**** BEGIN LOGGING AT Wed Oct 29 02:59:57 2008 Oct 29 03:21:35 what't the best MACHINE for PentiumIII? Oct 29 03:59:13 anybody done any x86 builds lately? Oct 29 05:26:50 Quick question -- is there a way to just run an incremental make in a given package's tree w/ all environment variables set from bitbake? I'm trying to figure out how best to do a quick-turnaround test cycle on a single build product out of a very large tree. Oct 29 06:03:37 lo all Oct 29 07:10:07 hey wich brach i need select in the git repo to htcuniversal? Oct 29 07:25:49 Hey Oct 29 07:26:04 lo Oct 29 07:26:52 srcrev ?= 5, what does the ?= operator do? Only set it, if not defined yet? Oct 29 08:02:03 yes, exactly Oct 29 08:02:38 mDuff: see devshell, or just do "bitbake -f -c compile" Oct 29 08:54:45 bonjour Oct 29 09:36:35 good morning Oct 29 09:41:53 hi ant Oct 29 09:42:33 hallo woglinde Oct 29 09:45:29 ok Oct 29 09:45:47 so... where on earth is info on oe's git dev repo Oct 29 09:45:51 ie how to check out and commit Oct 29 09:53:34 raster: I suggest to start here http://wiki.openembedded.net/index.php/GitPhraseBook Oct 29 09:53:50 raster: and follow there: http://git.or.cz/gitwiki/GitCheatSheet Oct 29 09:53:50 oooh Oct 29 09:54:05 i thoguht that'd just be a handy set of git magic incantations Oct 29 09:54:16 not the actual direct repo info Oct 29 09:57:23 aha Oct 29 09:57:34 as suspected.... oe.net is being... slow Oct 29 09:57:53 in fact... not even managing a connect :( Oct 29 09:57:56 poobuttox Oct 29 09:58:31 aaah now Oct 29 09:58:36 raster: probably you want to do something like this Oct 29 09:58:47 git clone git://git.openembedded.net/openembedded org.openembedded.dev Oct 29 09:58:56 already had that Oct 29 09:59:06 need git clone git@git.openembedded.net:openembedded Oct 29 09:59:16 git checkout -b my-branch origin/org.openembedded.dev Oct 29 09:59:23 i tried to git-remote rm origin.. and that just screwed up Oct 29 09:59:42 so i'm starting my tree afresh and will hand-import my patches Oct 29 09:59:54 raster: wait: Oct 29 09:59:56 git config user.name "Your Name" Oct 29 10:00:01 git config user.email "you@name" Oct 29 10:00:27 and you can easily commit on your local branch Oct 29 10:00:29 git commit -a Oct 29 10:00:37 there on the wiki :) Oct 29 10:01:16 been doing that Oct 29 10:01:21 on my "public" git tree Oct 29 10:01:33 this way, you can daily 'git pull --rebase' without loosing your changes Oct 29 10:01:35 i'm just growing a forest of patches that aren't making their way upstream to oe Oct 29 10:10:32 * ant_work did a bitbake -v -v -v and a found a lot of missing PREFERRED_PROVIDER and unbuildable dependencies... oebug 4784 Oct 29 10:11:38 basically for each python-xy runtime Oct 29 10:45:50 urg, day 3 with under 5 hours of sleep Oct 29 10:46:58 Crofton: your getting old man Oct 29 10:47:03 Crofton: get in more beers Oct 29 10:47:33 um beer Oct 29 10:47:39 or rum Oct 29 10:47:44 Baileys Oct 29 10:48:10 or nuka cola! Oct 29 10:48:21 Rum + Energy Drink :-D Oct 29 10:50:45 Paris for President Oct 29 10:50:52 Starbucks Oct 29 10:52:17 heh, would make a change for all presidential broadcasts to be done using night vision :-D Oct 29 10:52:33 paint the whitehouse pink Oct 29 10:53:20 Paris is the same level as bush Oct 29 10:53:29 i think it cannot be worse Oct 29 10:53:34 :) Oct 29 10:53:37 they both are follow by a dog Oct 29 10:53:38 heh, I was going to say scary thing is, she probably got +1 braincell on bush Oct 29 10:53:44 Paris looks better in a bikini Oct 29 10:53:50 and we are all very tired Oct 29 10:53:57 XorA : so it's the double ! Oct 29 10:54:13 some Canadians where gloating about their election happening in about 6 weeks Oct 29 10:54:16 at least Paris would have more success at negotiations :-) Oct 29 10:54:21 right Oct 29 10:54:31 yeap , good argues =) Oct 29 10:54:38 as long as she keeps her mouth closed anyway Oct 29 10:54:42 oh well, I need to crawl into a shower Oct 29 10:54:56 be sure to look for her latest campaign video Oct 29 10:58:40 :-D Oct 29 11:12:58 hmm, point for discussion out pref-ver-xorg files contain lots of app versions which are no longer part of an xorg release, should I strip them down a bit? Oct 29 11:19:00 if the entries are just plain wrong then yes, I guess you should remove them. Oct 29 11:19:07 can you give an example of one of the packages in question? Oct 29 11:19:35 PREFERRED_VERSION_oclock ?= "1.0.1" Oct 29 11:19:36 PREFERRED_VERSION_proxymngr ?= "1.0.1" Oct 29 11:19:45 neither app is in X11R7.4 release Oct 29 11:20:01 I guess they are still valid apps, just not in the release Oct 29 11:21:08 just wondering if they should move out to another .conf file just to make it a little clearer Oct 29 11:22:21 and there are a whole load Oct 29 11:22:43 I suppose you should probably just remove them from the X11R7.4 .conf file. Oct 29 11:22:54 (obviously not from the earlier versions, if they did indeed form part of those releases) Oct 29 11:23:33 Im guessing releasing in the form RX.X will probably end soon anyway Oct 29 11:23:35 I don't think there's much point in making a new config file for them if it's going to end up being basically just a random grab-bag of versions. Oct 29 11:23:54 yeah, probably Oct 29 11:24:41 anybody who wants to back-fill the X11R7.4 release by taking the X11R7.3 versions of packages that were dropped can already do that by just including the two config files after each other. Oct 29 11:25:08 so, I would just go ahead and remove them from the conf files where they are not applicable anymore. Oct 29 11:25:56 ah, yeah never thought of that, thanks Oct 29 11:32:33 mickeyl: good morning Oct 29 11:42:11 good morning pb__ Oct 29 11:49:33 is someone having trouble with firefox? Oct 29 11:49:37 | archivereader.o: In function `ArchiveReader::ExtractItemToStream(MarItem_ const*, _IO_FILE*)': Oct 29 11:49:40 | /home/otavio/hacking/ossystems/oe/tmp/work/i586-ossystems-linux/firefox-3.0.1-r7/mozilla/toolkit/mozapps/update/src/updater/archivereader.cpp:126: undefined reference to `BZ2_bzDecompressInit' Oct 29 11:49:44 | /home/otavio/hacking/ossystems/oe/tmp/work/i586-ossystems-linux/firefox-3.0.1-r7/mozilla/toolkit/mozapps/update/src/updater/archivereader.cpp:148: undefined reference to `BZ2_bzDecompress' Oct 29 11:49:48 | /home/otavio/hacking/ossystems/oe/tmp/work/i586-ossystems-linux/firefox-3.0.1-r7/mozilla/toolkit/mozapps/update/src/updater/archivereader.cpp:167: undefined reference to `BZ2_bzDecompressEnd' Oct 29 11:49:52 | /home/otavio/hacking/ossystems/oe/tmp/cross/i586/lib/gcc/i586-ossystems-linux/4.3.1/../../../../i586-ossystems-linux/bin/ld: updater: hidden symbol `BZ2_bzDecompress' isn't defined Oct 29 11:49:55 | /home/otavio/hacking/ossystems/oe/tmp/cross/i586/lib/gcc/i586-ossystems-linux/4.3.1/../../../../i586-ossystems-linux/bin/ld: final link failed: Nonrepresentable section on output Oct 29 11:56:10 ~seen rschuster Oct 29 11:56:13 rschuster was last seen on IRC in channel #oe, 12d 12h 37m 18s ago, saying: 'woglinde: I had to write a lesstif recipe ...'. Oct 29 12:00:18 mickeyl: hum well... You see that after doing such a nasty job he had to take some days off ;) Oct 29 12:02:05 hahaha Oct 29 12:02:11 understandable Oct 29 12:08:02 mickeyl: would you take a llok at the attachment to oebug 4784? It's about preferred_provider for python-modules Oct 29 12:09:09 ~oebug 4784 Oct 29 12:09:24 hmm... so much for the bot Oct 29 12:09:27 ant_work: will do Oct 29 12:09:37 i have a big overhaul on my list Oct 29 12:09:40 once we move to python 2.6 Oct 29 12:09:50 alas, it's very low on my list Oct 29 12:09:58 +of priority Oct 29 12:11:04 XorA: :-) Oct 29 12:11:18 mickeyl: pong Oct 29 12:12:03 thebohemian: hiho Oct 29 12:12:09 just a short ping wrt. e.V. Oct 29 12:12:13 where do we stand atm.? Oct 29 12:12:17 i lost track a bit Oct 29 12:12:54 does "bitbake console-image -c fetch" download all needed source for building? Oct 29 12:13:40 it should, yea Oct 29 12:13:58 where does it save all downloaded files? Oct 29 12:15:01 where DL_DIR points to Oct 29 12:15:38 mickeyl: about bots...no commits in gitweb since 18 hours? really? Oct 29 12:15:48 is it common that it fails to fetch a package? Oct 29 12:16:00 it did when doing a console-image earlier (after a couple of hours of crunching) Oct 29 12:16:04 cbus: unfortunately yes. upstream moves files from time to time Oct 29 12:16:21 ant_work: crossref the commit-list Oct 29 12:16:28 CIAbot somethings hangs Oct 29 12:16:33 but the e-mail commit list should rarely Oct 29 12:17:09 no new commits then Oct 29 12:17:47 moin Oct 29 12:17:59 :-) Oct 29 12:18:02 moo Oct 29 12:21:15 zecke: does bitbake have a way to list all packages depending on x? Something like 'equery depends x' in Gentoo? Oct 29 12:21:23 mickeyl: florian and me fixed the first 3 points the court mentioned Oct 29 12:21:32 mickeyl: those were easy Oct 29 12:22:03 mickeyl: the last point was about the 'wirschaftlicher/nicht-wirtschaftlicher verein' recognition Oct 29 12:22:37 thebohemian: ok, so that's still pending? Oct 29 12:22:46 mickeyl: we have to present a separate declaration how important it is for us to raise an income for the foundation Oct 29 12:22:54 ah Oct 29 12:23:06 ant_work: sure... Oct 29 12:23:06 ant_work: bitbake -g world Oct 29 12:23:11 mickeyl: basing on that the court will decide whether we are a commercial foundation or not Oct 29 12:23:20 hmm, we better get it right then Oct 29 12:23:21 ant_work: cat *.dot | grep "-> depends" :) Oct 29 12:23:33 mickeyl: obviously we don't want to be regarded as a commercial foundation Oct 29 12:23:35 ant_work: bitbake trunk has some proof of concepts for dependency browsers Oct 29 12:23:36 thebohemian: ok, do we have a) a deadline here, and b) guidelines? Oct 29 12:23:47 zecke: it's oebug 4784: something is pulling old dinosauria... Oct 29 12:24:00 mickeyl: because then the district court is not the right place to register it and we will not get tax exemption either Oct 29 12:24:04 and grep did not help Oct 29 12:24:10 mickeyl: mom Oct 29 12:24:25 mickeyl: I asked till jäger to help me with this separate declaration Oct 29 12:24:38 thebohemian: oh excellent. can you keep us all in the (mail)loop ? Oct 29 12:24:48 mickeyl: I hope to get some help to get the declaration into proper Amtsdeutsch ;) Oct 29 12:24:55 right Oct 29 12:24:56 mickeyl: sure Oct 29 12:25:03 mickeyl: he hasn't responded yet Oct 29 12:25:34 mickeyl: till is working for tarent and I got OK from elmar to ask him Oct 29 12:25:44 ah, excellent Oct 29 12:25:50 so that's "in progress" Oct 29 12:26:03 mickeyl: I think I wrote this in a mail to you, florian and holger already ... Oct 29 12:26:09 mickeyl: right Oct 29 12:26:30 probably, my scheduler has a bit too many tasks atm. Oct 29 12:26:41 round robin gives very few timeslots ;) Oct 29 12:27:34 ant_work: hmm, i guess we should finally kill python24 Oct 29 12:28:06 ant_work: i don't see a major problem with the providers though Oct 29 12:28:54 +1 for python >= 2.5 Oct 29 12:31:50 *nod* 2.5 has shown to be 99.9% compatible with 2.4 anyways Oct 29 12:32:00 2.6 will introduce some more breakage Oct 29 12:32:06 not to speak of 3.0 Oct 29 12:32:08 *sigh* Oct 29 12:32:14 ant_work: I use -g to track these kind of issues, it is not too nice though Oct 29 12:34:07 the only grep results in oetree for 'gcc-2.95' were about old embedix kernels and local.conf.sample Oct 29 12:36:27 hmm, doing a "bitbake -c fetch console-image" only tries to do 11 tasks Oct 29 12:36:30 isnt that a bit weird? Oct 29 12:38:53 zecke: good morning Oct 29 12:39:03 moin Oct 29 12:50:19 fetchall <3 Oct 29 12:51:17 I will merge this branch now.. (the one CIA is going to announce) Oct 29 12:53:42 Any ideas how PACKAGE_ARCHS could get duplicates? I'm not setting it explicitly but bitbake is complaining. Oct 29 12:54:35 broonie: you have fallen victim to koen's recent change to the cpu configuration files. Oct 29 12:54:53 most likely, your MACHINE.conf both adds to PACKAGE_ARCH itself and includes one of the tune-* files which adds the same thing Oct 29 12:55:31 Not in my own machine.conf... Oct 29 12:56:04 what MACHINE are you using? Oct 29 12:56:51 mx31ads, just grovelling about just now. Oct 29 12:57:24 is it armv6 that's duplicated? Oct 29 12:57:40 bitbake didn't say. Oct 29 12:57:52 oh, that's odd. what was the actual error output that you got? Oct 29 12:58:28 ah, I see, sanity.class now includes a check for that explicitly Oct 29 12:58:36 Error, Your PACKAGE_ARCHS field contains duplicates. Perhaps you set EXTRA_PACKAGE_ARCHS twice accidently through some tune file? Oct 29 12:58:37 but, unhelpfully, doesn't say what architectures actually have the problem Oct 29 12:58:57 broonie: use "bitbake -e | grep ^PACKAGE_ARCHS" to find out Oct 29 12:59:32 you might need to turn off sanity in order to get that to run Oct 29 12:59:39 empty output. looks like the mx31ads conf is duplicating the tune. Oct 29 13:00:14 ah, that would be no good Oct 29 13:00:52 I mean, duplicating the EXTRA_PACKAGE_ARCHS defined by the tune Oct 29 13:01:19 right, yes, that is the usual problem Oct 29 13:01:50 you probably just want to remove the definition from MACHINE.conf Oct 29 13:02:33 Thats' what I've done and it seems happier. Oct 29 13:02:55 03Holger Hans Peter Freyther  07org.openembedded.dev * r59fdade6ad 10openembedded.git/ (2 files in 2 dirs): app-restarter: Add the little helper utility to relaunch X11 apps Oct 29 13:02:55 03Graeme Gregory  07org.openembedded.dev * rcae3aa1ba1 10openembedded.git/packages/openmoko-tools/app-restarter_svn.bb: app-restarter: fix versions Oct 29 13:02:57 03Holger Hans Peter Freyther  07org.openembedded.dev * ra02570c1ce 10openembedded.git/ (2 files in 2 dirs): (log message trimmed) Oct 29 13:02:57 hal_0.5.11: Move back to the hal directory and make it default for Openmoko Oct 29 13:02:59 hal 0.5.11 is fixing many, many, many memory leaks and is required Oct 29 13:03:01 if someone wants to have the device powered on for more than two Oct 29 13:03:03 days. I heavily encourage other distributions to switch as well. Oct 29 13:03:05 I can not see the autoreconf errors mentioned in koen's move Oct 29 13:03:07 a765a58dfd410eef8147e2e6ad1d0e4cef922c58 he is free to send them to Oct 29 13:03:09 03Holger Hans Peter Freyther  07org.openembedded.dev * r1bcbcaea6f 10openembedded.git/packages/openmoko-tools/app-restarter_svn.bb: app-restarter: Add X11 to the dependencies. Oct 29 13:03:12 03Holger Hans Peter Freyther  07org.openembedded.dev * r7434683fda 10openembedded.git/ (2 files in 2 dirs): sane-srcrevs.inc: Upgrade app-restarter and fix the PV, bump the PE Oct 29 13:03:17 03Holger Hans Peter Freyther  07org.openembedded.dev * r12e84f237f 10openembedded.git/ (2 files in 2 dirs): Oct 29 13:03:22 openmoko: Add a Gtk+ utility to set the root password Oct 29 13:03:24 This utility will be installable through the feed. Oct 29 13:03:26 03Graeme Gregory  07org.openembedded.dev * rc5431547e5 10openembedded.git/packages/openmoko-tools/openmoko-set-root-password.bb: openmoko-set-root-password: fix versions Oct 29 13:03:29 03Holger Hans Peter Freyther  07org.openembedded.dev * r078b0e6335 10openembedded.git/packages/xserver-kdrive-common/ (xserver-nodm-init.bb xserver-nodm-init/xserver-nodm): Oct 29 13:03:32 xserver-nodm-init: Make sure to set a HOME and LOGNAME when launching Oct 29 13:03:34 Set a HOME and LOGNAME when none is being set. Oct 29 13:03:36 03John Lee  07org.openembedded.dev * r4688924045 10openembedded.git/packages/dropbear/dropbear.inc: Oct 29 13:03:41 dropbear: add /etc/default/dropbear in conffiles only on openmoko Oct 29 13:03:43 distro. thanks holger for pointing it out. Oct 29 13:03:51 03Tick Chen  07org.openembedded.dev * r80b160a11b 10openembedded.git/ (2 files in 2 dirs): Oct 29 13:03:54 packagekit: Update to a public version of packagekit Oct 29 13:03:56 Remove the private patches and use a public packagekit again Oct 29 13:03:58 03Julian_chu  07org.openembedded.dev * rcc8f1710d3 10openembedded.git/packages/apmd/apmd_3.2.2.bb: Oct 29 13:04:01 apmd: Correct the .la file Oct 29 13:04:03 Apmd install the .la file but the content was wrong. Oct 29 13:04:05 It should remove the STAGING path. Oct 29 13:04:07 03walkingice  07org.openembedded.dev * r3c2efdd0d3 10openembedded.git/packages/packagekit/packagekit_git.bb: Modify packagekit to make it patch successful Oct 29 13:04:12 03Julian_chu  07org.openembedded.dev * r41cef62881 10openembedded.git/packages/packagekit/ (files/repository-ping.patch packagekit_git.bb): Oct 29 13:04:15 pb__: Thanks, I'll send a patch if the build completes. Oct 29 13:04:17 packagekit: Add a patch to packagekit. Oct 29 13:04:19 The patch add repos_ping to the backend therefore Oct 29 13:04:20 ok, cool Oct 29 13:04:21 packagekit can ping repository via backend. Oct 29 13:04:23 03John Lee  07org.openembedded.dev * r294c894b7a 10openembedded.git/packages/dropbear/ (dropbear.inc dropbear/openmoko/default dropbear_0.51.bb): Oct 29 13:04:28 dropbear: improve dropbear security by listening only on usb0 by default Oct 29 13:04:30 trac #1853, originated by RuiSeabra Oct 29 13:04:32 03Julian Chu  07org.openembedded.dev * r2f2fce10c4 10openembedded.git/ (2 files in 2 dirs): Oct 29 13:04:35 packagekit: Change revision of packagekit. Oct 29 13:04:37 To make it buildable, add a patch to it. Oct 29 13:04:41 In next time we update the packagekit version, this patch doesn't need anymore. Oct 29 13:04:45 ( it should enter in upstream) Oct 29 13:04:47 03Julian_chu  07org.openembedded.dev * rd92e054547 10openembedded.git/packages/packagekit/ (files/03_group_infomation.patch packagekit_git.bb): Oct 29 13:04:50 (33 lines omitted) Oct 29 13:04:58 wow nice bot ^^ Oct 29 13:06:53 thebohemian: ping Oct 29 13:10:29 *sigh* gettext won't patch. Oct 29 13:10:54 broonie hu? Oct 29 13:11:05 broonie what you are building? Oct 29 13:11:18 angstrom x11-image for mx31ads Oct 29 13:11:30 hm gettext-native? Oct 29 13:12:21 yes Oct 29 13:12:39 because angstroem is using proxy-libintl at runtime Oct 29 13:13:01 broonie which version you gettext? Oct 29 13:13:13 0.17-rc3 Oct 29 13:13:23 rc3? Oct 29 13:13:26 It looks like it's further up than that, quilt is getting invoked with invalida rguments. Oct 29 13:13:33 whats this? Oct 29 13:13:34 zecke: pong Oct 29 13:13:43 r3,sorry Oct 29 13:14:05 I never checked in r3 Oct 29 13:14:13 where is this comming from? Oct 29 13:15:11 org.openembedded.dev HEAD. Oct 29 13:15:38 nope Oct 29 13:15:46 Hrm. The gettext.bb has r2 in it. Oct 29 13:15:56 How can I get bitbake to tell me what it's looking at? Oct 29 13:16:09 hm Oct 29 13:16:19 ah, -native has r3 Oct 29 13:16:45 broonie which patch is failing? Oct 29 13:17:06 autotools.patch - like I say, it's not even trying to apply the patch Oct 29 13:17:19 quilt is barfing on the command line Oct 29 13:18:08 broonie hm whats your bitbake version? Oct 29 13:19:02 svn head Oct 29 13:19:26 hm Oct 29 13:19:28 strange Oct 29 13:19:37 * broonie is investigating. Oct 29 13:19:42 must be local problem Oct 29 13:20:02 otherwise all guys had screamd at me Oct 29 13:20:17 or the autobuilders Oct 29 13:20:23 Or something further up, I think gettext is just the first package it hit. Oct 29 13:23:17 how can I check the current srcrev of a git branch ? Oct 29 13:27:12 dcordes git show? Oct 29 13:28:20 03Holger Hans Peter Freyther  07holger/openmoko-merges * rf7123c9a64 10openembedded.git/packages/e17/ (e-wm/Xsession.d/98enlightenment e-wm_svn.bb): e-wm: Start e with optional arguments we get from a config file Oct 29 13:28:23 03Holger Hans Peter Freyther  07holger/openmoko-merges * r54e4d97e8a 10openembedded.git/packages/e17/e-wm/Xsession.d/98enlightenment: e-wm: Remove left over, e gets started through the x-window-manager Oct 29 13:28:24 03Holger Hans Peter Freyther  07holger/openmoko-merges * r83c19ae31e 10openembedded.git/packages/e17/e-wm_svn.bb: e-wm: Fix the DESCRIPTION as of https://docs.openmoko.org/trac/ticket/1894 Oct 29 13:28:25 03Holger Hans Peter Freyther  07holger/openmoko-merges * r7cf8236642 10openembedded.git/packages/mono/mono_svn.inc: Oct 29 13:28:26 mono: Two svn checkouts are used, specify a SRCREV_FORMAT Oct 29 13:28:28 Currently the same SRCREV is used for mcs and mono so have a Oct 29 13:28:30 simple SRCREV_FORMAT. Oct 29 13:28:32 03Holger Hans Peter Freyther  07holger/openmoko-merges * r509a1ba26d 10openembedded.git/packages/ (6 files in 4 dirs): bashism: Move away from pushd/popd as it is only available for bash Oct 29 13:28:35 03Holger Hans Peter Freyther  07holger/openmoko-merges * rd62958c18c 10openembedded.git/packages/dbus/ (dbus-1.0.3/fedora-compile-fix.patch dbus-native_1.0.3.bb): Compile with recent glibc,gcc,headers (e.g. fedora 9) Add a patch to dbus to properly compile on fedora 9. Oct 29 13:28:41 03Holger Hans Peter Freyther  07holger/openmoko-merges * r3d702a5a96 10openembedded.git/packages/tasks/ (task-mipl.bb task-openmoko-asu.bb task-openprotium.bb): Oct 29 13:28:44 tasks: Set PACKAGE_ARCH so it has an effect Oct 29 13:28:46 Set PACKAGE_ARCH after inherit task so it is not getting lost. Do Oct 29 13:28:48 not set it to all as task.bbclass is already doing that. Oct 29 13:28:50 03Holger Hans Peter Freyther  07holger/openmoko-merges * r2813dc78d6 10openembedded.git/packages/clutter/moblin-proto_git.bb: Oct 29 13:28:55 moblin-proto: Use the right git server Oct 29 13:28:57 - moblin.org => git.moblin.org Oct 29 13:29:44 thebohemian: could you take a look at my sanity.bbbclass change and say if ecj still needs to mmap '0'? Oct 29 13:32:07 03Holger Hans Peter Freyther  07org.openembedded.dev * r83c19ae31e 10openembedded.git/packages/e17/e-wm_svn.bb: e-wm: Fix the DESCRIPTION as of https://docs.openmoko.org/trac/ticket/1894 Oct 29 13:32:08 03Holger Hans Peter Freyther  07org.openembedded.dev * r54e4d97e8a 10openembedded.git/packages/e17/e-wm/Xsession.d/98enlightenment: e-wm: Remove left over, e gets started through the x-window-manager Oct 29 13:32:08 03Holger Hans Peter Freyther  07org.openembedded.dev * r2813dc78d6 10openembedded.git/packages/clutter/moblin-proto_git.bb: Oct 29 13:32:09 moblin-proto: Use the right git server Oct 29 13:32:11 - moblin.org => git.moblin.org Oct 29 13:32:13 03Holger Hans Peter Freyther  07org.openembedded.dev * rd62958c18c 10openembedded.git/packages/dbus/ (dbus-1.0.3/fedora-compile-fix.patch dbus-native_1.0.3.bb): Compile with recent glibc,gcc,headers (e.g. fedora 9) Add a patch to dbus to properly compile on fedora 9. Oct 29 13:32:19 03Holger Hans Peter Freyther  07org.openembedded.dev * r7cf8236642 10openembedded.git/packages/mono/mono_svn.inc: Oct 29 13:32:22 mono: Two svn checkouts are used, specify a SRCREV_FORMAT Oct 29 13:32:24 Currently the same SRCREV is used for mcs and mono so have a Oct 29 13:32:26 simple SRCREV_FORMAT. Oct 29 13:32:28 03Holger Hans Peter Freyther  07org.openembedded.dev * r509a1ba26d 10openembedded.git/packages/ (6 files in 4 dirs): bashism: Move away from pushd/popd as it is only available for bash Oct 29 13:32:31 03Holger Hans Peter Freyther  07org.openembedded.dev * r3d702a5a96 10openembedded.git/packages/tasks/ (task-mipl.bb task-openmoko-asu.bb task-openprotium.bb): Oct 29 13:32:34 tasks: Set PACKAGE_ARCH so it has an effect Oct 29 13:32:36 Set PACKAGE_ARCH after inherit task so it is not getting lost. Do Oct 29 13:32:38 not set it to all as task.bbclass is already doing that. Oct 29 13:32:40 03Holger Hans Peter Freyther  07org.openembedded.dev * rf2bec9a8f6 10openembedded.git/: Merge branch 'holger/openmoko-merges' into org.openembedded.dev Oct 29 13:32:49 03Holger Hans Peter Freyther  07org.openembedded.dev * rf7123c9a64 10openembedded.git/packages/e17/ (e-wm/Xsession.d/98enlightenment e-wm_svn.bb): e-wm: Start e with optional arguments we get from a config file Oct 29 13:33:54 zecke: http://gitweb.openembedded.net/?p=openembedded.git&a=commit;h=c3d62194db40e759f2bfa22e22aea1ae0ee0583f ? Oct 29 13:34:37 i have a problem regarding the fetch of illume_svn Oct 29 13:35:15 the bb-file wants to fetch a revision's directory which is not available anymore Oct 29 13:35:27 and svn bugs out on that Oct 29 13:35:39 i was told to use the "peg revision" Oct 29 13:36:32 this would mean, doing something like svn co http://blabla/tree@36450 instead of specifying svn co -r 36450 Oct 29 13:37:01 could someone help me with the correct syntax in the bb-files here? Oct 29 13:37:03 woglinde, hm that shows the commit ids but where is the sourcerevision? I'm looking for the one that you put in SRCREV_pn_package to get the latest code from that branch. Oct 29 13:37:32 I always read it from the gitweb url.. Oct 29 13:40:32 thebohemian: yes? Not this commit, the mail I sent to the list :) Oct 29 13:42:33 zecke: ok Oct 29 13:43:34 03Elena ``of Valhalla'' Grandi  07org.openembedded.dev * r2b6bf90a32 10openembedded.git/packages/podget/ (podget-0.5.8/busybox_df.patch podget_0.5.8.bb): podget 0.5.8: new recipe Oct 29 13:43:40 zecke: I think the propblem you are experiencing with ecj-initial is actually cacao doing the mmap Oct 29 13:43:45 03Elena ``of Valhalla'' Grandi  07org.openembedded.dev * r211e659dc4 10openembedded.git/conf/checksums.ini: podget 0.5.8: added checksum.ini entry Oct 29 13:43:58 zecke: actually cacao-initial Oct 29 13:44:21 zecke: I can't provide a fix for this since we need to use a rather ancient cacao version for *-initial purposes Oct 29 13:44:38 zecke: you can work around this by using jamvm-initial instead Oct 29 13:44:55 zecke: that would be slower but avoids the need for mmaping the zero page Oct 29 13:45:26 zecke: I add this information to OE-wiki/Java Oct 29 13:48:31 zecke: on my current system I set vm_mmap_min_adr = 0 on order to get qemu to work correctly inside OE. Oct 29 13:48:34 thebohemian: oh well, if this is needed we can add the sanity.bbclass check Oct 29 13:49:06 zecke: so patch is ok for me Oct 29 13:49:11 thebohemian: I don't and have binary locales disabled so get bit by the java stuff Oct 29 13:49:11 zecke: wait Oct 29 13:49:21 zecke: it is better to say 'or cacao-initial' Oct 29 13:49:22 cool Oct 29 13:49:39 because that is the app that actually fails Oct 29 13:49:57 thanks, I will update that Oct 29 13:50:16 time for some assignments now Oct 29 13:50:53 anyone knows how to draw vector fields with gnuplot or octave? Oct 29 13:54:23 woglinde: blowing away my tmp directory appears to have fixed whatever was wrong. *sigh* Oct 29 13:54:25 lol bitbakery.org Oct 29 13:56:25 zecke, you registered that? Oct 29 13:56:35 no Oct 29 13:56:43 some people do autobuilds there, wicked :) Oct 29 13:57:59 zecke, on the machine where the site is hosted? seems there is no content Oct 29 14:02:35 oestats.bitbakery.org Oct 29 14:03:44 zecke, but they don't bake but toast. Oct 29 14:07:27 i have an issue with illume_svn.bb - it fails at checking out the svn-revision Oct 29 14:07:35 36450 Oct 29 14:18:03 re Oct 29 14:20:04 pretty logo for bitbake Oct 29 14:21:20 :-| Oct 29 14:21:33 no-one here who can help me fix a bb recipe? Oct 29 14:21:34 hmm, got any url for the bitbaker logo? :D Oct 29 14:27:12 tomboy64: well, your question is not specific enough... fetching fails... do you hit the connection limit on the e server? Oct 29 14:27:21 tomboy64: are you behind a firewall? Oct 29 14:27:25 zecke: i am Oct 29 14:27:29 but that's not the issue Oct 29 14:27:39 it's concerning illume Oct 29 14:28:09 it had some issues downloading the correct revision, i tinkered around with peg-revisioning, adding rev=$revision Oct 29 14:28:13 tomboy64: you need to be more specfic to gett help Oct 29 14:28:57 and now it wants to get a file from the angstrom repos Oct 29 14:29:04 i'll get the logs, sec please Oct 29 14:30:07 this file is not existent: it wants to fetch illume_svn.enlightenment.org_.svn.e.trunk_36450_.tar.gz Oct 29 14:30:31 from http://www.angstrom-distribution.org/unstable/sources/ Oct 29 14:31:31 zecke: http://rafb.net/p/eBNkKB79.html Oct 29 14:33:39 the file it's seeking is not available Oct 29 14:33:47 tomboy64: this not an issue Oct 29 14:33:52 and that is the stock build Oct 29 14:33:59 *bb file i mean Oct 29 14:34:01 Is it possible to configure bitbake to run clean after each sucessfull build of bitbake file? Or are files in tmp/work needed for other builds? Oct 29 14:34:21 tomboy64: it tries a download server first to take load away from the subversion/git/cvs... servers Oct 29 14:34:38 uhm ... Oct 29 14:35:00 tomboy64: if download fails it goes to the scm server Oct 29 14:35:16 ok, hang on Oct 29 14:37:26 zecke: could you have a look at this one please? http://rafb.net/p/i40qat17.html Oct 29 14:37:54 more context please Oct 29 14:38:52 http://rafb.net/p/uHDMpZ52.html Oct 29 14:39:00 i had to forcefully abort the download-attempts Oct 29 14:39:12 what happens is, svn needs a "peg revision" Oct 29 14:39:42 e.g. illume has been moved on the server and svn can't find it in the old revision without that peg-revision in the url Oct 29 14:40:14 03Elena ``of Valhalla'' Grandi  07org.openembedded.dev * r8ecf1848ce 10openembedded.git/conf/checksums.ini: geany 0.14: added checksum.ini entry Oct 29 14:40:37 it should look like svn co -r 36540 http://svn.enlightenment.org/svn/e/trunk/illume@36540 illume Oct 29 14:40:46 ah fun with subversion :) Oct 29 14:40:48 instead of just svn co -r 36540 http://svn.enlightenment.org/svn/e/trunk/illume illume Oct 29 14:40:53 yeah :-| Oct 29 14:41:15 yes, trunk/illume vanished from later revisions and svn needs a hint to find it... Oct 29 14:41:29 i tried the docs for bitbake ... not much of a help Oct 29 14:41:31 tomboy64: I fear you need to patch bitbake (lib/bb/fetch/svn.py) Oct 29 14:41:38 * tomboy64 sighs Oct 29 14:41:47 tomboy64: it should be fairly easy Oct 29 14:41:55 yes Oct 29 14:41:59 i'll try myself Oct 29 14:42:11 if i succeed, where should i file a bug/patch? Oct 29 14:42:42 bitbake-dev, this channel Oct 29 14:42:58 but I actually try hard to make my assignments... Oct 29 14:43:08 zecke haha Oct 29 14:43:11 bitbake-dev? that's the mailing list? Oct 29 14:43:26 zecke: you can't do everything :) Oct 29 14:43:37 yes, check developers.berlios.de/projects/bitbake Oct 29 14:43:41 alright Oct 29 14:48:46 if I want to change the config file for my kernel, where would the best place be do do so? Just change the defconfig in packages/linux/linux-omap-2,6,23-omap1 (which is what I want to build). Or is there a better/cleaner way ? Oct 29 14:49:56 tomboy64: cool Oct 29 14:55:43 pew pew Oct 29 15:00:15 hey Oct 29 15:08:46 g'day kergoth Oct 29 15:18:42 03Robert Schuster  07org.openembedded.dev * r2efc0d473e 10openembedded.git/conf/checksums.ini: checksums.ini: Added midpath 0.3rc1 entry. Oct 29 15:19:02 * kergoth tries to put together a list of books to read to do better at his job Oct 29 15:20:05 heh Oct 29 15:20:29 If you stick with x86, then you need "The Indispensable PC Hardware Book" Oct 29 15:20:53 Its out of print, but if you live near a reasonable used book store, they should be able to get one for you Oct 29 15:21:51 okay, thanks for the tip. i expect i'll be on x86 for a while, my customer only uses that Oct 29 15:22:17 Its one of those things you might not use for 9 months, and then you really really need to figure out how IDE works Oct 29 15:23:55 * kergoth nods Oct 29 15:24:28 i fix the customer's bugs... what the bug involves varies. 90% of the time i work on boring backports, but then bam, ugly panic issue with scsi or something Oct 29 15:24:35 heh Oct 29 15:24:53 The mindshare books are on a whole really exellent Oct 29 15:25:39 I've got like 7 of those on low priority on my amazon wishlist :) i'll get em eventually Oct 29 15:26:08 Together all of us have most of them Oct 29 15:26:12 cool Oct 29 15:26:15 its just a matter of finding out where in the bulding they are Oct 29 15:26:36 * kergoth 's next priority is to read a good asm book, to brush up so he can do better at core analysis Oct 29 15:26:42 then we'll see Oct 29 15:27:03 03Elena ``of Valhalla'' Grandi  07org.openembedded.dev * r3ff38e08f4 10openembedded.git/conf/checksums.ini: Oct 29 15:27:03 mcabber 0.9.9: added checksum.ini entry Oct 29 15:27:03 This closes bug 4764 Oct 29 15:27:03 http://bugs.openembedded.net/show_bug.cgi?id=4764 Oct 29 15:27:27 kergoth, CosmicPenguin: do either of you happen to have AT&T or BellSouth internet? Oct 29 15:27:34 fraid not Oct 29 15:27:40 one of our pesky customers is complaining of a problem that happens with those ISPs Oct 29 15:27:54 drat, oh well Oct 29 15:28:12 Sorry Oct 29 15:28:33 I am poorly serviced by one of the other baby bells Oct 29 15:30:35 zecke: so, my fso-image build now fails with "Cannot find package gst-plugin-mad", and the same for gst-plugin-sid. Oct 29 15:30:40 is there something I need to clean and rebuild? Oct 29 15:30:47 CosmicPenguin: thanks anyway Oct 29 15:31:58 pb__: nah, you need to kick me... Oct 29 15:32:03 * pb__ kicks zecke Oct 29 15:32:16 :) Oct 29 15:32:19 harder maybe? Oct 29 15:32:31 pb__: locally, remove that from your fso-image.bb Oct 29 15:32:43 * pb__ takes a long run-up and kicks zecke again Oct 29 15:32:47 zecke: okay, thanks :-) Oct 29 15:33:02 pb__: I will do some work to fix this kind of stuff properly Oct 29 15:33:08 excellent Oct 29 15:33:21 nevermind, if I can build images locally then I will be happy again for now :-) Oct 29 15:33:48 gonna see if I can get the whole of mickey's framework stack to run in a thumb world now. Oct 29 15:33:56 for some reason, last time I tried that the machine froze up Oct 29 15:36:11 how was your gcc error resolved? Oct 29 15:36:36 it seemed to be some bad effect of mixing gcc versions Oct 29 15:37:00 I had previously built gcc-4.2.4 in that tmpdir, and this seems to poison the well somehow in such a way that gcc-4.1.2 won't then compile against that staging directory Oct 29 15:38:58 so, basically I just deleted the whole of tmp and recompiled Oct 29 15:40:00 ok, great, image building works now Oct 29 15:40:36 zecke: by the way, have you ever experienced a thing where dropbear suddenly stops listening for new connections? this has started happening for me quite a lot, I'm not sure if it's a toolchain issue or something else. Oct 29 15:41:38 pb__: after what amount of time? For Openmoko we still have the older 0.49 in the field and din't get a bug report like this Oct 29 15:42:05 varies, anything from a few minutes to a few hours Oct 29 15:42:10 this is with 0.50 Oct 29 15:42:37 it seems that the "master" dropbear process is exiting for some reason. Oct 29 15:42:57 I suspect it's probably a toolchain or library bug but I'm not sure where, yet. Oct 29 15:43:38 pb__: hmm, attach gdb to it while it is running and see how it quit? Oct 29 15:44:01 yeah Oct 29 15:44:12 ah, it seems that it's getting an unexpected error from select() for some reason Oct 29 15:44:17 unhelpfully it doesn't say which one :-} Oct 29 15:45:18 pb__: I assume you use strace now? Oct 29 15:46:13 yeah, trying that now Oct 29 15:46:20 I'm just waiting for it to crash again Oct 29 15:46:26 -ENOSPACE :) Oct 29 15:46:30 heh Oct 29 15:46:54 seems odd that a thumb problem would cause this, there's nothing very special about select() that I remember. Oct 29 15:47:08 03Koen Kooi  07org.openembedded.dev * rdc39745f3c 10openembedded.git/packages/hal/ (hal-0.5.11/fix-configure.diff hal_0.5.11.bb): hal 0.5.11: add fix from hal git to fix libtool issues in configure Oct 29 15:55:30 zecke: as ever, now it does not fail :-} Oct 29 15:55:46 oh well, problem solved. I will just patch the init scripts to run dropbear inside strace! Oct 29 15:55:55 sounds like fun Oct 29 16:03:31 pb__: maybe just enable ptracing on every process? Oct 29 16:05:38 pb__: I know the deal of a cambridge master, do you know if something like this exists for a bachelor degree? Oct 29 16:06:00 zecke: I'm not sure I understand the question Oct 29 16:07:00 pb__: Search for a university that gives me a bachelor degree just because I have high school degree and comitted no crime :) Oct 29 16:07:15 ah, heh Oct 29 16:07:19 I don't know of anybody who does that :-) Oct 29 16:07:43 You clearly haven't read my spam folder lately Oct 29 16:08:02 CosmicPenguin: they want money :) Oct 29 16:08:49 I figured as much Oct 29 16:09:10 though the spammers could also be really concerned that I am as impotent, lonely, and degreeless as they think I am Oct 29 16:31:03 tomboy64: any progress? Oct 29 16:33:59 hello Oct 29 17:02:15 hmm, anyone play with android's 'repo' tool? Oct 29 17:02:23 * kergoth poked at it a little th eother day, seems interesting Oct 29 17:08:32 good morning Oct 29 17:09:33 has anyone done any x86 builds lately? got couple questions here :) Oct 29 17:10:06 as android , should we put a kill switch in our default image ^^ Oct 29 17:10:30 that feed buzz , make some ads for us :] Oct 29 17:17:06 denix: not terribly recently, but a few weeks ago. what are your questions? Oct 29 17:17:35 1. what's the best MACHINE for PentiumIII? I'm trying i686-generic now... Oct 29 17:17:49 that is probably the best Oct 29 17:18:03 bonne soirée Oct 29 17:18:36 2. I tried DISTRO=minimal and failed in libtool-cross. also tried DISTRO=angstrom-2008.1 and failed in glibc with some weird ASM errors... any clues? Oct 29 17:19:18 without seeing the actual errors, no, no idea Oct 29 17:19:44 I have never tried an x86 build with 4ng5tr0m. I think it ought to work in theory but I have no idea whether it really does in practice. Oct 29 17:20:00 minimal should be ok though Oct 29 17:20:39 ok, lets try minimal again :) although it uses some old packages, compared to angstrom Oct 29 17:24:49 so, if I use minimal distro, libtool-cross fails do_configure saying gcc cannot generate binary Oct 29 17:25:38 you'll have to check config.log to find out what's wrong with it Oct 29 17:26:25 I did Oct 29 17:26:31 wait 1 sec Oct 29 17:36:50 mickey|shower: ping Oct 29 17:37:10 zecke: he's been in the shower for about 8 hours. I fear he has been washed down the plughole by now. Oct 29 17:38:08 pb__: maybe there is internet on the other side? Oct 29 17:40:39 hmm, http://git.oblomov.eu/zit is kind of cute. i wonder what the use case is, though Oct 29 17:43:58 zecke: got to be - after all it's a "series of tubes"... :) Oct 29 17:43:59 zecke pasted "fix for tomboy?" at http://paste.lisp.org/display/69404 Oct 29 17:44:31 pb_: sorry for the delay - got sidetracked here Oct 29 17:45:50 pb_: the config.log says cannot run i686-linux-gcc Oct 29 17:46:24 pb_: and the strange thing there was no cross directory in temp, even though gcc-cross was built before Oct 29 17:46:42 that is rather odd Oct 29 17:46:49 I wonder where your cross directory has gone Oct 29 17:46:55 what version of gcc-cross are you using? Oct 29 17:47:33 hmm, did a clean build and got the same error, this time cross is in place though Oct 29 17:48:19 its gcc-4.1.1 Oct 29 17:49:31 that version should be fine, I would have thought Oct 29 17:50:27 http://pastebin.com/d7ef93e24 Oct 29 17:53:10 hmm, this time the config.log is little bit different, since it can now find i686-linux-gcc in the tmp/cross Oct 29 17:53:43 pb_: please see this - http://pastebin.com/d53256b83 Oct 29 17:54:30 undefined reference to `__libc_csu_fini', `__libc_csu_init', `__libc_start_main' Oct 29 18:04:19 03Koen Kooi  07org.openembedded.dev * rf36893c64e 10openembedded.git/packages/xorg-xserver/ (files/xcalibrate-xorg.diff xserver-xorg_1.4.2.bb): xserver xorg 1.4.2: try harder to get xcalibrate enabled Oct 29 18:05:11 pb_: should libtool-cross compile after gcc-cross and glibc-cross are done? Oct 29 18:06:22 it shouldn't need a crosscompiler for that, iirc Oct 29 18:07:19 so why it uses i686-linux-gcc built by gcc-initial? Oct 29 18:08:48 what's oe think your BUILD_SYS is? Oct 29 18:09:28 i686-linux Oct 29 18:10:04 see, it's not looking for the cross stuff at all. Oct 29 18:10:18 it's looking for your build machine's compiler, and will fall back to 'gcc' when it cant find i686-linux-gcc Oct 29 18:10:28 afaik anyway Oct 29 18:11:15 I don't have i686-linux-gcc as mine is i686-pc-linux-gnu-gcc Oct 29 18:12:05 and it finds i686-linux-gcc in staging where gcc-cross-initial just got installed Oct 29 18:16:34 03Koen Kooi  07org.openembedded.dev * re25e2f583c 10openembedded.git/packages/xorg-xserver/xserver-xorg/omap3evm/xorg.conf: xserver-xorg: use omapfb driver and tslib for omap3evm machine Oct 29 18:21:37 denix: sounds like your target is the same as your build machine, so some of your stuff isn't clear that its crosscompilation. I'd say set TARGET_VENDOR or BUILD_VENDOR to be absolutely certain there's no confusion Oct 29 18:21:42 * kergoth shrugs Oct 29 18:22:47 I'll try Oct 29 18:26:02 woglinde: holger/enterprise could you take a look at this branch? Oct 29 18:26:44 zecke what exactly should I look for? Oct 29 18:27:14 woglinde: missing PR Oct 29 18:27:22 03Holger Hans Peter Freyther  07holger/enterprise * r0d2c29eaff 10openembedded.git/ (53 files in 17 dirs): (log message trimmed) Oct 29 18:27:22 bluemchen wiese: Allow to disable certain ugly packages globally Oct 29 18:27:22 Adopt the notion of gstreamer and allow to easily not build certain Oct 29 18:27:22 ugly packages. People are constraints to not live on a bluemchen Oct 29 18:27:24 wiese can set ENTERPRISE_DISTRO="1" in their config and get ugly Oct 29 18:27:26 packages removed. Oct 29 18:27:28 This is implemented by a set of packages in the BBMASK (as a safety Oct 29 18:27:56 zecke: "ugly"? Oct 29 18:28:36 yeah gstreamer has good bad and ugly plugins Oct 29 18:30:39 03Koen Kooi  07org.openembedded.dev * r686b241399 10openembedded.git/contrib/angstrom/build-feeds.sh: angstrom feed builder: add qemux86 machine Oct 29 18:31:57 huh, the way google's repo tool handles upstream stuff is interesting Oct 29 18:32:25 could be worth considering as a way to handle our modifications to upstream Oct 29 18:33:11 kergoth: What do they do? Oct 29 18:33:33 kergoth: Until now I thought repo was just a multi git repo checkout tool Oct 29 18:33:37 like mr Oct 29 18:33:38 they fetch the upstream tarball from upstream, write that as a tree object into a git repo, and clone the rest of their repo for it from their site, which has everything else Oct 29 18:33:40 yeah thats part of it Oct 29 18:33:46 the rest is this "import" upstream thing Oct 29 18:33:57 hmm Oct 29 18:34:10 their git repositories for external code just contain their changes, they don't have the upstream tarball object, which keeps the repo small Oct 29 18:34:11 sounds a bit like some debian package teams handle this Oct 29 18:34:28 yeah, its similar to how lots of distro folk have considered handling things Oct 29 18:35:11 not sure how happy I'd be about continuing to rely on upstream urls for sources, but it would definatley keep the repo size down Oct 29 18:35:25 someone should throw together a proof of concept repo based oe metadata+code setup Oct 29 18:35:29 just to play around Oct 29 18:35:29 What I like is the integration with the review system. Even if it depends on the project if such a complex thing would be used Oct 29 18:35:50 haven't looked at what they're doing for the reviews, how does it work? Oct 29 18:36:09 * stefan_schmidt looks at his todo list and is sure that he is not this somebody. :) Oct 29 18:36:14 hehe Oct 29 18:36:16 kergoth: mom Oct 29 18:36:29 i might give it a whirl, I've disliked the way we organize things for a while Oct 29 18:37:52 heh, google really sucks if you search for stuff over them. Android, GSoC Oct 29 18:38:26 hehe Oct 29 18:38:37 reviewboard looks quite a bit nicer than google's review webapp imo Oct 29 18:38:38 kergoth: There you go: http://source.android.com/submit-patches/workflow Oct 29 18:38:46 ah, thanks Oct 29 18:51:58 hi, i tried to make a new -native.bb - it works fine, no errors but it just installs nothing. any idea why? the do_install.log is zero size Oct 29 19:07:00 ok, git-stash is useful Oct 29 19:07:42 very Oct 29 19:07:57 only thing i hate is i end up with a ton of old stashes laying aorund that i forgot about :) Oct 29 19:08:09 heh Oct 29 19:12:30 03Joshua Lock  07org.openembedded.dev * r714e2cace5 10openembedded.git/packages/xerces-c/xerces-c_2.8.0.bb: xerces-c: Correct SRC_URI Oct 29 19:12:30 03Stefan Schmidt  07org.openembedded.dev * r2086b01ec3 10openembedded.git/conf/checksums.ini: checksums.ini: Add sums for new xerces-c location Oct 29 19:14:27 zecke: I just verified that the hook decline that I can remove the stefan/xo-support branch. Some magic from you please. :) Oct 29 19:18:51 stefan_schmidt: like in every good hacker movie I'm waiting for the transaction to hit my account in lichtenstein Oct 29 19:19:08 zecke: cool Oct 29 19:19:25 zecke: Let me make sure that this transaction is not traceable Oct 29 19:23:54 hmm, anyone know where the schemas are for the xml files used in android? Oct 29 19:26:27 You're going to end up making me buy an Android, aren't you? Oct 29 19:34:17 hehe, i avhent even looked at the platform itself yet, just the tools ;) Oct 29 19:36:07 Hello everybody. I have mandriva2009, and i have next error while bitbake. Oct 29 19:36:07 NOTE: package perl-native-5.8.8: started Oct 29 19:36:07 ERROR: function do_compile failed Oct 29 19:36:07 make: *** No rule to make target `/usr/lib/gcc/i586-manbo-linux-gnu/4.2.3/include/float.h', needed by `miniperlmain.o'. Stop. Oct 29 19:36:07 FATAL: oe_runmake failed Oct 29 19:36:09 I just migrate from mandriva 2008, and some packets are built. Oct 29 19:40:09 Kalemas: start a clean build? Oct 29 19:52:50 Is there an "easy" way to determine by which dependency a package ends up in a certain image? I tried with "bitbake -g" but the dot-file does not seem to be correct/usable. Oct 29 19:53:55 Kalemas: the dot files work for me, at least they should have the right depends Oct 29 19:54:05 RP: could you review a bitbake patch for me? Oct 29 19:54:48 zecke: sorry can you tell me exact command, how to clean. Oct 29 19:55:07 zecke: hm ok thanks, I'll give it another spin. Have you ever tried with dot files for complete images? Oct 29 19:56:04 Kalemas: rm -rf tmp, or bitbake -cclean perl but I would suspect more packages changed? Oct 29 19:56:29 Kensan: no, I open them in vi and search for my things or -> Oct 29 19:56:48 Kensan: there is one bug that some "sugar" gets torn into single charachters... Oct 29 20:02:42 03Khem Raj  07org.openembedded.dev * r587358b294 10openembedded.git/ (12 files in 3 dirs): Oct 29 20:02:42 binutils checksum.ini: Add binutils 2.19 recipes Oct 29 20:02:42 Add new checksum entry Oct 29 20:16:47 Did someone built firefox or fennec lately? Oct 29 20:16:52 both fails to me Oct 29 20:22:24 otavio: koen does Oct 29 20:22:30 otavio: maybe msg him or send a mail Oct 29 20:23:57 zecke: what is his nickname? Oct 29 20:24:45 koen Oct 29 20:24:56 but he is busy studying Oct 29 20:25:37 can someone explain to me the difference between staging and installation? Oct 29 20:25:42 zecke: good; I've msged him and asked him to ping me when available Oct 29 20:27:11 or point me to documentation that explains it... Oct 29 20:30:06 http://comments.gmane.org/gmane.comp.version-control.git/99389 is prety nifty Oct 29 20:30:11 zecke: seen it? Oct 29 20:32:01 jerm^_: basically: the installation part is for image/package creation while staging is a common place from where other packages will look for libraries to include during its compilation/linkage Oct 29 20:32:11 kergoth: that is slick Oct 29 20:32:18 quite Oct 29 20:33:56 a little re-inventing the tinderbox wheel Oct 29 20:36:27 vivijim: thx Oct 29 20:36:45 CosmicPenguin: what's neat is the way it handles failures. notice how it automatically bisects to isolate the problem commit Oct 29 20:36:49 CosmicPenguin: that's pretty sexy right there Oct 29 20:37:11 course, you could use a continuous integration style setup instead, and never get to that point Oct 29 20:37:31 at least that not fine grained Oct 29 20:37:47 wonder how often it tests branch heads rather than tags Oct 29 20:37:54 if it can monitor or be poked by a trigger or what Oct 29 20:38:24 does it build every single commmit, or just every few commits and bisect if it needs? Oct 29 20:38:58 anyone can help me with a Makefile rule? Oct 29 20:39:12 zecke: can try Oct 29 20:39:15 and hi Oct 29 20:39:45 zecke: can you pastebin your question? Oct 29 20:40:23 zecke yeah tell us your problem Oct 29 20:40:34 kergoth: it seems nice but in our case it would be impossible to do one build for every single commit... because our build takes too long and there are so many commits here... Oct 29 20:40:41 I have a file with *.png and *.js files Oct 29 20:40:51 but the files often change... Oct 29 20:41:01 a file with files? Oct 29 20:41:09 hi woglinde Oct 29 20:41:10 denix: mm, dunno. like kergoth says, it does sound a bit like a cross-compile mixup Oct 29 20:41:15 he pb :) Oct 29 20:41:18 zecke: you mean, a dir with files? Oct 29 20:41:25 inspector/*.{js,png} Oct 29 20:41:31 k Oct 29 20:41:45 pb_: apparently, minimal distro is broken when host=target... Oct 29 20:41:47 vivijim: true. it's less of an issue with packaged staging though, you could build a single package for a set of targets leveraging previous build output to satisfy deps, so a commit to a single package could fire off its build Oct 29 20:41:53 I could write my rule with "install inspector/*.js" but I want proper depedency tracking Oct 29 20:42:03 Hi guys. I'd like to learn how to make opkg packages. They are quite similar to ipkg as far as I know. The reason I want to do this is that i need sympy, numpy, and scipy on my openmoko. Where to begin? Oct 29 20:42:05 e.g. if I remove one png and add another this rule is run again Oct 29 20:42:06 vivijim: wasn't really looking at it for oe though, just as another useful tool for devs to have available Oct 29 20:42:17 tomboy64: want to try a patch? Oct 29 20:42:22 :) Oct 29 20:42:36 i'd love to Oct 29 20:42:47 just hacking around myself a bit Oct 29 20:42:58 but never tried python before :-S Oct 29 20:43:07 zecke hm I saw something in the automake documentation Oct 29 20:43:19 pb_: I was able to build angstrom for x86, but had to use qemux86 machine instead of i686-generic Oct 29 20:46:07 zecke hm sorry that wasnt what you will achieve Oct 29 20:46:32 kergoth: sure! it is a nice tool... And I'd like to see one solution like that that fits OE... Oct 29 20:46:35 tomboy64: ah then try :) Oct 29 20:46:40 tomboy64: it will be fun Oct 29 20:46:42 vivijim: yeah.. i think we all would Oct 29 20:46:49 :-D Oct 29 20:47:02 * kergoth is saddened by the state of open source enterprise level build / release / CI tools Oct 29 20:47:18 tomboy64: http://paste.lisp.org/display/69404 Oct 29 20:47:33 woglinde: okay, how to make a rule execute every time? (e.g. to disable dep tracking?) Oct 29 20:48:45 lisp ... Oct 29 20:48:48 * tomboy64 shudders Oct 29 20:48:50 zecke thats an configure option Oct 29 20:48:51 :-D Oct 29 20:49:37 Automatic dependency tracking can be suppressed by putting no-dependencies in the variable AUTOMAKE_OPTIONS, or passing no-dependencies as an argument to AM_INIT_AUTOMAKE (this should be the preferred way Oct 29 20:54:09 kergoth: what's your reference for such tooling? Oct 29 20:54:31 http://github.com/apenwarr/cron2rss/tree/master Oct 29 20:54:32 cute Oct 29 20:54:32 just research I've done on my own when thinking about the possibilities of a bitbake/oe replacement Oct 29 20:54:38 That guy tinkers alot Oct 29 20:54:52 kergoth: ok, good. Oct 29 20:55:05 kergoth: closest to my ideal ci tool now is a nightle OE build :-) Oct 29 20:55:11 nightly Oct 29 20:55:30 zecke: that won't work Oct 29 20:55:31 kergoth: ideally have it deployed on the target automatically and tested against Oct 29 20:56:07 tomboy64: hehe, then make it work :) Oct 29 20:56:22 the way i see it, bitbake is just like any other enterprise level build/release tool. it builds lots of individual components to result in a whole. it's just one time, rather than daemonized, and is missing some other features.. so thats why i was doing such research Oct 29 20:56:26 ci tools are probably the number one need at the enterprise that rolls their own Oct 29 20:57:11 zecke: the paste swallowed "odule)" in the last +-line Oct 29 20:57:40 ah okay Oct 29 20:57:50 tomboy64: send a patch when you are done and tested it :) Oct 29 20:57:57 oki Oct 29 21:01:43 Sorry, I feel like I'm boosing in, but coul any of you guys point me in the right direction on how to make ipkg/opkg packages from an svn checkout. I'm a bit stumped, but would love to maintain some packages for the openmoko. Oct 29 21:26:55 finally Oct 29 21:27:00 * tomboy64 's illume builds Oct 29 21:27:11 hehe Oct 29 21:30:04 oh, how do I execute shell from a Makefile? Oct 29 21:30:12 huh? Oct 29 21:30:25 like sh somefile? Oct 29 21:30:27 zecke???? Oct 29 21:31:20 like A = `ls foo/*.ong` Oct 29 21:31:43 A = $(shell ls foo/*.ong) Oct 29 21:35:53 zecke: is that what you want? Oct 29 21:36:05 let me see :) Oct 29 21:42:30 morning all Oct 29 21:42:38 morning Oct 29 21:44:09 morning Oct 29 21:45:26 hi thesing Oct 29 22:30:44 hmm, why's there a link to openembedded.net in the topic when it just redirects to .org? and is .org acting up for anyone else? Oct 29 23:59:37 * otavio doing another firefox build Oct 30 00:00:13 otavio: for which arch? Oct 30 00:05:26 likewise: i586; it is failing due a bz2 linking issue. I've changed it to use system bzip2 and will give it a try. Let's see how it goes Oct 30 00:06:12 otavio: ok. we should bring our x86 support in shape again. it has been on my todo for long. Oct 30 00:07:18 otavio: you are working with x86 targets much? Oct 30 00:07:32 likewise: it is my only target ATM Oct 30 00:07:53 likewise: what about you? Oct 30 00:08:10 * otavio waits firefox to build ... it takes looooong time Oct 30 00:09:33 otavio: i would like to target core 2 duo and such Oct 30 00:09:42 likewise: wow! Oct 30 00:09:48 likewise: what kind of usage? Oct 30 00:10:01 likewise: and why not use a regular GNU/Linux system? Oct 30 00:10:19 03Felix Domke  07org.openembedded.dreambox * r455a0761da 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.opendreambox.org/git/obi/opendreambox-1.5 into org.openembedded.dreambox Oct 30 00:10:31 03Andreas Oberritter  07org.openembedded.dreambox * r3a3061c379 10openembedded.git/packages/wpa-supplicant/ (10 files in 2 dirs): Oct 30 00:10:31 wpa-supplicant: backport 0.6.3 from dev (disabled) Oct 30 00:10:31 * applied some patches from openwrt Oct 30 00:10:31 * ported zydas patch from 0.5.8, but doesn't seem to work Oct 30 00:10:32 * integrated ralink driver doesn't seem to work either Oct 30 00:10:41 03ghost  07org.openembedded.dreambox * rf041b76f95 10openembedded.git/packages/tuxbox/ (2 files in 2 dirs): tuxbox-tuxtxt-32bpp.bb: follow e2 eConsoleAppContainer change Oct 30 00:10:42 03Andreas Oberritter  07org.openembedded.dreambox * r56eff37943 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.openembedded.net/openembedded into org.openembedded.dreambox Oct 30 00:16:56 likewise: I just had quite a bad experience earlier today with x86 target - things like minimal distro broken for x86 etc. Oct 30 00:18:34 denix: I saw that. I expect more need for x86 now that the "embedded" market expands with all kinds of x86 toys. Oct 30 00:18:35 likewise: ended up building angstrom for qemux86 target, as i686-generic target was also failing for angstrom Oct 30 00:19:37 likewise: would be nice to have x86 stabilized Oct 30 00:23:05 denix: I have i586 working Oct 30 00:23:11 denix: it doesn't seem too bad for me Oct 30 00:23:21 denix: mine current system even has a working X Oct 30 00:23:25 otavio: yeah, that's what worked for me too Oct 30 00:23:43 denix: firefox fails though Oct 30 00:23:52 hi kergoth Oct 30 00:23:53 qemux86 is i586 and angstrom Oct 30 00:24:01 denix: I have XFCE, X and like Oct 30 00:24:03 otavio: which distro are you using? Oct 30 00:24:52 the problem was with setting target to i686 and -march=pentiumpro - glibc-2.6.1 fails to build with weird ASM errors Oct 30 00:26:44 denix: angstrom based Oct 30 00:26:58 denix: few changes against it but I do include it in our distro recipe Oct 30 00:27:47 denix: well, i build to i586 so I can not be hit by this bug Oct 30 00:28:47 did setting the vendor not work? Oct 30 00:28:54 that's how i always used to do that sort of build Oct 30 00:29:23 otavio: so, have you seen this issue also with glibc-2.6.1 failing for i686? Oct 30 00:29:48 ah, you got past the libtool one. what's the error on this one? Oct 30 00:32:20 kergoth: sorry, I haven't tried setting vendor for minimal distro, as I was able to get angstrom built by downgrading to i586 Oct 30 00:32:24 denix: as I said, I build to i586; it works here Oct 30 00:32:26 ah Oct 30 00:32:31 thatd do too Oct 30 00:32:52 Kensan: once I have time, I may go back to minimal with vendor setting, like you suggested Oct 30 00:33:05 sorry, last msg meant for kergoth Oct 30 00:33:23 well, top priority is to make it work, go from there :) Oct 30 00:34:06 otavio: Ok, I'm just trying to debug that strange issue with i686... Oct 30 00:35:10 good luck guys. I'll try to get x86 stuff building here soon and see along what I can fix. Oct 30 00:35:16 need sleep now -cya Oct 30 00:36:18 cool; firefox is built Oct 30 00:46:28 03Otavio Salvador  07org.openembedded.dev * r9bac655320 10openembedded.git/conf/checksums.ini: (log message trimmed) Oct 30 00:46:28 checksums.ini: new checksums for a set of packages Oct 30 00:46:28 - LibVNCServer-0.9.1 Oct 30 00:46:28 - MesaLib-7.0.2 Oct 30 00:46:28 - acpid-1.0.6 Oct 30 00:46:30 - console-tools-0.3.2 Oct 30 00:46:32 - dbh_1.0-18 Oct 30 00:46:36 03Otavio Salvador  07org.openembedded.dev * r7f8f1ac6f2 10openembedded.git/packages/xfce/thunar_0.9.0.bb: thunar: recommends shared-mime-info to properly support .desktop files Oct 30 00:46:39 03Otavio Salvador  07org.openembedded.dev * r45c74cbeb6 10openembedded.git/packages/mozilla/firefox-3.0.1/mozconfig: firefox: use system bz2 library to avoid missing symbols Oct 30 00:46:47 03Otavio Salvador  07org.openembedded.dev * rd569e2df3c 10openembedded.git/packages/xorg-xserver/xorg-xserver-common.inc: Oct 30 00:46:47 Revert "xorg-xserver-common.inc: set userid to allow Xorg to be run by users" Oct 30 00:46:49 This reverts commit 83176ca41497b63eab25babf275eb09b5de55a8d. Oct 30 00:50:09 03Otavio Salvador  07org.openembedded.dev * rfb43a0277a 10openembedded.git/packages/initscripts/initscripts-1.0/checkroot.sh: initscripts: do /proc remount instead of mounting it Oct 30 00:54:19 * denix is away: I'm not here Oct 30 00:54:31 lo Oct 30 00:54:56 someone can to tell me what is that or why in bitbake pls? ERROR: Please set the 'CACHE' variable. Oct 30 00:58:42 ALoGeNo: you need to configure your local.conf file and set BBPATH to the root of your OE environment Oct 30 00:59:15 i did it, hola otavio ;) Oct 30 00:59:23 hola! Oct 30 01:00:16 man, im crazy with local.conf, the stuff {HOME$}/stuff/ Oct 30 01:00:38 is like /home/user/stuff no? Oct 30 01:01:24 otavio ? Oct 30 01:01:25 ${HOME} you meant? Oct 30 01:01:37 yes sorry im too n00b Oct 30 01:01:50 ALoGeNo: local.conf is pretty easy; just read the comments carefully Oct 30 01:02:24 ALoGeNo: however I suggest to you to take a look on the wiki for the Getting Started document; otherwise you can get lost in the OE details pretty easily. Oct 30 01:02:31 yep i did many times, ${HOME} is like /home/user/ no? Oct 30 01:02:38 ALoGeNo: yep Oct 30 01:02:50 okey Oct 30 01:03:14 you are in the rigth i need to do that from the begginnig.. Oct 30 01:03:53 i was used two tutos, and i think that i mix something Oct 30 01:28:57 otavio still there? Oct 30 01:29:07 can you take a look please? http://pastebin.com/m635eb5f0 Oct 30 01:29:59 i think that is all ok , the falut maybe is the export stuff paths Oct 30 01:30:22 what is your opinion? Oct 30 01:31:07 from a fast looking it looks fine Oct 30 01:31:18 aham Oct 30 01:31:32 I believe you forgot to: export BBPATH=$(pwd) Oct 30 01:31:45 your rigth **** ENDING LOGGING AT Thu Oct 30 01:32:09 2008 **** BEGIN LOGGING AT Thu Oct 30 01:32:37 2008 Oct 30 01:33:04 how much space you need to build a full image? Oct 30 01:36:37 otavio when it say export BBPATH=/stuff/build:/stuff/openembedded its realy means BBPATH=/home/user/stuff/build:/home/user/stuff/openembedded no? Oct 30 01:36:50 i will do that like user or root? Oct 30 01:36:59 user; Oct 30 01:37:16 ^^' oki Oct 30 01:37:20 I usually do: cd ; export BBPATH=$(pwd) Oct 30 01:37:44 $(pwd) ? Oct 30 01:38:34 try pwd in your shell Oct 30 01:38:52 the i must use export BBPATH=/home/user/stuff/build:/home/user/stuff/openembedded no? Oct 30 01:39:03 then* Oct 30 01:39:32 OHH Oct 30 01:39:39 pwd niceeee... Oct 30 01:40:42 in this way? export BBPATH=$(pwd)/stuff/build:$(pwd)/stuff/openembedded no? Oct 30 01:41:19 ALoGeNo: I only do: export BBPATH=$(pwd) Oct 30 01:41:35 oh.. i see.. Oct 30 01:41:55 you have stuff user or something? Oct 30 01:42:10 http://wiki.openembedded.net/index.php/Getting_Started Oct 30 01:42:11 no Oct 30 01:42:26 yeah im in this place octavio :) Oct 30 01:43:02 but is a little confuse at the end (for my level of english.. :) ) Oct 30 01:43:55 where is your OE checkout? Oct 30 01:44:12 from clone git ;) Oct 30 01:44:16 yes Oct 30 01:44:23 but where in your disk? Oct 30 01:44:38 mine is at: ~/hacking/ossystems/oe Oct 30 01:44:42 oh is in /home/usr/stuff Oct 30 01:44:52 and inside of stuff has? Oct 30 01:44:58 oh wait Oct 30 01:45:06 oh is in /home/usr/stuff/openembedded Oct 30 01:45:17 ALoGeNo: so: cd /home/usr/stuff/openembedded Oct 30 01:45:23 ALoGeNo: export BBPATH=$(pwd) Oct 30 01:45:32 oki doki Oct 30 01:45:32 ALoGeNo: bitbake foo Oct 30 01:45:41 foooooo Oct 30 01:45:43 ^^ Oct 30 01:45:59 foo is just an stupid example; use a recipe you wish Oct 30 01:46:27 foo / bar / foobar are words used when we don't bother to find a real example Oct 30 01:46:29 ;P Oct 30 01:47:14 ALoGeNo: bitbake minimal-image Oct 30 01:47:22 ALoGeNo: this is a real example Oct 30 01:47:51 oh oh.. :( Oct 30 01:47:53 pi@ps3:~/stuff/openembedded$ bitbake foo Oct 30 01:47:54 ERROR: no files to build. Oct 30 01:47:54 NOTE: Parsing finished. 0 cached, 0 parsed, 0 skipped, 0 masked. Oct 30 01:47:54 NOTE: build 200810300252: started Oct 30 01:47:54 OE Build Configuration: Oct 30 01:47:54 BB_VERSION = "1.8.10" Oct 30 01:47:56 METADATA_BRANCH = "" Oct 30 01:47:58 METADATA_REVISION = "" Oct 30 01:48:00 TARGET_ARCH = "INVALID" Oct 30 01:48:02 TARGET_OS = "INVALID" Oct 30 01:48:04 MACHINE = "" Oct 30 01:48:06 DISTRO = "" Oct 30 01:48:08 DISTRO_VERSION = "" Oct 30 01:48:10 TARGET_FPU = "" Oct 30 01:48:12 ERROR: The following variable(s) were not set: TARGET_ARCH, TARGET_OS Oct 30 01:48:14 Please set them directly, or choose a MACHINE or DISTRO that sets them. Oct 30 01:48:16 pi@ps3:~/stuff/openembedded$ Oct 30 01:48:21 ALoGeNo: your local.conf is not properly set Oct 30 01:48:25 what? Oct 30 01:48:35 you was show after :P Oct 30 01:48:41 ALoGeNo: you need to define the distro to be used and the machine Oct 30 01:49:01 is in /home/usr/stuff/build/conf/local.conf Oct 30 01:49:06 ALoGeNo: BTW, use a pasting service ... do not paste it in chaneel Oct 30 01:49:15 oh sorry.. Oct 30 01:49:35 nervious excuse me ^^' Oct 30 01:49:44 ALoGeNo: man; I need to go to sleep; I believe you can sort it out now. Oct 30 01:49:53 ALoGeNo: tomorrow I have a busy day Oct 30 01:50:01 oh thanks in any case octavio asta la vista ;) Oct 30 01:50:10 ALoGeNo: asta la vista! Oct 30 01:50:30 ALoGeNo: go and read the tutorial carefully; it is where I started too Oct 30 01:50:56 okey dokey :) Oct 30 01:51:03 ALoGeNo: cya! Oct 30 01:51:08 cya folks Oct 30 01:51:13 bb and tnk u Oct 30 02:44:00 hmm **** ENDING LOGGING AT Thu Oct 30 02:59:57 2008