**** BEGIN LOGGING AT Fri Oct 24 02:59:57 2008 Oct 24 04:43:00 03Mike Westerhof  07org.openembedded.dev * r7ca90a798f 10openembedded.git/conf/machine/ (om-gta01.conf om-gta02.conf): Oct 24 04:43:00 FSO: om-gta0?.conf - remove duplicate armv4t in package architecture lists Oct 24 04:43:00 (Fixes breakage caused by Koen's commit to move PACKAGE_EXTRA_ARCHS into Oct 24 04:43:00 the tune files) Oct 24 05:18:33 03Mike Westerhof  07org.openembedded.dev * r05dd06fb5d 10openembedded.git/classes/package_ipk.bbclass: Oct 24 05:18:33 package_ipk.bbclass: add import bb where missing Oct 24 05:18:33 (fixes build failure for pointercal, and probably others) Oct 24 05:51:59 03Kalev Lember  07org.openembedded.dev * ra2f917e983 10openembedded.git/conf/machine/ (i586-generic.conf i686-generic.conf): Oct 24 05:51:59 i?86-generic.conf: remove duplicate entries from PACKAGE_EXTRA_ARCHS Oct 24 05:51:59 * fixes breakage caused by the commit which moved PACKAGE_EXTRA_ARCHS Oct 24 05:51:59 into the tune files Oct 24 07:23:01 mickey|munich: good morning Oct 24 07:27:17 good morning pb_ Oct 24 07:33:21 03Koen Kooi  07org.openembedded.dev * rf45c9735ef 10openembedded.git/conf/machine/davinci-dvevm.conf: davinci-dvevm: remove dsplink from RDEPENDS, it isn't strictly needed Oct 24 07:33:24 morning Oct 24 07:37:31 mickey|munich: how much longer will you stay in bavaria? Oct 24 07:46:12 bonjour Oct 24 07:49:54 pb_: taking the train back around 15:00 today Oct 24 07:50:07 attending a Mac/iPhone workshop tomorrow Oct 24 07:50:58 ah, very good Oct 24 07:51:06 one of my colleagues has just signed up to the iphone developer program thing Oct 24 07:51:25 he hasn't really done much with it yet though Oct 24 07:52:57 i signed up two months ago Oct 24 07:53:07 have been slowly starting to learn Oct 24 07:55:29 hope the workshop is useful. Oct 24 07:55:47 today I'm going to try to debug my thumb thing again. Oct 24 07:55:59 I had forgotten how bad it is to do thumb interworking on armv4t. Oct 24 07:56:31 still, it should be possible to make it work I think. Oct 24 08:00:50 * pb_ go to office now Oct 24 08:00:50 bbiab Oct 24 08:13:29 http://blog.haerwu.biz/2008/10/24/merging-stuff-from-poky-into-openembedded/ Oct 24 08:22:10 03Koen Kooi  07org.openembedded.dev * r821a525bf4 10openembedded.git/packages/tasks/task-base.bb: task-base: bump PR Oct 24 08:27:50 mickey|munich: ping Oct 24 08:28:22 mickey|munich: are we using ompower or empower in OE (fso image ?) Oct 24 08:32:04 re Oct 24 08:36:46 morning all Oct 24 08:37:57 hi RP Oct 24 08:50:58 argh Oct 24 08:51:01 * pb__ stabs u-boot Oct 24 08:51:13 who had this stupid idea of making an empty line mean "repeat the last command"? Oct 24 08:51:22 yep Oct 24 08:52:09 very annoying, I just flashed a new rootfs with dfu-util, then accidentally hit return in u-boot and erased it again. Oct 24 08:52:29 03Koen Kooi  07org.openembedded.dev * rfa0d83dad9 10openembedded.git/conf/machine/neuros-osd2.conf: neuros osd2: remove dsplink from machine recommends, they are not strictly needed Oct 24 08:52:32 * pb__ waits for slow download once more Oct 24 08:52:41 pb_, FAIL :P Oct 24 08:53:06 pb: s3c? Oct 24 08:53:18 yes Oct 24 08:53:33 I think u-boot does this on all platforms though, I have had the same trouble on one of our marvell boards Oct 24 08:53:45 * Hrw|n810 waits for i.mx31 hw Oct 24 08:54:06 yes, thats uboot inside Oct 24 09:00:22 johncylee: hi Oct 24 09:00:37 johncylee: neither one is used in fso. didn't even look at it yet Oct 24 09:00:50 johncylee: i think the framework handles almost everything ompower was designed to do Oct 24 09:00:53 mickey|munich: hi. i just found out neither ompower nor empower are in OE Oct 24 09:01:03 yes. we don't have anything that used it Oct 24 09:01:10 mickey|munich: understood. thanks. Oct 24 09:01:31 if you want to build om2008 out of OE, feel free to add it Oct 24 09:01:50 johncylee: FYI: the fso restructuring for Om is nearly finished. Oct 24 09:02:02 did most of the tasks already, just missing the final image recipes Oct 24 09:02:10 will send a mail once it is done Oct 24 09:11:46 phew, flashing finished Oct 24 09:11:50 * pb__ very careful not to press return again Oct 24 09:15:10 grr, stupid openssh and its host key checking Oct 24 09:15:12 * pb__ stabs it too Oct 24 09:19:40 patch your ssh config to ignore it for the Neo Oct 24 09:20:13 oh, you can do that? what's the config setting? Oct 24 09:22:02 let me check Oct 24 09:22:17 Host neo Oct 24 09:22:17 HostName 192.168.0.202 Oct 24 09:22:17 StrictHostKeyChecking no Oct 24 09:22:17 UserKnownHostsFile /dev/null Oct 24 09:22:17 User root Oct 24 09:22:36 (mutis mutandis) Oct 24 09:22:43 (mutatis mutandis, even) Oct 24 09:23:49 ah, clever Oct 24 09:23:59 I hadn't realised you could set the known hosts file to /dev/null for a particular host Oct 24 09:24:21 * pb__ <3 mickeyl Oct 24 09:27:34 doh, glibc's memmove() isn't interworking safe Oct 24 09:27:40 maybe this is why my thumb build keeps crashing Oct 24 09:32:10 d'oh Oct 24 09:32:21 same for clone() Oct 24 09:32:33 not that I think busybox is likely to be using clone(), but still sucks that it wouldn't work Oct 24 09:32:44 stupid armv4t Oct 24 09:36:19 cee you Oct 24 09:52:42 03Koen Kooi  07org.openembedded.dev * r54146066ad 10openembedded.git/packages/ppp-dialin/ppp-dialin_0.1.bb: pp-dialin: try to make it upgrade safe Oct 24 09:55:01 doh, busybox still crashes even with those two fixed Oct 24 09:55:06 must be something else lurking in there Oct 24 10:00:40 oh, memcpy() has the same problem Oct 24 10:00:40 drat Oct 24 10:06:50 nasty Oct 24 10:07:01 well, easily fixed Oct 24 10:07:09 I've corrected that one as well now and busybox does run at least Oct 24 10:07:45 not quite sure how I failed to spot memcpy the first time :-} Oct 24 10:09:55 * pb__ stabs armv4t again for good measure Oct 24 10:10:04 hail zecke Oct 24 10:12:25 moin, I see you have fun? Oct 24 10:12:39 you try to understand why 2008 silicon is slower than strong arm from 1999? Oct 24 10:13:00 heh. well, no, I am taking a break from that at the moment, trying to make my thumb build work on gta01. Oct 24 10:13:22 it turns out that glibc in oe lacks interworking ability for v4t in a few important functions :-} Oct 24 10:13:35 so, any call to memmove() or memcpy() is death Oct 24 10:14:08 ibot: ugt Oct 24 10:14:09 ugt is probably Universal Greeting Time. Created in #mipslinux, it is a rule that states that whenever somebody enters an IRC channel it is always morning, and it is always late when the person leaves. The local time of any other people in the channel, including the greeter, is irrelevant. http://www.total-knowledge.com/~ilya/mips/ugt.html Oct 24 10:14:35 on the performance front I will try to get one of my colleagues to benchmark one of our own s3c2410 boards and see whether it is as bad as the neo. Oct 24 10:14:53 not sure at the moment whether the go-slow is the cpu's fault or whether it is due to some special openmoko thing Oct 24 10:16:12 pb__: likely special openmoko magic :) Oct 24 10:17:06 fear Oct 24 10:21:39 hmm, I'm still puzzled why xorg decided that all my keycodes are different now :( Oct 24 10:45:32 03Jeremy Lainé  07org.openembedded.dev * r0d4559f79c 10openembedded.git/packages/gnutls/ (gnutls.inc gnutls_1.4.5.bb gnutls_1.6.3.bb gnutls_2.4.2.bb): gnutls: fixing the pkgconfig is not necessary for gnutls 2.4.2 Oct 24 11:03:50 03Florian Boor  07org.openembedded.dev * r8655c6c036 10openembedded.git/packages/linux/linux-bd-neon-2.6_2.6.22.bb: linux-bd-neon: Set COMPATIBLE_MACHINES Oct 24 12:39:55 03Cliff Brake  07org.openembedded.dev * r97f239bff8 10openembedded.git/: Merge commit 'openembedded.org/org.openembedded.dev' into org.openembedded.dev Oct 24 12:40:52 03Cliff Brake  07org.openembedded.dev * r868fe416fa 10openembedded.git/MAINTAINERS: MAINTAINERS: add vala to my recipe list Oct 24 12:43:19 03Cliff Brake  07org.openembedded.dev * ra57a4df1e3 10openembedded.git/ (5 files in 2 dirs): vala 0.4.0: add new version Oct 24 12:44:40 03Cliff Brake  07org.openembedded.dev * r703cea83a1 10openembedded.git/: Merge commit 'openembedded.org/org.openembedded.dev' into org.openembedded.dev Oct 24 12:49:43 re Oct 24 12:54:53 03Koen Kooi  07org.openembedded.dev * re16c6ec539 10openembedded.git/ (3 files in 2 dirs): vlc: update to 0.9.4 and add v4l1 support Oct 24 13:00:11 hmm for some reason my commits above seem to have dissapeared from org.openembedded.dev?? Oct 24 13:03:57 pushed again, and now my changes are there Oct 24 13:04:01 strange Oct 24 13:04:47 03Cliff Brake  07org.openembedded.dev * ra57a4df1e3 10openembedded.git/ (5 files in 2 dirs): vala 0.4.0: add new version Oct 24 13:05:10 03Cliff Brake  07org.openembedded.dev * r97f239bff8 10openembedded.git/: Merge commit 'openembedded.org/org.openembedded.dev' into org.openembedded.dev Oct 24 13:05:10 03Cliff Brake  07org.openembedded.dev * r703cea83a1 10openembedded.git/: Merge commit 'openembedded.org/org.openembedded.dev' into org.openembedded.dev Oct 24 13:05:11 03Cliff Brake  07org.openembedded.dev * r29ab598543 10openembedded.git/: Merge commit 'openembedded.org/org.openembedded.dev' into org.openembedded.dev Oct 24 13:05:23 03Cliff Brake  07org.openembedded.dev * r868fe416fa 10openembedded.git/MAINTAINERS: MAINTAINERS: add vala to my recipe list Oct 24 13:06:03 3 merges? :) Oct 24 13:10:08 Hrw|n810: I really don't know what is going on yet ... Oct 24 13:10:29 Hrw|n810: after the first set of commits, I saw my changes on http://git.openembedded.net/?p=openembedded.git;a=summary Oct 24 13:10:39 Hrw|n810: then they dissapeared Oct 24 13:12:32 no idea why Oct 24 13:15:39 Hrw|n810: well, I guess I'll keep an eye on them -- hope its not a git bug ... Oct 24 13:30:46 cbrake: it is a good idea to do "git pull --rebase" before pushing to reduce merges Oct 24 13:40:09 see you Oct 24 13:46:35 anyone is trying intrepid as well? Oct 24 13:59:17 03Koen Kooi  07org.openembedded.dev * r01658efef0 10openembedded.git/packages/vlc/ (vlc_0.9.2.bb vlc_0.9.4.bb): vlc: fix v4l option Oct 24 14:19:50 kalev: good idea, thanks Oct 24 14:35:20 Hi, some romanian or italian OE expert who can help me? Oct 24 14:48:38 03Mike Westerhof  07org.openembedded.dev * r4f984c778f 10openembedded.git/conf/distro/include/slugos.inc: Oct 24 14:48:38 SlugOS: slugos.inc - set preferred version for gnutls Oct 24 14:48:38 (our autotools is a bit too old at the moment) Oct 24 14:54:55 consider: Oct 24 14:54:55 PREFERRED_PROVIDERS += " virtual/${TARGET_PREFIX}gcc:gcc-cross" Oct 24 14:54:55 which real path "virtual/" directory represents? Oct 24 14:55:33 Claudiu85: any package can add that to their PROVIDES Oct 24 14:55:48 thats the whole point of it, it's not one specific package Oct 24 14:56:24 in that case you're saying you want gcc-cross as the specific provider of that functionality. it doesnt tell you where it resides Oct 24 14:56:42 but if I want to tell to OE where to find it? Oct 24 14:56:48 what? Oct 24 14:56:50 I built a toolchain Oct 24 14:56:53 gcc Oct 24 14:57:07 gcc-cross is in the openembedded repository, in packages/gcc/. Oct 24 14:57:09 what's the problem? Oct 24 14:57:41 I built a custom toolchain, and I want to link OE to it Oct 24 14:58:31 I found this in local.conf: Oct 24 14:58:31 # Uncomment this if you want to use a prebuilt toolchain. You will need to Oct 24 14:58:31 # provide packages for toolchain and additional libraries yourself. You also Oct 24 14:58:31 # have to set PATH in your environment to make sure BitBake finds additional binaries. Oct 24 14:58:31 # riga originale Oct 24 14:58:32 ASSUME_PROVIDED += "virtual/${TARGET_PREFIX}gcc virtual/libc" Oct 24 15:04:31 yes. that tells OE that your system already provides those, so it doesn't need to build the packages that provide them. Oct 24 15:07:14 ok, thanks ;) Oct 24 15:08:41 I am a novice with OE Oct 24 15:12:35 np Oct 24 15:18:06 hi Oct 24 15:18:50 what to do with for instance libncurses that fails the sanity test at the end of system build ? Oct 24 15:26:42 morning kergoth Oct 24 15:26:54 hey Oct 24 15:48:06 heh, I guess mickey must be going through a tunnel. Oct 24 15:50:13 heh Oct 24 17:06:54 Hmmm Oct 24 17:07:21 Has anyone played around with making the OE builds make use of the PKG_CONFIG_PATH / SYSROOT stuff the SDK exports use? Oct 24 17:09:00 Hello all Oct 24 17:34:17 hi mickey|ICExyz Oct 24 17:34:23 my rootfs is failing because opkg-cl reports the following error: Oct 24 17:34:23 ERROR: duplicate src declaration. Skipping: Oct 24 17:34:23 src oe-armv7a file:/OE/build/tmp/deploy/glibc/ipk/armv7a Oct 24 17:34:23 Though this error doesn't seem to be critical as if I disable the log_check_rootfs, everything goes through and is fine Oct 24 17:34:23 Do you know what is this error? Is it really critical? How to get more log info from opkg-cl? Oct 24 17:34:39 hi Oct 24 17:34:39 30 more minutes... Oct 24 17:34:44 * mickey|ICExyz pretty tired Oct 24 17:34:47 gregoiregentil: you probably have a duplicate entry in PACKAGE_ARCHS Oct 24 17:35:18 "ICE xyz" sounds pretty special. Have DB laid on a custom train for you? Oct 24 17:35:53 or did you just pick a train at random without looking at which one it was? Oct 24 17:36:06 hehehe, yeah, i'm a VIP, did you forget? Oct 24 17:36:06 the latter, actually Oct 24 17:36:12 there's chaos in germany Oct 24 17:36:17 "anywhere must be better than Munich, just get me out of here" Oct 24 17:36:34 oh dear, what sort of chaos? Oct 24 17:36:48 half the ICE trains are being maintained Oct 24 17:36:49 some special wheel problem Oct 24 17:36:59 so lots of trains got cancelled Oct 24 17:37:01 and i had to find my way to frankfurt somehow Oct 24 17:37:22 pb_: where is defined this PACKAGE_ARCHS? I can only find PACKAGE_INSTALL Oct 24 17:37:39 mickey|ICExyz: oh dear, that does sound annoying Oct 24 17:37:48 gregoiregentil: somewhere in your MACHINE.conf or its inclusions Oct 24 17:38:18 mickey|ICExyz: in better news, my thumb build does seem to boot now that glibc is repaired Oct 24 17:38:44 pb__: ah, that sounds good. so you basically had to rewrite glibc? Oct 24 17:38:53 heh, no, just change three assembler files Oct 24 17:39:02 I'll check in a patch later this evening Oct 24 17:39:37 due to stupid armv4t, you can't use "pop {rn, ..., pc}" to return from a function if interworking might be in use, because ldm doesn't change mode in that architecture. Oct 24 17:39:45 awesome Oct 24 17:40:04 this was fixed in armv5, which is one of the reasons that the eabi specifies v5 for interworking. Oct 24 17:40:06 hmm, ~lart armv4t. did you have a chance to do some benchmarks? Oct 24 17:40:35 no, not yet. might do that this evening if I get time. Oct 24 17:41:30 excellent. the patches can be applied unconditionally? Oct 24 17:42:19 yeah, they have "#if defined(__ARM_V4T__) && defined(__THUMB_INTERWORK__)" around all the changes Oct 24 17:44:02 * pb__ go home now Oct 24 17:44:03 bbl Oct 24 17:44:43 cu Oct 24 17:52:38 03Koen Kooi  07org.openembedded.dev * r0e2bb68527 10openembedded.git/conf/machine/omap3evm.conf: omap3evm: remove duplicate package archs Oct 24 18:47:12 What is the difference between DEPENDS, require and IMAGE_INSTALL in bb file? Oct 24 18:49:03 bitbake has extremely poor documentation. :( Oct 24 18:53:17 'require' is described at http://bitbake.berlios.de/manual/ch02.html#id869229 Oct 24 18:53:26 hubar: DEPENDS means your package needs another to be built Oct 24 18:53:38 did you actually read the bitbake manual? it certainly is not perfect but it does answer at least some of your questions. Oct 24 18:53:48 hubar: IMAGE_INSTALL is a list of packages to be installed in the rootfs Oct 24 18:53:52 IMAGE_INSTALL is part of oe, not bitbake, and gives a list of packages to install into the rootfs. Oct 24 18:56:31 <_bernie> has anyone ever seen an error like this? Oct 24 18:56:33 <_bernie> * ERROR: Cannot satisfy the following dependencies for task-base-extended: Oct 24 18:56:33 <_bernie> * libgnutls13 (>= 1.6.3) * libgnutls13 (>= 1.6.3) * libgnutls13 (>= 1.6.3) * libgnutls13 (>= 1.6.3) * libgnutls-extra13 (>= 1.6.3) * Oct 24 18:56:40 <_bernie> (sorry for the flood) Oct 24 18:57:02 _bernie: well, it appears that you have a problem with libgnutls13. does that package actually exist in your deploy directory? Oct 24 18:57:59 <_bernie> yes, it appears * appears * appears * appears * Oct 24 18:58:10 heh Oct 24 18:58:11 <_bernie> pb_: I grepped for it everywhere and can't find it Oct 24 18:58:25 _bernie: check the Packages file to see what packages actually depend on it Oct 24 18:58:41 <_bernie> is there a way to re-run bitbake scripts manually to trace errors? Oct 24 19:18:40 _bernie: who let you in here? Oct 24 19:18:53 * CosmicPenguin needs to have a word with the bouncer Oct 24 19:25:00 What is the difference between DEPENDS, require and IMAGE_INSTALL variable in bb file? Oct 24 19:25:36 hubar: you were just told. Oct 24 19:25:38 Er, alot :) Oct 24 19:25:40 hubar: i'd suggest you read Oct 24 19:25:50 kergoth: my machine crashed Oct 24 19:26:08 the screensaver crashed it. :( Oct 24 19:26:17 pb_: 'require' is described at http://bitbake.berlios.de/manual/ch02.html#id869229 Oct 24 19:26:17 [11:53am] denix: hubar: DEPENDS means your package needs another to be built Oct 24 19:26:18 pb_: IMAGE_INSTALL is part of oe, not bitbake, and gives a list of packages to install into the rootfs. Oct 24 19:26:36 pb_: did you actually read the bitbake manual? it certainly is not perfect but it does answer at least some of your questions. Oct 24 19:28:40 hey guys ... where i can find the flier we used at fosdem ? Oct 24 20:06:07 <_bernie> CosmicPenguin: oh you are a member of our exclusive club too? ;-) Oct 24 20:06:12 <_bernie> CosmicPenguin: I'm working on the beagleboarsd Oct 24 20:06:50 <_bernie> porting sugar to itsd Oct 24 20:08:18 <_bernie> CosmicPenguin: and what is amd using openembedded for? Oct 24 20:10:16 I've been using it exclusively on Geode for years Oct 24 20:11:30 hi blinder Oct 24 20:17:17 hi Oct 24 20:58:32 03Kalev Lember  07org.openembedded.dev * r71cb86307b 10openembedded.git/ (conf/checksums.ini packages/valgrind/valgrind_3.3.1.bb): valgrind 3.3.1: add new version Oct 24 21:07:40 so DEPENDS don't necessarily end up in the rootfs? Oct 24 21:09:37 indeed not, DEPENDS has no direct relationship with rootfs construction Oct 24 21:11:49 hmm Oct 24 21:12:09 What if the IMAGE_INSTALL needs to link against, say, zlib. Oct 24 21:12:24 Do I put zlib in DEPENDS or IMAGE_INSTALL? Oct 24 21:12:40 neither, it should all be automatic Oct 24 21:13:33 hmm? Oct 24 21:13:55 how does it know to compile zlib then? Oct 24 21:14:16 because the individual package will name it in DEPENDS. Oct 24 21:15:03 oh. Oct 24 21:15:15 hmm. Oct 24 21:15:44 But if zlib statically linked? or dynamically linked? Oct 24 21:15:52 with the individual package. Oct 24 21:16:04 I mean is zlib. Oct 24 21:16:10 it shouldn't make any difference to the rootfs. Oct 24 21:16:30 so long as the individual packages that go into it are correctly formed, and have the right dependencies, everything should "just work". Oct 24 21:16:53 if you are having a problem with this, it would probably be better for you to give a concrete example rather than talking about hypotheticals. Oct 24 21:23:20 hi pb Oct 24 21:23:23 hi hubar Oct 24 21:23:44 heh wog Oct 24 21:23:56 damnit how do I tell the htc blueangel to use the usb connection for inet Oct 24 21:26:45 What does "?=" mean? Oct 24 21:26:58 if set do not touch it Oct 24 21:27:05 otherwise set the variable Oct 24 21:27:22 oh! Oct 24 21:57:01 03Otavio Salvador  07org.openembedded.dev * rf33c7f9356 10openembedded.git/packages/vte/vte.inc: vte: depends on ncurses Oct 24 21:57:02 03Otavio Salvador  07org.openembedded.dev * rcfada083d1 10openembedded.git/packages/vnc/libvncserver_0.9.1.bb: libvncserver: fix build properly applying autoconf patch Oct 24 21:57:03 03Otavio Salvador  07org.openembedded.dev * rda70fd4cc1 10openembedded.git/packages/qmake/ (files/0001-fix-mkspecs.patch qmake2-native.inc): qmake2: properly set rcc utility to rcc4 in native stagging Oct 24 21:57:12 03Otavio Salvador  07org.openembedded.dev * r9e03174940 10openembedded.git/packages/matchbox-wm/matchbox-wm_1.2.bb: matchbox-wm: fix dbus support Oct 24 22:00:38 03Otavio Salvador  07org.openembedded.dev * rfdf64f8770 10openembedded.git/packages/pam/ (libpam_0.79.bb libpam_1.0.2.bb): (log message trimmed) Oct 24 22:00:38 pam (1.0.2 and 0.79): install headers in ${includedir}/security Oct 24 22:00:38 While building an O.S. Systems module that uses the PAM library this Oct 24 22:00:38 problem has been found since the code was compiling fine in a lot of Oct 24 22:00:38 different environments and failing on OE. Oct 24 22:00:42 Checking the source has been noticed that upstream explicitely set Oct 24 22:00:44 includedir in configure script but autotools class overrided it by Oct 24 22:00:46 03Otavio Salvador  07org.openembedded.dev * r3b6f9a308f 10openembedded.git/packages/pam/libpam_1.0.2.bb: Oct 24 22:00:49 pam 1.0.2: fix plugins split up Oct 24 22:00:51 The recipe was missing the PACKAGES_DYNAMIC field making impossible to Oct 24 22:00:53 include the plugins in images Oct 24 22:00:53 <_bernie> CosmicPenguin: indeed a really nice distro Oct 24 22:01:12 <_bernie> CosmicPenguin: well, bitbake is a bit hard to work with... Oct 24 22:02:29 03Otavio Salvador  07org.openembedded.dev * r8e522a5c03 10openembedded.git/conf/checksums.ini: checksums.ini: add gamin 0.1.8 Oct 24 22:05:45 :-) Oct 24 22:05:55 * otavio likes to push changes to upstream :-D Oct 24 22:09:02 _bernie: in what way would you say it's hard to work with? Oct 24 22:11:15 <_bernie> kergoth: perhaps I'm still unfamiliar with it, but I found it hard to see what commands were executed before an error so I can reproduce it on the command line Oct 24 22:11:38 <_bernie> you know, the usual "cd build_dir; ./foo" Oct 24 22:12:16 all shell tasks (configure, compile, etc) are emitted as shell scripts and executed, you can set a var to ensure they stick around. a log of the execution of that emitted script is also available Oct 24 22:13:20 BBINCLUDELOGS is what results in log emission, and that's in local.conf.sample Oct 24 22:13:29 I just build on my z. helps having a spitz mind Oct 24 22:13:30 as is OEDEBUG, which results in showing more info during the build Oct 24 22:18:01 <_bernie> kergoth: oh nice to know! this is the piece of information I was missing Oct 24 22:18:29 sounds like you should read local.conf.sample (it surprises me that you ever got up and running without doing so, since its there to show you how to create your initial local.conf...) Oct 24 22:18:49 <_bernie> everybody needs to know this at some point, maybe it should be advertised more. (such as in the output after an error occurs) Oct 24 22:19:13 the log is pointed out in the output after an error occurs, if BBINCLUDELOGS is set, iirc Oct 24 22:19:22 i haven't done a build in a while, but that's how it used to behave Oct 24 22:19:29 and as i said, BBINCLUDELOGS is set in the sample configuration Oct 24 22:21:19 kergoth: it still behaves the same way... :) Oct 24 22:23:09 kergoth: but, about reading the local.conf.sample file - there are many instructions on the Net, telling you how to create a simple few-lines local.conf, w/o even explaining or directing users to local.conf.sample... :( Oct 24 22:25:13 <_bernie> kergoth: yes I had it set too Oct 24 22:25:45 <_bernie> kergoth: still, I don't understand how to keep the shell scripts after the build fails Oct 24 22:41:38 <_bernie> kergoth: another issue I have with bitbake is that after updating a few times I get missing dependency errors for a package that does not exist any more. Oct 24 22:42:35 <_bernie> so presumably not all relevant packages got rebuilt after the update, and I don't know how to force a rebuild. Oct 24 22:44:53 _bernie: you can always remove your temp directory :) Oct 24 22:45:41 <_bernie> denix: I was afraid someone would suggest that :-( Oct 24 22:45:47 The metadata is moving rather fast - you should use caution when blindly updating the metadata Oct 24 22:45:58 <_bernie> it took like 3 hours to rebuild everything from scratch Oct 24 22:47:05 _bernie: if you know what you want to rebuild, bitbake package -c clean would clean, or -c rebuild Oct 24 22:47:07 <_bernie> CosmicPenguin: auto{conf,make,hell} also has the same problem of loosing its mind when you make certain changes. Oct 24 22:47:38 <_bernie> denix: trying Oct 24 22:47:45 <_bernie> thanks everybody for helping out Oct 24 22:48:37 CosmicPenguin: you mentioned you use oe with geode. do you have a distro for it? Oct 24 22:48:46 <_bernie> now it's rebuilding it for real. so it seems wiping out the stamps was not the way to fool it Oct 24 22:49:10 denix: what do you mean? Oct 24 22:49:59 CosmicPenguin: what do you use for DISTRO and MACHINE? Oct 24 22:50:40 for MACHINE I have a slightly tweaked version of conf/machine/geodelx.conf Oct 24 22:51:08 for distro, just use whatever makes you happy - the only difference between Geode and any other x86 image is the machine type and the kernel Oct 24 22:52:00 usually, distros specify preferred kernels, that's why I asked Oct 24 22:52:28 the macine should specify the preferred kernel Oct 24 22:53:16 denix: the oe metadata keeps them pretty independent. distro defines policy, not what's necessary to support a specific platform Oct 24 22:53:57 yes, you're right Oct 25 00:39:19 03Leon Woestenberg  07org.openembedded.dev * r69e1cede32 10openembedded.git/conf/machine/canyonlands.conf: canyonlands.conf: New machine for AMCC460EX (PowerPC) development board. Oct 25 01:09:20 03Khem Raj  07org.openembedded.dev * r7df20bae85 10openembedded.git/MAINTAINERS: Add myself to the list. Oct 25 01:53:00 hmm, if I set my bb image based on some other package, that package should be able to compile successfully right? Oct 25 01:53:09 | In file included from pcre.c:470: Oct 25 01:53:11 | chartables.c:21:27: error: pcre_internal.h: No such file or directory Oct 25 01:53:12 | pcre.c: In function 'pcre_fullinfo': Oct 25 01:53:14 | pcre.c:780: error: 'pcre_default_tables' undeclared (first use in this function) Oct 25 01:53:15 | pcre.c:780: error: (Each undeclared identifier is reported only once Oct 25 01:53:17 | pcre.c:780: error: for each function it appears in.) Oct 25 01:53:56 err I meant to including the package inside IMAGE_INSTALL. Oct 25 02:06:35 hmm, if I set my bb image IMAGE_INSTALL on some other package, that package should be able to compile successfully right? Oct 25 02:08:11 Are there actually people reading oe-user ML? **** ENDING LOGGING AT Sat Oct 25 02:59:57 2008