**** BEGIN LOGGING AT Fri Apr 01 02:59:58 2011 Apr 01 08:11:10 woglinde: he... Kopfwaschen für SOC Apr 01 08:30:42 ant yes Apr 01 08:31:04 good morning Apr 01 08:41:38 morning Apr 01 08:47:47 gdb tells me e.g. "warning: no loadable sections found in added symbol-file /lib/.debug/ld-2.9.so" when I debug a GStreamer app on my target. Same message for several libraries. Apr 01 08:48:20 I've found others with this question, but haven't seen a solution to it. Any ideas? Apr 01 08:51:17 hi hrw Apr 01 08:51:35 mikeul install the -dbg packages Apr 01 08:53:22 I have done. it finds the file /lib/.debug/ld-2.9.so. Apr 01 08:53:49 and opkg tells me that it belongs to glibc-dbg Apr 01 08:59:29 and 'nm' shows me a bunch of symbols in the .debug/ libraries. Apr 01 08:59:54 ldso is a diffrent package Apr 01 08:59:56 hm Apr 01 09:00:01 hello Apr 01 09:00:33 little question during bitbake invocation : this error ERROR: EOL while scanning string literal during parsing Apr 01 09:00:37 any idea ? Apr 01 09:01:43 morning all Apr 01 09:02:04 woglinde 'opkg search /lib/.debug/ld-2.9.so' returns 'glibc-dbg - 2.9-r37.4.6'. Apr 01 09:03:30 hm okay Apr 01 09:04:07 to take a different example, another one being complained about is '/usr/lib/.debug/libgstreamer-0.10.so.0.27.0', which is provided by gstreamer-dbg Apr 01 09:04:29 bluelightning: hey, do you still have spare time for OPIE fixing? Apr 01 09:04:47 there is that couple of legacy-staging recipes Apr 01 09:04:53 ant_work: of course Apr 01 09:04:59 hmm, I forgot about that Apr 01 09:05:12 about runtime, let me try to fix udev mess before Apr 01 09:05:23 IIRC there was one I looked at that was an obvious fix but the other I wasn't sure Apr 01 09:05:25 anybody for help me ? Apr 01 09:05:49 guilig: it means that you have an opening " without a closing one Apr 01 09:06:07 either that or a line in a multi-line string without a \ at the end Apr 01 09:06:40 in my local.conf file ? Apr 01 09:06:49 bluelightning: all' should just work' with latest udev and 2.6.38 but doesn't. I see there are a couple of mount.sh scripts and some blacklist for e.g. zaurus Apr 01 09:06:54 guilig: does the error say which file it is talking about? Apr 01 09:06:59 it should Apr 01 09:07:27 while parsing /home/guilig/dev/OE/openembedded/recipes/uclibc/uclibc_git.bb Apr 01 09:08:03 I just configured a config file, update git openembedded and try a test with bitbake nano Apr 01 09:08:17 guilig: does it say which line? Apr 01 09:08:39 and during the parsing, a lot of erros are listed like in the uclibc_git.bb Apr 01 09:09:15 sorry I didn't show you the full error Apr 01 09:09:21 guilig: use pastebin Apr 01 09:09:22 see here : Apr 01 09:09:28 ok Apr 01 09:09:33 guilig: thx :) Apr 01 09:09:36 gm kgilmer Apr 01 09:10:40 http://pastebin.com/2xpvx7kh Apr 01 09:12:51 hi woglinde, good morning to you. Apr 01 09:13:03 guilig: hmm, something is going wrong there Apr 01 09:13:12 guilig: which version of bitbake are you using? Apr 01 09:14:18 bluelightning: 1.10.2 Apr 01 09:25:40 guiling try 1.12 Apr 01 09:25:46 its faster at parsing Apr 01 09:25:59 but maybee its the last patch from khem Apr 01 09:28:27 hm no Apr 01 09:54:00 woglinde: after trying guilig's config it seems that it was due to setting DISTRO to angstrom-2008.1 with master Apr 01 09:54:29 somewhat undesirable behaviour in that it blew up on something else before getting a chance to error on DISTRO being invalid Apr 01 10:07:22 ant_work: so 2.6.38 is not OK on spitz yet? Apr 01 10:08:40 95% Apr 01 10:09:28 ant_work: what's still missing? Apr 01 10:10:07 touchscreen (wrong event( and sound Apr 01 10:10:26 hmm ok Apr 01 10:10:34 sound modules are compiled but not autoloaded Apr 01 10:10:36 bluelightning: hi, did you notice my mail on opie-devel ml about libSDL for opie? Apr 01 10:10:51 I don't know whether it's udev work now or modules-fpp* Apr 01 10:11:01 anarsoul: I did, sorry I have not yet had a chance to look at it or reply Apr 01 10:11:05 *modules-foo Apr 01 10:11:25 bluelightning: ok Apr 01 10:11:37 anarsoul: thx for working on SDL stuff :) Apr 01 10:12:03 bluelightning: no pb, just ask me if something's wrong :) Apr 01 10:13:59 bluelightning: btw task.base rdepends on modutils-initscripts (2.4 iirc) while task.base has different RDEPENDS for 2.4 and 2.6 Apr 01 10:14:32 so we end up with modules cruft I see Apr 01 10:14:57 ant_work: hmm that doesn't sound good Apr 01 10:16:35 , I'v eseen n900 has specific rules Apr 01 10:16:52 for loading sound modules Apr 01 10:20:32 Hi! ..... Do you guys know what this means ? http://paste.ubuntu.com/588179 Apr 01 11:00:38 03Steffen Sledz  07org.openembedded.dev * r23c3585c62 10openembedded.git/recipes/ (4 files in 4 dirs): Apr 01 11:00:38 linux-2.6.24: allow building kernel with newer GNU make Apr 01 11:00:38 GNU make >= 3.82 does not allow mixed implicit and normal rules Apr 01 11:00:38 Signed-off-by: Steffen Sledz Apr 01 11:00:38 Acked-by: Khem Raj Apr 01 11:01:05 hi woglinde Apr 01 11:01:14 I've a patch, I'm not sure I should send it Apr 01 11:04:39 hi dcordes Apr 01 11:22:43 sgh, not off the top of my head Apr 01 11:23:05 Crofton|work: apparently latest master gives me that error. Apr 01 11:24:32 I am not quite running that Apr 01 11:25:16 Crofton|work: hmmm ... it might be because of some overlay then. Apr 01 11:25:34 I meant I am not running master Apr 01 11:25:46 well, I ahven't pulled in a while Apr 01 11:37:00 Crofton|work: I had DISTRO = "angstrom-2008.1". Setting it to 2010.x fixes it. Apr 01 11:37:48 ah Apr 01 11:45:32 bluelightning: ping Apr 01 12:30:26 Hello; I sent 6 patches that have been in use here for merging. Can someone review them? Apr 01 12:30:40 I did a full build with those without problem. Apr 01 12:43:34 otavio: hey, how are you doing Apr 01 12:48:03 ant_work: doing better Apr 01 12:48:07 ant_work: :-) Apr 01 13:03:33 hi woglinde Apr 01 13:03:37 I've a big patch Apr 01 13:03:39 about 500k Apr 01 13:03:48 It needs review and can't be pushed Apr 01 13:38:59 ji mickeyl Apr 01 13:40:46 woglinde, can I send that patch to the list: http://gnutoo.homelinux.org/0001-Add-support-for-toaster72xx-machine.patch Apr 01 13:40:51 ? Apr 01 13:41:04 it's about 500k Apr 01 13:41:35 note that it can't go in Apr 01 13:41:40 I can't commit it directly Apr 01 13:42:52 dont know Apr 01 13:43:34 maybe I should reduce the patch to the machine config? Apr 01 13:43:44 because 500k is huge Apr 01 13:44:05 specially for that kind of patches Apr 01 13:57:45 woglinde_, I'll be back later Apr 01 13:57:58 maybe someone whould know if I can post theses kind of patches Apr 01 14:11:07 Anybody, please confirm or correct me: When building a package as a host native package the receipt shall have the name postfix "-sdk" and include "inherit sdk" Apr 01 14:11:11 Correct? Apr 01 14:14:14 that depends on what you mean by 'host native package' Apr 01 14:20:29 I have built log4cxx so that the target libs are placed in the SDK. Now I also want to have a host native log4cxx so that I can compile applications on host and link them with log4cxx. Apr 01 14:21:54 The other host native host tools are built fine into my SDK, but log4cxx does not link ok. Probably because some errors in ny log4cxx-sdk recipe. Comments? Apr 01 14:27:28 you'd likely need to elaborate on "does not link ok" :) Apr 01 14:29:03 kergoth: I have do_compile_xyz[depends] Apr 01 14:29:19 kergoth: I want to make it so that it becomes null op for uclibc Apr 01 14:29:31 kergoth: will override work here Apr 01 14:30:51 Sure. The log4cxx-sdk package compiles all packages and in the end it uses a libtool script that doesnt find the libgdb.la file. Apr 01 14:31:10 libgdm.la is built for i686. Apr 01 14:31:55 The libtool script does not earch for the newly built libtool.la but tries to find it in /usr/lib. !?!? Apr 01 14:32:20 log4cxx unmodified does nicely build for my ARM system. Apr 01 14:41:34 Anders__: the libtool script is old for the package Apr 01 14:42:00 Anders__: you might have to remove custom do_configure for this recipe Apr 01 14:42:06 and inherit autotools Apr 01 14:42:27 khem: overrides and flags don't mix well. case in point: FOO_libc-uclibc[bar] = "baz" -- that flag stays there, it doesn't get shifted to apply to FOO. Apr 01 14:42:34 it's something i've been wanting to do something about for a while Apr 01 14:42:36 but its non-trivial Apr 01 14:42:45 how should it behave? should flags combine, override one another, what? Apr 01 14:43:15 yes I want that Apr 01 14:44:03 I think if they behaved same Apr 01 14:44:17 like var overrides Apr 01 14:44:24 that would be nice Apr 01 14:44:30 i'd really really like to see override *blocks*, it'd make less confusion syntactically Apr 01 14:44:51 override libc-uclibc {\nFOO = "bar"\nFOO[depends] += "baz"\n} Apr 01 14:44:53 blocks Apr 01 14:44:59 also much less ugly than every damn var Apr 01 14:45:01 i see Apr 01 14:45:23 yes that would be better Apr 01 14:45:58 would avoid the need to rely on _append/_prepend too, since it'd give us conditional +=/=+ for free Apr 01 14:46:00 Ok, I'll check about libtool and autotools Apr 01 14:46:21 I changed the above to use DEPENDS Apr 01 14:46:24 anyway.. yeah, we don't have what you want today overrides wise Apr 01 14:46:27 * kergoth nods Apr 01 14:46:45 o_compile_perf[depends] =+ "elfutils:do_populate_sysroot" Apr 01 14:46:54 DEPENDS_perf_libc-glibc += "elfutils" Apr 01 14:47:00 but it seems not to work Apr 01 14:47:44 what's perf? Apr 01 14:47:56 its a yocto thing in kernel Apr 01 14:48:12 ah Apr 01 14:49:00 here its just a qualifier to differential do_compile Apr 01 14:49:25 kergoth: I thought that should have done the job Apr 01 14:49:48 since I see "elfutils" in the dep tree Apr 01 14:50:00 what's the behavior? Apr 01 14:50:27 it cant find headers that should be staged by elfutils Apr 01 14:50:46 so this did not build before do_compile_perf executed Apr 01 14:51:21 huh Apr 01 14:53:55 "linux-yocto.do_build" -> "elfutils.do_package_write" Apr 01 14:54:09 thats what I see in task dep .dot Apr 01 15:20:12 03Jason Kridner  07org.openembedded.dev * rba53dd199f 10openembedded.git/recipes/u-boot/ (2 files in 2 dirs): Apr 01 15:20:12 u-boot: BeagleBoard: add rev xM-C detection Apr 01 15:20:12 Signed-off-by: Jason Kridner Apr 01 15:20:12 Signed-off-by: Koen Kooi Apr 01 15:43:22 Is a md5sum error like this http://pastebin.com/7MYT5yvh routed at the .bb file, my OE configuration or something else? Apr 01 15:44:16 Is that an opkg you built yourself, or is it coming from the package feeds? Apr 01 15:44:33 bitbaked myself yes Apr 01 15:45:14 IIRC it's attempting to compare the feed md5sum with the local package -- try specifying "./.ipk" instead. Apr 01 15:49:49 mwester: Sorry little new to this, you mean "opkg install ./.ipk"? Apr 01 15:50:06 Yep Apr 01 15:50:21 Same error Apr 01 15:51:56 Well, you can always remove your package feeds -- rename the configuration files in /etc/opkg, do an opkg update to clear out the old info. Apr 01 16:06:05 mwester: that might be it, just going to configure server and see if it works ok. Apr 01 16:13:01 finally files a zipping through the ether, thanks for the help Apr 01 16:13:45 *are zipping I meant to say Apr 01 17:07:56 khem: hello, thank you for making fennec compiling under arm, but starting it on GTA02 (Freerunner) I get an Illegal Instruction, is there any fix available ? Apr 01 17:09:05 nschle85 you need to find which illgela instruction it is Apr 01 17:09:19 so we can see whats wrong with compile option Apr 01 17:10:37 woglinde: how do i do that ? My knowledge is on level : bitbake yxz.... Apr 01 17:13:00 woglinde: is it possible to create a thread dump on crash ? Apr 01 17:13:19 sorry I have to go Apr 01 17:13:39 nschle hm first you could look at do_compile log Apr 01 17:13:45 and the gcc args Apr 01 17:13:56 woglinde: ok Apr 01 17:13:59 there shouldnt be something like armv5 Apr 01 17:14:02 or amrv6 Apr 01 17:14:14 second grep source code for assembler Apr 01 17:14:32 woglinde: the patch contains assembler code Apr 01 17:15:09 woglinde: but with right compiler flags, this should result in an error i think Apr 01 17:15:35 woglinde: you are one of the Navit developers i think, thue ? Apr 01 17:17:20 hm not only Apr 01 17:17:26 but I have to go now Apr 01 17:17:30 till later Apr 01 18:01:38 nschle85: what arm processor does GTA02 run on ? Apr 01 18:02:13 armV4t Apr 01 18:02:49 hmm Apr 01 18:02:57 Nice SOC, but woefully ancient core. Apr 01 18:03:15 * mwester has a couple gta02s Apr 01 18:03:16 nschle85: ok and is that whats being passed in thru compile options Apr 01 18:03:21 mwester: hey Apr 01 18:03:39 mwester: I was thinking of porting over slugos distro to use oe-core Apr 01 18:03:53 mwester: we need to setup a git repo with right set of layers Apr 01 18:04:41 jftr, mozilla dropped support for everything below armv7 IIRC Apr 01 18:06:23 what mozilla code depends on the core? plugin ABI only? Apr 01 18:08:33 ouch for mozilla Apr 01 18:09:11 khem: i dont know, GNUtoo is helping me to find the issue, there is a ruourus that thumb mode is responsilble for that Apr 01 18:09:21 not sure Apr 01 18:09:25 rumourus Apr 01 18:09:25 that's what JaMa said Apr 01 18:09:42 but illegal instruction is often code for the wrong architecture Apr 01 18:10:06 GNUtoo: or currupt stack Apr 01 18:10:11 khem, when you get a chance, could you try building u-boot with an oe gcc 4.5? Apr 01 18:10:17 openoffice and mozilla are the kind of things that compiles for arm > v4 Apr 01 18:10:40 for android they had issues with v6 Apr 01 18:11:26 Crofton|work: which machine ? Apr 01 18:12:11 btw is it possible to send a patch-joke on the mailing list? Apr 01 18:12:31 (since the first april people makes joke at least in europe) Apr 01 18:12:46 GNUtoo: i installed dbg version of fennec but debugging symbols are not found Apr 01 18:12:50 GNUtoo: Thumb mode could cause it sure but since the processor is thumb capable I think it could be something else Apr 01 18:13:13 khem, indeed that what I tought Apr 01 18:13:23 nschle85: best is to enable USER_DEBUG=-1 on kernel cmdline Apr 01 18:13:24 nschle85, debug it, install -dbg and follow the guide I pointed to Apr 01 18:13:34 and compile kernel with relevant options Apr 01 18:13:45 khem, I too have been thinking of migrating to oe-core. But I know not where to create the SlugOS-specific git repo. Apr 01 18:13:52 then it will show you exact place where the illegal instr is happening Apr 01 18:14:04 mwester: oe.org Apr 01 18:14:08 or nslu.org Apr 01 18:14:11 whereever you wish Apr 01 18:14:28 mwester: oe-core is rock solid Apr 01 18:14:35 no nonsense breakages Apr 01 18:14:46 Then I shall start to work on that. Apr 01 18:14:55 I was thinking of something Apr 01 18:14:58 mwester: I think oe-core+meta-oe+meta-slugos Apr 01 18:15:03 is what we need for slug Apr 01 18:15:09 let's say there is oe-core+meta-oe+meta-shr Apr 01 18:15:10 meta-oe already exists as well? Apr 01 18:15:16 yep Apr 01 18:15:27 a recipe come in meta-shr with PR="r0" Apr 01 18:15:28 Hmm.. so meta-slugos would be very small. Apr 01 18:15:37 mwester: git clone git://git.openembedded.org/meta-openembedded Apr 01 18:15:38 then bugs are beeing fixed and the PR increase Apr 01 18:15:54 what PR should it have when it goes into meta-oe and is removed frmo meta-shr? Apr 01 18:15:58 mwester: meta-slugos will be mainly distro policies Apr 01 18:16:01 r0? Apr 01 18:16:17 GNUtoo: same pr Apr 01 18:16:21 ok Apr 01 18:16:22 thanks Apr 01 18:16:53 usually 0 is fine too since I expect the new oe-core to not be part of upgrade paths Apr 01 18:17:15 mwester: you could look at angstrom for reference Apr 01 18:20:48 nschle85: GNUtoo btw fennec is already compiled with ARM mode only Apr 01 18:20:59 it includes firefox.inc which has ARM_INSTRUCTION_SET = "arm" Apr 01 18:21:00 ok Apr 01 18:21:10 I'm installing fennec to debug it for nschle85 Apr 01 18:21:16 good Apr 01 18:21:18 ah it's not in shr feeds Apr 01 18:21:23 I'll have to compile it then Apr 01 18:21:33 fennec takes forever to compile on my box Apr 01 18:21:50 yes, fireox failled with a strange failure today Apr 01 18:22:04 coulnd't find glib.h Apr 01 18:22:12 altough the -I were right Apr 01 18:22:17 it should have found it Apr 01 18:22:23 and the #include was right too Apr 01 18:22:30 GNUtoo: we should upgrade to 4.0 for firefox as well Apr 01 18:22:35 ok Apr 01 18:22:38 but I dread the compile time Apr 01 18:22:49 + I am burried deep Apr 01 18:22:52 personally I fear more the faillures Apr 01 18:23:00 and the autootools issues Apr 01 18:23:08 that could take very long to resolve Apr 01 18:34:08 evening Apr 01 19:05:11 broonie, you will be at elc? Apr 01 19:20:13 why is 'socat' in the repository of recipes, but not part of the Angstrom feeds? Apr 01 19:22:05 koen hasn't added it to the feeds Apr 01 19:22:12 not all recipes make the feeds Apr 01 19:22:25 you can ask koen to add it though Apr 01 19:22:35 I've done that with a few I've found useful Apr 01 19:22:54 where is the list of files in the feeds managed? Apr 01 19:23:40 probably a build script koen has Apr 01 20:02:16 jkridner|work1: http://cgit.openembedded.org/cgit.cgi/openembedded/tree/contrib/angstrom/build-feeds.sh Apr 01 20:02:43 thanks denix Apr 01 20:11:18 hm.. Apr 01 20:11:36 on which language patchwork is written? Apr 01 20:18:13 django Apr 01 20:18:28 I think Apr 01 20:22:31 well.. Apr 01 21:27:11 Any infrastructure folks awake? The old ml for the old stable branch should probably be deactivated Apr 01 21:27:15 or at least hidden a bit Apr 01 21:28:29 mailing lists are on ltg Apr 01 21:28:43 florian or koen might be helpful Apr 01 21:42:44 or pb__ Apr 01 21:58:39 03Mark Hatle  07master * ree39de6856 10bitbake.git/lib/bb/build.py: (log message trimmed) Apr 01 21:58:39 build.py: Merge fakeroot changes from Poky Apr 01 21:58:39 Changes imported from poky repository, master branch as of commit: Apr 01 21:58:39 bcc817ee058bbab91d6e88a3e1e867a4ff09aeca Apr 01 21:58:39 In Poky fakeroot processing is handled exclusively by FAKEROOTENV. The Apr 01 21:58:40 fakeroot binary itself is never invoked directly. (This is due to Poky Apr 01 21:58:41 using pseudo, a fakeroot replacement that is more efficiently run by Apr 01 22:23:08 khem: ping Apr 01 22:35:34 conf-dbus fails to fetch Apr 01 22:35:39 *gconf-dbus Apr 01 22:36:54 i ran into that too Apr 01 22:41:30 ok Apr 01 22:41:33 is there a fix? Apr 01 22:42:09 no idea, haven't found one, just dug up the file on a mirror :\ Apr 01 22:44:43 ant_home: yes Apr 01 22:45:47 khem, I'm testing 'make savedefconfig' and that makes miracles Apr 01 22:46:01 hmm Apr 01 22:46:06 with uclibc or kernel Apr 01 22:46:27 kernel Apr 01 22:46:35 I've added a new task to my recipe Apr 01 22:46:50 but the right place would be where do_menuconfig lives Apr 01 22:47:01 I'll go bye Apr 01 22:47:07 too bad it'll fail to build Apr 01 22:48:15 khem: corgi shrinked from 36402 to 3328 bytes Apr 01 22:49:02 I'm re-diffing now easily the various devices Apr 01 22:53:17 ant_home: nice Apr 01 23:13:09 * Tartarus stabs php Apr 01 23:13:24 * Tartarus takes a swing at making php and php-native be in the same file and be less ancient Apr 02 00:15:11 03Siddharth Heroor  07master * r9628bcd27c 10openembedded.git/recipes/ti/ (ti-edma3lld-sysbios.inc ti-edma3lld-sysbios_02.10.03.04.bb): (log message trimmed) Apr 02 00:15:11 TI EDMA3LLD Sysbios: Move installer unpack to recipe version Apr 02 00:15:11 * The EDMA3LLD for Sysbios, does not ship as an installer Apr 02 00:15:11 but as a tarball. The older release 02.10.03.04 used an installer Apr 02 00:15:11 and this patch is in preparation of upcoming patches for new Apr 02 00:15:12 releases of EDMA3LLD. Apr 02 00:15:13 Signed-off-by: Siddharth Heroor Apr 02 00:15:19 03Siddharth Heroor  07master * ree06832fbe 10openembedded.git/recipes/ti/ti-edma3lld-sysbios.inc: Apr 02 00:15:20 TI EDMA3LLD Sysbios: Update recipe for SRC_URI Apr 02 00:15:20 * Add SRC_URI for tar.gz releases. Apr 02 00:15:20 * Bump PR. Apr 02 00:15:20 Signed-off-by: Siddharth Heroor Apr 02 00:15:20 Signed-off-by: Denys Dmytriyenko Apr 02 00:15:23 03Siddharth Heroor  07master * rf124b652d2 10openembedded.git/recipes/ti/ti-edma3lld-sysbios_02.11.01.02.bb: Apr 02 00:15:23 TI EDMA3LLD Sysbios: Add version 02.11.01.02 Apr 02 00:15:23 Signed-off-by: Siddharth Heroor Apr 02 00:15:23 Signed-off-by: Denys Dmytriyenko Apr 02 00:15:28 03Siddharth Heroor  07master * rae71894f24 10openembedded.git/recipes/ti/ti-cgt6x_7.2.2.bb: Apr 02 00:15:28 ti-cgt6x: Add release 7.2.2 Apr 02 00:15:28 Signed-off-by: Siddharth Heroor Apr 02 00:15:28 Signed-off-by: Denys Dmytriyenko Apr 02 02:32:11 Crofton|work: I'll be in SF and staying at Hotel Kabuki but not actually attending. Apr 02 02:32:36 Crofton|work: I'm in town for some customer stuff and also to present at Android builder's Summit Apr 02 02:32:43 Busy on Tuesday night. Apr 02 02:33:05 Crofton|work: Busy on Tuesday night. **** ENDING LOGGING AT Sat Apr 02 02:59:58 2011