**** BEGIN LOGGING AT Sun Aug 09 02:59:57 2009 Aug 09 03:03:26 zecke: course, for this class to work, I'd have to make changes to the native/cross classes in addition to base.. hacked it locally just to test m4-native->shasum-native Aug 09 03:03:28 heh Aug 09 03:04:13 okay, i think we need to make sure bitbake starts setting FILE to the *abs_fn* rather than *fn* Aug 09 03:04:20 I'm starting to get tired of having to bb.which() manually Aug 09 03:04:25 heh Aug 09 03:04:36 though, i guess i could see the relative one being useful for something Aug 09 03:19:32 03Khem Raj  07org.openembedded.dev * rcf58571e78 10openembedded.git/recipes/linux/ (2 files in 2 dirs): Aug 09 03:19:32 linux-titan-sh4-2.6.21: Provide sidivsi3_i4i and udivsi3_i4i Aug 09 03:19:32 Signed-off-by: Khem Raj Aug 09 03:19:42 03Khem Raj  07org.openembedded.dev * r9f6ae87076 10openembedded.git/recipes/lilo-sh/ (files/lilo-replace-PAGE_SIZE.patch lilo-sh_21.bb): Aug 09 03:19:42 lilo-sh_21.bb: Replace PAGE_SIZE Aug 09 03:19:42 PAGE_SIZE is not available to usersplace anymore Aug 09 03:19:42 struct partion in linux/genhd.h is defined conditional Aug 09 03:19:44 if CONFIG_BLOCK is defined. Aug 09 03:19:46 Signed-off-by: Khem Raj Aug 09 03:23:15 shit Aug 09 03:23:18 hrmph Aug 09 03:30:19 ah, there we are Aug 09 03:35:45 anyone know what the reasoning behind e1464b0c0fa9ca60be0a1209f32978639f633c8d was? Aug 09 03:41:46 kergoth: hehe Aug 09 03:42:39 * zecke wondered if an INSTALL="gst-plugins-mad" bitbake custom-image would be of any use Aug 09 03:43:48 huh, i didn't know about the cleandirs flag, cute. /me uses it.. Aug 09 03:44:56 kergoth: what flag are you referring to and where ? Aug 09 03:45:11 cleandirs, all dirs listed there get wiped clean and recreated before the task runs Aug 09 03:45:18 do_install has it set to ${D} to make sure its starting fresh Aug 09 03:45:38 * kergoth was manually clearing ${D} before he realized it was unnecessary because of that flag :) Aug 09 03:45:54 nice :-) Aug 09 03:46:33 zecke: so does what I'm doing here make sense to you, or am i completely crazy? :) Aug 09 03:46:51 * kergoth thinks its slick that all the sbindir, bindir, includedir, etc overrides in native/cross .bbclass can just go away Aug 09 03:49:57 kergoth: I think it makes sense, no way that some other PN-PV-PR sneaks in into the build (if I got this right) Aug 09 03:50:04 yeah, exactly Aug 09 03:50:09 if i dep on ncurses-5.4, thats exactly what i link against Aug 09 03:50:13 should be more deterministic Aug 09 03:50:27 kergoth: do_stage() will kind of die, there is the case where do_stage adds some extra scripts (I think of qmake foo) Aug 09 03:50:43 yeah, I'm planning on adding a var with callbacks for that. Aug 09 03:50:55 .la mangling, .pc stuff, extra scripts like you said, .. Aug 09 03:51:02 binconfig mangling, perhaps Aug 09 03:51:09 kergoth: sounds great Aug 09 03:51:22 OT: anyone toyed with lsb? is /ld-lsb.so always installed nowadays? Aug 09 03:51:39 * kergoth just needs a sounding board occasionally to make sure he's not doing something completely wacky :) Aug 09 03:51:51 * kergoth has happily avoided learning anything about lsb so far Aug 09 03:55:58 kergoth: could you check if your distro has /lib/ld-lsb.so.* installed? Aug 09 03:56:15 this debian machine doesn't Aug 09 03:56:36 * kergoth checks ubuntu karmic .. not there either Aug 09 03:56:47 kergoth: *sigh* Aug 09 03:56:58 course, debian/ubuntu split out their lsb stuff into a ton of packages Aug 09 03:57:04 maybe its buried in one Aug 09 03:57:06 heh Aug 09 03:58:12 lsb-base came up the other day for SlugOS folks. Aug 09 03:58:43 * zecke tries to built a lsb app... deployment is pretty shit when there is no easy install :} Aug 09 03:58:56 ushare has a dependency in its init script on lsb-base. Aug 09 03:59:01 we got bit by some /lib/lsb/init-functions or whatever last week Aug 09 03:59:04 mwester: heh :) Aug 09 04:00:28 an OE overlay can be to any level of completeness right ? Just to confirm this,I can have one with just package/image recipes and expect bitbake to pick up the machine confs etc from my original OE tree, right ? Aug 09 04:00:37 init-functions looks pretty basic; we should be able to get that done easily. We'd need a busybox version, and probably another for those distros with a bit more robust set of initial packages in their rootfs, I suspect. Aug 09 04:01:17 rkirti: BBPATH is how it finds conf/classes, BBFILES is how it finds recipes. if you use COLLECTIONS, it sets both to include them all Aug 09 04:01:24 short answer: yes, if you have it configured correctly Aug 09 04:02:00 is 4.4.3 smaller than 4.4.3a according to opkg? Aug 09 04:02:17 * kergoth thinks we should make bitbake function in an overlay/collections based way directly, instead of using BBPATH/BBFILES, someday, maybe Aug 09 04:02:28 i think so Aug 09 04:03:11 kergoth: naive question - how do overlays and collections differ ? Aug 09 04:03:16 same idea Aug 09 04:03:24 just differing names Aug 09 04:03:44 ok Aug 09 04:03:45 internally, bitbake supports priorities in overlays/collections via BBFILE_COLLECTIONS, so .. :) Aug 09 04:04:00 but some people like to think of them as overlays, since you sort of pile them on top one another Aug 09 04:04:09 so whatever you like to call it.. Aug 09 04:04:36 * rkirti reaches out for the bitbake manual, since the OE one doesn't have this stuff.. Aug 09 04:04:51 theres an example of BBFILE_COLLECTIONS usage in the bitbake manual Aug 09 04:05:18 the OE manual is great, but there are sections like this with the "please complete this" tag that stop it from being as awesome as it could Aug 09 04:05:20 if you use COLLECTIONS, it sets up BBFILE_COLLECTIONS for you, too Aug 09 04:05:33 kergoth: ok, thanks :-) Aug 09 04:05:37 * kergoth likes it, but is probably biased since he wrote conf/collections.inc for montavista Aug 09 04:05:39 :P Aug 09 04:06:22 export COLLECTIONS = "/some/path /another/path" .. its in priority order, highest to lowest Aug 09 04:06:31 * kergoth grumbles at this class Aug 09 04:08:17 wow, things explode at parse time if you run a distro that doesn't INHERIT a packaging class, due to the call of runtime_mapping_rename in image.bbclass Aug 09 04:09:37 gah, i keep finding more and more things i need to tweak for the staging changes. lots of dependencies on the old way of doing it Aug 09 04:09:42 * kergoth overrides another class or 12 Aug 09 04:18:29 kergoth: May I ask what are you trying to do with staging ? Aug 09 04:19:37 * rkirti is reminded of the crib on packaged staging she had and jumps to uservoice Aug 09 04:27:34 rkirti: I'm reimplementing it from scratch. see the first paragraph after the summary at http://gist.github.com/164598 Aug 09 04:27:35 stupid question... how to make patchwork change the state of a patch? Aug 09 04:27:58 * kergoth has no idea Aug 09 04:28:45 rkirti: each recipe gets its own staging area with only the bits from the recipes it depends on Aug 09 04:29:47 * rkirti grumbles,wondering why rosetta should be built each time an image is done. (on a daily basis rather) Aug 09 04:31:11 course, this stuff is going to be a royal pain in the ass to get upstream to oe, since it requires a bitbake change. think i'll try to keep it relatively independent for the time being Aug 09 04:32:15 hrmm Aug 09 04:32:20 * kergoth wanders off for a while Aug 09 04:32:27 kergoth: a lot of work indeed. But why do we need this ? Aug 09 04:32:38 make builds more deterministic Aug 09 04:32:50 right now things can get rather ugly if multiple versions of something are built in one run, for example Aug 09 04:33:06 staging will upgrade, but depending on which recipe links against it when, things can go bad Aug 09 04:33:10 just one example Aug 09 04:33:29 kergoth: ok, I am being stupid here, but is this like - right now there is a staging areaand what you do is pick up all that you need for that recipe from the staging area ? Aug 09 04:33:56 yep. its one big pile. its managed by package manager, which helps, but its still one big area for everything from every recipe for things that depend on them Aug 09 04:34:02 kergoth: yep. I have had infinite problems with wrong staging stuff being used due to version decrements and stuff Aug 09 04:34:08 03Holger Hans Peter Freyther  07org.openembedded.dev * re513208d78 10openembedded.git/recipes/meta/meta-toolchain-qte.bb: Aug 09 04:34:08 meta-toolchain-qte: Export some more paths Aug 09 04:34:08 Export RCC, UIC3, DBUSXML2CPP, DBUSCPP2XML environment variables. This Aug 09 04:34:08 is based on a patch from Ihar Hrachyshka. Aug 09 04:34:10 Signed-off-by: Ihar Hrachyshka Aug 09 04:34:10 right, this should take care of all of that Aug 09 04:34:18 03Holger Hans Peter Freyther  07org.openembedded.dev * rebacd72f73 10openembedded.git/recipes/qt4/qt4-tools-sdk_4.5.2.bb: qt4-tools-sdk: Use an install target to install the qt4 utilities Aug 09 04:35:19 hmm, i wonder if we could leverage something like what I'm doing with the cooker here to make the image builds a bit smarter. make it pull the *exact* package from the preferred built recipes as of *now*, so old packages, etc would be a non-issue Aug 09 04:35:26 could kill the global pile of ipks Aug 09 04:35:41 store a fn->ipks mapping in the sql db with persist_data Aug 09 04:36:11 * kergoth really wanders off this time Aug 09 04:36:28 wow. that would be awesome Aug 09 04:36:49 but how do you define/specify "exact package" ? :- Aug 09 04:36:56 :-) Aug 09 04:37:30 the way this works is it asks bitbake, given the current user preferences, what recipe filename is used in *this* build to satisfy my dependencies. Aug 09 04:37:38 itd work the same way for an image Aug 09 04:37:51 identify the files satisfying its deps & rdeps, map that to ipks, go Aug 09 04:56:37 anyone could do me a favor and try a static binary? Aug 09 05:03:12 has anyone seen a "ValueError: too many variables to unpack" when building glibc-2.9? Aug 09 05:06:05 kg4ysn: didn't try, but you should get a backtrace of some such... Aug 09 05:06:18 kg4ysn: or actually, bitbake tells you which task failed, then you can start looking from there Aug 09 05:06:55 zecke: Aug 09 05:06:57 NOTE: Running task 665 of 2238 (ID: 172, /root/workspace/openembedded/recipes/glibc/glibc_2.9.bb, do_package) Aug 09 05:06:57 ERROR: Error in executing: /root/workspace/openembedded/recipes/glibc/glibc_2.9.bb Aug 09 05:06:57 ERROR: Exception: Message:too many values to unpack Aug 09 05:06:58 ERROR: Printing the environment of the function Aug 09 05:06:58 ERROR: 0058: utf8_only = int(bb.data.getVar('LOCALE_UTF8_ONLY', d, 1) or 0) Aug 09 05:06:58 ERROR: 0059: encodings = {} Aug 09 05:07:00 ERROR: 0060: for l in supported: Aug 09 05:07:02 ERROR: 0061: l = l[:-1] Aug 09 05:07:04 ERROR: 0062: (locale, charset) = l.split(" ") Aug 09 05:07:06 ERROR: 0063: if utf8_only and charset != 'UTF-8': Aug 09 05:07:08 ERROR: 0064: continue Aug 09 05:07:10 ERROR: 0065: m = dot_re.match(locale) Aug 09 05:07:12 ERROR: 0066: if m: Aug 09 05:07:16 that's the backtrace Aug 09 05:08:05 zecke: python also kicked out a bunch of traceback information ... files/lines/etc. i can send that if better. Aug 09 05:08:22 kg4ysn: thanks, use paste.lisp.org/new/oe or such in the future.. Aug 09 05:08:37 kg4ysn: so which line do you think is wrong Aug 09 05:08:53 zecke: didn't know about that, thanks. Aug 09 05:09:10 zecke: not sure. all i know is that bitbake is tossing off an error, preventing my build from completing. Aug 09 05:09:32 zecke: ENABLE_LOCALES is set to "0" Aug 09 05:10:49 kg4ysn: yes, the key is called ENABLE_BINARY_LOCALES or such... Aug 09 05:11:08 kg4ysn: well, there are not too many "unpacks" in the above code... what about printing "l" and see how it loiks Aug 09 05:11:12 kg4ysn pasted "traceback output" at http://paste.lisp.org/display/85079 Aug 09 05:13:09 zecke: I had ENABLE_BINARY_LOCALES set to "0", but still had GLIBC_GENERATE_LOCALES set to "en.US.UTF-8". When I commented out the GLIBC_GENERATE_LOCALES field, all seems to be happy. are the two not related? Aug 09 05:20:11 yeah, that's the damnedest thing ... ENABLE_BINARY_LOCALES="0" and GLIBC_GENERATE_LOCALES="en.US.UTF-8" causes an error. is it an error worthy of a bug report? Aug 09 05:20:16 or is it expected behavior? Aug 09 05:20:32 kg4ysn: no idea, but if there is a problem maybe you should help to fix it? Aug 09 05:21:54 zecke: thanks, i often contribute to the kernel as well as other FOSS projects. in this case, however, it's a little out of my league to get up to speed on the code surrounding it. my contribution will be to report the condition, starting the chain of events. :-) Aug 09 05:22:06 zecke: appreciate the help! and the pointer to the lisp.org paste util ... didn't know about it, very neat. Aug 09 05:22:46 kg4ysn: yeah reporting is a good start, any kind of such errors should be there (e.g. non compatible options should generate a warning by sanity.bbclass or such) Aug 09 05:23:11 kg4ysn: I would be interested in what "supported" holds, and then "l" Aug 09 05:23:35 zecke: might need to prompt me more on how to get that info ... not following Aug 09 05:26:50 kg4ysn: it is python code so "print supported" and "print l" (with the right indention) in glibc-package.inc... let me check the filename Aug 09 05:27:34 kg4ysn: glibc-package.bbclass in the recipes/glibc directory Aug 09 05:28:39 kg4ysn: basically the l.split(" ") fails as there is no " " in 'l'... but I have no idea why Aug 09 05:29:05 * * OE Bug 5292 has been created by chrisv(AT)cyberswitching.com Aug 09 05:29:07 * * Setting ENABLE_BINARY_LOCALES="0" and GLIBC_GENERATE_LOCALES="en.US.UTF-8" causes glibc_2.9 to error when Aug 09 05:29:09 * * http://bugs.openembedded.org/show_bug.cgi?id=5292 Aug 09 05:29:24 zecke: ok i'll take a look and append to the bug when i get some more info. Aug 09 05:29:37 the build takes a little time before the error pops up Aug 09 05:35:02 zecke: thanks again for the help! Aug 09 05:47:31 kg4ysn|away: thanks Aug 09 06:28:59 03Holger Hans Peter Freyther  07org.openembedded.dev * r8e50ddc483 10openembedded.git/recipes/qt4/ (qt4-embedded.inc qt4-x11-free.inc): (log message trimmed) Aug 09 06:28:59 qt4-x11-free.inc: Do not override the var set by qt4-x11-free-gles Aug 09 06:28:59 Another attempt to fix the dependencies of qt4-x11-free-gles. Do not Aug 09 06:28:59 make libqt4-FOO-dev depend on qt4-x11-free-gles which will drag in Aug 09 06:29:00 more dependencies. Aug 09 06:29:02 A better fix would be to build the OpenVG graphicsengine unconditionally Aug 09 06:29:04 and build the screendriver out of this tree and install it as a plugin. Aug 09 08:05:49 ~seen Sup3rkiddo Aug 09 08:05:51 sup3rkiddo was last seen on IRC in channel #openmoko, 137d 16h 26m 41s ago, saying: 'its not booting, i read the wiki as well.. strangely enough it worked with my friend's battery'. Aug 09 08:05:58 bleh Aug 09 09:14:21 03Khem Raj  07org.openembedded.dev * r257821eeea 10openembedded.git/recipes/pciutils/pciutils_3.1.2.bb: Aug 09 09:14:21 pciutils_3.1.2.bb: Dont install gzipped versio of pci.ids its not there. Aug 09 09:14:21 Signed-off-by: Khem Raj Aug 09 09:14:22 03Khem Raj  07org.openembedded.dev * rdc2cbe1749 10openembedded.git/conf/machine/include/ (tune-sh3.inc tune-sh4.inc): Aug 09 09:14:22 tune-sh3.inc tune-sh4.inc: Define BASE_PACKAGE_ARCH Aug 09 09:14:24 Signed-off-by: Khem Raj Aug 09 09:14:26 03Khem Raj  07org.openembedded.dev * rcb58de6bef 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev Aug 09 09:14:29 03Khem Raj  07org.openembedded.dev * ra394010d44 10openembedded.git/classes/sanity.bbclass: Aug 09 09:14:32 sanity.bbclass: Correct typo to refer to PACKAGE_EXTRA_ARCHS Aug 09 09:15:19 We were using EXTRA_PACKAGE_ARCHS which was sort of misleading Aug 09 09:15:19 if you started to search for this variable. It was a typo Aug 09 09:15:19 hence fixed as obvious. Aug 09 09:15:20 Signed-off-by: Khem Raj Aug 09 10:11:55 03Holger Hans Peter Freyther  07org.openembedded.dev * rb9d8c34bc0 10openembedded.git/recipes/uclibc/ (uclibc-0.9.30.1/gcc-4.4-fixlet.patch uclibc_0.9.30.1.bb): Aug 09 10:11:55 uclibc-initial_0.9.30.1.bb: Fix do_stage for ubuntu karmic Aug 09 10:11:55 gcc4.4 and glibc 2.10 complain that getline already has Aug 09 10:11:55 a different signature. Rename the version in the unidef.c. Aug 09 10:40:08 03Holger Hans Peter Freyther  07org.openembedded.dev * r39f8b8ac1d 10openembedded.git/docs/usermanual/chapters/common_use_cases.xml: usermanual: Grammar fix. Aug 09 10:42:18 flo_lap: hey, you will get an email today... Aug 09 10:42:25 anyone has access to patchwork? Aug 09 10:43:47 hi all Aug 09 10:44:02 zecke: sounds very promising - cool! Aug 09 10:44:36 morning florian :-) Aug 09 10:47:05 hi rkirti Aug 09 10:49:47 flo_lap: :} Aug 09 10:49:55 flo_lap: I found out that 7-eleven can scan documents Aug 09 10:50:28 zecke: 7-eleven? Aug 09 10:50:51 flo_lap: convience stores in asia... can scan, print, fax documents Aug 09 10:51:23 zecke: ok great! Aug 09 10:57:01 flo_lap: how is the summer? Aug 09 10:59:00 so i've built a gpe-image rootfs for the sl-5600 and from the errors i get at boot using altboot i'm extrapolating that i might need to set up the fstab correctly. am i right? Aug 09 10:59:52 zecke: last week was pretty good... it deserved the name "summer" :) Aug 09 11:01:37 today is less sunny but warm enough to enjoy sitting in the garden (and the fact that this village has dsl now) Aug 09 11:01:54 Hello everyone Aug 09 11:02:50 * Jay7 is hoping to get free month one summer and go through europe :) Aug 09 11:06:34 flo_lap: dsl... finally :) Aug 09 11:06:51 i've got a question about my oe configuration, somthing is not quite working and i don't know how to solve this problem because i'm a linux newbie, would somebody be so kind and support me a little bit? Aug 09 11:07:19 just ask and if anyone has an answer they will help you out Nobody2303 Aug 09 11:07:19 Nobody2303: you can just ask your question here but without any warranties :) Aug 09 11:07:32 hehe Aug 09 11:07:52 i get an unable to parse error if i want to start bitbake nano Aug 09 11:08:09 pastebin' it Aug 09 11:09:01 ERROR: Unable to parse conf/bitbake.conf (Could not inherit file classes/base.bbclass) Aug 09 11:09:03 xcasex: about fstab - imho yes. and don't forget kernel's cmdline 'root=' parameter Aug 09 11:10:56 Jay7: eh the latter part ? Aug 09 11:11:54 I've not using altboot, but imho you should specify that your root is on SD/CF Aug 09 11:12:16 ah yes i did, but i will check out what configs i can find relating to booting off of sd Aug 09 11:12:17 I think it's a permission problem, i'm using kubuntu 9.04 and the oe folder is within home Aug 09 11:12:25 if altboot not does that automagically Aug 09 11:12:41 xcasex: kexecboot is modern way :) Aug 09 11:13:04 kexecboot? wow am i missing out, anything prebuilt for poodle? Aug 09 11:13:17 hmm.. Aug 09 11:13:33 http://www.angstrom-distribution.org/demo/kexecboot/ Aug 09 11:13:41 looks like Aug 09 11:14:08 it should work on collie at least Aug 09 11:14:14 someone was reported Aug 09 11:15:54 so kexecboot makes what easier? Aug 09 11:19:34 booting multiple images e.g. :) Aug 09 11:23:39 zecke, yas I have patchworks access what do you need? Aug 09 11:24:50 Jay7: aha! well i guess i'll have to try it out soon :D Aug 09 11:27:14 Crofton|work: I want to archive patches? is that the right thing to do? Aug 09 11:27:29 yeah Aug 09 11:27:36 I need to give you access hang on Aug 09 11:27:49 Crofton|work: zecke is my user account Aug 09 11:28:00 heh Aug 09 11:28:05 I could guess that :) Aug 09 11:29:14 ok check now Aug 09 11:29:33 site looks different :) Aug 09 11:30:09 great Aug 09 11:30:15 we have limited admin options Aug 09 11:30:20 so everyone is an admin Aug 09 11:31:54 Can anyone help me with this Problem? Aug 09 11:31:56 ERROR: Unable to parse conf/bitbake.conf (Could not inherit file classes/base.bbclass) Aug 09 11:32:08 Nobody2303: what instructions did you follow? Aug 09 11:33:05 i made the directions neccessary for the environment Aug 09 11:33:18 Nobody2303: have you read this: http://wiki.openembedded.net/index.php/Getting_started Aug 09 11:33:22 that's a good start Aug 09 11:33:29 and make sure you have no symlinks in your path Aug 09 11:33:52 Nobody2303: you have not exported BBPATH Aug 09 11:34:05 i don't even know how i can be sure that i don't have any symlinks in the path ^^ Aug 09 11:35:02 this is out of my source-me.sh that i should write Aug 09 11:35:05 export BBPATH="${OE_LOCAL}:${OE_BRANCH}" Aug 09 11:36:05 Nobody2303: did you source it Aug 09 11:36:14 try echo $BBPATH Aug 09 11:36:14 of course ^^ Aug 09 11:36:34 this is the whole file Aug 09 11:36:35 bbiab Aug 09 11:36:37 #!/bin/sh Aug 09 11:36:37 export OE_PATH="${HOME}/oe" Aug 09 11:36:37 export OE_BRANCH="§{OE_PATH}/openrobotix" Aug 09 11:36:37 export OE_LOCAL="${OE_PATH}/build" Aug 09 11:36:37 export BBPATH="${OE_LOCAL}:${OE_BRANCH}" Aug 09 11:36:39 export BB_ENV_EXTRAWHITE="OE_PATH OE_BRANCH OE_LOCAL" Aug 09 11:36:41 export PATH="${PATH}:${OE_PATH}/bitbake/bin" Aug 09 11:36:58 what says echo $BBPATH Aug 09 11:37:27 §{OE_PATH}/openrobotix Aug 09 11:37:34 thats the answer Aug 09 11:38:09 s/?/$/ Aug 09 11:38:18 Nobody2303: daft question but you are pushing that scritpt into your global env? i.e. . ./source-me. Aug 09 11:38:19 ah, you used a paragraph symbol not a $ in OE_BRANCH Aug 09 11:38:51 Ahhh, yep, that script is not good ;-) Aug 09 11:39:01 it lies in the build folder within /oe Aug 09 11:39:46 narf Aug 09 11:39:57 this is not good: Aug 09 11:39:57 export OE_BRANCH="§{OE_PATH}/openrobotix" Aug 09 11:40:14 should be a $ before {OE_PATH} Aug 09 11:40:14 thats the problemm when you write scripts at 4 am xD Aug 09 11:40:28 13.40 here Aug 09 11:40:53 np, similar things have happened to most of us as well Aug 09 11:41:32 here it is 13:40 also ^^ Aug 09 11:41:38 <--- german Aug 09 11:41:59 * eFfeM is in NL Aug 09 11:42:13 ah an neighbour ^^ Aug 09 11:42:30 yes Aug 09 11:42:43 as a duch proverb says: a good neighbour is better than a far friend :-) Aug 09 11:42:54 hehe ^^ Aug 09 11:43:13 and a new error Aug 09 11:43:15 -.- Aug 09 11:43:20 ERROR: Openembedded's config sanity checker detected a potential misconfiguration. Aug 09 11:43:20 Either fix the cause of this error or at your own risk disable the checker (seesanity.conf). Aug 09 11:43:20 Following is the list of potential problems / advisories: Aug 09 11:43:20 /proc/sys/vm/mmap_min_addr is not 0. This will cause problems with qemu so please fix the value (as root). Aug 09 11:43:43 i think i most of the time ignore that one, but you can do Aug 09 11:43:53 echo 0 >/proc/sys/vm/mmap_min_addr Aug 09 11:44:03 that will get it away Aug 09 11:44:07 i must say that i made the folder under / and copied it to home after i tried anything Aug 09 11:45:32 not sure if that matters Aug 09 11:46:06 the echo doesn't work Aug 09 11:46:27 did it with sudo but cdid not change anything Aug 09 11:46:42 after that cat the file to see if it is correct Aug 09 11:46:46 and had to remove the > Aug 09 11:47:03 you need the > Aug 09 11:47:09 otherwise you do not change the var Aug 09 11:47:20 might need echo -n though, not sure and not behind my system Aug 09 11:47:50 than he says permission denied even with sudo Aug 09 11:48:19 the > might bind wrong, do sudo bash then in the shell do the echo Aug 09 11:48:23 or apply quotes wisely Aug 09 11:48:40 but that is not one of my skills :-) Aug 09 11:49:10 shall i try it within sudo -i ? Aug 09 11:49:24 you can always try Aug 09 11:49:31 well.. current .dev build of x11-image for akita at least boots in qemu Aug 09 11:50:06 * Jay7 still can't build tosa's x11-image Aug 09 11:50:24 aaaah it worked out of sudo -i ;) Aug 09 11:50:54 Nobody2303: you should not install OE to use it under root Aug 09 11:51:28 there is no any reasons to do it Aug 09 11:51:50 but i need always sudo to do anything even to create an folder so they are always root or am i wrong Aug 09 11:52:06 Jay7: you need to be root temoprarily to set mmap_addr to 0 Aug 09 11:52:30 Nobody2303: apart from this setting everything should work as a regular user Aug 09 11:52:41 ah, yes Aug 09 11:52:47 assuming you started as a regular user (otherwise your permissions might be wrong) Aug 09 11:53:19 and ownerships Aug 09 11:53:31 i also set the permissions of the whole oe folder for everyone to read &write permission Aug 09 11:54:43 sakoman: ping, quick question, been playing with the gnome images but do you have a task and image bb anywhere for 'Sakomans Gnome image' I could try to check I am building in everything? I can't see it in your GIT but I may be missing the obvious (again) Aug 09 11:54:46 don't forget x for the dirs Aug 09 11:55:08 i used sudo dolphin an changed the permissions myself and the owner from root to my username Aug 09 11:57:23 thanks alot bitbake is handling the bibake files now and i hope there woun't be any problems anymore ^^ Aug 09 12:06:04 Nobody2303: good luck! Aug 09 12:06:38 thank you neighbour ^^ Aug 09 12:47:37 mickey|bbq: ping? Aug 09 13:07:37 03Stanislav Brabec  07org.openembedded.dev * reef4af6584 10openembedded.git/recipes/xcalibrate/xcalibrate_git.bb: xcalibrate: Use increment-able PV format as other git version recipes do. Aug 09 13:21:11 DJWillis: I'm slowly cleaning up a feeding all my changes upstream Aug 09 13:22:01 But there were more than 100 changes, so it is taking a while! Aug 09 13:28:17 mwester: ping Aug 09 13:38:23 this is odd, it boots into the gpe rootfs, but doesn't launch the gui Aug 09 13:38:33 and without any errors. Aug 09 13:38:36 very odd. Aug 09 13:38:38 any ideas? Aug 09 13:41:53 xcasex: is is trying to launch anything x related? Aug 09 13:43:12 there isnt anything xrelated in the image :S Aug 09 13:43:27 its just the oe/images/gpe-image image Aug 09 13:44:25 xcasex: well, you should have an xserver... Aug 09 13:45:28 03Klaus Kurzmann  07shr/import * ra2756e30cc 10openembedded.git/recipes/opkg/opkg_svn.bb: Aug 09 13:45:28 opkg: fix postinst for SHR Aug 09 13:45:28 Signed-off-by: Klaus Kurzmann Aug 09 13:51:07 zecke: yeah .. i guess i should and am setting up right now Aug 09 13:57:06 03Holger Hans Peter Freyther  07org.openembedded.dev * r7e5dace40c 10openembedded.git/recipes/tasks/task-qte-toolchain-target.bb: Aug 09 13:57:06 task-qte-toolchain-target: Do not install the full qt4-embedded Aug 09 13:57:06 Do not install debug packages and examples making the SDK Aug 09 13:57:06 a bit smaller. Aug 09 14:08:56 hi, is there a switch for Xfbdev to get mouse acceleration? Aug 09 14:11:36 zecke: pong Aug 09 14:17:58 03Steve Sakoman  07org.openembedded.dev * r70924feaec 10openembedded.git/ (6 files in 3 dirs): gnome-common: restructure recipes to use gnome-common.inc, add 2.26.0 Aug 09 14:17:59 03Steve Sakoman  07org.openembedded.dev * r4a402ce814 10openembedded.git/recipes/gnome/gnome-common-2.18.0/omf.patch: gnome-common: one more bit of cleanup Aug 09 14:18:00 03Steve Sakoman  07org.openembedded.dev * r2215a06d34 10openembedded.git/ (conf/checksums.ini recipes/gnome/gnome-python_2.26.1.bb): gnome-python: add 2.26.1 Aug 09 14:18:02 03Steve Sakoman  07org.openembedded.dev * r48d4343132 10openembedded.git/ (conf/checksums.ini recipes/gnome/gconf-editor_2.26.0.bb): gconf-editor: add 2.26.0 Aug 09 14:18:04 03Koen Kooi  07org.openembedded.dev * r96efe18731 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.org:openembedded into org.openembedded.dev Aug 09 14:28:45 jo Aug 09 14:28:56 hello woglinde Aug 09 14:29:14 he rkirti Aug 09 14:29:27 all clear now with oe-overlay? Aug 09 14:29:56 woglinde: yep. thanks :-) I got my own almost done. Only need to clean up some conf files. Aug 09 14:30:22 cool Aug 09 14:30:26 woglinde: thanks for the help. My earlier tree was making it very difficult for anyone to build and test my stuff Aug 09 14:31:50 erk, i didn't think about do_fetch[depends] = "shasum-native:do_populate_staging" .. there's no equivalent in my new model, since it isn't shasum-native that deploys its bits into staging.. hrmm. Aug 09 14:32:16 jo kergoth Aug 09 14:32:21 hey Aug 09 14:32:23 yesss strike Aug 09 14:32:36 I have updated gstreamer-ti to the latest dvsdk Aug 09 14:33:17 now I need a beagleboard to test it Aug 09 14:33:22 g'day kergoth, woglinde Aug 09 14:33:26 hey pb_ Aug 09 14:33:35 * pb_ returns from vacation! Aug 09 14:33:53 re pb Aug 09 14:34:03 hello pb_ :-) Aug 09 14:34:05 pb where did your trip went? Aug 09 14:36:49 hi rkirti Aug 09 14:36:51 woglinde: west wales Aug 09 14:37:28 * kergoth just broke bitbake Aug 09 14:37:46 ... more specifically, here: http://maps.google.com/maps?f=q&source=s_q&hl=en&geocode=&q=&sll=37.0625,-95.677068&sspn=26.535573,48.603516&ie=UTF8&ll=51.994374,-4.371475&spn=0.001255,0.002967&t=h&z=18 Aug 09 14:37:53 kergoth: doh Aug 09 14:38:47 pb_: http://kergoth.pastey.net/120612 Aug 09 14:38:54 ever seen that one? Aug 09 14:44:21 pb nice Aug 09 14:44:30 seems really quiet there Aug 09 14:44:41 pb_: http://gist.github.com/164598 - functional, but not yet complete, private staging implementation. at least, it was functional last night. this morning, bitbake explodes Aug 09 14:44:59 * kergoth grumbles Aug 09 14:45:37 kergoth bisect? Aug 09 14:47:18 * kergoth 's working outside the upstream git tree at the moment.. time for a diff Aug 09 14:48:55 kergoth: cute, I'll have a look at that presently Aug 09 15:02:04 woglinde, summer holidays Aug 09 15:02:12 things always get slow this time of year Aug 09 15:03:33 crofton? Aug 09 15:03:43 I have no summer holidays Aug 09 15:03:45 not really Aug 09 15:04:01 * rkirti loves summer holidays. Only time to dabble properly in projects one likes. They are over for me though :-/ Aug 09 15:12:22 cold winter nights are more productive ;) Aug 09 15:17:14 * kergoth thinks bb.data.inherits_class is stupid. what else would we inherit but a class? bb.data.inherits would make more sense Aug 09 15:19:35 * kergoth sighs Aug 09 15:24:25 hi Aug 09 15:28:51 I got a running version of openembedded - a kernel and a .img file. Now I would like to modify these. How do I find out how they were created to build upon them? Aug 09 15:28:59 I'm starting reading the oe manual Aug 09 15:29:21 a kernel and an img file isn't openembedded. its an image and a kernel built by openembedded. Aug 09 15:29:35 right Aug 09 15:30:05 you can't reproduce the OE build environment magically from a root filesystem, sorry Aug 09 15:30:55 can I find out whether it is angstrom, etc? Aug 09 15:31:04 check for /etc/angstrom-version Aug 09 15:31:09 assumign its angstrom Aug 09 15:31:13 which isn't the only distro in openembedded Aug 09 15:31:45 oh yeah ... it is openzaurus 3.5.3, /etc/issue and /etc/version Aug 09 15:31:58 kergoth: thanks, I guess I'll ask those people Aug 09 15:32:15 OZ isn't maintained anymore, to my knowledge Aug 09 15:32:32 :-/ Aug 09 15:32:47 get OE, built angstrom for your zaurus, see if it works Aug 09 15:32:49 thats your best bet Aug 09 15:32:54 hmm ... can I use angstrom or any other distro or does that very much depend on the hardware? Aug 09 15:33:09 in OE, distro and machine are quite independent Aug 09 15:33:11 I have an iPAQ (hx2790) Aug 09 15:33:20 oh cool Aug 09 15:33:29 distro decides policy, machine decides device specifics, kernel modules, etc Aug 09 15:33:37 so angstrom is kind of the most active at the moment? Aug 09 15:33:59 its the closest to what your expectations are, from what i can tell. Aug 09 15:35:24 I just need a terminal and would like to get a newer version of bluez or bluetooth support ... I think that already works in the OpenZaurus I have right now, but I will definitely need some tinkering environment Aug 09 16:07:20 grrrr Aug 09 16:11:48 probably a bit off topic so just tell me to get lost if i'm in the wrong place. but built oe for my device, want to be able to upgrade the whole root fs image from within the booted system. that is, it's booted from the root parititon, i want to be able to erase that and reflash it. anyone know of any resources (particularly a howto of sorts if possible) i can read on doing this successfully? I keep having to break into uboot and do it manually because the sys Aug 09 16:20:08 okay, that was a royal pain in the ass Aug 09 16:20:37 i now have it, at parse time, replacing do_populate_staging in task 'depends' flags with do_capture Aug 09 16:20:58 then i have an event catching the TaskStarted, checking the depends flag, and populating the private staging area with those packages Aug 09 16:21:17 had to implement locking on the private staging area, and tracking of what's already been staged, and.. Aug 09 16:21:19 * kergoth rolls eyes Aug 09 16:21:50 hrmph, staging.bbclass is up to 306 lines Aug 09 17:03:10 BB_NUMBER_THREADS and PARALLEL_MAKE, what should I set them to? I used -j2 as I only have one cpu, but how many threads? Aug 09 17:03:39 how many tasks do you want bitbake to run at the same time? Aug 09 17:04:10 kergoth: I don't know? Aug 09 17:04:42 well, if you don't, nobody else is going to be able to tell you. it depends on your cpu, disk, and how usable you want your system to remain other than the bake Aug 09 17:05:15 having 2 threads can be nice even on a single core single cpu box, just because then it can run a fetch or unpack in the background while something is compiling Aug 09 17:05:31 ok I used -j2 and 4 threads and the system is freezing half of the time Aug 09 17:05:41 that's not surprising. Aug 09 17:08:06 pb_: heh, the main problem i see with this thing is that it doesn't override the prefix, etc in native/cross, which means all of those will need to be relocatable, which is not the case today. either need to fix those, which would be worth doing, or i need to change how I'm doing this.. hmm Aug 09 17:08:08 * kergoth thinks Aug 09 17:09:08 kergoth: so 2 threads or 2 processes or both? what specs of my system do you need to know? Aug 09 17:09:33 again, it depends. try 2/2 and see how it behaves for you. it won't take long to figure it out Aug 09 17:10:13 I don't think it will be faster with more than 3 tasks at a time (swapping & scheduling) Aug 09 17:34:23 whew, this is a lot of work Aug 09 17:36:59 kergoth: what are you doing? Aug 09 17:37:24 rewriting the OpenEmbedded staging implementation Aug 09 17:38:34 staging is what part again? Aug 09 17:38:52 when one recipe needs something another recipe provides, thats where staging comes in Aug 09 17:39:02 libfoo puts files in staging, bar links against libfoo by pointing at staging Aug 09 17:39:33 dependency resolving? or similar to virtual packages in gentoo? Aug 09 17:39:39 no, neither. Aug 09 17:39:52 dynamic linking? Aug 09 17:39:52 there's no equivalent in desktop distros, because everything puts things in standard paths and everything looks in standard paths Aug 09 17:41:03 where can I read more about it? Aug 09 17:41:21 J13R: OE user manual Aug 09 17:41:57 J13R: http://docs.openembedded.org/usermanual/usermanual.html is good to begin with Aug 09 17:42:04 yes Aug 09 17:42:29 kergoth: sorry if I am interrupting, but wont your new model increase bitbake's runtime ? Aug 09 17:43:01 it will add a bit of delay during each staging population, but i think its worthwhile to increase reliability and determinism, since those are some of the biggest complaints Aug 09 17:43:07 most people complain about the parse time, not the run time Aug 09 17:43:10 also, we can cache it Aug 09 17:43:12 does oe do static or dynamic linking? Aug 09 17:43:14 i was just thinking about how to do it Aug 09 17:43:42 i was thinking that we could do something like.. if a staging area contains glibc, shasum-native, libtool-native, make an archive for that combination of deps, so anyone else with the same deps will sue it Aug 09 17:43:45 s/sue/use/ Aug 09 17:44:12 hash the staging idents / archive names for all the deps that go into staging, and archive the staging area for later use Aug 09 17:44:21 but thatll just be an optimization, caching mechanism Aug 09 17:44:41 J13R: generally everything is dynamically linked, with a couple recipes that are exceptions. you can change that if you choose to, but i doubt you'd want to Aug 09 17:45:49 oh well, 10% done. time to sleep a few hours. Will ask more stupid questions tomorrow :-] Aug 09 17:46:01 kergoth: iirc, you said the whole staging revamping is to let the user choose what versions to choose for staging. so for caching can't you just have a listing of packages/files with the corresponding versions to be used ? Aug 09 17:46:12 yeah, thats pretty much what I'm saying Aug 09 17:46:34 i have this thing i call a staging ident, its basically the package name, version ,revision + target arch/os, and for cross, + build arch/os Aug 09 17:46:44 very fair Aug 09 17:46:55 so if we hash the set of those for all of the bits we pull into staging, thatd be a great identifier for the cached staging archive Aug 09 17:47:17 but i want to profile builds with and without this to see how significant the difference really is, before i work on that :) Aug 09 17:47:45 this is really a proof of concept, i often dive into crazy rework like this, and at the end, it may or may not end up being worthwhile, but if i don't do it, it'll bug me until i do... Aug 09 17:48:00 this has been sitting in the back of my mind for like.. years now Aug 09 17:48:09 since before i left the project Aug 09 17:48:48 :-) Aug 09 17:49:20 the problem with stuff like this is you can never anticipate how itll break things. any core changes can break any number of random recipes that make assumptions about internals / implementation details Aug 09 17:49:37 thats why i never got my autotools class cleanups upstream, every 5th recipe exploded Aug 09 17:51:05 * rkirti wonders why/how a recipe would do something that depended on internals of bitbake tasks Aug 09 17:51:41 well, consider recipes that add their own tasks for example Aug 09 17:51:55 if i changed around what tasks run or what order they run in.. Aug 09 17:52:39 or consider those that do iffy things like override sbindir/bindir/prefix, which assumes that its able to do so without breaking things Aug 09 17:53:39 right, but in this case, we are only looking at do_stage_append/prepend and typically these custom tasks only add files to the staging area - so.with all my naiveness I fail to see how it could get affected if internals change Aug 09 17:54:10 * rkirti doesnt much about overrides - yet another section which is "please complete me" in the OE user manual.. Aug 09 17:54:37 well, right now, the native class changes prefix/etc so that its *built* to go into staging Aug 09 17:54:48 if i leave it that way, the archives will contain that deep path Aug 09 17:55:05 and then when it gets unpacked into staging, it'll be like tmp/.../staging/home/clarson/... Aug 09 17:55:33 * rkirti wants to pain the people here sometime, learn the details and finish those "please complete me" sections in the manual sometime Aug 09 17:55:40 if a recipe overrides prefix/sbindir/etc, the contents of the archive wont match up with the expected contents of staging Aug 09 17:56:42 ahh Aug 09 17:57:16 the original intent of our directly parsed file format was to make it less locked in than portage, since they're stuck with shell. but the problem is, our recipes end up making assumptions about the way we do things anyway, so a conversion to another format would still lose. though, it could get you half way there, the user would have to fix the rest... Aug 09 17:57:36 which is part of why I'd like to make it so recipes can't screw with how things work under the hood Aug 09 17:57:48 make them just define descriptive metadata and their configure, copmile, etc, and thats -it- Aug 09 18:00:09 hmm Aug 09 18:06:39 * kergoth grumbles, time to add the staging mangling callbacks/scripts Aug 09 18:08:23 re Aug 09 18:12:00 hey woglinde Aug 09 18:14:17 hmm Aug 09 18:17:37 kergoth I can help you Aug 09 18:17:39 ups cannt Aug 09 18:39:35 hi Aug 09 18:44:47 hi hrw Aug 09 18:51:35 03Thomas Zimmermann  07shr/import * rab64018b0b 10openembedded.git/recipes/epdfview/ (epdfview_0.1.7.bb epdfview_svn.bb files/acroread.png): Update ePDFView to 0.1.7 and add icon Aug 09 19:13:53 re Aug 09 19:18:57 so ehrm am i misunderstanding something or does the sl-5600 have 64mb of internal space to reflash with a rom? or err... durrr.. bit mushy after a whole day of crash courses w. it Aug 09 19:27:54 xcasex: afaik yes. sl-5600 have only 64Mb of nand Aug 09 19:28:05 florian: good evening Aug 09 19:29:15 hmm.. or even 16Mb Aug 09 19:29:38 Jay7: now you're back to confusing me again Aug 09 19:29:44 florian: ping Aug 09 19:29:48 ah, no, 5500 have 16Mb Aug 09 19:30:50 xcasex: yep Aug 09 19:30:56 xcasex: Memory: 64MB protected Flash, 32MB SDRAM Aug 09 19:31:00 hi pb_ Aug 09 19:32:07 jo florian Aug 09 19:35:52 florian: i really bought the wrong device eh ? ;) Aug 09 20:32:32 * kergoth goes back to battling OpenEmbedded Aug 09 20:33:41 okay, if do_stage dies, what should I name the mechanism for adding post-stage mangling snippets? Aug 09 20:33:50 i.e sed'ing .la files, sed'ing libc.so Aug 09 20:36:49 sed'ing binaries? Aug 09 20:37:00 libc.so is a linker script Aug 09 20:37:02 not a binary Aug 09 20:37:16 pulls in libc nonshared .a as well as libc.so.6 Aug 09 20:37:24 uses full paths by default Aug 09 20:38:12 hm okay Aug 09 20:38:24 do a cat /usr/lib/libc.so on your desktop :) Aug 09 20:38:29 yes did it Aug 09 20:38:31 already Aug 09 20:38:34 hehe Aug 09 20:44:19 jo stefan Aug 09 20:44:40 hi woglinde Aug 09 21:00:30 hmm, given the proliferation of "all" tasks.. fetchall, buildall, etc, it may be worth adding a bitbake commandline arguments for it Aug 09 21:37:25 hrw: do you have time for a few questions on the at91sam9263ek board using nandflash? Aug 09 21:55:34 I think hrw is at sleep Aug 09 22:00:25 woglinde: think you might be right. too bad, too ... i'm being driven nuts by this board. Aug 09 22:02:39 take a better one Aug 09 22:03:03 woglinde: if only i could :-) Aug 09 22:56:08 someone up? Aug 09 22:56:21 i need help with git format-patch command Aug 09 22:58:22 whats the problem? Aug 09 23:01:14 i made a commit in my local repo to my own branch Aug 09 23:01:49 i use git format-patch -CM --find-copies-harder Aug 09 23:02:00 and my commit never appears as a patch Aug 09 23:02:06 what could be the problem? Aug 09 23:02:26 i tested with origin, the original branch, the remote branch.... Aug 09 23:02:31 but nothing happes Aug 09 23:02:34 but nothing happens Aug 09 23:03:27 have you attached a log revision to it? Aug 09 23:03:50 yes Aug 09 23:04:10 i could see the commit with: git log, gitk, git gui .... Aug 09 23:04:15 but no way :( Aug 09 23:04:25 hm never used it Aug 09 23:04:43 you never used git format-patch? Aug 09 23:04:43 is it find without these option? Aug 09 23:04:57 not with -CM and -findcopies Aug 09 23:05:00 i try without this option, Aug 09 23:05:08 without luck Aug 09 23:05:12 it's really strange Aug 09 23:05:18 hm Aug 09 23:05:26 go one log back Aug 09 23:05:32 aeh lognumber Aug 09 23:06:13 how? i never tried this option Aug 09 23:06:38 aeh Aug 09 23:06:43 logrevision Aug 09 23:06:52 the checkin number Aug 09 23:09:15 with the commit id works Aug 09 23:10:13 if you want your commit in Aug 09 23:10:25 i've!! Aug 09 23:10:27 go one commit id before ypur commit Aug 09 23:10:33 and no its in? Aug 09 23:10:34 i know what options didn't work Aug 09 23:10:36 ups now Aug 09 23:10:39 -MC together Aug 09 23:10:44 hm Aug 09 23:10:46 hehe Aug 09 23:11:44 but i usually use in that way: git format-patch Aug 09 23:11:51 and usually it works Aug 09 23:12:49 me too Aug 09 23:13:13 well, it doesn't matter Aug 09 23:13:19 i'mgoing to sent it know Aug 09 23:13:26 let's see if you accept it ;) Aug 10 00:18:42 in case anyone else is working on the Atmel AT91SAM9263-EK board (or its processor) along with OpenEmbedded, I'm documenting a lot of the process needed to get OE building Angstrom images. The #oe room has been a great help in providing pieces of the puzzle, so I'd like to use this as an opportunity to give back. Feel free to follow the work and add helpful information: http://www.headnut.org/tag/at91sam9263 Aug 10 02:02:05 hi. qemu wouldn't build on my 64bit gentoo Aug 10 02:02:38 Which didn't surprise me because the gentoo package doesn't build either Aug 10 02:03:04 so I am going now for a ubuntu i386 in vmware Aug 10 02:03:25 although I'm open to suggestions Aug 10 02:35:03 * kergoth grumbles, can't decide how to get this mangling done cleanly Aug 10 02:41:31 it annoys me when i get all the hard stuff done in no time, then get hung up on stupid shit Aug 10 02:41:34 * kergoth shakes fist at the universe Aug 10 02:47:20 kergoth: that only happens to people who do real work :-] Aug 10 02:47:28 kergoth: what OS do you use? Aug 10 02:48:00 my main workstation for non-development work is a macbook, my main development machines are linux, and i game on a windows box at home. Aug 10 02:49:33 what linux and what architecture? Aug 10 02:50:11 debian and ubuntu, generally Aug 10 02:50:35 i386 Aug 10 02:50:56 or 64bit? Aug 10 02:51:27 who cares? Aug 10 02:51:54 i do, because qemu doesn't work on 64bit and I would like to know what you use Aug 10 02:55:29 i use qemu every day on my x86_64 box, so I'm not sure what you're talking about. Aug 10 02:55:38 i use it to test OpenEmbedded and mvl6 distros all the time Aug 10 02:56:02 hmm ... Aug 10 02:57:04 maybe you have a 32bit qemu running in compatibility mode, and my gentoo tries to build a 64bit. Aug 10 02:59:03 oe builds qemu Aug 10 02:59:16 And it's a host native bin, so 64bit **** ENDING LOGGING AT Mon Aug 10 02:59:57 2009