**** BEGIN LOGGING AT Tue Jul 28 02:59:57 2009 Jul 28 03:07:45 03Graham Gower 07org.openembedded.dev * r7f58c81ce1 10openembedded.git/docs/usermanual/chapters/recipes.xml: Jul 28 03:07:45 usermanual: Mostly grammar and spelling fixes Jul 28 03:07:45 Signed-Off-By: Holger Freyther Jul 28 05:11:21 th1: If you are still there (?), the answer to your question is to use bitbake -k Jul 28 05:11:33 th1: the -k allows it to proceed even if errors are generated Jul 28 06:45:30 good morning Jul 28 07:14:56 morning Jul 28 07:16:21 good morning hrw Jul 28 07:31:55 03Steve Sakoman 07org.openembedded.dev * ra0236d71a8 10openembedded.git/recipes/gnome/gnome-power-manager_2.24.2.bb: Jul 28 07:31:55 gnome-power-manager: fix typos in specification of FILES for doc package Jul 28 07:31:55 Signed-off-by: Koen Kooi Jul 28 07:31:56 03Steve Sakoman 07org.openembedded.dev * r72e8a6c97b 10openembedded.git/recipes/gtksourceview/gtksourceview2_2.6.0.bb: Jul 28 07:31:56 gtksourceview: package was missing a few important files, add them Jul 28 07:32:00 Signed-off-by: Koen Kooi Jul 28 07:32:02 03Steve Sakoman 07org.openembedded.dev * ra1adb7f8b6 10openembedded.git/recipes/policykit/ (policykit/PolicyKit.conf policykit_0.9.bb): policykit: install missing PolicyKit.conf Jul 28 07:56:47 How do I rebuild my kernel having added KERNEL_IMAGETYPE to my local.conf ? Jul 28 07:59:58 bitbake virtual/kernel -cclean;bitbake virtual/kernel Jul 28 08:01:58 hrw: of course ....... just got back from vacation so I guess my brain has got a kill -9. Thank you :) Jul 28 08:04:22 ;D Jul 28 08:08:17 morning Jul 28 08:13:10 morning florian Jul 28 08:14:54 good morning Jul 28 08:46:39 rkirti, thanks for the answer Jul 28 08:47:41 th1: you are welcome :- Jul 28 08:47:42 :-) Jul 28 08:48:04 I did get past that problem though after specifying the xserver explicitly. but now it won't build my xorg-video-vesa or xorg-input-keyboard because it can't find a header file Jul 28 08:48:41 th1: could you pastebin the log ? Jul 28 08:48:53 yes Jul 28 08:49:01 give me a second to log in :) Jul 28 08:49:33 I have to run it again my ssh session timed out .. but it won't take more than a minute Jul 28 08:49:50 ok Jul 28 08:52:35 http://pastebin.ca/1509787 Jul 28 08:54:44 How do I ensure that "bitbake virtual/kernel" will do_compile_append in linux.inc ? Jul 28 08:55:08 th1: Not sure. did you bitbake xserver-xorg earlier ? Jul 28 08:55:20 yes I think I did Jul 28 08:55:57 th1: maybe bitbake -c clean and try again ? Jul 28 08:56:14 it's worth a try Jul 28 08:56:35 but the same happens for xorg-driver-keyboard Jul 28 08:56:58 th1: and check your staging directory if the header file is there. Jul 28 08:57:14 th1: if not, something's wrong with your xserver-xorg build Jul 28 08:57:41 so maybe that's the one I should clean/rebuild? Jul 28 08:58:14 th1: yes Jul 28 08:58:18 th1 Jul 28 08:58:35 tmp/staging/i686p4c-angstrom-linux/usr/include/xorg/xorg-server.h Jul 28 08:58:35 if the staging dir doesnt have the header, you really should rebuild Jul 28 08:59:00 so it's there Jul 28 08:59:47 th1: ok. Sorry, I dont now then why it doesnt find it. Jul 28 09:00:58 where does it store the full make output with gcc command line? Jul 28 09:01:55 th1: in the work directory, cd to the package name, and you shoulld have a dir. name temp there. that dir. contains all the logs. check log.do_compile Jul 28 09:02:39 ok Jul 28 09:02:52 it doesn't have -I...staging-dir/usr/include/xorg Jul 28 09:10:58 how can I add an include path specifically in xorg-driver-common.bb? Jul 28 09:11:28 I know I shouldn't have to but I just want to get this to work Jul 28 09:11:42 th1: usually oe takes care of the correct includes. Jul 28 09:12:04 but not in this case it seems Jul 28 09:12:48 well I will try rebuilding xserver-xorg first and if it doesn't help then find a way to add the include dir to configure Jul 28 09:24:40 03Holger Hans Peter Freyther 07org.openembedded.dev * r3aeba2d2d2 10openembedded.git/recipes/xorg-lib/ (libx11-trim/makekeys-update.patch libx11-trim_1.1.5.bb): Jul 28 09:24:40 libx11-trim: Update makekeys.c to fix the compile Jul 28 09:24:40 Without the updated makekeys-update.patch the makekeys.c Jul 28 09:24:40 will not work and the build will be broken. Jul 28 09:24:41 03Holger Hans Peter Freyther 07org.openembedded.dev * r840b7178f2 10openembedded.git/recipes/e2fsprogs-libs/ (4 files): Jul 28 09:24:45 e2fsprogs-libs: inherit gettext to fix uclibc build Jul 28 09:24:47 Inherit gettext to get the DEPENDS of gettext-native and to Jul 28 09:24:49 add --disable-nls/--enable-nls based on the settings. This Jul 28 09:24:51 is unbreaking the uclibc build. Jul 28 09:26:23 Has anybody had success building wvdial/wvstreams here ? Jul 28 09:27:25 03Dmitry Eremin-Solenikov 07org.openembedded.dev * rb8ffa54b8a 10openembedded.git/: Merge branch 'org.openembedded.dev' of git://git.openembedded.org/openembedded into org.openembedded.dev Jul 28 09:27:26 03Dmitry Eremin-Solenikov 07org.openembedded.dev * ra5c9f24b60 10openembedded.git/conf/distro/angstrom-2008.1.conf: Jul 28 09:27:26 angstrom: default to nogpg opkg Jul 28 09:27:26 Since feeds aren't signed anyway, default to nogpg opkg version. Jul 28 09:27:28 Signed-off-by: Dmitry Eremin-Solenikov Jul 28 09:27:30 03Dmitry Eremin-Solenikov 07org.openembedded.dev * r6ab46c6e2d 10openembedded.git/classes/oestats-client.bbclass: Jul 28 09:27:33 oestats-client: restore support for proxies Jul 28 09:27:35 Signed-off-by: Dmitry Eremin-Solenikov Jul 28 09:27:36 rkirti: yes, about a year ago... and it didn't work on arm :) Jul 28 09:27:37 03Dmitry Eremin-Solenikov 07org.openembedded.dev * r0945dd486a 10openembedded.git/: Merge branch 'org.openembedded.dev' of git://git.openembedded.org/openembedded into org.openembedded.dev Jul 28 09:27:40 03Dmitry Eremin-Solenikov 07org.openembedded.dev * r6e75ec0e0d 10openembedded.git/: Merge branch 'org.openembedded.dev' of git://git.openembedded.org/openembedded into org.openembedded.dev Jul 28 09:27:43 03Dmitry Eremin-Solenikov 07org.openembedded.dev * r527d3b71c9 10openembedded.git/: Merge branch 'org.openembedded.dev' of git://git.openembedded.org/openembedded into org.openembedded.dev Jul 28 09:27:50 03Dmitry Eremin-Solenikov 07org.openembedded.dev * r922a5b21ab 10openembedded.git/: Merge commit 'origin/org.openembedded.dev' into org.openembedded.dev Jul 28 09:28:16 zecke: :-( Jul 28 09:30:15 rkirti: what problem do you facE? Jul 28 09:31:52 Why is KERNEL_IMAGETYPE overridded when I set it to uImage in my local.conf ? kernel.bbclass still thinks it is zImage :( Jul 28 09:32:12 zecke: wvstreams fails to compile Jul 28 09:32:18 zecke: http://pastebin.com/d2af6404d Jul 28 09:32:53 03Steve Sakoman 07org.openembedded.dev * r7af11a2eff 10openembedded.git/recipes/icecast/icecast_2.3.2.bb: icecast: add 2.3.2 Jul 28 09:32:54 03Koen Kooi 07org.openembedded.dev * r1d5ab38e31 10openembedded.git/recipes/xserver-common/ (files/xtscal-fix.patch xserver-common_1.25.bb): xserver-common: run tslib-calibrate for angstrom instead of xtscal (patch from gumstix) Jul 28 09:32:55 03Steve Sakoman 07org.openembedded.dev * r758afabaff 10openembedded.git/recipes/xmms/xmms_1.2.10.bb: xmms: tweak .desktop file so that it works with gnome Jul 28 09:32:57 03Koen Kooi 07org.openembedded.dev * r19862d4b2f 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev Jul 28 09:43:25 any idea how I can build the xf86-video-vesa and friends when it doesn't find the include files in $(STAGING)/usr/include/xorg? Jul 28 09:43:49 can I do something like "EXTRAINCLUDES=..." and if so where? Jul 28 09:44:53 output is here: http://pastebin.ca/1509787 Jul 28 09:44:57 any help would be much appreciated Jul 28 09:49:42 rkirti: looks fairly simple, how do you want to solve it? Jul 28 09:50:45 th1_: did you build kdrive by any chance? Jul 28 09:55:02 zecke: simple it is. Only that I dont know if their INT_MAX is a standard value or something they want. and why the inclusion of limits.h (i did that in uniconf.h)doesnt work Jul 28 09:56:47 rkirti: well, it is a standard one... otherwise they are too fucked :) Jul 28 09:56:50 rkirti: stdint.h? Jul 28 09:57:25 zecke: will check. Jul 28 09:59:26 too bad their latest packages (4.5) dont build even natively Jul 28 10:23:14 XorA, I think I built it previously yes Jul 28 10:23:27 th1_: rebuild xorg then Jul 28 10:23:36 th1_: kdrive fscks up some of the staging files Jul 28 10:23:48 I did "bitbake -c clean xserver-xorg" Jul 28 10:24:06 and then rebuilt xserver-xorg Jul 28 10:24:14 that shoulda fixed it Jul 28 10:24:19 but it didn't :( Jul 28 10:24:44 * XorA burns something random Jul 28 10:25:12 burn kdrive :) Jul 28 10:25:46 * XorA wishes he had NIN Burn to hand Jul 28 10:27:09 is there anything else you can suggest apart from nuking entire tmp and starting again? Jul 28 10:27:21 because otherwise that's what I'll do Jul 28 10:28:23 th1_: the header location comes from pkgconfig, so I would look at the xorg.pc file Jul 28 10:28:34 * XorA doesnt have a build machine on hand Jul 28 10:30:18 http://pastebin.ca/1509842 Jul 28 10:31:25 and headers are in ${STAGING}/usr/include/xorg/ ? Jul 28 10:31:29 yes Jul 28 10:31:33 odd then Jul 28 10:31:35 well this is after rebuilding a 2nd time Jul 28 10:31:42 let's see what happens after it rebuilds cache Jul 28 10:31:52 although 1.6.1 is new, maybe there is some incompatability with some drivers Jul 28 10:32:55 http://pastebin.ca/1509843 is my machine config if it matters. I don't specify any X versions specifically Jul 28 10:35:00 its still the same ... Jul 28 10:38:59 does PARALLEL_MAKE and BB_NUMBER_THREADS cooperate Jul 28 10:39:13 or does it fire up BB_NUMBER_THREADS * PARALLEL_MAKE gcc's? Jul 28 10:39:23 th1_: the second Jul 28 10:40:03 so on a quadcore is 2, 2 reasonable values? Jul 28 10:40:28 now that I have to rebuild everything I want to optimize it as much as possible ;) Jul 28 10:40:35 I think I use 3,4 on my quad Jul 28 10:40:40 to flatline all CPU's Jul 28 10:41:23 ok Jul 28 10:41:32 3 for THREADS or MAKE? Jul 28 10:41:44 3 bb threads Jul 28 10:42:21 it seems one thread is always waiting on IO, so in general seems to end up with 8x gcc running Jul 28 10:42:27 re Jul 28 10:43:32 sgh: you have to set KERNEL_IMAGETYPE in machine config or in recipe Jul 28 10:47:34 XorA, I took your values and that seems ok Jul 28 10:47:46 after a bit getting started Jul 28 10:51:00 i'm running it all under script this time so I can paste a very big log if it fails ;) Jul 28 10:53:03 hi Jul 28 11:10:42 hrw: it does not work .... it is overwritten somewhere else. Even if I set it in kernel.bbclass is fails to realize that it is not zImage Jul 28 11:13:47 which machine which kernel? Jul 28 11:18:25 hrw: well .... if I replace =? with = in kernel.bbclas it seems to work better Jul 28 11:18:45 hrw: gamecube .... the kernels is the one that it builds out of the box Jul 28 11:19:48 sgh: you set IMAGETYPE and then inherit kernel? Jul 28 11:19:57 ssh: Could not resolve hostname git.openembedded.org: Name or service not known Jul 28 11:21:14 hrw: don't you mean KERNEL_IMAGETYPE? Jul 28 11:24:50 KERNEL_IMAGETYPE I mean yes Jul 28 12:02:39 zecke: ping ? Jul 28 12:03:47 pong Jul 28 12:04:48 zecke: I got through with the build. Adding the limits.h worked, and then I ran into a gcc4.3 related error. I would like to put a patch for that, but how do I put a patch for something thats related to the compiler ? Jul 28 12:05:11 zecke: the wvstreams pakcage has an error related to this description : http://d.hatena.ne.jp/TELL/20090208/p1 Jul 28 12:05:21 rkirti: well, it can be applied uncondtionally Jul 28 12:05:42 zecke: cool. I wasn't sure of that. thanks Jul 28 12:20:07 rkirti, it seems like gcc4.3 patches do not break on older toolchains Jul 28 12:20:28 Crofton|work: ok, will submit it to the ml then Jul 28 12:31:18 xserver-kdrive 1.3.0.0 strikes me again Jul 28 12:41:28 hi Cliff Jul 28 12:41:43 hi Marcin Jul 28 12:43:59 cbrake: you have experience with different ARM cpus. have you ever met device with ST8815 cpu? Jul 28 12:44:19 hrw: no I have not -- looking up ... Jul 28 12:44:45 hrw: hmm google does not return something quickly -- must be fairly obscure? Jul 28 12:45:03 cbrake: search for 'st nomadik' Jul 28 12:47:34 hrw: looks like an interesting cpu Jul 28 12:49:48 hrw: anybody using it? Jul 28 12:50:11 cbrake: thats unknown to me too Jul 28 12:50:25 cbrake: iirc nokia n96 uses it Jul 28 12:50:38 cbrake: I have NHK15 (ST devboard) on desk Jul 28 13:57:51 What is the first the kernel write to the screen. I am in doubt weither my kernel actually boots or uboot plays a trick on me. ? Jul 28 14:02:35 sgh: uncompressing kernel? but this can be done by uboot too Jul 28 14:03:13 zecke: how do I avoid having ubot do it uboot-mkimage on a non-gzip compressed file ? Jul 28 14:03:30 no idea Jul 28 14:03:37 zecke: ok Jul 28 14:26:37 Can anyone give me some pointers on creating a bb file for an application that relies on a -native package? Jul 28 14:27:45 I'm trying to create a bb file for Debian's realpath. It requires po4a to build - I've created a native package for that, which seems to get put into my staging tree. Jul 28 14:29:04 However the realpath build appears to be looking for stuff in my systems directories rather than staging Jul 28 14:32:37 configure it right Jul 28 14:33:41 :-) Yeah Jul 28 16:38:14 have a nice day Jul 28 16:40:16 RP: ping Jul 28 16:46:05 jo kergoth Jul 28 16:46:10 hey Jul 28 16:46:19 hey Jul 28 16:46:37 kergoth: Can I bother you some more with the witpa thing? Jul 28 16:46:54 The 404 experiences on do_fetch are really annoying Jul 28 16:46:59 and too numerous Jul 28 16:47:25 witpa thing? Jul 28 16:47:38 I'd be more than willing to help out where I can, but I got stuck with witpa-native Jul 28 16:47:54 "where is that package at" from likewise Jul 28 16:47:58 * Laibsch digs for the URL Jul 28 16:50:33 kergoth: http://article.gmane.org/gmane.comp.handhelds.openembedded/17776 Jul 28 16:50:50 where is pb? Jul 28 16:50:55 ~seen pb Jul 28 16:50:57 pb was last seen on IRC in channel #tomcat, 277d 22h 7m 32s ago, saying: 'arghh'. Jul 28 16:51:02 ~seen pb_ Jul 28 16:51:03 pb_ was last seen on IRC in channel #oe, 1d 7h 47m 4s ago, saying: 'florian: good morning'. Jul 28 16:51:17 ~seen pb__ Jul 28 16:51:18 pb__ was last seen on IRC in channel #oe, 15d 22h 17m 48s ago, saying: 'Jay7: sounds like you need a different DISTRO :-}'. Jul 28 16:51:40 kergoth: rings a bell? Jul 28 16:52:09 we discussed the integration with do_fetch about a month ago Jul 28 16:52:11 yep, i remember looking at it. i won't be able to look into doing anything with it for at least a few weeks, we're approaching a deadline soon at work, the pressure is on Jul 28 16:52:20 darn Jul 28 16:52:52 I'm not the only one who could add it, though. the fetcher stuff is all just python :) Jul 28 16:54:50 I know you're not the only one. I've already looked around for pb who had previously expressed an interest. He's not around. Neither is likewise. Jul 28 16:55:11 I'd do it myself if I knew how to. But I'm already stuck with just witpa-native bb. Jul 28 17:33:41 Laibsch: ah Jul 28 17:43:25 kergoth: IIRC the package failed in do_packge Jul 28 17:43:46 which is a strange place to fail and I think I normally should be able to fix it, but I wasn't Jul 28 17:43:49 :( Jul 28 17:54:50 that's odd. native recipes should have a no-op package generation phase. is PACKAGES empty for that recipe? check bitbake -e Jul 28 18:07:06 What does the message in do_rootfs : "found in Packages, but mtime differs - rereading" imply ? Jul 28 18:09:43 re Jul 28 18:10:24 rkirti: thats from ipkg-make-index tool. it means that name of package file is the same but file itself was changed since last run Jul 28 18:10:38 hrw: ok Jul 28 18:10:53 thanks Jul 28 18:36:24 re Jul 28 18:41:23 http://www.entropykey.co.uk/ Jul 28 18:48:43 doing builds in virtualbox is handy Jul 28 18:57:14 * hrw -> off Jul 28 19:31:48 hi, how do I configure OE to build with eglibc instead of glibc ? Jul 28 19:32:06 setting PREFFERED_PROVIDER_virtual/libc = "eglibc" doesnt seem to make the trick for me Jul 28 19:33:39 what distro? Jul 28 19:33:49 It's mostly a matter of setting LIBC=eglibc Jul 28 19:33:55 in your local.conf Jul 28 19:34:02 thanks Jul 28 19:34:04 angstrom Jul 28 19:34:52 yeah, that obeys LIBC Jul 28 19:35:31 thanks a lot :-) Jul 28 19:35:56 This reminds me that the toolchain targets should follow this selection too Jul 28 19:36:23 what toolchain targets? Jul 28 19:43:31 iirc the last time i tried meta-toolchain always pulled in glibc Jul 28 19:43:53 even if LIBC is set to eglibc Jul 28 19:47:40 Must have been a while ago then :) Jul 28 19:48:11 The toolchain task keys off of the libc, prior to the libc-{glibc,uclibc} override being added Jul 28 19:48:24 It didn't used to, but I did fix that a while ago Jul 28 19:51:37 true Jul 28 19:51:41 ok, perfect :) Jul 28 19:51:47 good guy :) Jul 28 19:58:12 Tartarus: angstrom uses ANGSTROMLIBC Jul 28 19:58:42 koen should start using sane-toolchain conf files prolly Jul 28 19:59:11 khem, angstrom uses both Jul 28 19:59:22 ANGSTROMLIBC defaults to LIBC value Jul 28 19:59:38 oh that would work Jul 28 20:00:30 Tartarus: distro/include/angstrom.inc doesnt seem to do that Jul 28 20:01:15 Er Jul 28 20:01:28 I didn't delete ANGSTROMLIBC since I don't have their docs :) Jul 28 20:01:38 But it's not used anywhere else in the angstrom confs now Jul 28 20:01:51 So it's what I meant to say, ANGSTROMLIBC will just set LIBC, or you can use LIBC directly Jul 28 20:01:57 like every other distro :) Jul 28 20:02:13 (I also tried to not break the other distro that did thiss hook) Jul 28 20:02:29 ah ANGSTROMLIBC is only for backward compatibility Jul 28 20:07:10 is git.openembedded.org down ? Jul 28 20:07:59 khem: I can ping it Jul 28 20:09:42 hmm I can ping 140.211.169.165 the IP but not by name Jul 28 20:10:46 khem: I pinged by name. Jul 28 20:11:34 seems something is wrong with my DNS Jul 28 21:27:56 anyone having trouble with oe.org dns? Jul 28 21:27:58 I am Jul 28 21:28:17 I checked the dns service provider and nothing obviously wrong, all paid up etc Jul 28 21:29:58 Server not found Jul 28 21:29:58 Jul 28 21:29:58 Jul 28 21:29:58 Jul 28 21:30:00 Jul 28 21:30:02 Jul 28 21:30:04 Jul 28 21:30:08 Jul 28 21:30:10 Jul 28 21:30:12 Firefox can't find the server at www.openembedded.org. Jul 28 21:30:27 dns seems ok here atm... Jul 28 21:30:56 ping www.openembedded.org Jul 28 21:30:56 ping: unknown host www.openembedded.org Jul 28 21:31:23 www.oe.org works here Jul 28 21:31:31 hm here its working Jul 28 21:31:35 it works from my hosted server, but not at home Jul 28 21:31:54 (the internet continues to degrade) Jul 28 21:32:13 ~blame Crofton Jul 28 21:32:14 * ibot blames Crofton (and Canada) for all the evil in the world Jul 28 21:32:32 lol Jul 28 21:33:07 stupid dotster Jul 28 21:33:11 Crofton: I had trouble like an hour back but now it seems ok Jul 28 21:33:24 if it does not resolve itself soon, I'll complain Jul 28 21:33:48 Crofton: they probably sent a duf update around Jul 28 21:34:02 duf update Jul 28 21:34:04 ? Jul 28 21:34:08 mine issues were with git.openembedded.org Jul 28 21:34:14 duff Jul 28 21:36:47 hm ah first ion netbook will be soon on sale Jul 28 21:54:45 hmm Jul 28 21:55:01 wonder if we should switch bitbake to setuptools rather than just distutils Jul 28 21:55:08 woglinde, hi what's an ion netbook? Jul 28 21:55:44 Crofton: I see the same dns problem here... Jul 28 21:57:14 ok, off to the customer support site Jul 28 21:59:14 http://pastebin.ca/1510442 Jul 28 22:00:25 oops Jul 28 22:02:18 yeah Jul 28 22:02:29 if it does not get better tomorrow, I will complain Jul 28 22:08:13 hmmm Jul 28 22:08:18 I'm very new to OE Jul 28 22:08:38 welcome Jul 28 22:08:43 I'm trying to build a mipsel minimal-ulibc as a starting point for a dev board I have Jul 28 22:09:33 and running into an error in ulibc-initial Jul 28 22:09:42 "../../Rules.mak:525: *** Your binutils don't support --hash-style option, while you want to use it. Stop." Jul 28 22:11:14 googling didn't really help, although I found some discussion with the error in october last year Jul 28 22:13:39 I thought upgrading binutils-cross might help Jul 28 22:14:04 but I'm not sure where minimal-uclibc derives its binutils-cross version from (2.18) Jul 28 22:16:04 hmm Jul 28 22:16:19 i had something like that Jul 28 22:16:28 * timtimred checks logs... Jul 28 22:17:19 i think i disabled gnu link hash style Jul 28 22:17:22 somehow Jul 28 22:19:35 thaytan: its a binutils bug IIRC Jul 28 22:19:35 using angstrom i added.... Jul 28 22:19:38 TARGET_LINK_HASH_STYLE�=�"${@['-Wl,--hash-style=gnu',''][bb.data.getVar('TARGET_ARCH',d,1)�in�['mips',�'mipsel']]}" Jul 28 22:19:52 to uh, angstrom-uclibc.inc Jul 28 22:20:11 yes that should have taken care of not using this option Jul 28 22:23:02 Is it OK to throw that in minimal-uclibc.conf then? Jul 28 22:23:35 and, is there a way to trace how it's deciding to use version 2.18 binutils? Jul 28 22:24:01 I think no Jul 28 22:24:12 binutils should be setup in your distro config somewhere Jul 28 22:24:36 I have PREFERRED_VERSION_binutils-cross ?= "2.18.50.0.7" in there Jul 28 22:24:41 at it's still building vanilla 2.18 Jul 28 22:24:52 which makes me think it's getting an explicit setting somewhere else? Jul 28 22:24:58 but I can't see where :) Jul 28 22:25:00 pass, im not in front of an up to date copy atm Jul 28 22:25:15 its in there somewhere. grep etc ;) Jul 28 22:25:23 probably from distro/include/sane-toolchain.inc Jul 28 22:25:39 if you want to override remove ? Jul 28 22:25:47 PREFERRED_VERSION_binutils-cross = "2.18.50.0.7" Jul 28 22:26:34 aha Jul 28 22:26:41 I grepped binutils, not BINUTILS Jul 28 22:27:00 option timtimred suggested should have come from distro/include/sane-toolchain-uclibc.inc for you Jul 28 22:28:20 thaytan: btw are you using .dev or stable branch ? Jul 28 22:28:21 thank you Jul 28 22:28:29 khem: starting on stable Jul 28 22:28:38 ugh Jul 28 22:28:57 ugh that ulibc-minimal doesn't build on stable? :) Jul 28 22:29:07 or ugh, don't use that branch? Jul 28 22:29:18 I'm also having trouble accessing git.openembedded.org and www.openembedded.org Jul 28 22:29:52 thaytan: stable is probably wise choice but my suggestions so far were based on trunk Jul 28 22:29:55 or .dev Jul 28 22:30:13 khem: the 2.18 seems to be in those files Jul 28 22:30:26 I'm happy to assign it explicitly, I just wanted to understand where the value derived from Jul 28 22:30:37 and it seems to be building the correct one now Jul 28 22:30:44 yeah sane-toolchain.inc should specify it Jul 28 22:30:51 well, "correct" -> "the one I wanted" Jul 28 22:32:58 john3909: seems to be a general problem many people are seeing right now Jul 28 22:35:51 khem: I read about this above. I haven't been able to access the OE url all day. It is strange that some can access the site and some cannot. Looks like a DNS problem. Jul 28 22:36:40 khem: probably clear up in 24 hours Jul 28 22:41:10 you can add 85.214.40.226 www.openembedded.org to your /etc/hosts as a workaround Jul 28 22:44:10 aha, mipsel-oe-linux-uclibc-ld: .gnu.hash is incompatible with the MIPS ABI Jul 28 22:44:20 I guess I'll give up on trying to upgrade things to fix it then :) Jul 28 22:45:41 thaytan: it should not have been selected for you Jul 28 22:45:44 you can tell it not to use gnu hash iirc... Jul 28 22:45:45 at first place Jul 28 22:46:01 thaytan: post your local conf somehwere Jul 28 22:46:16 but that should be the default for mipsel anyway.... Jul 28 22:48:07 khem: it's just DISTRO=minimal-uclibc MACHINE=custom Jul 28 22:48:27 custom? Jul 28 22:48:36 and a new machine description file Jul 28 22:49:00 a copy of db1200.conf Jul 28 22:51:20 mimimal-uclibc -> minimal.conf -> sane-toolchain.inc -> sane-toolchain-uclibc.inc thats where it should have defined not to include gnu hash for mips uclibc Jul 28 22:52:08 woglinde: I finally compiled my binary against klibc, but I had to use mtd-user.h Jul 28 22:52:33 khem: indeed - that's in there for me Jul 28 22:52:51 I don't understand yet if I have to stage the hacked headers or not... Jul 28 22:54:52 khem: so, perhaps my LDFLAGS do include the right flags, and it's something else causing uclibc to mis-detect things Jul 28 22:55:07 khem: I got the ABI error running ld myself, which wouldn't have used the same flags Jul 28 22:57:27 thaytan: I think I see the bug Jul 28 22:58:52 happy to test something :) Jul 28 22:59:20 can you set PREFERRED_LIBC = "uclibc" in your local.conf ? Jul 28 23:00:48 OK Jul 28 23:05:21 woglinde: I see mtd-utils stages own /include/mtd/* headers. I'll do like them Jul 28 23:12:45 ant cant help you but make it Jul 28 23:12:50 khem: this is going to take a while Jul 28 23:13:00 under debian mts-user.h is linux-libc-dev Jul 28 23:13:35 ah Jul 28 23:13:52 woglinde: wie gehts Jul 28 23:14:08 khem hm okay Jul 28 23:14:09 thaytan: you can also try .dev Jul 28 23:14:38 have you ever used svn merge ? Jul 28 23:14:57 khem: let's see how this build goes Jul 28 23:15:17