**** BEGIN LOGGING AT Fri Aug 20 02:59:56 2010 Aug 20 04:55:04 finally for those watching: http://tinderbox.openembedded.net/builds/83632/ (beagleboard-demo-image building, minus the qt tools -- need to submit the patch) Aug 20 05:06:21 I think I need to turn on rm_work again... my build seems to have grown to >400GB! Aug 20 05:06:55 heh Aug 20 05:07:03 400GB ough Aug 20 05:07:06 ggdb3 Aug 20 05:07:10 indeed Aug 20 05:07:39 or maybe some simlink-cp bug too? Aug 20 05:09:10 ggdb3 is main bloat in my experience others are little less Aug 20 05:09:39 long time ago someone had an issue with git Aug 20 05:09:52 there was huge bloat in git-native Aug 20 05:10:03 some other people didn't have the same issue Aug 20 05:10:06 I keep all objects and an arm/x11-image takes around 30G Aug 20 05:10:13 thats using -g -O Aug 20 05:10:17 ok Aug 20 05:10:27 -ggdb3 include complete source code in dwarf2 ? Aug 20 05:11:07 it includes debug information for macros and so much more Aug 20 05:11:29 and if u peek into preprocessed files you will see what you are throwing at compiler Aug 20 05:11:40 sometimes the sourcefiles are very very huge Aug 20 05:11:45 after preprocessing Aug 20 05:11:56 tonnes of defines Aug 20 05:12:11 ok Aug 20 05:12:16 and now all will be having debug info records Aug 20 05:12:30 good to have if you dont care for size it occupies Aug 20 05:12:35 it has its own benefits Aug 20 05:12:36 I meant, for instance let's say someone has to debug on target but didn't build the sources Aug 20 05:12:50 for instance he downloaded SHR Aug 20 05:12:59 I think you still will need srcs Aug 20 05:13:05 ah ok,too bad Aug 20 05:13:22 but you will be able to step into macros :) Aug 20 05:13:26 ok Aug 20 05:13:56 thanks a lot for the infos Aug 20 05:14:48 no worries Aug 20 05:15:00 try a small example on your box with -ggdb3 Aug 20 05:15:12 a little helloworld.c Aug 20 05:15:35 and then rm -rf helloworld.c after compiling it with -ggdb3 Aug 20 05:15:41 load a.out in gdb Aug 20 05:15:49 and now ask for source list Aug 20 05:15:54 it will want the src file Aug 20 05:19:40 sleep time Aug 20 05:19:43 gn all Aug 20 05:34:47 ok Aug 20 05:37:46 03Roman I Khimov  07org.openembedded.dev * r03bc424d62 10openembedded.git/recipes/clamav/ (3 files in 2 dirs): Aug 20 05:37:46 clamav: update 0.96.1 to 0.96.2 Aug 20 05:37:46 * bug fixes mostly, considered as safe upgrade for 0.96.x line Aug 20 05:37:46 Signed-off-by: Roman I Khimov Aug 20 05:56:18 03Roman I Khimov  07org.openembedded.dev * r3a1681c2f7 10openembedded.git/recipes/squid/ (4 files in 2 dirs): Aug 20 05:56:18 squid: update 3.1.4 to 3.1.6 Aug 20 05:56:18 * bug fixes, considered as safe upgrade Aug 20 05:56:18 Signed-off-by: Roman I Khimov Aug 20 05:57:17 03Martin Jansa  07org.openembedded.dev * r398c1ea247 10openembedded.git/recipes/ (libexosip2/libexosip2_3.1.0.bb linphone/linphone_3.1.0.bb): linphone,libexosip2: bump PR for rebuild after openssl upgrade Aug 20 06:03:51 03Martin Jansa  07org.openembedded.dev * r1e78d7a70c 10openembedded.git/recipes/navit/ (navit-icons_svn.bb navit_svn.bb): Aug 20 06:03:51 navit(-icons): bump SRCREV a bit for latest fixes Aug 20 06:03:51 Signed-off-by: Martin Jansa Aug 20 06:07:01 hi JaMa|Wrk Aug 20 06:07:28 was the city search for navit fixed? Aug 20 06:07:49 gm Aug 20 06:08:02 hi Aug 20 06:11:58 GNUtoo|laptop: if you mean this, then yes http://trac.navit-project.org/ticket/575 Aug 20 06:38:27 03Dmitry Artamonow  07org.openembedded.dev * r78c84d68c9 10openembedded.git/recipes/jpeg/jbig2dec_0.11.bb: Aug 20 06:38:27 jbig2dec-0.11: new recipe Aug 20 06:38:27 jbig2dec is a decoder implementation of the JBIG2 image compression Aug 20 06:38:27 format. JBIG2 is designed for lossy or lossless encoding of monochrome Aug 20 06:38:27 images at moderately high resolution, in particular scanned paper documents. Aug 20 06:38:27 Signed-off-by: Dmitry Artamonow Aug 20 06:38:28 Signed-off-by: Martin Jansa Aug 20 06:38:40 03Dmitry Artamonow  07org.openembedded.dev * r6dc630a920 10openembedded.git/recipes/jpeg/ (openjpeg-1.3/fix_installdir.patch openjpeg_1.3.bb): Aug 20 06:38:40 openjpeg-1.3: new recipe Aug 20 06:38:40 The OpenJPEG library is an open-source JPEG 2000 codec written Aug 20 06:38:40 in C language. Aug 20 06:38:40 Signed-off-by: Dmitry Artamonow Aug 20 06:38:40 Signed-off-by: Martin Jansa Aug 20 06:38:41 03Michael Kurz  07org.openembedded.dev * r9231eedfb0 10openembedded.git/recipes/freesmartphone/fso-specs_git.bb: Aug 20 06:38:41 fso-specs_git.bb: sync PV with upstream Aug 20 06:38:42 Signed-off-by: Michael Kurz Aug 20 06:38:42 Signed-off-by: Martin Jansa Aug 20 06:38:43 03Michael Kurz  07org.openembedded.dev * rcd50f53c6d 10openembedded.git/recipes/freesmartphone/libfso-glib_git.bb: Aug 20 06:38:43 libfso-glib_git.bb: sync PV with upstream Aug 20 06:38:44 Signed-off-by: Michael Kurz Aug 20 06:38:44 Signed-off-by: Martin Jansa Aug 20 07:20:32 JaMa|Wrk, I've been looking at mupdf yesterday and today... i assume you have also? Aug 20 07:21:15 Dmity's patch against it for fixing graphical corruption didn't seem to work for me. But a dev version of mupdf from git works nicely Aug 20 07:23:16 03Martin Jansa  07org.openembedded.dev * r74f4939c8f 10openembedded.git/recipes/mupdf/mupdf_0.6.bb: Aug 20 07:23:16 mupdf: fix build on systems without all dependencies already installed (use -native stuff) Aug 20 07:23:16 Signed-off-by: Martin Jansa Aug 20 07:23:18 03Dmitry Artamonow  07org.openembedded.dev * rcf8d370464 10openembedded.git/recipes/mupdf/ (3 files in 2 dirs): Aug 20 07:23:18 mupdf-0.6: new recipe Aug 20 07:23:19 MuPDF is a slim and fast PDF viewer/toolkit. Aug 20 07:23:19 Signed-off-by: Dmitry Artamonow Aug 20 07:23:19 Signed-off-by: Martin Jansa Aug 20 07:23:25 grg: yup :) Aug 20 07:23:30 03Martin Jansa  07org.openembedded.dev * rd8cfef03be 10openembedded.git/recipes/tasks/task-shr-feed.bb: Aug 20 07:23:30 task-shr-feed: add mupdf and readd eve Aug 20 07:23:30 Signed-off-by: Martin Jansa Aug 20 07:23:44 grg: feel free to upgrade it a bit more if needed Aug 20 07:24:54 JaMa|Wrk, i'll take a look next week. Certainly it should be using build=release, not the default debug Aug 20 07:27:08 grg: ok, thanks Aug 20 07:27:42 it shits all over xpdf for speed though, even with debug build Aug 20 07:29:44 home time for me... Aug 20 07:29:46 bye Aug 20 07:35:10 bye Aug 20 07:43:13 pb_: ping Aug 20 07:44:03 03Martin Jansa  07org.openembedded.dev * raa5fae1ad0 10openembedded.git/recipes/xalan-j/xalan-j_2.7.1.bb: Aug 20 07:44:03 xalan-j: install and stage both jars, xalan and serializer Aug 20 07:44:03 * there is --with-xalan2-serializer-jar=${STAGING_DATADIR_JAVA}/serializer.jar Aug 20 07:44:03 in openjdk configure, but serializer was installed only to target Aug 20 07:44:03 sysroot Aug 20 07:44:03 Signed-off-by: Martin Jansa Aug 20 08:05:34 JaMa|Wrk: saw your mupdf patch for -native (http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=74f4939c8f85b4d2055a552661bfd984cd3862fd) Aug 20 08:05:57 is this really needed, I would expect the stuff it needs to come from staging dir Aug 20 08:06:58 what does e.g openjpeg-native deliver that is not staged by openjpeg Aug 20 08:07:30 btw and I guess a dependency on libpng is missing there is a depends on libpng-native but not on libpng, is that ok ? Aug 20 08:09:33 eFfeM_work: cmapdump and fontdump are build with native compiler and are linked against native libs Aug 20 08:09:51 eFfeM_work: that's why I had to add -native versions and then add those to DEPENDS Aug 20 08:09:57 ah ok, understood Aug 20 08:10:20 no problem with it just didn't understand and wanted to know Aug 20 08:10:24 what about libpng ? Aug 20 08:11:29 not sure if non-native stuff is also linked with libpng Aug 20 08:12:22 eFfeM_work: and it should IIRC be in jbig2dec DEPENDS instead Aug 20 08:13:04 ah ok, don't know all the details of all packages Aug 20 08:13:16 bitbake@jama ~/tmpdir-dev/sysroots/x86_64-linux/usr/lib $ ldd libjbig2dec.so linux-vdso.so.1 => (0x00007fffe65ff000) libpng12.so.0 => /OE/tmpdir-dev/sysroots/x86_64-linux/usr/lib/././libpng12.so.0 (0x00007f5b9e220000) Aug 20 08:18:02 03Martin Jansa  07org.openembedded.dev * rfe366ec1bb 10openembedded.git/recipes/pimlico/dates_0.4.8.bb: Aug 20 08:18:02 dates: add 0.4.8 version which compiles with newer gettext Aug 20 08:18:02 Signed-off-by: Martin Jansa Aug 20 08:47:49 hi florian Aug 20 08:48:37 good morning Aug 20 08:59:56 hm, trying to build with TARGET_OS set to linux-gnuspe, now native recipes call configure with: --build=x86_64-linux --host=x86_64-linux --target=x86_64-linux-gnuspe Aug 20 09:00:25 this does not seem to be ok should for native recipes target and host not be equivalent ? Aug 20 09:05:49 morning florian, got interesting post from Berlin: "Zuständigkeitsprüfung Körperschaftssteuer". Basically given the address changes (from Robert (berlin) to me (frankfurt)) they wonder whether Berlin is still the place where "the decisions are made" Aug 20 09:05:58 florian: what shall i respond? Aug 20 09:06:23 usually the place is where the office rooms are Aug 20 09:06:27 alas, we don't have an office :D Aug 20 09:13:01 mickey|office: heh... when did we send the address change? soemthing liek 8 or nine months ago... Aug 20 09:13:18 hehe Aug 20 09:13:24 mills of justice are sllllllllllllllow Aug 20 09:13:49 in fact taxes depertments are incredibly slow this year Aug 20 09:15:26 mickey|office: actually its up to them - the e.v. is registered in berlin and if they want to hand it over to frankfurt they will have to make an official statement Aug 20 09:16:19 hmm Aug 20 09:16:31 well, first i have to tell them where our main office is... Aug 20 09:16:50 can we just chose zecke's or Robert's home to not have to change anything? Aug 20 09:17:09 would be an option yes Aug 20 09:18:56 ok, i'll put up a question to the board mailing list and cc zecke and Robert Aug 20 09:19:07 okay great Aug 20 09:19:33 mickey|office: this reminds me that we might want to have the assembly at oedem this year :-) Aug 20 09:20:09 hmm Aug 20 09:20:12 is there going to be an OEDEM? Aug 20 09:20:18 i't already almost september Aug 20 09:22:17 we had a discussion on the mailinglist to have one together with elc europe Aug 20 09:22:54 that's end of october in cambridge Aug 20 09:23:19 hmm, right. that might be convenient for most of us Aug 20 09:23:29 not for me actually, from Frankfurt it's quite a trip to Cambridge Aug 20 09:24:03 no cheap flights to luton? :) Aug 20 09:24:14 no. only city or heathrow Aug 20 09:24:24 took me 8h last time from door to door Aug 20 09:24:31 ah nasty Aug 20 09:24:46 Frankfurt/Main doesn't serve luton anymore, which is annoying Aug 20 09:25:01 and Frankfurt/Hahn is like... miles away from here Aug 20 09:25:13 maybe its better... luton sucks ;) Aug 20 09:25:50 cbrake: did you add Simon yet or shall I? Aug 20 09:47:57 mickey|office: good morning Aug 20 09:49:30 florian: good morning Aug 20 09:50:27 hey pb_ Aug 20 09:51:22 mickey|office: maybe you could ride your favourite ice to köln/bonn airport. germanwings fly from there to stansted, which is only 30 mins from Cambridge. Aug 20 09:52:29 * florian thinks about using this one as well Aug 20 09:53:40 pb_: oh yes, that sounds like a good plan Aug 20 09:53:45 good morning, btw. Aug 20 09:54:26 * eFfeM_work does not know if he will ge going to elc, but if so going from ein to stansted, not sure whether to take car or train from there Aug 20 09:54:56 either car or train is equally easy. depends if you want to have the car at your disposal while you're in the uk, I guess. Aug 20 09:55:05 I wouldn't bother renting a car just to drive from the airport. Aug 20 09:55:44 depends on whether i need the car locally Aug 20 09:56:32 oedem might help in convincing mgmt to go to cambridge Aug 20 09:56:47 taht is to send me, not go there themselves :-) Aug 20 09:57:06 ah good, I was slightly worried for a moment there Aug 20 09:57:17 Aug 20 09:57:31 that reminds me though, must pick up my suit from the dry cleaners today Aug 20 09:58:46 i seem to recall from the minutes that oedem 2009 was not really convenient w/o car (unless you enjoyed walking) Aug 20 10:24:54 Hi eFfeM_work Aug 20 10:25:09 hi noor Aug 20 10:25:20 need some help Aug 20 10:26:34 speak up (i have 5 minutes before lunch, but there may be others as well who can help you) Aug 20 10:26:38 I have build uicmoc4-native successfully but now its do_compiles is giving me an error Aug 20 10:27:03 Noor: fahad_usman: btw I've pushed some of your recipes, didn't get to all of them though Aug 20 10:27:17 (too many things at hand) Aug 20 10:27:27 n file included from ssl/qsslcertificate.cpp:116: | ssl/qsslsocket_openssl_symbols_p.h:264: error: variable or field 'q_sk_free' declared void | ssl/qsslsocket_openssl_symbols_p.h:264: error: 'STACK' was not declared in this scope | ssl/qsslsocket_openssl_symbols_p.h:264: error: 'a' was not declared in this scope | ssl/qsslsocket_openssl_symbols_p.h:265: error: 'STACK' was not declared in this scope | ssl/qsslsocket_openssl_symbols_p.h:265 Aug 20 10:28:06 hm, you build the same recipe properly before ? Aug 20 10:28:34 yeah Aug 20 10:28:36 from what date is your tree and what openssl version are you using Aug 20 10:29:25 openssl version at the host? Aug 20 10:29:32 without going into detail: you might try to do a bitbake -cclean openssl-native and retry Aug 20 10:29:53 03Martin Jansa  07org.openembedded.dev * r5b1ad58d47 10openembedded.git/recipes/phoneme/ (3 files in 2 dirs): Aug 20 10:29:53 phoneme-advanced: add patch for build with (e)glibc-2.12 and fix QA issue with missing GNU_HASH Aug 20 10:29:53 Signed-off-by: Martin Jansa Aug 20 10:30:02 03Martin Jansa  07org.openembedded.dev * r108c9ae2e5 10openembedded.git/ (3 files in 2 dirs): Aug 20 10:30:02 xf86-input-evdev: add 2.5.0 RC1 with few mem leaks fixed Aug 20 10:30:02 Signed-off-by: Martin Jansa Aug 20 10:30:02 03Martin Jansa  07org.openembedded.dev * rfb43f19aec 10openembedded.git/recipes/tasks/task-shr-feed.bb: Aug 20 10:30:02 task-shr-feed: reaad phoneme-advanced-foundation Aug 20 10:30:02 Signed-off-by: Martin Jansa Aug 20 10:30:03 03Martin Jansa  07org.openembedded.dev * rf20c60f141 10openembedded.git/recipes/mupdf/mupdf_0.6.bb: Aug 20 10:30:06 mupdf: source archive name was changed, update SRC_URI Aug 20 10:30:06 * thanks soltys for reporting! Aug 20 10:30:06 Signed-off-by: Martin Jansa Aug 20 10:30:26 OK Aug 20 10:30:57 depends on what version you have of openssl and what you have now Aug 20 10:31:37 OK i'll check it out and let u know the result after rebuilding Aug 20 10:32:12 angstrom moved from 0.9.8m or so to 1.0.0 that might also be the reason, it coudl be that the 1.0.0 openssl does not export this symbol Aug 20 10:32:29 that is why I was asking what openssl-native version you were building Aug 20 10:33:39 I have 1.0.0 Aug 20 10:36:34 03Robert Schuster  07org.openembedded.dev * r2f99d4fff3 10openembedded.git/recipes/jdepend/ (jdepend-native_2.9.bb jdepend_2.9.bb): Aug 20 10:36:34 jdepend-native: Deleted. Aug 20 10:36:34 jdepend: BBCLASSEXTEND="native"-ified. Aug 20 10:36:36 03Robert Schuster  07org.openembedded.dev * r23fc60aa3b 10openembedded.git/recipes/junit/ (4 files): Aug 20 10:36:36 junit-native: Deleted. Aug 20 10:36:36 junit4-native: Deleted. Aug 20 10:36:36 junit: BBCLASSEXTEND="native"-ified. Aug 20 10:36:36 junit4: Dito. Aug 20 10:36:36 03Robert Schuster  07org.openembedded.dev * rf019b3ce8b 10openembedded.git/recipes/mozilla/ (rhino-native_1.7r1.bb rhino_1.7r1.bb): Aug 20 10:36:37 rhino-native: Deleted. Aug 20 10:36:37 rhino: BBCLASSEXTEND="native"-ified. Aug 20 10:36:38 03Robert Schuster  07org.openembedded.dev * rcd9cf593e7 10openembedded.git/recipes/servlet-api/ (servlet2.3-native_4.1.37.bb servlet2.3_4.1.37.bb): Aug 20 10:36:38 servlet2.3-native: Deleted. Aug 20 10:36:39 servlet2.3: BBCLASSEXTEND="native"-ified. Aug 20 10:36:39 03Robert Schuster  07org.openembedded.dev * rb940fff06a 10openembedded.git/recipes/bcel/ (bcel-native_5.2.bb bcel_5.2.bb): Aug 20 10:36:40 bcel-native: Deleted. Aug 20 10:37:55 anyone with patches to update tzcode/tzdata already? Aug 20 11:16:38 ao2, why not give it a try to make some Aug 20 11:24:08 eFfeM_work, done. It took a minute, I will push them as long as the build completes. I just asked to save some few seconds :) Aug 20 11:24:33 ah ok, np, Aug 20 11:24:48 thanks alot Aug 20 11:26:38 eFfeM_work: openssl does not built during building uicmoc4-native Aug 20 11:27:20 noor, strange, for the error message I thougt it should Aug 20 11:28:13 the ssl folder is present in qt-embedded-linux-opensource Aug 20 11:28:51 then qt-embedded probably has the dependency on openssl Aug 20 11:31:13 have you tried rebuilding openssl-native yourself ? Aug 20 11:32:59 Noor: the error you gave before is that the first one? no error of a missing .h file or so ? Aug 20 11:34:44 yes that is the first error Aug 20 11:39:07 Noor: I am out of ideas on this one, did the original recipe buiild before you touched it ? Aug 20 11:39:37 as far as I remember yes Aug 20 11:40:00 i build this packed converted to new style staging Aug 20 11:40:05 rebuilt it Aug 20 11:40:09 btw Noor and fahad_usman: if you encounter QA issues or get messages that files are not properly installed in a package, please also look into these or report them (that was the root cause of the sgmlspl problem I encountered) Aug 20 11:40:14 never mind Aug 20 11:40:21 i'll dig into it Aug 20 11:41:25 what did you identify in sgmlapl? Aug 20 11:42:06 the file sgmlspl or whatever the name was was not staged, it was not before you touched it either Aug 20 11:42:11 that caused iputils to fail Aug 20 11:42:37 fixed it yesterday, simple change, see the current recipe and compare with yours Aug 20 11:44:21 deploy/glibc/sources sure is light on content. Aug 20 11:44:36 * jkridner|work1 was thinking I could tar it up for GPL compliance. Aug 20 11:45:22 hi eFfeM_work: Aug 20 11:45:30 eFfeM_work: thanks for updating the recipes for me Aug 20 11:46:08 eFfeM_work: now i will update other recipes in this way and submit the patches Aug 20 11:49:37 hi fahad_usman Aug 20 11:50:10 appreciate your work, will try to look at the pending patches this weekend, let me know after the weekend if I forgot some Aug 20 11:50:27 eFfeM_work: okay, sure Aug 20 11:51:33 eFfeM_work: i will try to push a couple of more patches today Aug 20 11:52:26 great, thanks Aug 20 13:14:23 hi Aug 20 13:14:37 i cannot seem to build any native recipes Aug 20 13:14:56 s/native recipes/targat recipes/g Aug 20 13:15:21 native recipes are building fine Aug 20 13:16:13 i get error like Aug 20 13:16:15 ERROR: function do_compile failed ERROR: log data follows (/home/fahad/systemBuilder/oe-git/build/tmp/work/armv7a-angstrom-linux-gnueabi/makedevs-1.0.0-r8/temp/log.do_compile.23168) | Assembler messages: | Fatal error: Invalid -march= option: `armv7-a' NOTE: Task failed: /home/fahad/systemBuilder/oe-git/build/tmp/work/armv7a-angstrom-linux-gnueabi/makedevs-1.0.0-r8/temp/log.do_compile.23168 Aug 20 13:17:19 hi eFfeM_work Aug 20 13:17:27 fahad_usman: git grep armv7-a conf/ and also your local.conf Aug 20 13:17:28 hi Aug 20 13:18:34 conf/machine/include/tune-armv7.inc:# valid options for -march: `armv7', `armv7-a', `armv7-r', `armv7-m' conf/machine/include/tune-cortexa8.inc:TARGET_CC_ARCH = "-march=armv7-a -mtune=cortex-a8 -mfpu=neon ${ARM_FP_OPT}" conf/machine/include/tune-cortexa9.inc:TARGET_CC_ARCH = "-march=armv7-a -mtune=cortex-a9 -mfpu=neon -mfloat-abi=softfp" Aug 20 13:18:51 fahad_usman: are you using tune-cortexa[89].inc? which gcc version do you have? Aug 20 13:19:19 4.3.3 Aug 20 13:19:37 machine? distro ? Aug 20 13:20:18 DISTRO = "angstrom-2008.1" Aug 20 13:20:28 MACHINE = "beagleboard" Aug 20 13:20:39 you're not using bitbake -b , are you? Aug 20 13:21:22 this time i was running it in interactive mode Aug 20 13:21:40 fahad_usman: hmm 4.3.3 is new enought to support that march Aug 20 13:22:04 i guess you are using the wrong gcc Aug 20 13:22:07 eFfeM_work: but i think the behavior is the same without interactive mode Aug 20 13:22:26 eFfeM_work: and i am not using bitbake -b Aug 20 13:23:03 JaMa|Wrk: i haven't changed anything particular to gcc Aug 20 13:23:07 someone reported similar problem "[oe] configure error for xproto" Aug 20 13:23:31 is this for a specific recipe or for all recipes Aug 20 13:23:38 can you do a bitbake helloworld without error Aug 20 13:23:43 just from the cmd line? Aug 20 13:23:53 fahad_usman: just for confirmation run arm-angstrom-linux-gnueabi-gcc -v Aug 20 13:23:59 fahad_usman: from interactive mode Aug 20 13:26:24 eFfeM_work: no, helloworld also fails Aug 20 13:26:25 ERROR: function do_compile failed ERROR: log data follows (/home/fahad/systemBuilder/oe-git/build/tmp/work/armv7a-angstrom-linux-gnueabi/helloworld-1.0.0-r1/temp/log.do_compile.23396) | Assembler messages: | Fatal error: Invalid -march= option: `armv7-a' NOTE: Task failed: /home/fahad/systemBuilder/oe-git/build/tmp/work/armv7a-angstrom-linux-gnueabi/helloworld-1.0.0-r1/temp/log.do_compile.23396 ERROR: TaskFailed event exception, abo Aug 20 13:26:51 then please try the suggestion from JaMa|Wrk Aug 20 13:28:14 JaMa|Wrk: you mean build arm-angstrom-linux-gnueabi-gcc -v? Aug 20 13:28:55 no, in interacive mode give the command Aug 20 13:29:05 arm-angstrom-linux-gnueabi-gcc -v Aug 20 13:29:31 it says, its not a valid command Aug 20 13:29:32 arm-angstrom-linux-gnueabi-gcc should be already built as gcc-cross, but maybe for some reason you have wrong version in PATH Aug 20 13:31:40 so, how can i verify that? Aug 20 13:32:44 fahad_usman: check last command which was executed in log.do_compile.23396 Aug 20 13:34:04 eFfeM_work JaMa|Wrk arm-angstrom-linux-gnueabi-gcc -march=armv7-a -mtune=cortex-a8 -mfpu=neon -mfloat-abi=softfp -mthumb-interwork -mno-thumb -isystem/home/fahad/systemBuilder/oe-git/build/tmp/sysroots/armv7a-angstrom-linux-gnueabi/usr/include -fexpensive-optimizations -frename-registers -fomit-frame-pointer -O2 -ggdb3 -L/home/fahad/systemBuilder/oe-git/build/tmp/sysroots/armv7a-angstrom-linux-gnueabi/usr/lib -Wl,-rpath-link,/home/f Aug 20 13:35:27 than arm-angstrom-linux-gnueabi-gcc must exist Aug 20 13:35:51 if you're in interactive mode with environment set right Aug 20 13:38:24 is there a compiler if you do Aug 20 13:38:32 ls tmp/cross/*/bin Aug 20 13:39:01 e.g. does tmp/cross/*/bin/arm-angstrom-linux-gnueabi-gcc exist ? Aug 20 13:40:11 yes, it exist Aug 20 13:40:12 tmp/cross/armv7a/bin/arm-angstrom-linux-gnueabi-gcc Aug 20 13:40:49 eFfeM_work: and tmp/cross/armv7a/bin/arm-angstrom-linux-gnueabi-gcc -v gives me 4.3.3. Aug 20 13:41:04 ok Aug 20 13:42:11 fahad_usman: then you should probably remove tmp and rebuild from scratch as tmp/cross should be gone Aug 20 13:42:12 fahad_usman: what does tmp/cross/armv7a/bin/arm-angstrom-linux-gnueabi-gcc -dumpmachine say ? Aug 20 13:42:26 JaMa|Wrk: disagree Aug 20 13:42:31 fahad_usman: if you're using org.openembedded.dev Aug 20 13:43:27 OELAYOUT bump wasn't just for fun: http://git.openembedded.org/cgit.cgi/openembedded/commit/?id=ff701435705b6d0bb28a051e5e3c524b2badc834 Aug 20 13:43:34 ok, I eat my words Aug 20 13:43:52 i was checking in a tree I was building, but this was not git head Aug 20 13:44:13 hi,just to tell: Aug 20 13:44:32 fahad_usman: I agree with JaMa|Wrk but with an additional step: Aug 20 13:44:39 | No package 'eina' found Aug 20 13:44:41 in: Aug 20 13:44:48 do a git pull then rm tmp then rebuild from scratch Aug 20 13:44:57 libeflvala-1_0.5.0+svnr51235-r1 Aug 20 13:44:57 GNUtoo|laptop: wait for next EFL bump (already in my queue) Aug 20 13:45:03 the workarround is that one: Aug 20 13:45:11 s/eina/eina-0 Aug 20 13:45:19 in configure.ac: change eina to eina0 Aug 20 13:45:21 oops Aug 20 13:45:24 eina-0 Aug 20 13:45:26 I know Aug 20 13:45:28 ok Aug 20 13:45:32 so you are aware of it Aug 20 13:45:32 but eina is newer.. that's why it was changed (and fails for you because of libeflvala autorev Aug 20 13:45:34 so fine.... Aug 20 13:45:39 ok Aug 20 13:45:41 I'll wait then Aug 20 13:45:49 I really need to talk to mickeyl Aug 20 13:45:57 I'll go in #openmoko-cdevel Aug 20 13:46:28 GNUtoo|laptop: you can merge my branch (efl changes) now, or wait till I push it (hopefully today night or tomorrow - if eve works ok) Aug 20 13:46:36 I'll wait Aug 20 13:46:37 eFfeM_work: i just did git pull few hours back Aug 20 13:46:39 I'm on 3g.... Aug 20 13:47:23 fahad_usman: then you can skip that, but I do suggest a full rebuild from scratch (so after revming tmpdir) Aug 20 13:47:49 and i also cleaned gcc, gcc-cross, gcc-cross-intermediate and gcc-cross-initial Aug 20 13:48:02 okay, thanks Aug 20 13:48:11 hm that is very strange,\ Aug 20 13:48:30 and you pulled from dev head (and not e.g. from the stable branch) Aug 20 13:48:33 eFfeM_work: thanks, but i must leave now Aug 20 13:48:40 ok, cya Aug 20 13:48:49 eFfeM_work its end of day here, thanks for your help Aug 20 13:49:05 ok, i suggest to rm tmp and kickoff a new build before you leave :-) Aug 20 13:49:09 have a nice weekend Aug 20 13:49:19 eFfeM_work and sorry, because i got stuck in this, i couldn't push any changes Aug 20 13:49:26 np Aug 20 13:49:31 monday there's another day Aug 20 13:49:31 thanks, same to you Aug 20 13:50:01 eFfeM_work, ya, that's right Aug 20 14:05:17 morning Aug 20 14:35:23 kergoth: RP: (and ofc anyone else knowledgeable on our toolchain): could you please look at the following patches from me and ack/nak them: Aug 20 14:35:24 http://patchwork.openembedded.org/patch/2635/ Aug 20 14:35:46 http://patchwork.openembedded.org/patch/2649/ Aug 20 14:36:11 http://patchwork.openembedded.org/patch/2725/ (this one is only sent 15 mins or so ago) Aug 20 14:36:35 but the first two are there for a while; would like to close them either as applied or rejected Aug 20 14:37:55 eFfeM_work: the last if in 2649 should IMHO stay there Aug 20 14:38:12 eFfeM_work: if you don't have checksums at all it's better to show it there Aug 20 14:39:05 eFfeM_work: but maybe there is similar check + error message somewhere else IIRC Aug 20 14:39:16 yes there is,was looking for it Aug 20 14:39:26 ok good Aug 20 14:39:31 eFfeM_work: I'm ok with the first two but I've hardly reviewed them enough for an acked-by Aug 20 14:39:37 That last one looks wrong Aug 20 14:40:09 RP, I can wait a while with them, np, thougth they were forgotten Aug 20 14:41:03 JaMa|Wrk: there is still an error message in base_chk_file_checksum, that is the code that spits out the "Your checksums" message Aug 20 14:41:59 RP: wrt the 3rd one: I'm open to other suggestions; this got it working for me, but I was not too sure about the fix, hence the long explanation Aug 20 14:42:13 I hope the issue is clear Aug 20 14:42:59 but it could well be I fixed the wrong thing and only fixed the symptoms, not the root cause Aug 20 14:43:18 eFfeM_work: I'm pretty sure there is a deeper problem Aug 20 14:43:38 could well be, I've dug into it and could not find the root cause. Aug 20 14:44:50 the issue is if I set the TARGET_OS in my local.conf I get this, if I move to angstrom/eglibc there is also code that sets TARGET_OS to linux-gnuspe in eglibc.inc and there it works (and the dir does not get the -gnuspe suffix) Aug 20 14:45:17 eFfeM_work: The darwin hacks in gmp-native.inc look suspect to me Aug 20 14:45:44 RP, agree, saw them too, guess they had a similar problem Aug 20 14:46:18 btw my build host is ubuntu 10.04, not darwin, so that does not apply to me (but didn't introduce a similar hack) Aug 20 14:49:38 eFfeM_work: I don't know what the problem is exactly but you need to look deeper Aug 20 14:49:51 Adding things like that to native.bbclass will just cause pain in the long run Aug 20 14:50:00 eFfeM_work: What distro are you building for where ppc e500 didn't just work? Aug 20 14:50:09 sane-toolchain.inc should be setting TARGET_OS right (it does here) Aug 20 14:50:38 RP, the problem is that configure is called with an arg to --target that is not equal to --host and gmp-native didn't like that Aug 20 14:51:01 is this the thing you sent a patch for to the list? Aug 20 14:51:05 Tartarus: distro is angstrom, hw is a board we're developing (based upon calamari) Aug 20 14:51:07 yes Aug 20 14:51:10 pb_: yes Aug 20 14:51:50 if so, the proximate cause is that your _local override is clobbering TARGET_OS after native.bbclass sets it. the real fix would be to stop that from happening. Aug 20 14:52:12 if you just assign TARGET_SYS directly then you will simply create a parallel class of problems where TARGET_OS gets ignored when it shouldn't. Aug 20 14:52:18 If you're setting TARGET_OS_local = "" you're asking for trouble... Aug 20 14:52:39 indeed, that is exactly what he is doing. Aug 20 14:52:52 i would like to avoid it... what I need is done in angstrom.inc as follows: Aug 20 14:53:26 sorry meant angstrom-eglibc.inc Aug 20 14:53:44 TARGET_OS = "linux" Aug 20 14:53:44 TARGET_OS .= "${@['','-gnuspe'][bb.data.getVar('TARGET_ARCH',d,1) in ['ppc', 'powerpc'] and bb.data.getVar('MACHINE',d) in ['calamari']]}" Aug 20 14:54:21 i'm in the same league as this code, I also have an e500 ppc Aug 20 14:54:24 WHy isn't the machine just setting TARGET_OS correctly? Aug 20 14:54:44 and actually if I replace calamari with my machine name it also works Aug 20 14:54:59 RP, tried that too, same problem with configure Aug 20 14:55:16 thougth it was being overridden, that's why i did the local.conf thing Aug 20 14:56:41 the other native recipes seemed ok, it was gmp-native that complained, but at that point it already did a lot of other native recipes Aug 20 14:57:00 ah well, i think I'll dig into this next week with a fresh mind Aug 20 14:57:15 calling it weekend (although I might show up tonight from home) Aug 20 14:57:39 thanks for your suggestions and if you happen to have an idea please leave a message Aug 20 15:01:38 Fix angstrom to use sane-toolchain.inc's TARGET_OS magic? :) Aug 20 15:23:02 pb_: time for a question about package naming? Aug 20 15:24:30 03Khem Raj  07org.openembedded.dev * r538ccedebb 10openembedded.git/recipes/gloox/ (4 files in 2 dirs): Aug 20 15:24:30 gloox: Upgrade to 1.0 Aug 20 15:24:30 * Remove unbuildable old versions. Aug 20 15:24:30 Signed-off-by: Khem Raj Aug 20 15:30:28 khem: I see Ubuntu packages klcc in libklibc-dev Aug 20 15:31:53 khem: and i.e. klibc-qGbyaURo657ZIvp232JjfGSjx40.so as libklibc Aug 20 15:32:42 ant_work: sure Aug 20 15:32:53 hello pb_ Aug 20 15:33:00 is about klibc Aug 20 15:33:13 it happens that recipe is incompatible with new staging Aug 20 15:33:24 we have to split in more packages Aug 20 15:33:31 how tocall them? Aug 20 15:33:54 khem talked about klcc-native bt I see debian naming differs Aug 20 15:33:55 what goes in each of them? Aug 20 15:34:11 klcc is a crosscompiler, goes to sysroots Aug 20 15:34:27 the klibc-utils are for target Aug 20 15:35:14 klcc-native sounds about right to me. Aug 20 15:35:38 or klcc-cross. one could imagine "klcc" being a compiler to be installed on the target, cf gcc and gcc-cross Aug 20 15:36:18 yea, klcc-cross perhaps Aug 20 15:37:48 it's interesting, though, the old do_stage could install both for target and buildhost in a single package (like gcc-cross is doing) Aug 20 15:38:27 yeah, indeed. that doesn't work very well with new staging, need to split them up. Aug 20 15:38:50 as a general rule it is more wholesome to have each recipe build for just one target in any case. Aug 20 15:39:39 yes, more insulated Aug 20 15:40:31 pb_: so, it's ok to forget the debian naming Aug 20 15:40:35 in this case Aug 20 15:41:12 (libklibc-dev) Aug 20 15:41:53 I think it would be sensible to keep the debian naming for the actual -dev files, but the compiler shouldn't go in that package. Aug 20 15:42:16 debian doesn't (need to) worry about cross builds so it is ok for them to ship mixed binary packages like that. Aug 20 15:47:51 ok, and we can let the .so in klibc-dev or should be renamed libklibc? Aug 20 15:48:31 you can call it what you like, the autonamer will sort it out Aug 20 15:50:02 ok, at first was not packaged, then we did Aug 20 15:50:10 PACKAGES = "${PN} ${PN}-dev " FILES_${PN}-dev = "${base_libdir}/klibc.so" Aug 20 15:50:28 to shut up the QA Aug 20 15:51:36 * ant_work is reading about the autonamer Aug 20 15:51:42 pb_: thx Aug 20 16:04:15 bbl Aug 20 16:26:05 re Aug 20 16:29:04 03Antonio Ospite  07org.openembedded.dev * rfccf01fafe 10openembedded.git/recipes/linux/openezx-kernel_git.bb: Aug 20 16:29:04 openezx-kernel_git: bump SRCREV Aug 20 16:29:04 This revision brings in some fixes to pcap-regulator. Aug 20 16:29:04 Signed-off-by: Antonio Ospite Aug 20 16:29:07 03Antonio Ospite  07org.openembedded.dev * rab2f7ec845 10openembedded.git/recipes/tzcode/tzcode-native_2010k.bb: Aug 20 16:29:07 tzcode-native: bump version to 2010l Aug 20 16:29:07 Signed-off-by: Antonio Ospite Aug 20 16:29:09 03Antonio Ospite  07org.openembedded.dev * ref9fd49baf 10openembedded.git/recipes/tzdata/tzdata_2010k.bb: Aug 20 16:29:09 tzdata: bump version to 2010l Aug 20 16:29:09 Signed-off-by: Antonio Ospite Aug 20 16:54:49 03Andrea Adami  07org.openembedded.dev * re6eda83fd6 10openembedded.git/recipes/klibc/klibc-static-utils_1.5.19.bb: klibc: readd renamed static utilities (missing in previous commit) Aug 20 16:55:14 jesus, fresh recipe with no deps other than the stock ones - 650 tasks, including python-native? Aug 20 16:55:23 we really need to go through this shit with a fine toothed comb Aug 20 16:55:40 wow Aug 20 16:56:16 it's no wonder folks used to buildroot and firmware linux and stuff look at OE as this monstrous thing Aug 20 16:56:23 heh Aug 20 16:57:23 glib-2.0-native now? Aug 20 16:57:24 bite me. Aug 20 17:01:00 03Tom Rini  07org.openembedded.dev * r5b0d271683 10openembedded.git/recipes/u-boot/ (3 files): Aug 20 17:01:00 u-boot: Move the config target into do_configure Aug 20 17:01:00 Signed-off-by: Tom Rini Aug 20 17:29:09 03Tom Rini  07org.openembedded.dev * rc5a6d64427 10openembedded.git/recipes/time/time.inc: Aug 20 17:29:10 time: Add BBCLASSEXTEND = "native" Aug 20 17:29:10 Signed-off-by: Tom Rini Aug 20 17:29:13 03Tom Rini  07org.openembedded.dev * r4e239b1db5 10openembedded.git/recipes/tar/tar.inc: Aug 20 17:29:13 tar: Add BBCLASSEXTEND = "native" Aug 20 17:29:13 Signed-off-by: Tom Rini Aug 20 17:29:13 03Tom Rini  07org.openembedded.dev * rd420df7cf6 10openembedded.git/recipes/u-boot/u-boot-bug_svn.bb: Aug 20 17:29:14 u-boot-bug: Drop do_stage (not needed), add package_stagefile_shell to deploy Aug 20 17:29:14 Signed-off-by: Tom Rini Aug 20 17:31:38 kergoth glib-2.0-native is brought in by insane.bbclass as dep to check if desktop files are ok Aug 20 17:32:44 ugh Aug 20 17:32:51 that's really stupid Aug 20 17:34:25 kergoth_ don't tell me Aug 20 17:34:52 i think the best way is to split insane.bbclass Aug 20 17:36:08 that would be a good first step to making sense of that madness Aug 20 17:38:01 yup Aug 20 17:38:20 jkridner have you found a solution for gnome-games? Aug 20 17:38:48 it builds now. Aug 20 17:39:06 I don't recall what the solution was... Aug 20 17:39:11 but, I didn't have to patch it... Aug 20 17:39:19 it was a dependency issue... Aug 20 17:39:41 didn't document it cleanly, but I will get back to it building from scratch again. Aug 20 17:39:44 ah ok, i know the what was the name again bookkeeper or so Aug 20 17:42:11 khem/pb/RP found some more info on my configure problem, actually by putting TARGET_OS in the machine file as it is supposed to be gmp-native buils, but gcc-cross-initial failed: Aug 20 17:42:13 checking for suffix of object files... configure: error: cannot compute suffix of object files: cannot compile Aug 20 17:46:51 kergoth_: btw preparing the runqueue also seems to take longer nowadays Aug 20 17:47:11 hmm, shouldn't Aug 20 17:47:19 will look into it at some point Aug 20 17:47:25 what branch are you on? Aug 20 17:49:44 this is with 1.8.18, i think 1.10 is the same Aug 20 17:50:53 highly unlikely, 1.10+ is known to have a commit from RP that improves runqueue generation speed substantially Aug 20 17:51:00 it wasn't brought back to 1.8 Aug 20 17:51:12 i'll move to 1.10 Aug 20 17:51:32 thought i already had, but bitbake --version told differently Aug 20 17:53:07 hmm... the use of a 'maint' branch sounds like a nicer approach than maintaining long lived per-major-version stable branches Aug 20 17:53:08 * kergoth_ ponders Aug 20 17:54:03 wie weit bist du denn mit deiner doofenunterstützun? Aug 20 17:56:29 hmmm... I put USE_VT = "1" and SYSVINIT_ENABLED_GETTYS = "0" in my machine/.conf file and did 'bitbake -c clean base-files', but inittab still doesn't have a 'tty0' entry. :( Aug 20 17:56:52 :( Aug 20 17:59:08 hm: cc1: error: not configured for ABI: 'spe' Aug 20 17:59:32 how should I configure gcc for spe abi ? Aug 20 18:00:10 jkridner btw scrollkeeper fails for me guess somewhere scrollkeeper-native is needed Aug 20 18:00:32 oh wait, inittab is in sysvinit... I wonder if I did a -c clean of that! Aug 20 18:01:36 eFfeM: most certainly. denix explained how to fix that. I described in one of the e-mails why the problem is still there and then denix provided an e-mail from over a year ago describing the right way to remove the scrollkeeper dependency. Aug 20 18:01:46 it requires hacks to the autoconf sources. Aug 20 18:02:14 ah ok, saw the mail, assumed it was committed Aug 20 18:09:05 03Tom Rini  07org.openembedded.dev * rd6f49fe115 10openembedded.git/recipes/dbench/dbench_4.0.bb: Aug 20 18:09:05 dbench: Add popt to DEPENDS, fix GNU_HASH, add clients.txt to tbench Aug 20 18:09:05 Signed-off-by: Tom Rini Aug 20 18:09:07 03Tom Rini  07org.openembedded.dev * r034dce2d55 10openembedded.git/recipes/linux/linux-smdk2443_2.6.20+git.bb: Aug 20 18:09:07 linux-smdk2443: Switch to u-boot-mkimage-native Aug 20 18:09:07 Signed-off-by: Tom Rini Aug 20 18:09:08 03Tom Rini  07org.openembedded.dev * r73d5a2c506 10openembedded.git/recipes/update-modules/update-modules_1.0.bb: Aug 20 18:09:09 update-modules: Add INHIBIT_DEFAULT_DEPS = "1" Aug 20 18:09:09 Signed-off-by: Tom Rini Aug 20 18:09:10 03Tom Rini  07org.openembedded.dev * r09c0bb89e2 10openembedded.git/recipes/u-boot/u-boot-utils-native_1.2.0.bb: Aug 20 18:09:10 u-boot-utils-native: Only provided mkimage, drop now that it's unused Aug 20 18:09:10 Signed-off-by: Tom Rini Aug 20 18:09:14 03Tom Rini  07org.openembedded.dev * r65f7fb71bf 10openembedded.git/recipes/u-boot/ (60 files in 2 dirs): Aug 20 18:09:14 u-boot-mkimage-openmoko-native: Unused now, remove. Aug 20 18:09:14 Signed-off-by: Tom Rini Aug 20 18:09:17 03Tom Rini  07org.openembedded.dev * r6059d55945 10openembedded.git/recipes/linux/linux-gumstix_2.6.15.bb: Aug 20 18:09:17 linux-gumstix: Call uboot-mkimage and just use normal hook for DEPENDS Aug 20 18:09:18 Signed-off-by: Tom Rini Aug 20 18:09:19 03Tom Rini  07org.openembedded.dev * rb78bda26b7 10openembedded.git/recipes/linux/gumstix-linux.inc: Aug 20 18:09:19 gumstix-linux.inc: Switch to u-boot-mkimage-native Aug 20 18:09:19 Signed-off-by: Tom Rini Aug 20 18:09:23 03Tom Rini  07org.openembedded.dev * rf484884f7e 10openembedded.git/recipes/linux/linux-bd-neon-2.6_2.6.22.bb: Aug 20 18:09:27 linux-bd-neon-2.6: Switch DEPENDS to u-boot-mkimage-native, call uboot-mkimage Aug 20 18:09:27 Signed-off-by: Tom Rini Aug 20 18:09:27 03Tom Rini  07org.openembedded.dev * r5dcdc30b57 10openembedded.git/recipes/linux/linux-turbostation_2.6.20.2.bb: Aug 20 18:09:27 linux-turbostation: Drop extra DEPENDS as normal hook works Aug 20 18:09:28 Signed-off-by: Tom Rini Aug 20 18:09:28 03Tom Rini  07org.openembedded.dev * re3e1991133 10openembedded.git/classes/magicbox-image.bbclass: Aug 20 18:09:29 magicbox-image.bbclass: Call uboot-mkimage not mkimage, add DEPENDS Aug 20 18:09:29 Signed-off-by: Tom Rini Aug 20 18:23:11 03Tom Rini  07org.openembedded.dev * r82088cc7fc 10openembedded.git/classes/ (packaged-staging.bbclass staging.bbclass): Aug 20 18:23:11 staging, packaged-staging.bbclass: Update call to fastpath Aug 20 18:23:11 Call fastpath if pstageactive, and don't need to check it again in Aug 20 18:23:11 pstaging's fastpath. Aug 20 18:23:11 Signed-off-by: Tom Rini Aug 20 18:24:34 03Martin Jansa  07org.openembedded.dev * r1b452f02fd 10openembedded.git/recipes/efl1/ewl_svn.bb: Aug 20 18:24:34 ewl: update SRC_URI, was moved to BROKEN Aug 20 18:24:34 Signed-off-by: Martin Jansa Aug 20 18:24:36 03Martin Jansa  07org.openembedded.dev * rb03b1f6016 10openembedded.git/recipes/ (6 files in 4 dirs): Aug 20 18:24:36 recipes: bump PR after efreet -> libefreet1 PN change Aug 20 18:24:36 Signed-off-by: Martin Jansa Aug 20 18:24:37 03Martin Jansa  07org.openembedded.dev * r0e5d23b3f4 10openembedded.git/recipes/tasks/task-shr.bb: Aug 20 18:24:37 task-shr: use eve as new default browser, keep ventura in feeds Aug 20 18:24:37 Signed-off-by: Martin Jansa Aug 20 18:24:38 03Martin Jansa  07org.openembedded.dev * r5ffe7981a3 10openembedded.git/recipes/e17/ (eve/default.theme.remove.min.size.patch eve_svn.bb): Aug 20 18:24:38 eve: update PV and SRC_URI for rewritten version Aug 20 18:24:39 Signed-off-by: Martin Jansa Aug 20 18:24:39 03Martin Jansa  07org.openembedded.dev * rd32c89e80a 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Aug 20 18:24:40 EFL: bump SRCREV for alpha release and rewritten eve Aug 20 18:24:40 Signed-off-by: Martin Jansa Aug 20 18:24:41 03Martin Jansa  07org.openembedded.dev * rae1c822dbb 10openembedded.git/recipes/intone-video/ (intone-video/depends.eina.patch intone-video_svn.bb): Aug 20 18:24:42 intone-video: search for eina not eina-0 Aug 20 18:24:42 Signed-off-by: Martin Jansa Aug 20 18:27:43 hollisb: Going to kick off m-t for qemuppc, see what happens here... Aug 20 18:29:54 03Martin Jansa  07org.openembedded.dev * r57a1c3a667 10openembedded.git/recipes/xorg-xserver/ (xserver-xorg-conf/spitz/xorg.conf xserver-xorg-conf_0.1.bb): Aug 20 18:29:54 xserver-xorg-conf: add xorg.conf for spitz Aug 20 18:29:54 Signed-off-by: Martin Jansa Aug 20 18:35:06 Tartarus: did you manage to build minimal for calamari, this was on the list may 8, 2010 Aug 20 18:36:22 I don't recall Aug 20 18:36:31 I know in general ppc stuff is in good shape Aug 20 18:38:16 hm, angstrom builds fine for me but has a TARGET_OS line in angstrom-eglibc.inc Aug 20 18:38:26 but can't get it to build without it Aug 20 18:38:38 sec Aug 20 18:38:42 ok Aug 20 18:38:54 wondering if it should pick up spe automatically Aug 20 18:39:06 angstrom is broken, yes Aug 20 18:39:12 sane-toolchain.inc gets this right Aug 20 18:39:22 compute_os_portion_of_target_triplet Aug 20 18:39:31 Is what angstrom should get changed to use Aug 20 18:39:39 ah ok, will try to build minimal Aug 20 18:42:11 kicking off clean build with minimal, afk for a while but will report back once it is done, Tartarus thanks for your help Aug 20 18:47:47 Tartarus: cool. I'm just not sure if I need some other magic variable set in my local.conf, or what Aug 20 18:47:55 argh Aug 20 18:48:04 is there really no way to apply a patch in a different directory than ${S} right now? Aug 20 18:48:07 this is really lame Aug 20 18:48:17 * kergoth_ glares at his patch code Aug 20 18:53:57 * kergoth_ tests enhancement for this Aug 20 18:55:17 i wonder how well a maint/master/next branching system would work for OE.. guess the difficulty would be in the lack of releases Aug 20 18:55:29 yeah Aug 20 18:55:43 getting people to test next will be an issue :) Aug 20 18:56:25 true, but i think we could find some. if something unstable gets into master for their setup, its their own fault for not testing next :) Aug 20 18:56:32 blame offloaded! Aug 20 18:56:35 ;) Aug 20 18:56:37 heh Aug 20 18:56:44 sounds like my week Aug 20 18:56:48 well, we clearly need to do something Aug 20 18:56:50 this just isn't working. Aug 20 18:57:21 well, I suspect clever people are already doing this for their own targets Aug 20 18:57:27 yeah, definitely Aug 20 18:58:03 it's just nearly impossible to test every combination. people doing development just can't make sure it works everywhere, so something like next would make a good staging area for that, if people are wililng to test their own particular combinations against it Aug 20 19:02:06 meh Aug 20 19:02:58 crofton: i think the key difference in policy with maint/master/next, other than the existance of maint, is the periodic *merge* of next into it, rather than trying to cherry pick fixes indefinitely, which just isn't sustainable Aug 20 19:03:07 yeah Aug 20 19:03:24 i think that's the better approach, but the question is how to determine when next is sufficiently stable Aug 20 19:03:29 we know the stable method does not work forever, but it does stop the target from mvoing Aug 20 19:03:39 yeah Aug 20 19:03:55 i'm thinking specific indiividuals volunteer to test their particular combination and submit the result somewhere we can look at to make sure its building fine Aug 20 19:04:04 i sort of view it as an end user responsibilty Aug 20 19:04:08 distro/machine combination testers, of a sort Aug 20 19:04:40 yeah, i agree for the most part, its easy to offload ensuring XYZ works to whomever's the userbase of XYZ Aug 20 19:04:50 but i can see the advantages of doing that in upstream branches Aug 20 19:04:54 vs everyone having to maintain their own Aug 20 19:05:16 and people that can't handle this should talk to your corporate overlords Aug 20 19:05:24 i dunno, we're just a very different project from normal open source projects Aug 20 19:05:28 no Aug 20 19:05:32 not really :) Aug 20 19:06:04 well, more so in the number of projects all playing in the same repository Aug 20 19:06:16 which was one of hte goals of OE from the beginning, to consolidate common metadata Aug 20 19:06:51 true Aug 20 19:07:06 but even regular sw projects have similar issues Aug 20 19:07:39 true, but as things stand today, a *large* number of the users of a piece of software end up maintaining their own patchsets Aug 20 19:07:42 basically every distribution Aug 20 19:08:25 you don't have the debian maintainer adn rpm maintainer for a given project playing in its repository, thank goodness Aug 20 19:08:26 hehe Aug 20 19:14:48 http://tinderbox.openembedded.org/public/logs/task/7297730.txt Aug 20 19:14:54 fbreader is failed Aug 20 19:15:49 http://tinderbox.openembedded.org/public/logs/task/7298263.txt Aug 20 19:15:52 libbonobo Aug 20 19:16:02 fails on ubuntu Aug 20 19:16:48 Ubuntu 10.04.1 LTS Aug 20 19:17:25 ah.. Aug 20 19:17:39 a bit outdated branch.. Aug 20 20:02:37 | /home/openembedded/tmp/sysroots/i686-linux/usr/bin/perl: /lib/i686/cmov/libc.so.6: version `GLIBC_2.2.5' not found (required by /home/openembedded/tmp/sysroots/i686-linux/usr/bin/../lib/perl/5.8.8/CORE/libperl.so) Aug 20 20:02:45 hm.. is it known bug? Aug 20 20:03:14 * Jay7 is on debian testing now Aug 20 20:05:23 ah... seems I understand Aug 20 20:05:42 I've rsync'ed tmp too.. Aug 20 20:11:20 from one machine to another, yeah Aug 20 20:11:30 doesn't work w/o some more stuff that's likely to upset folks :) Aug 20 20:12:27 hollisb: So, angstrom + qemuppc + meta-toolchain just built Aug 20 20:12:38 ubuntu 9.10 32bit host Aug 20 20:12:51 Tartarus: could the host really matter? Aug 20 20:13:26 Shouldn't, no Aug 20 20:13:37 Try qemuppc instead of ebony for now Aug 20 20:13:46 It's possible there's crazy bitrot elsewhere due to ppc440 Aug 20 20:13:53 hmm ok Aug 20 20:14:01 thanks Aug 20 20:14:03 Tartarus: whats the issue Aug 20 20:14:05 with ppc Aug 20 20:14:24 khem: See hollis' post to the ML Aug 20 20:14:43 oh I havent checked mail today Aug 20 20:14:48 03Tom Rini  07org.openembedded.dev * r2f40e35f34 10openembedded.git/recipes/busybox/ (busybox.inc files/hwclock-default files/hwclock.sh): Aug 20 20:14:48 busybox: Add and check /etc/default/hwclock for hwclock initscript Aug 20 20:14:48 Signed-off-by: Tom Rini Aug 20 20:15:00 basically I can't build meta-toolchain Aug 20 20:18:39 hollisb: strange Aug 20 20:18:55 I just blew away my tmp dir to rebuild fresh Aug 20 20:19:13 may be try as Tartarus suggested use something thats known to work Aug 20 20:19:19 and then alter the knobs one by one Aug 20 20:19:27 I had deleted part of it earlier, and maybe "bitbake -c clean" didn't really do what was needed Aug 20 20:20:21 hollisb: what target? Aug 20 20:21:04 not sure what you mean. angstrom-2008.1? Aug 20 20:21:10 likewise: howdy Aug 20 20:21:36 khem: howdy there khem Aug 20 20:22:14 khem: Bad news for you: http://nos.nl/video/177774-de-bierfiets-is-te-breed.html Aug 20 20:22:29 khem: we have to work on the problem before you can visit .nl Aug 20 20:23:19 likewise: heh Aug 20 20:24:03 hollisb: target usually means the target hw, so the ebony stuff you added Aug 20 20:25:49 hollisb: yes I meant the machine, more specifically the architecture you are building for (sorry for being too vague). Aug 20 20:26:32 hi florian Aug 20 20:26:42 re Aug 20 20:27:09 I run into this one: http://tinderbox.openembedded.net/public/logs/task/7299475.txt Aug 20 20:27:47 gives me a deja vu Aug 20 20:28:32 which package Aug 20 20:29:00 likewise: ah, I added a machine definition for Ebony, a PPC 440GP eval board Aug 20 20:29:09 likewise: I pasted that conf (it's short) into my mail Aug 20 20:29:49 hollisb: ok, I'll test my amcc460ex dev kit build (MACHINE=canyonlands) Aug 20 20:30:27 likewise: oh ok, cool Aug 20 20:35:23 forgot to preserve my download directory. I guess this will take a while... Aug 20 20:38:00 hollisb: multithreaded bitbake? Aug 20 20:38:53 hollisb: BB_NUMBER_THREADS larger than 1 and the downloads occur in a seperate thread Aug 20 20:39:20 ah, cool, I put that in local.conf? Aug 20 20:39:26 or site.conf, yes Aug 20 20:39:56 I think the number can be set to 4 or 5, as each thread handles one TYPE of task (fetch, unpack, patch, ...) Aug 20 20:40:18 ok Aug 20 20:40:33 hollisb: adding this *does* detect errors in the OE metadata faster, I usually set it back to 1 when debugging OE itself. Aug 20 20:41:03 hollisb: it should max out your machine, especially if you have multiple cores and fast disks Aug 20 20:41:12 that I don't want :) Aug 20 20:41:18 I'll keep it small Aug 20 20:57:19 good nite all Aug 20 20:58:00 hollisb: helloworld builds for me for canyonlands Aug 20 20:58:17 hollisb: didn't read your email, your prob was meta-toolchain?? Aug 20 20:58:18 like|zzz: thanks for the info Aug 20 20:58:20 yes Aug 20 20:58:53 hollisb: let's see if it builds while I close things down Aug 20 20:59:03 :) Aug 20 21:00:21 sanity.conf should check for == "dash"; it should check for != "bash" Aug 20 21:00:28 sanity.conf should NOT check for == "dash"; it should check for != "bash" Aug 20 21:00:46 (but I always disable that check, dash works for me) Aug 20 21:06:29 hollisb: sry, will report tomorrow Aug 20 21:06:32 nite all Aug 20 21:42:37 Anyone know why we test -x before doing a u-a for pivot_root? Aug 20 21:43:26 to be sure? :) Aug 20 21:43:53 heh Aug 20 21:44:00 Well, I see it very much not working right now Aug 20 21:44:09 now that u-a is able to finish running, on the host side Aug 20 21:44:14 * Tartarus fixed a few other problems in the recipe Aug 20 21:46:39 03Tom Rini  07org.openembedded.dev * rdd9969cbac 10openembedded.git/recipes/util-linux-ng/util-linux-ng.inc: Aug 20 21:46:39 util-linux-ng: Fix vipw/vigr/reset/readprofile U-A's. Aug 20 21:46:39 Signed-off-by: Tom Rini Aug 20 21:49:36 * Jay7 have quick-try to boot OE's image for Ben NanoNote from SD Aug 20 21:49:39 w/o success Aug 20 21:49:49 ! Aug 20 21:51:10 Tartarus: thanks again for the minimal hint, gcc builds now fine for spe, i seem to have an issue with busybox, but that it probably something I can sort out (tomorrow, calling it a day for now) Aug 20 21:51:18 k Aug 20 21:51:23 fwiw, spe stuff builds fine here Aug 20 21:51:32 for some subset of what i think is at least in console-image Aug 20 21:52:06 is Kristoffer Ericson here? :) Aug 20 21:52:09 i see a fail for libusb and busybox, but haven't peeked at the logs Aug 20 21:52:18 yeah, both of those are fine here Aug 20 21:52:20 will do so tomorrow or so Aug 20 21:52:47 nite all & have fun! Aug 20 21:53:34 too much zzZZzz's from mickey :) Aug 20 21:54:58 maybe he hit the keyboard with his head when he fell asleep Aug 20 21:55:18 * florian should get some sleep in order to avoid this to happen Aug 20 22:00:19 good idea.. Aug 20 22:01:32 i'm really getting sick of seeing exception tracebacks *twice* in bitbake master Aug 20 22:01:38 due to the python task log output Aug 20 22:01:40 hrm Aug 20 22:18:25 Jay7, ? Aug 20 22:18:39 kristoffer: hey :) Aug 20 22:18:58 kristoffer: did you still have building for Ben NanoNote? Aug 20 22:19:14 Jay7, yeah, what you need? Aug 20 22:19:23 git log show you as machine conf creator Aug 20 22:19:42 kristoffer: just to be sure that all is working :) Aug 20 22:19:54 Jay7, We got alot of stuff to push but wont happen within the next couple of weeks though. Got developer meeting in one week. Aug 20 22:19:59 I've quick try to boot OE's image from SD w/o success Aug 20 22:20:05 did not investigated yet Aug 20 22:20:19 Jay7, have you tried the userland image? Aug 20 22:21:04 I've tried angstrom-opie Aug 20 22:21:33 Jay7, I ment the muffinman release Aug 20 22:21:34 not sure I understand what is 'userland image' :\ Aug 20 22:21:50 where can I get it? Aug 20 22:26:33 build we build cmake (for the target yet) ? Aug 20 22:26:49 * Crofton is compiling cmake on an overo AGAIN Aug 20 22:26:53 and getting anoyed Aug 20 22:29:38 03Tom Rini  07org.openembedded.dev * rf554f67dfd 10openembedded.git/contrib/oe-stylize.py: Aug 20 22:29:38 oe-stylize.py: Add INC_PR Aug 20 22:29:38 Signed-off-by: Tom Rini Aug 20 22:29:48 03Tom Rini  07org.openembedded.dev * r280a17bc1d 10openembedded.git/recipes/debianutils/ (4 files): Aug 20 22:29:48 debianutils: Move run-parts and tempfile to ${base_bindir}, use INC_PR Aug 20 22:29:48 We move run-parts to match busybox and usage and tempfile to match Aug 20 22:29:48 normal usage (and upstream). Run oe-stylize.py while in here. Aug 20 22:29:48 Signed-off-by: Tom Rini Aug 20 23:02:00 03Tom Rini  07org.openembedded.dev * r4bbc0df43b 10openembedded.git/recipes/udev/ (9 files in 4 dirs): Aug 20 23:02:00 udev 141/151/154: Move rm'ing /tmp/uname, etc, to udev-cache Aug 20 23:02:00 Otherwise we'll remove these files before udev-cache is run. Aug 20 23:02:00 Signed-off-by: Tom Rini Aug 20 23:02:08 03Tom Rini  07org.openembedded.dev * r4e20a53382 10openembedded.git/recipes/udev/ (udev-141/init udev-154/init): Aug 20 23:02:08 udev 141/154: Move udevadm settle timeout to 8 to match 151 Aug 20 23:02:08 Signed-off-by: Tom Rini Aug 20 23:28:02 03Tom Rini  07org.openembedded.dev * r219b13ab95 10openembedded.git/recipes/busybox/ (6 files): Aug 20 23:28:02 busybox (various): Add hwclock-default to versions to overwrite SRC_URI Aug 20 23:28:02 Signed-off-by: Tom Rini Aug 21 00:13:53 03Tom Rini  07org.openembedded.dev * r8925db69ff 10openembedded.git/recipes/busybox/ (4 files in 3 dirs): Aug 21 00:13:53 busybox 1.15.3 / 1.16.2: Drop run-parts.in.usr-bin.patch Aug 21 00:13:53 Along with the change I made earlier to debianutils (which wasn't Aug 21 00:13:53 matching upstream Debian wrt locations), we don't need this change Aug 21 00:13:53 and since run-parts can be used before /usr is mounted, it's not Aug 21 00:13:53 right technically. Aug 21 00:13:54 Signed-off-by: Tom Rini Aug 21 00:19:50 khem: I'm testing the klcc-cross recipe, from scratch Aug 21 00:23:23 03Tom Rini  07org.openembedded.dev * re325cc7b38 10openembedded.git/recipes/busybox/ (7 files in 2 dirs): Aug 21 00:23:23 busybox: Add 1.17.1, D_P = -1 for all for now Aug 21 00:23:23 Signed-off-by: Tom Rini Aug 21 00:57:40 03Chris Larson  07org.openembedded.dev * r2b7ea6f52d 10openembedded.git/classes/ (base.bbclass utils.bbclass): Aug 21 00:57:40 Fix the package_arch=machine_arch logic to use FILESPATH* Aug 21 00:57:40 - Use FILESPATHPKG and FILESPATHBASE to locate the machine specific dirs Aug 21 00:57:40 - Use the fetch urldata rather than poking at urls ourselves Aug 21 00:57:40 - Move the logic into utils.bbclass as def'd python functions Aug 21 00:57:40 - Simplify Aug 21 00:57:41 Signed-off-by: Chris Larson Aug 21 01:50:39 zecke, gm Aug 21 02:48:36 Hi, I am getting the following error "ERROR: IO Error: file 'conf/bitbake.conf' not found " with bitbake and openembedded. Can anyone help me? Aug 21 02:49:09 your BBPATH is either not set correctly, or not set at all Aug 21 02:49:39 I have already set BBPATH. I have checked it with echo $BBPATH Aug 21 02:49:58 And the file exists under build/conf/ Aug 21 02:50:18 bitbake.conf is in OE, not build Aug 21 02:50:39 Ok, thank you, I'll check and get back to you Aug 21 02:50:42 the fact that you echo'd it doesn't necessarily mean anything at all Aug 21 02:50:54 in addition to being in your shell, it has to be exported Aug 21 02:50:54 a Aug 21 02:50:54 nd i Aug 21 02:50:59 and its value has to be correct Aug 21 02:51:08 that is, including both the path to your build directory and the path to the oe directory, in that order Aug 21 02:52:19 Thank you so much! Aug 21 02:52:25 That was the problem, it had to be exported! Aug 21 02:52:42 May I ask what the difference is between exporting and defining it under .bashrc? Aug 21 02:55:39 an exported variable makes it into the environment of child processes Aug 21 02:55:45 without that, bitbake can't see it, its only in the shell Aug 21 02:59:34 Great, thank you very much! Are you one of the developers of oe? I am Brazilian and I am developing a custom board based on the AT91SAM9XE processor Aug 21 02:59:52 as my senior year project at the univeristy Aug 21 02:59:55 yep, i'm one of the founders of the project **** ENDING LOGGING AT Sat Aug 21 02:59:57 2010