**** BEGIN LOGGING AT Tue Nov 25 02:59:57 2008 Nov 25 05:57:59 Hello, I get this error during the configure: configure.in:29: error: possibly undefined macro: LT_CURRENT Nov 25 05:57:59 http://pastebin.com/d215938f4 . Any idea what I should do? Nov 25 05:57:59 I have "inherit autotools pkgconfig" and nothing for do_configure. The package compiles on the host machine Nov 25 05:58:21 s/nothing/nothing\ special/ Nov 25 07:05:04 * * OE Bug 4871 has been created by openembedded-dpe(AT)temporaryforwarding.com Nov 25 07:05:06 * * libxml does not build with Ubuntu Intrepid Nov 25 07:05:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4871 Nov 25 07:51:01 03Jeremy Lainé  07org.openembedded.dev * r0f7e675c3d 10openembedded.git/packages/opensync/ (10 files in 2 dirs): Nov 25 07:51:01 wbxml2: update to version 0.9.2 Nov 25 07:51:01 * package latest upstream release Nov 25 07:51:01 * fix description and license (LGPL, not GPLv2) Nov 25 07:51:01 * apply patches included in Debian package Nov 25 08:26:41 botmail read Nov 25 08:26:45 ~botmail read Nov 25 08:27:02 morning Nov 25 08:27:05 ~seen pb Nov 25 08:27:07 pb was last seen on IRC in channel #tomcat, 32d 13h 43m 42s ago, saying: 'arghh'. Nov 25 08:27:16 ~seen pb_ Nov 25 08:27:17 pb_ is currently on #oe (3h 38m 35s) #openmoko (3h 38m 35s) #elinux (3h 38m 35s), last said: 'germann: that would be a question for #angstrom, I suppose'. Nov 25 08:44:02 hi lorn Nov 25 08:44:27 hi woglinde Nov 25 08:45:54 ping kergoth Nov 25 08:46:04 he pb Nov 25 08:46:11 pb yes I goggled after this bug Nov 25 08:46:15 and foud it too Nov 25 08:46:26 its the normal gcc-4.2 in angstroem Nov 25 08:47:04 All, I get this error in configure: configure: error: conditional "am__fastdepCXX" was never defined. Nov 25 08:47:04 Any idea how I can avoid such problem? I google it but nothing interesting Nov 25 08:47:29 gregoiregentil hm which package? Nov 25 08:48:00 A package that I'm creating (a git version of cairo) Nov 25 09:14:44 OK. I miraculously fixed the problem. Changing a random CXX (configure files are chinese to me!) Nov 25 09:15:32 I have an easy question now: how can I force to add an include path to my gcc cross compilation? Nov 25 09:16:09 gregoiregentil oe or general? Nov 25 09:16:19 oe Nov 25 09:16:38 in a recipe. Something like EXTRA_OEMAKE Nov 25 09:16:38 setting CFLAGS += Nov 25 09:16:43 or Nov 25 09:16:43 OK Nov 25 09:16:47 CXXFLAGS Nov 25 09:16:54 there are some recipes using i Nov 25 09:16:55 t Nov 25 09:16:57 just grep Nov 25 09:17:04 OK. thanks. Nov 25 09:54:30 morning Nov 25 09:55:16 mogning Nov 25 09:56:53 hi hrw Nov 25 10:12:54 hi xora Nov 25 10:13:03 hey woglinde Nov 25 10:13:10 * XorA is freezing Nov 25 10:16:11 you are not the only one Nov 25 10:18:05 * * OE Bug 4872 has been created by david(AT)tuxbrain.com Nov 25 10:18:07 * * packages/e17/illume_svn.bb file doesn't fetch source from svn Nov 25 10:18:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4872 Nov 25 10:43:51 http://www.youtube.com/watch?v=45LtTe9pUfI Nov 25 12:18:32 03Koen Kooi  07org.openembedded.dev * ra49777cf04 10openembedded.git/packages/e17/e-wm_svn.bb: e-wm: attempt to unbreak uclibc builds Nov 25 13:14:00 elo mickey|ezx Nov 25 13:16:02 morning hrw Nov 25 13:19:14 does someone here tried to sync with opensync? Nov 25 13:21:13 hi zecke Nov 25 13:21:34 moin Nov 25 13:25:44 hey mickey|ezx Nov 25 13:25:49 What about these commit rights? :-) Nov 25 13:26:08 soon Nov 25 13:26:11 (by the way, I restarted to build the image from scratch, git-native and python-etk are failing) Nov 25 13:26:18 that's sad Nov 25 13:26:20 git-native because the patch doesn't apply Nov 25 13:26:24 and python-etk for the eina dependency Nov 25 13:26:31 Corrected locally anyway Nov 25 13:26:38 strange, should both be fixed iirc Nov 25 13:26:46 anyways, working on other things today :) Nov 25 13:26:50 Maybe I didn't update my local repository Nov 25 13:27:00 mickey|ezx: Are all of raster patches applied now? Nov 25 13:27:05 pretty much, ya Nov 25 13:27:11 Pretty much means yes or no? :-) Nov 25 13:27:17 means i don't know Nov 25 13:27:19 stefan_schmidt should know Nov 25 13:27:25 Ok I'll ask him then :-) Nov 25 13:27:32 stefan_schmidt: ping Nov 25 13:28:05 Ainulindale: git native is still failing... Nov 25 13:28:17 denix: do you plan to remove the patch? Nov 25 13:28:34 zecke: yeah, I saw that this morning Nov 25 13:28:37 fwiw, lets remove 1.5.x Nov 25 13:28:43 i don't see why we should keep it Nov 25 13:28:59 mickey|ezx: hehe, talk to to pb_ on removing "old" recipes :) Nov 25 13:29:12 And by the way Nov 25 13:29:33 Do you think there should be a way to remove all the data related to packages arch-dependent switching to non arch-dependent? Nov 25 13:29:54 (Had this problem with one of my task, which I switched from PACKAGE_ARCH="${MACHINE_ARCH}" to nothing Nov 25 13:29:57 ) Nov 25 13:30:01 zecke: i know his opinion Nov 25 13:30:04 but i have mine as well Nov 25 13:30:09 and it happens to be different Nov 25 13:30:10 hehe Nov 25 13:32:24 ah okay git-native was fixed... and koen even changed e-wm Nov 25 13:32:39 Good :-) Nov 25 13:32:53 Still missing this eina/python-etk dependency then? Nov 25 13:33:15 Ainulindale: is your build failure on the oestats server? Nov 25 13:33:27 Yep Nov 25 13:33:48 http://tinderbox.openembedded.net/public/logs/2235771.txt Nov 25 13:33:55 Ah, damn, not this one Nov 25 13:34:12 http://tinderbox.openembedded.net/public/logs/2235653.txt <= this one Nov 25 13:34:22 hello; how exactly is it decided what package definitely goes into resulting image? I have one image where iconv is installed, although it is pulled by glib-2.0, which is not being installed by default Nov 25 13:34:40 (and I'm not aware of any other dependency that could pull it into image) Nov 25 13:34:52 mmp: well, opkg info iconv will tell you that Nov 25 13:35:06 (or opkg remove iconv, anyway) Nov 25 13:35:10 Ainulindale, well, I don't have iconv Nov 25 13:35:15 and theargh, opkg :) Nov 25 13:35:33 the problem is that the image itself now doesn't fit to flash Nov 25 13:35:38 WEll, you just told us iconv was installed Nov 25 13:35:47 Try to see the dependency graph then Nov 25 13:35:47 Ainulindale, sorry typo ;) Nov 25 13:35:53 Ainulindale: eina... I assume it was found Nov 25 13:36:02 Ainulindale: s/found/build/ Nov 25 13:36:02 zecke: found? The bug? Nov 25 13:36:06 No Nov 25 13:36:08 iconv is installed, I don't have opkg installed, as it pulls another quite big dependencies Nov 25 13:36:11 I knew this problem zecke Nov 25 13:36:19 It has been this way for two weeks IIRC Nov 25 13:36:27 I just built eina, then cleaned python-etk, rebuilt Nov 25 13:36:29 And it worked Nov 25 13:36:47 Ainulindale: eet depends on eina, evas depends on eet, python-evas depends on evas, python-etk depends on python-evas Nov 25 13:37:03 zecke: then either I'm not up to date or something is borking Nov 25 13:37:43 Ainulindale: e.g. evas is unlikely to build without Eina (checks that) Nov 25 13:38:05 zecke: hmmm Nov 25 13:38:12 python-etk has a RDEPENDS on python-evas Nov 25 13:38:17 Shouldn't it be DEPENDS? Nov 25 13:38:29 (if the dependency you're talking about is right) Nov 25 13:38:33 (that or eina in the DEPENDS) Nov 25 13:38:37 mickey|ezx: is there a way to install all python modules in one go? Nov 25 13:38:53 good point Nov 25 13:39:15 Ainulindale: but then python-etk is including python-efl.inc Nov 25 13:39:21 Ainulindale: which has eina in DEPENDS :) Nov 25 13:39:37 Ainulindale: I suspect python-etk needs a pkg-config --cflags eina Nov 25 13:39:50 Then why would it work after a rebuild? Nov 25 13:39:51 hrw: no, might be a good addition to the generate-manifest script Nov 25 13:40:00 And anyway, why doesn't it build eina anyway before python-etk? Nov 25 13:40:05 mickey|ezx: with perl it is just 'perl-modules' Nov 25 13:40:12 like with ernel Nov 25 13:41:28 Ainulindale: it built eina Nov 25 13:41:34 Ainulindale: otherwise evas does not build Nov 25 13:41:40 Ainulindale: and you have the evas headers installed Nov 25 13:42:14 Ainulindale, looks like gnutls depends on it somehow (quite strange dependency, though) Nov 25 13:42:31 zecke: well, bitbake eina resulted in a build Nov 25 13:42:42 It shouldn't have if the dependencies were managed correctly, wouldn't it? Nov 25 13:42:53 (Or I missed something) Nov 25 13:43:12 zecke: well anyway Nov 25 13:43:13 http://tinderbox.openembedded.net/builders/shr.bearstech.com/ Nov 25 13:43:15 see the log Nov 25 13:43:30 (Seems you're right anyway) Nov 25 13:43:33 (It's listed as built) Nov 25 13:43:41 (Could it be a staging problem?) Nov 25 13:47:08 morning all Nov 25 13:49:18 Ainulindale: and for the third time. it is more a missing pkg-config --cflags eina in python etk... Nov 25 13:50:07 zecke: What bugs me with that, is why it would work after a bitbake eina then Nov 25 13:52:14 03Robert Schuster  07org.openembedded.dev * r46a903ec13 10openembedded.git/packages/classpath/ (classpath-0.97.2/gtk-fix.patch classpath_0.97.2.bb): classpath 0.97.2: Added gtk performance fix. Nov 25 13:52:15 03Robert Schuster  07org.openembedded.dev * r6b6df778d8 10openembedded.git/packages/antlr/ (antlr-native_2.7.5.bb antlr_2.7.5.bb files/install.patch): antlr 2.7.5: Moved to obsolete since it does not build. Nov 25 13:52:16 03Robert Schuster  07org.openembedded.dev * re25498c050 10openembedded.git/packages/antlr/ (antlr-native_2.7.7.bb antlr_2.7.7.bb files/runantlr): Nov 25 13:52:19 Replace non-working antlr recipe with working ones from Jalimo repo. Nov 25 13:52:21 antlr 2.7.7: New recipe. Nov 25 13:52:23 antlr-native 2.7.7: New recipe. Nov 25 13:53:56 03Robert Schuster  07org.openembedded.dev * r01d8c64f7b 10openembedded.git/conf/distro/include/preferred-om-2008-versions.inc: preferred-om-2008-versions.inc: Set preference for antlr[-native} to 2.7.7. Nov 25 14:07:47 Ainulindale: no idea Nov 25 14:17:17 is there any way to get rid of libiconv for all? Nov 25 14:17:52 USE_NLS="0" or such in the config Nov 25 14:18:02 zecke: it's set Nov 25 14:18:15 glib-2.0 depends on virtual/iconv Nov 25 14:18:26 and the only provider for it (at least in my setup) is iconv itself Nov 25 14:19:24 problem is that some packages, finding iconv in staging, will automagically use it (e.g. libgpg-error), creating unwanted dependencies Nov 25 14:20:48 mmp: it looks like you'd need to provide a patchset to properly support it. Nov 25 14:21:15 mmp: i believe many people would like to get it however people not wanting it shouldn't be affected (like me :P) Nov 25 14:21:30 mmp: but this is indeed a good thing to be properly supported. Nov 25 14:21:55 well, I can at least fix libgpg-error... Nov 25 14:22:23 mmp: maybe you might build a package that provided a null-iconv. something like: fake-iconv.bb Nov 25 14:22:46 mmp: and it could provide virtual/iconv; so you could take it using a preffered version Nov 25 14:23:12 mmp: that way I think you solve your problem in an easy way while to not make noise for others ;-) Nov 25 14:23:21 mmp: give it a try and see if it works for you. Nov 25 14:23:59 yes, it's possible, but it is a bit weird... why there is virtual/iconv, if the only package providing it is iconv itself... Nov 25 14:24:12 I guess I should grep packages/ for virtual/iconv :) Nov 25 14:24:19 mmp: i believe glibc provides iconv too Nov 25 14:24:31 mmp: git grep is your friend ;-) Nov 25 14:24:36 :-) Nov 25 14:24:39 mmp: 'git grep virtual/iconv' Nov 25 14:24:56 mmp: faster and easier Nov 25 14:25:14 mmp: so it greps only the metadata; ignoring your tmp and like Nov 25 14:25:33 hmm, nice feature, didn't know about this one Nov 25 14:25:41 mmp: git is neat ;-) Nov 25 14:25:54 * otavio loves git Nov 25 14:36:26 Talking about git Nov 25 14:36:36 DOes anyone know if there's a way to unpack only a subdirectory of a git repo? Nov 25 14:36:53 (such as module for svn) Nov 25 14:38:53 03Robert Schuster  07org.openembedded.dev * r2ed3c60d8b 10openembedded.git/packages/lesstif/lesstif_0.95.0.bb: lesstif 0.95.0: Provide faked -x-include argument to fix compilation of configure tests. Nov 25 14:42:12 03Jeremy Lainé  07org.openembedded.dev * r00c45ab409 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Nov 25 14:42:13 03Jeremy Lainé  07org.openembedded.dev * r34f28f6963 10openembedded.git/packages/dnsmasq/ (dnsmasq.inc dnsmasq_2.46.bb): dnsmasq: create ${sysconfidir}/dnsmasq.d config directory Nov 25 15:01:30 gregoiregentil: pong Nov 25 15:12:19 does busybox have an irc channel? Nov 25 15:17:37 03Jeremy Lainé  07org.openembedded.dev * r7d8900bcb1 10openembedded.git/conf/machine/boc01.conf: boc01: fix erase block size for jffs2 Nov 25 15:22:20 gregoiregentil: i've seen that error before, but its been years, having trouble remembering the cause. thinking it may have had something to do with the use of automake conditionals Nov 25 15:22:24 gregoiregentil: not certain Nov 25 15:47:33 i found a nice issue considering the mount implementation of util-linux and busybox ... when my system boots from a non volatile media (harddisk / flash) i can do mount --bind /a /b quiet well (udev for instance needs this) but as soon as i boot the system with the same rootfs from an initramfs the mount command will supply me with a nice invalid argument error Nov 25 15:47:42 anybody has seen such behavior before? Nov 25 15:50:25 AvengerMoJo: no, it is not possible for now. AFAIK they're working to get it to be supported in future. Nov 25 16:24:10 03Jeremy Lainé  07org.openembedded.dev * r22ea6e5fbe 10openembedded.git/packages/dnsmasq/dnsmasq.inc: dnsmasq: fix permissions on dnsmasq.conf (644, not 755) Nov 25 16:31:34 03Jeremy Lainé  07org.openembedded.dev * r313d0662ee 10openembedded.git/packages/dnsmasq/dnsmasq-dbus_2.46.bb: dnsmasq-dbus: add a version of dnsmasq with DBus support Nov 25 16:54:13 03Michael 'Mickey' Lauer  07org.openembedded.dev * rd930861216 10openembedded.git/packages/ (python/python-rpyc_3.00.bb tasks/task-python-everything.bb): python-rpyc: new recipe; remote procedure call package for python Nov 25 17:37:00 kergoth: thanks for the tip. Yes, I saw a thread on google about it mentioning your name. I played with my configure.in (adding some CXX) and managed to fix it. Thanks... Nov 25 17:37:36 ah, it was probably doing something based on a conditional that was never set up, or something Nov 25 17:37:38 np Nov 25 17:49:06 * * OE Bug 4861 has been RESOLVED (FIXED) by thommycheck(AT)gmx.de Nov 25 17:49:08 * * checksums.ini update for xqt2 and blackbox Nov 25 17:49:10 * * http://bugs.openembedded.net/show_bug.cgi?id=4861 Nov 25 17:50:04 * * OE Bug 4582 has been RESOLVED (FIXED) by thommycheck(AT)gmx.de Nov 25 17:50:06 * * X/Qt2 fixes (screen rotation screen resolution) Nov 25 17:50:08 * * http://bugs.openembedded.net/show_bug.cgi?id=4582 Nov 25 17:50:16 * * OE Bug 4603 has been RESOLVED (FIXED) by thommycheck(AT)gmx.de Nov 25 17:50:18 * * xqtlauncher integrates X/Qt2 nicely into opie Nov 25 17:50:20 * * http://bugs.openembedded.net/show_bug.cgi?id=4603 Nov 25 17:52:19 gregoiregentil: i'm wondering if OpenEmbedded should have more things like that in a buildsystem tips area of our wiki or something. common problems with specific buildsystems, tips for conversion of old autoconf version configure.in's to new style configure.ac's, or at least links to resources on such things.. hmm Nov 25 17:52:32 gregoiregentil: given a good % of what we do is deal with buildsystem issues Nov 25 17:52:45 heh Nov 25 17:53:17 kergoth sounds like a good idea Nov 25 17:53:46 just seems silly to not share that sort of information, those sorts of lessons learned Nov 25 17:54:07 kergoth please add some stuff to the wiki :) Nov 25 17:54:23 * kergoth rolls eyes Nov 25 17:54:41 i'd be lucky to remember 10% of the stuff that needs to be documented, too rusty Nov 25 17:54:47 heh Nov 25 17:55:35 kergoth every thing added to the wiki is better than nothing even if it is just a structure for others to fill out later Nov 25 17:55:39 kergoth: oh yeahhh! I do agree. But from a certain point of view, google is your wiki Nov 25 17:55:58 yeah, thats true, a lot of the info is out there, but it can be scattered Nov 25 17:56:10 yes. I agree Nov 25 18:29:19 What are all of the uses of conditionals in recipes (not configuration files) in OE, today? distributionisms, machine specific patches (rare), machine specific configuration files.. Nov 25 18:29:47 variants of the package (change autoconf flags based on desired features, etc).. Nov 25 18:36:16 bbl Nov 25 18:50:22 re Nov 25 18:58:31 woglinde hi Nov 25 18:58:40 hi mr_nice Nov 25 18:58:49 now you can commit all the stuff yourself Nov 25 19:05:53 woglinde hm, maybe there is no mail about sending keys or something. but it would be awesome :) Nov 25 19:06:14 mr_nice hm Nov 25 19:06:16 okay Nov 25 19:06:23 then I puzzeld it Nov 25 20:45:23 Hello. I just launched x11-gpe-image on magician, there is a lot of work , right? virtual keyboard does not insert a character, no hardwarekeys configuration, system info fails... Nov 25 21:03:01 Hey guys Nov 25 21:03:14 For PREFERRED_VERSION, in the case of _svn packages, putting "svn" will work as well? Nov 25 21:04:44 i'd say check the version .inc's in openembedded/conf/, i'm sure at least some of them prefer svn versions of things Nov 25 21:04:52 pretty sure you'd have to put the PV, the same way you do for everything else. so svn+${SRCREV} or whatever it is for that package, but i'm not certain Nov 25 21:06:21 Hmmmm Nov 25 21:07:22 kergoth: You were right it seems :-) Nov 25 21:12:45 hoi kergoth Nov 25 23:35:01 03Khem Raj  07org.openembedded.dev * rae274a5269 10openembedded.git/packages/uclibc/uclibc_0.9.30.bb: uclibc_0.9.30.bb: Use renamed ldso_hash.patch for avr32 Nov 25 23:37:32 khem it isnt working anyway Nov 25 23:37:51 avr32 binutils does not have gnu-hashstyle Nov 26 02:57:04 03Junqian Gordon Xu  07org.openembedded.dev * r4fe6b38d07 10openembedded.git/conf/checksums.ini: checksums.ini: add linux-rp-2.6.24 **** ENDING LOGGING AT Wed Nov 26 02:59:57 2008