**** BEGIN LOGGING AT Wed Dec 22 02:59:57 2010 Dec 22 04:04:47 grg: cool nice to know Dec 22 04:05:07 kergoth_: around ? Dec 22 04:05:22 kergoth_: I am seeing a issue which I think may relate to bb Dec 22 04:07:34 kergoth: for uclibc there is config.machine fragment of .config which is stored in uclibc/uclibc-git// dir I thought bb would prefer the file from $machine dir thats how it used to be but now with bb master I am seeing thats its picking it up from uclibc-git/$arch/ dir Dec 22 04:08:52 concrete example when I have MACHINE=qemuarm I expected it to use uclibc-git/qemuarm/uClibc.machine even thought there exists uclibc-git/armv5te/uClibc.machine Dec 22 04:09:03 but it picks up the latter Dec 22 05:12:28 hmmm FILESPATH has "/home/kraj/work/oe/openembedded/recipes/uclibc/uclibc-git/armv5te at beginning Dec 22 05:13:09 even before uclibc/uclibc-git/local Dec 22 05:16:22 hmm I see ${FEED_ARCH} appears to be the last in OVERRIDES Dec 22 05:16:26 thats why Dec 22 05:18:53 here is the culprit in sane-toolchain.inc OVERRIDES .= ":${FEED_ARCH}" Dec 22 07:29:40 good morning Dec 22 08:12:09 hi, anyone an idea how to change the font size for pidgin? I've upgraded from 11.2 to 11.3 and the font became much smaller, tried all kind of things but to no avail Dec 22 08:12:26 ah wrong chat (and gm) Dec 22 08:15:31 JaMa: cleaned and rebuild zlib, now my .la file is similar to yours but still fontconfig fails Dec 22 08:16:07 you probably have more broken .la files in sysroot Dec 22 08:16:51 eFfeM: probably freetype, expat or something similar Dec 22 08:18:08 it is still an error with libtool and libz Dec 22 08:18:16 (on phone) Dec 22 08:18:27 do you have /usr/lib/libz.la on your system? Dec 22 08:18:36 no Dec 22 08:18:41 hm ok Dec 22 08:18:54 please try clean build Dec 22 08:19:04 will do Dec 22 08:19:31 seems there is libtool stuff in fontconfig (e.g. ltman.sh) Dec 22 08:19:47 which version are you building? Dec 22 08:20:14 but I have built all ok (just for test when I was changing SRC_URI etc) Dec 22 08:20:38 grr git.kernel.org doesn't respond here :/ Dec 22 08:22:18 2.8.0 Dec 22 08:23:06 i'm rebuilding from scratch actually it seemed working for armv5, then i moved to armv7 and it failed, but rebuilding is needed anyway Dec 22 08:23:12 takes most of the day though :-( Dec 22 08:23:55 I'm building arm[4567] and built ok on all Dec 22 08:25:04 but I also bumped DISTRO_PR imediately after switch to force everyone to clean rebuild Dec 22 09:03:47 morning Dec 22 09:12:55 hi hrw Dec 22 09:41:31 03Eric BENARD  07org.openembedded.dev * r8e0e91d87f 10openembedded.git/recipes/asterisk/ (29 files in 6 dirs): Dec 22 09:41:31 asterisk: move non working recipes to nonworking directory Dec 22 09:41:31 Signed-off-by: Eric Bénard Dec 22 09:41:31 Acked-by: Khem Raj Dec 22 09:41:31 Acked-by: Frans Meulenbroeks Dec 22 09:41:44 03Eric BENARD  07org.openembedded.dev * r6b4e635e64 10openembedded.git/recipes/ (14 files in 2 dirs): Dec 22 09:41:44 asterisk: fix SRC_URI Dec 22 09:41:44 * keep the higher version which fetch and patch Dec 22 09:41:44 * fix SRC_URI for all the asterisk recipes Dec 22 09:41:45 Signed-off-by: Eric Bénard Dec 22 09:41:45 Acked-by: Khem Raj Dec 22 10:14:38 hi florian Dec 22 11:06:25 hiho Dec 22 11:06:41 hi Dec 22 11:08:10 mhm, is there some specific thread on the ML regards the split up into branches/layers ? have been away for ~1 month from OE and I think there have been a lot of changes ? ;) Dec 22 11:08:45 yes Dec 22 11:13:47 any hints on the layering? just found some threads in the ml but no real direction given from the tsc Dec 22 11:17:08 mlip, there is no real direction yet, actually the discussion on layering started (afaik) only a few days ago) Dec 22 11:18:46 eFfeM: thx, that may be a reason why I can't find anything specfic ;) Dec 22 11:19:11 mlip are you on the members ML ? Dec 22 11:22:46 ah nevermind, it was on the dev list Dec 22 11:23:14 mlip I feel this is the latest proposal (from hrw): http://article.gmane.org/gmane.comp.handhelds.openembedded/40762 Dec 22 11:23:18 i am on the casual dev lis Dec 22 11:23:37 already reading this atm Dec 22 12:07:15 jo jama Dec 22 12:12:10 jo woglinde Dec 22 12:14:31 * woglinde tries an android build Dec 22 12:26:50 something wrong with pusing to git now? http://paste.pocoo.org/show/308323/ Dec 22 12:27:47 rebase? Dec 22 12:28:56 hi jconnolly Dec 22 12:29:07 hello woglinde Dec 22 13:03:03 JaMa: rebuild is beyond fontconfig, it builds fine now Dec 22 13:03:23 but while peeking in sysroots I noticed the following issues: Dec 22 13:03:24 frans@linux-suse:~/oe/tmp_minimal/sysroots/armv7a-oe-linux-gnueabi/usr/lib> grep sysroot *.la Dec 22 13:03:25 libobjc.la:libdir='/home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/armv7a/arm-oe-linux-gnueabi/lib' Dec 22 13:03:25 libssp.la:libdir='/home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/armv7a/arm-oe-linux-gnueabi/lib' Dec 22 13:03:25 libssp_nonshared.la:libdir='/home/frans/oe/tmp_minimal/sysroots/i686-linux/usr/armv7a/arm-oe-linux-gnueabi/lib' Dec 22 13:03:48 no idea about hese recipes and who makes them but apparently they have their own libtool Dec 22 13:04:00 btw should this be a QA check ? Dec 22 13:04:15 they are all from gcc Dec 22 13:04:22 where we dont touch libtool Dec 22 13:11:07 woglinde: ah ok, thanks Dec 22 13:11:33 not sure if it harms, it just defines libdir and if eventually things end up at the right place I guess it is ok Dec 22 13:11:55 (but we could e.g. do a sed on the .la file in a do_install_append) Dec 22 13:59:44 jo ph5 Dec 22 14:01:46 hej woglinde Dec 22 14:06:54 hey guys .. since some time now , there is no staging directory anymore .. but i used a extened $PATH to it to get the compilers also used by hand .. how can i get this back ? Dec 22 14:16:13 rob_w, look in sysroot Dec 22 14:16:28 you will find the copilers there now Dec 22 14:21:18 thanks Crofton|work Dec 22 14:33:34 hi jay7 Dec 22 15:04:54 is crost posting between oe-devel and yocto OK for discussing multiple git repository support in a build? Dec 22 15:04:57 *cross Dec 22 15:08:37 hey does anyone have a recipe for protocol buffers from google? Dec 22 15:10:17 cbrake: something wrong with pusing to git now? http://paste.pocoo.org/show/308323/ Dec 22 15:12:17 JaMa: hmm, someone seems to keep eating this description file Dec 22 15:12:19 fixing ... Dec 22 15:15:30 cbrake, the number of lists is exploding Dec 22 15:15:36 not sure what to do :) Dec 22 15:26:25 kergoth_: wrt the python queue error i got yesterday; after this mornings fresh boot the problem is gone, no idea what it was but can't reproduce it for now Dec 22 15:26:43 weird. okay. let me know if it crops up again Dec 22 15:34:12 will do Dec 22 15:50:17 03Bernhard Reutner-Fischer  07master * r5cde120003 10bitbake.git/lib/bb/runqueue.py: Dec 22 15:50:17 runqueue: fix typo in documentation of calculate_task_weights Dec 22 15:50:17 Signed-off-by: Bernhard Reutner-Fischer Dec 22 15:59:59 JaMa|Off: ok, should be fixed now Dec 22 16:16:14 03Cliff Brake  07master * r97fe8e5853 10openembedded.git/README: README: test commit Dec 22 16:37:10 hmm, if i have multiple threads and do a fork, i wonder how that affects things. do both threads stick around? i'd guess so Dec 22 16:37:23 * kergoth_ busts out advanced programming in the unix environment Dec 22 16:41:23 kergoth there is chapter about python? Dec 22 16:42:13 hehe, no, not that it matters Dec 22 16:43:12 03Martin Jansa  07master * rcdaf1640f8 10openembedded.git/recipes/xorg-doc/ (3 files): Dec 22 16:43:12 xorg-docs,xorg-sgml-doctools: upgrade from 1.5.99.901 to 1.6 Dec 22 16:43:12 Signed-off-by: Martin Jansa Dec 22 16:43:13 cbrake: yes, pushed, thanks Dec 22 16:43:22 03Martin Jansa  07master * re816f74b86 10openembedded.git/recipes/linux/ (4 files in 3 dirs): Dec 22 16:43:22 linux(-kexecboot): bump SRCREV a bit more to rc7 Dec 22 16:43:22 Signed-off-by: Martin Jansa Dec 22 16:51:18 kergoth_: only the calling thread is replicated in the child process, courtesy of "The Linux Programming Interface" Dec 22 16:51:36 ah, right, thanks. i was hoping that was the case Dec 22 16:51:55 * foerster loves this book, though I haven't actually had the chance to read it yet Dec 22 16:52:24 getting an odd 'dictionary size changed during iteration' while saving the codeparser cache from poky, even though it should in theory have only been used during the up front parse, or during the recipe parse, which would be in the worker process, which shouldn't affect the cache being saved.. Dec 22 16:53:24 hmm. Dec 22 16:53:29 I remember seeing that at one point Dec 22 16:53:31 * kergoth_ scratches head Dec 22 16:53:32 but we got it fixed somehow Dec 22 16:55:35 bitbake start has become very very long, is it normal? Dec 22 16:55:57 mckoan: what do you mean? Dec 22 16:56:05 and what bitbake? Dec 22 16:56:19 BB_VERSION = "1.10.0" Dec 22 16:56:20 METADATA_BRANCH = "master" Dec 22 16:56:40 I launch the 'usual' bitbake minimal-image and it takes ages Dec 22 16:57:21 does not appears any message indicating it is workingù Dec 22 16:57:41 'ages'? Dec 22 16:57:49 at most it should be a couple second delay while it loads the cache from disk Dec 22 16:58:28 kergoth_: yea it was what I was used to Dec 22 16:58:32 but not now Dec 22 17:00:05 MACHINE=spitz time bitbake minimal-image Dec 22 17:00:08 15" Dec 22 17:00:17 NOTE: Handling BitBake files: \ (7255/7255) [100 %] Dec 22 17:00:17 Parsing of 7255 .bb files complete (6833 cached, 422 parsed). 7620 targets, 336 skipped, 2 masked, 0 errors. Dec 22 17:00:22 anymore Dec 22 17:00:31 no prompt no messages Dec 22 17:01:00 mckoan do you have tinderbox activated? Dec 22 17:01:08 tinderbox is slow these days Dec 22 17:01:09 does it eat cpu cycles? Dec 22 17:01:12 disable it for now Dec 22 17:01:22 htop shows no significant activity Dec 22 17:01:32 mckoan just try it Dec 22 17:01:39 disbale tinderbox support Dec 22 17:02:11 woglinde: doing Dec 22 17:02:32 * mwester just assumed tinderbox updates were done by a background thread... Dec 22 17:02:56 mwester not the initial Dec 22 17:04:13 mckoan: i'd also be curious to hear what bitbake master does Dec 22 17:04:25 if it acts differently in this way or not. it shouldn't, if its tinderbox Dec 22 17:04:56 w/o tinderbox is ok Dec 22 17:05:01 heh Dec 22 17:05:03 thank you guys! :-) Dec 22 17:05:09 you should try master anyway, i need testers ;) Dec 22 17:05:30 kergoth_: master branch? Dec 22 17:05:37 yeah Dec 22 17:05:37 i can't even fathom using 1.8 or 1.10 ever again -- too painfully slow :) Dec 22 17:05:43 foerster: agreed Dec 22 17:05:44 kergoth_: I'm on it ;-) Dec 22 17:06:08 * foerster hugs parallel parsing Dec 22 17:08:11 have a nice rest of the day Dec 22 17:11:47 okay, we need to fix this 'eol while scanning string literal' when machine isn't set Dec 22 17:11:49 have a nice xmas and xmas-ny week Dec 22 17:12:00 happy holidays hrw Dec 22 17:12:03 hrw thanks Dec 22 17:12:16 merry christmas to you and your family Dec 22 17:12:23 thx Dec 22 17:12:37 I will have english lesson in ~hour Dec 22 17:13:04 have to ask why 'we wish you a merry christmas' and not 'the merry christmas' Dec 22 17:13:25 hehe Dec 22 17:13:31 because english makes no sense Dec 22 17:13:49 as those are wishes for this particular christmas not for any Dec 22 17:14:00 kergoth_: ;D Dec 22 17:14:16 kergoth_: get mentor to sponsor your trip for fosdem Dec 22 17:16:08 when is it again? Dec 22 17:36:19 morning Dec 22 17:36:29 * Jay7 is back Dec 22 17:36:32 hey Jay7 Dec 22 17:36:38 kergoth_: 4 & 5 feb if I recall correctly Dec 22 17:36:43 sat+ sun Dec 22 17:39:46 https://github.com/kergoth/openembedded/compare/master...tasks-reduction - comments? Dec 22 17:40:00 avoids the need to build coreutils-native and stagemanager-native Dec 22 17:41:30 Ah, a move of stagemanager Dec 22 17:42:09 yep. it doesn't require a build, and we need it for just about anything, so makes sense to just drop it somewhere we can always run it directly Dec 22 17:44:37 Does anyone know how to build a kernel in openembedded's meta-toolchain cross compiler environment? Dec 22 17:44:47 bitbake virtual/kernel Dec 22 17:44:56 bitbake linux should work too Dec 22 17:45:04 *g* Dec 22 17:45:29 sorry, I should have been more clear Dec 22 17:45:32 not using bitbake Dec 22 17:45:48 I've installed the meta-toolchain cross-compiler Dec 22 17:45:57 so I just want to build my tree Dec 22 17:46:15 unfortunately the environment-setup script tramples on CPATH Dec 22 17:46:17 that's a kernel build question, you cross compile it the way you do with any crosscompiler from anyone Dec 22 17:46:22 use CROSS_COMPILE= on the make line Dec 22 17:46:29 which causes the host compiler to barf Dec 22 17:47:07 in particular, it seems that host utilities like fixdep and depmod are built against the target abi Dec 22 17:47:23 if I run make oldconfig before sourcing the cross-compiler environment Dec 22 17:47:38 so don't source it, set CROSS_COMPILE= and ARCH= (unless its a new enough kernel to select in the config) Dec 22 17:47:39 then start the build then things are fine until depmod Dec 22 17:48:02 hmm, sometimes it's the simple things you miss I guess Dec 22 17:48:06 let me give that a try Dec 22 17:48:16 make CROSS_COMPILE=/path/to/foo/bar/bin/arm-linux- ARCH=arm Dec 22 17:48:18 or whatever Dec 22 17:48:21 the compiler doesn't need any of the other environment provided by environment-setup? Dec 22 17:48:27 not for a kernel it doesnt. Dec 22 17:48:39 kernel only needs gcc tough Dec 22 17:48:41 the kernel takes care of its own copmiler options Dec 22 17:48:56 cool Dec 22 17:49:03 alright, well that's far simpler than I thought Dec 22 17:49:08 sorry about the noise Dec 22 17:49:11 np Dec 22 17:49:38 I'm starting a page on the wiki describing use of the cross-compiler environment Dec 22 18:16:54 http://www.openembedded.org/index.php/Cross_Compiler Dec 22 18:17:05 it would be nice if others could contribute what they know Dec 22 18:17:42 I suspect a note about the risk of compiling against incorrect library versions residing on the host would be nice Dec 22 18:17:56 but I don't know enough about this sort of thing to comment Dec 22 19:06:19 hi all Dec 22 19:11:37 03Chris Larson  07master * ra148e6a63c 10bitbake.git/lib/bb/utils.py: Dec 22 19:11:37 utils: show the actual exception in better_exec Dec 22 19:11:37 Signed-off-by: Chris Larson Dec 22 19:11:39 03Chris Larson  07master * ra4230811ff 10bitbake.git/lib/bb/data_smart.py: Dec 22 19:11:39 data_smart: handle None values in expandVarrefs Dec 22 19:11:39 Signed-off-by: Chris Larson Dec 22 19:11:40 03Chris Larson  07master * r457171b5b2 10bitbake.git/lib/bb/build.py: Dec 22 19:11:40 build: pass the correct filename to better_{compile,exec} Dec 22 19:11:40 Signed-off-by: Chris Larson Dec 22 19:11:41 03Chris Larson  07master * r6fb448a6cf 10bitbake.git/lib/bb/pysh/pyshyacc.py: Dec 22 19:11:41 pysh: add missing os.path import Dec 22 19:11:41 Signed-off-by: Chris Larson Dec 22 19:12:18 03Khem Raj  07master * re2f8a0db6b 10openembedded.git/conf/distro/include/sane-toolchain.inc: Dec 22 19:12:18 sane-toolchain.inc: Use BASE_PACKAGE_ARCH instead of FEED_ARCH and prepend to OVERRIDES Dec 22 19:12:18 Currently it ends up after local override and hence appears first in FILESPATH Dec 22 19:12:18 and for buiding uclibc_git.bb on qemuarm it picks wrong uClibc.machine file from Dec 22 19:12:18 uclibc-git/armv5te where as it should have picked from uclibc-git/qemuarm Dec 22 19:12:19 Signed-off-by: Khem Raj Dec 22 19:12:20 03Khem Raj  07master * r80317a3dcd 10openembedded.git/recipes/gcc/ (19 files in 2 dirs): Dec 22 19:12:20 gcc-4.5: Apply patches upto 2010.12 release of linaro gcc Dec 22 19:12:21 Signed-off-by: Khem Raj Dec 22 19:12:42 go go khem Dec 22 19:13:34 err forgot the acks Dec 22 19:13:38 but nevermind Dec 22 19:14:29 new gcc patches again Dec 22 19:15:20 woglinde_: yes Dec 22 19:15:41 woglinde_: should just work usually Dec 22 19:15:52 its tested on different platforms Dec 22 20:28:28 khem: I'm trying to build gcj 4.5. in my understanding, setting JAVA = ",java" should be enough. i did that in amend.inc, and i can see that my amend.inc gets picked up, because it also contains PR_append "-something". however, gcj doesn't get compiled. Dec 22 20:29:16 i can't find 'java' in gcc-4_5-branch/build.i686-linux.mipsel-oe-linux/config.status either Dec 22 20:29:39 config.status contains --enable-languages=c,c++,lto Dec 22 20:30:31 but i didn't add lto to LANGUAGES Dec 22 20:30:42 does this make any sense? Dec 22 20:30:55 setting JAVA = ",java" worked for gcc-4.4.4 Dec 22 20:41:52 hm... i think the problem is that JAVA_mipsel gets set to "" Dec 22 20:42:21 khem: i think we should remove all JAVA_foo ?= "" from gcc-configure-common.inc Dec 22 20:43:02 every gcc-4.x.y.inc does override the variable anyway Dec 22 20:59:48 kergoth: managed to get this one again: Dec 22 20:59:49 File "/usr/lib/python2.6/multiprocessing/queues.py", line 233, in _feed Dec 22 20:59:49 debug('feeder thread got sentinel -- exiting') Dec 22 20:59:55 but now I know how I got there Dec 22 21:00:08 I happened to hit ^C when it was at parsing recipes 0% Dec 22 21:00:31 because I made a mistake in my cmd line and wanted to abort and restart the cmd Dec 22 21:05:09 eFfeM: hmm, we did a lot of testing of ^C before parsing, during parsing, and after parsing, trying to find and fix things like that. will have to look into th at one Dec 22 21:05:18 hmm Dec 22 21:05:49 ^C doesn't make it to the server at all anymore. the UI gets it, then shuts down the server. maybe you sent it before the UI started..hmm Dec 22 21:06:29 btw I am using the cmd line version not the gui one Dec 22 21:09:24 obi why you want gcj instead of openjdk? Dec 22 21:10:38 kergoth: rechecking now if rebuilding the cache by touching local.conf will fix it Dec 22 21:11:33 it is also strange that yesterday it got resolved by a reboot, this seems to indicate some stale data in mem or /tmp that is cleared upon a reboot (or ignored) Dec 22 21:11:59 btw Dec 22 21:12:14 I've had some lzma and gtk+ failures on clean builds Dec 22 21:12:26 but it was about 4 days ago Dec 22 21:12:37 so I should recheck before reporting again Dec 22 21:17:02 eFfeM: do let me know if you find steps to reproduce, these sorts of things are a pain to nail down Dec 22 21:17:48 woglinde: because i want to create a native executable from java code Dec 22 21:18:23 obi hu? Dec 22 21:18:34 so you need gij Dec 22 21:18:36 okay Dec 22 21:19:01 kergoth: will do Dec 22 21:20:06 woglinde: we're using projectx (without the gui) to convert DVB to DVD Dec 22 21:20:20 ah projectx Dec 22 21:46:28 gcj seems to require alsa nowadays. i wonder how one could possibly integrate that into OE Dec 22 21:50:33 obi? Dec 22 21:50:40 alsa is there Dec 22 21:53:22 yes, it's there, but not before gcc-cross is built Dec 22 21:53:38 and gcj(-cross) is part of gcc-cross Dec 22 21:55:37 ah right Dec 22 21:55:46 hm Dec 22 22:16:33 03Chris Larson  07master * rc510e54c39 10bitbake.git/: Dec 22 22:16:33 Merge remote branch 'foerster/depexp-sorting' Dec 22 22:16:33 * foerster/depexp-sorting: Dec 22 22:16:33 depexp: add sorting of all package lists Dec 22 22:16:45 03Bob Foerster  07master * r80e3d3a130 10bitbake.git/lib/bb/ui/depexp.py: Dec 22 22:16:45 depexp: add sorting of all package lists Dec 22 22:16:45 Without sorting, it's very difficult to find the information you're Dec 22 22:16:45 looking for. Now, the lists are all sorted alphabetically for easy Dec 22 22:16:45 viewing. Dec 22 22:16:46 Signed-off-by: Bob Foerster Dec 22 22:21:52 calling it a day Dec 22 22:21:58 nite effem Dec 22 22:48:46 obi: gcc recipes gets clutterred very much. So I am all ears for cleanups Dec 22 22:48:50 propose them Dec 22 22:51:38 khem: what happened to that cleanup you proposed a while ago? I recall there was some consensus what is ok to clean, wasn't it? Dec 22 22:52:07 yes some sort Dec 22 22:52:11 no one voiced Dec 22 22:52:18 except Tartarus Dec 22 22:52:31 which meant the other option "plan in my head" gets implemented Dec 22 22:55:18 khem: will do Dec 22 22:55:55 khem: i've disabled alsa now, and gcj built a little further. but now it tried to link against the wrong zlib Dec 22 22:56:10 there's an option in our recipe: --with-system-zlib Dec 22 22:56:27 what is this supposed to do? Dec 22 22:56:37 in our case Dec 22 22:57:29 i assume the system's zlib will never exist before gcc-cross is built Dec 22 22:57:44 obi: java is a unfathomed terrain and even more on mips Dec 22 22:57:46 keep in mind Dec 22 22:58:09 --with-system-zlib means gcc will depend on the libz installed in rfs Dec 22 22:58:18 and not compile its own captive version Dec 22 22:59:54 i know. well... we've already used an older version of gcj for our distro successfully, but it has never really been integrated into our OE flavor Dec 22 23:00:25 khem it chooses wrong libz because we are not touching libtool in gcc build Dec 22 23:00:38 ups meant obi Dec 22 23:01:14 woglinde: I think we can not avoid that for long Dec 22 23:01:31 unfortunately that means we need to upgrade to 2.4 before upstream does Dec 22 23:01:41 and gcc libtool usage is not faint hearted Dec 22 23:01:49 woglinde: maybe, but - as with alsa - i don't have a cross-zlib at this stage ;) Dec 22 23:02:01 woglinde, sorry for bothering you, but do you remember if midpath has support for network-enabled applications (opera mini etc)? Dec 22 23:02:21 unless, of course, gcc would build one Dec 22 23:02:37 I managed to get phoneme + midpath working, it's a bit faster than jamvm, but both setups cannot connect to the internet Dec 22 23:02:44 but it sounds like --with-system-zlib is meant for the x86 executables Dec 22 23:02:57 b_lizzard check the configure log Dec 22 23:03:05 b_lizzard normaly it has netsupport Dec 22 23:03:12 OK... Dec 22 23:03:13 obi: we have libz recipe Dec 22 23:03:14 I can remeber running opera mini with it Dec 22 23:03:18 khem yes Dec 22 23:03:20 so alsa better depend on it Dec 22 23:03:24 if it needs it Dec 22 23:03:26 but not at gcc-cross stage Dec 22 23:03:31 thats the problem Dec 22 23:03:36 khem: gcj-cross depends on alsa and on libz Dec 22 23:03:42 seems we need seperate recipe for gcj Dec 22 23:03:57 on gcj alsa wht Dec 22 23:04:19 hm Dec 22 23:04:21 why would gcj need alsa I wonder Dec 22 23:04:27 it includes a copy of classpath, which implements java sound stuff Dec 22 23:05:04 well, it's possible to disable that and i already did that Dec 22 23:05:21 but now i still have to investigate how to solve the dependency on zlib Dec 22 23:07:17 ok then gcj needs to be part of gcc Dec 22 23:07:22 not gcc-cross I think Dec 22 23:07:47 gcc-cross is last stage of cross compiler bootstrap Dec 22 23:07:56 and no target packages exist before it Dec 22 23:08:02 woglinde, I think network support is enabled, but I get an "exception in network request thread", trying to run opera mini, whatever that means Dec 22 23:08:10 I'll just ask the developer dude Dec 22 23:08:25 guillaum1 Dec 22 23:08:40 khem: "gcc" creates target binaries, right? i need to create a cross-gcj Dec 22 23:08:52 cross-gcj Dec 22 23:08:55 hmmm Dec 22 23:09:07 then you have to spin a different rcipe Dec 22 23:10:22 or you can divide it up Dec 22 23:10:41 where you generate whatever it can along with cross-gcj along with cross-gcc Dec 22 23:10:56 and build other runtime as part of separate recipe to complement it Dec 22 23:11:01 and that would work too Dec 22 23:11:43 i'm trying without --with-system-zlib first Dec 22 23:13:18 hm? Dec 22 23:13:25 args Dec 22 23:13:32 cross-compiling Dec 22 23:13:41 hm hm Dec 22 23:13:43 wait Dec 22 23:13:53 it should zlib from native Dec 22 23:13:57 take Dec 22 23:16:04 yes, for x86 binaries. but for its java libraries (or whatever it is building) it shouldn't. with-system-zlib is also an option for libjava (inside gcc's source), so maybe this will work Dec 22 23:19:35 obi: you should disable that option Dec 22 23:34:28 did e2fsprogs get fixed yet or not Dec 22 23:34:33 http://tinderbox.openembedded.org/public/logs/task/14976522.txt Dec 22 23:45:14 i have this patch, which resolves a different issue with e2fsprogs: http://git.opendreambox.org/?p=obi/openembedded.git;a=commitdiff;h=d87d337f2d2693f3859de64a8b09c5b4b362abdd Dec 22 23:45:41 but i'm waiting for bernhard to test it Dec 23 00:00:00 the gcj-cross build succeeded now, but do_qa_staging reports a "failed sanity test (workdir)" in sysroots/i686-linux/usr/mipsel/lib/gcj-4.5.3-11/libjvm.la. shouldn't that location be covered by inherit autotools? other .la files get rewritten correctly. Dec 23 00:00:18 obi not for gcc-cross Dec 23 00:00:24 you have to do it manually Dec 23 00:02:17 http://www.google.com/search?source=ig&hl=en&rlz=&=&q=gcc+precompiled+headers&btnG=Google+Search&aq=f&oq= Dec 23 00:02:27 check out the advertisement Dec 23 00:02:46 hm. there are other .la files from gcc-cross, which are rewritten correctly Dec 23 00:03:08 but they are all in sysroots/mipsel-oe-linux/usr/lib/, not sysroots/i686-linux/usr/mipsel/lib/ Dec 23 00:10:01 good nite Dec 23 01:23:04 03Paul Eggleton  07master * rb5b1990612 10openembedded.git/recipes/libopie/ (libopie2/gcc45_opiemm_include.patch libopie2_1.2.5.bb): Dec 23 01:23:04 libopie2: fix compilation of 1.2.5 version with gcc 4.5 Dec 23 01:23:04 Adds some missing includes in owavrecorder.cpp. Dec 23 01:23:04 Signed-off-by: Paul Eggleton **** ENDING LOGGING AT Thu Dec 23 02:59:58 2010