**** BEGIN LOGGING AT Sun Oct 17 02:59:58 2010 Oct 17 04:21:38 have a question Oct 17 04:21:43 might be a stupid one Oct 17 04:22:20 if I put ubuntu on something like a beagleboard Oct 17 04:22:35 and I am using mono Oct 17 04:23:00 would compiling on ubuntu on host pc let me use the same mono executable in beagleboard? Oct 17 04:28:20 off topic. this channel is for the openembedded project Oct 17 04:28:27 perhaps try the mono channel or #edev Oct 17 04:28:52 * kergoth grumbles, a *lot* of sed -i usage in OE to fix for OSX Oct 17 04:47:40 * kergoth tests a workaround Oct 17 05:02:29 okay, think i've got a fix for sed portability for OSX builds Oct 17 05:02:30 next.. Oct 17 05:20:27 well Oct 17 05:20:32 more relevant question Oct 17 05:20:37 installed mono-mcs Oct 17 05:20:47 except it doesn't find it ... Oct 17 05:21:09 not in /usr/bin... or /usr/sbin Oct 17 05:22:21 opkg list mono-mcs Oct 17 05:23:06 there's only one Oct 17 05:23:43 http://pastebin.com/ZFQumEdn Oct 17 05:30:20 tried removing and reinstalling, didn't help Oct 17 05:36:44 ooh Oct 17 05:36:53 I can just compile on windows and copy over the exec Oct 17 05:36:55 awesomeness Oct 17 05:40:43 er, i meant opkg files mono-mcs, lists the files installed by the package Oct 17 05:42:02 ah Oct 17 05:42:04 thats nice to know Oct 17 05:42:16 but apparently I can just code in VS on windows and copy exec over :D Oct 17 05:42:22 thanks much Oct 17 05:42:50 kergoth: hi, please feel free to push those overrides/filespath patches if they gets some acks, I'll be off today Oct 17 05:43:11 i'm about to test a commit which sanitizes OVERRIDES at ConfigParsed event Oct 17 05:43:20 drops duplicates and empty elements Oct 17 05:43:25 may help Oct 17 05:43:44 though.. hmm Oct 17 05:43:50 might have to use RecipeParsed for that, dangit Oct 17 05:43:57 hrmph Oct 17 05:47:56 Looks like the libsamplerate0.1.7 archive checksum changed. I'll format a patch for it. Does that sort of thing happen often? Oct 17 05:49:23 xobs: no, but better to find original archive somewhere and attach "diff -rq" to show what exactly changed in new archive Oct 17 05:50:16 jama: Alright. Strange, libatomic-ops also seems to have changed. Oct 17 05:50:45 xobs: and if "diff -rq" shows nothing, then compare ie "ls -la" of archives (sometimes only file permissions are changed) Oct 17 05:51:20 I mean unpacked archives in both cases :) Oct 17 05:51:40 jama: Right. I'm sure the problem is that I'm not using angstrom-mirrors.bbclass. Oct 17 05:53:52 if it was really changed then it's right to fix those checksums, no matter which mirrors.bbclass you're using Oct 17 05:58:09 03Chris Larson  07master * ra241106ca4 10openembedded.git/conf/bitbake.conf: Oct 17 05:58:09 bitbake.conf: sanitize FILESPATH Oct 17 05:58:09 - reverse OVERRIDES Oct 17 05:58:09 - drop empty overrides Oct 17 05:58:09 - uniq the list Oct 17 05:58:09 - use nested variable references rather than getVar Oct 17 05:58:10 Signed-off-by: Chris Larson Oct 17 05:58:14 03Chris Larson  07master * r16011461a1 10openembedded.git/ (classes/utils.bbclass lib/oe/utils.py): Oct 17 05:58:15 utils, oe.utils: add 'uniq' function Oct 17 05:58:16 Ignore duplicates in an iterable. Oct 17 05:58:17 Signed-off-by: Chris Larson Oct 17 06:04:08 Yep, something definitely did change. Not much, though. I'll format that patch. Oct 17 06:18:50 * kergoth is making good progress on oe builds on osx :) Oct 17 06:19:10 * ka6sox is happy to hear that. Oct 17 06:19:22 i had an idea to handle the argument portability issues Oct 17 06:20:18 things can get a little wierd with porting to osx. Oct 17 06:20:41 http://gist.github.com/630556 Oct 17 06:22:56 thats nice and clean. Oct 17 06:23:32 I'm thinking something similar should work for the cp argument madness wrt recursion -- at the very least, it will ensure an immediate failure if anyone uses args that aren't portable, regardless of platform Oct 17 06:23:38 which i think is a definite improvement Oct 17 06:24:31 failing sooner than later is always an improvement if it gets you to fix issues. Oct 17 06:24:38 * kergoth should set up a freebsd VM Oct 17 06:25:31 at least now its easy to setup a new "machine" instead of having to get all the hardware first. Oct 17 06:27:40 bleh Oct 17 06:27:46 mono-devel is also apparently broken Oct 17 06:30:24 is anyone doing regression on freebsd? Oct 17 06:32:43 i've heard someone was attempting to do builds on it, but i don't know whether they completed it or test it regularly -- would be nice to improve our portability in general Oct 17 06:32:54 ideally, it'd be nice to support freebsd, osx, cygwin, and potentially even msys Oct 17 06:34:27 ya, the more platforms we can support effectively would make the penetration better. Oct 17 06:35:58 i think windows is a particularly important one Oct 17 06:37:44 yep. Oct 17 06:46:49 cygwin isnt too bad to setup. Oct 17 06:58:29 03Khem Raj  07master * ra89e9c41f6 10openembedded.git/recipes/libgee/libgee.inc: Oct 17 06:58:30 libgee.inc: Use libtool macros from installed libtool Oct 17 06:58:30 * Do not use libtool macros that come with libgee Oct 17 06:58:30 * Default to autoconf provided do_configure Oct 17 06:58:30 Signed-off-by: Khem Raj Oct 17 07:24:28 khem: your libgee/libtool commit seems to break libgee-native for me Oct 17 07:24:50 | gee/Makefile.am:95: HAVE_INTROSPECTION does not appear in AM_CONDITIONAL Oct 17 07:51:22 03Koen Kooi  07org.openembedded.dev * rdfc23a1bbc 10openembedded.git/recipes/x-load/x-load-rsalveti_git.bb: x-load-rsalveti: add x-load 1.41 built from Salveti's tree Oct 17 07:51:23 03Koen Kooi  07org.openembedded.dev * r90536c28b5 10openembedded.git/recipes/angstrom/ (2 files in 2 dirs): angstrom-led-config: add pandaboard support Oct 17 07:51:26 03Koen Kooi  07org.openembedded.dev * r6728579502 10openembedded.git/recipes/powervr-drivers/ (2 files in 2 dirs): omap4-sgx-modules: add L24.9 Oct 17 07:51:26 03Koen Kooi  07org.openembedded.dev * r32e3e870ea 10openembedded.git/recipes/linux/ (4 files in 3 dirs): linux-omap4: add 2.5.35.3 from ubuntu git Oct 17 07:51:32 03Koen Kooi  07org.openembedded.dev * r870db17aec 10openembedded.git/recipes/gnome/gnome-screensaver_2.30.2.bb: gnome-screensaver: add 2.30.2 Oct 17 07:51:32 03Koen Kooi  07org.openembedded.dev * r84f26093c9 10openembedded.git/recipes/imagemagick/imagemagick_6.4.4-1.bb: Oct 17 07:51:32 imagemagick: add missing lcms dependency Oct 17 07:51:32 see http://gitorious.org/angstrom/angstrom-testlab/commit/22416165ec14512909041503be3aaa6490a62d89 Oct 17 07:51:55 03Koen Kooi  07org.openembedded.dev * r8d914cdf12 10openembedded.git/recipes/gnome/ (3 files in 2 dirs): system-tools-backends: update to 2.10.1 and make it work with libtool 2.4 Oct 17 07:51:55 03Koen Kooi  07org.openembedded.dev * r970de24799 10openembedded.git/recipes/ti/ti-dmai.inc: Oct 17 07:51:57 ti-dmai: pass c6accel location to makefile Oct 17 07:51:58 Signed-off-by: Koen Kooi Oct 17 07:51:58 03Koen Kooi  07org.openembedded.dev * r2451a24537 10openembedded.git/conf/machine/omap4430-panda.conf: omap4430-panda: update machine config to match latest MLO and linux providers Oct 17 08:07:29 i'm having an issue with libtool (for certain recipes) adding -L/usr/lib to the linker flags, which causes issues because i'm running arch linux and it stores libgcc_s.so in /usr/lib, not in /lib Oct 17 08:07:39 so a number of recipes are failing due to using the wrong libgcc_s.so Oct 17 08:07:59 i saw this mentioned in some #oe chat logs, but no solution was given - can anyone help? Oct 17 08:10:30 ka6sox: I'm interested in testing on freebsd Oct 17 08:10:33 and I'll Oct 17 08:10:45 next week I hope Oct 17 08:14:16 khem: you were actually the one talking about it, mentioning how libsdl included "libdir=/usr/lib" and on archlinux, things like that caused issues with libgcc_s.so Oct 17 08:16:06 03Koen Kooi  07org.openembedded.dev * rd9468afc58 10openembedded.git/conf/distro/angstrom-2010.x.conf: angstrom next: bump DISTRO_PR for overrides and filespath changes Oct 17 09:58:01 he zecke Oct 17 10:06:05 woglinde: hey Oct 17 11:25:35 gm Oct 17 11:25:51 hi all Oct 17 11:45:08 hi crofotn and effem Oct 17 13:48:49 I have very strange problem building linux kernel and packing it, I can't debug.. who could help me to find a problem? Oct 17 13:50:21 do_populate_staging) failed with 256 Oct 17 14:00:41 anyone here use mono on OE? Oct 17 14:51:17 khem, hi Oct 17 15:50:13 damnit Oct 17 15:50:22 ftdi is being weird too Oct 17 15:53:10 my recipe depends on lua5.1-native, how could I change the hardcoded path to /usr/bin/luac to that luac from lua5.1-native which is in sysroots Oct 17 15:53:32 ${sysroot}s/usr/bin/luac ? Oct 17 15:54:00 yes Oct 17 15:54:45 anyone use FTDI serial ports with beagleboard/OE? Oct 17 15:54:56 or is it enough to just use luac? it seems like the sysroots path is in PATH Oct 17 15:55:04 or ever seen weird shit with FTDI serial ports not liking \r or \n Oct 17 15:55:34 (err... sorry ynezz... should have prob waited...) Oct 17 15:56:39 ftdi likes \r\n Oct 17 15:56:54 as opposed to one or the other? Oct 17 15:56:58 it's probably problem somewhere else Oct 17 15:57:25 like if I take out 0x0As and 0x0Ds, data keeps coming in fine Oct 17 15:57:39 otherwise, I get data until the first 0x0A or 0x0D and then BAM... nothing Oct 17 15:57:52 wrong port settings or so Oct 17 15:58:04 it's not a hardware problem Oct 17 15:58:41 yeah its definitely some sort of config issue Oct 17 15:58:49 cause realterm on me winpc sees anything fine Oct 17 15:59:07 just cant see what, baud rate/parity etc are all right Oct 17 15:59:32 ok, you should ask on #beagle about that, this is #oe Oct 17 15:59:58 just wanted to figure out if there were any issues in OE about this, guess I shall bug people in beagle Oct 17 16:00:00 Jay7 are you going to test locally? Oct 17 16:00:36 ka6sox: I'm preparing buildserver to do QA testing for OE :) Oct 17 16:00:54 oooh, fancy Oct 17 16:41:49 Jay7, are you planning on using something automated to do that testing...like Hudson or Tinderbox? Oct 17 16:42:08 ka6sox: I'm thinking about this slowly :) Oct 17 16:42:31 yes, it is a Sunday... Oct 17 16:42:43 I'm thinking about tinder-bake :) Oct 17 16:42:52 bitbake mode to do automated testing Oct 17 16:43:01 but this is just crazy idea :) Oct 17 16:44:43 I am looking at Hudson-bake too. Oct 17 16:46:22 either one should work across multiple platforms. Oct 17 16:46:56 I think that Zecke has tinderbake working already on melo. Oct 17 16:58:09 he rp Oct 17 17:11:56 khem, hi Oct 17 17:12:00 woglinde_, hi Oct 17 17:12:05 he gnutoo Oct 17 17:12:17 gm debian is doing now a armhf port Oct 17 17:12:42 it has 40% performance gain on most floatingpoint stuff Oct 17 17:13:09 what is hf? Oct 17 17:13:15 hardware floaring point? Oct 17 17:13:17 hardfloat Oct 17 17:13:18 yes Oct 17 17:13:33 topic came up in ac100 today Oct 17 17:13:37 ok Oct 17 17:16:05 I wonder, why it's not creating -dbg package and doesn't strip the shared .so library http://pastebin.com/gkkgk9WU Oct 17 17:16:26 I use it in other recipes in the same way and it works, but in this one it doesn't Oct 17 17:16:37 any idea? Oct 17 17:17:23 it's 200kB lib so strip would help a bit Oct 17 17:21:34 ynezz aeh? Oct 17 17:27:11 woglinde_: hi Oct 17 17:27:52 jo Oct 17 17:28:10 * woglinde_ now try to setup some stuff for the ac100 Oct 17 17:30:25 woglinde_: do you know how does the creating of the -dbg package works? Oct 17 17:30:41 not 100% Oct 17 17:31:24 hi Oct 17 17:31:28 I've compiled library, then something creates .debug/library.so and strips the library.so Oct 17 17:31:29 lib is copied to .debug and than stripped Oct 17 17:31:47 yep, but for some reason it's not doing so Oct 17 17:32:01 I would like to find why, is the answer in the package.bbclass ? Oct 17 17:32:48 I hope that it's not, I'm lost there :D Oct 17 17:33:13 woglinde_, at evolvis some people cared about n900, was it for compatibility reason only(maemo-compat for java etc...) is there some other people caring about n900, I would like to merge some patch for booting on microsd Oct 17 17:33:53 gnutoo hm Oct 17 17:35:51 basically it's for the SHR distro Oct 17 17:35:57 we plan to bring it to n900 Oct 17 17:36:10 hm Oct 17 17:36:21 and I wonder if it would conflicts with some other people's goal Oct 17 17:36:30 you dont need maemo-compat Oct 17 17:36:39 me, not at all Oct 17 17:36:58 but I was wondering if what we do could break maemo-compat in some ways Oct 17 17:37:06 because we will boot from microsd Oct 17 17:37:07 in extras someone copied icedtea binarys from debian Oct 17 17:37:09 thats it Oct 17 17:37:15 ah ok Oct 17 17:37:18 better extras-devel Oct 17 17:37:21 ok Oct 17 17:37:31 so the maemo deb-src is whole binary Oct 17 17:37:34 what a shame Oct 17 17:37:47 so I could go on and and add support for the n900 for SHR? Oct 17 17:37:57 why you ask med? Oct 17 17:37:57 that would include xorg.conf, kenrel defconfig etc... Oct 17 17:38:00 just do it Oct 17 17:38:03 ok Oct 17 17:38:16 I cannt deciced what shr is doing or not Oct 17 17:38:28 SHR wants to merge that stuff into oe Oct 17 17:38:44 but I was wondering if it could hurt someone if I merge it Oct 17 17:38:46 *sigh* Oct 17 17:38:48 ok Oct 17 17:39:09 I will first merge some other stuff I did Oct 17 17:39:11 and then do it Oct 17 17:39:18 with the help of mr_nice|work Oct 17 17:39:19 oops Oct 17 17:39:21 mrmoku, Oct 17 17:52:59 jo jama Oct 17 17:55:05 jo Oct 17 18:01:17 03Thomas Zimmermann  07org.openembedded.dev * r6fffe3eb48 10openembedded.git/recipes/e17/notification_svn.bb: Oct 17 18:01:18 notification: add new recipe Oct 17 18:01:18 * notification is a notification gadget for E Oct 17 18:01:18 Signed-off-by: Thomas Zimmermann Oct 17 18:01:29 03Thomas Zimmermann  07org.openembedded.dev * r4006aa07d0 10openembedded.git/recipes/e17/ (efm-nav_svn.bb efm-path_svn.bb efm-pathbar_svn.bb): Oct 17 18:01:30 efm-{nav, path, pathbar}: add new recipes Oct 17 18:01:30 * this is a E module for the E file manager Oct 17 18:01:30 Signed-off-by: Thomas Zimmermann Oct 17 18:09:15 woglinde_: hardfloat can be done in OE too Oct 17 18:09:21 woglinde_: our gcc supports it Oct 17 18:09:31 just I dont have enough time Oct 17 18:09:45 gm all Oct 17 18:09:53 morning khem Oct 17 18:11:17 ka6sox: hi Oct 17 18:20:42 hm anyone knows how to say SRC_URI please save the download as file moo Oct 17 18:22:35 ah kergoth Oct 17 18:22:40 -> hm anyone knows how to say SRC_URI please save the download as file moo Oct 17 18:23:18 someone sent a patch to do that to bitbake-dev -- i'd thought there was already a way, but not sure Oct 17 18:23:38 hms Oct 17 18:23:48 seems I need to look into the source Oct 17 18:24:41 I've sent that patch Oct 17 18:24:51 he ynezz :) Oct 17 18:24:59 http://www.mail-archive.com/bitbake-dev@lists.berlios.de/msg00862.html Oct 17 18:25:10 hm its pushed? Oct 17 18:25:28 woglinde: I don't think so :) Oct 17 18:25:30 dont think so Oct 17 18:25:57 damn gitrous download Oct 17 18:26:08 use git then :p Oct 17 18:26:18 i have 2mbit Oct 17 18:26:22 ah Oct 17 18:26:23 and its kernel source Oct 17 18:26:25 noway Oct 17 18:26:37 good morning Oct 17 18:28:24 khem: can you please reply? :) http://www.mail-archive.com/bitbake-dev@lists.berlios.de/msg00862.html Oct 17 18:29:02 jo florian Oct 17 18:34:08 ugh which package needs flex for target es dep Oct 17 18:41:55 where is defined isexec() function, which is used in package.bbclass ? Oct 17 18:42:09 -rwxr-xr-x 1 ynezz ynezz 103032 2010-10-17 20:30 lua51-lanes.so Oct 17 18:42:17 google? Oct 17 18:42:18 git grep is your friend Oct 17 18:42:25 it return 0 on this file Oct 17 18:42:59 yep, found it in debian.bbclass but I don't know if it could be used in package.bbclass Oct 17 18:43:12 or if it's the correct one Oct 17 18:43:33 I don't know the internals that much, sorry Oct 17 18:43:34 def'd functions are global, not local Oct 17 18:43:42 in metadata Oct 17 18:44:25 hm, but if I run the content of that function manually, it returns 64 so it works Oct 17 18:45:29 there's this if not os.path.islink(file) and not os.path.isdir(file) and isexec(file): in package.bbclass which fails for me and so I don't get -dbg version of the package Oct 17 18:45:58 ie, no .debug is created, because isexec() think it's not executable file Oct 17 18:57:50 hm Oct 17 18:58:00 why the hell no ext3 image is generated Oct 17 18:58:19 only in the menu its really visibleIMAGE_FSTYPES = "ext2 ext3" Oct 17 18:58:25 hm Oct 17 18:58:31 bitbake -e Oct 17 18:58:42 maybee someone overwrites it Oct 17 19:03:40 hm problem found Oct 17 19:05:01 damn, found it Oct 17 19:05:09 good Oct 17 19:05:24 hmm Oct 17 19:05:27 I need to install every .so lib to /usr/lib to get .debug Oct 17 19:05:31 so I think it WAS a hardware problem Oct 17 19:05:37 cause I can read USB serial fine now Oct 17 19:05:48 I think maybe loose connections for power on my USB hub Oct 17 19:05:52 or something Oct 17 19:05:53 ynezz? Oct 17 19:06:08 ynezz where did you the other install? Oct 17 19:06:16 if it's installed directly to /usr/lib say it's in /usr/lib/local/lua/5.1/ Oct 17 19:06:28 it's not Oct 17 19:06:33 local? Oct 17 19:06:46 where the hell is local comes from? Oct 17 19:06:50 we dont use it Oct 17 19:06:51 lua-lanes-2.0.6-r0/package/usr/local/lib/lua/5.1/lua51-lanes.so Oct 17 19:06:59 dead dead Oct 17 19:07:08 I dont think we HAVE a local Oct 17 19:07:09 *sigh* Oct 17 19:07:11 but Lua use it for libs Oct 17 19:07:12 or I haven't seen one Oct 17 19:07:17 change lua Oct 17 19:07:22 its only a dir Oct 17 19:07:27 oh no :( Oct 17 19:07:30 usr/lib/lua Oct 17 19:07:35 not autoconf? Oct 17 19:07:35 and all packages... Oct 17 19:07:40 otherwise lua suckz Oct 17 19:07:44 Lua is crap in this, yep Oct 17 19:07:46 what is /usr/local supposed to be for anyway? Oct 17 19:07:57 aditya local stuff the user is doing Oct 17 19:08:06 without interferring packaged stuff Oct 17 19:08:13 ah Oct 17 19:08:16 damn Oct 17 19:08:22 ~curse Lua Oct 17 19:08:22 May you be reincarnated as a Windows XP administrator, Lua ! Oct 17 19:08:23 so if you personal install qt again Oct 17 19:08:31 you wouldnt do it under /usr/lib Oct 17 19:08:32 I serioulsy have a love/hate relationship with linux Oct 17 19:08:46 and you have cool stuff like stow Oct 17 19:08:50 how much ever I learn, there is still like mounds of shit I have no clue about Oct 17 19:08:54 which dont pollutes you fs Oct 17 19:09:14 there is mounts of shit I dont know in windows Oct 17 19:13:56 hi eFfeM_work Oct 17 19:18:07 hi woglinde Oct 17 19:18:11 does that sounds right? Oct 17 19:18:13 cp -fR ${S}/arch/arm/include/asm/* $kerneldir/include/asm/ Oct 17 19:18:52 besides the missing {} Oct 17 19:19:10 I wondered because the kernel staging changed recently Oct 17 19:19:22 sure Oct 17 19:19:37 bescaue we didnt wanted to be spamed Oct 17 19:19:44 from insane.bbclass Oct 17 19:19:46 I know Oct 17 19:19:59 it took 20min to stage Oct 17 19:20:01 for me Oct 17 19:20:04 now it's fast Oct 17 19:20:22 like for instance I wanted to go out and power off the computer, and end up waiting 20min Oct 17 19:20:55 anyway if the change is ok I'll push it Oct 17 19:21:19 I'll pastebin the change Oct 17 19:21:30 http://pastebin.com/FYaJaGN2 Oct 17 19:22:51 that's the change I'm talking about (thanks eFfeM_work for finding the rev) : http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=b6cc4bc217e695801db060a1366915a870cab5db Oct 17 19:24:21 ? Oct 17 19:24:46 basically the staging method of kenrel changed Oct 17 19:25:13 so I wondered if that line needed to be converted somehow Oct 17 19:25:16 cp -fR ${S}/arch/arm/include/asm/* $kerneldir/include/asm/ Oct 17 19:25:25 in the linux-bug recipe Oct 17 19:27:19 I bet eFfeM_work knew the response but he's not there Oct 17 19:30:35 khem: i saw some old chat logs from #oe where you were mentioning issues on arch linux where recipes would include -L/usr/lib in the linker flags (from libtool) and it would cause issues because it would use the native libgcc_s.so instead of the target one. i'm currently running into that issue - any ideas? Oct 17 19:31:07 hm its no arch linux support channel Oct 17 19:31:15 hi bluelightning Oct 17 19:32:03 woglinde: yes, but this seems to be more of an OE issue (some config setting is causing it to set -L/usr/lib Oct 17 19:32:35 vorporeal: well. which recipe is giving you the issue? Oct 17 19:32:40 dont use arch linux Oct 17 19:32:57 a few - freetype, libvisual, a couple others Oct 17 19:33:18 vorporeal: hello. Whats your distro ? Oct 17 19:33:24 vorporeal: well, for freetype look into the config.log and figure out which test finds -L/usr/lib Oct 17 19:33:28 he khem Oct 17 19:33:32 arch linux x86_64 Oct 17 19:33:36 woglinde: hello . Oct 17 19:33:44 hi khem Oct 17 19:33:51 vorporeal: OEs target distro Oct 17 19:33:55 GNUtoo|laptop: hello Oct 17 19:34:06 ah, angstrom (for the pandora handheld) Oct 17 19:34:08 GNUtoo|laptop: any more info on cairo issue Oct 17 19:34:40 khem, no new infos yet but there is a patch from you that caused us some issues, let me find it Oct 17 19:34:50 a89e9c41f6ea2df686bcfb7c01b39b23685b4261. Oct 17 19:34:52 hi mr_nice|work Oct 17 19:34:53 oops Oct 17 19:34:56 again.... Oct 17 19:35:01 mrmoku, hi Oct 17 19:35:25 GNUtoo|laptop hmm Oct 17 19:35:48 GNUtoo|laptop: did you try what Koen suggested to paul's mail ? Oct 17 19:35:49 here's the message: libgee.inc: Use libtool macros from installed libtool Oct 17 19:35:55 I'll look Oct 17 19:36:08 I'm pushing fix atm Oct 17 19:36:37 ah ok Oct 17 19:36:44 03Martin Jansa  07master * rdf932a2769 10openembedded.git/recipes/libgee/libgee_0.6.0.bb: Oct 17 19:36:44 libgee_0.6.0: add gobject-introspection-native to DEPENDS again Oct 17 19:36:44 Signed-off-by: Martin Jansa Oct 17 19:36:50 thanks Oct 17 19:41:54 I'll push my linux-bug and android-rpc changes Oct 17 19:43:15 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r2d9d8d68c6 10openembedded.git/recipes/android/android-rpc_git.bb: Oct 17 19:43:15 android-rpc_git: fix libgee link failure Oct 17 19:43:15 without the fix we have: Oct 17 19:43:15 | .../sysroots/x86_64-linux/usr/armv4t/lib/gcc/arm-oe-linux-gnueabi/4.5.2/../../../../arm-oe-linux-gnueabi/bin/ld: cannot find -lgee Oct 17 19:43:15 | collect2: ld returned 1 exit status Oct 17 19:43:15 | make[3]: *** [fso-msm-gps-activator] Error 1 Oct 17 19:43:16 Signed-off-by: Denis 'GNUtoo' Carikli Oct 17 19:43:25 03Denis 'GNUtoo' Carikli  07org.openembedded.dev * r0fc54bae7b 10openembedded.git/recipes/linux/linux-bug_2.6.27.2.bb: (log message trimmed) Oct 17 19:43:25 linux-bug: fix installation Oct 17 19:43:25 without the fix it does that: Oct 17 19:43:25 | cp: target `.../work/bug-angstrom-linux-gnueabi/linux-bug-2.6.27.2+svnr10746-r32/image/kernel//include/asm/' is not a directory Oct 17 19:43:26 NOTE: package linux-bug-2.6.27.2+svnr10746-r32: task do_install: Failed Oct 17 19:43:26 Thanks to eFfeM on #oe for help: Oct 17 19:43:26 you copy to $kerneldir/include/asm but that is not guaranteed to exist, ust add an install -d (install -d is better than mkdir as it will also create inbetween dirs if needed) Oct 17 19:45:39 zecke: i couldn't find anything mentioning -L/usr/lib, although at the bottom of the log, it says libdir=/usr/lib Oct 17 19:45:47 although -L/usr/lib isn't part of LDFLAGS Oct 17 19:46:07 vorporeal: well, it must be there. do you see it in the Makefile? Oct 17 19:46:37 [dv@dvpc oe_sam_dv2]$ bitbake --interactive Oct 17 19:46:49 bitbake -i dont works Oct 17 19:46:52 at the moment Oct 17 19:46:59 ah Oct 17 19:47:04 how can I debug? Oct 17 19:47:09 -e Oct 17 19:47:12 -DDD -v Oct 17 19:47:26 depends what you want to debug Oct 17 19:47:33 it gives nothing. I see an error return code == 256, that's all Oct 17 19:48:03 stap is do_stage while building virtual/kernel Oct 17 19:48:28 which bitbake and which branch you are using? Oct 17 19:48:39 shell script that bitbake called to do an action returns success if I call it manually Oct 17 19:48:52 origin/stable/2009 Oct 17 19:48:54 zecke: it's not explicitly listed in the makefile either - LDFLAGS and libdir are the same as in config.log Oct 17 19:49:00 and it's not mentioned anywhere else Oct 17 19:49:06 dv use master Oct 17 19:50:03 vorporeal: well, so it is not in the link line passed to libtool? Oct 17 19:50:04 bitbake __version__ = "1.10.0" is too old too? Oct 17 19:50:11 vorporeal: in that case it must find a .la file somewhere Oct 17 19:50:22 for stable its to new Oct 17 19:50:29 use latest 1.8 Oct 17 19:50:35 which is in this branch Oct 17 19:51:58 zecke: here's the failing call (and bitbake error lines following it) - http://pastebin.com/q2EYhKDs Oct 17 19:52:06 libtool converts any -rpath commands to -L commands during relink Oct 17 19:52:15 the -L/usr/lib is near the bottom of the gcc linking call Oct 17 19:52:23 and if you have -rpath /usr/lib it will appear -L/usr/lib Oct 17 19:52:29 and that will cause problems Oct 17 19:52:39 khem: the only -rpath commands i've seen correctly reference the target directories, i haven't seen an "-rpath /usr/lib" Oct 17 19:52:40 I have fixed this behavior in libtool 2.4 Oct 17 19:53:11 vorporeal: whats inside .la also matters Oct 17 19:55:58 khem: well, it seems i'm using libtool 2.2.6b Oct 17 19:56:01 guess that could be a problem Oct 17 19:56:40 vorporeal: I dont know Oct 17 19:57:05 vorporeal: if you have time you could try using angstrom-2010 Oct 17 19:57:12 that uses libtool 2.4 Oct 17 19:57:25 khem you should change minimal too Oct 17 19:58:04 woglinde: I could now actually Oct 17 19:58:33 earlier main images were not building Oct 17 19:58:36 but now they do Oct 17 20:03:49 hm I will go reading Oct 17 20:03:52 so good nite Oct 17 20:03:59 nite Oct 17 20:04:00 woglinde: gn Oct 17 20:04:41 in workdir/src/lua51-lanes.so does have +x executable bit set Oct 17 20:05:10 then it's copied to workdir/image and workdir/package but without that +x bit set Oct 17 20:05:25 so the runstrip function never runs on that file Oct 17 20:06:11 -rwxr-xr-x 1 ynezz ynezz 103032 Oct 17 22:00 tmp/work/armv4t-angstrom-linux-uclibceabi/lua-lanes-2.0.6-r0/lanes-2.0.6/src/lua51-lanes.so Oct 17 20:07:19 I've modified the package.bbclass http://pastebin.com/KpmbEXp1 Oct 17 20:07:37 NOTE: XXXX2 file:/media/data/devel/oe/ts72xx.git/tmp/work/armv4t-angstrom-linux-uclibceabi/lua-lanes-2.0.6-r0/package/usr/lib/lua/5.1/lua51-lanes.so isxec:0 Oct 17 20:07:50 how to fix that? Oct 17 20:08:51 if not os.path.islink(file) and not os.path.isdir(file) and (isexec(file) or file.endswith('.so'): ? Oct 17 20:09:22 or use some package_something_prepend and chmod +x that file here? Oct 17 20:09:54 or is the root of the problem somewhere else? Oct 17 20:16:19 damn :) 188 install -m 644 $(_TARGET_SO) $(LUA_LIBDIR) Oct 17 20:20:43 which bitbake is better to use for production: 1.10 or 1.8 ? Oct 17 20:27:02 I use master and it's ok Oct 17 21:04:25 http://wiki.gudinna.com/770 ångstrom booting up on the Toshiba ac100 Oct 17 21:08:21 03Philip Balister  07org.openembedded.dev * r54edb574ee 10openembedded.git/recipes/gnuradio/ (gnuradio.inc gnuradio_git.bb): gnuradio: Update GNU Radio recipe. Clean up configuration settings. Oct 17 21:10:37 03Khem Raj  07master * r980a1ea8d5 10openembedded.git/recipes/openldap/openldap_2.4.23.bb: (log message trimmed) Oct 17 21:10:38 openldap_2.4.23.bb: Fix build when using libtool 2.4 Oct 17 21:10:38 * It uses a prebuilt version of libtool which does not know Oct 17 21:10:38 about sysroot and worse it does not use automake and autoconf in such Oct 17 21:10:38 a way that one could use autoreconf to regenerate stuff. So we Oct 17 21:10:38 regenerate the things needed for libtool macros. It should work Oct 17 21:10:39 with any version of libtool now. Oct 17 21:12:00 Crofton: hey, thanks for gnuradio update Oct 17 21:12:05 no problem Oct 17 21:12:07 03Philip Balister  07org.openembedded.dev * r7ecfe36f9f 10openembedded.git/recipes/gnuradio/ (8 files in 2 dirs): gnuradio: Remove old recipes. I promise to deal with all user fall out promptly. Oct 17 21:12:13 I'm just removing the old stuff Oct 17 21:12:18 Crofton: now you can use libtool 2.4 to build gnuradio Oct 17 21:12:22 I need people to beat on the current code :) Oct 17 21:12:25 yep Oct 17 21:12:26 I have fixed other recipes Oct 17 21:12:39 things should be better Oct 17 21:12:40 cool Oct 17 21:13:06 btw. did u get part that error you were getting at runtime Oct 17 21:13:25 I think there may be issues with some kernels Oct 17 21:13:31 hmm Oct 17 21:13:49 next weeks issue Oct 17 21:14:08 there is new release of gnuradio Oct 17 21:14:08 I also have a srctree recipe in an overlay for code dev Oct 17 21:14:09 did u see Oct 17 21:14:17 3.3? Oct 17 21:14:20 old :) Oct 17 21:14:55 yes 3.3 Oct 17 21:14:58 we need to get some NEON stuff upstreamed before picking a stable release again Oct 17 21:15:14 ok Oct 17 21:15:32 also have to tst the uhd interface to get away from the libusb-0.12 issues Oct 17 21:15:42 yeah thats nice Oct 17 21:15:50 the old usrp interface used private structs in the old library Oct 17 21:16:18 ok Oct 17 21:16:25 time to do a little yard work Oct 17 21:16:51 ok Oct 17 21:17:03 I thought you got rid of usb Oct 17 21:17:06 but its still used Oct 17 21:17:16 Crofton|work: which recipe is recommended Oct 17 21:17:20 right now it picks 3.2.1 Oct 17 21:17:42 new recipe is the git one Oct 17 21:17:48 oh cool Oct 17 21:18:12 hopefully there are not a bunch of people using the old one :) Oct 17 21:18:27 03Ulf Samuelsson  07master * r98d0eb19ca 10openembedded.git/recipes/linux/ (14 files in 12 dirs): Add AT91SAM9 linux-2.6.30 from www.linux4sam.org Oct 17 21:18:28 03Ulf Samuelsson  07master * r1b4882558c 10openembedded.git/conf/machine/ (9 files in 2 dirs): Use common files for AT91SAM9 configuration Oct 17 21:18:31 03Ulf Samuelsson  07master * rfee7520020 10openembedded.git/recipes/xorg-xserver/ (4 files in 3 dirs): Add X-Windows support for AT91SAM9G45 LCD controller Oct 17 21:18:32 03Ulf Samuelsson  07master * r9f2d379829 10openembedded.git/recipes/images/ (console-audio-image.bb x11-at91sam9-image.bb): Add X11 images with more multimedia support (mplayer etc.) Oct 17 21:18:46 03Ulf Samuelsson  07master * rffd3f97c6b 10openembedded.git/recipes/u-boot/ (15 files in 2 dirs): (log message trimmed) Oct 17 21:18:46 Add u-boot-2009.11 support for AT91SAM9 Oct 17 21:18:46 Add support for dataflash for SAM9G45/M10 Oct 17 21:18:46 Add support for SD-Card Oct 17 21:18:46 Add mmc/mux command Oct 17 21:18:46 Workaround for problem with UBOOT_VERSION beeing overwritten Oct 17 21:18:47 Add CRC for dataflash environment Oct 17 21:36:31 I am setting up a new build machine thats 64 bit and reading through some setup guides I have noticed some say in sysctl add vm.mmap_min_addr = 0 and some say nvm.mmap_min_addr = 0 , which would be the correct one ? Thanks Oct 17 21:36:52 oh forgot its ubuntu 10.10 Oct 17 21:44:57 you dont need any of those with 10.10 Oct 17 21:47:35 so no changes are needed to 10.10 on sysctl ? Oct 17 21:48:27 because I also tried this one vm.vdso_enabled = 0 and said I couldnt add it Oct 17 21:58:33 03Khem Raj  07master * rab83ea3d6e 10openembedded.git/recipes/openjade/ (3 files in 2 dirs): Oct 17 21:58:33 openjade_1.3.2.bb: Update configury to make it autreconf'able Oct 17 21:58:33 * With the autoconf and makefile patches now we can Oct 17 21:58:33 run autreconf and which means the appropriate libtool Oct 17 21:58:33 macros are generated Oct 17 21:58:33 Signed-off-by: Khem Raj Oct 17 23:01:46 bitbake world on the last testing branch has gotten to the end (presumably; its at 65k tasks) and now appears to have hung Oct 17 23:02:11 top shows the python process has been sitting there for some time now Oct 17 23:03:40 my bitbake is fully up to date Oct 17 23:04:03 any clues how to "break" the python process and find out what its doing? (i'm not a python geek) Oct 17 23:15:45 grg: hmm Oct 17 23:16:31 khem, strace shows its waiting to read on a pipe Oct 17 23:17:10 and looking at other processes, the world have not have come to an end. A recipe build may have hung... investigating further Oct 17 23:17:11 hmmm its possible Oct 17 23:17:40 it should use subprocess to communicate thru pipe Oct 17 23:19:13 i see a few 'tee' processes still sitting on the log.do_* files in lirc-0.8.6+cvs20100613-r0 Oct 17 23:20:15 /usr/X11R6/bin/xmodmap not found Oct 17 23:20:28 thats the reason x11-image is not coming up Oct 18 00:03:10 can someone else bitbake lirc and see if it hangs after do_package? Oct 18 00:07:13 grg: is it lirc package causing isues ? Oct 18 00:07:25 khem, i suspect so Oct 18 00:07:36 it hangs when i bitbake it Oct 18 00:08:15 i'm going to try cutting some stuff out of my local.conf to see if any bbclasses are responsible Oct 18 00:09:40 grg: ok let me try it Oct 18 00:10:17 i removed all of my random INHERITs and it still hangs Oct 18 00:10:45 just started bitbake lirc lets see Oct 18 00:11:41 run.do_package is empty Oct 18 00:11:48 grg: hmmm seems to have built ok here Oct 18 00:14:34 odd. I'm going to kick off a second bitbake world, from the same testing-next branch and compare the two Oct 18 00:16:10 no no Oct 18 00:16:13 too early Oct 18 00:16:20 tossed the hot potato Oct 18 00:16:30 its doing do_package_stage Oct 18 00:16:34 and seems to have hung Oct 18 00:16:46 :) Oct 18 00:16:51 i'm not insane Oct 18 00:17:58 seems not Oct 18 00:19:35 recipe seems simple Oct 18 00:23:23 and it hasn't changed since it last built successfully Oct 18 00:25:24 hmmm... here's Frans' build from a couple of days ago: http://tinderbox.openembedded.net/packages/856272/ Oct 18 00:25:27 Hi Oct 18 00:25:31 r90? mine is r0! Oct 18 00:25:45 i made a xorg-image and i need a xorg.conf file Oct 18 00:26:05 my device is a1200 Oct 18 00:26:14 it need 240x320 resolution with 24bpp Oct 18 00:29:48 any help? Oct 18 00:30:41 angelo__, take a working xorg.conf from somewhere and edit it to your needs - with the xorg.conf manual page open Oct 18 00:31:11 all xorg.conf's that i get is of pc's and too larges Oct 18 00:31:32 gnutoo had give to me long time ago but i forge Oct 18 00:31:38 forget* Oct 18 00:32:23 you have any ? Oct 18 00:32:45 angelo__, i have one for my workstation Oct 18 00:33:30 can you give to me please? Oct 18 00:34:14 angelo__, http://pastebin.com/iff4Hwt2 Oct 18 00:34:24 thank you Oct 18 00:35:04 change the driver to "fb" and tweak the monitor and screen sections. I guess. Oct 18 00:35:17 nv dont need? Oct 18 00:35:48 my workstation is using the proprietary nv driver Oct 18 00:35:55 i suspect your device is not Oct 18 00:36:11 yes Oct 18 00:36:15 uses only fb Oct 18 00:40:30 grg: are you using libtool 2.4 Oct 18 00:40:43 khem, no not for this build Oct 18 00:40:53 ok Oct 18 00:40:55 mine is Oct 18 00:41:03 so its not related to libtool Oct 18 00:41:06 good to rule that one out :) Oct 18 00:41:15 do you have rm_work Oct 18 00:41:18 i do Oct 18 00:41:21 me too Oct 18 00:41:46 i commented out all of the INHERITs and tried to biuld again though. and it still failed Oct 18 00:42:18 could it needs something eternal that has been rm'd ? Oct 18 00:45:19 grg: which process is hung Oct 18 00:45:49 khem, i see do_package start, but never success or failure Oct 18 00:46:00 do_package ? Oct 18 00:46:04 hmmm Oct 18 00:46:08 * grg shrugs Oct 18 00:46:25 there are like 4 tee Oct 18 00:46:29 processes Oct 18 00:46:48 khem, yes. And the run.* files for those processes are all 0 bytes Oct 18 00:47:01 for all ? Oct 18 00:47:06 yep Oct 18 00:47:08 grg: worked here thanks Oct 18 00:47:09 hmm thats something Oct 18 02:56:04 grg: around ? Oct 18 02:56:12 khem, yep Oct 18 02:56:19 grg: what is 'p' attribute of a file any idea ? Oct 18 02:56:29 prw-r--r-- 1 kraj kraj 0 2010-10-17 19:13 lircd Oct 18 02:56:30 khem, pipe? Oct 18 02:56:45 named pipe i guess Oct 18 02:57:05 ok so I know where its choking Oct 18 02:58:14 grg@eye7:/tmp$ mkfifo /tmp/fifo Oct 18 02:58:15 grg@eye7:/tmp$ ls -l /tmp/fifo Oct 18 02:58:15 prw-r--r-- 1 grg grg 0 2010-10-18 13:25 /tmp/fifo| Oct 18 02:59:19 ok so we are using shutils copytree Oct 18 02:59:30 to copy image files to packages Oct 18 02:59:46 and it chokes if one of the file is a named pipe Oct 18 02:59:57 http://docs.python.org/dev/library/shutil.html **** ENDING LOGGING AT Mon Oct 18 02:59:57 2010