**** BEGIN LOGGING AT Tue Oct 27 02:59:56 2009 Oct 27 04:13:56 hey all Oct 27 04:14:59 I'm trying to build Angstrom with stable/2009, and I'm having problems with the initramfs images it's producing Oct 27 04:15:07 I don't get a /init, nor do any of the initramfs modules get installed Oct 27 06:25:41 03Rolf Leggewie  07org.openembedded.dev * rbf6846eb5c 10openembedded.git/recipes/tzcode/tzcode-native_2009k.bb: tzcode: update to 2009p Oct 27 06:33:09 eh? tzcode-native_2009k.bb uses 2009p? Oct 27 06:39:12 no Oct 27 06:39:22 I'm in the process of fixing it Oct 27 06:39:42 but tzcode-native apparently uses tzdata2009p code Oct 27 06:43:53 03Rolf Leggewie  07org.openembedded.dev * r9c0f554e9e 10openembedded.git/recipes/tzcode/ (tzcode-native_2009k.bb tzcode-native_2009p.bb): tzcode: only update tzdata SRC_URI to 2009p. Fixes previous commit. Oct 27 07:26:59 hi all Oct 27 07:27:32 do we have a machine definition that fit on a eeetop and/or a generic Atom machine ? Oct 27 07:57:39 03Koen Kooi  07org.openembedded.dev * reb6844453a 10openembedded.git/conf/checksums.ini: squashfs-tools: add checksum Oct 27 08:04:30 good morning Oct 27 08:15:20 morning Oct 27 08:15:32 gremlin[it]: tell me if you find one Oct 27 08:15:41 hrw: hi Oct 27 08:20:09 hi mckoan ... Oct 27 08:20:14 mckoan, find what ? Oct 27 08:26:56 gremlin[it]: what are you looking for: a machine definition that fit on a generic Atom machine Oct 27 08:29:24 morning folks Oct 27 08:36:30 moin Oct 27 08:41:31 mckoan: grab "netbook" machine from Poky Oct 27 08:45:21 mckoan, ahh ok ;) Oct 27 09:10:40 I have a local src tree and would like bitbake to fetch src from this tree. Is there a way to specify that bitbake should fetch the entire tree instead of specifying all files and folders from the tree? Oct 27 09:11:41 i had some dev recipes that just modified ${S} to point to the local source Oct 27 09:12:48 then fetched a dummy file (e.g. SRC_URI = "file://dummy.txt") Oct 27 09:13:08 Does this mean that bitbake will work on local src dir instead of downloading and unpacking into work/ folder? Oct 27 09:13:14 yes Oct 27 09:13:35 I've seen at least one howto (possibly the Beagleboard one) that said to do it like this Oct 27 09:13:36 Thanks, just what I'm looking for. Oct 27 09:13:55 Do you have any links to it? Oct 27 09:14:06 but bitbake -c clean doesn't run a make clean on your sources, so I found I sometimes had to do that manually Oct 27 09:14:37 But it will not remove your local sources Oct 27 09:14:46 no Oct 27 09:15:27 http://elinux.org/BeagleBoardOpenEmbeddedDevelopment Oct 27 09:15:29 near the bottom Oct 27 09:16:27 Thanks, this is what I was looking for. Oct 27 09:17:54 now my turn... what is the status of the "multimachine" stuff? Oct 27 09:21:06 Sorry, don't know:) Oct 27 09:22:23 mike_cw: multimachine? it is default on Oct 27 09:24:35 hrw: my colleague posted me a quote from somewhere saying that it can be used on a "per distribution basis". does that mean it can't be used with Angstrom? Oct 27 09:28:30 mike_cw: iirc multimachine cannot be turned off now Oct 27 09:29:47 03Koen Kooi  07org.openembedded.dev * rf624db1d3c 10openembedded.git/conf/bitbake.conf: bitbake conf: put static libs in ${PN}-static Oct 27 09:29:53 03Koen Kooi  07org.openembedded.dev * r3499e5b252 10openembedded.git/classes/package.bbclass: package bbclass: don't strip static libs so aggressively Oct 27 09:29:54 03Koen Kooi  07org.openembedded.dev * r9c4fbd1f70 10openembedded.git/recipes/glibc/glibc-package.bbclass: (log message trimmed) Oct 27 09:29:54 glibc package bbclass: tweak library packaging Oct 27 09:29:54 libcidn and libmemusage have there own package now Oct 27 09:29:56 Static libs move around a bit: Oct 27 09:29:58 glibc-dev: Oct 27 09:30:00 /usr/lib/libpthread_nonshared.a Oct 27 09:30:02 /usr/lib/libc_nonshared.a Oct 27 09:30:04 03Koen Kooi  07org.openembedded.dev * rf5d9adcea8 10openembedded.git/recipes/bison/bison_2.3.bb: Oct 27 09:30:07 bison: put liby.a in ${PN} Oct 27 09:30:09 Phil Blundell said: Oct 27 09:30:11 >> bison-2.3-r0: /usr/lib/liby.a Oct 27 09:30:13 That one is a special case: it wants to stay in the main bison package, Oct 27 09:30:15 since bison itself is a development tool. I'm not sure why it is a Oct 27 09:30:17 static-only library; that might be an error. Oct 27 09:30:21 03Koen Kooi  07org.openembedded.dev * r4b276575de 10openembedded.git/recipes/flex/ (flex.inc flex_2.5.31.bb): (log message trimmed) Oct 27 09:30:24 flex: put libfl into ${PN} Oct 27 09:30:28 Phil Blundell said: Oct 27 09:30:30 > > flex-2.5.31-r4: /usr/lib/libfl.a Oct 27 09:30:32 This is like bison. Oct 27 09:30:34 >> bison-2.3-r0: /usr/lib/liby.a Oct 27 09:30:36 That one is a special case: it wants to stay in the main bison package, Oct 27 09:30:40 03Koen Kooi  07org.openembedded.dev * r16726220cd 10openembedded.git/recipes/libxml/ (libxml2.inc libxml2_2.7.3.bb): libxml2: stop overriding PACKAGES Oct 27 09:30:43 03Koen Kooi  07org.openembedded.dev * rf257d5a8e8 10openembedded.git/conf/distro/include/ (angstrom-eglibc.inc angstrom-glibc.inc): angstrom: change *glibc build to -Os Oct 27 09:30:46 03Koen Kooi  07org.openembedded.dev * rf6fb860bff 10openembedded.git/conf/distro/angstrom-2008.1.conf: angstrom: switch to linux-libc-headers 2.6.31 Oct 27 09:30:55 03Koen Kooi  07org.openembedded.dev * r44ec0ef1a3 10openembedded.git/conf/distro/include/ (angstrom-glibc.inc angstrom-ldflags.inc): Oct 27 09:31:00 angstrom: add an option to use -Wl,--as-needed for glibc builds Oct 27 09:31:02 Console-image builds and seems to work, but this needs lots of testing before getting enabled Oct 27 09:35:24 good morning Oct 27 09:35:40 good morning Oct 27 09:43:21 hello Oct 27 09:43:25 FATAL: Task /net/wintermute/vol1/home/anish/openembedded/recipes/update-rc.d/update-rc.d_0.7.bb (do_package) has circular dependency on /net/wintermute/vol1/home/anish/openembedded/recipes/glibc/glibc_2.9.bb (do_package) Oct 27 09:43:38 how to resolve this circular dependency error Oct 27 09:44:59 03Enric Balletbò i Serra  07org.openembedded.dev * rda37e9d04c 10openembedded.git/recipes/netbase/ (netbase/igep0020/interfaces netbase_4.21.bb): Oct 27 09:44:59 netbase: Add version for IGEP v2 board Oct 27 09:44:59 Signed-off-by: Enric Balletbo i Serra Oct 27 09:44:59 Signed-off-by: Koen Kooi Oct 27 09:45:11 03Enric Balletbò i Serra  07org.openembedded.dev * rcae0cfd5c0 10openembedded.git/recipes/xorg-xserver/ (2 files in 2 dirs): Oct 27 09:45:11 xserver-xf86-config: Add version for IGEP v2 board Oct 27 09:45:11 Signed-off-by: Enric Balletbo i Serra Oct 27 09:45:11 Signed-off-by: Koen Kooi Oct 27 09:47:42 03Jeremy Lainé  07koen/static-libs-rework * r1f7a632a96 10openembedded.git/classes/update-rc.d.bbclass: Oct 27 09:47:42 update-rc.d.bbclass: make prerm and postrm dpkg-compatible Oct 27 09:47:42 * prerm : only stop service at the "upgrade" and "remove" steps Oct 27 09:47:42 * postrm : only remove symlinks at the "remove" and "purge" steps, after Oct 27 09:47:42 checking the init script no longer exists Oct 27 09:47:44 03Frans Meulenbroeks  07koen/static-libs-rework * r4882a87cc5 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Oct 27 09:47:49 03Xerxes RÃ¥nby  07koen/static-libs-rework * r66c4fef12c 10openembedded.git/ (5 files in 2 dirs): Oct 27 09:47:52 llvm2.6: Updated to use llvm 2.6 release tarball. Oct 27 09:47:54 llvm2.6-native: Likewise. Oct 27 09:47:56 checksums.ini: add sums for llvm-2.6.tar.gz Oct 27 09:47:58 03Frans Meulenbroeks  07koen/static-libs-rework * rd7e9f09e29 10openembedded.git/recipes/tgt/tgt_0.9.10.bb: tgt: updaded recipe to use ${PV} for download Oct 27 09:48:01 03Steffen Sledz  07koen/static-libs-rework * r249a10ce18 10openembedded.git/conf/checksums.ini: ltrace_0.5.3: missing checksum added Oct 27 09:48:04 03Steffen Sledz  07koen/static-libs-rework * r93443046f1 10openembedded.git/conf/checksums.ini: ltrace_0.5.3: some more missing checksums added Oct 27 09:48:07 03Xerxes RÃ¥nby  07koen/static-libs-rework * rb39a67f8b6 10openembedded.git/conf/checksums.ini: llvm2.6: fix checksum Oct 27 09:48:18 03Frans Meulenbroeks  07koen/static-libs-rework * r2683426742 10openembedded.git/ (3 files in 2 dirs): open-iscsi: added recipe; still needs stress-testing Oct 27 09:48:21 03Michael Smith  07koen/static-libs-rework * rdba8dec409 10openembedded.git/classes/gitver.bbclass: (log message trimmed) Oct 27 09:48:24 gitver.bbclass: fix git dependency tracking Oct 27 09:48:26 .git/HEAD doesn't usually change on a commit, but the ref it points to Oct 27 09:48:28 (e.g. .git/refs/heads/master) should. Handle this and another couple of Oct 27 09:48:30 cases: if a tag is added without a new commit, or if the ref is in Oct 27 09:48:32 packed-refs. Oct 27 09:48:34 Signed-off-by: Michael Smith Oct 27 09:48:36 03Florian Boor  07koen/static-libs-rework * r7ef1593130 10openembedded.git/recipes/prelink/ (prelink_20061027.bb prelink_20071009.bb): prelink: Change postrm script to prerm. Calling 'prelink' after removing the binary is likely to fail. Oct 27 09:48:50 03Frans Meulenbroeks  07koen/static-libs-rework * rf8e861e80c 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Oct 27 09:48:53 03Frans Meulenbroeks  07koen/static-libs-rework * r1164525915 10openembedded.git/recipes/open-iscsi/open-iscsi-user_2.0-871.bb: open-iscsi-user: added RRECOMMENDS Oct 27 09:48:56 03Frans Meulenbroeks  07koen/static-libs-rework * r8afa9e311d 10openembedded.git/ (conf/checksums.ini recipes/tgt/tgt_0.9.10.bb): tgt (Linux SCSI target framework) : recipe created Oct 27 09:48:59 03Marcin Juszkiewicz  07koen/static-libs-rework * r6cb806430f 10openembedded.git/recipes/linux/ (10 files in 2 dirs): Oct 27 09:49:02 linux: added NHK15 support for 2.6.20 version Oct 27 09:49:06 staged headers are needed for x11 acceleration Oct 27 09:49:12 03Marcin Juszkiewicz  07koen/static-libs-rework * r63fd61ae8a 10openembedded.git/conf/machine/nhk15.conf: nhk15: added machine config for ST Microelectronics NHK15 devboard Oct 27 09:49:15 03Marcin Juszkiewicz  07koen/static-libs-rework * r083679cd96 10openembedded.git/recipes/xorg-xserver/ (3 files in 2 dirs): Oct 27 09:49:20 xserver-kdrive-nomadik: added 1.3.0.0 version Oct 27 09:49:22 This version has hardware acceleration working. Oct 27 09:49:24 NOTE: patches were applied and refreshed - thats why they differ from ST Oct 27 09:49:26 ones. Oct 27 09:49:28 (59 lines omitted) Oct 27 09:55:41 03Rolf Leggewie  07org.openembedded.dev * rdfa6b034ba 10openembedded.git/recipes/libsdl/ballfield_1.0.bb: ballfield: fix spelling mistake that prevents a build. s/vrtual/virtual/ Oct 27 09:58:55 Reminder: This error comes in the starting only: FATAL: Task /net/wintermute/vol1/home/anish/openembedded/recipes/update-rc.d/update-rc.d_0.7.bb (do_package) has circular dependency on /net/wintermute/vol1/home/anish/openembedded/recipes/glibc/glibc_2.9.bb (do_package) What should be done in this case? Oct 27 10:16:58 some image recipes seem to have packages listed as DEPENDS, whereas others just list everything in INSTALL_IMAGE. What is the right way to do this? Oct 27 10:17:06 s/INSTALL_IMAGE/IMAGE_INSTALL/ Oct 27 10:21:44 florian: good morning Oct 27 10:21:53 mickey|office: good morning Oct 27 10:23:19 ~curse crappy ethernet on boards Oct 27 10:23:21 May you be reincarnated as a Windows XP administrator, crappy ethernet on boards ! Oct 27 10:34:33 hmm. can I use multimachine to (in one go) build a generic rootfs plus a number of board specific kernels? Oct 27 10:37:50 pb__: morning pb__. i have uploaded the problematic rootfs @ http://linuxtogo.org/~mickeyl/misc/minimal-fso-console-image-eglibc-ipk-.dev-snapshot-20091026-om-gta02.rootfs.tar.bz2 Oct 27 10:39:30 ok, awesome Oct 27 10:39:30 03Marcin Juszkiewicz  07org.openembedded.dev * r592aec1153 10openembedded.git/: Merge branch 'org.openembedded.dev' of git.openembedded.org:openembedded into org.openembedded.dev Oct 27 10:39:35 03Marcin Juszkiewicz  07org.openembedded.dev * r03ed4b979c 10openembedded.git/ (conf/machine/nhk15.conf recipes/linux/linux_2.6.20.bb): nhk15: added uImage parameters Oct 27 10:39:46 I need to write some java code (!) first but I will look at that later this morning Oct 27 10:40:43 cool, thanks (d'oh, java... good speed) Oct 27 10:51:53 yeah, all xml and xpath and stuff, very enterprise class Oct 27 10:52:33 I'm going to have to start wearing a suit at this rate. Oct 27 10:57:47 heh Oct 27 10:59:25 * florian managed to book a flight Oct 27 11:00:06 very good Oct 27 11:00:08 where to? Oct 27 11:00:29 the tropics I hope Oct 27 11:00:39 Luton :) Oct 27 11:00:52 heh, not quite so exotic Oct 27 11:01:09 pb_, you need to do something about java, and even worse "suits" Oct 27 11:01:53 03Klaus Kurzmann  07shr/import * rd937cae60f 10openembedded.git/recipes/libgee/libgee-native_git.bb: Oct 27 11:01:53 libgee-native: new recipe Oct 27 11:01:53 Signed-off-by: Klaus Kurzmann Oct 27 11:01:57 It took me some time to arrange everything... the weeks around OEDEM are full of events and to make it more complicated my wife insisted on visiting Cambridge. Oct 27 11:02:47 ah, I guess that means you need to bring your children as well Oct 27 11:04:04 nope... they will go on holiday to the Westerwald :-) grandma lives there... that's more fun for the kids. Oct 27 11:05:51 My wife will have fun... English was her worst subject at school and she never really had to make use of it :) Oct 27 11:06:42 florian: ugh Oct 27 11:09:31 pb__: I just read you can take a plane from London to Cambridge :) Oct 27 11:12:25 hmm, who stole stdio.h > Oct 27 11:12:27 ? Oct 27 11:12:28 florian: heh, I didn't know that Oct 27 11:12:30 /home/frans/oe/tmp_angstrom/work/armv7a-angstrom-linux-gnueabi/libgsm-1.0.12-r1/temp/log.do_compile.10142 Oct 27 11:13:25 oops wrong paste Oct 27 11:13:33 src/add.c:13:19: error: stdio.h: No such file or directory Oct 27 11:13:55 libgsm package Oct 27 11:14:12 Oh oh, http://www.phdcomics.com/comics.php?n=1243 Oct 27 11:14:16 pb__: The Express Holiday Inn website mentiones this. Oct 27 11:14:22 is this due to some of the changes in linux libc headers ? Oct 27 11:17:36 btw. this hotel has a special offer for the oedem weekend for two persons or more. Oct 27 11:18:48 eFfeM: this looks evil... but stdio.h should not be part of linux libc headers... or? Oct 27 11:19:40 florian, the file has #include and yes there are plenty of stdio.h files in my work dir :-) Oct 27 11:19:49 heh Oct 27 11:20:22 I wonder if the change could have messued up the whole toolchain somehow. Oct 27 11:20:25 could someone try a bitbake libgsm to see if this is specifc to my setup or a general problem (preferably someone whose local git is at the HEAD Oct 27 11:20:56 I remember some strange effects like this with multimachine builds Oct 27 11:20:57 could of course also be something in the recipe Oct 27 11:21:36 i do build different machines regularly (beagleboard, sheevaplug, openrd), never had this before Oct 27 11:21:49 03Jeremy Lainé  07org.openembedded.dev * reeac84ae32 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Oct 27 11:21:50 03Jeremy Lainé  07org.openembedded.dev * rcc80e9a2ec 10openembedded.git/recipes/linux/ (4 files in 2 dirs): linux-2.6.31: remove SPI patch for boc01, use devicetree instead Oct 27 11:21:53 of course I don't do parallel builds, that seems too daring Oct 27 11:22:32 pb__: Looking at the oedem wiki page and travel descriptions we should add something: The estimated price for a taxi from the railway station :) Oct 27 11:22:47 florian: oh right, yes Oct 27 11:22:55 I think it usually costs about GBP 7 Oct 27 11:23:02 florian: maybe we all could share a taxi Oct 27 11:23:11 :-) Oct 27 11:23:29 03Klaus Kurzmann  07shr/import * rb35a3b44b8 10openembedded.git/conf/distro/include/preferred-shr-versions.inc: Oct 27 11:23:29 preferred-shr-versions.inc: prefer vala 0.7.7+fso5 Oct 27 11:23:29 Signed-off-by: Klaus Kurzmann Oct 27 11:23:30 03Klaus Kurzmann  07shr/import * r29e85950ba 10openembedded.git/conf/distro/include/shr-autorev-unstable.inc: Oct 27 11:23:32 shr-autorev-unstable.inc: add and set libgee-native to AUTOREV Oct 27 11:23:32 pb__: ok that's not that bad... Oct 27 11:23:34 Signed-off-by: Klaus Kurzmann Oct 27 11:23:36 03Klaus Kurzmann  07shr/import * rab3fd2a3e7 10openembedded.git/ (3 files in 2 dirs): Oct 27 11:23:38 vala(-native): bump to 0.7.7+fso5 Oct 27 11:23:40 Signed-off-by: Klaus Kurzmann Oct 27 11:23:50 * mckoan booked an hotel nearby the YHA Oct 27 11:24:12 florian: yeah, it should definitely be less than GBP 10. I think the cheapest I have ever had is about GBP 4 but it depends rather on the time of day: obviously if the traffic is worse then it ends up more expensive. Oct 27 11:24:39 pb__ btw I won't be able to make it for oedem (pity but have too many things in that weekend) Oct 27 11:25:12 pb__: Sounds like a good option... with the fact in mind that the bus is way more complicated and costs as well. Oct 27 11:25:49 what about to rent a bike ? Oct 27 11:26:23 florian: right, if there are at least two of you then the taxi is probably your best option. Oct 27 11:26:46 the bus fare for 2 people would be about GBP 5 anyway, and it is much slower as well as being more complicated Oct 27 11:27:28 pb__: sounds like the bus is comarably expensive to Siegen. Oct 27 11:27:46 mckoan: yes, there is a bike hire shop at the station, and another next door to the yha Oct 27 11:28:14 http://www.stationcycles.co.uk/Services/Hire.html is the station one, I can't remember what the other is called Oct 27 11:28:44 how many of us must leave on early Sunday afternoon to go to airpotr? Oct 27 11:29:44 It's not really early afternoonf for us I hope. I still need to check about trains to LTN on Sunday. Oct 27 11:30:22 for LTN you might be better off taking the bus than the train. Oct 27 11:30:35 pb__: what about STN ? Oct 27 11:30:49 for STN the train is usually better, though there isn't a lot in it. Oct 27 11:31:23 there is a direct train service from cambridge to STN, it takes about half an hour. Oct 27 11:31:27 florian: fyi: http://www.nationalexpress.com/Airport/luton.cfm Oct 27 11:32:01 pb__: thanks! Oct 27 11:34:01 Cambridge (CBG) to Stansted Airport (SSD) 16:10 Oct 27 11:34:15 the only thing to be wary of with the train is that sometimes they close that line for engineering work at weekends, so you need to check that there is no scheduled maintenance on the sunday you want to travel Oct 27 11:34:37 it is annoying to arrive at the station only to find that all trains are cancelled. I have fallen for this trick several times and it is no fun. Oct 27 11:35:36 03Klaus Kurzmann  07shr/import * r11b5e27e67 10openembedded.git/conf/distro/include/shr-autorev-unstable.inc: Oct 27 11:35:36 shr-autorev-unstable.inc: set vala-dbus-binding-tool-native to AUTOREV Oct 27 11:35:36 Signed-off-by: Klaus Kurzmann Oct 27 11:35:43 pb__: I thought it happened only in Italy :-D Oct 27 11:35:53 heh Oct 27 11:36:41 * mckoan -> lunch Oct 27 11:38:12 pb__: Do you know offhand where the bus leaves and if it is useful to book it in advance? Oct 27 11:41:03 03Martin Jansa  07shr/merge * r55d22887b5 10openembedded.git/ (5 files in 4 dirs): Oct 27 11:41:03 Latest xorg stuff for xorg-7.5 2009-10-27 Oct 27 11:41:03 Signed-off-by: Klaus Kurzmann Oct 27 11:41:04 03Martin Jansa  07shr/merge * r3aa646f138 10openembedded.git/recipes/ (28 files in 6 dirs): Oct 27 11:41:04 shr/merge: Remove xorg stuff added to shr, pre xorg-7.5 versions Oct 27 11:41:08 Signed-off-by: Klaus Kurzmann Oct 27 11:42:28 florian: the airport bus leaves from the centre of town. I think it is a good idea to book in advance (you can print your own ticket) but not essential: they are hardly ever full Oct 27 11:44:41 pb__: ok - many thanks! Oct 27 11:47:06 ok... quite some busses sunday afternoon. not a single on that fits really good but better than expected :) Oct 27 11:53:58 yeah, the airport bus runs fairly frequently. they have more or less 24-hour service although it is a bit less frequent in the middle of the night Oct 27 12:00:15 hi all, what is the recommended bitbake for dev branch, please? Oct 27 12:01:00 I guess 1.8.8, isn't it? Oct 27 12:01:49 not .10? Oct 27 12:03:28 .. and .12 ? Oct 27 12:11:57 recalcati_: see sanity.conf Oct 27 12:19:05 * * OE Bug 5330 has been created by yeyo(AT)poczta.onet.pl Oct 27 12:19:07 * * Problem with building kernel image Oct 27 12:19:09 * * http://bugs.openembedded.org/show_bug.cgi?id=5330 Oct 27 12:22:19 ich muß erstmal ins office Oct 27 12:22:52 wrong window, sorry Oct 27 12:41:55 hi all, toolchain build bails here: http://pastebin.com/d262ff495 Oct 27 12:42:02 any ideas? Oct 27 12:44:10 your binutils is (ironically) too new Oct 27 12:44:31 there is a patch for glibc to fix that but I suspect it has either not been applied to 2.6, or you haven't pulled since it was checked in. Oct 27 12:44:48 I'm not quite sure how that happened! :) Oct 27 12:44:51 I just git pulled Oct 27 12:46:58 yeah, there looks to be a patch for 2.9, will see what I will do Oct 27 13:00:16 florian: how goes maemo-oe integration from gsoc student? Oct 27 13:01:06 hrw: that's a good topic Oct 27 13:01:22 ibot: seen rkirti? Oct 27 13:01:24 rkirti was last seen on IRC in channel #oe, 5d 17h 58m 33s ago, saying: 'hello woglinde'. Oct 27 13:01:36 tsts Oct 27 13:02:21 florian: I was asked few times during ELC-E about maemo5 in OE Oct 27 13:04:06 hrw: yep... shouldn't be too complicated but there are some bits that need an update. Oct 27 13:04:18 as usual Oct 27 13:04:19 hrw: That's something I would like to see too. Oct 27 13:05:10 florian: and DISTRO="maemo5-compat" then Oct 27 13:06:03 hrw: *if* we want to build compatible to the distro on the tablets yes. but that should be optional Oct 27 13:06:17 good morning Oct 27 13:07:18 hi recalcati Oct 27 13:07:38 florian: should, but sooner or later will land Oct 27 13:08:53 bitbake 1.8.12: ./setup.py build and tan install . no more ? Oct 27 13:18:02 * mckoan re Oct 27 13:29:20 I'm builing an image based on the console-image. It contains the netbase package which I want to get rid of because I want to manage network configuration my own way. I have tried to figure out what package that depends or includes netbase but cannot find it.. Can anyone point it out for me? Oct 27 13:37:05 what does "opkg whatdepends netbase" say? Oct 27 13:39:18 probably task-base Oct 27 13:48:20 http://marcin.juszkiewicz.com.pl/2009/10/27/elc-e-2009/ - finally wrote it Oct 27 13:52:35 anybody getting failures build flex since a few hours ago? Oct 27 14:20:01 I have a recipe for a package I need to build using scons. In .bb I need to set stuff like LD_SEARCH_PATH, INC_DIR, CFLAGS, and let SConstruct pick this up. How should this be done? Using environment variables or command line arguments to scons are two options, I guess. Oct 27 14:21:13 tasslehoff: tried "inherit scons"? Oct 27 14:21:53 hrw: yep. that works well, but isn't the problem :) Oct 27 14:23:48 CFLAGS is in the environment, so if I set it in .bb, scons can pick it up with os.environ['CFLAGS']. can I easily export other environment variables from a bb-file? is it as simple as : export SCONS_LD_FLAGS maybe? :) Oct 27 14:24:38 tasslehoff: I would start from checking other recipes which use scons and then by just setting TARGET_CLAGS etc Oct 27 14:25:02 pb_: task-boot Oct 27 14:27:12 hrw: I've checked the recipes, but they don't do the same. I'll look into how I can pass arguments to scons. Oct 27 14:28:38 tasslehoff: if only CFLAGS then use TARGET_CFLAGS Oct 27 14:29:05 03David-John Willis  07org.openembedded.dev * r6de8a003f9 10openembedded.git/classes/xfce46.bbclass: xfce46.bbclass: Add a helper class for XFCE 4.6 (based on the existing xfce.bbclass). Oct 27 14:29:06 03Enric Balletbo i Serra  07org.openembedded.dev * r5a2e1ba41c 10openembedded.git/ (4 files in 3 dirs): Oct 27 14:29:06 u-boot: add IGEP v2 support to 2009.08 release Oct 27 14:29:06 Signed-off-by: Enric Balletbo i Serra Oct 27 14:29:08 Signed-off-by: Koen Kooi Oct 27 14:29:10 03David-John Willis  07org.openembedded.dev * rcf508f496d 10openembedded.git/recipes/xfce/ (28 files in 4 dirs): XFCE: Rename xfce to xfce-base to support the import of the new XFCE 4.6.1 builds. Oct 27 14:29:17 03David-John Willis  07org.openembedded.dev * r599bbb2eaf 10openembedded.git/ (43 files in 9 dirs): XFCE: Import all the 4.6.1 recipes (and a newer version of xfce-terminal). Oct 27 14:29:19 03Enric Balletbo i Serra  07org.openembedded.dev * r3cecf685fc 10openembedded.git/ (3 files in 3 dirs): Oct 27 14:29:22 linux-igep2: Fix audio capture, the capture result is always silence Oct 27 14:29:24 Signed-off-by: Enric Balletbo i Serra Oct 27 14:29:26 Signed-off-by: Koen Kooi Oct 27 14:29:45 hrw: the issue is that I can't use only that, since scons wants flags and incdirs separated to handle dependencies. think I have found a way though Oct 27 14:30:15 iam a newbee in openembedded. run openembedded on the htc dream? Oct 27 14:35:27 hi, how can In force OE to rebuild a bb file Oct 27 14:35:28 w Oct 27 14:35:29 ?? Oct 27 14:35:46 clean it Oct 27 14:36:00 how? Oct 27 14:36:07 bitbake -c clean package Oct 27 14:36:19 pretty easy Oct 27 14:36:19 thanks Oct 27 14:37:14 jovox, are you using beagle? Oct 27 14:37:41 yes and no.. I got one but at the moment I'm working on an x86 board Oct 27 14:38:40 hmm...having some usb issues...although the kernel should have dm9601 built in it doesn't bring it up...so I changed the defconf just to buils it apart, along with mii.ko and usbnet.ko...if I get something I'll post Oct 27 14:39:55 I mwan, OE defconf since patching th ekernel by hand would be tremendously boring.... Oct 27 14:41:41 anyone working wiht beagle knows about the usb issuesin kernel 2.6.29? I downloaded it from google project page...and usb is not patched....is there a place to download the most recent working, stable one? don't wanna wait for it to be built.... Oct 27 14:42:22 lim_away: -> #beagle please Oct 27 14:42:28 ok Oct 27 14:43:42 hrw: nice your review, I also made I little one, but in pdf format, because it was in intranet wiki, http://www.opensurf.it/ELC-2009.pdf . Oct 27 14:44:58 03David-John Willis  07org.openembedded.dev * re26a80201b 10openembedded.git/recipes/tasks/task-xfce46-extras.bb: task-xfce46-extras: Don't try and ship themes that are not in 4.6 Oct 27 14:45:00 03David-John Willis  07org.openembedded.dev * r5642836ea5 10openembedded.git/recipes/images/xfce46-image.bb: xfce46-image: Add very simple Xfce 4.6 image Oct 27 14:45:03 hello run OE on the htc dream? Oct 27 14:49:07 recalcati: just read that Oct 27 14:49:20 master_main: no, but OE supports htcdream as target device Oct 27 14:49:31 what do you think about? Oct 27 14:50:47 nice one Oct 27 14:50:51 * hrw -> out Oct 27 15:14:40 re Oct 27 15:26:34 should bitbake -f unconditionally rebuild all dependencies? Oct 27 15:27:05 -f forces execution of the requested task, that's all. that one task, for the target you're requesting it be run against Oct 27 15:27:45 ah right. that would be the problem. so is there a recursive equivalent? Oct 27 15:27:55 or even a way to recursively clean Oct 27 15:28:15 pretty sure i pushed a cleanall task Oct 27 15:28:22 try bitbake -c cleanall foo to clean foo & all of its deps Oct 27 15:28:29 bitbake -n -c cleanall foo to be safe, dry run first Oct 27 15:28:31 :) Oct 27 15:30:15 heh. my workspace is in a mess. if it doesn't work i'll just wipe out tmp and come back in the morning :) Oct 27 16:15:18 am I the only one with problems with and friends ???? Oct 27 16:15:22 ls Oct 27 16:21:22 good night Oct 27 16:23:02 Hi. As an experiment, I tried building "qt4-embedded" and it built a large number of packages (>40), more than I recall it needing the last time I built it (non-oe build system). Is there a lighter-weight qt recipe? I'm thinking of using Qt for a non-X11 UI app, not for media player etc. Oct 27 16:24:24 it seems that not all the packages actually need to be installed Oct 27 16:25:05 qt4-embedded, qt4-embedded-fonts and qt4-embedded-plugin-decoration-default seems to be fine Oct 27 16:25:11 plus any imageformat plugins you need Oct 27 16:26:04 i have a non-X11 UI running with that, plus a custom gfxdriver for SDL Oct 27 16:26:17 Yeah, I might need libjpeg and similar, but don't want all of gnome, gstreamer, and some toher stuff Oct 27 16:26:47 So the "libX11" I saw in my build log is just an intermediate thing, not needed on target? Oct 27 16:27:44 can't say if my image built any X11 libs during staging because it's in the middle of a clean build right now, but it didn't have any on the target Oct 27 16:28:28 target is arm9 Oct 27 16:29:44 ok. Maybe installing and running a hello-Qt after the qte ipks will show what other runtime stuff is needed Oct 27 16:33:12 hoj, I might be doing something wrong, but I found that it didn't automatically install e.g. the fonts Oct 27 16:33:40 it seems reasonable to leave it up to the developer to install the required imagedriver plugins, though, which I also had to do Oct 27 16:34:09 mike_cw: ok I can burn that bridge when I get to it :) Oct 27 16:35:08 :) Oct 27 16:50:44 how do i make sure that bitbake compiles a kernel module for a specific kernel? Oct 27 16:52:43 bye for now Oct 27 16:53:53 rodgema: a module which is inside the linux kernel? Oct 27 16:54:04 its fuse Oct 27 16:54:09 kernel-module-fuse Oct 27 16:54:28 its not part of the kernel its a module Oct 27 16:54:40 is that a separate receipe? Oct 27 16:55:07 Weaselweb: its supposed to show up in lsmod, but it gets added to a different /lib/modules/kernelversion directory Oct 27 16:55:25 Weaselweb: so it doesnt show up with lsmod because it was compiled for the wrong kernel Oct 27 16:55:37 and insmod doesnt work cause its compiled wrong Oct 27 16:55:46 have you tried modprobe or insmod? Oct 27 16:55:49 ah, ic Oct 27 16:56:25 modprobe fuse saus module not found obviously Oct 27 16:56:41 well it doesnt say "obviously" :) Oct 27 16:57:08 rodgema: so are you using the kernel recipe that matches what you have installed? Oct 27 16:57:37 how can I tell that it matches per se? Oct 27 16:57:43 mh, no idea. i'm not that familiar with oe recipes. maybe you need to set or make sure STAGING_KERNEL_DIR has the correct value Oct 27 16:57:50 I'm just typing bitbake fuse Oct 27 16:58:05 its supposed to pull from the recipes directory the correct package no? Oct 27 16:59:23 rodgema: it is, what machine? I suspect your virtual/kernel is set to something that is now what is on your device. Oct 27 17:00:07 magician Oct 27 17:00:31 MACHINE=magician, OS=angstrom2007.1 Oct 27 17:00:47 rodgema: using an OpenEmbedded kernel. Hmmm, why so old? Oct 27 17:01:06 requirement Oct 27 17:01:10 :) Oct 27 17:01:36 EXTRA_OECONF = " --enable-kernel-module --with-kernel=${STAGING_KERNEL_DIR}" Oct 27 17:01:46 hmmmm Oct 27 17:02:24 rodgema: but with a recent openembedded tree? For something that old you may well be on your own. Oct 27 17:02:37 no its an old oe tree Oct 27 17:02:57 this is why i have lots o problems Oct 27 17:04:03 where should STAGING_KERNEL_DIR normally be set? Oct 27 17:04:09 rodgema: that goes without saying ;-), tried a bitbake virtual/kernel to see what gets built and packaged up, you can try that on your device. Oct 27 17:04:30 rodgema: that gets set to the build location of your kernel so other recipies can use it. Oct 27 17:05:29 yeah but is it normally set in the build/conf/local.conf? Oct 27 17:06:20 rodgema: it would be set when it is built up, you don't need to ever manually set it up. Oct 27 17:07:26 hi, i am still hunting what happened to my include files, gcc -v says --includedir=/home/frans/oe/tmp_angstrom/cross/armv5te/include but this dir does not exist; the only stdio.h file in cross/armv5te is ./lib/gcc/arm-angstrom-linux-gnueabi/4.3.3/include/ssp/stdio.h Oct 27 17:07:31 /lib/modules$ ls Oct 27 17:07:33 2.6.21-hh20 2.6.25 Oct 27 17:07:42 any one an idea ? Oct 27 17:47:30 03Jeremy Lainé  07org.openembedded.dev * rf07cd5560e 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://git@git.openembedded.net/openembedded into org.openembedded.dev Oct 27 17:47:37 03Jeremy Lainé  07org.openembedded.dev * r6f36bd6e28 10openembedded.git/recipes/klibc/ (4 files in 2 dirs): Oct 27 17:47:37 klibc-1.5.15: fix build against 2.6.31 linux headers Oct 27 17:47:37 * apply patches from Ubuntu to build against 2.6.31 linux headers Oct 27 18:17:06 03Koen Kooi  07org.openembedded.dev * r1a4bd59fda 10openembedded.git/ (conf/checksums.ini recipes/xfce-base/xfce4-mixer_4.6.1.bb): xfce 4.6.x: more SRC_URI and checksum tweaks Oct 27 18:34:51 hi, im trying to build the atngw100 target with current stable/minimal distro and receive the following error Oct 27 18:34:54 http://pastebin.com/m3f05d839 Oct 27 18:35:15 any advice for:NOTE: Handling BitBake files: \ (5086/6728) [75 %]ERROR: Information not available for target 'avr32-linux' Oct 27 18:35:27 hey everyone. I have found references to a python-pyusb_0.4-1.bb on the oe mailing list but I can't find the package anywhere. I'm trying to get the python bindings to libusb onto a gumstix oe build. Oct 27 18:36:05 we have built the usb.so file, but unfortunately when i import it I get a "usb.so: file or directory not found" when importing usb Oct 27 18:40:49 mickeyl: your rootfs does boot for me under qemu Oct 27 18:41:43 so, that's a bit of a pain. I guess that would suggest that there is some armv5 assumption in the code somewhere. Oct 27 18:41:55 if that's true, it would tend to point the finger of blame towards eglibc Oct 27 18:42:39 pb_: ok, thanks for inspecting. khem? Oct 27 18:42:58 annoyingly I don't think qemu has any armv4t emulation targets Oct 27 18:43:38 have you tried building a uclibc image? that might be worth a go, although glibc definitely did work for me on gta01 a few months ago Oct 27 18:44:05 actually, isn't gta02 armv5te anyway? Oct 27 18:44:22 or am I mixing that up with one of the "gta03" concepts Oct 27 18:45:04 pb_: I don't know if mickeyl said it, but that problem was fixed with this commit http://patchwork.dev.bearstech.com/patch/379/ Oct 27 18:45:19 pb_: we stayed on same gcc/binutils/eglibc combo and now it works ok Oct 27 18:45:25 yeah, that basically just turns off thumb Oct 27 18:45:34 so, that's not really a "fix", just a workaround that happens to stop the bug showing up Oct 27 18:46:03 pb_: gta02 is s3c2442 Oct 27 18:46:28 oh, right, so that is still armv4t Oct 27 18:46:35 *nod* *sigh* Oct 27 18:46:42 silly old samsung Oct 27 18:47:12 I guess maybe I am getting confused with my own products, the first gen of our board was arm920t but second gen is arm926 :-} Oct 27 18:47:16 hehe Oct 27 18:48:52 so in that case yeah, I would suggest trying an older glibc (not eglibc) and/or uclibc Oct 27 18:49:16 I think I was using glibc before for my gta01 thumb builds, so I am fairly certain that that has worked at some time. Oct 27 18:49:39 more recently I have been using uclibc for most things but I think all the targets I have tested on there have been armv5te or higher Oct 27 18:49:43 guys nobody with problems ? Oct 27 18:52:04 oh, this is no good: Oct 27 18:52:15 6e884: e8bd8011 pop {r0, r4, pc} Oct 27 18:52:32 i think that's in memmove() though it is a bit hard to tell from the stripped binary Oct 27 18:52:54 can you put the libc_pic.a somewhere for me to look at? Oct 27 18:54:04 sure, give me a minute Oct 27 18:54:53 oh yeah, there it is in memmove.S Oct 27 18:55:08 crazy eleet assembler dudes Oct 27 18:55:51 you could try replacing that with a thumb-safe sequence and see if the situation improves, something like "ldmfd sp!, {r0, r4, pc}; bx lr" Oct 27 18:56:07 er, s/pc/lr/ Oct 27 18:56:45 http://linuxtogo.org/~mickeyl/misc/libc_pic.a Oct 27 18:56:57 ok, will give that modification a try Oct 27 18:57:32 there might be a few other instances of the same thing, you might need to grep for them Oct 27 18:58:15 so, pop is bad? Oct 27 18:58:46 pop {..., pc} is bad Oct 27 18:59:19 it will work on armv5 but it fails on armv4t Oct 27 18:59:40 for armv4t, you need to pop into another register (lr is usually ok) and then use "bx" to return Oct 27 19:00:38 bbiab, my turn to bath emma this evening Oct 27 19:00:52 hehe Oct 27 19:00:54 have fun Oct 27 19:01:13 pb_: nice... enjoy! Oct 27 19:20:49 re Oct 27 19:20:56 right, bathtime is over, milk time now Oct 27 19:22:10 mickeyl: it looks like files/armv4t-interworking.patch is what you need Oct 27 19:22:22 for some weird reason this doesn't seem to be applied to anything newer than glibc 2.6.1 Oct 27 19:22:44 and it isn't in recipes/eglibc at all, though it still appears to be necessary Oct 27 19:25:45 cd beagle Oct 27 19:26:05 pb_: ooh, great find. thanks! Oct 27 19:26:19 * mickeyl applies and rebuilds the rootfs Oct 27 19:26:31 mickeyl: funnily enough it appears that I wrote that patch almost exactly one year ago :-} Oct 27 19:26:57 re Oct 27 19:27:02 hi, I'm using angstrom kernel 2.6.29 built from narcissus on a BB revB5 and it is frozen at [38.133941] IrCOMM protocol (Dag Brattli) .. Does anyone have any idea what does it mean? Oct 27 19:27:17 pb_: hehe Oct 27 19:27:22 lim_: that would probably be a question for #beagle or #angstrom Oct 27 19:27:24 all good things come back Oct 27 19:27:33 yeah, it seems Oct 27 19:27:44 ok Oct 27 19:27:46 cd angstrom Oct 27 19:27:48 ops Oct 27 19:27:48 it is a little bit frustrating having to fix the same bugs again and again, but I guess it gets easier each time Oct 27 19:27:51 heh Oct 27 19:44:46 hi, I'm using angstrom kernel 2.6.29 built from OE on a BB revB5 and it freezes during boot in NET: Registered protocol family 10 ....ideas? Oct 27 19:45:01 ops, 2.6.28 Oct 27 19:48:07 lim_: that would probably be a question for #beagle or #angstrom Oct 27 19:48:10 shit.. Oct 27 19:48:22 even though built from oe... Oct 27 19:48:22 xserver-xorg 1.5.3 has lot of align traps on armv6 Oct 27 19:48:35 lim_: even Oct 27 20:29:36 * kergoth ponders Oct 27 20:37:41 anything in particular, or just general pondering? Oct 27 20:39:12 PACKAGES_DYNAMIC, persistence of emitted content, general pre-build acquisition of data.. the ui folk want to be abl eto see ahead of time what packages will be available to go into an image and the like, ideally it'd be able to get that info very quickly, so the ui is responsive... Oct 27 23:06:32 I'm not really sure if this is the right place to speak up but I'll do it anyways. I'm fallowing this wikipage: http://wiki.davincidsp.com/index.php/DaVinci_PSP_3.00_Release_Notes by TI trying to get a Kernel proper for my DaVinci precssor. Anyways when I try to git checkout the git tells me that some stuff would be overwritten. So I have to --force it. Why is this, or what am I doing wrong? Oct 27 23:12:34 Is this the wrong place to talk about embedded kernels pulled from git? Oct 27 23:26:40 hi! Oct 27 23:27:11 I'm trying to use the stable branch... with minimal-uclibc, but I've got an error saying the distro is not found... Oct 28 00:00:47 Rovanion: probably the proper place would be #arago... anyway, what is your problem? Oct 28 00:01:50 denix: When I'm trying to checkout with git, git tells me that some files already exst, all the way trough the guide Oct 28 00:03:55 you can checkout to a branch... any specifics? can you post the output at pastebin.org? Oct 28 00:07:27 denix: Well I've gone trough everything with a -f after git. #git told me so. Anyways, here's the output: git checkout DEV.DaVinciPSP.03.00.00.00 Oct 28 00:07:27 error: Entry 'CREDITS' would be overwritten by merge. Cannot merge. Oct 28 00:08:48 how about "git checkout -b psp30 DEV.DaVinciPSP.03.00.00.00"? Oct 28 00:09:25 Same result Oct 28 00:30:57 Rovanion: your local tree may not be clean, I don't see this error. try resetting the tree "git reset --hard HEAD" Oct 28 00:31:17 denix: Will do Oct 28 00:33:00 And I should now fallow the guide top > down again? denix ? Oct 28 00:34:56 reset and checkout Oct 28 00:49:01 denix: It's working correctly now. Thank you so much. Too bad I'm planning on returning my device tomorrow Oct 28 02:07:21 I'm trying to build with the sourcery toolchain as stated in a TI guide but when I try I get this error: http://pastebin.org/48858 . I don't understand the error: What do I do? **** ENDING LOGGING AT Wed Oct 28 02:59:58 2009