**** BEGIN LOGGING AT Thu Jan 05 02:59:57 2012 Jan 05 06:16:04 i am getting an error for oe_runmake install func , http://pastebin.com/3PmKf7j9 does anyone has any idea Jan 05 06:34:01 Is there a method of cleaning all tasks before it starts bitbake again from scratch Jan 05 07:03:14 cleansstate Jan 05 08:13:25 morning all Jan 05 09:53:50 hello Jan 05 09:54:13 khem ? Jan 05 09:55:36 I m building console-image with minimal-uclibc distro..... for samsung-soc-utils ..... I m getting error of fetch ...... svn: OPTIONS of 'http://svn.openmoko.org/trunk/src/target/gpio': Could not read status line: Connection reset by peer (http://svn.openmoko.org) Jan 05 09:56:25 how can I fix this error..... or how can I discard samsumg-soc-utils Jan 05 10:24:21 Hi. I am trying to build java in the Yocto project (java-related recipes are ported from OE-core). Build fails with message about missing libstdc++. Seems that there is no package provided native libstdc++. Same problem is described here: http://gumstix.8.n6.nabble.com/Error-building-openjdk6-icetea6-native-td663441.html#a31305851. Any suggestions? Jan 05 10:26:54 jekhor: are you using meta-java ? Jan 05 10:27:30 as linked from http://www.openembedded.org/wiki/LayerIndex Jan 05 10:28:29 hmm... No. Packages were picked from OE core only. Jan 05 10:30:12 Do you recommend using of meta-java? Jan 05 10:58:39 jekhor: given that it's maintained and I'm sure people have had success with it, yes Jan 05 10:58:58 OK, will try Jan 05 10:59:13 jekhor: when you say OE-Core you mean OE-Classic? AFAIK we don't have java in OE-Core Jan 05 10:59:48 probably :) I mean git://git.openembedded.org/openembedded Jan 05 11:00:06 right, that's OE-Classic Jan 05 11:00:34 hmm. I need to learn more about new trends in OE :) Jan 05 12:24:17 Same error... Jan 05 12:26:43 hmm, woglinde's not about, I guess he is still on holiday Jan 05 12:26:55 jekhor: if he was around I'm sure he'd be able to help you Jan 05 12:27:48 What package should provide native libstdc++? Jan 05 12:35:52 if i change the uclibc version during the bitbake process for a particular package them will it affect the gcc cross compiler ? Jan 05 12:37:22 erwt: yes, the toolchain will be rebuilt Jan 05 12:37:33 fuck Jan 05 12:37:43 sorry Jan 05 12:38:00 i made that mistake , Jan 05 12:38:01 np, I'm European :) Jan 05 12:38:14 jekhor: native? you mean on your host machine or for the target device? Jan 05 12:38:34 for the host machine it's whatever your distro provides; for the target device the package is calleed Jan 05 12:38:38 libstdc++ Jan 05 12:38:38 host machine. It needed for icedtea-native Jan 05 12:38:49 jekhor: depends on your distro then Jan 05 12:38:53 likewise, one command for rebuilding all packages with my new uclibc version Jan 05 12:39:15 erwt: is that a question? Jan 05 12:39:31 ya if u know it Jan 05 12:39:32 * bluelightning -> office, brb Jan 05 12:40:00 erwt: you normally bitbake , that will build all dependencies required for that image. Jan 05 12:40:24 erwt: in other words, you define an image which has "all packages" Jan 05 12:40:40 erwt: and there used to be "world" (really all packages), not sure if that exists anymore. Jan 05 12:41:39 ok but when i started i got this beautiful error checking whether the C compiler works... no Jan 05 12:42:04 it fails on configuring Jan 05 12:48:09 likewise, The oe must be keeping a track of all files bitbaked in the system ,eg stamps/arch..../*.ipk files ,is there anything such ? Jan 05 12:50:05 erwt: if you want to throw away everything and rebuild, you can throw away sstate* and tmp-*, and pseudodone Jan 05 12:50:39 likewise, ok Jan 05 13:04:20 03Richard Purdie  07master * rcff2c258b7 10bitbake.git/lib/bb/runqueue.py: Jan 05 13:04:20 bitbake/runqueue: Improve the setscene logging Jan 05 13:04:20 When debugging setscene problems it was found that some extra debug messages were useful. Jan 05 13:04:20 This patch adds them. Jan 05 13:04:20 Signed-off-by: Richard Purdie Jan 05 13:04:31 03Richard Purdie  07master * r4483dc4bc4 10bitbake.git/lib/bb/cooker.py: Jan 05 13:04:31 cooker.py: Convert a bb.data.expand refernce to the updated syntax Jan 05 13:04:31 Signed-off-by: Richard Purdie Jan 05 13:04:31 03Richard Purdie  07master * rbc954f7462 10bitbake.git/lib/bb/cooker.py: (log message trimmed) Jan 05 13:04:31 cooker.py: Add missing BB_WORKERCONTEXT conditionals Jan 05 13:04:31 Its desireable to run be able to execute bitbake tasks directly, only Jan 05 13:04:32 loading as much of the cooker is as necessary to do this. This use case Jan 05 13:04:32 can be identified by the BB_WORKERCONTEXT variable. There were a couple Jan 05 13:04:33 of places cooker code needed to be conditional on this but wasn't. This Jan 05 13:04:33 brings in the missing conditionals from poky where those code paths Jan 05 13:37:08 icedtea-native links libstdc++ statically, not dynamically. Resolved by installing libstdc++-static Jan 05 13:48:36 jekhor: that sounds like a dependency missing in icedtea-native Jan 05 13:48:51 jekhor: or a bug in our library depency magic Jan 05 14:36:19 hi Jan 05 14:46:18 i've got some problems building images for the netbook-pro (2.6.17 kernel and angstrom-2010.x), compile of images other than minimal-gpe-image fail with compilation of schroorc.c in schroedinger_1.0.9.bb, spewing out a load of 'expected before 'var4'' error messages, where is usually a ')' , ';' or '=' Jan 05 14:46:28 is there anything i can do about that? Jan 05 14:47:22 not sure if this is a bug or my own fault :D Jan 05 14:50:35 examples would be opie-image, angstrom-x-image , x11-image and gpe-image Jan 05 15:15:27 cs_nbp: no idea really but 2.6.17 and angstrom 2010 is probably not a combo that is often build as 2.6.17 is from 2006 Jan 05 15:16:06 might also be related to your C or C++ compiler (guess that one is old too) Jan 05 15:16:43 well the problem is 2.6.17 is the last kernel w a recipe for the netbook-pro (or im overseeing something) Jan 05 15:17:07 my gcc is almost up to date (ubuntu 10.04) Jan 05 15:17:49 target gcc, that is *not* the gcc on your build system Jan 05 15:18:04 yes sry i just realized -.- Jan 05 15:18:10 np Jan 05 15:18:15 target is old of course, yes Jan 05 15:18:44 so id have to make a kernel recipe for a newer kernel? Jan 05 15:26:13 cs_nbp: have a look at this http://lists.linuxtogo.org/pipermail/openembedded-devel/2011-April/031966.html Jan 05 15:26:17 eFfeM_work: hi Jan 05 15:26:28 hi ant Jan 05 15:26:31 doing ok? Jan 05 15:26:41 sort of ;) Jan 05 15:27:03 are you jumping on the oe-core wagon in 2012? Jan 05 15:28:42 cs_nbp: quick hack the orc-stdint.h Jan 05 15:29:58 how do i do that? seems to be an 'intentionally' empty file Jan 05 15:30:50 yes, just fill the missing types Jan 05 15:31:42 like that? http://code.google.com/p/ossbuild/source/browse/trunk/Shared/Build/Windows/Win32/include/orc-0.4/orc/orc-stdint.h?r=857 Jan 05 15:32:00 ah no thats not what u meant Jan 05 15:32:46 I'm kind of (but not really) a newb to C, could u clarify a bit please Jan 05 15:34:12 you see th eerror? chroorc.c:93:3: error: 'int16_t' undeclared (first use in this function) Jan 05 15:34:55 yes and some uint32 too Jan 05 15:36:05 so I need to declare what type eg int32_t should be ionn orc-stdint.h? Jan 05 15:36:18 no, just copy over the 9 lines Jan 05 15:36:24 on the empty file Jan 05 15:36:31 it will build Jan 05 15:36:55 waita sec, gotta untar and retar that Jan 05 15:37:26 btw, this is a recent breakage Jan 05 15:37:53 I mean, appeared when oe-classic was almost abandoned by devs Jan 05 15:38:22 I've seen this last week and had a déjà-vu Jan 05 15:40:47 so its a new feature^^ Jan 05 15:41:09 ^_^ Jan 05 15:42:41 ok editing and repacking done, bitbake started aaaand same errors o_O: schroorc.c:93:3: error: 'int16_t' undeclared (first use in this function) Jan 05 15:43:05 imma check if i packed it right -.- Jan 05 15:44:02 yes, I packed it right Jan 05 15:44:39 orc-stdint.h looks like in the ossbuild link Jan 05 15:47:15 cs_nbp: try this way: 1) bitbake -c clean schroedinger 2) bitbake -c patch schroedinger 3) hack the header 4) bitbake schroedinger Jan 05 15:47:33 ant_work: was not paying attention to irc, might be that a next project will be using oe-core Jan 05 15:48:35 be prepared to some roughness adding layers Jan 05 15:48:54 but is definitely a big improvement Jan 05 15:51:22 ant_work: yeah, well not too sure if we need additional layers (apart from a bsp one), Jan 05 15:51:34 it really depends on what my new project will be and on what hw Jan 05 15:51:44 that's my point, we need more hands on meta-oe Jan 05 15:52:11 srey but schroorc.c:93:3: error: 'int16_t' undeclared (first use in this function) Jan 05 15:52:39 cs_nbp: are you changing the header in the right place? Jan 05 15:53:19 has anyone seen this error before: Jan 05 15:53:20 https://gist.github.com/1565817 Jan 05 15:55:07 what im doing is: unpack orc-0.4.16.tar.gz, then edit orc-0.4.16/orc/orc-stdint.h, putting this http://code.google.com/p/ossbuild/source/browse/trunk/Shared/Build/Windows/Win32/include/orc-0.4/orc/orc-stdint.h?r=857 under the "intentionally empty" text, then tarring it back together and putting orc-0.4.16.tar.gz back into the sources folder Jan 05 15:55:41 cs_nbp: sorry I forgot to say thuis patch is for orc Jan 05 15:56:03 change my 1-3 points to orc Jan 05 16:00:01 FATAL: orc-0.4.16: http://code.entropywave.com/download/orc/orc-0.4.16.tar.gz cannot check archive integrity Jan 05 16:00:09 guess i have to edit some md5sum Jan 05 16:00:49 cs_nbp: sorry but here in office I don't have my OE Jan 05 16:01:07 np its got the same name as the file Jan 05 16:01:11 don't touch the recipes, it is not necessary Jan 05 16:01:32 it's just that orc recreates its types Jan 05 16:02:14 you don't need to go in depth, -c clean orc (the standard one in oe) Jan 05 16:03:49 k deleted all orcs incl (md5), ran bitbake -c clean and bitbake -c patch on orc Jan 05 16:03:54 and now just retry? Jan 05 16:04:09 ok, now check in workdir of orc and modify the header Jan 05 16:04:50 the one in tmp/work? Jan 05 16:04:55 yes Jan 05 16:07:31 ...building... Jan 05 16:07:38 suxezz Jan 05 16:07:40 !! Jan 05 16:07:46 txhalot ant Jan 05 16:08:09 think i just understood smtin Jan 05 16:09:48 :) Jan 05 16:10:18 maybe someone should send a patch for orc in oe-classic... Jan 05 16:13:56 bah now its screamin FATAL: bluez4-4.95: http://kernel.org/pub/linux/bluetooth/bluez-4.95.tar.gz cannot check archive integrity Jan 05 16:14:24 Expected MD5: '341294b2849a04a4afff5c96bfbf30b2' and Got: '3e6e77ae06aad73bbedbabed3f0d508e' Jan 05 16:14:42 he he Jan 05 16:15:01 that'svery recent breakage, actually discussed on both mailing list Jan 05 16:15:09 guess its anutha recent feature :D Jan 05 16:15:18 k imma try to take a look there Jan 05 16:15:38 as before, this comes from upstream...not our fault Jan 05 16:16:00 txh gawd for upstream -.- Jan 05 16:16:03 some genius decided to unpack and recompress the original tarballs Jan 05 16:16:22 yeh i like to do that alot too Jan 05 16:17:08 quoting kergoth "/me thinks maintainers should tar -cvO | gzip -n if they're going to use gzip ;)" Jan 05 16:17:32 yeah, those kernel.org lunatics. Jan 05 16:17:40 it does seem that no possible action is too crazy for them Jan 05 16:17:48 who gets handed an archive from upstream and decides to unpack and repack it? Jan 05 16:17:48 i dint knew there was anutha waz Jan 05 16:17:49 boggles the mind Jan 05 16:17:53 -z+y Jan 05 16:18:00 quite Jan 05 16:18:29 perhaps someone wanted to check the contents rofl\ Jan 05 16:25:28 best joke is i just stepped on my mini-bt stick and now dont need bluetooth anyways *whistle* Jan 05 16:25:58 k laterZ, gotta get some work done .,. Jan 05 16:54:28 My boss has told me I can spend $800-$1100 on a new build machine for working on my openembedded builds. does anyone have recommendations on a build machine? Jan 05 16:56:02 gchiii: yeah, get the cheapest from the nearest discounter and spend the rest on booze and girls :> Jan 05 16:56:33 lol Jan 05 16:56:38 bye i7 Jan 05 16:56:42 ups buy Jan 05 16:56:50 or i5 Jan 05 16:56:54 and a ssd Jan 05 16:57:00 for ccache Jan 05 16:57:06 fast drive speeds up things significantly Jan 05 16:57:18 and ram Jan 05 16:57:22 ram is cheap these days Jan 05 16:57:28 8 gig for 30 euros Jan 05 16:57:35 and a quad core at least :> Jan 05 16:59:06 building android is worser in disk and time usage than oe Jan 05 16:59:31 so quad core i7, ssd, 16 gig ram? Jan 05 16:59:43 yeah Jan 05 16:59:43 is it actually possible to build android using OE? that'd be cool I guess Jan 05 16:59:48 no Jan 05 17:00:14 till later Jan 05 17:02:03 so skip the AMD or Xeon options Jan 05 17:02:33 something like that sounds fairly reasonable. personally I would skip the ssd and spend my money on more RAM, though. Jan 05 17:03:02 we went through a phase of fitting SSDs to our build hosts here but it didn't seem to be worth the grief and we are not doing that anymore. Jan 05 17:04:00 want disk-io? go with JBOD or RAID0 Jan 05 17:04:38 well, yeah, though want to keep your data? go with raid10 I suppose. Jan 05 17:05:17 build stuff is transient :) Jan 05 17:05:30 yes, but still annoying to lose. Jan 05 17:06:24 I guess it depends how you work and how much you value your time, but if a failed disk might cause me to lose half a day's work then I would much rather spend a few dollars more on a redundant drive Jan 05 17:07:19 it's not so much that the data is irreplaceable, more a case of avoiding the downtime associated with swapping the disk, rebuilding the filesystem, and recreating whatever build artifacts I lose in the crash Jan 05 17:44:27 l8r Jan 05 17:55:56 does anyone here have experience whit how to put together a nBkProOs.img ? Jan 05 17:56:21 i#ve got the kernel, the initrd, the modules and mk-nBkProOsImg.c is compiled Jan 05 17:56:36 but it's input is already a raw image Jan 05 17:56:43 yes, though it must be nearly a decade since I did one of those Jan 05 17:56:43 heh Jan 05 17:56:55 how do I obtain that raw image? Is it simply the kernel zImage? Jan 05 17:57:13 talkin bout netbook-pro from psion btw Jan 05 17:57:20 iirc it is a slightly preprocessed version of the kernel image. florian might remember more. Jan 05 17:58:14 thanx I will try to contact him again :D Jan 05 17:59:00 I will have a quick look and see if I can find any of those bits in my old working trees Jan 05 18:01:34 that's really nice Jan 05 18:03:25 looks like its input is meant to be the kernel Image file, padded to 4MB and then with the initrd appended Jan 05 18:04:36 thanx. but what's 'padded'? I only know this from clothing (<- not a native speaker) Jan 05 18:05:21 4MB of zeroes in front? Jan 05 18:06:35 padded - some set of bytes before or after the data you care about Jan 05 18:07:24 ah k, so I could go w 4MB of 0's Jan 05 18:07:27 ? Jan 05 18:07:51 sorry posted the second half of my message to a different channel Jan 05 18:07:54 in this case, 4MB of "something" before the data really begins.. Jan 05 18:07:59 np;) Jan 05 18:09:42 hi all Jan 05 18:10:00 hi Jan 05 18:11:15 what would that 'something' be in my special case? perhaps someone can remember? Jan 05 19:21:16 ooh... oe-core requires now python 2.7 due to 3e9c368a8045044736fc7e348404060498c7491b? Jan 05 19:24:32 it's a bug.. it'll get fixed.. Jan 05 19:24:46 for now disable the QA warnings. Jan 05 19:25:47 (put the following in your local.conf to work around it) WARN_QA = "ldflags useless-rpaths rpaths" Jan 05 19:28:54 ok, thx Jan 05 19:29:27 Just to be clear.. python 2.6 is still the expected minimum version of python required Jan 05 19:29:42 (we just happened to have found a gap in our testing) :( Jan 05 19:56:37 I typed files:// instead of file:// when specifying a patch and bitbake went berserk Jan 05 19:56:46 http://paste.ubuntu.com/794124/ Jan 05 19:57:06 and hung Jan 05 20:00:13 hi khem Jan 05 20:04:39 woglinde: hi Jan 05 20:32:21 khem: probably the outstanding random hang on parse error issue with bitbake, I'm working on it now Jan 05 20:33:34 Khem, I found that gcc-runtime will ICE w/ master, if -feliminate-dwarf2-dups is added to the "FULL_OPTIMIZATION" flag.. :| Jan 05 20:33:48 http://bugzilla.pokylinux.org/show_bug.cgi?id=937 Jan 05 20:47:29 Is there a way to do something like "bitbake -c rm_work world" ? Jan 05 21:05:04 dinc++ -nostdlib++ -x c++-header -nostdinc++ -O2 -pipe -g -feliminate-unused-debug-types -feliminate-dwarf2-dups Jan 05 21:05:04 -fpermissive -I/home/mhatle/git/oss/poky/build-test/tmp/work/x86_64-poky-linux/gcc-runtime-4.6.2+svnr181430-r20/g Jan 05 21:05:04 cc-4_6-branch/build.x86_64-poky-linux.x86_64-poky-linux/libstdc++-v3/include/x86_64-poky-linux -I/home/mhatle/git Jan 05 21:05:04 -linux.x86_64-poky-linux/libstdc++-v3/include -I/home/mhatle/git/oss/poky/build-test/tmp/work-shared/gcc-4.6.2+sv Jan 05 21:05:31 sorry -- finger slipped.. Jan 05 21:27:38 woglinde: java only supports mipsel and not mips ? Jan 05 21:27:51 woglinde: i.e. big endian mips Jan 05 21:30:09 khem I never tried it on mips Jan 05 21:30:25 but I think cacao-vm should support both Jan 05 21:30:41 ok Jan 05 21:44:19 can you not put anonymous python functions into machine.conf files? Jan 05 21:50:07 msm: they're valid in .inc/.bb/.bbclass Jan 05 21:50:18 you can do a require/include of a .inc from a .conf Jan 05 21:51:22 kergoth: got it, thanks Jan 05 22:15:31 woglinde: I sent couple of patches for meta-java Jan 05 22:15:42 woglinde: with them now I can parse angstrom atleasr Jan 05 22:15:47 for qemumips Jan 05 22:23:01 hmm, I do not get it, I have an autotools based package, when I look at the config log it says configure: exit 0, at the same time bitbake tells me: Task xxx (..do_configure) failed with exit code '1' Jan 05 22:23:07 this does not really make any sense.. Jan 05 22:23:14 any ideas what I might be missing? Jan 05 22:25:48 did you look at the log and the run task? In tmp/work/.../package/temp Jan 05 22:26:04 it's possible that something in the run file will show you what is wrong Jan 05 22:26:53 I do not see anything suspicious there... Jan 05 22:28:02 also, if I manually run ../temp/do_configure.xxx and afterwards echo $? - the return code is 0 Jan 05 22:28:25 so I'm really puzzled... Jan 05 22:28:47 it could be a do_configure_append task or something? Jan 05 22:29:31 I had those and got rid of them, moved to do_compile_prepend Jan 05 22:29:45 could do_compile_prepend be treated same way as configure_append? Jan 05 22:29:50 well let me comment that stuff out just to be sure Jan 05 22:29:52 ya.. no idea then.. sorry Jan 05 22:32:26 nope, the do_compile_prepend task has nothing to do with it Jan 05 22:32:33 it still fails Jan 05 22:33:16 only other thing I can think of is to check the log for warning/error messages that might have come outside of the run file Jan 05 22:33:36 | configure: WARNING: unrecognized options: --disable-silent-rules Jan 05 22:33:39 or try again when you are sober :) Jan 05 22:34:03 khem: lol :) it's not me who got budget for an expensive PC ;) (not sure if you read the conversation earlier) Jan 05 22:34:25 is bitbake parsing the log and choking on the WARNING? Jan 05 22:35:17 i've not had a problem w/ that warning before.. Jan 05 22:35:20 Jin^eLD: yes I did Jan 05 22:35:26 gchiii: are you in US ? Jan 05 22:35:30 http://pastebin.mozilla.org/1435668 Jan 05 22:35:32 I've seen messages in the past where the config.log was scanned and errors caused based on contents Jan 05 22:35:52 gchiii: costco has a good system from zt systems Jan 05 22:35:54 buy that Jan 05 22:35:58 its cheap and powerful Jan 05 22:36:44 only thing I can suggest is to add a "echo $? Jan 05 22:36:53 'er.. echo $? as the last thing in your do_configure Jan 05 22:36:56 see what it says Jan 05 22:37:01 (hell see if it runs) Jan 05 22:37:11 http://www.costco.com/Browse/Product.aspx?prodid=11329189&whse=BC&topnav=&cm_sp=RichRelevance-_-categorypageHorizontalTop-_-TopRatedProductsInCategory&cm_vc=categorypageHorizontalTop|TopRatedProductsInCategory Jan 05 22:37:22 fray: when I run the do configure script manually, i.e. from workdir ../temp/run.do_configure - and echo then, it's 0 Jan 05 22:37:34 so I really don't know where bitbake is taking that 1 from Jan 05 22:37:55 drop it in your recipe and see.. thats hte only thing I can think of Jan 05 22:38:33 mhm ok Jan 05 22:38:38 I never had a problem like that Jan 05 22:38:51 Jin^eLD: bitbake -DDD Jan 05 22:38:58 khem: that one is running right now Jan 05 22:39:12 khem: and it did not reveal anything new Jan 05 22:39:32 do cleanall on the recipe and rebuild Jan 05 22:40:07 khem: did that a couple of times already :P Jan 05 22:40:21 which recipe is it Jan 05 22:40:38 btw its a recipe from git that inherits gitpkgv if that matters Jan 05 22:40:54 khem: one of my own recipes... it used cmake before, we switched to autotools Jan 05 22:41:27 paste it somewhere Jan 05 22:44:25 http://pastebin.mozilla.org/1435682 Jan 05 22:45:01 it's not really doing anything fancy, well maybe besides the AUTOREV thing.. Jan 05 22:45:05 but that should not affect configure? Jan 05 22:51:21 is bitbake trying to parse some logs afterwards or is it trying to be super smart or something? Jan 05 22:51:31 I should see that with -DDD though but there's nothing interesting there Jan 05 22:58:06 khem: any further clues? Jan 05 23:00:54 Jin^eLD: it looks overall ok Jan 05 23:01:02 Jin^eLD: does it fetch ok ? Jan 05 23:01:06 yes Jan 05 23:01:07 from gitorious ? Jan 05 23:01:13 I also tried with a non git package now, with a release tarball Jan 05 23:01:17 and I have the same problem there Jan 05 23:01:21 I have had problems with gitorious in past Jan 05 23:01:29 it tends to be notorious Jan 05 23:01:42 hmm Jan 05 23:01:53 is it on oe-core ? Jan 05 23:01:57 yes Jan 05 23:02:07 are you uptodate ? Jan 05 23:02:15 automake got upgraded Jan 05 23:02:20 and that can be nuisance Jan 05 23:02:40 but since you dont do funky stuff in recipe it may not be an issue Jan 05 23:02:54 oh hmm, no I'm not up to date, I'm on some 1 month old snapshot.. Jan 05 23:03:08 hmm Jan 05 23:03:16 other autoconf based stuff works though Jan 05 23:03:27 well you could try to update Jan 05 23:03:47 what somehow bothers me that there is no indication to what the error could be Jan 05 23:04:02 I mean, if I knew it's a problem in the configure script or macro or bitbake itself I could check and fix it Jan 05 23:04:06 but configure returns 0 Jan 05 23:04:13 and bitbake things do_configure returns 1 Jan 05 23:04:30 running ../temp/run.do_configure manually (from workdir) also returns 0 Jan 05 23:04:52 so I somehow have doubts that it's autotools fault... Jan 05 23:05:04 has the autotools.bbclass been updated as well recently? Jan 05 23:05:17 maybe this is what I should try if it did.. let me check the git log Jan 05 23:05:23 how about you upgrade bitbake Jan 05 23:05:32 I remember having this issue once Jan 05 23:05:39 and in the end it was my fault Jan 05 23:05:44 thats why I dont remember Jan 05 23:05:46 I'm on 1.15.0 Jan 05 23:05:50 so I can commit it again Jan 05 23:05:58 use git master Jan 05 23:07:18 ok trying with bitbake master now.. Jan 05 23:09:26 git master of bitbake did not help Jan 05 23:12:53 ha! Jan 05 23:12:56 I think I finally found it Jan 05 23:13:08 would be worth rechecking with older bitbake, if I just did not notice Jan 05 23:13:31 http://pastebin.mozilla.org/1435723 Jan 05 23:13:51 ERROR: This autoconf log indicates errors <- this is printed inmid of the log somewhere Jan 05 23:14:40 so mystery solved... just bitbake's error reporting was "hidden" Jan 05 23:14:48 for no idea what reason Jan 05 23:17:05 would be also helpful if it printed the actual path it complains about :P Jan 06 00:53:33 somebody getting an error "tmp/work/armv5te-angstrom-linux-gnueabi/python-2.6.6-ml12.3/Python-2.6.6/python: cannot execute binary file" too when trying to build angstrom-x-image? Jan 06 01:33:06 k gnZ **** ENDING LOGGING AT Fri Jan 06 02:59:57 2012