**** BEGIN LOGGING AT Thu Jun 02 02:59:56 2011 **** ENDING LOGGING AT Thu Jun 02 03:41:39 2011 **** BEGIN LOGGING AT Thu Jun 02 03:42:14 2011 Jun 02 05:06:36 http://pastebin.ca/2073533 Jun 02 05:06:40 anyone seen this? Jun 02 05:06:47 bitbake tanked Jun 02 05:10:00 hmph, i think its gcc related Jun 02 05:10:05 * CcSsNET keeps looking Jun 02 05:15:26 hmm maybe glibc related Jun 02 05:17:51 meh couldnt figure it out Jun 02 06:01:04 nvm its a python bug Jun 02 06:05:03 hello... Jun 02 06:07:25 can anyone help me..... while compiling kernel.... from where and how that pengiun logo is inserted from the kernel source into the kernel image? Jun 02 06:07:47 linux.inc depending on logo in SRC_URI Jun 02 06:09:42 JaMa, I didn't get you? Jun 02 06:14:57 sanket: # logo support, if you supply logo_linux_clut224.ppm in SRC_URI, then it's going to be used Jun 02 06:15:30 JaMa, yes ... but from where and how that logo image is generated.... Jun 02 06:16:53 no I didn't get you... do you need to know how .ppm format is edited and created? Jun 02 06:19:46 JaMa, yes.... Jun 02 06:20:07 and how is this question related to OE? Jun 02 06:22:16 I had builded a filesystem using OE... Jun 02 08:25:58 morning all Jun 02 08:29:33 morning, bluelightning Jun 02 08:29:39 hi anarsoul Jun 02 08:30:08 anarsoul: good news, after switching mmc to PIO the h1940 booted :) Jun 02 08:30:17 Hi everybody Jun 02 08:30:37 I am new at oe and trying to discover Jun 02 08:30:46 anarsoul: there's still a touchscreen issue (probably just missing device) to sort out, plus for some reason I couldn't get usbnet to work which made debugging hard Jun 02 08:31:09 I have a question, I will appreciate if anybody help me Jun 02 08:31:24 bluelightning: hmm, usbnet should work nicely Jun 02 08:32:01 I want to use different version packages, how can I select specific version of package in a bb? Jun 02 08:32:22 I did some search and found PREFERRED_VERSION which set in distro configurations Jun 02 08:32:40 is there any other way for packet version selection? Jun 02 08:32:42 bluelightning: touchscreen probably requires fake pressure patch Jun 02 08:32:51 bluelightning: but then it's tslib issue Jun 02 08:33:04 I dont want to create a new distro config file for packet version selecting Jun 02 08:33:39 kays_: you can set PREFERRED_VERSION_xxxx in your local.conf Jun 02 08:34:27 should I add to all my preferred version into local.conf? Jun 02 08:34:56 Can I add this into bb file which I define dependency Jun 02 08:35:17 for example I need to use alsa-lib.xxx for alsa-utils Jun 02 08:35:36 in alsa-utils.bb I define alsa-lib as DEPENDS Jun 02 08:35:59 Can I also define PREFERRED_VERSION_alsa-lib = "xxx" in alsa-utils.bb Jun 02 08:36:00 ? Jun 02 08:36:20 or should I define all preferred version in configuration files? Jun 02 08:40:33 kays_: no you can't specify it for dependencies in recipes Jun 02 08:40:53 you can specify it in a distro conf file or your local.conf Jun 02 08:41:06 hm ok Jun 02 08:41:25 then what about undefined PREFERRED_VERSION packages Jun 02 08:41:45 they take the latest version (apart from DEFAULT_PREFERENCE) Jun 02 08:41:58 hm ok Jun 02 08:42:09 thanks in advance Jun 02 08:42:26 often when new versions of recipes are added, a DEFAULT_PREFERENCE = "-1" is added so that the new (and perhaps untested) version doesn't break things for other people Jun 02 08:42:39 PREFERRED_VERSION_xxxx will override DEFAULT_PREFERENCE though Jun 02 08:42:51 hmm ok, I got it Jun 02 10:11:37 hi zecke Jun 02 10:11:45 hi pb Jun 02 10:11:49 hi woglinde Jun 02 10:11:51 nice holiday Jun 02 10:51:10 jo zecke Jun 02 11:40:10 doh, libc6-dev getting installed on target system again Jun 02 11:40:12 * pb_ stabs eglibc Jun 02 11:42:19 uhm? Jun 02 11:42:31 I know why I use uclibc Jun 02 11:42:46 and why I even factored out backtrace Jun 02 11:46:21 hi gnutoo Jun 02 11:47:17 hi woglinde Jun 02 12:48:35 hi all Jun 02 12:48:57 open embedded changed the file structure of cross compiling recently eh? Jun 02 12:49:21 does it no longer package up a tarball? Jun 02 12:49:33 of the root filesystem Jun 02 12:57:04 CoSsNET: you can add tar or tar.gz to IMAGE_FSTYPES at your local.conf Jun 02 13:05:02 o Jun 02 13:05:05 :) Jun 02 13:05:28 to late now dont feel like dealing with a new compile haha Jun 02 13:05:41 i can just tarball it myselfi guess Jun 02 13:14:20 03Paul Menzel  07master * rc55a2839e1 10openembedded.git/recipes/lynx/lynx_2.8.7.bb: (log message trimmed) Jun 02 13:14:20 lynx: inherit class `gettext` to fix build with `minimal-uclibc` Jun 02 13:14:20 Task configure failed with the following error message using `minimal-uclibc` for `MACHINE = "beagleboard"`. Jun 02 13:14:20 checking for __argz_stringify... no Jun 02 13:14:20 checking for __argz_next... no Jun 02 13:14:21 checking for iconv... no, consider installing GNU libiconv Jun 02 13:14:22 checking for nl_langinfo and CODESET... yes Jun 02 13:18:07 CcSsNET: if you already build the image it will not compile everything again if you build with another file system afterwards Jun 02 13:18:48 ok Jun 02 13:34:00 so does open embedded not make a /dev? Jun 02 13:34:14 its no big deal i guess i can try and chroot? lol Jun 02 13:34:32 o right wrong arch... Jun 02 13:42:11 trying again this is getting odd Jun 02 16:30:02 pb_: updating my oe-core trees Jun 02 16:30:15 but I would swear up and down that as of 3.7.5 the Makefile called tclsh Jun 02 16:31:15 kergoth: around ? Jun 02 16:43:00 Also, uh, what? sqlite3-native-3.6.23.1-r4.0 is being used here for psuedo Jun 02 16:43:03 not 3.7.x Jun 02 16:43:16 er, no, nm me, too many builds at once Jun 02 16:49:38 yes 3.6 is still de default Jun 02 16:49:42 dont know why Jun 02 16:51:07 in oe-core? Jun 02 16:52:10 ah Jun 02 16:52:13 oe-core Jun 02 17:28:20 uh wtf gettext uses git Jun 02 17:30:33 jo crofton Jun 02 17:31:19 hi Jun 02 17:57:12 03Tom Rini  07master * r28cb4be9d2 10openembedded.git/recipes/lmbench/lmbench_3.0-a9.bb: Jun 02 17:57:13 lmbench: Add missing end-quote Jun 02 17:57:13 Noticed by Paul Menzel. Jun 02 17:57:13 Signed-off-by: Tom Rini Jun 02 18:09:25 would people be interested in a version of the meta-toolchain env setup script which doesn't hardcode its location, and therefore is relocatable? Jun 02 18:09:32 (naturally just one piece of it, but..) Jun 02 18:32:56 perl 5.10.1 build fails with latest OE Jun 02 18:33:06 problem also described here: http://permalink.gmane.org/gmane.comp.handhelds.openembedded/45640 Jun 02 18:33:10 Anyone familiar with this? Jun 02 18:35:36 yes pick the patch from oe-core Jun 02 18:45:49 woglinde_, not sure what the difference between the oe and oe-core? and why does it exist? Jun 02 18:46:11 woglinde_, if you could point to some documentation regarding that, would appreciate Jun 02 18:49:57 joelagnel: http://www.yoctoproject.org/projects/openembedded-core is one page about it Jun 02 18:50:07 it's basically a much more strict, controlled subset of oe, based on poky Jun 02 18:50:13 which we then layer more metadata on Jun 02 18:50:30 kergoth, thanks. why not just have both together? Jun 02 18:50:36 so the oe repository is being broken up into layers, by distro/bsp, on top of oe-core Jun 02 18:50:41 what do you mean? Jun 02 18:50:49 breaking it up is the point Jun 02 18:51:00 oh Jun 02 18:51:17 we get more fine grained control to the layers, avoid parsing crap people don't need, makes it much easier to pick and choose the bits you need for what you're doing Jun 02 18:51:59 in the process, we also ditch some old things that nobody needs anymore, fix things up as they get migrated Jun 02 18:52:21 oe-core also has a superior mechanism for handling prebuilt recipes Jun 02 18:52:26 kergoth, those fixes are ported back to oe? Jun 02 18:52:28 and tighter control on licenses Jun 02 18:52:37 some things are in both, some are in one and not the other Jun 02 18:52:55 kergoth, how do I know which one to use when Jun 02 18:52:56 in the very long term, the oe repo will go read only and bits will be pulled over into oe-core and meta-oe on an as needed basis Jun 02 18:53:38 oe-core is still in development, it's in progress, not everyone is using it yet Jun 02 18:53:50 if you want to try it out, you're welcome to, but oe is still where most people are Jun 02 18:54:06 so it is indeed worth pulling certain bugfixes from oe-core to oe for the time being Jun 02 18:54:54 kergoth, ah ok I see :) Jun 02 18:55:37 so I should look in oe-core for whether the perl recipe there doesn't break? Jun 02 18:56:36 I guess I should, nm. Jun 02 18:59:25 03Paul Menzel  07master * r0fe0f7d562 10openembedded.git/recipes/gammu/gammu_1.29.0.bb: (log message trimmed) Jun 02 18:59:26 gammu: add `curl` to `DEPENDS` Jun 02 18:59:26 Task compile failed with the following error message [1]. Jun 02 18:59:26 Linking C executable gammu Jun 02 18:59:26 /oe/build-angstrom-next/angstrom-dev/sysroots/i686-linux/usr/lib/libcurl.so: could not read symbols: File in wrong format Jun 02 18:59:26 collect2: ld returned 1 exit status Jun 02 18:59:27 make[3]: *** [gammu/gammu] Error 1 Jun 02 19:00:51 huh Jun 02 19:01:23 CIA-8, I guess pastbin would be more pleasant Jun 02 19:01:34 *pastebin Jun 02 19:01:35 joelagnel: that's a bot Jun 02 19:01:46 really Jun 02 19:01:56 all commits are broadcast in the channel Jun 02 19:02:22 ok :-) Jun 02 19:03:07 joelagnel: http://cia.vc/ Jun 02 19:03:52 ah Jun 02 19:14:24 03clarson  07master * rceaae812da 10openembedded.git/classes/sourceipk.bbclass: Jun 02 19:14:24 sourceipk: deal with compileconfigs only when it exists Jun 02 19:14:24 Signed-off-by: Chris Larson Jun 02 19:14:24 git-svn-id: http://svr-alm-svn-01.alm.mentorg.com:8080/svn/ea/easi_oe/system-builder/release-2011.1/sb-core@4918 6a149ff2-cb3c-0410-b4c8-c88968701766 Jun 02 19:14:28 03clarson  07master * rec05290d0a 10openembedded.git/classes/sourceipk.bbclass: Jun 02 19:14:28 sourceipk: use tar to get the files copied, rather than cp Jun 02 19:14:28 Signed-off-by: Chris Larson Jun 02 19:14:28 git-svn-id: http://svr-alm-svn-01.alm.mentorg.com:8080/svn/ea/easi_oe/system-builder/release-2011.1/sb-core@4917 6a149ff2-cb3c-0410-b4c8-c88968701766 Jun 02 19:14:29 03clarson  07master * r944b7e1bb7 10openembedded.git/classes/sourceipk.bbclass: Jun 02 19:14:29 sourceipk: allow override of the section Jun 02 19:14:29 Signed-off-by: Chris Larson Jun 02 19:14:30 git-svn-id: http://svr-alm-svn-01.alm.mentorg.com:8080/svn/ea/easi_oe/system-builder/release-2011.1/sb-core@4916 6a149ff2-cb3c-0410-b4c8-c88968701766 Jun 02 19:14:31 03clarson  07master * r1ffd6fce9a 10openembedded.git/classes/sourceipk.bbclass: Jun 02 19:14:31 sourceipk: exclude temp dirs when S == WORKDIR Jun 02 19:14:31 Signed-off-by: Chris Larson Jun 02 19:14:34 damnit Jun 02 19:14:37 * kergoth mutters Jun 02 19:14:40 git-svn-id: http://svr-alm-svn-01.alm.mentorg.com:8080/svn/ea/easi_oe/system-builder/release-2011.1/sb-core@4922 6a149ff2-cb3c-0410-b4c8-c88968701766 Jun 02 19:14:41 cbrake: ping Jun 02 19:14:42 ? Jun 02 19:14:53 kergoth: yes Jun 02 19:14:53 what is that? Jun 02 19:14:56 forgot a rebase -i to fix that up Jun 02 19:15:03 cbrake: can you revert master back before these commits, please? Jun 02 19:15:12 cbrake: to 0fe0f7d5629578839192153c012a05a192dc2b41 specifically Jun 02 19:15:27 kergoth: openembedded? Jun 02 19:15:33 cbrake: yeah Jun 02 19:19:36 cbrake: hah, i was just doing what you just did to fix it (the gitolite.conf change). that works. thanks. didn't think about it at first Jun 02 19:21:02 kergoth: ok, I think its fixed now -- can you verify? Jun 02 19:21:21 yep, its good Jun 02 19:21:23 thanks, appreciatei t Jun 02 19:21:24 kergoth: I'll probably revert my gitolite.conf change to prevent mishaps :-) Jun 02 19:21:35 * kergoth nods, thats what he was thinking of doing too Jun 02 19:21:38 cool Jun 02 19:21:58 lets try this again.. Jun 02 19:22:04 03Chris Larson  07master * r17fd4202c0 10openembedded.git/classes/sourceipk.bbclass: Jun 02 19:22:04 sourceipk: allow override of the section Jun 02 19:22:04 Signed-off-by: Chris Larson Jun 02 19:22:14 03Chris Larson  07master * raca8df4471 10openembedded.git/classes/sourceipk.bbclass: Jun 02 19:22:14 sourceipk: use tar to get the files copied, rather than cp Jun 02 19:22:14 Signed-off-by: Chris Larson Jun 02 19:22:15 03Chris Larson  07master * r964f73b2ba 10openembedded.git/classes/sourceipk.bbclass: Jun 02 19:22:15 sourceipk: deal with compileconfigs only when it exists Jun 02 19:22:15 Signed-off-by: Chris Larson Jun 02 19:22:16 03Chris Larson  07master * rd85c4522b3 10openembedded.git/classes/sourceipk.bbclass: Jun 02 19:22:16 sourceipk: exclude temp dirs when S == WORKDIR Jun 02 19:22:17 Signed-off-by: Chris Larson Jun 02 19:22:26 whew. Jun 02 19:22:27 hmm Jun 02 19:27:36 hmm indeed. Jun 02 19:28:59 I need to figure out if there's a way to run the package configuration steps during the build process instead of the device's first boot. Jun 02 19:33:52 mewyn: that's always done when possible Jun 02 19:34:06 not everything can be done at image creation time (e.g. things that run the target binaries) Jun 02 19:34:14 see the pkg_postinst's in teh recipes Jun 02 19:34:27 hm Jun 02 19:34:35 if $D is defined when the postinst is run, then its image construction time, not first boot time Jun 02 19:34:45 Also, how've you been, kergoth? :) It's been a long time since I've been in here. Jun 02 19:34:47 so if something needs to run at first boot time, it checks that and exists if non-empty Jun 02 19:37:10 I'll take a look at what's going on in the packages in question. Maybe I'll have to produce sane defaults in a seperate pacakage. Jun 02 19:38:19 yeah, htat might be a decent route. or just use a local layer and bbappend to replace the postinst and add to SRC_URI / whatever, to drop in your versions Jun 02 19:38:26 would avoid issues with packages stepping on one another that way Jun 02 19:38:41 *nod* Jun 02 19:39:06 been well, busy busy, working on oe-based products first at montavista and now at mentor graphics Jun 02 19:39:07 you? Jun 02 19:39:42 I've gone back to school at Univ. of Illinois and for the summer I'm working at HP/Palm. Jun 02 19:39:49 ah, nice Jun 02 19:40:19 Trying to pick up my old ArcadiaOS project too. Jun 02 19:41:16 Which is supposed to be an OE-based distro aimed at my arcade cabinet. Jun 02 19:49:05 pidgin busted? Jun 02 19:53:24 tcl zapp Jun 02 21:35:22 03Chris Larson  07master * r17a64df7e9 10bitbake.git/lib/bb/exceptions.py: Jun 02 21:35:22 bb.exceptions: add to_string convenience function Jun 02 21:35:22 Signed-off-by: Chris Larson Jun 02 21:35:25 03Otavio Salvador  07master * r0db267b948 10bitbake.git/lib/bb/codeparser.py: Jun 02 21:35:26 codeparser.py: fix syntax error in exception handling Jun 02 21:35:26 Commit 036cf3cd11b3a6836b77f5ffa760ceee6b71b1ef missed the needed Jun 02 21:35:26 brackets to handle more then a type of exception. Jun 02 21:35:26 Signed-off-by: Otavio Salvador Jun 02 21:35:26 03Chris Larson  07master * rdbf405f1f7 10bitbake.git/lib/bb/exceptions.py: (log message trimmed) Jun 02 21:35:27 bb.exceptions: don't show a repr of 'self' Jun 02 21:35:27 Rather than treating self like an ordinary argument, showing a repr of its Jun 02 21:35:28 value in the function spec when formatting the traceback entry, now we show Jun 02 21:35:32 the class name for the method as a part of the function name. Example: Jun 02 21:35:33 Old: bar(self=, f=5) Jun 02 21:35:33 New: Fooclass.bar(f=5) Jun 02 21:35:33 03Richard Purdie  07master * r08f53145c5 10bitbake.git/lib/bb/codeparser.py: Jun 02 21:35:33 codeparser.py: Ignore incomplete cache files Jun 02 21:35:33 Signed-off-by: Richard Purdie Jun 02 21:35:34 03Chris Larson  07master * r8c33f50eb6 10bitbake.git/lib/bb/cooker.py: Jun 02 21:35:34 cooker: pass traceback back from parsing thread Jun 02 21:35:35 Uses bb.exceptions to get a traceback back from the parsing thread to the main Jun 02 21:35:35 thread, where it is then formatted. Also enables 3 lines of context for the Jun 02 21:35:36 formatted traceback, and limits the number of entries displayed to 5. Jun 02 21:35:36 Signed-off-by: Chris Larson Jun 02 21:35:41 whee Jun 02 21:35:46 hrm Jun 02 21:42:45 kergoth: sounds good Jun 02 21:47:27 03Chris Larson  07master * r12aef37618 10bitbake.git/lib/bb/cooker.py: Jun 02 21:47:27 cooker: handle ExpansionError the same way we do ParseError Jun 02 21:47:27 Signed-off-by: Chris Larson Jun 02 21:47:38 03Chris Larson  07master * rec8a5a495b 10bitbake.git/lib/bb/siggen.py: (log message trimmed) Jun 02 21:47:38 siggen: don't choke with traceback when data is None Jun 02 21:47:38 Given we use bb.error, not bb.fatal, here, it seems this was intended to be Jun 02 21:47:38 non-fatal, yet we'd end up trying to concatenate None. Fix this by setting an Jun 02 21:47:39 empty task to the empty string, for the purposes of hashing. Also str() the Jun 02 21:47:39 value we get from the datastore, just in case something other than a string Jun 02 21:47:40 was stored there. Jun 02 21:47:40 03Chris Larson  07master * r8e1be0ca41 10bitbake.git/lib/bb/parse/ast.py: Jun 02 21:47:41 parse: pass a useful (if mangled) key to bb.data.expand for := Jun 02 21:47:41 Signed-off-by: Chris Larson Jun 02 21:47:58 kergoth: ill' test very soon above changes... Jun 02 21:48:15 whew, we're slowly but surely improving the exception/traceback situation Jun 02 21:48:16 advanced versioning like 1.5+1.6-rc1 breaks vercmp Jun 02 21:48:31 so I'll trigger some nasty output Jun 02 21:48:37 we need to stop having the two version comparison functions still at some point Jun 02 21:48:40 :\ Jun 02 21:48:58 I can live with our own hack but... Jun 02 22:50:59 hi, is there anyone left here Jun 02 22:51:04 I still have questions on udev Jun 02 22:51:13 now I should add my own udev rules Jun 02 22:51:20 but that's for udev 171 Jun 02 22:51:29 can I pin udev 168 for a machine Jun 02 22:51:30 ? Jun 02 22:51:39 like put that in machine or distro config Jun 02 22:51:50 since udev is already machine arch for that machine that could be done Jun 02 22:51:53 right? Jun 02 22:52:36 machine arch is independent. there's no reason you can't set preferred version wherever you want to Jun 02 22:52:38 bitbake doesn't give a crap Jun 02 22:53:02 I'll rephrase then Jun 02 22:53:08 is it acceptable to do that? Jun 02 22:53:23 if so sould it go in distro pinning or in machine config? Jun 02 22:58:07 acceptable for people not for bitbake Jun 03 00:12:18 kergoth: IIRC you once looked into a canadian cross for win32? Jun 03 00:14:49 I'm getting a failure building eglibc-inital in oe-core.. anyone else having that? Jun 03 00:15:32 fray: related to garbage in a linker script? Jun 03 00:15:36 problem is in installing gnu/lib-names.h Jun 03 00:15:49 | /usr/bin/install: cannot stat `/home/mhatle/git/oss/oe-core/build/tmp-eglibc/work/i586-oe-linux/eglibc-initial-2.13-r1+svnr13356/build-i586-oe-linux/gnu/lib-names.h': No such file or directory Jun 03 00:16:48 never saw that one Jun 03 00:18:08 ya, I haven't either.. Jun 03 00:18:15 I'm going to disable parallel builds and see if it resolves it Jun 03 00:20:47 nope same error.. just occurs slower.. :( Jun 03 00:34:34 trying 2.12 lets see if this works Jun 03 01:07:51 just got hte same failure on 2_12.. ugh Jun 03 01:08:22 I suspect it's something else then Jun 03 02:44:02 hiya HopsNBarley Jun 03 02:44:15 I'll be up your way next thursday... Jun 03 02:44:32 hiya ka6sox-farfarawa !! Jun 03 02:44:41 same bat-number - just give a ring **** ENDING LOGGING AT Fri Jun 03 02:59:57 2011