**** BEGIN LOGGING AT Sun Sep 13 02:59:56 2009 Sep 13 03:09:21 http://gist.github.com/186063 Sep 13 03:48:08 * kergoth experiments with an 'events' flag for event handlers Sep 13 03:50:34 hmm Sep 13 04:56:50 hmmmm Sep 13 10:47:20 03Brijesh Singh  07org.openembedded.dev * r1cf11450bf 10openembedded.git/recipes/ti/ (3 files): lpm, sdma, cmem recipes: rename to match PV Sep 13 10:47:21 03Brijesh Singh  07org.openembedded.dev * r5c635b365f 10openembedded.git/recipes/ti/ (5 files in 2 dirs): gstreamer-ti: bump SRCREV and apply patch from tracker 462 Sep 13 10:47:21 03Koen Kooi  07org.openembedded.dev * r8d8f4b1fdc 10openembedded.git/ (conf/checksums.ini recipes/gnome/libsoup-2.4_2.27.4.bb): libsoup-2.4: update to 2.27.91 Sep 13 10:47:32 03Brijesh Singh  07org.openembedded.dev * r65a3216b1d 10openembedded.git/recipes/ti/README: dvsdk: update README to use ce 2.24.01 Sep 13 10:47:32 03Koen Kooi  07org.openembedded.dev * r3a3fd40fec 10openembedded.git/: Merge branch 'org.openembedded.dev' of ssh://dominion.thruhere.net/var/cache/git/openembedded into org.openembedded.dev Sep 13 11:57:18 http://oe.linuxtogo.org/ - not working ^_^ Sep 13 12:17:30 | /local/pkg/oe/om-gta02/tmp/cross/armv4t/lib/gcc/arm-oe-linux-gnueabi/4.4.1/../../../../arm-oe-linux-gnueabi/bin/as: unrecognized option '--fix-v4bx' Sep 13 12:17:34 khem|afk: this with 4.4.1 Sep 13 12:17:38 anyone seen? Sep 13 12:57:45 hi Sep 13 12:57:54 I m a newbie in OE Sep 13 12:58:12 can someone tell me how to add a package to console-image? Sep 13 13:09:14 hi Sep 13 13:09:39 mickeyl, no... Sep 13 13:10:12 SeldanB, you must edit .bb file manually in recipes dir Sep 13 13:10:38 console-image.bb? Sep 13 13:23:51 yes Sep 13 13:30:31 fpga: sorry, but can you tell me what to do if I want to add alsa for example? Sep 13 13:31:29 sorry, I dont know :) Sep 13 13:32:12 I trying to find some info about creation of my own packages, but the only way I see is to learn by examples... Sep 13 13:35:33 same here, I want to add alsa, ethernet and my own package, so far, I m not able to do that >_< Sep 13 13:49:28 hmm. any idea how to extract an initrd which 'file' says is in uImage format? Sep 13 13:51:58 crz_, dd if=uImage of=initrd bs=1 skip=64 Sep 13 13:52:05 mike^, cool, thanks Sep 13 13:53:30 SeldanB, you can add alsa and other existing packages using ANGSTROM_EXTRA_INSTALL in your local.conf Sep 13 13:55:17 do somebody know what is remotfs? It uses FUSE and much more easier than CIFS and NFS... Sep 13 13:55:55 fpga, this? http://remotefs.sourceforge.net/ Sep 13 13:55:58 I want to add remotefs server deamon and client to OpenEmbedded... but do not know how... Sep 13 13:56:03 yes, this Sep 13 13:56:06 mike^: ANGSTROM_EXTRA_INSTALL = "alsa" ? Sep 13 13:56:40 yep Sep 13 13:56:43 mike^, thanks again, worked like a charm. Sep 13 13:57:06 I have many many troubles and problems, I hate NFS and CIFS, it works but with problems, so I want RFS, remotefs ^_^ Sep 13 13:57:43 SeldanB, sorry, I think alsa is split into several packages, it should be rather 'alsa-lib' and 'alsa-utils' Sep 13 14:00:05 anyone have any idea why you would need to override do_configure with {; oe_runconf; } ? Sep 13 14:01:45 mike^: thx Sep 13 14:05:03 anyone know who the linphone guy is Sep 13 14:11:08 03Martin Jansa  07shr/import * r89c3e5f472 10openembedded.git/conf/distro/include/preferred-shr-versions.inc: Fix nonexistent preferred-shr-versions Sep 13 14:17:24 hi, I was wondering why some patches sent from git seems to be deep in a thread in the archive page http://lists.linuxtogo.org/pipermail/openembedded-devel/2009-September/thread.html Sep 13 14:17:39 s/seems/seem Sep 13 14:56:52 <_whitefox> Crofton|work hi Sep 13 14:57:18 <_whitefox> I got openbts bitbaked Sep 13 14:57:21 cool Sep 13 14:59:56 03Klaus Kurzmann  07shr/import * r0ae524d82f 10openembedded.git/conf/distro/include/ (shr-autorev-unstable.inc shr-autorev.inc): Sep 13 14:59:56 shr-autorev(-unstable): update some refs for testing and sync both Sep 13 14:59:56 Signed-off-by: Klaus Kurzmann Sep 13 15:02:23 <_whitefox> Crofton|work if I want to bitbake it for gumstix instead of beagleboard, I just need to change the MACHINE setting in local.conf? Sep 13 15:02:39 <_whitefox> we have tried a few things... but we are new to gumstix env Sep 13 15:03:02 mickey_away: what version of binutils is in use ? Sep 13 15:04:04 khem: 2.18 Sep 13 15:04:09 that too old for 4.4.1? Sep 13 15:07:47 <_whitefox> hi khem Sep 13 15:08:49 _whitefox, THAT SHOULD WORK Sep 13 15:09:26 <_whitefox> what should the MACHINE setting be changed to? or I just use the same one for both beagleboard and gumstix? Sep 13 15:09:27 * mwester covers his ears Sep 13 15:09:45 look in conf/machine Sep 13 15:10:00 I think the machine is "overo" Sep 13 15:10:07 look for the overo.conf file Sep 13 15:10:09 mickey_away: let me have a look Sep 13 15:14:06 morning Sep 13 15:16:59 mickey_away: 2.18 has this option hmm Sep 13 15:17:55 mickey_away: which machine is it for Sep 13 15:18:07 mickey_away: I assume distro is minimal Sep 13 15:18:23 kergoth: _whitefox g'morning Sep 13 15:18:35 <_whitefox> hi Sep 13 15:18:51 morning Sep 13 15:19:15 hi chouimat Sep 13 15:19:23 NOTE: The MD5Sums did not match. Wanted: 'b64a8ae8592b4d32ac9baebf0b915c1e' and Got: 'd41d8cd98f00b204e9800998ecf8427e' Sep 13 15:19:26 NOTE: Task failed: Checksum of 'http://projects.linuxtogo.org/frs/download.php/222/atd-0.80.tar.gz' failed Sep 13 15:19:29 interesting Sep 13 15:20:12 hmm Sep 13 15:20:20 I wonder if this has to do with moving ltg? Sep 13 15:21:13 did url change Sep 13 15:21:45 http://gist.github.com/186063 Sep 13 15:22:02 made and experimented with a prototype of that last night Sep 13 15:24:33 khem: yes, minimal Sep 13 15:26:32 mickey_away: machine ? Sep 13 15:26:49 <_whitefox> khem: my build problem is solved Sep 13 15:27:25 mickey_away: and failing package plz Sep 13 15:27:42 _whitefox: good how did it happen Sep 13 15:27:52 <_whitefox> help from Crofton Sep 13 15:28:06 <_whitefox> then I figured some files need to be removed in tmp for it to recompile properly Sep 13 15:28:18 khem: x86_64 building for MACHINE=om-gta02, failing package is the kernel (openmoko-kernel) Sep 13 15:30:34 mickey_away: I think one problem could be that defaults are armv5t for EABI in compiler Sep 13 15:31:01 let me try with your config and see if I can reproduce it Sep 13 15:31:05 ok, thanks a lot Sep 13 15:51:09 03Klaus Kurzmann  07shr/import * r8252c92d31 10openembedded.git/ (10 files in 3 dirs): Sep 13 15:51:09 eliminate shr.bbclass Sep 13 15:51:09 Signed-off-by: Klaus Kurzmann Sep 13 16:38:26 Hmm, any opinion on a possible long term deprecation of anonymous python functions in favor of event handlers that responds to a new RecipeParsed event? It's functionally the same, but makes it much more clear and explicit Sep 13 17:32:58 mickey_away: your config worked ok for me with binutils cvs Sep 13 17:33:10 mickey_away: I am trying with 2.18 now Sep 13 17:50:31 kergoth: I'd be in favour Sep 13 17:50:43 i'm trying to focus on some of our usability issues Sep 13 17:50:56 trying, anyway... i have little experience with hci :) Sep 13 17:51:09 RP: http://gist.github.com/186063 Sep 13 17:51:09 kergoth: :) Sep 13 17:52:29 kergoth: Hmm, doesn't seem to buy us much IMO :/ Sep 13 17:52:55 * RP needs to think about it a bit Sep 13 17:53:12 'addhandler' could go away, along with all the common code thats in -every- event handler right now.. getname / isinstance Sep 13 17:53:26 just a proof of concept, something to consider Sep 13 17:53:36 s/all/most/ Sep 13 17:53:55 :) Sep 13 17:54:08 kergoth: That is true, our currently handler code sucks... Sep 13 17:55:28 this deprecates bb.event's registration/firing code entirely, adds a single 'fire_event' method to data_smart, removes the passing of the data into the event handler, and passes both event and the data_smart instance into the handler Sep 13 17:57:10 guh, i really need to read more about object oriented design Sep 13 17:59:00 kergoth: Did you see that layout_* email? Sep 13 17:59:20 yeah, seems reasonable to me Sep 13 18:12:40 this is just odd. i'm not unable to reflash the damn poodle Sep 13 18:12:57 *i'm not able Sep 13 18:37:05 hi all Sep 13 18:41:41 hi florian Sep 13 18:52:27 * chouimat|nap hates Sundays Sep 13 19:12:43 florian: md5 checksum for atd does not match anymore Sep 13 19:22:04 03Khem Raj  07org.openembedded.dev * rdbecbf49e2 10openembedded.git/recipes/sysstat/sysstat_8.1.6.bb: Sep 13 19:22:04 sysstat_8.1.6.bb: Fix compilation on uclibc Sep 13 19:22:04 * Add dependency on virtual/libintl Sep 13 19:22:04 Signed-off-by: Khem Raj Sep 13 19:22:08 03Khem Raj  07org.openembedded.dev * r34e6a1ce23 10openembedded.git/recipes/nfs-utils/ (2 files in 2 dirs): Sep 13 19:22:10 nfs-utils_1.1.2.bb: Fix compilation on uclibc Sep 13 19:22:12 Signed-off-by: Khem Raj Sep 13 19:22:35 hi Sep 13 19:23:27 can anyone give me a hand with some bitbake HEAD problems? Sep 13 19:23:48 i try to set up a new machine with minimal distro Sep 13 19:24:08 is current HEAD supposed to work with oe.dev? Sep 13 19:24:22 ulitu: yes it should Sep 13 19:24:40 firt thing, bitbake didn't find included/required files Sep 13 19:24:48 ^^^ first^^^ Sep 13 19:25:02 i had to revert a commit Sep 13 19:25:12 have you set up your local conf Sep 13 19:25:16 which commit Sep 13 19:25:17 commit f71171f445fc7d846772ce5a87caf488df006cfc Sep 13 19:25:17 Author: Chris Larson Sep 13 19:25:17 Date: Wed Apr 8 17:37:30 2009 +0000 Sep 13 19:25:17 Make the bbpath change for includes temporary rather than permanent. Sep 13 19:25:19 yep Sep 13 19:25:34 # Use this to specify where BitBake should place the downloaded sources into Sep 13 19:25:34 DL_DIR = "${HOME}/sources" Sep 13 19:25:34 BB_NUMBER_THREADS = "1" Sep 13 19:25:34 BBFILES = "openembedded/recipes/*/*.bb" Sep 13 19:25:34 DISTRO = "minimal" Sep 13 19:25:35 MACHINE = "s101" Sep 13 19:26:16 hmm, i'll see about reverting that, i think we had to temporarily revert it for mvl6 too. need to revisit that, as that change really should go in Sep 13 19:26:29 after reverting that commit the complaints went away Sep 13 19:27:24 but now, bitbake complaints about file:// in SRC_URI with : Sep 13 19:27:24 NOTE: Running task 3 of 1491 (ID: 29, openembedded/recipes/shasum/shasum-native.bb, do_unpack) Sep 13 19:27:24 NOTE: Unpacking tmp/work/i686-linux/shasum-native-1.0-r1/main.c to tmp/work/i686-linux/shasum-native-1.0-r1/ Sep 13 19:27:24 cp: cannot stat `/home/luckas/Projekte/ePaper/oe-build/tmp/work/i686-linux/shasum-native-1.0-r1/main.c': No such file or directory Sep 13 19:27:24 NOTE: Task failed: Sep 13 19:28:25 main.c is there, in openembedded/recipes/shasum/files/ Sep 13 19:29:33 * kergoth does builds with bitbake master branch pretty regularly without problems Sep 13 19:29:34 i source this script: Sep 13 19:29:35 BASE=`pwd` Sep 13 19:29:35 export BBPATH=${BASE}:${BASE}/openembedded/ Sep 13 19:29:35 export PATH=${BASE}/bitbake/bin:$PATH Sep 13 19:31:46 maybe it's my fault. i do have some experience with oe stable. but i haven't looked into dev for a while Sep 13 19:32:56 kergoth: have you done a built from scratch lately? Sep 13 19:37:02 ulitu: yes as of this morning Sep 13 19:37:14 using minimal .dev and git master Sep 13 19:37:20 for bitbake Sep 13 19:37:44 any changes in bitbake other then reverting f71171f445fc7d846772ce5a87caf488df006cfc ? Sep 13 19:37:56 or even without reverting? Sep 13 19:39:05 no reverting to bitbake Sep 13 19:39:14 its plain master whatever is there Sep 13 19:47:39 03Stefan Schmidt  07org.openembedded.dev * r430db434a7 10openembedded.git/conf/distro/ (include/moko-autorev.inc openmoko.conf): Sep 13 19:47:39 moko-autorev.inc: Remove no longer used file. Sep 13 19:47:39 This file was only used during the developemnet of the OM2007.x releases which Sep 13 19:47:39 are dead by now. If you have this file in your local.conf it can be safely Sep 13 19:47:39 removed. Sep 13 19:47:43 03Stefan Schmidt  07org.openembedded.dev * r848446695a 10openembedded.git/recipes/efl1/emotion_svn.bb: emotion_svn: Remove not needed sed hack for edje_cc. Sep 13 19:47:48 03Stefan Schmidt  07org.openembedded.dev * rf16e563197 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Sep 13 19:48:06 03Klaus Kurzmann  07shr/import * rbb0a144dd2 10openembedded.git/conf/distro/include/preferred-shr-versions.inc: Sep 13 19:48:06 preferred-shr-versions.inc: bump vala-native Sep 13 19:48:06 Signed-off-by: Klaus Kurzmann Sep 13 19:50:58 hi Sep 13 19:53:45 there are sombody? Sep 13 19:59:57 hi sushisan Sep 13 20:00:42 hi ulitu Sep 13 20:01:00 can I make you some question about bitbake? Sep 13 20:06:05 sushisan: just ask Sep 13 20:11:31 well . The bottleneck in the bitbake process is when try to resolve any missing task queue dependencies. If I'm shure abut this...there is some flag to avoid this check? (if I made a -c clean and a remake of an a image I have 2 checks!!) Sep 13 20:12:48 is it possible to instruct bitbake where I want to unpack source archive? Sep 13 20:12:55 sushisan: I think you can't eliminate the time of bitbake 'bootstrapping' Sep 13 20:13:11 JaMa: why do you need it? Sep 13 20:13:39 when there is tar.gz without directory inside its difficult to pack the conntent when S==${WORKDIR} Sep 13 20:13:51 sushisan: bitbake -b recipie.bb Sep 13 20:14:08 blah I mean to install content of package without temp image patches etc Sep 13 20:14:58 my situation got worse :-( Sep 13 20:15:12 now i get tons of:ERROR: opening openembedded/recipes/opie-embeddedkonsole/opie-embeddedkonsole_1.2.3.bb: file openembedded/recipes/opie-embeddedkonsole/opie-embeddedkonsole_1.2.3.bb not found Sep 13 20:15:53 how does bb know opie-embeddedkonsole_1.2.3.bb is there and then can't open it??? Sep 13 20:16:06 and yes, opie-embeddedkonsole_1.2.3.bb is readable ;-) Sep 13 20:16:56 this message comes for (probably) every single bb file Sep 13 20:17:39 ulitu: txs!!! Sep 13 20:17:46 :-) Sep 13 20:18:16 ulitu: what is your build box running (OS) Sep 13 20:18:30 I beleave there are not any reference to this in the BB manual Sep 13 20:19:50 ulitu: about unpacking? but maxrev parametr of patch is not mentioned there too :/ Sep 13 20:19:58 sushisan: ^^ Sep 13 20:20:12 debian unstable Sep 13 20:20:25 python2.6 from experimental Sep 13 20:20:26 ulitu: what is your shell Sep 13 20:20:31 ls -ltr /bin/sh Sep 13 20:20:31 bash Sep 13 20:20:42 ok Sep 13 20:20:51 and Do you have BBFILES set properly Sep 13 20:20:54 # ls -ltr /bin/sh Sep 13 20:20:55 lrwxrwxrwx 1 root root 4 2009-08-31 00:44 /bin/sh -> bash Sep 13 20:21:16 BBFILES = "openembedded/recipes/*/*.bb" Sep 13 20:21:25 openembedded is my git clone Sep 13 20:21:26 khem: you are connected here 7x24? :-) Sep 13 20:22:10 sushisan: heh not 24x7 but mostly Sep 13 20:22:10 is there any kind of file browser... like nautilus for ubuntu in the angstrom demo images? Sep 13 20:22:11 are relative paths in BBFILES OK? Sep 13 20:22:20 I dont think so Sep 13 20:22:45 ulitu: Sep 13 20:22:55 I have absolute Sep 13 20:23:21 morning Sep 13 20:23:36 khem: I'll try Sep 13 20:24:00 khem: I tested your patch about armv5te-linux bogus dir..it has gone :) Sep 13 20:24:07 that helped Sep 13 20:24:13 ant__: cool ack it Sep 13 20:24:24 khem: nice catch...it's there at least since a couple of years iirc :) Sep 13 20:24:59 khem: is that a bug or a feature? Sep 13 20:25:10 shouldn't relative paths work for BBFLIES? Sep 13 20:25:27 relative to what ? Sep 13 20:25:28 anyway, thanks alot Sep 13 20:25:41 my current working dir Sep 13 20:27:03 well I guess parsing of the oe tree should not depend on my cwd. all right Sep 13 20:28:02 ulitu: :) Sep 13 20:29:01 TMPDIR and BBFILES are absolute Sep 13 20:29:34 i thought Einstein said everything was relative ;-) Sep 13 20:30:35 by the way, oe saves a lot of my time every day. thanks guys Sep 13 20:30:49 heh einstein was a moron Sep 13 20:33:04 ant__: there is another bogus thing $TMPDIR/usr Sep 13 20:33:24 opkg-native creates it for some reason I think Sep 13 20:33:29 I will chase it down one day Sep 13 20:33:54 khem do you have any interest in gnupg_1.4.10.bb? Sep 13 20:34:17 good night guys Sep 13 20:34:21 khem: a, yes, that's recent Sep 13 20:34:52 m4t: Whats wrong with it Sep 13 20:34:56 I'm investigating opkg too..it's failing to find some packages during rebuild from pstage Sep 13 20:35:07 i get some link errors when i try to build it Sep 13 20:35:11 it 'doesn't exist' Sep 13 20:35:17 ant__: this might be related Sep 13 20:35:22 m4t: what link errors Sep 13 20:35:22 trying to bump the version to the latest stable in 1.x series Sep 13 20:35:27 khem: that's my suspect Sep 13 20:35:28 lessee Sep 13 20:35:40 content of that dir has some staging information Sep 13 20:35:43 http://pastebin.com/m6b7d7440 shows the build error Sep 13 20:35:47 khem: btw I'm investigating on the relation btw OELAYOUT_ABI and DISTRO_PR Sep 13 20:36:37 ../../powerpc-angstrom-linux-uclibc/bin/ld: Dwarf Error: Offset (553) greater than or equal to .debug_str size Sep 13 20:36:52 these are in PSTAGE_EXTRAPATH... nice thing is both are overwritten by distro settings..// Sep 13 20:37:46 m4t: which binutils ? Sep 13 20:38:15 binutils-cross_2.19.51.0.3.bb Sep 13 20:39:29 m4t: I see you're using uclibc. fwiw with glibc NOTE: Running task 782 of 782 (ID: 0, /oe/org.openembedded.dev/recipes/gnupg/gnupg_1.4.2.2.bb, do_rm_work_all) Sep 13 20:39:47 is there any kind of file browser... like nautilus for ubuntu in the angstrom demo images? Sep 13 20:39:52 must be an header mismatcg Sep 13 20:40:09 julemore: mc Sep 13 20:40:15 mc? Sep 13 20:40:18 just mc? Sep 13 20:40:19 ok Sep 13 20:40:29 you can't be wrong with it ..:) Sep 13 20:40:41 lol k Sep 13 20:40:56 julemore: with a mionor editing mc opens oe's ipk too Sep 13 20:41:33 ant__: oh this is 1.4.10 Sep 13 20:41:40 mc: command not found Sep 13 20:41:46 1.4.2.2 compiles okay Sep 13 20:42:06 there isnt any mc Sep 13 20:42:06 m4t: ah, I see Sep 13 20:42:24 julemore: ? Sep 13 20:42:49 in the rootshell i type mc Sep 13 20:42:55 and its not found Sep 13 20:43:07 opkg install mc? Sep 13 20:43:16 http://pastebin.com/d186acb70 <--- is 'gnupg_1.4.10.bb', it probably wont work under glibc either Sep 13 20:43:53 m4t: do you need this particualr version Sep 13 20:44:01 there 2.0.13 out there Sep 13 20:44:22 khem the 2.x is 'modularized' and seems like it'd be a big leap to get it packaged etc. Sep 13 20:44:32 m4t: if you paste the .bb somewhere I can test it against glibc Sep 13 20:44:49 ant__ check that last pastebin url ^^ Sep 13 20:45:07 oopss.was writing at the same mom :) Sep 13 20:46:06 khem i believe the 1.4.x series is still being actively developed Sep 13 20:46:19 svn has the same issue, btw Sep 13 20:46:37 is there any kind of file browser BY DEFAULT... like nautilus for ubuntu in the angstrom demo images? Sep 13 20:47:52 m4t: could you use binutils 2.19.1 Sep 13 20:48:20 the HJ Lu's binutils releases are good for x86 only Sep 13 20:48:33 sushisan: look again. http://bitbake.berlios.de/manual/ch04s02.html#id870114 Sep 13 20:48:41 sushisan: "Example 4.1. Executing a task against a single .bb" Sep 13 20:48:49 i could...i dont want to mess up my cross toolchain though (ie. i need to be able to revert easily) Sep 13 20:49:22 khem this is really the only problem ive run into with that binutils release as well Sep 13 20:52:07 ant__: what 's mc? Sep 13 20:52:35 midnight commander, its a curses based file manager shell Sep 13 20:52:46 blue and white, iirc Sep 13 20:53:37 kergoth: txs, yes, I know the use of bb -b but in the manual there are not a notice about the avoid of the dependency checking in this mode.... Sep 13 20:54:33 seems pretty obvious to me. it tells you it's parsing only that one file, and building a task for just that file Sep 13 20:54:44 if it's just that file, how would it be able to build anything else? magic? Sep 13 20:55:08 it would be good to make it more explicit though, that's a good point Sep 13 20:55:31 * kergoth thinks we should use something other than docbook, docbook is rather painful Sep 13 20:56:02 kergoth: true docbook is evil ... Sep 13 20:56:20 i was thinking about converting to restructuredtext Sep 13 20:56:43 kergoth: I stopped at latex and context ... they are good enough for my needs Sep 13 21:01:17 xml Sep 13 21:05:15 I have Beagle Board ARM 600mhz right? i want to encode frames from a webcam... what do u think is the best resolution may I use? 320? 640? 1024? i dont want to collapse the cpu.. Sep 13 21:05:33 kergoth: do you have 2 mins for 2 rapid-straight questions? Sep 13 21:08:04 kergoth: well , wasn't too obvious to me. Maybe are my problem (another one!!!) or a newbie case... a more explicit paragraph are not a problem and cat help a lot of people not very expert in OE Sep 13 21:08:56 kergoth: there aren't a translate gorupe for the manuals? Sep 13 21:09:05 groupe Sep 13 21:17:03 kergoth: 1) after sync-with-poky I have in packaged-staging.bbclass PSTAGE_EXTRAPATH ?= "/${OELAYOUT_ABI}/${DISTRO_PR}/" but in OE is overwritten (in some DISTRO's) by e.g. ./distro/include/angstrom.inc:PSTAGE_EXTRAPATH = "${LIBC}" Sep 13 21:17:03 2) ./abi_version.conf:OELAYOUT_ABI = "3" how does it cope with DISTRO_PR? Isn't one just enough to mark binary incompatibility?OELAYOUT_ABI seems unused in OE Sep 13 21:17:03 3) (I lied, 3 questions :) I see strange things for some packages e.g. * Package staging-kexecboot-armv5te-angstrom-linux-gnueabi version 0.5-r3+gitr0d4d33d85baa6251de3b1e20bd7bd464f11c5097 has no architecture specified, defaulting to i686-linux. But then | opkg-cl -f -> * Cannot find package kexecboot. Sep 13 21:17:10 khem: ^^ Sep 13 21:36:03 hi Sep 13 21:36:08 I've some patches Sep 13 21:36:47 some I have to have prior discussion before pushing them Sep 13 21:39:02 I've a temporary fix for mdev recipe that is not part of busybox Sep 13 21:39:34 +TARGET_CC_ARCH += "-I${STAGING_DIR_HOST}/usr/include/xorg" in xorg-driver-common Sep 13 21:39:51 and another patch for xfbdev 0.4.0 Sep 13 21:39:58 and another patch that I can apply Sep 13 21:40:15 the patch for xfbdev is a patch for the include...it replace "" by <> Sep 13 21:40:54 TARGET_CC_ARCH is not how you add include paths. Sep 13 21:40:59 that's what CPPFLAGS/CFLAGS is for Sep 13 21:43:45 yes I know it's a dirty hack Sep 13 21:44:00 I'll try another way with CPPFLAGS/CFLAGS Sep 13 21:44:18 it was done on some recipes because it was the only way(for cc_arch) Sep 13 21:47:11 khem, kergoth: If you're interested in toolchain improvements at some point OE should look at removing the -L option from LDFLAGS and the -isystem from CFLAGS Sep 13 21:47:30 (Since sysroot takes care of that now - Poky has done this for a while) Sep 13 21:48:26 RP: ah, good idea. thanks Sep 13 21:48:30 hmm Sep 13 21:52:09 RP: hi there Sep 13 21:53:37 RP: the 3 questions above would be for you of course...the poor kergoth is my secondary-target ;) Sep 13 21:58:40 RP: its only the old compilers we need to support if we agree to not support gcc 3.3 and 3.4 I can do that Sep 13 22:04:46 khem: How about we just advise anyone using those compilers that they need to tweak the variables? Sep 13 22:05:03 khem: and I thought it was just 3.3 that was an issue, 3.4 has sysroot? Sep 13 22:05:40 Is it possible to instruct bitbake to unpack source archive to specified directory? Sep 13 22:06:16 Didn't found it in OE manual and no luck in bitbake code :/ Sep 13 22:07:00 JaMa: I don't think that is possible Sep 13 22:07:34 ant__: I thought OELAYOUR_ABI was used in sanity.bbclass? Sep 13 22:07:56 ant__: As for the other issues, I don't know :/ Sep 13 22:08:09 RP: ok, thanks Sep 13 22:08:24 RP: I tried removing all but pstage and leavinf the whole deploy dir Sep 13 22:08:31 JaMa: We probably could improve the unpack function mind... Sep 13 22:08:37 there are packages failing Sep 13 22:08:52 in the first case there are 3 more... Sep 13 22:08:59 ant__: Hmm :/ Sep 13 22:09:07 I'll pastebin Sep 13 22:09:28 ant__: I don't have time atm to look into something like that :/ Sep 13 22:09:51 RP: np, I just find strange so many failures Sep 13 22:09:53 http://fr.pastebin.ca/1564700 Sep 13 22:10:05 I knew zaurus-updater and kexecboot are evil ;) Sep 13 22:10:23 btw all fine building from scratch Sep 13 22:10:36 it's just packaged staging terribly broken Sep 13 22:10:51 I'm using poky's latest rev with some OE patches Sep 13 22:11:02 (packaged-staging.bbclass) Sep 13 22:11:06 RP: my case can be solved with upstream packing source to directory package-version.. so its ok for me.. when its not just config param to SRC_URI line :) Sep 13 22:11:41 JaMa: You can always override ${S} Sep 13 22:11:55 ant__: Check Poky's zaurus-updater Sep 13 22:12:08 ok, but for the others? Sep 13 22:12:16 they fails in both cases Sep 13 22:13:05 I mean even with full deploy dir (pstage inside) Sep 13 22:13:23 ant__: Its odd its saying they have no arch specified Sep 13 22:13:23 see..task-base ! Sep 13 22:13:33 ant__: Have a look at the packages and see if that is set? Sep 13 22:13:43 yes, that's a long time it emits these notes Sep 13 22:13:59 RP: problem is in custom setup.py in source.. it tries to install all stuff in source directory... which includes image and then loops endlessly Sep 13 22:14:20 JaMa: ouch Sep 13 22:14:44 RP: btw what do you thing about this small patch? http://article.gmane.org/gmane.comp.handhelds.openembedded/26150/match=martin+jansa i'm using it for some time and consider it usefull Sep 13 22:15:03 s/thing/think/ Sep 13 22:15:46 JaMa: I've seen that and quite like it. Should go to the bitbake list though ;-) Sep 13 22:16:11 JaMa: We need to find a better way to write that if statement mind... Sep 13 22:18:11 RP: just split it for PV, PR, PE? Sep 13 22:18:53 JaMa: Probably write a function which returns true/false or something Sep 13 22:19:00 RP: ouch...looking at udev Sep 13 22:19:06 make it readable... Sep 13 22:19:34 RP: should I do it and send that patch again, or should someone more expirienced with python and bitbake? Sep 13 22:19:54 in the staging-ipk I see /pkgdata//c7x0-angstrom-linux-gnueabi and staging armv5te-angstrom-linux-gnueabi Sep 13 22:20:02 ? Sep 13 22:20:22 JaMa: Either, but sent it to the bitbake list as well as the oe one Sep 13 22:20:42 RP: ok Sep 13 22:21:07 ant__: hmm, I don't have the context or the time to work out the context, sorry :( Sep 13 22:21:43 khem: Have you ever seen glibc enter and infinite loop entering and exitting the nptl directory? Sep 13 22:22:19 RP: this 'has no architecture specified' ..I have to see the code how Sep 13 22:22:45 but if the staging- ipk has b0rked paths... Sep 13 22:31:51 is there any bitbake for ffmpeg? Sep 13 22:32:27 or is it a opkg install ffmpeg? Sep 13 22:34:41 RP: done.. sending in few minutes after test Sep 13 22:35:05 RP: :[ Sep 13 22:35:10 Package: staging-udev-c7x0-angstrom-linux-gnueabi Sep 13 22:35:17 Architecture: i686-linux Sep 13 22:36:31 ant__: Something is badly broken then... Sep 13 22:36:47 ant__: It would be interesting to see if Poky has the same problem Sep 13 22:37:02 * RP -> Zzzz Sep 13 22:37:11 'nite Sep 13 22:37:40 RP: thats new but it can happen if you have dangling symlinks in the tree **** ENDING LOGGING AT Mon Sep 14 02:59:57 2009