**** BEGIN LOGGING AT Mon Oct 20 02:59:57 2008 Oct 20 06:33:45 morning Oct 20 07:23:15 03Phil Blundell  07org.openembedded.dev * ra62e78edb4 10openembedded.git/packages/busybox/busybox.inc: busybox: tolerate empty layout_execprefix Oct 20 07:23:25 03Phil Blundell  07org.openembedded.dev * r2cab657e1a 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Oct 20 07:54:24 bonjour Oct 20 08:48:46 hi all Oct 20 08:50:08 morning Oct 20 08:59:38 I m compiling freetype2, for arm, It is saying " configure: error: cannot run C compiled programs." what could be the problem ? Oct 20 09:03:18 nakul_: it would appear that configure doesn't realise you are trying to cross compile. are you using the unmodified .bb files from the repository? Oct 20 09:09:45 No. I have created .bb file for freetype, & its do_configure is do_configure(){oe_runconf} Oct 20 09:10:34 Why are you not using the existing .bb files? Oct 20 09:12:33 I am using different source code for Freetype2 Oct 20 09:13:11 This is the error :: | checking whether the C compiler works... configure: error: cannot run C compiled programs. Oct 20 09:13:11 | If you meant to cross compile, use `--host'.| See `config.log' for more details.| make: *** [setup] Error 1 Oct 20 09:23:28 nakul_: I think you would do better to start with the .bb files that work, rather than trying to invent your own from scratch. Oct 20 09:23:56 as you can see from the log fragment you posted, the configure script clearly doesn't think you are setting it up for cross compiling. Oct 20 09:24:05 florian_: good morning Oct 20 09:24:33 good morning Oct 20 09:25:00 huhu Oct 20 09:29:22 are mtn commits disabled ? Oct 20 09:29:37 yes, they have been for some time Oct 20 09:29:47 see the mailing list Oct 20 09:31:41 * steliosk thought OE was still transitioning... Oct 20 09:32:05 pb__ : Thanks. Haven't look at the ml for a week or so Oct 20 10:30:26 03Koen Kooi  07org.openembedded.dev * r92002b2813 10openembedded.git/packages/nfs-utils/nfs-utils_1.1.2.bb: Oct 20 10:30:26 nfs-utils: add e2fsprogs to DEPENDS to satisfy libblkid dep Oct 20 10:30:26 * why do we have conflicting e2fsprogs and e2fsprogs-libs recipes? Oct 20 11:17:44 morning all Oct 20 11:45:36 03Koen Kooi  07org.openembedded.dev * r49ff305433 10openembedded.git/packages/libffi/libffi_2.0+gcc4.3.2.bb: libffi: update to gcc 4.3.2 Oct 20 12:09:04 moin Oct 20 12:09:30 hi zecke Oct 20 12:09:38 moin Oct 20 12:10:15 hi zecke Oct 20 12:44:45 hail zecke Oct 20 12:46:14 yo zecke Oct 20 12:47:42 Is the stable branch supposed to be compatible with latest bitbake 1.8? Oct 20 12:49:11 it certainly should be, yes. Oct 20 12:49:25 I'm pretty sure I am using the stable branch successfully with bitbake 1.8.something. Oct 20 12:51:15 I recently updated to latest from svn and stable suddenly fails. Oct 20 12:51:50 oh, right, latest from svn. I am using the latest released one. Oct 20 12:51:50 But I have to admit that I do not have stable from git yet, that's latest mtn version. Oct 20 12:52:03 of course, latest from svn also ought to work in theory. Oct 20 12:52:13 if bitbake has changed in a backwards-incompatible way, that would be quite a grave bug. Oct 20 12:52:52 * florian tries latest release Oct 20 12:58:19 ~lart setup.py for a lacking uninstall target Oct 20 12:58:19 * ibot hurls dozens of incontinent, insomniac, hungry kittens with tiny little razor-sharp claws and a wide variety of contagious intestinal parasites at setup.py for a lacking uninstall target Oct 20 13:04:03 morning Oct 20 13:05:36 morning .. Oct 20 13:05:54 do someone have any idea why xfce .desktop items doesn't work? Oct 20 13:06:02 i'm going to debug it now Oct 20 13:14:29 Chttp://pl.youtube.com/watch?v=tP-reW1eLYE Oct 20 13:18:02 ops Oct 20 13:20:57 does anyone know what a "Fault in unaligned fixup" mean? google revealed a lot of logs but nothing conclusive Oct 20 13:30:20 What platform are you seeing that on? Oct 20 13:30:26 03Thomas Kunze  07org.openembedded.dev * r0eb44cec07 10openembedded.git/packages/linux/ (3 files in 2 dirs): linux-rp_2.6.26: add patch for collie usb Oct 20 13:30:37 03Thomas Kunze  07org.openembedded.dev * rf932f18774 10openembedded.git/packages/xorg-xserver/ (2 files in 2 dirs): Oct 20 13:30:37 xserver-kdrive_1.4.090: fix rotated tslib input Oct 20 13:30:37 * close #3741 Oct 20 13:31:01 At a guess, it means that you took a second page fault while trying to handle an initial alignment trap fault. Oct 20 13:32:29 morning a;; Oct 20 13:32:32 and all Oct 20 13:35:35 hey RP Oct 20 13:37:40 hi RP Oct 20 13:39:03 hi RP Oct 20 13:42:21 good morning Oct 20 13:42:55 I've been using the trial git repo now for a few months Oct 20 13:43:03 pb_: I can confirm that it works with the 1.8.10 release Oct 20 13:43:11 florian: What failed exactly? Oct 20 13:43:26 florian: Something environment related by any chance? Oct 20 13:43:48 Is it possible to use some git magic to switch now to the official repo in order to maintain my change history? Oct 20 13:44:07 sakoman: What kind of changes did you make? Oct 20 13:44:22 sakoman: Lots of merges? Oct 20 13:44:28 Mostly file additions Oct 20 13:44:36 not very many mods at all Oct 20 13:44:47 you could try and rebase your tree onto the new repo Oct 20 13:45:15 how does one do that, I'm not a git guru Oct 20 13:45:51 pull both trees into one repo, checkout your changes, then "git rebase master" assuming "master" is the official repo head Oct 20 13:46:04 I really don't know how well it will work though Oct 20 13:50:23 RP: by "pull both trees" do you mean to use something like 'git remote add" of the new oe url? Or is there a way to pull the new repo directly? Oct 20 13:50:59 hasn't been obvious to me from reading the git docs Oct 20 13:51:35 sakoman: "git fetch someurl" should pull in the revisions Oct 20 13:52:05 I usually hack the .git/config by hand Oct 20 13:52:57 I tried just hacking in the url of the new repo, but that resulted in disaster :-) Oct 20 13:53:08 git remote add just does a config file edit for you. Oct 20 13:53:10 (in .git/config) Oct 20 13:53:25 I'll try fetch and see what happens Oct 20 13:53:33 (all you need to do is clone the origin block & give it a new name.) Oct 20 13:53:46 sakoman: You probably just need to make sure it doesn't try to merge it anywhere Oct 20 13:53:57 pull = fetch+merge Oct 20 13:53:57 fetch is safe for that, it won't do merges. Oct 20 13:54:12 hello Oct 20 13:54:18 I need a little help Oct 20 13:54:26 However if you don't add a remote presumably you'll need a commit ID to find any of the fetched blobs. Oct 20 13:54:41 apitrus: what kind of help do you require? Oct 20 13:55:09 sakoman: git remote add new-repo git://wherver ; git remote update Oct 20 13:55:28 Hi Phil, I'm trying to do a qemux86 build Oct 20 13:55:52 broonie: will try that now -- creating a new copy to work on just in case another mess ensues :-) Oct 20 13:56:18 RP: No other changes except of bitbake itself. Every package I built failed with either failing to find autoconf or executing gcc. configure.log only had the initial check for gcc in there... Oct 20 13:57:14 RP: Just downgraded bitbake and everythign is fine Oct 20 13:57:31 pb__:http://debian.pastebin.com/d748d4090 Oct 20 13:57:41 florian: ouch :( Oct 20 13:57:53 what am I doing wrong ? Oct 20 13:58:08 florian: I'd be interested to know if "BB_PRESERVE_ENV=1 bitbake foo" worked Oct 20 13:58:52 * RP was fairly sure others were using OE with bitbake 1.8 svn Oct 20 13:59:21 RP: ok, I can try this but I have to finish some stuff first Oct 20 13:59:29 * RP suspects PATH might not be marked as export in bitbake.conf Oct 20 14:00:10 florian: np, no rush. I would like to get that fixed though Oct 20 14:00:32 RP: I built some stuff from .dev using bitbake svn, so it seems to work for the newer stuff. Oct 20 14:01:00 oe-source-checker.py is good tool Oct 20 14:02:06 speaking of using svn bitbake... what is a status of using bitbake trunk with OE? Oct 20 14:02:16 apitrus: hm, apparently nothing. I guess you need to figure out what's become of patch-2.6.26-rc4.bz2. Oct 20 14:02:28 I think this can happen if you deleted the file from the downloads directory but left the md5sum there. Oct 20 14:02:39 apitrus: how old metadata you use? Oct 20 14:03:12 ok. we need to switch qemux86 to 2.6.26 at least Oct 20 14:03:23 hrw: It should roughly work Oct 20 14:03:37 * RP remembers koen mentioned a bug that was fixed in poky and not upstream yet Oct 20 14:03:39 I don't know what do you mean with "how old metadata you use?" sorry Oct 20 14:04:10 BTW I deonloaded all from scratch yesterday Oct 20 14:05:14 ok Oct 20 14:05:22 apitrus: I am pushing fix Oct 20 14:05:50 03Marcin Juszkiewicz  07org.openembedded.dev * ra5f93f8b43 10openembedded.git/packages/linux/linux-rp_2.6.25+2.6.26-rc4.bb: linux-rp: fix location of patch-2.6.26-rc4 Oct 20 14:07:19 03rpurdie * r1109 10/ (2 files in 2 dirs): runqueue.py: Fix nostamp flag handling (from Poky) Oct 20 14:10:14 03rpurdie * r1110 10bitbake/lib/bb/cooker.py: cooker.py: Fix an init problem (from poky) Oct 20 14:11:26 hrw: That should be enough to make bitbake trunk run again Oct 20 14:12:37 cool Oct 20 14:13:24 hrw: do I have to wait? Oct 20 14:13:47 git pull said : Oct 20 14:13:49 packages/linux/linux-rp_2.6.25+2.6.26-rc4.bb | 2 +- Oct 20 14:13:50 packages/linux/linux-rp_2.6.26.bb | 3 +- Oct 20 14:14:12 apitrus: thats fix Oct 20 14:15:12 hrw: if I do bitbake console-image it still fails Oct 20 14:15:38 apitrus: ERROR: linux-rp-2.6.25+2.6.26-rc4: http://kernel.org/pub/linux/kernel/v2.6/testing/v2.6.26/patch-2.6.26-rc4.bz2 has no entry in conf/checksums.ini, not checking URI Oct 20 14:15:42 that way? Oct 20 14:16:59 03Marcin Juszkiewicz  07org.openembedded.dev * r34331f2934 10openembedded.git/conf/checksums.ini: checksums.ini: make linux-rp, fakeroot, base-passwd happy Oct 20 14:17:18 apitrus: git pull and build Oct 20 14:17:56 finally console-iamge will build here Oct 20 14:23:10 <_bernie> hey, how do I get bitbake to pass -j4 to gnu make? Oct 20 14:23:42 _bernie: PARALLEL_MAKE = "-j 8" Oct 20 14:23:49 <_bernie> an env var? Oct 20 14:23:52 well, would be 4 in your case Oct 20 14:24:13 I have this in my site.conf but I think an env var or setting it in local.conf should work too Oct 20 14:24:28 this one is also useful: BB_NUMBER_THREADS = "4" Oct 20 14:24:48 <_bernie> Jin^eLD: oh there's also a site.conf? where does it go? Oct 20 14:25:06 _bernie: this will probably depend on your overall setup Oct 20 14:25:28 so you might not have a dedicated site.conf Oct 20 14:25:37 just put it in your local.conf for starters, should do the trick Oct 20 14:26:45 hrw: I dit git pull git checkout Oct 20 14:26:56 bitbake console-image Oct 20 14:27:03 ... Oct 20 14:27:06 NOTE: package linux-rp-2.6.25+2.6.26-rc4-r5.1: task do_patch: started Oct 20 14:27:06 NOTE: Applying patch 'patch-2.6.26-rc4' (/home/apitrus/oe/tmp/work/qemux86-angstrom-linux/linux-rp-2.6.25+2.6.26-rc4-r5.1/patch-2.6.26-rc4) Oct 20 14:27:06 NOTE: Task failed: Error: /home/apitrus/oe/tmp/downloads/patch-2.6.26-rc4.bz2 not found. Oct 20 14:27:24 do I have to clean something before ? Oct 20 14:28:59 apitrus: rm /home/apitrus/oe/tmp/downloads/patch-2.6.26-rc4.bz2*;bitbake -cclean linux-rp Oct 20 14:30:28 re Oct 20 14:32:40 hrw: excellent! it's building Oct 20 14:36:05 I planned to build a new PC for OE daily builds, do you think a AMD Phenom 4 core would be enough to speed up the build? Oct 20 14:36:27 What are you out there using? Oct 20 14:38:31 I'd like to point up another minor issue: http://puzzle.dl.sourceforge.net/sourceforge/expat/expat-1.95.7.tar.gz is extremely slooooow Oct 20 14:39:01 maybe would be better changing repo? Oct 20 14:39:10 sure, do you have a better url? Oct 20 14:39:35 if you have a preferred sourceforge mirror, you can configure it in local.conf Oct 20 14:39:43 cool! Oct 20 14:40:27 this goes to timeout repeatedly Oct 20 14:40:29 apitrus: I use core2quad 2.4GHz with 4GB ram Oct 20 14:41:04 apitrus: http://blog.haerwu.biz/2008/04/07/speeding-up-bitbake-builds/ also worth read Oct 20 14:41:19 thank you :-D Oct 20 14:49:37 03Cliff Brake  07org.openembedded.dev * rcf470a6ec3 10openembedded.git/contrib/weekly-changelog-report.py: contrib/weekly-changelog-report.py: add script to generate Oct 20 14:55:42 cbrake: |grep -v "Merge branch" would be nice addon? Oct 20 14:55:56 hrw: thanks good idea Oct 20 14:56:39 cbrake: but number after names does not match amount of lines then Oct 20 14:58:10 hrw: I'm reading your website Oct 20 14:58:28 ...about "Updates" Oct 20 14:58:52 hrw: maybe I can remove the count after the name" Oct 20 15:00:38 hey guys Oct 20 15:01:15 kergoth: hi Oct 20 15:02:25 hi kergoth Oct 20 15:02:56 hi kergoth Oct 20 15:03:24 03Cliff Brake  07org.openembedded.dev * r059a65f61c 10openembedded.git/contrib/weekly-changelog-report.py: weekly-changelog-report.py: remove "Merge Branch" entries, add TODO list Oct 20 15:08:57 hrw: do you mind if I write you privately ? Oct 20 15:15:10 cbrake: will you setup a cron job to get the weekly changelog to oe-devel? Oct 20 15:19:33 thesing: yes Oct 20 15:20:26 really cool. Oct 20 15:23:33 that newletter is great, good work Oct 20 15:28:43 I git cloned the org.openembedded.dev branch last week and did my first image and kernel compile(took a couple hours). now I should be using the "openembedded" branch, right? Can somehow convert my old clone to the new one without having to recompile? Forgive my git ineptness. Oct 20 15:30:02 cbrake: Thanks for bringing back the changelog! Oct 20 15:32:57 cbrake: it does help to people see what was done in a week period of time; it is impressive :-D Oct 20 15:33:31 cbrake: check python-git and python-bzutils stuff for weekly-changelog stuff Oct 20 15:33:46 bzutils? Oct 20 15:33:58 hrw: what bzutils does? Oct 20 15:34:13 python-bzutils - Python module to interact with bugzilla servers Oct 20 15:34:23 hrw: wow; great! :-D Oct 20 15:34:42 yeah, I'm hoping the changelog will be good for project reputation -- lets people know a lot is happening :-) Oct 20 15:34:45 so we can get list of fixed/opened bugs etc Oct 20 15:35:08 hrw: excellent -- I'll note that and try the next round Oct 20 15:36:01 I do not like days when I have to pay taxes... Oct 20 15:37:14 * NAiL usually gets taxes back Oct 20 15:38:55 NAiL: my company pay more then get back Oct 20 15:43:21 hrw & NAiL: you aren't the only one who hate taxes ;-) Oct 20 15:56:20 hmm.. ${STAGING_LIBDIR} is sh4-angstrom-linux/usr/lib/ but what is sh4-angstrom-linux/lib? Oct 20 16:00:36 anyone Oct 20 16:00:38 ? Oct 20 16:00:41 I can't find it in bitbake.conf Oct 20 16:01:04 Jin^eLD: and this is empty? Oct 20 16:01:17 hrw: what is emtpy? Oct 20 16:02:05 hrw: I have quite some stuff in staging/sh4-angstrom-linux/lib and it seems I need to pass the LDFLAGS for that to some obscure library, for which I wrote a .bb recipe Oct 20 16:02:35 but I was not able to find a variable for staging/sh4-angstrom-linux/lib, everything in bitbake.conf seems to be pointing to usr/lib Oct 20 16:04:10 which leads me to another question... if there is no ${STAGING...} variable for it, why is there so much stuff in there? (provided that I did not miss the variable when reading the .conf file) Oct 20 16:04:42 sh4-angstrom-linux/lib dir Oct 20 16:04:53 ah ok Oct 20 16:05:24 I'm talking about the dirs in work/staging/ Oct 20 16:05:25 libc stage itself into */lib Oct 20 16:05:29 yes Oct 20 16:05:48 and I need to pass that to my app (else it borks out not finding libpthread) Oct 20 16:05:48 and iirc no other lib should Oct 20 16:06:29 Jin^eLD: gcc should have it in default libdir Oct 20 16:06:36 hrw: there is a bunch of them there, but I guess libc related... http://pastebin.mozilla.org/558054 Oct 20 16:07:16 same here Oct 20 16:07:19 the /lib one has libpthread.so.0 -> libpthread-2.6.1.so while the /usr/lib one has libpthread.so Oct 20 16:07:31 and /usr/lib also has libpthread.a Oct 20 16:07:41 but the linker was only happy when it got the libpthread.so.0 from /lib Oct 20 16:08:41 I never was good at toolchain details Oct 20 16:09:23 me neither, I fed the path to /lib in my .bb recipe and that fixed it, but I was confused that there was no ${STAGING_xxxDIR} variable for it Oct 20 16:13:18 Jin^eLD: STAGING_DIR_TARGET="/home/hrw/devel/build/angstrom/tmp/staging/armv5te-angstrom-linux-gnueabi" Oct 20 16:13:28 bbl Oct 20 16:15:54 hrw|gone: not ${STAGING_DIR_HOST} ? but adding /lib in the end is anyway required Oct 20 18:14:45 greetings! is there any way to set a > or >= in an RDEPENDS? Oct 20 18:18:02 yes, I think it should work just as you'd expect Oct 20 18:18:12 RDEPENDS = "pkg (>= version)" Oct 20 18:18:26 pb_: oh very cool, I'll give it a try thanks Oct 20 18:38:13 Anyone going to the Embedded Systems Conference (ESC) in Boston next week? Oct 20 18:38:34 !me Oct 20 18:40:58 :) cool. Turns out I'll be there for a couple of days too. Oct 20 18:41:30 how can I replace a binary provided by busybox? seems like a really obvious thing to do but I can't find anything Oct 20 18:41:56 I know I can manually deleted the link, but I figure there must be a procedure Oct 20 18:42:29 Just install the real binary, it should replace the busybox symlink. Better behaved installers will use the alternatives mechanism, but I don't think that's completely implemented for busybox. Oct 20 18:44:06 mwester: ok thanks very much Oct 20 18:57:54 why does bitbake helloworld-image build the git-native package? Is it building a git client for the target system? Oct 20 18:59:28 morning Oct 20 18:59:39 jimsheldon: use update-alternatives system Oct 20 18:59:53 git-native is a git client for the host Oct 20 19:00:03 slothlove: "bitbake -g helloworld-image' will tell you what req git-native Oct 20 19:00:22 slothlove: some component used in build process is fetched from git so thats why git-ative is built Oct 20 19:00:24 probably it is needed to download some package from a git repo Oct 20 19:00:30 hrw: right, I'm actually trying to make the ifupdown package install, following update-alternatives Oct 20 19:01:01 hrw: so I'm changing the .bb file and using the wget .bb file as a guide Oct 20 19:01:20 why doesn't bitbake just use the git client I already have installed? Oct 20 19:01:52 as a rule bitbake tries to use as little as possibile from the host Oct 20 19:02:13 unless told to do otherwise Oct 20 19:03:28 has anybody tried rebasing their old gittrial repo to a new git one? Oct 20 19:03:39 hrw: so I get the error "update-alternatives: Error: cannot register alternative ifup to /usr/bin/ifup since it is already registered to /sbin/ifup", I assume I'm doing something stupid? Oct 20 19:04:07 hrw: I think I just answered my own question... Oct 20 19:04:11 morning ! Oct 20 19:04:23 khem: hi Oct 20 19:04:25 has anyone put together any documentation about which native packages are pretty safe to use from a normal ubuntu install? I'm looking for ways to reduce compile time Oct 20 19:04:44 jimsheldon: l:) Oct 20 19:05:09 slothlove: git, gcc3 are safe Oct 20 19:07:05 hrw: ${sbindir} != ${bindir} :) all set now Oct 20 19:09:19 hrw: thanks Oct 20 19:10:26 hi Oct 20 19:18:47 hi woglinde Oct 20 19:18:52 hi pb :) Oct 20 19:19:51 hi woglinde :) Oct 20 19:20:29 he kalev Oct 20 19:20:36 kalev please send ssh key to the masters Oct 20 19:20:49 oh, did you talk to them? Oct 20 19:21:47 thanks! Oct 20 19:21:55 I talked with zecke about the git stuff Oct 20 19:26:00 did he figure out what we did wrong with branches? Oct 20 19:26:17 hm seems I pushed wrong Oct 20 19:26:30 i has to use push kalev-tmp Oct 20 19:26:33 had Oct 20 19:27:46 ah Oct 20 19:28:39 I assumed that automagicly the brach I am in is use Oct 20 19:28:40 d Oct 20 19:30:25 makes sense, so you got the message "already up to date" because it tried to push your org.openembedded.dev branch instead Oct 20 19:30:44 yes Oct 20 19:31:04 denix: did u solve your issue ? Oct 20 19:31:43 huhu khem Oct 20 19:31:57 woglinde: wie gehts ! Oct 20 19:32:14 khem uclibc is working nice with hash-style Oct 20 19:32:25 I hope Oct 20 19:32:30 I only compiled sofar Oct 20 19:32:41 yes it should work Oct 20 19:32:48 it has worked well in part year or so Oct 20 19:33:04 s/part/past Oct 20 19:33:22 khem I saw another issue on uclibc in our repos Oct 20 19:33:30 we have ful debug support on Oct 20 19:33:39 which means uclibc is compiled with -o0 Oct 20 19:33:50 woglinde: hmm Oct 20 19:33:53 instead of -Os and the other features Oct 20 19:34:15 I want to turn it of Oct 20 19:34:20 if you dont mind Oct 20 19:34:47 you can better would be to have two configs one for debug and one for release Oct 20 19:35:14 its very handy for debugging especially ld.so Oct 20 19:35:18 khem we need a sane config merhing Oct 20 19:35:25 merging Oct 20 19:35:34 too Oct 20 19:35:43 we have machine arch distro ... Oct 20 19:35:51 hi thesing Oct 20 19:36:05 hi Oct 20 19:36:46 woglinde: I think release or debug is a distro feature imo Oct 20 19:40:10 woglinde: machine and arch could be one Oct 20 19:41:02 khem: sorry, stepped out Oct 20 19:41:52 khem: yeah, there was one change overlayed by me, which I didn't notice Oct 20 19:43:06 denix: cool nice what was it just curious Oct 20 19:43:44 woglinde: ok, key sent Oct 20 19:48:54 khem: Koen's staging changes Oct 20 19:49:55 khem: one of the changed files was overlayed in my setup... Oct 20 19:51:30 hmmm but that bumped up ABI version it should have barfed Oct 20 19:52:29 khem: when it should have barfed? Oct 20 19:53:43 during build Oct 20 19:54:26 but not when you do clean build though Oct 20 19:54:38 yeah Oct 20 20:05:13 * denix is away: I'm not here Oct 20 20:13:06 have a nice evening Oct 20 20:14:13 nite hrw Oct 20 20:30:33 hi likewise Oct 20 20:55:16 hi, Is there something like a minimal OS setup in packages? Oct 20 20:55:23 for x86 that is. Oct 20 20:55:36 bitbake minimal-image Oct 20 20:56:43 03Koen Kooi  07org.openembedded.dev * r0a274f5eb6 10openembedded.git/packages/dsplink/dsplink_1.51.00.08.bb: dsplink: disable 128MB patch in 1.51, it doesn't apply Oct 20 20:57:30 hm don't have that. Oct 20 20:59:57 I used git clone git://git.openembedded.net/openembedded to check the tree out. Oct 20 21:00:28 yeah Oct 20 21:00:29 and $ git pull --rebase to update. Oct 20 21:01:01 Follow closely with the instructions on wiki. Oct 20 21:01:11 openembedded/packages/images/minimal-image.bb Oct 20 21:01:17 there it is Oct 20 21:01:31 and if you run bitbake minimal-image Oct 20 21:01:34 it will be build Oct 20 21:04:59 why is it for other packages, it is /prefix/packages/blah/blah.bb ? Oct 20 21:07:59 I do not under stand your question Oct 20 21:08:21 i think he wants to know why all the image bb files are in images/ instead of their own directories Oct 20 21:10:07 Yeah that is what I meant. :) Oct 20 21:10:26 because noby teached this to bitbake Oct 20 21:10:56 hm args Oct 20 21:10:58 sorry Oct 20 21:11:04 because It makes no sense Oct 20 21:11:13 images is the place for all images Oct 20 21:11:31 you do not need a patch or something else in the images recipe Oct 20 21:20:19 logically, images and tasks should not be in packages directory... Oct 20 21:20:27 denix: heh, i was about to say that :) Oct 20 21:20:31 conceptually anyway Oct 20 21:20:57 should really have been named recipes/ instead of packages/, since the term 'package' is way too overloaded Oct 20 21:21:00 ah well Oct 20 21:21:05 historical reasons aside, any progress on a new structure? http://wiki.openembedded.net/index.php/Structure_of_.dev_branch Oct 20 21:23:01 what's proposed in that wiki page was specifically decided against when originally structuring the repo, because such groupings are pretty arbitrary. who decides what goes in core? Oct 20 21:23:17 maybe people are moving back in that direction, i'm not involved anymore, just giving a bit of historical info :P Oct 20 21:24:01 some of those items don't make much sense, but others do - toolchain, for example Oct 20 21:24:03 hmmm. Oct 20 21:24:10 17:23:12 ] wweng /WorkSpace/oe/stuff $ bitbake minimal-image Oct 20 21:24:12 ERROR: Unable to open conf/bitbake.conf Oct 20 21:24:26 But the wiki told me to edit local.conf Oct 20 21:24:32 I'd say images and tasks should be decoupled from packages first :) Oct 20 21:27:21 hubar seems that you did not edit local.conf correct Oct 20 21:27:42 best way is to put your local.conf outside the oe tree Oct 20 21:29:57 hubar: what version of bitbake do you have? Oct 20 21:30:16 1.8? Oct 20 21:30:37 1.8.10 Oct 20 21:31:11 ok. do you have BBPATH set? Oct 20 21:33:12 kergoth: do you plan to get more involved in oe again? Oct 20 21:33:25 Yes. Oct 20 21:34:05 [ 17:30:30 ] wweng /WorkSpace/oe/stuff/build $ echo $BBPATH Oct 20 21:34:06 /WorkSpace/oe/stuff/build Oct 20 21:34:34 I start to think maybe it is because my BBPATH didn't end with a slash... Oct 20 21:34:45 it should be stufff/openembedded added Oct 20 21:35:09 otherwise bitbake can't find you oe checkout Oct 20 21:35:46 What do you mean? Oct 20 21:36:18 I have my OE checked out in /WorkSpace/oe/stuff/openembedded/ Oct 20 21:36:39 hubar: BBPATH=/WorkSpace/oe/stuff/openembedded/ Oct 20 21:37:11 hmm then how does bitbake know where to find the configuration file? Oct 20 21:37:33 you need to have both in ther. Oct 20 21:37:36 if you have your build directory, then BBPATH=/WorkSpace/oe/stuff/build:/WorkSpace/oe/stuff/openembedded Oct 20 21:37:53 BBPATH=/home/tkunze/devel/zaurus/build:/home/tkunze/devel/zaurus/openembedded Oct 20 21:39:50 kergoth: do you plan to get more involved in oe again? Oct 20 21:40:05 * denix seconds that question Oct 20 21:40:30 hi... i saw the restructure page Oct 20 21:40:30 [ 17:39:57 ] wweng /WorkSpace/oe/stuff/build $ bitbake minimal-image Oct 20 21:40:32 ERROR: Please set the 'CACHE' variable. Oct 20 21:40:47 restructuring is good btu with that layout you still have most stuff in packages/ Oct 20 21:41:23 [ 17:40:54 ] wweng /WorkSpace/oe/stuff/build $ echo $BBPATH Oct 20 21:41:25 /WorkSpace/oe/stuff/build:/WorkSpace/oe/stuff/openembedded Oct 20 21:41:30 if stuff is going to get moved around significantly proper categorisation is probably the only way to make a difference? Oct 20 21:41:48 my tuppence... :) Oct 20 21:46:21 Do I have to set CACHE? Oct 20 21:47:07 # Uncomment this to disable the parse cache (not recommended). Oct 20 21:47:08 # CACHE = "" Oct 20 21:47:30 It already says it is not recommended... Oct 20 21:48:38 And is it ok if I just comment out the last line, not remove it? Oct 20 21:48:41 # EDIT THIS FILE and then remove the line below before using! Oct 20 21:48:42 # REMOVE_THIS_LINE:="${@bb.fatal('Read the comments in your conf/local.conf')}" Oct 20 21:49:21 should be Oct 20 21:49:37 i'm not sure about the cache issue Oct 20 21:49:47 if it's not recommended, what would make you think you'd have to do it? Oct 20 21:49:54 of course not, leave it alone unless you're told otherwise Oct 20 21:50:37 hubar: have you set machine/distro? Oct 20 21:51:21 MACHINE = "x86" Oct 20 21:51:38 DISTRO = "minimal" Oct 20 21:53:04 hmm, I haven't tried minimal distro myself... Oct 20 21:53:46 do you have /WorkSpace/oe/stuff/openembedded/conf/bitbake.conf? Oct 20 21:55:12 [ 17:54:44 ] wweng /WorkSpace/oe/stuff $ file /WorkSpace/oe/stuff/openembedded/conf/bitbake.conf Oct 20 21:55:14 /WorkSpace/oe/stuff/openembedded/conf/bitbake.conf: ASCII English text, with very long lines Oct 20 21:58:16 ok, what do you have in your build directory? Oct 20 21:59:38 [ 17:58:48 ] wweng /WorkSpace/oe/stuff/build $ ls -ad * Oct 20 21:59:39 conf/ Oct 20 21:59:40 [ 17:59:02 ] wweng /WorkSpace/oe/stuff/build $ cd conf Oct 20 21:59:42 [ 17:59:13 ] wweng «Space/oe/stuff/build/conf $ ls Oct 20 21:59:44 bitbake.conf local.conf local.conf~ Oct 20 22:00:08 hmm, why do you need your own bitbake.conf? Oct 20 22:00:11 the conf/bitbake.conf is the same file as conf/local.conf Oct 20 22:00:38 WoW Oct 20 22:00:38 ???? Oct 20 22:00:39 if it's the same, that's your problem, I think Oct 20 22:00:47 YUP YUP! Oct 20 22:00:49 local.conf is not bitbake.conf Oct 20 22:00:58 It starts running now. :) Oct 20 22:00:58 remove it or move away Oct 20 22:01:18 bitbake.conf is inside oe tree Oct 20 22:01:36 local.conf.sample is the only file to put outside as local.conf Oct 20 22:01:50 hubar i386 build can have some glitches Oct 20 22:02:01 woglinde: not necessary, but in his case that's true Oct 20 22:02:06 especially with minimal Oct 20 22:02:12 you mean minimal distro? or machine x86? Oct 20 22:02:17 x86 Oct 20 22:02:36 if not works try the angstroem distribution Oct 20 22:02:49 and bitbake minimal-image Oct 20 22:03:34 ok! Oct 20 22:03:54 How do I remove these built trees? Oct 20 22:04:04 Just rm -rf WorkSpace? Oct 20 22:04:47 * florian is a hero - he always finds a better way to make oe produce bullshit Oct 20 22:04:49 I mean they are not somehow installed as packages in the system, right? Oct 20 22:05:07 he florian Oct 20 22:05:10 florian: ))) Oct 20 22:05:14 liblibunicapgtk_0.9.3-r0 Oct 20 22:05:35 how does something like this happen? Oct 20 22:05:44 hm Oct 20 22:05:44 hubar: nope, everything is within your /WorkSpace/oe/stuff/ Oct 20 22:05:48 wworng .bb Oct 20 22:06:07 maybe at PACKAGES *= line Oct 20 22:06:10 ups + Oct 20 22:06:58 well - one bb, tree library packages Oct 20 22:07:41 two named correct and one wrong Oct 20 22:08:19 * florian smells a bug Oct 20 22:11:24 I suppose the renamer gets confused with libunicap and libunicapgtk Oct 20 22:22:58 hmm. Oct 20 22:23:17 What do I do if it times out in the middle of getting a package? Oct 20 22:23:21 18:21:05 (11.63 KB/s) - Data connection: Connection timed out; Oct 20 22:23:47 NOTE: fetch ftp://ftp.gnu.org/gnu/gettext/gettext-0.17.tar.gz Oct 20 22:23:55 Only got 98%. Oct 20 22:29:17 Should I restart? Oct 20 23:07:00 03Koen Kooi  07org.openembedded.dev * r96198fd96b 10openembedded.git/packages/mozilla/fennec_hg.bb: fennec: update to 1.0a1 Oct 20 23:14:10 * denix is away: I'm not here Oct 20 23:42:45 Anyone here??? Oct 20 23:56:32 I guess people all call it a day already. **** ENDING LOGGING AT Tue Oct 21 02:59:57 2008