**** BEGIN LOGGING AT Wed Jan 12 02:59:58 2011 Jan 12 03:00:16 Lopi, re firewall... the angstrom source mirrors will probably be of use to you. Try adding INHERIT += "angstrom-mirrors" to your local.conf Jan 12 04:55:29 Tartarus: yes it will happen on all arches Jan 12 04:55:41 Tartarus: its that bison assumes sched.h is from glibc Jan 12 05:38:28 k Jan 12 05:39:16 * Tartarus tests the ac_cv_func_dlopen thing on native-sdk-image + minimal and will push in the am Jan 12 05:47:58 Having some trouble compiling SHR for the iPhone 3G. Anyone mind taking a look at my pastebin? http://pastebin.com/y3Wnn73D Jan 12 05:59:41 Lopi, looks like a SRCREV was bumped without checking that all the patches apply to all the recipes using it Jan 12 06:01:25 Lopi, see http://cgit.openembedded.org/cgit.cgi/openembedded/commit/?id=f3b77424a8568acd0e5563c080384bbcfce80bc4 - Your failing recipe, libeflvala_svn.bb, uses EFL_SRCREV Jan 12 06:02:07 You could send a message to the list, or wait for JaMa|Off to come online. Jan 12 06:02:38 ah okay thanks grg Jan 12 06:02:57 or you could try to fix it yourself. Either the patch has been applied upstream, or it needs to be reworked to apply to the new revision Jan 12 06:03:57 Lopi, "Patch 0001-eflvala-adapt-to-elementary-API-change-from-r55869.patch can be reverse-applied" Jan 12 06:04:09 i suppose that indicates the patch has been applied upstream Jan 12 06:04:20 you can try removing it from the recipe and see if that works for you Jan 12 06:04:35 grg: howdy Jan 12 06:04:41 grg: happy ny Jan 12 06:04:41 khem, hello Jan 12 06:04:48 thanks Jan 12 06:04:52 didnt see u for sometime Jan 12 06:05:00 had a bit of a break Jan 12 06:05:07 understandable Jan 12 06:05:12 the world gets in the way sometimes Jan 12 06:05:15 :) Jan 12 06:05:16 summer break :) Jan 12 06:05:30 yeah. Lots of sun and lots of booze Jan 12 06:05:47 Did you stop at all? Jan 12 06:05:55 heh a bit Jan 12 06:06:30 I am planning a vacation in French Polynesia Jan 12 06:06:44 Nice. Jan 12 06:06:58 or hawaii Jan 12 06:07:05 might be in Feb Jan 12 06:07:16 but now at work something is pressing Jan 12 06:07:16 v. interesting places for evolutionary biologists to investigate Jan 12 06:09:27 hmm... linux.conf.au could be cancelled due to flooding! Jan 12 06:11:37 its scheduled to be held at the uni, right in the middle of this map: http://www.abc.net.au/news/infographics/qld-floods/brisbane.htm Jan 12 06:38:53 Lopi: hi.. it's because you have fso-autorev.inc included.. Jan 12 06:39:07 Lopi: drop that patch or wait till i remove libeflvala from fso-autorev.inc Jan 12 06:39:42 JaMa|Off: so just remove it from the recipe? Jan 12 06:41:03 mmt pushing Jan 12 06:45:24 03Martin Jansa 07master * r9a42af313c 10openembedded.git/conf/distro/include/fso-autorev.inc: Jan 12 06:45:24 fso-autorev: drop libeflvala as it's in EFL repository now Jan 12 06:45:24 * and SRCREV there is changed too often without any libeflvala change Jan 12 06:45:24 Signed-off-by: Martin Jansa Jan 12 06:47:31 Lopi: ^^ you can build again Jan 12 06:49:35 JaMa|Off: thanks building again :) Jan 12 06:57:54 Tartarus: which distro do you use for qemuppc? I've used SHR to reuse all -native stuff and it fails quite often in QA ie: ERROR: Error executing a python function in /OE/dev/recipes/update-rc.d/update-rc.d_0.7.bb: Jan 12 06:57:58 KeyError: 'linux-gnu' Jan 12 07:00:04 Tartarus: I'll also add mesa-7.10 today with D_P -1 so you can try it in ppc build Jan 12 07:21:31 JaMa|Off: he has figured the real problem its not related to mesa Jan 12 07:21:35 but config Jan 12 07:25:46 ah ok, then I'll amend 7.10 commit a bit :) Jan 12 07:29:13 grr I forgot to amend before send-email.. so Tartarus, khem please ignore ppc note in mesa-7.10 patch on ML Jan 12 07:44:08 03Bernhard Reutner-Fischer 07master * r0673c3c410 10openembedded.git/recipes/xz/xz.inc: Jan 12 07:44:09 xz: Configure with --enable-small --disable-lzma-links Jan 12 07:44:09 Signed-off-by: Bernhard Reutner-Fischer Jan 12 07:44:09 Signed-off-by: Khem Raj Jan 12 07:47:25 gm Jan 12 07:50:12 gm Jan 12 08:10:31 03Paul Menzel 07master * r57c2ec8554 10openembedded.git/recipes/libcap/libcap_1.10.bb: (log message trimmed) Jan 12 08:10:31 Convert `libcap_1.10` to new staging system by removing `do_stage()` [1]. Jan 12 08:10:31 This change is build tested with the following configuration and `PREFERRED_BINUTILS_VERSION = "2.21"`. Jan 12 08:10:31 Build Configuration: Jan 12 08:10:31 BB_VERSION = "1.10.2" Jan 12 08:10:31 METADATA_BRANCH = "org.openembedded.dev" Jan 12 08:10:32 METADATA_REVISION = "224c193" Jan 12 08:10:38 03Paul Menzel 07master * rac35c02c47 10openembedded.git/recipes/libcap/ (libcap2_2.16.bb libcap_1.10.bb): (log message trimmed) Jan 12 08:10:38 libcap, libcap2: use automatic packaging to fix QA errors Jan 12 08:10:38 Since commit 6883f6 [1] packages have been populated manually. Jan 12 08:10:38 FILES_${PN} = "${libdir}" Jan 12 08:10:38 FILES_${PN}-dev = "${includedir}" Jan 12 08:10:38 This gives the following QA errors for `libcap` and `libcap2` (which is based on `libcap` [2]) [3]. Jan 12 08:10:39 • `libcap`: Jan 12 08:10:43 03Paul Menzel 07master * rf4f7638828 10openembedded.git/recipes/vdr/ (10 files in 4 dirs): (log message trimmed) Jan 12 08:10:43 vdr: update to 1.7.16 from 1.7.10 Jan 12 08:10:43 Several of Hennig’s patches were applied upstream and are not necessary anymore. Jan 12 08:10:43 • `recipes/vdr/files/cplusplus.patch`: replaced by `recipes/vdr/vdr-1.7.16/0002-tools.h-include-stdarg.h.patch` Jan 12 08:11:36 1. `canonicalize_file_name()` is available in uClibc now. Jan 12 08:11:36 2. Deprecated C headers will remain in VDR instead of C++ headers [1]. Jan 12 08:11:36 • `recipes/vdr/files/disable_plugin.patch`: removed, since ncurses is found correctly now. Jan 12 08:12:13 khem: my assignment result http://build.shr-project.org/tests/jama/binutils.overflow.test.tgz Jan 12 08:12:27 khem: failing ld call is included in ld.call.sh Jan 12 08:25:49 good morning Jan 12 08:26:00 hi mckoan Jan 12 08:35:52 03Martin Jansa 07master * re7e75a2d19 10openembedded.git/ (3 files in 3 dirs): Jan 12 08:35:52 e-base: bump EFL SRCREV a bit again to drop libeflvala patch Jan 12 08:35:52 Signed-off-by: Martin Jansa Jan 12 08:41:56 JaMa: ok thx Jan 12 08:42:15 I will see whats happening Jan 12 08:44:38 JaMa: does it happen only on armv4t ? Jan 12 08:45:06 can you try to add ARM_INSTRUCTION_SET = "arm" to boost and see if it happens then too Jan 12 08:51:33 khem: I'll recheck but sofar I've seen it only on armv4t Jan 12 08:53:07 ok Jan 12 08:53:22 I think ARM_INSTRUCTION_SET = "arm" will make you move forward Jan 12 08:53:24 then Jan 12 08:53:35 arm build is already running :) Jan 12 08:53:50 JaMa: I sent a patch to eglibc which will impact all distros using it Jan 12 08:53:58 please review/test it if possible Jan 12 08:54:09 along with the gcc patch I posted earlier :) Jan 12 08:54:26 ofcourse both patches are unrelated Jan 12 08:54:58 gcc patch already built and tested in efl build Jan 12 08:55:03 k Jan 12 08:55:24 whenever it goes thru your builds sign it off Jan 12 08:55:27 pand eglibc you mean that about features which was applied few days ago? Jan 12 08:55:36 yes Jan 12 08:55:39 s/pand/and/g Jan 12 08:55:42 that was just beginning Jan 12 08:55:50 I posted one just now Jan 12 08:55:54 which really enables it Jan 12 08:56:17 ok my inbox is waiting :) Jan 12 08:56:17 for most distos it should work as usual Jan 12 08:57:02 distros like micro and slugos will be happy to have this feature Jan 12 08:57:15 as they can turn off the bogus knobs they dont need Jan 12 08:57:21 and reduce the code size Jan 12 08:57:48 yup it looks nice Jan 12 08:58:42 for SHR I'll probably keep default as smalles device we support has 128mb flash, but it's still interesting for memory footprint as we have little ram Jan 12 08:59:03 yes Jan 12 08:59:34 you can play around by defining your own set of DISTRO_FEATURES_EGLIBC in distro conf or local conf Jan 12 09:00:01 and leave out the features your distro doesnt need Jan 12 09:00:26 03Koen Kooi 07org.openembedded.dev * r99c91adb06 10openembedded.git/recipes/ntp/ntp.inc: Jan 12 09:00:26 ntp-ssl: fix PACKAGES Jan 12 09:00:26 Signed-off-by: Koen Kooi Jan 12 09:01:39 khem: forcing arm mode in boost made it built Jan 12 09:02:51 JaMa: there we go Jan 12 09:03:00 thats your plan B :) Jan 12 09:03:14 post a patch to enable arm mode for now Jan 12 09:03:21 with comment why we do it Jan 12 09:03:26 until linker is fixed Jan 12 09:03:56 * khem -> sleep () Jan 12 09:03:59 gn Jan 12 09:04:31 JaMa: and go ahead with all Xorg patches Jan 12 09:04:36 look ok to me Jan 12 09:04:53 nite khem Jan 12 09:05:30 khem: I'll wait for 2nd ack as we moved to 7.6 only few days ago and maybe few people will be interested in 7.5 -> 7.6 diff Jan 12 09:16:56 03Martin Jansa 07master * r5511dc1939 10openembedded.git/recipes/e17/ (2 files in 2 dirs): Jan 12 09:16:57 elmdentica: add patch for azy api change from r56034 Jan 12 09:16:57 Signed-off-by: Martin Jansa Jan 12 09:17:04 03Martin Jansa 07master * rb1a15f7dc1 10openembedded.git/recipes/boost/boost_1.45.0.bb: Jan 12 09:17:04 boost-1.45.0: force ARM mode for now, thumb makes ld to buffer overflow Jan 12 09:17:04 * can be removed as soon as binutils are fixed Jan 12 09:17:04 * testcase is here http://build.shr-project.org/tests/jama/binutils.overflow.test.tgz with failing ld call included in ld.call.sh Jan 12 09:17:05 Signed-off-by: Martin Jansa Jan 12 09:17:15 03Martin Jansa 07master * rf578c96b87 10openembedded.git/recipes/mesa/ (6 files in 2 dirs): Jan 12 09:17:15 mesa: add new 7.10 version, with D_P -2 Jan 12 09:17:15 * someone please test uclibc build if old uclibc.patch is enough Jan 12 09:17:15 Signed-off-by: Martin Jansa Jan 12 12:14:34 jo obi Jan 12 12:39:39 hello here. does someone already see this kind of error building gettext-native http://tinderbox.openembedded.org/builds/118113/ Jan 12 12:40:06 http://tinderbox.openembedded.org/public/logs/task/18953738.txt Jan 12 12:41:56 usr/bin/per ?? that looks like someone typoed Jan 12 12:44:47 XorA|gone: yope but I still don't find where and visibly I am the only one who have it despite the fact that it is on a native target Jan 12 12:45:11 yeah its a mystery to me Jan 12 12:48:52 XorA|gone: I have launch grep -R "bin/per " tmp/ like ten minutes ago and it is still not comed back :) Jan 12 12:50:02 GarthPS: you should start in tmp/work/x86*/gettext-native :) Jan 12 12:51:37 JaMa: hmmmm yeap :) Jan 12 12:52:46 JaMa: but I have some dishes to do :) Jan 12 12:53:09 :) Jan 12 13:11:15 XorA|gone, thanks for running the vote Jan 12 13:14:04 hi Jan 12 13:14:11 hi ericben Jan 12 13:14:18 hi florian Jan 12 13:14:51 GarthPS: I also meet perl problems with configure on native recipes when building from scratch but not this error Jan 12 13:14:55 hi florian hi ericben Jan 12 13:15:07 hi woglinde Jan 12 13:18:18 is it possible to select a patch conditionnaly ? Example : with linux 2.6.37 input abi changed, thus we need 2.6.37 kernel headers. But now this creates a problem in strace where a patch is needed to compile against kernel headers >= 2.6.32 : is there an easy way to apply this patch if kernel headers' version is >= 2.6.32 ? Jan 12 13:19:22 ericben no Jan 12 13:22:32 ericben: did you run into non-working tslib with 2.6.37? Jan 12 13:24:44 JaMa: yes Jan 12 13:25:23 have you tried if evdev works fine? Jan 12 13:25:37 JaMa: message is something like "this device is not a touchscreen I understand", only because input has increased its ABI number Jan 12 13:26:02 https://github.com/radekp/linux-2.6/commit/84af470fc07070fd5a129bfdcbd7c25978f8e778 :) Jan 12 13:26:07 JaMa: I solved the problem by siwtching to kernel headers 2.6.37 and having tslib compiled against that Jan 12 13:26:21 JaMa: yes I also did that before :-) Jan 12 13:26:29 but ths's not clean for the future Jan 12 13:26:34 ericben: and how did you fix it ? Jan 12 13:26:59 GarthPS: kernel headers 2.6.37 + patch for strace to make it compile against these new headers Jan 12 13:27:06 ericben: is it backward compatible? after you rebuild tslib against .37? Jan 12 13:27:35 JaMa: is what backward compatible ? Jan 12 13:27:41 ericben: about perl error ?? Jan 12 13:28:03 GarthPS: in my case it was libperl.so not found. So relaunching bitbake solves the error Jan 12 13:28:19 but in your case it's very strange as the l of perl is missing Jan 12 13:28:19 ericben: if you build tslib against .37 headers, does it still work on target with ie .34 kernel? Jan 12 13:28:32 JaMa: no idea but I think that will fail Jan 12 13:28:44 JaMa: because ABI won't be in sync Jan 12 13:29:17 I've seen patch somewhere checking ABI version not for equal but just >= Jan 12 13:30:35 JaMa: I don't know why the ABI number was increased but for touchscreen tslib seems to work fine without any change Jan 12 13:30:57 so a >= could be a solution for tslib Jan 12 13:31:07 because just rebuilding against 2.6.37 is not an option if you support multiple devices where some are running older kernels than .37 Jan 12 13:32:23 JaMa: that's right Jan 12 13:34:57 any idea how I can tune the SRC_URI to fetch something like this : http://git.busybox.net/busybox/patch/?id=8030a1484917d5b71d5ccd1a1d28a29da7a3d7f0 ? Jan 12 13:36:17 problem is that the file is fetched as "index.html\?id\=8030a1484917d5b71d5ccd1a1d28a29da7a3d7f0" and the fetcher is not happy with the name Jan 12 13:37:43 ericben: https://patchwork.kernel.org/patch/331091/ Jan 12 13:42:52 JaMa: seems like a hack to get something working fast :-) Jan 12 13:43:20 yes and that's the same as what I just did in my OE checkout :) Jan 12 13:53:11 03Richard Purdie 07master * r259dd464ab 10bitbake.git/lib/bb/ui/knotty.py: Jan 12 13:53:11 bitbake/knotty.py: Ensure task note messages are only surpressed at default log levels Jan 12 13:53:11 (From Poky rev: 3f08b7f1eadc8d181c76339d14c7abb4f49cad57) Jan 12 13:53:11 Signed-off-by: Richard Purdie Jan 12 14:06:13 JaMa: this suppose the ABI is kept compatible. Not sure that's the case as I didn't check why it was increased. Jan 12 14:06:51 ok for my fetch problem from git.busybox.net : simply removing the "?" fixe it Jan 12 14:06:58 just sent it as RFC.. maybe someone will need to get something working fast too :) Jan 12 14:47:19 hi, here's my issue: Jan 12 14:47:29 I'm trying to get a more recent version of linphone Jan 12 14:47:36 in order to see if it has more audio controls Jan 12 14:47:40 so.... Jan 12 14:47:47 it has nested configure files Jan 12 14:48:02 one of the nexted configure files in a subdir has a libtool mismatch Jan 12 14:48:10 what's the best way to handle this: Jan 12 14:48:19 un-nest the configure? Jan 12 14:48:24 like for instance: Jan 12 14:48:41 configuring each dir separately and hacking the main configure.ac not to run them? Jan 12 14:49:15 ha JaMa gave me a tip that I'll try Jan 12 14:49:25 the thing is that I was not willing to autoreconf manually Jan 12 14:49:35 hey JaMa Jan 12 14:49:37 without autoreconf Jan 12 14:49:40 Tartarus: hey Jan 12 14:49:57 like in the bbclass Jan 12 14:50:22 there's the fix for mesa+ppc, once CIA sees it Jan 12 14:51:00 s/bbclass/autotools bbclass/ Jan 12 14:51:01 03Tom Rini 07master * r223ce2db77 10openembedded.git/recipes/apache2/apache2_2.2.14.bb: Jan 12 14:51:01 apache2: Fix invoking buildconf Jan 12 14:51:01 It may not have execute set so start it via sh Jan 12 14:51:01 Signed-off-by: Tom Rini Jan 12 14:51:13 03Tom Rini 07master * r4735968287 10openembedded.git/site/powerpc-common: Jan 12 14:51:13 site/powerpc-common: Drop ac_cv_func_dlopen Jan 12 14:51:13 This is something that can be done at compile time and more importantly Jan 12 14:51:13 was set wrong (this should be no and then another one yet for Jan 12 14:51:14 getting -ldl, but that also wouldn't live in the -common one). Jan 12 14:51:14 Signed-off-by: Tom Rini Jan 12 14:52:25 Tartarus: nice :) Jan 12 15:00:48 ok solved Jan 12 15:00:53 thanks JaMa Jan 12 15:04:09 JaMa: building with your tslib patch & 2.6.31 headers for a 2.6.37 kernel. Jan 12 15:14:52 JaMa: you have go tan idea ? http://pastebin.com/YsuFzG1x Jan 12 15:15:17 JaMa: perhaps I nead to flush my build host a bit ? Jan 12 15:18:50 GarthPS: I guess that path is part of varible in this case probably STAGING_BINDIR_NATIVE, so look only for "per " without bin/ Jan 12 15:19:17 JaMa: hm right! Jan 12 15:21:52 JaMa: ok EV_VERSION was increased to support "large scancodes" in commit 8613e4c2872a87cc309a42de2c7091744dc54d0e so only keyboard seems to be affected by the ABI bump Jan 12 15:24:21 ericben: good Jan 12 15:31:07 JaMa: ... "per " is too commun.. hav go tons of it. I am good form flushing my build i think... Jan 12 15:31:22 hej, anyone has a stable 2.6.36+2.6.37-rc8 kernel? just did a clean build for beagleboard/angstrom and got this kernel; it doesn't boot (atleast not with any serial output) correctly, however Jan 12 15:33:08 mlip: I'm was using it few days since I added, then upgraded to 2.6.37 release on spitz Jan 12 15:34:15 JaMa: do you get any output on the debug uart? I did build a MLO/uboot/uimage for my beagleboard-xm and can't get past: "Uncompressing Linux... done, booting the kernel." Jan 12 15:34:38 u-boot seems to be fine, since it can boot the kernel on the shipped sd card (2.6.32 i think) Jan 12 15:35:22 mlip: I'm booting it from angelic (sharp) bootloader as linux-kexecboot and then 2nd stage and with quiet in CMDLINE so no much output there Jan 12 15:36:33 JaMa: ah, ok; but atleast nothing is screwed up configwise i guess; was just wondering about the version, which is really at the bloody edge imho Jan 12 15:42:03 03Martin Jansa 07master * re9bacee8f5 10openembedded.git/recipes/e17/e-wm_svn.bb: Jan 12 15:42:03 e-wm: bump SRCREV a bit more than EFL_SRCREV is now Jan 12 15:42:03 * ie illume-keybard makes E segfault in some configurations Jan 12 15:42:03 Signed-off-by: Martin Jansa Jan 12 15:47:53 JaMa: is there is liter flushing version of rm -rf tmp/ ? Jan 12 15:50:20 that depends on what you're trying to do. Jan 12 15:57:15 FOSDEM RSN Jan 12 15:57:18 eek Jan 12 15:57:20 hmm Jan 12 15:57:24 better get a ticket Jan 12 15:57:44 got a sponsor (I don't) ? Jan 12 15:57:52 not really Jan 12 15:57:54 myself Jan 12 15:58:26 yeah and the ticket price is infinitely high compared to the entrance price... Jan 12 15:58:31 yes Jan 12 15:58:32 hi hrw Jan 12 15:58:39 how is Texas Jan 12 15:58:46 even in absolute terms. Jan 12 15:58:55 hrw in Texas? Jan 12 15:59:32 yep Jan 12 16:00:53 which of the semi's? Jan 12 16:01:46 03Chris Larson 07master * r3d96044bd1 10openembedded.git/recipes/devmem2/ (devmem2.bb devmem2/devmem2-fixups-2.patch): Jan 12 16:01:46 devmem2: add freescale alignment fixup patch Jan 12 16:01:46 Signed-off-by: Chris Larson Jan 12 16:01:47 03Chris Larson 07master * r05df2e6157 10openembedded.git/recipes/rng-tools/rng-tools_2.bb: Jan 12 16:01:48 rng-tools: dep on argp-standalone for uClibc Jan 12 16:01:48 argp is hierarchical argument parsing built on top of getopt, but is part of Jan 12 16:01:48 glibc. Niels Möller created a standalone version, so make rng-tools depend Jan 12 16:01:48 upon that when we're using uClibc. Jan 12 16:01:49 Signed-off-by: Chris Larson Jan 12 16:14:59 kergoth: fs's patch to devemem seems to comment out write in case 'w' Jan 12 16:15:26 kergoth: is that a wanted fix ? Jan 12 16:17:58 * kergoth double checks Jan 12 16:18:39 good catch, thanks :) Jan 12 16:18:39 ++// *((unsigned long *)virt_addr) = write_val; Jan 12 16:18:42 * kergoth nods Jan 12 16:18:47 that doesn't look right to me :) Jan 12 16:18:53 fsl's patches are usually to be considered with care ... Jan 12 16:19:40 will be more careful in the future, thanks for the tip Jan 12 16:20:09 03Chris Larson 07master * r435e8eebb3 10openembedded.git/recipes/devmem2/ (devmem2.bb devmem2/devmem2-fixups-2.patch): Jan 12 16:20:09 devmem2: remove comment of word write from fsl patch Jan 12 16:20:09 Signed-off-by: Chris Larson Jan 12 16:20:30 np, I've lost so many hours thanks to some of their patches :-D Jan 12 16:21:04 hehe Jan 12 16:21:11 sounds "fun" Jan 12 16:21:59 most interesting one was a BSP with a several hundreds pages validation documents ... and a faulty get_gpio_level function :-D Jan 12 16:22:27 haha Jan 12 16:22:29 smooth Jan 12 16:23:09 who's the slugos guy in here again? Jan 12 16:23:19 mwester? Jan 12 16:23:26 ah, thanks Jan 12 16:23:27 I think so Jan 12 16:24:13 * kergoth verifies rng-tools build for uClibc Jan 12 16:29:06 * florian shoots samba Jan 12 16:29:19 or gcc rather Jan 12 16:31:51 https://gist.github.com/776390 Jan 12 16:33:10 kergoth: looks nice Jan 12 16:33:35 of course, ideally bitbake would have more sensible defaults in this regard itself, but.. Jan 12 16:33:38 :) Jan 12 16:33:56 (other than for parallel parsing, which does) Jan 12 16:34:06 bbiab Jan 12 16:34:32 kergoth: are these factors (2 for the compilers and 1.5 for the threads) know to be good? Jan 12 16:35:16 they seem to be, but i'm sure it depends on disk performance and the like as well. some of us use these defaults Jan 12 16:35:50 maybe we need up front benchmarks ;) Jan 12 16:36:24 :-) Jan 12 16:46:01 khem, minimal-uclibc.conf has this (now inappropriate) setting: # glib-2.0 needs ipv6 and there is no knob to control it Jan 12 16:46:02 DISTRO_FEATURES += "ipv4 ipv6" Jan 12 16:46:40 you could potentially turn ipv6 off again now, JFYI Jan 12 17:00:16 03Klaus Kurzmann 07org.openembedded.dev * r0cc07b0127 10openembedded.git/recipes/freesmartphone/cornucopia.inc: Jan 12 17:00:16 cornucopia.inc: bump SRCREV for latest fixes Jan 12 17:00:16 Signed-off-by: Klaus Kurzmann Jan 12 17:04:03 re Jan 12 17:16:59 So... Jan 12 17:17:11 We do a lot of hoop-jumping to not use the lzo2 recipe we (now have?) for mtd-utils Jan 12 17:17:33 i guess due to library upgrade stuff Jan 12 17:33:31 kergoth: Patch devmem2-fixups-2.patch does not apply (enforce with -f) Jan 12 18:30:06 kergoth: Patch devmem2-fixups-2.patch does not apply (enforce with -f) Jan 12 18:32:24 03Koen Kooi 07org.openembedded.dev * r163019f750 10openembedded.git/recipes/u-boot/u-boot-sakoman_git.bb: Jan 12 18:32:24 u-boot-sakoman: move to 2010.12 Jan 12 18:32:24 Tested on pandaboard EA1 and A1 ES2.0 boards Jan 12 18:32:24 Signed-off-by: Koen Kooi Jan 12 18:32:38 03Koen Kooi 07org.openembedded.dev * refb7475cdf 10openembedded.git/recipes/ti/ti-cgt6x_7.0.4.bb: Jan 12 18:32:38 ti-cgt: add 7.0.4 Jan 12 18:32:38 This adds support for mini64 aka 'Tesla' Jan 12 18:32:38 Signed-off-by: Koen Kooi Jan 12 18:36:28 03Koen Kooi 07org.openembedded.dev * r6e749b6713 10openembedded.git/recipes/linux/ (126 files in 10 dirs): Jan 12 18:36:29 linux-omap: move to 2.6.37 final Jan 12 18:36:29 There is no 2.6.37-omap1 tag and linus' tree is missing fixes, so patch up -rc8 with the patches from mainline Jan 12 18:36:29 Signed-off-by: Koen Kooi Jan 12 18:38:38 blindvt`: right ok Jan 12 18:41:47 gm all Jan 12 18:41:48 khem, btw.. i'm a bit puzzled why nobody else seems to need something like http://paste.debian.net/104444/ Or did i miss a different, magic single spot where that's propagated already? Jan 12 18:43:06 blindvt: you are assuming that enable|disable options are same for all packages Jan 12 18:43:10 thats not the cas Jan 12 18:43:43 and in some cases the flag is not there but assuptions are Jan 12 18:43:54 so we turn the assumption off ourselves in the recipe Jan 12 18:43:59 based on DISTRO_FEATURE Jan 12 18:44:27 khem, autoconf configure just ignores it if a flag isn't supported, so it's fine to pass to packages that don't handle it. Either way, my intermediate compiler fails to bootstrap in the lto-plugin if i don't configure it properly Jan 12 18:45:07 blindvt: thats ok though Jan 12 18:45:15 it wont har Jan 12 18:45:17 m Jan 12 18:47:28 03Koen Kooi 07org.openembedded.dev * rba886172c9 10openembedded.git/recipes/linux/linux-omap-2.6.37/dvfs/0004-OMAP3-Overo-Tide-enable-upto-720MHz-OPP.patch: Jan 12 18:47:28 linux-omap 2.6.37: check in updated overo patch Jan 12 18:47:28 Signed-off-by: Koen Kooi Jan 12 18:48:25 khem, we could as well add global ${DISABLE_NLS} / ${DISABLE_LARGEFILE} etc and pass that to the packages that support it, but that's much more work and error-prone to miss on updates. We could also scan configure{,.ac} for the mention of the respective macros and only supply it if supported. That would work but i'm not sure if it's better. What do you think? Jan 12 18:54:37 khem, re: so we turn the assumption off ourselves in the recipe Jan 12 18:56:05 khem, yea, i've seen that and i'm not too happy about it since that's rather buggy atm as the seed for the cache doesn't take any configuration into account :( Jan 12 18:56:55 in any case adding it to auttools wont hurt Jan 12 18:57:30 and we need ac cache seeding fixed Jan 12 19:07:35 khem: oh ya, where is the example fix for bison+uclibc? Jan 12 19:07:44 you were saying there was another app with that problem Jan 12 19:07:48 and it got fixed :) Jan 12 19:08:59 Tartarus: let me remember Jan 12 19:09:05 whats a usable webserver for embedded stuff? Jan 12 19:09:13 doesn't need to have bells and whistles Jan 12 19:09:20 lighttpd? Jan 12 19:09:53 busybox? :) Jan 12 19:09:55 BlindMan: yeah or cherokee Jan 12 19:10:10 Tartarus: no busybox system ... but it has one? wow Jan 12 19:10:23