**** BEGIN LOGGING AT Thu Feb 24 02:59:57 2011 Feb 24 02:59:59 03Enrico Scholz  07master * r95896c5c12 10openembedded.git/recipes/gcc/ (2 files in 2 dirs): Feb 24 02:59:59 gcc-4.5: added thumb2 build patch Feb 24 02:59:59 Fix for http://gcc.gnu.org/bugzilla/show_bug.cgi?id=43999 and is Feb 24 02:59:59 required to build gcc for Thumb2 only architectures (e.g. Cortex-M3). Feb 24 02:59:59 Signed-off-by: Enrico Scholz Feb 24 03:00:00 Signed-off-by: Khem Raj Feb 24 04:53:28 03Mike Westerhof  07org.openembedded.dev * r9aad70c6dd 10openembedded.git/recipes/meta/slugos-packages.bb: SlugOS: slugos-packages - reinstate groff, man, and man-pages. Feb 24 06:46:23 gm Feb 24 07:07:30 morning eFfeM_work Feb 24 07:20:07 time to sleep Feb 24 07:20:11 gn all Feb 24 07:20:21 gn khem Feb 24 07:43:29 good morning Feb 24 07:43:38 moin Feb 24 08:06:58 hello! In the updated openembedded, I'm trying to add usb-gadget-ethernet support. I get: NoProvider: kernel-module-g-ether (working on beagleboard) Feb 24 08:35:33 03Martin Jansa  07master * rd91b13c8c3 10openembedded.git/recipes/tasks/task-shr-feed.bb: Feb 24 08:35:33 task-shr-feed: add valacompass Feb 24 08:35:33 Signed-off-by: Martin Jansa Feb 24 08:35:34 03Martin Jansa  07master * r6ba453f600 10openembedded.git/recipes/freesmartphone/ (5 files): Feb 24 08:35:34 freesmartphone: bump SRCREVs Feb 24 08:35:34 Signed-off-by: Martin Jansa Feb 24 08:35:37 03Martin Jansa  07master * rff542aeb52 10openembedded.git/recipes/webkit/webkit-efl_svn.bb: (log message trimmed) Feb 24 08:35:37 webkit-efl: bump SRCREV a bit more Feb 24 08:35:37 * be aware there is issue in eve which can be built only after removing Feb 24 08:35:37 ewk_view_context_menu_forward_event Feb 24 08:35:37 http://patchwork.dev.bearstech.com/patch/856/ Feb 24 08:35:37 but better solution should be found Feb 24 08:35:38 I've already asked #edevelop people, but only answer was that ewk/eve Feb 24 08:35:41 03Michele Brocco  07master * ra47c9f3c5c 10openembedded.git/recipes/valacompass/ (hmc5843_git.bb lis302dl_git.bb valacompass_git.bb): Feb 24 08:35:41 valacompass: new recipe for valacompass and deps hmc5843 and lis302dl Feb 24 08:35:41 * Works with the hmc5843 module Feb 24 08:35:41 * Calibration is required Feb 24 08:35:41 Signed-off-by: Michele Brocco Feb 24 09:09:58 cpg, by any chance any feedback for me on the proggie i send to the list yesterday (my) evening (10 hrs or so ago) Feb 24 09:10:07 wrong chat Feb 24 09:21:56 03Martin Jansa  07master * r9be60082fb 10openembedded.git/recipes/qt4/qt4-x11-free.inc: Feb 24 09:21:56 qt4: bump INC_PR Feb 24 09:21:56 * 459110a17e59594c4fd9183a5684413ffa1479dd added generate_qt_config_file task Feb 24 09:21:56 which is executed in empty workdir if builder is using rm_work Feb 24 09:21:56 Signed-off-by: Martin Jansa Feb 24 09:46:47 hi broonie Feb 24 11:31:20 hi obi Feb 24 11:31:37 hi! Feb 24 13:17:56 hi hrw Feb 24 13:31:38 re **** BEGIN LOGGING AT Thu Feb 24 17:26:17 2011 Feb 24 17:27:47 yay it works, good job! (at least the wiki does) Feb 24 17:28:46 not quite yet... Feb 24 17:28:52 finishing up the fixing.. Feb 24 17:28:57 it will be spotty for a few minutes Feb 24 17:58:38 Tartarus: ping Feb 24 18:05:15 hi, any known solutions for configure.ac:12: error: AC_INIT should be called with package and version arguments Feb 24 18:05:23 I couldn't google anything relevant Feb 24 18:06:05 ? Feb 24 18:06:18 03Andreas Oberritter  07master * r2cfbab2f0f 10openembedded.git/recipes/libgpod/libimobiledevice_1.0.0.bb: (log message trimmed) Feb 24 18:06:18 libimobiledevice-1.0.0: depend on libtasn1 Feb 24 18:06:18 * Fixes the following error: Feb 24 18:06:18 | checking for libtasn1... configure: error: Package requirements Feb 24 18:06:18 (libtasn1 >= 1.1) were not met: Feb 24 18:06:18 | Feb 24 18:06:18 | No package 'libtasn1' found Feb 24 18:06:20 look at the autoconf documenataton Feb 24 18:06:20 or rather this one (right before the one above): error: cannot get project version Feb 24 18:06:35 woglinde: I am trying to build a recent dpkg Feb 24 18:06:48 woglinde: do you think that they have broken autoconf stuff? Feb 24 18:07:06 filip I cannt believe they wracked autotools support Feb 24 18:07:16 maybee they switched to cmake? Feb 24 18:08:09 hmmm: ./git/get-version: echo "error: cannot get project version." 1>&2 Feb 24 18:08:12 this is interesting Feb 24 18:08:48 aah, I know now. Can i prevent git fetcher from removing .git? Feb 24 18:09:52 hm Feb 24 18:09:59 khem may know Feb 24 18:10:28 khem: do you :)? Feb 24 18:12:27 gm Feb 24 18:12:44 eFfeM: hi Feb 24 18:13:45 filip: set add scmdata=keep to SRC_URI Feb 24 18:13:58 obi ah Feb 24 18:14:55 obi: I've just found that out, thanks :) Feb 24 18:17:59 khem, any idea on how to attack debugging a gcc 4.1.2 ice (target neek (arch == nios2))? With dbus I get: Feb 24 18:18:00 dbus-auth.c:2760: error: Attempt to delete prologue/epilogue insn: Feb 24 18:18:08 dbus-auth.c:2760: internal compiler error: in propagate_one_insn, at flow.c:1699 Feb 24 18:19:25 pastebin the code Feb 24 18:21:12 will do Feb 24 18:21:27 maybee wrong usage of isnan Feb 24 18:22:34 http://www.pastie.org/1603113 Feb 24 18:23:33 guess so, the code is not *that* mature, then again it did manage to build some 25 other packages before Feb 24 18:24:40 why does the bitbake version checker think that '0.0 < 0'? Was this behaviour added by a bug and not fixed to avoid breaking compatibility? Feb 24 18:38:09 denix, pong Feb 24 18:41:53 hm, is there any way OE prevents building of locale files? Feb 24 18:42:02 yeah Feb 24 18:42:21 ENABLE_BINARY_LOCALE_GENERATION = "0" Feb 24 18:42:27 will move generation to the target Feb 24 18:42:39 or do you mean limit the ones done? Feb 24 18:42:53 Tartarus: well I can't see .mo files built/installed yet I don't have this variable defined Feb 24 18:43:02 Tartarus: and I do want them in locale packages Feb 24 18:43:19 not sure then Feb 24 18:43:28 Hi Feb 24 18:43:54 StefanD_: hi Feb 24 18:49:56 --disable-nls where the hell does it come from? Feb 24 18:50:39 Tartarus: so, I guess when you say "I've never not had to set" you mean you always set? :) Feb 24 18:51:43 Er, yes :) Feb 24 18:52:38 I'm just getting started with OE and embedded Linux in general. I followed the getting started guide, and now I tried to "bitbake helloworld". But I get Errors like these: http://pastebin.com/MNPMpGLV Feb 24 18:52:57 I first tried it with bitbake 1.12.0 and it aborted after the first error (Line 4). I've now downgraded to bitbake 1.8.18 and it continues further. But now it fails later. Feb 24 18:54:26 Can someone tell me which bitbake version he is using, that is known to work with the testing-release-2011.03 release? Feb 24 18:54:54 lemme see... Feb 24 18:55:14 Can you pastebin your local.conf please Feb 24 18:55:36 looks like a not correctly chosen distro/machine, most likely Feb 24 18:56:29 Yep http://pastebin.com/B9bYUwqY Feb 24 18:57:29 wow, 285,068 unique filename/task combinations from OE today Feb 24 18:59:36 StefanD_, yeah, that doesn't look totally right Feb 24 18:59:50 First, minimal not Minimal :) Feb 24 18:59:58 Second, take a look at openembedded/conf/local.conf.sample Feb 24 19:00:17 And right now, while filip is working on deb support, I don't know if it's fully functional yet Feb 24 19:00:52 So it would be INERIT += "package_deb", I wouldn't recommend changing from ipk, which is what minimal will pull in Feb 24 19:01:19 Finally, once you do that, you can and should switch back to 1.12.0 :) 1.8.18 isn't supported anymore (min is 1.10.2) and 1.12.0 is much faster, especially on the initial parse if you have >1 core Feb 24 19:02:42 filip: gettext.bbclass, most likely Feb 24 19:02:50 (re: disable-nls) Feb 24 19:03:32 Tartarus: Oh thanks :-) output looks much better now Feb 24 19:03:57 kergoth`: but bitbake -e doesn't show USE_NLS variable Feb 24 19:04:15 Tartarus: so, should we bring the rest of --sysroot changes from Poky? and, btw, I see cross.bbclass clearing TOOLCHAIN_OPTIONS, but nothing else. will it work for native? Feb 24 19:04:34 denix, no, that's due to poky not having 'native' anymore really, just nativesdk Feb 24 19:04:38 So we need to clear it there too Feb 24 19:04:48 kergoth`: so I guess the function in gettext.bbclass doesn't do this? Feb 24 19:05:05 Is it also save to switch to the current git master of OE instead of testing-release-2011.03? Feb 24 19:05:14 Yes Feb 24 19:09:21 kergoth`: it's not from gettext.bbclass, any idea about another source? Feb 24 19:15:19 filip: gettext.bbclass is one of two classes that set that option, everything else comes from the recipes.. Feb 24 19:15:22 there's no magic here Feb 24 19:15:23 grep is your friend Feb 24 19:15:38 git grep even more so :) Feb 24 19:15:58 well I've grepped out nothing usable... Feb 24 19:16:00 damn Feb 24 19:16:46 well, there are only so many places metadata comes from. Feb 24 19:19:03 should I avoid separating non-native and native recipe when sending patches? Feb 24 19:21:15 Tartarus: I finished the bisect for the freetype issue. Please see http://pastebin.com/PdTRjXx0 Feb 24 19:21:34 hm freetype issue? Feb 24 19:21:47 yeah - it won't build for me Feb 24 19:21:55 how the hell strongswan? Feb 24 19:22:01 started monday/tuesday-ish Feb 24 19:22:02 no idea Feb 24 19:22:07 it never ever to do with freetype Feb 24 19:22:15 Someone has any clue with a sem_post could return with success but do not "change" the semaphore? Feb 24 19:22:19 Um, ok, wow Feb 24 19:22:40 or tartarus changed some freetype Feb 24 19:22:44 recipe inside Feb 24 19:22:48 but that I doubt Feb 24 19:22:52 let me dig up the original build error paste Feb 24 19:22:52 yeah, no Feb 24 19:22:58 that's just adding a new recipe Feb 24 19:23:15 besides strongswan dont uses freetype at all Feb 24 19:24:10 http://pastebin.com/ssK3yQte Feb 24 19:24:46 thats libtool 2.4 stuff Feb 24 19:24:55 taking libz.so from host sys Feb 24 19:25:34 woglinde: yeah - that was the original prognosis Feb 24 19:26:07 my whole build world kind of fell apart two weeks ago with all of the RPATH stuff, got better and then I hit this on Monday Feb 24 19:26:25 prior to that, everything has been fine for months and months Feb 24 19:26:53 but not for all recipes Feb 24 19:27:00 and branch is called dev Feb 24 19:27:31 guess freetype is with-sysroot missing Feb 24 19:27:39 will look at it now Feb 24 19:28:04 hrms seems infrastructre is not fully recovered Feb 24 19:28:26 hm my crappy inet connection again Feb 24 19:31:36 tartarus wasnt you the last one which touches freetype? Feb 24 19:35:15 martinmeba are you using libtool2.2? Feb 24 19:35:53 libtool-native 2.2.6 Feb 24 19:36:06 (angstrom 2008.1) Feb 24 19:36:09 yes Feb 24 19:36:10 okay Feb 24 19:36:29 needs the same fix as db and flac and sdl-mixer Feb 24 19:36:57 So, yes, back in oct I added 2.4.3 and dropped many older freetypes Feb 24 19:37:35 I will fix it now Feb 24 19:37:41 cool - Thanks! Feb 24 19:37:57 maybee that stuff should go into autotools class Feb 24 19:38:02 so many recipes Feb 24 19:39:17 re effem Feb 24 19:41:15 eFfeM: that error means that there is some insn pattern for nios which is not having correct conditions Feb 24 19:41:36 03Tom Rini  07master * ra5951200ff 10openembedded.git/recipes/pulseaudio/ (4 files): Feb 24 19:41:36 libcanberra: Add gconf to DEPENDS Feb 24 19:41:36 Signed-off-by: Tom Rini Feb 24 19:41:38 hi khem Feb 24 19:41:44 eFfeM: that kind of stuff needs professional support :) Feb 24 19:41:46 03Tom Rini  07master * r4efc9b0446 10openembedded.git/recipes/gnome/gnome-doc-utils.inc: Feb 24 19:41:46 gnome-doc-utils: Fix xml2po and long path to python Feb 24 19:41:46 Signed-off-by: Tom Rini Feb 24 19:41:50 03Tom Rini  07master * rafebb05633 10openembedded.git/recipes/gthumb/ (gthumb_2.10.11.bb gthumb_2.6.9.bb): Feb 24 19:41:50 gthumb: Fix long path to perl in local intltools scripts Feb 24 19:41:50 Signed-off-by: Tom Rini Feb 24 19:41:52 03Tom Rini  07master * raea42e2f74 10openembedded.git/recipes/loudmouth/loudmouth_1.4.3.bb: Feb 24 19:41:52 loudmouth 1.4.3: Switch from gnutls to openssl Feb 24 19:41:52 This version of loudmouth wants gnutls 1.4 rather than 2.0 so switch Feb 24 19:41:52 to openssl instead. Feb 24 19:41:52 Signed-off-by: Tom Rini Feb 24 19:41:54 03Tom Rini  07master * r5396e0868c 10openembedded.git/recipes/icon-naming-utils/icon-naming-utils-native_0.8.7.bb: Feb 24 19:41:54 icon-naming-utils-native: Fix long path to perl in icon-name-mapping Feb 24 19:41:54 Signed-off-by: Tom Rini Feb 24 19:41:58 woglinde: hello Feb 24 19:42:28 fwiw, all of that gets angstrom-gnome-image building under my matrix, for angstrom at least Feb 24 19:42:40 It's not a sane target to build on minimal*, so dropping from next iteration Feb 24 19:43:26 martinmeba are you sure you are under latest dev? Feb 24 19:44:00 no problems here compiling freetype under libtool2.2 Feb 24 19:44:49 i believe so Feb 24 19:45:56 with host libz installed Feb 24 19:46:19 could you check that you have libz.so under sysroot? Feb 24 19:46:29 evening Feb 24 19:46:33 and try building from scratch? Feb 24 19:46:36 hi jay7 Feb 24 19:47:33 in ~/oe/tmp/sysroots/x86_64-linux/usr/lib I have a sequence of libz.so variants Feb 24 19:47:47 i will update and rebuild and let you know Feb 24 19:48:33 hm intressting that it is in do_install Feb 24 19:49:19 hm oh okay I see the problem now Feb 24 19:49:26 /devel/arm/oetmp-ang2/sysroots/i686-linux/usr/armv5te/lib/gcc/arm-angstrom-linux-uclibceabi/4.3.3/../../../../arm-angstrom-linux-uclibceabi/bin/ld: skipping incompatible /usr/lib/libz.so when searching for -lz Feb 24 19:49:37 okay Feb 24 19:49:41 thats why we dont see it Feb 24 19:50:06 I will do the fix like the others Feb 24 19:51:25 woglinde: all, hi (actually my client disconnected and reconnected without me being aware of it and not paying too much attention to chat) Feb 24 19:52:44 khem, I can imagine that there is some oddity in the insn pattern for the nios (actually already found a different issue half a year or so ago) Feb 24 19:53:08 would it be possible for you to peek into it? Feb 24 19:53:35 issue ls also that a line like Feb 24 19:53:54 CC libdbus_1_la-dbus-auth.lo Feb 24 19:54:01 what the hell is going on Feb 24 19:54:13 effem see khem msg Feb 24 19:54:18 does not really give me a good handhold to reproduce the message Feb 24 19:54:29 with freetyoe Feb 24 19:55:00 libtool isnt copied Feb 24 19:55:02 woglinde: saw the 2 messages after i returned Feb 24 19:55:20 this was my reply to it :-) Feb 24 19:55:49 and why the hell we dont call autoreconf on it Feb 24 19:57:39 hm ah because of crappy pseudo build system Feb 24 19:58:00 khem, ping re http://patches.openembedded.org/patch/256/ Feb 24 20:04:30 Tartarus: I havent had time to debug it Feb 24 20:04:49 Tartarus: if you want to pull it into release its fine Feb 24 20:05:02 Tartarus: but on master leave it open Feb 24 20:05:50 hrms where the hell the -L/usr/lib comes from Feb 24 20:06:20 hi, I'm having problems building several packages due to this error: http://paste.pocoo.org/show/344039/ Feb 24 20:06:38 woglinde: mostly stale libtool Feb 24 20:07:00 khem I used the sed trick Feb 24 20:07:12 pespin, what machine? Feb 24 20:07:17 woglinde: is it angstrom 2008 Feb 24 20:07:26 pespin: are you using external toolchain ? Feb 24 20:07:32 khem yes I used the sed like the others Feb 24 20:07:39 Tartarus, a new one I'm adding, an exopc slate tablet Feb 24 20:07:40 but I will look now deeper Feb 24 20:07:49 someone used libtoolize Feb 24 20:07:52 khem, I'm using bitbake if you mean that ;) Feb 24 20:07:57 and patched out LT_INIT Feb 24 20:08:08 woglinde: hmm Feb 24 20:08:12 not a good idea Feb 24 20:08:12 Tartarus, here's the files which I'm, using to build the machine: http://shr-project.org/trac/wiki/exoPCSlate Feb 24 20:08:16 thats useless anyway Feb 24 20:08:19 bye Feb 24 20:08:30 I'm planning to send them upstream when they work. Feb 24 20:08:47 note all this configurations work in shr-testing branch, but not in master Feb 24 20:09:13 hm Feb 24 20:09:16 sys_lib_dlsearch_path_spec="/lib /usr/lib /usr/local/lib /lib/i486-linux-gnu /usr/lib/i486-linux-gnu /usr/local/lib " Feb 24 20:09:26 pespin, ah, ok Feb 24 20:09:27 pespin: which distro are you using Feb 24 20:09:42 khem, shr (shr-unstable) Feb 24 20:09:48 ok Feb 24 20:10:30 That branch seems a bit crusty Feb 24 20:10:43 Can you post the whole git diff you' Feb 24 20:10:45 shr-unstable uses OE master Feb 24 20:10:54 ve got for brinigng back into master? Feb 24 20:11:12 pespin, yes, but it hasn't been updated since may Feb 24 20:11:17 shr/unstabl Feb 24 20:11:17 e Feb 24 20:11:52 Tartarus, shr-unstable since may? it's updated probable every day I would say ;) Feb 24 20:12:01 Tartarus: ppc603 is from freescale right Feb 24 20:12:25 pespin, what branch name? Feb 24 20:12:29 khem, yes Feb 24 20:12:31 org.openembedded.dev Feb 24 20:12:33 now I wonder if we should differentiate the toolchains for ibm ppc and fscale ppc Feb 24 20:12:49 pespin, heh Feb 24 20:13:05 since the Os fix that went in into gcc specifically fixes eabi toolchains Feb 24 20:13:21 khem, huh? Feb 24 20:14:23 ibm'ish ppc is now just 4xx and 970 and whatever magic names they have for POWER4/whatever on up Feb 24 20:14:40 And fsl has various code names for stuff, but it's everything else, heh Feb 24 20:14:42 I get same error from recipes such as linux-libc-headers_2.6.32.bb, libtool-cross_2.4.bb, update-rc.d_0.7.bb Feb 24 20:14:45 But eabi doesn't make much sense Feb 24 20:14:49 Tartarus: PowerPC G3 (750) is also from freescale ? Feb 24 20:15:05 khem, back when anyone made 750s, both Feb 24 20:15:11 hmm Feb 24 20:15:18 So, what are you driving at? Feb 24 20:15:21 Lemme see the fix Feb 24 20:15:47 hm, I have trouble understanding difference of PROVIDES vs RPROVIDES Feb 24 20:15:57 well the problem with those savegprs etc. is specific to fscale ppc it seems Feb 24 20:16:10 filip: its the same as DEPENDS vs RDEPENDS. Feb 24 20:16:12 khem, lemme see the fix commit(s) please :) Feb 24 20:16:17 kergoth`: hmm, ok :) Feb 24 20:16:32 kerogoth`, that's true and equally confusing at times to us mere mortals ;) Feb 24 20:16:42 * Tartarus kicks a lack of tab complete Feb 24 20:16:55 Tartarus: http://gcc.gnu.org/viewcvs?view=revision&revision=170443 Feb 24 20:17:13 look at config.host changes Feb 24 20:17:17 the fix is correct Feb 24 20:17:28 but I want to make sure that we are using right triplets Feb 24 20:17:31 in toolchain Feb 24 20:17:39 so its configured well Feb 24 20:18:02 now kexecboot have log viewing ability Feb 24 20:18:04 filip: they're two independent namespaces. One of the packages in recipe foo may RDEPENDS on a package emitted by recipe bar (listed in bar's PACKAGES, or is rprovided by one of those) Feb 24 20:18:09 khem, um Feb 24 20:18:12 that long awaited feature Feb 24 20:18:16 That doesn't look like they fixed linux for anything Feb 24 20:18:20 just the bare metal stuff Feb 24 20:18:38 Should I go see if I can find Nathan? :) Feb 24 20:19:04 filip: R* is what gets emitted into the binary packages, to put it simpler yet. R is runtime, think in terms of how the packages will behave when constructing the root filesystem, or on the target Feb 24 20:19:14 filip: its just taht bitbake has some awareness of this to avoid having to duplicate metadata Feb 24 20:19:15 kergoth`: hm, If I am having an update-alternatives package in dpkg native recipe, should I do RPROVIDES_${PN}-update-alternatives = "update-alternatives" or RPROVIDES_dpkg-update-alternatives-native = "update-alternatives" ? Feb 24 20:19:41 Tartarus: my bad t-ppccomm is already included for lnx configs Feb 24 20:19:41 again, R is runtime. native is the opposite of that.. Feb 24 20:19:45 kergoth`: is it acceptable if 'native' is not at the end of the package? Feb 24 20:19:50 why would it be? Feb 24 20:20:04 kergoth`: to use the first syntax, not the second Feb 24 20:20:06 Tartarus: hmmm so it still is broken Feb 24 20:20:33 filip: RPROVIDES operates against a package listed in PACKAGES. if RPROVIDES_foo and foo isn't in PACKAGES, it will do jack squat Feb 24 20:20:37 kergoth`: so that if I have a foo-native recipe providing bar, should it be foo-native-bar (for common prefix) or foo-bar-native (for native at end) Feb 24 20:20:55 khem, is there a PR for it being broken on linux? heh Feb 24 20:21:22 kergoth`: no no, I mean something else Feb 24 20:21:32 kergoth`: I'm just asking about package calling convention Feb 24 20:21:53 Tartarus, khem I'm going to have dinner, please let me know if you find some error in my config or you fix the problem ;) Feb 24 20:22:00 kergoth`: foo-native-bar vs foo-bar-native (generated from foo-native_version.bb) Feb 24 20:22:28 we're in #oe, not #poky. OE doesn't mangle runtime packages for native yet Feb 24 20:22:30 :P Feb 24 20:22:40 see native.bbclass. Feb 24 20:22:54 kergoth`: god I am lost in our conversation Feb 24 20:22:57 khem, I guess you need to open a new PR, heh Feb 24 20:23:08 once again, RPROVIDES only makes sense relative to an item in PACKAGES Feb 24 20:23:10 Since there's no way that change can fix linux, if that file is already included... Feb 24 20:23:13 if you don't know what's in PACKAGES, *LOOK* Feb 24 20:23:17 bitbake -e is your friend Feb 24 20:23:25 kergoth`: let's reset this Feb 24 20:24:11 pespin, the question is why your TARGET_OS is set to linux-gnu not linux, basically Feb 24 20:24:12 kergoth`: what I need to know - if I have a foo-native_version.bb, can it provide foo-native-bar, or should it rather provide foo-bar-native? Feb 24 20:24:22 let me put it another way. native.bbclass resets PACKAGES, so RPROVIDES is *irrelevent* Feb 24 20:24:25 PROVIDES is not Feb 24 20:24:35 there was a short time where everything had linux-gnu not linux (except well, the various important other cases, ARM, some powerpc, etc) Feb 24 20:26:06 kergoth`: I am just not sure about the position of 'native' adjective in a package name Feb 24 20:26:38 yes, and I can't clarify because you're not being clear about what you're referring to. Feb 24 20:26:54 native recipes don't emit binary packages, so 'package' in that context makes absolutely no sense Feb 24 20:27:20 kergoth`: ok, I'll try to write it down :) Feb 24 20:27:35 khem, Natahn is on #gcc, if we want to go bug him :) Feb 24 20:27:47 gah, can't type this afternoon Feb 24 20:29:04 so Feb 24 20:29:06 lets see Feb 24 20:29:11 now I should have fixed it Feb 24 20:29:24 adding another sed Feb 24 20:29:24 ok, pastebin is down Feb 24 20:29:53 there are likely hundreds of good pastebins around. off the top of my head, pastey.net, pastebin.com, pastebin.ca, gist.github.com Feb 24 20:29:57 take your pick :) Feb 24 20:31:49 args no Feb 24 20:31:57 where the hell the crappy rpath comes now Feb 24 20:32:10 * kergoth` sighs Feb 24 20:33:20 hm ah Feb 24 20:33:28 my sed didnt worked Feb 24 20:35:24 kergoth`: now I get how my question was wrong :) Feb 24 20:37:16 I wouldn't say wrong, just unclear, at least to me. Feb 24 20:37:18 * kergoth` shrugs Feb 24 20:37:18 hm why my groupstuff dont works Feb 24 20:37:20 damn sed Feb 24 20:37:41 hm ah I know Feb 24 20:47:21 03David Lanzendörfer  07org.openembedded.dev * rfaa2c615c8 10openembedded.git/ (3 files in 3 dirs): Feb 24 20:47:21 conf/machine/beaglehyb.conf: new machine Goldelico's Beagle Board Hybrid Feb 24 20:47:21 linux-gta04: add kernel for Beagle Board Hybrid and the forthcoming Goldelico GTA04 Machine Feb 24 20:49:11 hm I have no clue where the -rpath comes from in install and relinking Feb 24 20:49:59 hm ah Feb 24 20:50:04 its in the .la file Feb 24 20:52:23 hm Feb 24 20:52:38 maybee its using the unmodified libtool Feb 24 20:58:49 hm hm why gets the .la file the -rpath option and the .so lib not Feb 24 21:03:39 dpkg-deb: warning: '/home/builds/jornada7xx/tmp/work/armv4-oe-linux-gnueabi/dpkg-1.15.8.10+git-r0/packages-split/dpkg-perl/DEBIAN/control' contains user-defined field 'OE' Feb 24 21:03:46 can I get rid of this field? Feb 24 21:14:34 filip: what does full control file look like Feb 24 21:17:38 khem are you familiar with the .lai files from libtool? Feb 24 21:17:58 the .lai files points to /usr/lib/libz.la Feb 24 21:18:10 the .la file to the correct sysroot one Feb 24 21:19:33 khem: http://pastebin.ubuntu.com/571922/ Feb 24 21:39:11 when I specify FILES_package-subpackage, do I have to remove those files from FILES_package? Feb 24 21:39:23 I mean the default mask for a package Feb 24 21:40:30 naah, I shouldn't have to Feb 24 21:40:42 nope Feb 24 21:40:46 But order in PACKAGES matters Feb 24 21:41:03 ie if something would normally be caught by the main package Feb 24 21:41:18 you need to prepend PACKAGES not append packages Feb 24 21:41:45 So, all Feb 24 21:41:45 aaaaah Feb 24 21:41:49 that explains a lot :) Feb 24 21:41:55 1h20m or so and I'll update the testing branch again Feb 24 21:42:08 So anything you want in, get it in quickly please or lemme know you need more time Feb 24 21:43:51 damn it Feb 24 21:43:57 stupid freetype defs Feb 24 21:44:07 INK_LIBRARY = $(LIBTOOL) --tag=CC --mode=link $(CCraw) -o $@ $(OBJECTS_LIST) \ -rpath $(libdir) -version-info $(version_info) \ Feb 24 21:44:36 hrms Feb 24 21:44:39 okay Feb 24 21:44:43 but now I can fix it Feb 24 21:45:17 and I will fix the patches disaster Feb 24 21:48:32 hm I think we should drop chinook support Feb 24 21:55:04 *sigh* why always I have to do the dirty work Feb 24 22:02:39 can I have an empty directory in a package? Feb 24 22:04:24 Tartarus: I need to get in a gcc update Feb 24 22:04:28 I mean patch Feb 24 22:04:34 ok Feb 24 22:04:51 will do it in an hour or so Feb 24 22:04:59 I am test building an image Feb 24 22:05:28 k Feb 24 22:05:56 hmmm the build has gone past the failure point Feb 24 22:06:04 so should be ok but I still would like to wait Feb 24 22:07:42 ok Feb 24 22:08:47 03Richard Purdie  07master * r1faa0bd624 10bitbake.git/lib/bb/fetch2/__init__.py: Feb 24 22:08:47 bitbake/fetch2: Ensure SRCREV_pn-PN is checked for a revision when the SRC_URI is unnamed Feb 24 22:08:47 (From Poky rev: 707132c60a84cdcf50da056c68f2da7402d3b2e8) Feb 24 22:08:47 Signed-off-by: Richard Purdie Feb 24 22:08:50 03Richard Purdie  07master * r99c38e8e31 10bitbake.git/lib/bb/ (data_smart.py parse/ast.py): (log message trimmed) Feb 24 22:08:50 bitbake/data_smart: Improve the way lazyassignment works Feb 24 22:08:50 Currently, if a variable has been set with ??= and the code looks it up Feb 24 22:08:50 before the data finalisation phase, no value is found. This is causes Feb 24 22:08:51 serious problems for anonymous python functions which manipulate data, or Feb 24 22:08:51 for the fetcher revision handling code where revisions can be set with Feb 24 22:08:52 ??=. Feb 24 22:08:52 03Richard Purdie  07master * r90fc09ee05 10bitbake.git/lib/bb/build.py: (log message trimmed) Feb 24 22:08:53 build.py: Operate logfile handling at the task level Feb 24 22:08:53 Previously, not all task output could end up in the logfile and Feb 24 22:08:54 sub functions run using exec_func() calls would end up logging to Feb 24 22:08:54 /dev/null. This is extrememly confusing for users who want to find Feb 24 22:08:55 the output from a given task as there are several possible log files, Feb 24 22:08:55 or it might not be there at all. Feb 24 22:09:00 03Richard Purdie  07master * r44dcec236f 10bitbake.git/lib/bb/ (build.py runqueue.py): (log message trimmed) Feb 24 22:09:00 build.py: Improve exec_task standalone usage Feb 24 22:09:00 This commit consolidates all the task execution code and exception Feb 24 22:09:00 handling into one place, the exec_task funciton where it should be. Feb 24 22:09:00 This function now returns an exit code value and handles all exceptions. Feb 24 22:09:11 (From Poky rev: 56a92105fe6b779c69bccd44c2cff8f21cafdfbd) Feb 24 22:18:41 Tartarus, I tried setting TARGET_OS = "linux" in machine config file and removed tmp dir and rebuilded from scratch, and I still have the same problem with linux-gnu, so I suppouse there's some kind of recipe which changes that env var? Feb 24 22:19:30 So, are you using an external toolchain by chance? Feb 24 22:19:44 Tartarus, bitbake Feb 24 22:19:57 I mean are you using CSL for gcc & co Feb 24 22:19:58 or no? Feb 24 22:20:32 Tartarus, uhm I don't even know what's that hehe Feb 24 22:21:49 ok Feb 24 22:24:24 the first error of this kind usually happens when building linux-libc-headers_2.6.32.bb Feb 24 22:24:59 does anything package? Feb 24 22:25:08 or just not 'all' type stuff? Feb 24 22:26:27 Tartarus, sorry, I don't understand you, can you rephrase please? Feb 24 22:30:17 So, what's failing is package_qa Feb 24 22:30:34 Do you have anything being written to tmp/deploy/ipk/i686/ or whatever? Feb 24 22:32:21 Tartarus, $ ls tmp/deploy/ Feb 24 22:32:22 sources Feb 24 22:32:35 so the answers seems to be "no" :) Feb 24 22:35:04 ok Feb 24 22:35:27 can you email /pastebin me your whole set of changes? Feb 24 22:35:56 Tartarus, which ones? the machine config, kernel defconfig, etc.? Feb 24 22:36:06 Everything you've got different :) Feb 24 22:36:23 if you want to do a temp commit, git send-email and then uncommit that's fine Feb 24 22:36:41 Tartarus, it'¡s everything here -> http://shr-project.org/trac/wiki/exoPCSlate#OEintegration Feb 24 22:36:53 I can send you the whole thing as a patch if you prefer though Feb 24 22:37:00 yeah, send me a patch pls Feb 24 22:37:04 ok :) Feb 24 22:37:13 I'll see if I can reproduce that bit you're seeing Feb 24 22:37:14 Tartarus, which mail addr¿ Feb 24 22:37:24 tom_rini@mentor.com Feb 24 22:37:27 ok! Feb 24 22:39:12 Tartarus, is git diff --cached enough for you? :) Feb 24 22:39:44 if everything is in the index, yes :) Feb 24 22:40:07 nah wait, I'll use format-patch + send-email Feb 24 22:41:02 hrms I think now I fixed it Feb 24 22:42:58 Tartarus, sent :) Feb 24 22:43:05 thanks for looking at it! Feb 24 22:43:13 hm or not Feb 24 22:44:33 okay good nite Feb 24 22:44:42 will check it in tomorrow Feb 24 22:56:53 Tartarus, I'm going to sleep, gnight and thanks :) Feb 24 23:03:25 03Khem Raj  07master * re12b55f89f 10openembedded.git/recipes/gcc/ (gcc-4.5.inc gcc-4.5/tune-xscale-infinite-loop-PR45177.patch): Feb 24 23:03:25 gcc-4.5: Fix GCC found on xscale with Os Feb 24 23:03:25 * This enables linaro patch 45177 and a patch on top Feb 24 23:03:25 to fix an ICE discovered because of this backport Feb 24 23:03:25 Signed-off-by: Khem Raj Feb 24 23:03:39 ah good, thanks Khem Feb 24 23:03:57 Just digging at pespin's problem quick Feb 24 23:04:00 then i'll update the branch Feb 24 23:10:23 ah ha Feb 24 23:10:34 * Tartarus mutters about a thinko Feb 24 23:14:29 * Tartarus tests Feb 24 23:17:09 03Tom Rini  07master * rc25762d71d 10openembedded.git/conf/distro/include/sane-toolchain.inc: Feb 24 23:17:09 sane-toolchain.inc: Fix a thinko in compute_os_portion_of_target_triplet Feb 24 23:17:09 Reported by by Pau Espin Pedrol Feb 24 23:17:09 Signed-off-by: Tom Rini Feb 24 23:17:46 03Andrea Adami  07org.openembedded.dev * r5dbd8d611f 10openembedded.git/recipes/klibc/klibc-1.5.21/fstype-sane-vfat-and-jffs2-for-1.5.patch: Feb 24 23:17:46 klibc: update jffs2 and vfat patch. Fix gcc warnings. Feb 24 23:17:46 Signed-off-by: Andrea Adami Feb 24 23:18:54 OK, updating Feb 24 23:46:38 khem: I have received your Ack directly, but don't see it in patchworks - http://patches.openembedded.org/patch/803/ Feb 24 23:48:27 missing it as well, yeah Feb 24 23:48:34 dunno why, was gonna push it :) Feb 24 23:49:39 Tartarus: so, are you fine with the current patch? do you want to simplify the regexp or combine the 2 patches? any preferences? Feb 24 23:51:33 i'm fine with as is Feb 24 23:57:44 03Denys Dmytriyenko  07master * r2ee7e82e8f 10openembedded.git/recipes/busybox/ (5 files in 3 dirs): (log message trimmed) Feb 24 23:57:44 busybox: fix udhcp check for NFS rootfs for new style /proc/mounts Feb 24 23:57:44 * Newer kernels change the format of /proc/mounts slightly when rootfs is Feb 24 23:57:44 mounted over NFS. Feb 24 23:57:44 * Old style: Feb 24 23:57:45 /dev/root / nfs 0 0 Feb 24 23:57:45 * New style: Feb 24 23:58:30 Tartarus: is release testing branch updated? Feb 24 23:59:52 yes Feb 25 00:03:52 Tartarus: I'll fire testbuilder up then :) Feb 25 00:05:39 * Jay7 -> sleep Feb 25 00:14:23 Tartarus: I was gonna ask you to review my last 3 sysroot patches, but it appears #1 hasn't made it to the list yet... Feb 25 02:06:42 03Denys Dmytriyenko  07master * rebd905685a 10openembedded.git/ (3 files in 2 dirs): (log message trimmed) Feb 25 02:06:42 bitbake.conf, native*.bbclass: pass --sysroot to TOOLCHAIN_OPTIONS Feb 25 02:06:42 Commit 9c5b9db29738a89837917b1a2d8ee5a1cae0b4cc (bitbake.conf: Simplify Feb 25 02:06:42 TARGET_CPPFLAGS/LDFLAGS) potentially breaks external toolchains, when their Feb 25 02:06:42 default sysroot is different from our sysroot/staging location. Feb 25 02:06:43 Pass default --sysroot to compiler/linker through TOOLCHAIN_OPTIONS, avoiding Feb 25 02:06:44 native and nativesdk packages. This matches the corresponding changes in Poky. Feb 25 02:06:51 03Denys Dmytriyenko  07master * r47a8210e4a 10openembedded.git/recipes/gstreamer/gst-ffmpeg_0.10.11.bb: Feb 25 02:06:52 gst-ffmpeg: remove 4-year old hack that sets CC directly Feb 25 02:06:52 The hack that sets CC variable directly is 4 year old and no longer required. Feb 25 02:06:52 On the other hand, it now breaks setups when --sysroot needs to be passed to Feb 25 02:06:52 the toolchain. Hence, remove the hack. Feb 25 02:06:52 Signed-off-by: Denys Dmytriyenko Feb 25 02:06:53 Acked-by: Tom Rini Feb 25 02:06:56 03Denys Dmytriyenko  07master * r93bb226b98 10openembedded.git/recipes/ffmpeg/ (ffmpeg_0.5.bb ffmpeg_0.6.1.bb ffmpeg_git.bb): Feb 25 02:06:56 ffmpeg: pass --sysroot to configure script directly Feb 25 02:06:56 ffmpeg's custom configure script fails some checks for specific headers, even Feb 25 02:06:56 though we pass --sysroot through CFLAGS. Use configure's specific --sysroot Feb 25 02:06:56 option to fix that. Feb 25 02:06:56 Signed-off-by: Denys Dmytriyenko Feb 25 02:06:56 Acked-by: Tom Rini **** ENDING LOGGING AT Fri Feb 25 02:59:57 2011