**** BEGIN LOGGING AT Sat Mar 28 02:59:57 2009 Mar 28 08:33:57 hi all ... Mar 28 08:34:11 what's the best branch to test minimal distro ? Mar 28 08:34:19 .dev Mar 28 08:34:48 thanks hrw|gone ! Mar 28 08:43:12 morning Elena ! Mar 28 11:13:24 morning all Mar 28 12:13:26 03Koen Kooi  07org.openembedded.dev * r5554bd6fde 10openembedded.git/ (5 files in 3 dirs): packagekit: add 0.4.5 Mar 28 12:13:39 03Koen Kooi  07org.openembedded.dev * r9d5b227e67 10openembedded.git/ (conf/checksums.ini recipes/gnome/gnome-keyring_2.26.0.bb): gnome-keyring: update to 2.26, adjust packaging accordingly Mar 28 12:36:16 03Koen Kooi  07org.openembedded.dev * r5b6bb392f2 10openembedded.git/conf/distro/include/sane-srcrevs.inc: Mar 28 12:36:16 sane-srcrevs: bump opkg revision Mar 28 12:36:16 * runtime tested on beagleboard Mar 28 12:36:27 03Koen Kooi  07org.openembedded.dev * radc3749d8c 10openembedded.git/ (4 files in 3 dirs): Mar 28 12:36:27 webkit-gtk: bump SRCREV Mar 28 12:36:27 * runtime tested on beagleboard with midori and epiphany Mar 28 12:58:04 hi, anybody with zaurus c1000 ? Mar 28 12:59:37 hi all Mar 28 12:59:38 I am trying to compile x264 with OE. I put in .bb in EXTRA_OECONF = "--extra-cflags=\"-O3 -Wall\"" .. but when I try to build the package using bitbake ./configure takes only the first argument .. Here it will take only -O3 and say Unknown option -Wall", ignored (in log.configure) Mar 28 12:59:43 If I ./configure it from the command line (without bitbake).. --extra-cflags can take lots of arguments... Mar 28 13:01:19 flameman: hey Mar 28 13:01:21 flameman: I'm Mar 28 13:01:46 how pretty is the kernel 2.6 support ? Mar 28 13:01:54 i am buying a unit from ebay Mar 28 13:02:59 flameman: I have old 2007.1x release of angstrom with 2.6.23 Mar 28 13:03:04 all works fine Mar 28 13:03:18 2.6.26 have issue with charging Mar 28 13:03:30 support is good enough Mar 28 13:03:41 :) good Mar 28 13:03:55 at least now you will be not alone ;) Mar 28 13:03:57 the c3200 is hard to be found ... so i buy this c1000 Mar 28 13:05:07 Sinky: try to use single quotes inside (--extra-cflags='-.... ') Mar 28 13:06:03 Jay7 ... umm i haven't seen yet: which is the zaurus kernel support from ? from hh ? Mar 28 13:06:36 i am trying to build a beagle demo image, i get the error during compilation of ffmpeg. please see has anyone faced problem building beagle demo in angstrom distro, this happens during compilation of ffmpeg: http://www.pastebin.ca/1374810. does anyone know how to get around this problem? Mar 28 13:06:52 flameman: no, linux-rp is Mar 28 13:06:58 by RP (tm) :) Mar 28 13:07:37 Jay7 uuuuuu good :D Mar 28 13:07:47 Jay7 I tried Mar 28 13:08:14 Jay7 I have no issue with another package - ffmpeg Mar 28 13:14:59 the issue i reported is from gstreamer/gst-ffmpeg package Mar 28 13:16:13 how can i generate the dependency graph for my image.bb file Mar 28 13:26:55 likewise Hi Mar 28 13:29:03 hi Mar 28 13:29:34 cupcake: bitbake -g IIRC Mar 28 13:29:45 bitbake --help should list the options Mar 28 13:29:57 03Philip Balister  07org.openembedded.dev * rfca3d79da6 10openembedded.git/recipes/iperf/ (iperf.inc iperf_2.0.2.bb): iperf_2.0.2.bb : Convert to use .inc file Mar 28 13:30:07 03Philip Balister  07org.openembedded.dev * r9bdbc57bcd 10openembedded.git/ (3 files in 2 dirs): iperf_2.0.4 : Add new version of iperf. Iperf measures network bandwidth. Mar 28 13:32:06 like2wise yesterday I asked the question about --extra-cflags .. do you remember Mar 28 13:32:13 Sinky: ah yes Mar 28 13:32:17 for x264 Mar 28 13:32:24 yes Mar 28 13:32:25 (16:14) likewise !(n=Leon_Woe@atwork-193.r-212.178.107.atwork.nl): Sinky: how would you type it from the command line. Can you use pastebin.ca to show me? Mar 28 13:32:31 Sinky: I would like to help, I want to test x264 as well Mar 28 13:32:57 Sinky: yeah, I probably went off being occupied with something else (lay-offs) Mar 28 13:33:05 cool.. I still didn't solve the issue Mar 28 13:33:11 do u have some time now Mar 28 13:33:29 Sinky: yes, sure. you already have a recipe that *almost* works? Mar 28 13:34:40 like2wise yes, almost :) Mar 28 13:34:54 just want to do some optimizations Mar 28 13:35:17 like2wise here it is from command line: http://stancho.com/pastebin/1903 Mar 28 13:35:27 it is ok Mar 28 13:37:00 Sinky: and a bitbake recipe? Mar 28 13:37:06 or do we have it in OE already? Mar 28 13:37:39 yes Mar 28 13:38:03 btw I have successfully compiled and runed x264.. just have an issue with --extra-cflags Mar 28 13:38:44 like2wise here it is log.do_configure Mar 28 13:38:45 http://stancho.com/pastebin/1904 Mar 28 13:39:03 Sinky: I don't see the recipe in oe Mar 28 13:39:16 w8 Mar 28 13:40:16 like2wise http://stancho.com/pastebin/1905 Mar 28 13:41:43 Sinky: have you tried: EXTRA_CFLAGS = '--extra-cflags="-O3 -Wall"' Mar 28 13:41:58 single quotes surrounding the stuff Mar 28 13:42:58 Sinky: in the meantime, I'm making some HD space and a new checkout. Mar 28 13:45:47 like2wise I tried lots of variants.. I will trhis now Mar 28 13:48:34 like2wise this way I get: /home/bitbake/oe/tmp/work/armv7a-angstrom-linux-gnueabi/x264-0.0-r0/temp/run.do_configure.3576: line 515: -Wall : command not found Mar 28 13:51:59 Sinky: ok let me see if there are other recipes which do this Mar 28 13:52:32 Sinky: ffmpeg Mar 28 13:53:50 like2wise http://stancho.com/pastebin/1906 Mar 28 13:54:03 you can take in mind TARGET_CFLAGS variable Mar 28 13:54:43 Sinky: just set ECFLAGS first, and use --extra-cflags="${ECFLAGS}" and see what it does Mar 28 13:55:07 Sinky: I'm also just guessing here Mar 28 13:58:19 like2wise you mean: Mar 28 13:58:19 ECFLAGS = "-O3 -Wall"" Mar 28 13:58:19 EXTRA_OECONF = "--extra-cflags=${ECFLAGS}" Mar 28 13:58:29 correct Mar 28 14:01:52 like2wise http://stancho.com/pastebin/1907 ... I see -O3 -Wall is not in quotes ? Mar 28 14:02:18 I will try with EXTRA_OECONF = "--extra-cflags="${ECFLAGS}"" Mar 28 14:02:51 or ECFLAGS = ""-O3 -Wall"" Mar 28 14:06:12 like2wise it is strange that in ffmpeg.bb I do not have to put char escaping on "${TARGET_CFLAGS}" Mar 28 14:06:53 Sinky: note that ffmpeg recipe override do_configure() Mar 28 14:07:00 override*s* Mar 28 14:07:15 yes Mar 28 14:08:03 like2wise in that case I will also try to override configure Mar 28 14:13:32 Sinky: frustrating indeed... Mar 28 14:14:30 :( Mar 28 14:17:03 Sinky: I am building ffmpeg and seeing what the temp/run.* stuff looks like Mar 28 14:20:12 like2wise cool Mar 28 14:20:24 like2wise I am trying with this recipe: http://stancho.com/pastebin/1908 Mar 28 14:22:50 like2wise this way I got this log.do_configure: Mar 28 14:22:59 http://stancho.com/pastebin/1909 Mar 28 14:23:06 like2wise do you think it should be ok ? Mar 28 14:24:30 can In generate opkg packages containing the headers for libs and libc etc? Mar 28 14:24:54 for use on target or some "sdk" like target Mar 28 14:26:44 Sinky: you can inspect how "./configure" was run in the /tmp/work/.../.../temp/run.do_configure* file Mar 28 14:27:28 like2wise the last pastebin is exact this log Mar 28 14:28:07 Sinky: no that's log.do not run.do Mar 28 14:28:33 the temp/run. is what you would type on a shell, the temp/log. is the output Mar 28 14:30:14 like2wise op yes sorry Mar 28 14:30:53 Sinky: in the meantime, ffmpeg build is gonna take a while here.... Mar 28 14:31:23 like2wise ok Mar 28 14:31:30 like2wise this is http://stancho.com/pastebin/1910 Mar 28 14:31:36 from run.do_configure Mar 28 14:31:45 i think it should be ok... Mar 28 14:31:54 Sinky: yup looks good! Mar 28 14:32:45 Sinky: I would fold STD_CONF into EXTRA_OECONF Mar 28 14:33:19 sinky: forget that, your STD_CONF is illegal, target specific Mar 28 14:33:41 in x264 ? Mar 28 14:33:59 shouldn't EXTRA_OECONF = "--extra-cflags="${RND_TARGET_CFLAGS}"" just work? Mar 28 14:42:35 Sinky: you're building this for beagleboard, right? Mar 28 14:43:46 how did the #oe-stable discussion go? Mar 28 14:44:25 like2wise right Mar 28 14:45:33 jkridner: dunno of such a discussion...?! Mar 28 14:47:00 it was discussed on the oe mailing list yesterday. Mar 28 14:47:09 there was an IRC meeting the day before. Mar 28 14:47:53 crofton, hrw, denix, and others were part of the mail thread. Mar 28 14:48:41 not sure why they broke it out on another channel. Mar 28 14:52:09 like2wise, see list Mar 28 14:52:15 will do Mar 28 14:52:17 tnx Mar 28 14:52:18 minimize chaos :) Mar 28 14:52:55 Sinky: I have a recipe that might be acceptable for commit. I need to reboot firefox/irc because the firefox upgrade made my running firefox/irc/pastebin broken... Mar 28 14:53:00 see you in a sec Mar 28 14:55:10 Sinky: http://stancho.com/pastebin/1911 Mar 28 14:55:26 notice ECFLAGS_beagleboard Mar 28 14:55:40 Sinky: also, the disable-asm should be set by default, but not for asm-capable x264 compiles Mar 28 14:55:49 Sinky: this requires some more foo Mar 28 14:59:20 yes like2wise it should be ok now Mar 28 14:59:21 thanks Mar 28 14:59:28 Sinky: did you enable disable-asm? Mar 28 14:59:33 like2wise yes Mar 28 14:59:39 like2wise now I will check vlc for these updates :) Mar 28 14:59:46 in vlc we have CFLAGS="" Mar 28 15:00:33 Sinky: could you test this recipe?? http://www.stancho.com/pastebin/1912 Mar 28 15:01:08 like2wise of course Mar 28 15:01:12 like2wise just a sec Mar 28 15:01:56 Sinky, do you want your email address or name in the x264 recipe commit or not? Mar 28 15:04:32 like2wise Stanislav Popov Mar 28 15:04:35 thanks Mar 28 15:08:46 like2wise ok, it is working Mar 28 15:13:58 like2wise btw isnt't it better to be : EXTRA_OECONF = "${X264_DISABLE_ASM} --extra-cflags='${X264_ECFLAGS}'", but not EXTRA_OECONF = "${X264_DISABLE_ASM} --extra-cflags="${X264_ECFLAGS}"" Mar 28 15:14:02 (with single quotes) Mar 28 15:14:24 maybe, why? Mar 28 15:16:21 like2wise I see in run.do_conf: cfgcmd=" .... --disable-asm --extra-cflags=" ... " " Mar 28 15:16:22 ... I think it is more correct to be in single quotes in case that there are no character escapes Mar 28 15:20:21 morning Mar 28 15:20:30 gm Mar 28 15:20:43 hello, the config.h is normaly generated from the autotools and included for defining things used in the compilation no? Mar 28 15:21:35 In wesnoth 1.6a it includes SDL_config.h and that conflocts with wesnoth's config.h... Mar 28 15:21:49 so my question is: Mar 28 15:22:03 should SDL_config.h be included when compiling wesnoth? Mar 28 15:22:22 Sinky: I can not commit the "-s -save-temps -static " part Mar 28 15:22:49 Sinky: that is, we usually do not want static built binaries Mar 28 15:23:02 like2wise oh yes, it is not so important! Mar 28 15:23:42 about neon and so on is what matters.. Mar 28 15:24:11 lik2wise you can add --enable-shared in the commitment Mar 28 15:33:55 should I care about 16k of memory for kexecboot? Mar 28 15:34:20 * Jay7 will spend this amount for caching values while parsing xpm images Mar 28 15:35:04 03Leon Woestenberg  07org.openembedded.dev * r5e0cdc8eac 10openembedded.git/recipes/vlc/x264_r2245.bb: Mar 28 15:35:04 x264: now uses upstream revision number. added --extra-cflags for beagleboard. Mar 28 15:35:04 Allows for easier optimizing the build using ECFLAGS. Mar 28 15:35:04 Tested-by: Leon Woestenberg Mar 28 15:35:04 Tested-by: Stanislav Popov Mar 28 15:35:07 seems not too big value but.. Mar 28 15:35:19 Sinky: I hope I didn't f*ck up :-) Mar 28 15:36:17 like2wise don't worry :) this package was f*cked up by default :) Mar 28 15:38:31 not many users I think Mar 28 15:39:47 Sinky: what encode speed do you get using those flags? Mar 28 15:43:52 I will make the tests later.. because the results were not very good before and now I am trying to optimize: libx264, ffmpeg and vlc Mar 28 15:45:35 x264 needs lots of horsepower Mar 28 15:46:48 of course Mar 28 15:54:48 Gnutoo: regarding wesnoth. I don't understand your problem. can you elaborate? Mar 28 15:55:41 rschuster, it's a known prolbem...I've 2 options: or use a more recent libsdl or try cmake or scons...i'm trying cmake Mar 28 15:57:50 Gnutoo: that is what I recommend as well Mar 28 15:57:58 ok Mar 28 15:58:06 Gnutoo: ever worked with cmake? Mar 28 15:58:11 no Mar 28 15:58:13 lol Mar 28 15:58:36 I saw a presentation about it but I don't remember it well Mar 28 15:58:43 Gnutoo: its much like autotools Mar 28 15:58:50 ah ok Mar 28 15:59:02 Gnutoo: first you call cmake and that will create you a bunch of ordinary makefiles Mar 28 15:59:17 yes but I don't remember how to see all options Mar 28 15:59:21 Gnutoo: the directory in which you call cmake is the build dir Mar 28 15:59:30 it also contains a file named CMakeCache.txt Mar 28 15:59:34 that one is important Mar 28 15:59:35 ah ok Mar 28 15:59:42 since it contains all the possible build options Mar 28 15:59:47 thanks Mar 28 16:00:00 it is advisable that you do Mar 28 16:00:00 "bitbake -c configure wesnoth" first Mar 28 16:00:16 ok Mar 28 16:00:18 then look into that CMakeCache file and decide what would make sense to modify for OE Mar 28 16:00:25 thanks a lot Mar 28 16:00:54 I needed it because the ebuild of gentoo doesn't define things such as editor Mar 28 16:00:57 in the recipe you can do EXTRA_OECMAKE do define your extra options Mar 28 16:00:58 thanks a lot Mar 28 16:01:14 I mean, give EXTRA_OECMAKE some value :) Mar 28 16:01:17 yes I've already that in my recipe Mar 28 16:01:31 as usual I grepped for knowing how it works Mar 28 16:01:52 ok. syntax is like C preprocessor macros -D whatever Mar 28 16:01:59 there is also one thing I just recently added Mar 28 16:02:08 you can do so called out of tree builds Mar 28 16:02:19 that is builddir != srcdir Mar 28 16:02:46 llvm 2.5 requires this. so I had no choice Mar 28 16:03:01 usually programs allos both worlds Mar 28 16:03:31 out-of tree building should be preferred because it will not mess around in the directories in which the sources lie Mar 28 16:03:31 yes I saw that on the mailing list Mar 28 16:03:45 have a look at the llvm_2.5.bb recipe to see how it works Mar 28 16:03:45 ok Mar 28 16:03:53 ok thanks a lot Mar 28 16:04:17 you need to apply a trick with EXTRA_OEMAKE (the value for normal make) to get it work Mar 28 16:05:20 ah ok Mar 28 16:05:28 its not hard, you will see :) Mar 28 16:05:32 ok thanks Mar 28 16:05:39 you're welcome :) Mar 28 16:06:02 we should document stuff like this ... Mar 28 16:06:09 it's not hard unless I have problems related to cmake(and not to oe)...I hope I won't have any Mar 28 16:06:55 the biggest change you will notice is that cmake does not provide the usual Unixy directory layout Mar 28 16:07:27 that is. with automake people can say that a specific file belongs to e.g. localstatedir Mar 28 16:07:46 and that will be translated to /usr/share/appname at some point Mar 28 16:07:54 (AFAIK) Mar 28 16:08:18 in cmake they will usually hardcode those paths Mar 28 16:09:41 the lack of this formalism is my biggest grief with cmake, since this creates more burden for us Mar 28 16:10:39 otherwise it quite nice and it makes those pesky windows people happy since it supports their proprietary toolchains ... Mar 28 16:16:27 like2wise any idea about that http://stancho.com/pastebin/1915 ... http://stancho.com/pastebin/1916 Mar 28 16:38:04 ok thanks Mar 28 16:52:05 Sinky: tried with double quotes? Mar 28 16:52:13 like2wise yes, the same Mar 28 16:52:33 Sinky: the configure command is interpreting the CFLAGS Mar 28 16:52:48 CFLAGS should probably be somewhere else. Mar 28 16:53:16 Sinky: I think in the compile stage as in CFLAGS="foobar" make Mar 28 16:53:30 Sinky: again, easiest is to learn from existing recipes here Mar 28 16:54:34 Sinky: for example, recipes/phalanx/phalanx_22.bb:13: oe_runmake CC="${CC}" CFLAGS="${CFLAGS}" STRIP=echo LD="${CC}" Mar 28 16:55:04 Sinky: --extra-cflags was a configure-stage item, CFLAGS is a compile-stage item Mar 28 17:01:15 like2wise yes, maybe I should set: CFLAGS_append Mar 28 17:34:41 mmm...I think I've a problem with cmake...it seems that I need to export *all* variables to make it work such as: export HOST_SYS="${HOST_SYS}" in the configure Mar 28 17:34:48 s/configure/do_configure Mar 28 17:35:00 obviously I use do_configure_prepend(){ Mar 28 17:39:08 ah it worked with only a few variables...wow Mar 28 17:45:14 03Koen Kooi  07org.openembedded.dev * r308bbb0a55 10openembedded.git/recipes/vlc/x264_r2245.bb: Mar 28 17:45:14 x264: fix a few bugs Mar 28 17:45:14 * quoting EXTRA_OECONF = "something"$FOO"" is wrong Mar 28 17:45:14 * passing --enable-shared and then doing CFLAGS += -static is wrong as well Mar 28 17:45:15 * the beagleboard 'optimizations' were duplicating flags already there, and worse passed in O3 when O4 was the default Mar 28 17:45:17 * adding machine optimizations without changing package arch is wrong as well Mar 28 18:26:22 ant__: hello Andrea :) Mar 28 18:27:50 good morning Mar 28 18:29:02 morning Mar 28 18:29:10 hi hrw Mar 28 18:30:12 ~ugt Mar 28 18:30:13 i heard ugt is Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html Mar 28 18:34:45 hrw: about stable...well, my daily build from scratch for c7x0 is almost finished...one hour remaining (console, opie, gpe images). Sounds not too bad. Mar 28 18:35:02 no errors until now....fingers crossed Mar 28 18:37:29 btw, who will maintain stable branch for zauruses? :) Mar 28 18:38:06 Jay7: you? Mar 28 18:39:00 hrw: sorry, no time now/yet Mar 28 18:39:11 I have baby born one week ago :) Mar 28 18:39:45 * Jay7 points to ant__ ;) Mar 28 18:50:44 bye Mar 28 18:58:10 congratulations, Jay7 ! Mar 28 18:58:32 mwester: thanks :) Mar 28 18:58:41 damn, did I miss a birthday party? Mar 28 18:59:15 I was adding my disk a minute ago; hdparm -t /dev/sdc shows 120 MB/s, my two previous disks did 57 and 59 MB/s. Mar 28 18:59:22 Yes, Jay7 has one that has just celebrated their zeroth birthday. Mar 28 18:59:47 Jay7: congrats! all programmers start counting at 0 Mar 28 19:00:02 so it must be a double party for you :-) Mar 28 19:08:03 like2wise: thanks too :) Mar 28 19:08:15 * Jay7 have daughter :) Mar 28 19:10:35 It's changing my liferhythm.. Mar 28 19:10:36 so I have 'productivity hole' now Mar 28 19:21:13 Jay7: you will take better decisions now, because things become more relative. Mar 28 19:21:33 Jay7: but you probably won't be getting much sleep :-) Mar 28 19:23:09 that's right :) Mar 28 19:24:46 * Jay7 is looking into his linkedin profile and thinking.. Mar 28 19:25:56 hello,I've problems with boost...again...-DBOOST_SP_USE_PTHREADS doesn't seem to work because I have: undefined reference to `__sync_add_and_fetch_4' Mar 28 19:26:04 it's boost 1.36.0 Mar 28 19:38:56 maybe It's not related to that...but to another issue...I'll try something Mar 28 19:45:53 03Koen Kooi  07org.openembedded.dev * recf551ef71 10openembedded.git/ (3 files in 3 dirs): Mar 28 19:45:53 midori: update to 0.1.5 and workaround an issue with the abomination called 'waf' Mar 28 19:45:53 * runtime tested on beagleboard Mar 28 19:46:03 03Koen Kooi  07org.openembedded.dev * r4733329637 10openembedded.git/recipes/gtk-webcore/midori_git.bb: midori git: use git instead of snapshots Mar 28 19:52:15 Gnutoo: __sync_add_and_fetch_4 -> this needs to be in your GCC Mar 28 19:52:53 Gnutoo: support for those functions varies between different architectures Mar 28 19:53:12 if you compile for arm it is likely that you are missing those functions Mar 28 19:53:55 Gnutoo: to fix it you need to patch boost and provide an implementation Mar 28 19:54:37 Gnutoo: you can steal the code for example from a very recent GCC source Mar 28 19:54:37 rschuster1, yes but that was only one part of the error message Mar 28 19:54:47 rschuster1, wow good idea Mar 28 19:54:57 Gnutoo: well, this is what everyone does Mar 28 19:55:09 thanks a lot Mar 28 19:56:29 Gnutoo: the other isssue you see is probably the other well known atomic operation: compare and swap Mar 28 19:56:46 rschuster1, the other issue is that one: boost::exception_detail::exception_clone Mar 28 19:57:20 Gnutoo: http://pastebin.com/m3985d89e Mar 28 19:57:55 Gnutoo: that is a bunch of atomic operations as used in icedtea. they are itself from GCC (everythin is GPL) Mar 28 19:58:05 wow thanks a lot Mar 28 19:59:46 ( http://rafb.net/p/EchTVD91.html ) for the full error message Mar 28 20:00:26 Gnutoo: but it is always: >::release() const]+0x18): undefined reference to `__sync_add_and_fetch_4' Mar 28 20:00:36 yes Mar 28 20:00:49 so maybe I need both fixes Mar 28 20:01:35 Gnutoo: before you try fixing it by patching boost have a look at your GCC whether it really does not contain those Mar 28 20:01:35 the strange thing is that according to http://lists.boost.org/Archives/boost/2008/04/135597.php -DBOOST_SP_USE_PTHREADS should work Mar 28 20:01:45 ok thanks a lot Mar 28 20:01:54 wow.. tosa is donated in angstrom-users.. Mar 28 20:02:01 oops, -devel Mar 28 20:02:33 Gnutoo: it is possible that some C, CXX or LDFLAG has gone wild (or is just not correct) Mar 28 20:02:49 yes that's what I thought Mar 28 20:04:03 if you tried the workaround with BOOST_SP_USE_PTHREADS you should make sure the preprocessor definition was really available when the respective source file was compiled Mar 28 20:04:05 maybe I should (cross)compile boost with debug and capture the CPPFLAGS Mar 28 20:04:11 ok Mar 28 20:04:28 rschuster1, I didn't try it...it was already there Mar 28 20:04:54 the last time I looked at boost it looked to me that it is very easy to do something without achieving an actual effect Mar 28 20:04:59 ok Mar 28 20:05:37 you can also have a look at the generated binaries (shared objects) with readelf/objdump (from your tmp/cross/bin/... dir) Mar 28 20:05:48 lol Mar 28 20:05:59 I think I'dd do that Mar 28 20:06:27 Gnutoo: the tools should tell you that __sync_fetch_and_add_4 is an undefined reference Mar 28 20:06:32 ok Mar 28 20:06:42 thanks a lot... Mar 28 20:06:53 so I won't have to compile wesnoth to see Mar 28 20:06:56 good idea Mar 28 20:07:05 Gnutoo: I don't know which tool uses which syntax but it is easily comprehensible when you see it (there is a U .... function_name somewhere) Mar 28 20:07:11 Gnutoo: yes Mar 28 20:07:22 ok Mar 28 20:07:38 objdump/readelf basically do the same stuff btw :) Mar 28 20:08:40 Gnutoo: http://pastebin.com/m9c94fb9 Mar 28 20:09:09 Gnutoo: this is how they use the replacement function in openjdk. Mar 28 20:09:16 ok Mar 28 20:10:30 Gnutoo: as you see it uses the same parameters as the normal gcc function so you should not have trouble fixing boost Mar 28 20:10:42 Gnutoo: btw. big thanks for caring about wesnoth :) Mar 28 20:10:58 big thanks to you for helping me Mar 28 20:11:00 Gnutoo: can't wait to build it for my devices :) Mar 28 20:11:37 I want to buy a touchbook when it is available for europe Mar 28 20:11:46 yes that's a great game...specially when you don't have or don't want to use 3d(for instance if you have a SGX) Mar 28 20:12:34 and then playing wesnoth on this thing would be cool Mar 28 20:12:34 I think I'll buy something like this too(touchbook,pandora,beagleboard...) Mar 28 20:12:39 yes Mar 28 20:12:44 ~lart SGX Mar 28 20:12:44 * ibot puts on some milking gloves. "All right, now, SGX, this won't hurt a bit...." Mar 28 20:13:15 lol Mar 28 20:13:17 this reverse engineering effort sounded good Mar 28 20:13:21 yes Mar 28 20:13:28 but it does not look like they continued Mar 28 20:13:33 ah ok Mar 28 20:13:48 :( Mar 28 20:14:22 hopefully SGX will become so pervasive that some more people will stand up and utter their discontent Mar 28 20:15:22 and reverse-engeener it... Mar 28 20:15:27 maybe... Mar 28 20:16:44 rschuster1: hopefully the company will open up their drivers Mar 28 20:17:16 rschuster1: most atom based intel boards are sgx-based and lack good linux support as well... Mar 28 20:22:57 yeah, the poulsbo ... Mar 28 20:23:04 we have one in the office Mar 28 20:23:21 I mean, a development board, containing one Mar 28 20:23:37 there is no accelerated driver for this since a few kernel versions Mar 28 20:23:57 once upon a time canonical worked on a free driver even Mar 28 20:24:10 but that halted Mar 28 20:24:41 I was told that ubuntu wants to fix the situation in (karmic koala +1) Mar 28 20:24:48 thus 09.04 Mar 28 20:25:00 well, we'll see how it works out Mar 28 20:25:35 rschuster1: how can they fix it? it really all depends on a driver being provided in some open-enough form. Mar 28 20:27:13 rschuster1, wasn't it nm? Mar 28 20:28:32 anyway the symbol are stripped...I must tell oe not to strip them...i'll look if I can find something about it Mar 28 20:28:40 s/symbol/symbols/ Mar 28 20:29:45 Gnutoo: nm does not work on stripped files but objdump/readelf do Mar 28 20:29:59 ah ok wow ....thanks Mar 28 20:30:24 fyi, if you want to stop OpenEmbedded from stripping the files, see local.conf.sample, it has commented lines to do exactly that Mar 28 20:30:29 :P Mar 28 20:32:15 thanks keesj Mar 28 20:32:16 oops Mar 28 20:32:21 thanks kergoth Mar 28 20:32:23 np Mar 28 20:32:45 Gnutoo: Mar 28 20:32:45 file libjvm.so Mar 28 20:32:45 libjvm.so: ELF 32-bit LSB shared object, ARM, version 1 (SYSV), dynamically linked, stripped Mar 28 20:32:52 nm libjvm.so Mar 28 20:32:52 nm: libjvm.so: no symbols Mar 28 20:33:00 yes that's what I had Mar 28 20:33:07 I'm trying objdump now Mar 28 20:33:24 readelf -s libjvm.so | less Mar 28 20:33:24 Symbol table '.dynsym' contains 376 entries: Mar 28 20:33:24 Num: Value Size Type Bind Vis Ndx Name Mar 28 20:33:24 0: 00000000 0 NOTYPE LOCAL DEFAULT UND Mar 28 20:33:24 1: 00007874 0 SECTION LOCAL DEFAULT 9 Mar 28 20:33:25 2: 00117908 0 SECTION LOCAL DEFAULT 18 Mar 28 20:33:27 3: 00000000 0 NOTYPE WEAK DEFAULT UND data_start Mar 28 20:33:29 4: 00000000 132 FUNC GLOBAL DEFAULT UND open@GLIBC_2.4 (3) Mar 28 20:33:31 and thousands more Mar 28 20:33:45 with readelf the option you need is -s (symbols) Mar 28 20:34:01 undefined ones have "UND" before the name Mar 28 20:34:06 I think it was readelf Mar 28 20:34:24 with -a it also gives the symbol Mar 28 20:35:32 Gnutoo: sure, it means 'print all information' :) Mar 28 20:35:41 yes that's what the manual said Mar 28 20:36:00 mmm....UND __sync_add_and_fetch_4 Mar 28 20:36:12 :) Mar 28 20:37:03 now when you patch its use away and use the implementation I posted __sync_add_and_fetch wont even show up as a symbol: you won :) Mar 28 20:38:19 ok thanks a lot!!! Mar 28 20:38:59 btw. you get an introduction to all these tools in the book 'Embedded Linux primer' by some guy from Montavista Mar 28 20:39:34 ok I think I'll buy it or download it if it's freely available or both Mar 28 20:39:52 I really enjoyed that one. Mar 28 20:40:01 no its 'classically' licensed ;) Mar 28 20:40:44 ok Mar 28 20:41:54 Gnutoo: if you are interested in FOSS courseware on embedded linux check out http://free-electrons.com/ Mar 28 20:42:18 its all CC licensed AFAICT of good quality Mar 28 20:43:12 http://free-electrons.com/docs/openembedded/ :D Mar 28 20:45:15 florian: hey :) Mar 28 20:45:21 rschuster1, i've some slides from them Mar 28 20:47:29 hi all Mar 28 20:47:47 hi florian Mar 28 20:48:02 florian: seems ltg is fast enough now :) Mar 28 20:48:56 florian: what was main bottleneck? Mar 28 20:56:49 Jay7: Disk I/O obviously, moving the databases helped... Mar 28 20:57:08 hi Mar 28 20:57:53 woglinde: wb :) Mar 28 20:58:12 florian: cool :) Mar 28 21:20:22 what about adding slides into OE's contrib/presentations ? Mar 28 21:26:00 rschuster1: maybe a good idea Mar 28 21:34:32 Hey all. New OE fan here! :-) Mar 28 21:35:27 *g* Mar 28 21:35:29 welcome Mar 28 21:36:27 So far, this thing ROCKS. Mar 28 21:36:51 I ran into an issue with the ctorrent recipe though. Not messed with it enough to figure it out yet. Mar 28 21:39:04 SDuensin: explain your issue and show log on pastebin Mar 28 21:40:55 Sure thing. (I just looked in the bug db - didn't see anything there.) Mar 28 21:41:47 SDuensin: nice to hear Mar 28 21:42:14 http://jaeger.pastebin.com/m1fb8bee9 Mar 28 21:42:48 | NOTE: make DESTDIR=/opt/oe/build/tmp/work/armv5te-angstrom-linux-gnueabi/ctorrent-3.3.1-r0/image install Mar 28 21:42:48 | make: *** No rule to make target `install'.  Stop. Mar 28 21:42:48 | FATAL: oe_runmake failed Mar 28 21:43:02 means that you call 'make' but there is no makefile Mar 28 21:43:24 check if it that 'make' is called from within the correct directory Mar 28 21:43:38 Yea, but "I" is the recipe. Personally, I don't know enough yet to make a bb file. Mar 28 21:44:03 ? Mar 28 21:44:23 ... /opt/oe/openembedded/recipes/ctorrent/ctorrent_3.3.1.bb <- that is the recipe Mar 28 21:44:41 pretty sure, yea Mar 28 21:44:58 checking whether things have gone wrong can be done by looking into logfiles Mar 28 21:45:13 check: Mar 28 21:45:13 /opt/oe/build/tmp/work/armv5te-angstrom-linux-gnueabi/ctorrent-3.3.1-r0/temp Mar 28 21:45:27 Only log is that make error. Mar 28 21:45:53 Well, "Nothing to compile" and "Nothing to configure" are there. Mar 28 21:46:12 you can try bitbake devshell and use it to check build by hands Mar 28 21:46:56 but I think there are some troubles with configure or Makefile.. Mar 28 21:47:06 may be new autotools affect this Mar 28 21:47:15 It's thinking about devshell. Mar 28 21:47:48 03Robert Schuster  07org.openembedded.dev * r40639b1e81 10openembedded.git/contrib/presentations/README: Mar 28 21:47:48 - added contrib/presentations folder Mar 28 21:47:48 - added a README for presentations Mar 28 21:47:48 03Robert Schuster  07org.openembedded.dev * r39c3a28928 10openembedded.git/contrib/presentations/ (2 files): Mar 28 21:47:50 Using_OpenEmbedded_to_build_embedded_Linux_distribution_May2009_FreeElectrons.odp: Mar 28 21:47:52 - new presentation Mar 28 21:47:54 - updated variant of http://free-electrons.com/doc/openembedded.odp Mar 28 21:47:55 SDuensin: ok same issue here Mar 28 21:47:56 cross-compilation_with_OpenEmbedded_-_LinuxTag2008_-_Robert_Schuster.odp: Mar 28 21:47:58 - new presentation from me in 2008 Mar 28 21:48:13 rschuster1 - That makes me feel a bit better. Mar 28 21:48:29 ah ok Mar 28 21:48:56 source dir is different from what the recipe expects Mar 28 21:49:14 That was my guess. Just don't know enough to fix it. Mar 28 21:49:54 the variable that controls this is named S Mar 28 21:51:04 many packages out there use the name of their tarball without the tar.bz2/tar.gz extension Mar 28 21:51:25 that is why OE sets S by default to the package name + version Mar 28 21:52:29 btw. why is -c rebuild not doing a 'clean' immediately anymore? Mar 28 21:52:45 So is S a full path or just "ctorrent_3.3.1"? Mar 28 21:52:53 S = "${WORKDIR}/${PN}-dnh${PV}" Mar 28 21:53:04 put this into your ctorrent_3.3.1.bb Mar 28 21:53:08 * SDuensin has no idea why anything does anything yet. Just found this cool tool last night. :-) Mar 28 21:53:31 SDuensin: beware this cool tool bites Mar 28 21:53:44 SDuensin: everyone knows that OE has a hard learning curve Mar 28 21:53:59 What doesn't these days? :-) Mar 28 21:54:18 SDuensin: but if you are going to do embedded development for some time (or different devices over the years) you will learn to love it ;) Mar 28 21:54:54 SDuensin: well ... Mar 28 21:55:09 SDuensin: tried extending LLVM a few days ago ... and failed Mar 28 21:55:36 LLVM is on my "to play with" list, too. Mar 28 21:56:26 SDuensin: if you grok the tablegen and DAG stuff feel free to write a cool tutorial :) Mar 28 21:56:58 I added the S and it still hates me. Mar 28 21:57:26 Are you an OE contributor? Should I put this in bugzilla? Mar 28 21:57:40 SDuensin: yeah, this is because OE does not know that you modified the recipe and starts over from scratch Mar 28 21:57:52 SDuensin: issues bitbake -c clean ctorrent Mar 28 21:57:58 then bitbake ctorrent Mar 28 21:58:07 then your change is in effect Mar 28 21:58:40 Excellent. Much thanks! Mar 28 21:58:50 if we fix something in a recipe that has worked before we increase the so called PR (package revision). in this case your build would have been restarted automatically Mar 28 21:59:09 yes, I am contributor and this fix will be in shortly Mar 28 21:59:13 Makes sense. Mar 28 21:59:21 Ok, I won't file it then. Thanks! Mar 28 21:59:35 Awesome tool, helpful people - damn weird! :-P Mar 28 21:59:44 :) Mar 28 22:00:52 03Robert Schuster  07org.openembedded.dev * r67deceb1c4 10openembedded.git/recipes/ctorrent/ctorrent_3.3.1.bb: Mar 28 22:00:52 /me wonders how could that have been slipped in ... Mar 28 22:00:52 ctorrent 3.3.1: Fixed S. Mar 28 22:01:07 * SDuensin cheers! Mar 28 22:01:22 SDuensin: on some days, we are unhelpful and also break the tool. Just to keep our FOSS appearance. Mar 28 22:01:28 hehe Mar 28 22:01:29 and again the world has become a little better ;) Mar 28 22:01:45 SDuensin: yeah, that happens also Mar 28 22:02:30 SDuensin: however if I am not totally under full workload I try to help where possible Mar 28 22:02:50 Sounds like me. Except in this case, I don't know crap yet. :-D Mar 28 22:02:56 SDuensin: and you seemed so enthuasiatic that I did not want you become annoyed so quickly ;) Mar 28 22:03:02 hehehe Mar 28 22:03:07 * SDuensin will remember to stay enthused! Mar 28 22:03:34 omg.. busybox developers are insane? :) Mar 28 22:03:38 * Jay7 reading ML Mar 28 22:03:42 I had been working all week on cross compiling things. Guy last night asked me WTF I was doing and gave me the OE URL. Been happy ever since! Mar 28 22:03:58 jay7 whats the problem? Mar 28 22:04:07 SDuensin: I was introduced to OE three years ago Mar 28 22:04:27 minimal-image does not respect ANGSTROM_EXTRA_INSTALL. I'm not sure what to do with that. Keep it minimal? :-) Mar 28 22:04:40 woglinde: http://lists.linuxtogo.org/pipermail/openembedded-devel/2009-March/009225.html Mar 28 22:04:48 SDuensin: the first year I was fighting with everything Mar 28 22:05:00 that is not first report about busybox 'instability' Mar 28 22:05:20 SDuensin: today I am mostly maintaining the Java stuff ... Mar 28 22:05:51 SDuensin: ... where it worries me that I am probably the one person knowing how it sticked together :$ Mar 28 22:05:52 I like Java, too. Just can't wrap my head around using it on an embedded device yet. Mar 28 22:06:09 SDuensin: do you have a sheevaplug already? Mar 28 22:06:21 I wish. Mar 28 22:06:38 Lining up my ducks for when I do. Mar 28 22:06:57 SDuensin: well, this gives you a little time to understand how you can set up the Jalimo overlay Mar 28 22:07:18 SDuensin: it contains some recipes which eventually all go into OE but are not ready yet Mar 28 22:07:21 The WHAT? Weren't we using English a minute ago? Mar 28 22:07:57 SDuensin: Robert is already talking marketing buzz to you :-) Mar 28 22:08:12 SDuensin: it contains the stuff to build OpenJDK Mar 28 22:08:26 I am just interested to see someone building OPenJDK for a Sheevaplug Mar 28 22:08:42 Ah. I looked briefly at the Java page on the wiki. Didn't really get to read any of it. Mar 28 22:08:44 its armv5. I dont have these ... :) Mar 28 22:10:18 and then I already have too much gadgets here and can't dare to get another Mar 28 22:10:44 Now for the challenging one... Ever hear of "Kepler"? Mar 28 22:10:55 * SDuensin wants more toys, but he's in the same boat. Mar 28 22:11:05 My garage and basement are a gaming museum. Mar 28 22:11:28 * Jay7 have only akita and nokia phone :)) Mar 28 22:12:12 rschuster1: I can build it for sheeva soon Mar 28 22:12:33 but there is around about 4 VIA mini-itx boards.. Mar 28 22:12:36 rschuster1: we are also creating sheeva derived hardware Mar 28 22:12:42 freerunner, n770, beagleboard: mine Mar 28 22:12:42 buglabs bug, n800: work Mar 28 22:12:54 like2wise: sweet. can you already tell what it is going to be? Mar 28 22:13:05 BeagleBoard looks fun. Mar 28 22:13:22 like2wise: should I wait with my order for a Touch Book ? Mar 28 22:13:30 rschuster1: no Mar 28 22:13:36 SDuensin: BB is the dogz ;) Mar 28 22:13:46 it will be a more open development board. Mar 28 22:14:14 SDuensin: really with this piece of hardware I have done so many embedded development things like with nothing else Mar 28 22:14:39 SDuensin: it is my primary test system for the openjdk stuff Mar 28 22:14:53 My warped brain saw "OpenGL ES" and went "VIDEO GAME!" Mar 28 22:15:18 SDSuen: here more marketing buzz ;) http://rschuster.blogs.evolvis.org/2008/12/21/serving-cross-compiled-openjdk-with-icedtea/ Mar 28 22:15:26 SDuensin: our warped brains say "bad driver support" Mar 28 22:16:06 * rschuster1 s built-in vrms cries out loudly when hearing SGX Mar 28 22:16:14 -> http://rschuster.blogs.evolvis.org/2009/03/27/fedora-giving-foss-graphics-driver-nouveau-a-try/ Mar 28 22:16:24 btw. we need a planet.openembedded.org Mar 28 22:16:38 planet.linuxtogo.org Mar 28 22:16:39 I just saw cbrake's blog today Mar 28 22:16:53 is planet.openembedded.org Mar 28 22:17:07 woglinde: ahh Mar 28 22:17:13 Hey, if you ever get insanely bored, crank out a recipe for the Einstein emulator. I really want that on my N800. Mar 28 22:17:17 this is where all the cool guys are. :) Mar 28 22:17:30 so whom to ask to syndicate my blog? Mar 28 22:17:42 rschuster1: hey, the cacao screen says SWT - is that true? Mar 28 22:18:03 What are those screenshots running on? Mar 28 22:18:29 like2wise: http://rschuster.blogs.evolvis.org/files/2008/12/2-cacao-openjdk.png ? Mar 28 22:18:45 like2wise: sure, openjdk+cacao running an SWT example Mar 28 22:18:54 SDuensin: its a neo freerunner Mar 28 22:19:01 SDuensin: the one on my desk Mar 28 22:19:10 wow.. Mar 28 22:19:20 * Jay7 have subscribed to planet rss :) Mar 28 22:19:36 rschuster1: it uses the swt libraries (swt is from Eclipse, right??) Mar 28 22:19:48 like2wise yes Mar 28 22:19:51 Cool. Mar 28 22:19:51 like2wise: yes Mar 28 22:19:59 rschuster1: nice Mar 28 22:20:11 like2wise: I can even take debian's azureus and run it on the device :) Mar 28 22:20:26 like2wise: i've taken this for my presentation at FOSDEM Mar 28 22:20:31 Sweet! Mar 28 22:20:53 rschuster1: damn, I was manning the OE booth then Mar 28 22:22:00 like2wise: slides if you are interested: http://wiki.debian.org/Java/DevJam/2009/Fosdem?action=AttachFile&do=view&target=cross+compiling+OpenJDK+with+OpenEmbedded+-+FOSDEM2009+-+Robert+Schuster.pdf Mar 28 22:27:55 * Jay7 -> sleep Mar 28 22:32:29 SDuensin: regarding the Einstein emulator Mar 28 22:32:39 Yes? Mar 28 22:33:54 SDuensin: omg ... Mar 28 22:34:06 Yes? Yes? Mar 28 22:34:13 SDuensin: I thought it would be a nice simple makefile based thing but it is actually using jam Mar 28 22:34:33 It's a mess. I've tried a few times to build it for my N800. No dice. Mar 28 22:34:35 SDuensin: so it might be too difficult Mar 28 22:34:51 SDuensin: I wanted to propose that you could do this as a starter Mar 28 22:35:01 There's where the "insanely bored" part comes in. :-P Mar 28 22:35:10 SDuensin: but messing around with uncommon build system in the beginning is a bit too tough Mar 28 22:36:15 SDuensin: feel free to put a request in the bugzilla. but I have to warn you that the bugtracker isn't maintained very well Mar 28 22:36:36 Bugzilla probably scares 'em off. Mar 28 22:36:52 SDuensin: we are just to few people and most of us work on specific things Mar 28 22:38:33 uh oh Mar 28 22:38:44 this Einstein uses a library named K Mar 28 22:38:57 and this one isn't maintained since 2004 ... Mar 28 22:39:11 expect fun with GCC 4.x on this :) Mar 28 22:40:28 SDuensin: however if you ever want to bring some software into OE that is not already there chose one that has a simple build system and is preferably already packaged for debian or ubuntu Mar 28 22:40:49 then get the source package for debian/ubuntu and see how they do it Mar 28 22:43:31 I'll let you know what I find. :-) Mar 28 22:44:08 03Michael 'Mickey' Lauer  07org.openembedded.dev * rf917c733f3 10openembedded.git/recipes/linux/ (linux-msm7xxxx/htcraphael/defconfig linux-msm7xxxx_git.bb): linux-msm7xxx: add more useful options to htcraphael defconfig Mar 28 22:44:19 03Michael 'Mickey' Lauer  07org.openembedded.dev * r3b49064805 10openembedded.git/recipes/zope/zope-3.3.1.inc: zope-3.3.1.inc: -> zope.inc Mar 28 22:46:56 anyone know what the 'J' in arm6TEJ stands for? Mar 28 22:47:02 jazelle Mar 28 22:47:15 java extension Mar 28 22:51:23 ah, thanks Mar 28 22:51:27 hmm, nice Mar 28 22:52:03 its NDA so it isnt nice at all Mar 28 22:54:21 ah, bummer Mar 28 22:56:34 woglinde: ^_^ Mar 28 22:57:13 pheww..long logs today... Mar 28 23:01:17 mickeyl: about Bugzilla...can we close old 'won't compile/won't work' bugs if we have evidence in tinderbox the issue is solved and if there are no confirmations? Mar 28 23:02:19 yes Mar 28 23:04:12 ok Mar 28 23:05:02 ant__, good idea Mar 28 23:05:04 mickeyl: bigger issue...I have a list of recipes using '-s' to strip during install Mar 28 23:05:16 this breacks dbg pckgs... Mar 28 23:05:28 -c Mar 28 23:05:42 our coreutils is patched to ignore -s Mar 28 23:05:49 at least on of the versions Mar 28 23:05:52 one Mar 28 23:06:14 hm...there was a patch from khem after autotools bump Mar 28 23:06:24 'strip' was not found Mar 28 23:06:50 afterward I think the -s is NOT ignored...probably I'm wrong... Mar 28 23:08:40 mickeyl: didn't know that, but indeed, we have been patching a few packages/recipes lately to remove the -s.... Mar 28 23:09:12 ya, i guess we should rather use this coreutils patch than removing the -s Mar 28 23:09:16 looks like less work to me Mar 28 23:10:02 this was the list couple of weeks ago: Mar 28 23:10:02 agreed Mar 28 23:10:20 mom..file is too big for pastebin... Mar 28 23:10:23 I'm pushing to commits out, but get this: error: failed to push some refs to 'git@git.openembedded.net:openembedded' Mar 28 23:10:27 s/to/two/ Mar 28 23:10:55 sounds like you need to pull --rebase Mar 28 23:11:07 someone has pushed in the meantime Mar 28 23:11:41 I did already do a pull --rebase, let me redo Mar 28 23:12:20 03Leon Woestenberg  07org.openembedded.dev * re5df9d634b 10openembedded.git/conf/checksums.ini: Mar 28 23:12:20 yasm: New recipe. x86 assembler. Mar 28 23:12:20 This assembler is used (as yasm-native) by the x264 build for Mar 28 23:12:20 x86 targets. Other packages may want this for similar reasons. Mar 28 23:12:21 03Leon Woestenberg  07org.openembedded.dev * r0ee45495fb 10openembedded.git/recipes/vlc/x264_r2245.bb: Mar 28 23:12:25 x264: Enable using yasm-native for x86 machine. Mar 28 23:12:27 Results in using optimized assembler functions instead of Mar 28 23:12:29 their C equivalent. The x264 encoder has many x86/SSE Mar 28 23:12:31 optimizations, this will increase its performance much. Mar 28 23:12:33 03Leon Woestenberg  07org.openembedded.dev * r7b383197a2 10openembedded.git/recipes/yasm/ (yasm-native_0.7.2.bb yasm.inc yasm_0.7.2.bb): yasm: New recipe. Mar 28 23:13:29 hm reminds me to put in the video accel patches for ffmpeg and libva self Mar 28 23:13:51 mickeyl: strange, I had a local change to minimal-image, which is not in any commit. I had to revert it (which is "checkout" in git) then pull --rebase first. Mar 28 23:14:13 I'm still not grasping git... need to re-read the git community book Mar 28 23:14:21 you can't pull --rebase if your tree is dirty Mar 28 23:14:23 i always do Mar 28 23:14:24 git stash Mar 28 23:14:26 git pull --rebase Mar 28 23:14:28 git stash apply Mar 28 23:14:33 ah thanks Mar 28 23:14:33 git stash rocks Mar 28 23:15:02 does stash stack, like quilt does? Mar 28 23:15:10 so..this was just grepping for 'INSTALL' and '-s' Mar 28 23:15:12 http://pastebin.ca/1375222 Mar 28 23:15:25 heh Mar 28 23:15:31 never tried that Mar 28 23:15:35 you can name stashes Mar 28 23:16:23 I like this cool option editing commits which missed some files Mar 28 23:16:25 * mwester has a number of stashes... Mar 28 23:16:28 it seems you can: http://book.git-scm.com/4_stashing.html Mar 28 23:16:29 :-O Mar 28 23:16:35 food stash? Mar 28 23:17:06 coffee, spare disk drives, and a few git stashes. Mar 28 23:17:14 the fuzz took my stash Mar 28 23:17:26 duuuude!!! Mar 28 23:18:10 That reminds me of "Duuude, I'm like 99% percent parshed out here, can someone fetch me some cooolaa?" Mar 28 23:21:44 whahah Mar 28 23:21:55 whats uo with the www server again Mar 28 23:22:02 ah Mar 28 23:22:27 seems to have sime hickups Mar 28 23:25:16 btw, do you receive mails from openembedded-commits? Mar 28 23:25:33 anymore here Mar 28 23:26:09 yes Mar 28 23:27:15 hm...I got last one in february Mar 28 23:27:18 Matches: to:(openembedded-commits) Mar 28 23:27:18 Do this: Skip Inbox, Apply label "openembedded-commits" Mar 28 23:27:31 never changed Mar 28 23:30:14 doh..I subscribed again Mar 28 23:30:46 hm Mar 28 23:31:01 I have 2 branches of org.openembedded.dev Mar 28 23:31:14 and I only want to push one of them Mar 28 23:31:20 is git push origin org.openembedded.dev Mar 28 23:31:25 the right command? Mar 28 23:31:41 within the branch to be pushed? Mar 28 23:32:14 should be git push origin your-local-branch:org.openembedded.dev Mar 28 23:32:33 okay Mar 28 23:32:36 will try this Mar 28 23:32:50 theur name differ, isn't ? Mar 28 23:32:52 and then point to ant in the case breakse Mar 28 23:33:08 I'm insured ;-) Mar 28 23:35:29 03woglinde  07org.openembedded.dev * r51d9060d45 10openembedded.git/recipes/taglib/ (4 files in 2 dirs): taglib: update to version 1.5 and using cmake for build Mar 28 23:35:30 03woglinde  07org.openembedded.dev * rdc54d7a733 10openembedded.git/recipes/libogg/libtheora_1.0.bb: Mar 28 23:35:30 libtheora: don't build the examples Mar 28 23:35:30 * zaps the dependency on sdl Mar 28 23:35:30 * bump PR Mar 28 23:35:33 03woglinde  07org.openembedded.dev * ra0498ecd1e 10openembedded.git/classes/scons.bbclass: scons: honor PARALLEL_MAKE Mar 28 23:35:36 03woglinde  07org.openembedded.dev * rfd4f6dc10e 10openembedded.git/recipes/linux/linux-2.6.28/wrap/defconfig: wrap: update defconfig, dont use input Mar 28 23:36:05 okay Mar 28 23:36:08 thanks ant Mar 28 23:36:10 pheww Mar 28 23:36:34 I was joking..I always use local branches :-) Mar 28 23:36:40 commands are tested Mar 28 23:36:55 ??? Mar 28 23:37:06 s/should be/ IS Mar 28 23:37:51 btw do you checkout in a custom dir? Mar 28 23:38:13 ???? Mar 28 23:38:20 03Michael 'Mickey' Lauer  07org.openembedded.dev * r8f8efbe43a 10openembedded.git/conf/distro/include/sane-srcrevs.inc: sane-srcrevs: bump EFL and add known-good for vala-native-git Mar 28 23:38:35 it checkout -b your-local-branch origin/org.openembedded.dev Mar 28 23:38:55 yes Mar 28 23:39:01 as the git page says Mar 28 23:39:24 what it did not say is you can clone in a specific dir Mar 28 23:39:35 r.g. #git clone git@git.openembedded.net:openembedded /oe/org.openembedded.dev Mar 28 23:41:01 hm I dont need this Mar 28 23:41:24 I'm lazy..this was the old monotone tree Mar 29 00:01:53 hm..digging in Bugzilla I found this ..http://bugs.openembedded.net/show_bug.cgi?id=5002 Mar 29 00:02:13 patch for ' Graphviz deps tree generation broken ' Mar 29 00:05:41 ant__: trying that patch Mar 29 00:06:13 thx, is touching a core file... Mar 29 00:06:23 I don't dare Mar 29 00:06:28 :-) Mar 29 00:07:12 well..after all is a print... Mar 29 00:12:51 we had such nice graphs before.... when I also had this patch: http://bugs.openembedded.net/show_bug.cgi?id=1488 Mar 29 00:17:44 * florian just noticed that its great that using a digital camera is dirt cheap - luckily since his son found out how to make photos Mar 29 00:18:49 likewise: here patch seems fine Mar 29 00:19:51 now is grep-bonanza again Mar 29 00:20:56 florian: CHDK ? Mar 29 00:22:02 ok..baby screaming...have to go Mar 29 00:22:05 'nite Mar 29 00:22:49 ant__: CHDK? Mar 29 00:22:58 ant__: enjoy ;) Mar 29 00:26:08 good nite Mar 29 01:41:41 Crofton|work: ping **** ENDING LOGGING AT Sun Mar 29 02:59:57 2009