**** BEGIN LOGGING AT Mon Apr 13 02:59:58 2009 Apr 13 08:13:33 hello I've got problems with getopt from my platform when used by oe Apr 13 08:13:54 quilt-native do not configure saying it does not handle long options Apr 13 08:14:34 it also says I can use the quilt provided version configuring it with --without-getopt Apr 13 08:15:19 but then there might be a problem with the quilt provided script Apr 13 08:15:40 so either I correct the one from my platform, either I correct the one provided by quilt in oe Apr 13 08:16:35 it seems to me the latter would be better and adding --without-getopt to the configure in oe as this would abstract more from the host platform Apr 13 08:16:41 what do you think ? Apr 13 08:39:02 when i try to compile libxcalibrate: i get the error: xcalibrate.c:233: error: 'xXCalibrateScreenToCoordReq' undeclared (first use in this function) Apr 13 08:39:18 i also get the same error when i try to compile virtual/xserver Apr 13 08:44:36 does anyone know what could be the problem? Apr 13 09:06:50 * Jay7 needed zaurus'es touchscreen or tslib guru Apr 13 09:07:00 seems hid-driver in 2.6.26 at least does not emit BTN_TOUCH for TS touching Apr 13 09:07:17 That's not guaranteed. Apr 13 09:07:59 tslib should be able to cope with either that or ABS_PRESSURE. Apr 13 09:15:35 well... Apr 13 09:16:02 broonie: but I'm not sure that bringing up tslib into kexecboot is good idea.. Apr 13 09:16:35 Either way, the point is you can't rely on BTN_TOUCH. Apr 13 09:17:03 ok, I'll look at tslib sources then :) Apr 13 09:35:29 answer is simple - just wait for release Apr 13 10:14:45 hello, I've my Package.gz that is not beeing created with bitbake package-index...what could be the cause Apr 13 10:14:45 ? Apr 13 10:20:54 inside /home/embedded/oetmp/deploy/glibc/ipk/arm-oabi it doesn't get created but inside /home/embedded/oetmp/deploy/glibc/ipk/arm-oabi it is...strange Apr 13 10:22:18 anybody with zaurus c1000 here ? Apr 13 10:28:16 where is ipkgarchs set? Apr 13 10:29:21 good morning Apr 13 10:33:14 found it Apr 13 12:09:20 any good video streaming applications which uses gstreamer on oe? Apr 13 12:30:25 03Koen Kooi  07stable/2009 * r697c7afd84 10openembedded.git/recipes/bluez/ (bluez4_4.31.bb files/bluetooth.conf): Apr 13 12:30:25 bluez4: add less restrictive (and hence less secure) bluetooth dbus config Apr 13 12:30:25 * this makes bluetooth actually work for non-root users and when using ssh/vnc/etc Apr 13 12:30:25 * this also overwrites the bluetooth.conf in the files/ that wasn't used anymore Apr 13 12:30:27 * this also includes the fixup commit from .dev named "bluez4: fix dbus service file" Apr 13 12:30:29 Signed-off-by: Koen Kooi Apr 13 12:30:31 Acked-by: Khem Raj Apr 13 12:53:08 hi Apr 13 12:53:59 hi Apr 13 12:54:10 how do I prevent the kernel from beeing compiled? Apr 13 12:54:25 why would you ? Apr 13 12:54:40 if you want to you can compile individual packages Apr 13 12:54:44 because for now I want a chroot image Apr 13 12:55:01 or how do I compile it with an external toolchain Apr 13 12:55:09 but the userland with oe's toolchain Apr 13 12:55:25 ASSUME_PROVIDED += "virtual/kernel" Apr 13 12:55:49 ok thanks a lot... Apr 13 12:56:11 if you build the kernel with oe, the patched source will end up in your work dir and you can copy that and compile with whatever you want to (after doing make clean) Apr 13 12:56:11 or you might need to put both linux and virtual/kernel Apr 13 12:57:45 eFfeM, i've already a recipe for the kernel Apr 13 12:58:06 the only problem is that I've proprietary modules...so I bet I can't change the compiler... Apr 13 12:58:58 i guess you can by hacking the recipe and redefining CC or GCC or so Apr 13 12:59:30 ah ok thanks a lot Apr 13 13:00:25 exporting it within the recipe...and the external toolchain can not be relocated...so it's easy thanks.... Apr 13 13:09:40 the assume provided doesn't work Apr 13 13:09:48 I'll try with the external toolchain Apr 13 13:16:01 <_cpo_> anyone knows for a source of uboot for the lite5200 eval board? Apr 13 13:16:21 <_cpo_> im sorry for the OT question Apr 13 13:21:12 _cpo_: google? ask in a powerpc channel? Apr 13 13:24:06 how do I filter the kernel's CFLAGS? Apr 13 13:26:53 Gnutoo: by putting something like this in the .bb file: Apr 13 13:26:54 CXXFLAGS := "${@oe_filter_out('-fvisibility-inlines-hidden', '${CXXFLAGS}', d)}" Apr 13 13:27:14 but then of cours ewith CFLAGS and the option you want to filter out Apr 13 13:27:22 thanks a lot Apr 13 13:27:34 yw Apr 13 13:27:37 I did put export CFLAGS = "" but it didn't work Apr 13 13:28:35 not sure how it fully works Apr 13 13:29:23 if it doesn't work I would have not to build the kernel Apr 13 13:30:01 or I could also try with a python replace if it doesn't work Apr 13 13:51:12 doesn't work Apr 13 13:51:25 I'll try extra oemake Apr 13 13:54:10 s/extra/ Apr 13 13:54:16 I'll override do_compile Apr 13 14:02:01 anybody with zaurus c1000 here ? Apr 13 14:10:55 the kernel doesn't compile Apr 13 14:11:01 assume provided doesn't work Apr 13 14:11:12 how do I bypass the kernel compilation? Apr 13 14:11:36 I'll try assume provided of my particular kenrel Apr 13 14:11:54 flameman, I don't have one...but better ask the question directly Apr 13 14:16:50 i am hacking the hw, wondering about the IO connector on the back: any doc around ? Apr 13 14:17:15 i'd like to add an uart/rs232 (console) to c1000 Apr 13 14:17:33 wondering where is the ttyS0 routed in that connector Apr 13 14:19:17 how do I bypass the kernel compilation? Apr 13 14:19:53 bypass ? Apr 13 14:20:46 if I bitbake minimal-image it wants to build the kenrel Apr 13 14:20:49 and It fails Apr 13 14:20:57 so I don't want to build it Apr 13 14:21:03 and I don't need to build it Apr 13 14:21:16 cause I want a chroot Apr 13 14:21:17 to try Apr 13 14:23:14 flameman, do you know #angstrom? maybe they know better or there could be other channels maybe... Apr 13 14:25:17 #zaurus Apr 13 14:26:04 XorA: ping Apr 13 14:26:21 XorA: did you have zaurus serial console cables? flameman needs one :) Apr 13 14:26:34 * Jay7 remember that there was some donated Apr 13 14:28:50 anyone? Apr 13 14:28:55 I'm desesperating Apr 13 14:29:05 should I make a dummy kernel? Apr 13 14:30:09 anyone? Apr 13 14:30:15 :( Apr 13 14:31:44 Jay7: I have one Apr 13 14:32:01 Jay7: was Laibsch (Rolf) I think who ordered them from USA Apr 13 14:32:25 ah, yes, Laibsh.. Apr 13 14:33:01 well.. seems he is not on irc this time Apr 13 14:33:01 Jay7: I would like to not post mine, I want to get collie working one day :-) Apr 13 14:33:15 XorA: np :) Apr 13 14:33:52 I abandonned Apr 13 14:34:00 I will build the default kenrel Apr 13 14:34:11 s/kenrel/kernel Apr 13 14:34:27 but it will be a mess Apr 13 14:37:09 Jay7: probably best email oe devel and ask if anyones is spare Apr 13 14:37:16 Jay7: lot of people stopped supporting the Z Apr 13 14:37:18 flameman: ^^ Apr 13 14:38:35 Jay7 that guy didn't add an other uart port to the c1000 Apr 13 14:38:50 jay7 he simply used the second uart in order to add wifi Apr 13 14:39:34 flameman: no, wifi was soldered to free usb port Apr 13 14:39:38 jay7 i ordered a special I/O connector-usb-a from ebay .... it is for c1000, this way i will be able to pinout the connector Apr 13 14:40:14 jay7 well it is what i said: he soldered the USB port, he didn't soldered the uart(serial line) ! Apr 13 14:41:09 flameman: bluetooth was soldered to free uart Apr 13 14:41:22 ttyS2 Apr 13 14:41:30 bluetooth is usb Apr 13 14:41:57 flameman: yes, but he uses only data lines and solder it to btuart of pxa270 :) Apr 13 14:41:58 anyway i will investigate it by myself :P Apr 13 14:42:07 I have asked him some time ago :) Apr 13 14:42:20 ah, ok ... not so clear from his page Apr 13 14:42:28 directly to pxa, ok Apr 13 14:43:12 well very good, so .. i can arrange a specific hw driver in order to export uart to the rs232 line Apr 13 14:47:51 hi all ! Apr 13 14:48:04 what's a good distro for NGW100 board ? Apr 13 14:59:43 re Apr 13 15:00:52 how should I pass library options to ld via oe_runmake ? Apr 13 15:01:05 can I just add -lgettext ? Apr 13 15:02:27 linux-libc-headers_2.6.22.bb seems to miss gettext library Apr 13 15:09:09 in fact it misses -lintl Apr 13 15:10:32 ptitjes, maybe with a SOMETHIN += " -lgettext" ? Apr 13 15:11:01 Gnutoo: yeah but what SOMETHIN... ? :) Apr 13 15:11:47 ptitjes, maybe LDFLAGS? Apr 13 15:12:04 I'm trying LDADD now Apr 13 15:12:51 ok Apr 13 15:13:07 or LDFLAGS_prepend Apr 13 15:15:02 or MAKEFLAGS Apr 13 15:16:52 03Frans Meulenbroeks  07org.openembedded.dev * r0557aecd17 10openembedded.git/recipes/zope/python-zopeinterface_3.5.1.bb: python-zopeinterface: fixed building Apr 13 15:20:52 Gnutoo: I do that: oe_runmake MAKEFLAGS="-lintl" allnoconfig ARCH=$ARCH Apr 13 15:21:08 which produces: NOTE: make -e MAKEFLAGS= MAKEFLAGS=-lintl allnoconfig ARCH=arm Apr 13 15:21:21 but this does no do it :( nor LDFLAGS Apr 13 15:21:56 sorry I don't know then Apr 13 15:22:07 Gnutoo: do you think I should directly patch scripts/kconfig/conf ? http://pastebin.com/m83a5809 Apr 13 15:22:50 don't know...but I've a recipe that pass flags to make: Apr 13 15:22:54 Google showed me someone doing that for cygwin Apr 13 15:23:27 Gnutoo: yeah what is that recipe ? Apr 13 15:23:30 http://rafb.net/p/lOl6Gz20.html Apr 13 15:24:59 haha Apr 13 15:25:06 trying that immediatly Apr 13 15:36:24 without success :( Apr 13 15:40:41 :( Apr 13 15:40:57 it seems that everybody has some problem today Apr 13 15:41:09 (I had things that didn't work too...) Apr 13 15:41:21 and I abandoned the idea of making them work Apr 13 16:17:46 Gnutoo: succeeded Apr 13 16:17:58 wow... Apr 13 16:18:01 what did you do? Apr 13 16:18:22 patched scripts/kconfig/Makefile Apr 13 16:18:26 90a91,92 Apr 13 16:18:26 > HOSTLOADLIBES_conf = -lintl Apr 13 16:18:26 > Apr 13 16:18:34 :D Apr 13 16:19:17 I'm at glibc now... still building Apr 13 16:27:25 ok Apr 13 16:29:22 I've some problem...my hello world works with oe's toolchain but not a statically linked gdbserver Apr 13 16:29:37 I use a 2.4.19 kernel with 2.6 headers Apr 13 16:29:41 I was told it was ok Apr 13 16:30:10 I'm trying to cross-compile gdbserver by hand with the vendor's toolchain but it doesn't work Apr 13 16:33:23 ok found a wokaround Apr 13 16:39:46 ptitjes: eh, you don't need libintl with glibc, only with uclibc. see gettext.bbclass. Apr 13 16:40:00 ptitjes: if you're hardcoding that -lintl, you're sacrificing portability Apr 13 16:45:10 i am trying to get xtscal working on xorg-xserver. the xcalibrate extention init was not called in the function which calls all the extentions init. but even if i call xcalibrate extention init it does not get included in the list of extentions Apr 13 16:45:27 does any one know/has got xtscal working with xorg-xserver? Apr 13 16:45:38 by the way how can I tell if a binary is hard float or soft float under oabi? Apr 13 16:47:19 if I compiled ldd with hardfloat and that it runs on the target does it mean that it's a hardfloat Apr 13 16:47:20 ? Apr 13 16:54:47 any know who is familiar with xorg-xserver here? Apr 13 17:12:42 kergoth: that is the linux headers that fails Apr 13 17:13:01 kergoth: the conf to be more exact. see that config log: http://pastebin.com/m83a5809 Apr 13 17:14:41 kergoth: I noticed that linux-libc-headers are builded before glibc, so how should I do than patch scripts/kconfig/Makefile ? Apr 13 17:15:38 kergoth: also I've seen on google that some people must do this to build the kernel headers on Cygwin. This might be related in some way. Apr 13 17:24:53 kergoth: any thoughts on this ??? Apr 13 17:37:37 hello, what does that mean? Inconsistency detected by ld.so: dynamic-link.h: 62: elf_get_dynamic_info: Assertion `! "bad dynamic tag"' failed! Apr 13 17:37:45 I've 3 kinds of binaries: Apr 13 17:37:59 *hello world => they work Apr 13 17:38:08 *illegal instruction programs Apr 13 17:38:12 *and that one Apr 13 17:38:20 strange isn't it Apr 13 17:39:01 I use 2.6 headers with 2.4.19 kernel...I heard it was ok...but is it realy? Apr 13 17:48:10 Hello; can someone help me to figure out why a PREFERRED_VERSION is not working? Apr 13 17:48:29 I've taken a lot at the building log using -DDD for the bitbake Apr 13 17:49:16 And the most interesting issue is that it does looks to be properly choosen on the log; however when I go and check the installed package in the installed-packages.txt it is the wrong version Apr 13 17:59:07 otavio: is there a + in the PV? Apr 13 17:59:26 XorA: no :( Apr 13 17:59:56 XorA: I'm finishing a clear log build to post here Apr 13 18:00:04 re Apr 13 18:01:12 XorA: is PREF_VER broken with + in PV? Apr 13 18:01:40 opkg/ipkg-utils sometimes seems to do odd things with a + Apr 13 18:02:38 ah Apr 13 18:03:01 but not bitbake? i.e. it shouldn't break PREFERRED_VERSION? Apr 13 18:03:07 XorA: http://pastebin.com/f48221acc Apr 13 18:04:34 I was told it was because of nss...because It's statically linked Apr 13 18:04:48 otavio: not sure I understand that log at all Apr 13 18:04:57 but Im off to play games in Vista Apr 13 18:05:05 (nss is dlopened by glibc and so beeing statically linked it can't find it) Apr 13 18:05:31 speaking of which, is there any way to set PREFERRED_PROVIDER for a _runtime_ package dependency? Apr 13 18:05:34 http://pastebin.com/f165d19a4 Apr 13 18:05:36 XorA: http://pastebin.com/f165d19a4 Apr 13 18:05:40 XorA: a more clear one Apr 13 18:05:50 XorA: the P_V is 9.0.95 Apr 13 18:05:52 oops Apr 13 18:05:54 9.0.94 Apr 13 18:06:02 but it installs 9.0.94.20... Apr 13 18:06:16 because thats higher version Apr 13 18:06:30 remember opkg/ipkg does the installs, and it selects highest in feed Apr 13 18:06:33 yeah, that's opkg selecting a newer version Apr 13 18:06:39 wow did someone had problems with gutils.h in glib Apr 13 18:06:40 ? Apr 13 18:09:07 humm Apr 13 18:09:17 otavio: http://wiki.openembedded.net/index.php/Versioning_Policy Apr 13 18:09:22 if opkg does the install how I could "cleanup" the feed? Apr 13 18:09:37 rm should work :) Apr 13 18:09:53 __GNUC_PREREQ is not known... Apr 13 18:10:01 rm /home/otavio/hacking/ossystems/oe/tmp/deploy/glibc/ipk/i686-osthindesktop-lite/osthindesktop-manager-applet-xabout_9.0.94.20 Apr 13 18:13:03 denix: there's no easy way to get it to "morgue" the not preferred ones? Apr 13 18:25:31 otavio: not sure. I believe only older versions go to morgue when newer added to the feed, not the other way around Apr 13 18:28:25 denix: I though that when building it would grab the preferred version for the installation Apr 13 18:28:41 denix: and only if we were doing an update it would install the newest version Apr 13 18:28:50 denix: i mean on the rootfs task Apr 13 18:32:31 ptitjes: you prolly need this patch http://www.nabble.com/-PATCH--*-lib-safe-alloc.h-%28__GNUC_PREREQ%29%3A-define-if-not-already-defined-td22300286.html Apr 13 18:33:55 khem: arf I have gone for this one: http://sources.redhat.com/ml/libc-alpha/2004-03/msg00264.html Apr 13 18:33:59 and this seems to be ok Apr 13 18:36:08 ptitjes: ok Apr 13 18:36:27 khem: should I propose some patches to oe ? Apr 13 18:36:41 is it in gnulib ? Apr 13 18:36:57 in glib Apr 13 18:37:19 otavio: the rootfs is populated by opkg and it doesn't know anything about bitbake variables (like PREFERRED_VERSION) Apr 13 18:38:52 ptitjes: try searching upstream first Apr 13 18:38:57 i'm having an issue using the x11-gpe-image. Everything builds fine and I can boot to it, however, when it starts /etc/init.d/gpe-dm it crashes with an unhandled page fault. I can write directly to the framebuffer and change the color of the screen. Why would gpe crash like that? I'm using the newest OE from git along with an Atmel AT91SAM9261ek board. Apr 13 18:38:58 for glib package Apr 13 18:39:51 khem: this is my third fix in the day Apr 13 18:40:10 I unfortunately won't be able to ping upstream for all Apr 13 18:40:36 as lots of work waits for my oe tree to work Apr 13 18:40:38 :( Apr 13 18:58:09 humm Apr 13 18:58:10 denix: got it Apr 13 18:58:18 * otavio needs to see how to handle it Apr 13 19:01:38 gm Apr 13 19:05:36 hello, when I chroot I have: Illegal instruction...what's the best way to debug that? Apr 13 19:05:51 as I can't chroot I can't install programs Apr 13 19:06:39 denix: is there an easy way to "split" the feeds? Apr 13 19:07:15 denix: like put the .ipk on a two group of feeds, so I could split it up and add it or not, depending on if it is a debug image or not Apr 13 19:11:09 hi all Apr 13 19:11:26 do we still have an openembedded-essentials (for Ubuntu/Debian) somewhere? Apr 13 19:12:00 hi likewise, what u mean ? Apr 13 19:13:35 otavio: you can have 2 separate deploy directories... Apr 13 19:14:10 not sure if you can do that with a single build, unless your are building for different machines/archs Apr 13 19:14:59 denix: my idea is to set it in build time Apr 13 19:15:18 denix: so the ones that are built using our development "snapshot" gets into this feed Apr 13 19:20:45 likewise: hey are you lookin for something like http://wiki.openembedded.net/index.php/OEandYourDistro Apr 13 19:21:26 khem, gremlin[it]: yes, but the wiki page does not mention an apt repository for openembedded-essential anymore, it used to. Apr 13 19:21:58 likewise: I think we have the rpm but not the .deb :( Apr 13 19:22:04 ah i understand ... don't know the details ... i'm a Fedora guy ... Apr 13 19:23:18 ah...hello world also has that...I'll investigate Apr 13 19:24:10 khem: I think hrw and Laibsch hosted their own .deb. Apr 13 19:27:32 otavio: you can probably set the DEPLOY_DIR based on some parameters, like whether it's "dev" build or not Apr 13 19:29:21 likewise: yeah could be Apr 13 19:31:05 can someone please consider http://bugs.openembedded.net/show_bug.cgi?id=5083 Apr 13 19:32:46 denix: humm do you know where to add it for it to be added on the "feeds" set? Apr 13 19:37:14 strange...: http://rafb.net/p/EqeCmz40.html <- is the cause of the illegal instruction here: 0x00008868 <__libc_setup_tls+292>: cmn r0, #4096 ; 0x1000 ? Apr 13 19:37:54 cmn is on evry version of arm...that is strange... Apr 13 19:41:59 yes strange Apr 13 19:42:37 here's the story: Apr 13 19:42:57 I've an old machine with 2.4 kernel and oabi and probably hard float Apr 13 19:43:08 worst case scenario...lol Apr 13 19:43:14 ep93xx ? Apr 13 19:43:18 so I'm trying to change the userland Apr 13 19:43:20 no pma430 Apr 13 19:43:59 I built a toolchain...building a hello world from the devshell and running it on the machine works Apr 13 19:44:25 but not the bitbaken hello world ipk Apr 13 19:44:43 chroot doesn't work(also illegal instruction) Apr 13 19:45:29 as chroot I meant chrooting...(the chroot program comes from the original distro) Apr 13 19:45:31 mhhh i suggest to check the options used to build libc ... Apr 13 19:46:08 gremlin[it], ok...I've built it against a 2.6 kernel headers but I was told it was ok Apr 13 19:47:18 mhhh yes maybe and surely could cause completely different kind of errors Apr 13 19:47:54 mmm...the hello world recipe is static... Apr 13 19:48:04 the other hello world isn't static Apr 13 19:48:06 could we consider bumping pax-utils ? 0.1.13 is actually failing on some plateforms MacOS X, BSD, ... Apr 13 19:48:36 ptitjes: gentoo folks already patched it a lot. and they are at version 0.1.18 Apr 13 19:48:44 gremlin[it], what kind of error...what do you mean? Apr 13 19:48:52 and thanks a lot for the indications Apr 13 19:50:12 Gnutoo, use the "wrong" kernel-headers should not cause illegal instruction ... Apr 13 19:50:33 ok thanks a lot for the info Apr 13 19:52:48 nothing ... Apr 13 19:56:03 the dynamicaly linked does the same thing...maybe it's a gcc flag Apr 13 19:57:21 oops...I forget to remove -static Apr 13 19:58:23 yes gotcha (probably) ! Apr 13 20:03:53 mmm...only dynamic thing work Apr 13 20:04:08 but that's with the old glibc Apr 13 20:06:23 what should I do now? Apr 13 20:06:41 maybe step by step execution? Apr 13 20:14:38 Hi valhalla ! Apr 13 20:15:02 hi gremlin[it] Apr 13 20:15:32 hi valhalla, I tried to update wesnoth but I run into boost bug for arm...I have to try boost's trunk... Apr 13 20:16:39 (I said that because maybe you are still interested in the wesnoth recipe) Apr 13 20:16:48 hi Gnutoo, I've been reading something about it on the ml Apr 13 20:16:58 ok Apr 13 20:17:16 and yes, I'm still interested inthe wesnoth recipe :) Apr 13 20:17:23 that part on the boost bug was not on the ml...it seems unrelated to the assembly Apr 13 20:17:33 s/assembly/assembly part Apr 13 20:18:53 here's a gdb session of wesnoth: http://192.168.1.1/wesnoth/backtrace_0001.txt Apr 13 20:18:57 oops again Apr 13 20:19:12 night all Apr 13 20:19:22 http://gnutoo.homelinux.org/wesnoth/backtrace_0002.txt Apr 13 20:19:58 (at the moment I was trying to find the cause of illegal instructions in a very old platform) Apr 13 20:21:41 ok, at the moment I'm not able to read it, but I'll look at it :) Apr 13 20:21:53 half asleep, sorry Apr 13 20:22:11 np Apr 13 20:22:26 some time I'm in worse conditions Apr 13 20:27:29 mmm...maybe my illegal instruction is related to pthread... Apr 13 20:27:50 I thought about pthread before...but it was said that it worked with 2.6 headers so... Apr 13 20:28:12 (nptl) Apr 13 20:34:05 if __libc_setup_tls is related to nptl Apr 13 20:45:25 florian, hi, would you mind applying two bugfixes for me ? Apr 13 20:45:57 marex: no problem if you have some minutes Apr 13 20:46:17 mmm...if: Apr 13 20:46:21 florian, I can mail them to you ... I'm just under heavy buildload and I dont feel like using firefox ;) Apr 13 20:46:40 *the program is statically linked to the cross-compiled glibc => illegal instruction Apr 13 20:46:41 florian, can you PM me your mail please? thanks Apr 13 20:47:14 *the program is dynamically linked to the cross-compiled glibc but...use the cross-compiled glibc => illegal instruction Apr 13 20:47:57 else if *the program is dynamically linked to the cross compiled glibc but runs on the original glibc => works Apr 13 20:48:48 Gnutoo: sounds like a problem with your cross compiled libc. Apr 13 20:49:32 florian, yes I think it is: it fails after __libc_setup_tls => it's related to threads... => ntpl doens't work on 2.4 kernel doesn't it? Apr 13 20:49:37 s/ntpl/nptl Apr 13 20:50:06 so maybe it's that but I'd like to verify it before rebuilding the whole setup Apr 13 20:50:20 and can nptl be disabled on the new as 2.6.1 glibc? Apr 13 20:50:39 s/setup/tmpdir Apr 13 20:50:48 I've also a gdb session about that Apr 13 20:51:05 and thanks a lot for your help Apr 13 20:52:56 florian, you have them in your inbox ... Apr 13 20:53:00 I hope it helps Apr 13 20:54:26 2.4 probably needs to be patched for nptl if its vanilla? Apr 13 20:54:34 oopa nypl Apr 13 20:54:52 ok, you need to fix my spelling since im done with the backspace game. Apr 13 20:56:45 timtimred, couldn't I desactivate glibc's nptl...and are you shure it's that? Apr 13 20:57:45 dont know, but i think you may need to patch a vanilla kernel for nptl if you are trying to make it work Apr 13 20:58:18 for 2.4s Apr 13 20:58:26 ok Apr 13 20:59:05 so I think I'll have to build the serial cable... Apr 13 20:59:55 can someone tell me the startegy oe uses for file fetching ? Apr 13 21:00:11 I have SRC_URI = "${GENTOO_MIRROR}/pax-utils-${PV}.tar.bz2;md5sum=98f6b9fe17a740a8cc577255422c6103" Apr 13 21:01:28 it seems GENTOO_MIRROR does not point to the gentoo servers but to ibiblio Apr 13 21:02:58 ha!! it is defined in bitbake conf. Shouldn't this point to gentoo ? Apr 13 21:04:43 ptitjes, maybe ibiblio is a gentoo mirror Apr 13 21:05:37 hum... Apr 13 21:13:19 florian, how are the patches? btw. linphone is broken :/ Apr 13 21:13:29 ptitjes: which version of paxutils are you building Apr 13 21:14:02 khem: 0.1.18 does build correctly Apr 13 21:14:13 0.1.19 is still masked on gentoo Apr 13 21:15:19 khem: but you have to configure GENTOO_MIRROR as a real gentoo mirror (for instance ftp://ftp.free.fr/mirrors/ftp.gentoo.org/distfiles/) Apr 13 21:16:34 men I don't know what to choose... can someone help me ? http://pastebin.com/m495f8b32 Apr 13 21:17:27 also did someone tried to replace the series of case in the openssl.inc by using "./config" as adviced at the end on the paste ? Apr 13 21:26:18 mmm...I think it's ntpl: http://www.codesourcery.com/archives/arm-gnu/msg00293.html Apr 13 21:27:17 so I've some options: Apr 13 21:27:29 *)find ntpl patches for my kenrel Apr 13 21:27:35 *)disable ntpl Apr 13 21:27:43 *)use an old glibc Apr 13 21:27:52 (with ntpl disabled) Apr 13 21:38:32 Gnutoo: use uclibc? Apr 13 21:38:37 :D Apr 13 21:39:37 ok I think I hit the hell problem... Apr 13 21:40:04 I have to reformat my hard disk so that it is case-sensitive :] Apr 13 21:48:44 timtimred, lol I never thought of that...would it work? Apr 13 21:51:43 ptitjes, what filesystem do you use? Apr 13 21:52:48 * Gnutoo wonder what is more compatible between uclibc and glibc-2.5 Apr 13 21:52:57 because there are issues on uclibc Apr 13 21:52:58 Journalized HFS+ but not case-sensitive :) Apr 13 21:53:33 ptitjes, ah ok...does oe work well on macOSX apart the case-sensitive thing? do you have bash on macOSX? Apr 13 21:54:32 Gnutoo: hey man MacOS X is a fully-fledged unix Apr 13 21:54:44 Gnutoo: you have bash by default for sure Apr 13 21:54:52 LOL Apr 13 21:55:00 because in FreeBSD there is no bash but another shell Apr 13 21:55:12 by default Apr 13 21:55:16 so there are problems Apr 13 21:55:25 Gnutoo: anyway I'm using Gentoo-Prefix over MacOS X Apr 13 21:55:29 ok Apr 13 22:01:05 OE does not work well on OS X Apr 13 22:01:37 that's what I wondered Apr 13 22:01:42 ....even on case sensitive HFS+ Apr 13 22:01:45 or UFS Apr 13 22:01:49 I think I'll try glibc-2.5 Apr 13 22:01:52 with a patch Apr 13 22:02:19 i started the OS X bit in OE and your distro. be good for someone to put more work in on it tho :) Apr 13 22:02:49 apparently there was another howto on the old wiki. i had (have) no idea where that is/was :) Apr 13 22:03:25 ptitjes : do you know who is in charge of the ecore python bindings ? Apr 13 22:10:23 timtimred: there is a git branch that koen has started for mac osx origin/koen/darwin-host-fixes Apr 13 22:19:28 * Jay7 -> sleep Apr 13 22:21:36 ahhh. Apr 13 23:12:51 Gnutoo: why do you need glibc 2.5 Apr 13 23:12:55 why not something newer Apr 13 23:20:18 nytowl: sorry not at all Apr 13 23:56:41 Hey, quick question. Is it python, or python-native, that I want to include in target images? Apr 14 00:09:52 Nevermind, figured it out. Apr 14 00:16:42 I'm having problems building gdbm. Apr 14 00:16:52 ERROR: QA Issue: No GNU_HASH in the elf binary: '/OE/build/angstrom-dev/tmp/work/armv7a-angstrom-linux-gnueabi/gdbm-1.8.3-r2/install/gdbm/usr/lib/libgdbm.so.3.0.0' Apr 14 00:23:30 I take it that nobody else has gotten this rather strange error? Apr 14 02:39:18 I recently meet some GNU_HASH QA issues while bitbaking a lot of packages. Apr 14 02:39:19 I had took a look of classes/insane.bbclass but having no idea why check gnu_hash? and why this way? **** ENDING LOGGING AT Tue Apr 14 02:59:57 2009