**** BEGIN LOGGING AT Sat Oct 25 02:59:57 2008 Oct 25 08:49:03 zecke: hail zecke Oct 25 08:51:09 zecke: what do I do if git tells me that I am "in the middle of a conflicted merge"? Oct 25 08:51:29 I have, I think, fixed the conflict in the file it is upset about, but it doesn't seem to want to let me proceed. Oct 25 09:04:51 hm, maybe I should just throw this working tree away and check out a fresh one Oct 25 09:09:46 hail zecke_ Oct 25 10:12:23 03Philip Blundell  07org.openembedded.dev * r0e578dfc60 10openembedded.git/packages/glibc/ (files/armv4t-interworking.patch glibc_2.6.1.bb): glibc-2.6.1: fix interworking for armv4t Oct 25 10:22:43 ah, whoops, forgot to set my correct email address Oct 25 10:22:49 ~lart me Oct 25 10:22:49 * ibot farts in pb_'s general direction Oct 25 10:23:21 03Kalev Lember  07org.openembedded.dev * ref4996ca83 10openembedded.git/ (3 files in 3 dirs): Oct 25 10:23:21 flashrom svn: new recipe Oct 25 10:23:21 * add makefile patch to not strip executable during install, we do that while packaging Oct 25 10:44:16 03Andreas Oberritter  07org.openembedded.dreambox * rfe12a7a495 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.openembedded.net/openembedded into org.openembedded.dreambox Oct 25 10:44:23 03ghost  07org.openembedded.dreambox * r47765381d3 10openembedded.git/packages/enigma/ (enigma/opendreambox/userbouquet.33fc5.tv enigma_cvs.bb): update default bouquet Oct 25 10:44:24 03ghost  07org.openembedded.dreambox * rae7313ea54 10openembedded.git/packages/dreambox/dreambox-dvb-modules.bb: update dm500plus/dm600pvr driver modules Oct 25 10:44:24 03ghost  07org.openembedded.dreambox * r2fe153b614 10openembedded.git/packages/dreambox/dreambox-blindscan-utils.bb: update dm500plus / dm600pvr blindscan utils Oct 25 10:44:25 03Andreas Oberritter  07org.openembedded.dreambox * r418cab5db0 10openembedded.git/packages/initscripts/ (5 files in 5 dirs): (log message trimmed) Oct 25 10:44:28 initscripts-opendreambox: consolidate initscripts Oct 25 10:44:30 - don't load lcd.ko (already loaded by rcS) Oct 25 10:44:32 - don't mount proc (already mounted by rcS) Oct 25 10:44:34 - don't hide errors by redirecting stderr Oct 25 10:44:36 - use /dev/discs/disc0/disc for the primary hdd Oct 25 10:44:38 - dm7020: fix boot logo Oct 25 10:44:40 03ghost  07org.openembedded.dreambox * r3740dcb4d3 10openembedded.git/packages/dreambox/dreambox-secondstage.bb: update dm500plus,dm600pvr,dm800 secondstage loader Oct 25 10:45:01 03Felix Domke  07org.openembedded.dreambox * r8b5423f421 10openembedded.git/: Merge branch 'org.openembedded.dreambox' of git://git.opendreambox.org/git/obi/opendreambox-1.5 into org.openembedded.dreambox Oct 25 10:45:01 03ghost  07org.openembedded.dreambox * r3e1889b2be 10openembedded.git/packages/dreambox/dreambox-secondstage.bb: dreambox-secondstage: bump dm800 second to 67 Oct 25 11:21:06 morning Oct 25 11:23:12 hi Oct 25 11:23:16 morninng hrw Oct 25 11:26:53 03Koen Kooi  07org.openembedded.dev * rc1bd0fc549 10openembedded.git/conf/machine/include/ (7 files): Oct 25 11:26:53 arm tune files: include tune-thumb by default Oct 25 11:26:53 * it defaults to non-thumb, so no functional change Oct 25 13:11:00 pb_: when you solves a git conflict you need to git add it; so git knows it is fixed and will get you go ahead Oct 25 14:12:25 <[gnubie]> am i the only one seeing this site => http://wiki.openembedded.net/index.php/OEandYourDistro that looks like it has been defaced? Oct 25 14:12:54 [gnubie]: typical wiki spam Oct 25 14:13:15 <[gnubie]> i see.. Oct 25 14:13:37 [gnubie]: I think there are spam crawlers that traverse the internet looking for specific wikis. Anything they're equipped to auto-post on really Oct 25 14:14:12 <[gnubie]> i see Oct 25 14:17:38 03Otavio Salvador  07org.openembedded.dev * rf0c403283a 10openembedded.git/conf/checksums.ini: checksums.ini: add libpng 1.2.31 Oct 25 14:17:39 03Otavio Salvador  07org.openembedded.dev * r59e375ec7a 10openembedded.git/conf/checksums.ini: checksums.ini: add flex 2.5.31 Oct 25 14:17:40 wiki changes can be undone on most wikis anyway. I just looked how to do it myself, but I'm not familiar with mediawiki Oct 25 14:17:49 03Otavio Salvador  07org.openembedded.dev * rde5d675749 10openembedded.git/conf/checksums.ini: checksums.ini: add freetype 2.3.6 Oct 25 14:17:50 03Otavio Salvador  07org.openembedded.dev * rb222ee6682 10openembedded.git/conf/checksums.ini: checksums.ini: add expat 1.95.7 Oct 25 14:17:50 03Otavio Salvador  07org.openembedded.dev * rb57972de43 10openembedded.git/conf/checksums.ini: checksums.ini: add expat 2.0.0 Oct 25 16:44:30 hi people Oct 25 16:45:01 no do_builtin_initramfs anymore - there something else that does this or do i have to do it myself now? Oct 25 16:59:00 03Koen Kooi  07org.openembedded.dev * r1a0a603ec7 10openembedded.git/classes/package.bbclass: package bbclass: only run ldconfig when there's a ld.so.conf present Oct 25 19:11:20 03Holger Hans Peter Freyther  07org.openembedded.dev * rf4fcadb061 10openembedded.git/packages/gstreamer/ (3 files): Oct 25 19:11:20 gstreamer: mp3 was never needed for Openmoko it was only there accidently Oct 25 19:11:20 ogg/vorbis is actually what we want! Oct 25 19:43:05 anyone here? @@ Oct 25 19:43:35 probably pretty many Oct 25 19:47:49 Does anyone have suggestions to this problem I have been having? http://article.gmane.org/gmane.comp.handhelds.openembedded.user/565 Oct 25 20:42:55 <_bernie> aleph_6510: hey there Oct 25 20:44:11 <_bernie> aleph_6510: what was the pixie powder to make bitbake bake multiple things at once? Oct 25 20:45:45 BB_NUMBER_THREADS=2 or 3 or 4... in local.conf Oct 25 20:46:44 PARALLEL_MAKE="-j2" to pass -j2 to make and build multiple things at once Oct 25 20:47:00 <_bernie> aleph_6510: this should spawn multiple jobs at once, right? orthogonal with make -j2 Oct 25 20:47:05 the latter is more useful I think Oct 25 20:47:31 but the former parallelize even things like ./configure Oct 25 20:49:24 A number of the bitbake recipes will override PARALLEL_MAKE because the particular package fails when build that way; if you have a number of those, or one very large one, then the BB_NUMBER_THREADS is the most useful. However, both have the ability to completely load your system :) Oct 25 20:49:29 <_bernie> aleph_6510: also, you should do OE_ALLOW_INSECURE_DOWNLOADS = "yes" Oct 25 20:49:51 <_bernie> mwester: oh thanks the tip Oct 25 20:51:06 <_bernie> mwester: I run bitbake prefixed with "please", which is: please() { nice ionice -c2 -n6 "$@" } Oct 25 20:51:12 <_bernie> but it doesn't help for memory pressure Oct 25 20:51:16 mwester: I understand... so BB_NUMBER_THREADS works always. nice Oct 25 20:51:24 <_bernie> there should be a "swap priority" attached to processes Oct 25 20:51:45 <_bernie> aleph_6510: no, only when bitbake finds two things that can be built at once Oct 25 20:52:12 <_bernie> aleph_6510: probably not during the first steps when the cross toolchain needs to be built Oct 25 20:52:50 Nor during the last step while the image itself is being built. Oct 25 20:53:47 PARALLEL_MAKE will help for the toolchain construction in the beginning, but AFAIK there is nothing to be done for the image generation step -- that will just run along using a single processor :( Oct 25 20:54:24 _bernie: I think we could use GLIBC_GENERATE_LOCALES="en_US.UTF-8" to avoid generating a billion of language ipkg Oct 25 20:59:14 <_bernie> aleph_6510: I did that already :) Oct 25 20:59:44 <_bernie> I get spurious md5 failures that disappear if I rerun bitbake a second time Oct 25 20:59:58 <_bernie> and on another host, package downloads seem to hang Oct 25 21:25:12 _bernie: What is OE_ALLOW_INSECURE_DOWNLOADS for? To bypass the checksum check? Oct 25 21:26:36 <_bernie> likewise: yes. there are many source packages in the bb files with missing checksums Oct 25 21:49:21 hi, i wonder how i can force bb to build a package native (i686 for example) too - some packages like compilers are native already, how does bb decide which ones? Oct 25 21:50:11 Someone creates a recipe for native builds. Oct 25 21:51:46 mwester: how? is that if a DEPENDS is in it, not a RDEPENDS? Oct 25 21:52:07 I'm not sure what you are asking? How what? Oct 25 21:52:32 i need a package build for my host arch. how can i achieve that? Oct 25 21:53:23 You build (by convention) "-native" --- but -native.bb must have already been created by someone. Oct 25 21:54:08 or by me ;) Oct 25 21:54:40 Sure. There's nothing hard about them, really. Oct 25 21:55:01 inherit native is what's needed? Oct 25 21:55:29 in the minimal case, I think that would do it. Oct 25 22:04:03 okay, understood & was able to fix the scummvm build problem, thnx mwester :D Oct 25 22:04:55 (which was a different prblem than the one because i asked) Oct 25 22:08:39 :) Oct 25 22:08:45 mwester: can you also tell me the cause for the new conf/checksum.ini? Oct 25 22:08:56 healthy paranoia Oct 25 22:09:11 how can i come around this? Oct 25 22:09:42 ..there is a checklsum in source/... already, hm... strange Oct 25 22:10:48 http://wiki.openembedded.net/index.php/Dealing_with_checksum_mismatch Oct 25 22:11:45 The most common problem (and a serious one) is that upstream will change the content of a release tarball or patch -- this breaks configuration management because you and I might build the same thing, but get totally different results. Oct 25 22:12:05 So the checksums approach ensures that when this happens, at least the community here is aware that someone did that. Oct 25 22:12:32 The last one I ran into recently was a patch to busybox -- they changed the patch after they had it published for some time. Oct 25 22:13:05 The checksums are stored in openembedded/conf somewhere, in checksums.ini Oct 25 22:13:17 (they're actually md5 and sha1 sums. Oct 25 22:13:19 ) Oct 25 22:19:27 yes, i saw it, i added sums for my packages. problem is, i change my packages often without always changing the name because i have bb in my dev cycle - that would explode the version numbers Oct 25 22:19:44 is there a command to update checksum.ini? Oct 25 22:27:46 Nope, no automagic command. Oct 25 22:29:33 shit, the native stuff does not work that simple in my case. i try to build py-e bindings and they use PN as a part of the svn-url... :/ Oct 26 02:54:49 g'evening **** ENDING LOGGING AT Sun Oct 26 02:59:57 2008