__ rather than every fetcher to an extent Nov 08 13:48:11 mickeyl: If tonights autobuilds get on with it ok, I'll commit into bitbake svn then Nov 08 13:49:41 mickeyl: I presume there is no way I can avoid that global? Nov 08 13:49:45 these alignment traps drive me nuts.. ls (->busybox) causes an alignment trap at PC=0x40002800. If I run it through gdb and set a breakpoint at that address it never gets there Nov 08 13:51:19 RP: which one? Nov 08 13:51:47 mickeyl: bb.fetch.srcrev_internal_call Nov 08 13:52:18 well Nov 08 13:52:25 mickeyl: I wondered about some kind of function overloading but couldn't find anything that seemed to fit... Nov 08 13:52:27 at least please make it function local Nov 08 13:52:37 function global,even Nov 08 13:52:43 i.e. Nov 08 13:52:50 get_srcrev.internal_call = False Nov 08 13:53:02 (you can add attributes to functions) Nov 08 13:53:17 mickeyl: right, I didn't know that :) Nov 08 13:53:22 :) Nov 08 13:53:31 new since python 2.something Nov 08 13:53:40 pretty cool for some things Nov 08 13:54:02 RP: right, you do not know python ;) Nov 08 13:54:19 hrw: and Dodji doesn't believe me... Nov 08 13:54:37 and he thinks that I also know python ;D Nov 08 13:55:30 has anyone tested http://www.angstrom-distribution.org/unstable/images/collie/20071107/ yet? Nov 08 13:56:05 * * OE Bug 3271 has been RESOLVED (FIXED) by koen(AT)openembedded.org Nov 08 13:56:07 * *  epiphany-2.19.6-r1-do_configure Nov 08 13:56:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3271 Nov 08 13:56:50 time to shower Nov 08 13:59:01 koen no the battery is dead :) Nov 08 14:14:26 hm..are only certain bugs announced in here by cdbot2 ? was just wondering why the bug I filed earlier on never made it to the channel Nov 08 14:15:53 i think it only reports opened bugs and resolved bugs, nothing in between Nov 08 14:16:11 i did open a new one Nov 08 14:16:16 and about to do another Nov 08 14:16:24 might be lag Nov 08 14:16:28 or a bug Nov 08 14:17:12 mickeyl ping Nov 08 14:18:56 mickeyl: maybe the one I filed now makes it through ;-] Nov 08 14:20:51 or maybe not Nov 08 14:25:00 gilligan_: the bot parses the email bugzilla send AFAIK Nov 08 14:25:19 and email isn't exactly known for speedy delivery Nov 08 14:26:52 koen ;] Nov 08 14:48:04 * * OE Bug 3273 has been created by tobias.pflug(AT)gmx.net Nov 08 14:48:06 * * gdb missing 'uclibc-dev' dependency Nov 08 14:48:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3273 Nov 08 14:51:22 there we go Nov 08 14:51:42 just a bit of a delay Nov 08 14:52:29 hmm.. now to retain symbol information for uclibc... Nov 08 14:59:20 cbrake: hey there Nov 08 15:00:06 cbrake: slowly getting closer to fixing some of the uclibc fishyness Nov 08 15:00:57 gilligan_: great, what are you finding? Nov 08 15:01:49 cbrake: gdb was segfaulting because of a missing dependency.. actually you could fix that right now if you don't mind.. gdb just needs uclibc-dev added to dependencies (OE bug 3273) Nov 08 15:04:18 cbrake: next thing is that uclibc segfautls in some init function.. i'm on that right now (had to fiddle around in the uclibc makefiles to disable stripping of symbol tables) Nov 08 15:05:07 gilligan_: but we don't want gdb to depend on uclibc-dev for glibc builds ... Nov 08 15:06:41 RPDEPENDS_append_uclibc = "" + RPDEPENDS_append_uclibc-gnueabi = "" Nov 08 15:07:05 s/RPD/RD/g Nov 08 15:07:24 that's what overrides are for :) Nov 08 15:09:17 right Nov 08 15:09:28 not to depends Nov 08 15:09:44 ;] Nov 08 15:09:55 makes sense Nov 08 15:12:17 cbrake: i am still in the dark about tracking down the cause of the alignment trap though.. the address reported by the kernel trap info lies somewhere in a memory region reserved for internal uclibc linking stuff.. trying to break at the given address leads to nothing Nov 08 15:14:39 gilligan_: yeah, that one sounds difficult to track down. I wonder if the buildroot folks have fixed this already? Nov 08 15:15:02 cbrake: maybe i can get some useful information by posting something on uclibc mailing list Nov 08 15:15:11 gilligan_: nod Nov 08 15:15:26 gilligan_: the .29 release is basically broken, you need to apply the fixes from svn Nov 08 15:15:42 the uclibc dudes refuse to release a 0.29.1 with the fixes :( Nov 08 15:15:47 x_X Nov 08 15:15:55 cbrake: do you think compulab took your "updates over usb" idea for the em-270? Nov 08 15:16:02 and current svn is broken as well.... Nov 08 15:16:54 hm.. but if I remember correctly i had the same issues with angstrom2008 which uses..hm.. dunno which version, but obviously not .29 Nov 08 15:17:21 2008 should use .29 Nov 08 15:17:25 if not, it's a bug :) Nov 08 15:17:30 jsheldon: could be :-) Nov 08 15:17:44 oh Nov 08 15:17:45 hehe Nov 08 15:17:59 cbrake: I wish they would incorporate that functionality into armon Nov 08 15:18:48 cbrake: have you ever looked into replacing armmon with, say uboot ? Nov 08 15:20:36 which version of binutils should be used with 2.6.24? Nov 08 15:21:09 2.18 should be ok Nov 08 15:21:17 that's what I use... Nov 08 15:21:19 hmm Nov 08 15:21:33 I get this when trying to load the kernel via tftp from redboot Nov 08 15:21:35 *** Abort! Attempt to load ELF data to address: 0x20038000 which is not valid Nov 08 15:21:53 the OE setup works fine for an older version, 2.6.16, no problem there Nov 08 15:22:05 a collegue who is not using OE sent me his vmlinux and its fine too Nov 08 15:22:10 so there must be something with my build... Nov 08 15:22:24 arm board? Nov 08 15:22:27 yes Nov 08 15:22:33 does your .24 rc have the binutils notes patch? Nov 08 15:22:47 hmmm.. good question Nov 08 15:22:54 probably not Nov 08 15:23:11 try applying packages/linux/linux-2.6.23/binutils-buildid-arm.patch Nov 08 15:23:24 thanks, let me see Nov 08 15:23:59 otherwise it will append 3Gbyte of zeros to vmlinux Nov 08 15:24:02 cbrake: oh and i figured how to get binutils to compile for uclibc if that is of any use for you Nov 08 15:24:12 which compresses to a few megs, so you don't really notice Nov 08 15:25:07 ooh ok Nov 08 15:25:11 *rebuilding* Nov 08 15:25:36 * koen echo ANGSTROM_MODE=uclibc >> conf/auto.conf ; bitbake minimal-image Nov 08 15:26:16 hrm.. how hard can it be to to get a /lib/ld-uClibc-0.9.29.so with symbol table.. *ngggrRgh* Nov 08 15:26:56 try replacing all occurences of 'STRIP' and 'strip' with echo in the source tree Nov 08 15:28:08 i changed the definition of do_strip to "true;" but it seems that wasn't enough Nov 08 15:28:38 that's in OE Nov 08 15:28:55 you want the makefile to stop stripping (if that's the culprit) Nov 08 15:29:04 koen: thanks a lot! I think that was it, it loads via tftp now Nov 08 15:29:25 Jin^eLD: that patch should be in 2.6.24rc2 AFAIK Nov 08 15:29:40 hmm Nov 08 15:29:47 gilligan_: you can stop OE stripping by putting INHIBIT_PACKAGE_STRIP=1 in local.conf Nov 08 15:30:50 I thought I was using rc2.. hmm Nov 08 15:31:31 I said AFAIK, so I could be wrong :) Nov 08 15:31:48 well me to :) Nov 08 15:31:58 I'll have a look Nov 08 15:32:35 koen: ah okay Nov 08 15:33:04 * * OE Bug 3275 has been created by autobuild(AT)openembedded.org Nov 08 15:33:06 * * mplayer-0.0+1.0rc2+svnr24587-r5-do_configure Nov 08 15:33:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3275 Nov 08 15:33:16 * * OE Bug 3274 has been created by tobias.pflug(AT)gmx.net Nov 08 15:33:18 * * uclibc 0.9.29-r8 segfaults Nov 08 15:33:20 * * http://bugs.openembedded.org/show_bug.cgi?id=3274 Nov 08 15:34:52 koen: meh.. okay.. the changes I did were overwritten ;] Nov 08 15:37:32 can we move /usr/bin/matchbox-session from matchbox-common to matchbox-wm? Nov 08 15:38:01 folks might want to use matchbox-wm but without the vfolder structure matchbox-common opposes Nov 08 15:38:33 the only thing hindering that is /usr/bin/matchbox-session, which we can either duplicate among different matchbox-common substitutes or (preffered imo) move it to matchbox-wm Nov 08 15:44:18 koen: nope, the patch is not in rc2 yet Nov 08 15:49:22 mickeyl: no objection to that, but please have o-hand fix lib-taku instead of moving the problem under the carpet by reshuffling matchbox-common Nov 08 15:50:13 well, it's bit more complicated Nov 08 15:50:19 i'll leave matchbox-common for now Nov 08 15:50:40 the stuff it installs is not getting used here anways Nov 08 15:50:59 i agree though we should make it a bit more easy to switch between different window managers Nov 08 15:51:15 although matchbox seems to make most people happy, right? Nov 08 15:52:16 where does the split between matchbox-wm and matchbox-common come from anyways? Nov 08 15:52:18 historic? Nov 08 15:52:22 dunno Nov 08 15:52:33 both are fetched from o-hand svn/tarballs Nov 08 15:53:03 IIRC matchbox-session does 'exec x-window-manager' which is under u-a control Nov 08 15:53:07 Has anyone build webkit-gtk recently? Nov 08 15:53:33 Tartarus: this morning for arm and powerpc Nov 08 15:53:46 koen, what distro? Nov 08 15:54:04 I'm trying to backport to an older snapshot and i'm seeing none of the autogenerated files in WebCore being built, basically Nov 08 15:56:11 DISTRO=anstrom-2007.1 and DISTRO=angstrom-2008.1 Nov 08 15:57:26 ok, thanks Nov 08 15:58:36 what's a valid machine for that? Nov 08 16:02:01 03mickeyl 07org.oe.dev * r988a52b9... 10/ (3 files in 2 dirs): matchbox-common: fix PACKAGE_ARCH (all) Nov 08 16:02:06 03mickeyl 07org.oe.dev * r501fc85e... 10/ (3 files in 2 dirs): matchbox-common: remove duplicated SECTION Nov 08 16:02:11 03mickeyl 07org.oe.dev * r33d934f3... 10/ (5 files in 3 dirs): Nov 08 16:02:11 openmoko-session2: provide own matchbox-session file, get rid of matchbox-common Nov 08 16:02:11 (which installs vfolders we don't want, we have openmoko-today2-folders for that) Nov 08 16:02:17 03mickeyl 07org.oe.dev * rabd46ad1... 10/ (1 packages/tasks/task-openmoko.bb): task-openmoko: remove dependency on matchbox-common Nov 08 16:13:24 mickeyl: http://www.openembedded.org/~koen/wm.pdf Nov 08 16:23:14 bye Nov 08 16:35:46 03koen 07org.oe.dev * r3f1764c8... 10/ (1 conf/distro/angstrom-2008.1.conf): angstrom 2008: use uclibc 0.9.29 Nov 08 16:49:47 question about monotone: when i do "checkout" - do I fetch stuff from the web or is it just unpacked from OE.mtn that is already on my local drive? Nov 08 16:53:11 co is a local operation Nov 08 16:53:24 and pull not Nov 08 16:53:28 only 'push', 'pull' and 'sync' are network operations Nov 08 16:54:19 thats nice to know Nov 08 16:54:20 thanks Nov 08 16:55:03 I am creating an svn repo with custom things and want to have a snapshot that I can always reproduce Nov 08 16:55:07 so checking in all source files Nov 08 16:55:28 and was wondering if it was enough just to have OE.mtn there or if I should package up the checked out tree, but is co is local - thats cool :) Nov 08 18:44:00 hello all Nov 08 18:46:13 Is Koen the one who maintains the angstrom feed? Nov 08 19:05:35 mreimer: a group of people looks after the feed Nov 08 19:07:11 mreimer: if something is wrong or missing, please file a bug so someone can look after it Nov 08 19:27:12 hey koen! Nov 08 19:27:34 koen|away: just wanted to see how I go about getting some of the python packages updated. It looks like some still depend on python 2.4 Nov 08 20:06:58 So, I've finally gotten OE to work on Mac OS X. Must say I'm impressed with the error messages it gives when you don't have the right programs installed :D Nov 08 20:54:14 i'm trying to reconfigure the kernel, and i'm using 'bitbake cirtual/kernel -c devshell' though it is failing. Any help would be nice. Thanks. . h Nov 08 20:54:30 help or ideas on what to try :) Nov 08 20:58:54 I beleive there is now a -c menuconfig Nov 08 20:59:19 ooo ill have to look into that one. I wasn't aawre of it. Nov 08 20:59:38 I saw someone the other day mention it, I hope it wasnt a wishlist Nov 08 20:59:53 ahh well, one way to find out. Nov 08 21:01:01 classes/kernel.bbclass:do_menuconfig() { Nov 08 21:01:01 classes/kernel.bbclass: export SHELLCMDS="make menuconfig" Nov 08 21:01:05 looks hopeful :-D Nov 08 21:01:21 trying it now, seems t o be doing something. Nov 08 21:03:27 hrmm | (gnome-terminal:17169): Gtk-WARNING **: cannot open display: Nov 08 21:03:31 doh! Nov 08 21:08:12 Hmm... anyone else having trouble compiling gettext-native? Nov 08 21:16:05 hi Nov 08 21:16:21 mreimer: right - someone needs to build all python and put into feeds Nov 08 21:16:44 hrw: so I could build and upload somewhere? Nov 08 21:18:21 mreimer: iirc first you need to send ssh key to koen to get access to upload space Nov 08 21:18:31 hrw: ok Nov 08 21:18:34 I do not remember whole procedure Nov 08 21:18:43 hrw: should I ask koen? Nov 08 21:18:52 hrw: for the procedure, that is Nov 08 21:18:58 would be good Nov 08 21:19:06 hrw: thanks for your help Nov 08 21:19:07 and if you can then please document it somewhere then Nov 08 21:19:11 no problem Nov 08 21:19:16 hrw: ok Nov 08 21:19:38 I think that I will build some stuff today night too - but for alix (geodelx) instead of arm Nov 08 21:20:21 hrw: I have access to a 4x4x3.2GHz box, but I don't think they'd appreciate me using it for oe builds :-) Nov 08 21:20:30 ;)) Nov 08 21:20:42 4 quad xeons? Nov 08 21:20:58 hrw: yes, as I recall Nov 08 21:21:12 I just upgraded to athlon64 x2 2.2GHz Nov 08 21:21:35 hrw: it's still an upgrade :-) Nov 08 21:22:04 my build box is a 1.6GHz Athlon, which is an step up from my old PIII 450MHz Nov 08 21:22:17 oh the pain of the older version of mtn on that box! Nov 08 21:24:13 mtn <0.26 was arghhh Nov 08 21:24:26 I had 1.8GHz athlon then Nov 08 21:24:46 yes Nov 08 21:24:58 and bitbake recipe parsing was painful then too Nov 08 21:25:12 made for high latency bug fixing Nov 08 21:27:06 parsing was hardcore before this Nov 08 21:27:16 I bought second 512M ram for bitbake Nov 08 21:27:27 yeah Nov 08 21:28:05 heh.. now I built poky-image-sato from scratch in ~2-3 hours Nov 08 21:28:20 wow Nov 08 21:28:33 DRAM manufacturers probably saw a drop in sales due to bitbake improvments Nov 08 21:28:38 lol Nov 08 21:29:44 * mreimer lunch Nov 08 21:57:28 I would document the 'procedures' if people would just respond to the RFCs I post in the angstrom ml Nov 08 21:58:20 * koen|away calls it a day and goes to bed Nov 08 22:28:55 http://www.youtube.com/watch?v=B1O2jcfOylU Nov 08 22:32:12 lol hrw Nov 08 22:36:06 I wondered why cores have 4-8% of usage during build.. Nov 08 22:36:24 then broke build, enabled 4 bbthreads and parallel_make Nov 08 22:36:36 now both cores got ~100% use Nov 08 23:07:04 * * OE Bug 3276 has been created by andrea.adami(AT)gmail.com Nov 08 23:07:06 * * toolchain badness ? Nov 08 23:07:08 * * http://bugs.openembedded.org/show_bug.cgi?id=3276 Nov 09 00:31:05 * * OE Bug 3277 has been created by dfarning(AT)gmail.com Nov 09 00:31:07 * * replace oenote with oe_note Nov 09 00:31:09 * * http://bugs.openembedded.org/show_bug.cgi?id=3277 Nov 09 01:01:40 urg Nov 09 01:01:48 does newer OE have a run.do_compile file? Nov 09 01:44:12 http://lkml.org/lkml/2007/10/25/466 haha Nov 09 01:44:20 Tartarus: what do you mean? Nov 09 01:44:38 Tartarus: a file with that name is generated on the fly when do_compile tasks need to be executed Nov 09 01:52:34 i'm using an older snapshot where they stick around Nov 09 01:52:46 trying to debug why webkit-gtk builds in current but fails in my backport Nov 09 01:52:53 was going to diff the run.do_compile's Nov 09 02:06:17 Tartarus: this sentence didn't make sense but if you have a compile issue. You should look at the gcc error instead of randomly two files Nov 09 02:07:53 and I have no idea what you mean with current and backport Nov 09 02:08:12 anyway I'm going home now Nov 09 02:29:54 for anyone else that may be listening Nov 09 02:30:06 the problem is that none of the auto generated webkit files are being made Nov 09 02:44:44 can OE build openbsd packages? Nov 09 02:45:51 for the arm architechture? Nov 09 02:47:12 i'll ask again in the morning, it is quite late. heh **** ENDING LOGGING AT Fri Nov 09 02:59:57 2007