**** BEGIN LOGGING AT Mon Jan 07 02:59:59 2013 Jan 07 08:02:13 what would be best pratice to overwrite a bbclass under */classes/*.bbclass in layer ? Jan 07 08:19:54 morning Jan 07 08:29:11 good morning Jan 07 09:37:14 anyone has seen this ? configure.ac:131: option `dist-lzip' not recognized Jan 07 09:45:01 apparently my automake is a bit too old Jan 07 12:03:33 what would be best pratice to overwrite a bbclass under */classes/*.bbclass in a layer ? Jan 07 12:45:23 rob_w: just copy original one and adapt to your needs? ugly but works Jan 07 12:52:47 thx hrw Jan 07 13:36:55 sat-solver-0.0-gitAUTOINC+0a7378d5f482f477a01cf1690d76871ab8bdcc32-r14 - why AUTOINC there instead of some kind of numbeR? Jan 07 13:42:14 hrw: it's converted to number later in PKGV Jan 07 13:42:21 thx Jan 07 13:43:29 hrw: do you care about upgrade path? Jan 07 13:43:58 https://bugzilla.yoctoproject.org/show_bug.cgi?id=3071 Jan 07 13:44:26 JaMa: we use build/boot/forget scheme Jan 07 13:44:57 JaMa: aarch64 model is too slow to use package management on it Jan 07 13:45:26 ok Jan 07 13:47:15 JaMa: probably would be faster to build from scratch then update week old image Jan 07 13:47:47 on same x86 box Jan 07 14:00:42 another day, another things merged into OE Jan 07 17:35:40 ant_work: around ? Jan 07 17:36:00 yes Jan 07 17:36:04 aafc: e12fff1e bxlr Jan 07 17:36:10 bx lr Jan 07 17:36:14 thats the problem Jan 07 17:36:16 thumb Jan 07 17:36:19 the code is not thumb Jan 07 17:36:28 its all arm Jan 07 17:36:53 but I guess processor you are on does not support interworking and hence bx Jan 07 17:37:03 what is this machine ? Jan 07 17:37:14 is it armv4 ? Jan 07 17:37:23 yes, strongarm Jan 07 17:37:24 http://wiki.debian.org/ArmEabiPort#Choice_of_minimum_CPU Jan 07 17:37:30 :/ Jan 07 17:37:31 hmmm Jan 07 17:37:44 OK what is that disassembly of ? Jan 07 17:37:49 is it klibc ? Jan 07 17:38:00 yes, it's ls from klibc_2.0.2 Jan 07 17:38:04 ok Jan 07 17:38:17 klibc-static-utils-ls is the package Jan 07 17:38:18 now show me a compiler commandline from a .c file in klibc Jan 07 17:38:36 I want to see how it is being compiled Jan 07 17:38:48 I think compiler is compiling for armv4t Jan 07 17:38:55 thats my hunch Jan 07 17:39:02 no, eglibc bins are running Jan 07 17:39:06 it's just klcc Jan 07 17:39:18 hmmm Jan 07 17:39:27 anyway thats what I meant Jan 07 17:39:28 we don't pass the fix for BX to it I think Jan 07 17:39:36 to klcc ? Jan 07 17:39:49 oh then thats an obvious problem Jan 07 17:40:13 I guess you need to modify klcc accordinly Jan 07 17:40:33 btw. you need that option for linker and assembler mainly Jan 07 17:41:06 are you on oe-core/master latest ? Jan 07 17:41:13 from saturday Jan 07 17:41:30 see, I'm heading home. Il'be there in one hour. I can open sshd Jan 07 17:42:00 or just have your guide online if you have time Jan 07 17:42:04 thx in advance Jan 07 17:42:50 hmm ping me Jan 07 17:42:53 if I have time Jan 07 17:43:02 and am sitting in a boring meeting I will help oyu Jan 07 17:43:08 ;) Jan 07 17:43:14 bbl Jan 07 18:35:51 khem, can you save the bits you want from Agate? Jan 07 19:57:34 Crofton|irssi: what's the channel again? Jan 07 20:03:56 khem: rebuilding from scratch. now 220/2682, ready in a few mins if you mind Jan 07 20:04:15 heh 2200/2682 Jan 07 20:09:23 ant_home: ok Jan 07 20:09:27 I have like 21 mins Jan 07 20:09:41 ka6sox: there ? Jan 07 20:10:16 ka6sox: can you give me a place to park the changes ?? Jan 07 20:10:46 khem: hi, is built now Jan 07 20:11:04 ant_home: ok Jan 07 20:11:06 I've added ARM_INSTRUCTION_SET = "arm" to klibc.inc right now Jan 07 20:11:10 k Jan 07 20:11:16 thats not the problem through Jan 07 20:11:17 dunno if it helps Jan 07 20:11:21 yes Jan 07 20:11:27 I asked for cmdline Jan 07 20:11:29 options Jan 07 20:11:36 so I can see if it liked correctly Jan 07 20:18:26 khem: this is klibc Jan 07 20:18:26 http://paste.debian.net/222611/ Jan 07 20:19:44 bitbake says Jan 07 20:19:46 TUNE_FEATURES = "armv4" Jan 07 20:19:46 TARGET_FPU = "soft" Jan 07 20:21:52 while a normal eglibc prog like stat has Jan 07 20:21:54 NOTE: make -j 5 -e MAKEFLAGS= Jan 07 20:21:54 arm-oe-linux-gnueabi-gcc -march=armv4 -mthumb-interwork --sysroot=/oe/oe-core/build/tmp-eglibc/sysroots/collie -O2 -pipe -g -feliminate-unused-debug-types -o stat stat.c -Wl,-O1 -Wl,--hash-style=gnu -Wl,--as-needed Jan 07 20:22:26 ant_home: ok there is a problem there Jan 07 20:24:30 ant_home: canyou add TARGET_LD_KERNEL_ARCH to LDFLAGS ? Jan 07 20:24:49 ant_home: also show me the linker cmdline Jan 07 20:30:06 hm.. maybe you mean this? http://paste.debian.net/222612/ Jan 07 20:33:05 in Kbuild.klibc there is mention of EXTRA_KLIBCCFLAGS Jan 07 20:33:30 KLIBCLDFLAGS := Jan 07 20:33:30 KLIBCCFLAGS := Jan 07 20:33:35 and co. Jan 07 20:42:39 bbl Jan 07 20:43:34 hi khem Jan 07 21:09:26 khem: rebuilding world with eglibc-2.17 on garnet to see how much stuff will fail Jan 07 21:16:07 * Daemon404 still thinks it's dismal when a new version of libc makes thigns fail in strange ways... Jan 07 21:23:33 does anyone think the bitbake manual warrants a short introduction to Python within the context of BitBake/OE? Jan 07 21:24:13 it should certainly cover soe of the bitbake python pacakge which is useful in that context Jan 07 21:24:30 great, thanks Jan 07 21:41:40 JaMa: cool. I have now booted all core-image-sato on qemu machines fine Jan 07 21:41:51 JaMa: btw. you need to pick the latest from my branch Jan 07 21:42:05 JaMa: since I pushed a udev fix too Jan 07 21:42:13 which is only needed for oe-core Jan 07 22:58:36 khem: aarch64 with eglibc 2.17 started Jan 07 22:58:50 I mean build - tomorrow will check booting Jan 07 23:06:05 hrw: here too :) NOTE: Running task 5008 of 22861 Jan 07 23:06:39 JaMa: cool. we probably build different targets. Linaro images + world here Jan 07 23:06:54 yes here world on qemuarm Jan 07 23:15:44 shit. gmp fails Jan 07 23:15:51 but it failed with 2.16 as well Jan 07 23:18:09 | {standard input}:318: Error: operand 3 should be an integer register -- `adc x8,x9,0' Jan 07 23:18:12 argh Jan 08 00:10:57 hrw: cool **** ENDING LOGGING AT Tue Jan 08 02:59:58 2013