**** BEGIN LOGGING AT Mon Mar 14 02:59:57 2011 Mar 14 06:25:48 after running bitbake virtual/kernel, how do i make sure that automatic cleanup is not enabled? and what is the location of uImage once the command finishes running? Mar 14 06:45:13 NotTooDumb3: I get my uImage from $TMPDIR/deploy/, eg tmp/deploy/eglibc/images/beagleboard/uImage-2.6.37-r100a+gitrfa3b4e23ec20cfc944db7cc2b30b0d82c20e4472-beagleboard.bin Mar 14 06:46:34 we get .bin? not just uImage? Mar 14 06:47:03 yeah it's a uImage you can use with u-boot Mar 14 06:47:38 there's a "work" directory where most things seem to get built Mar 14 06:48:08 in my case: tmp/work/beagleboard-angstrom-linux-gnueabi/linux-omap-2.6.37-r100a+gitrfa3b4e23ec20cfc944db7cc2b30b0d82c20e4472/ Mar 14 06:50:51 oh..ok i am building Mar 14 06:52:02 i am just running now, and there are no folders in tmp except cache, they may get generated slowly i think right Mar 14 06:52:57 if you're starting from scratch it might take a while Mar 14 06:56:00 yes i am starting from scratch Mar 14 07:37:40 gm Mar 14 08:17:00 good morning Mar 14 08:25:16 good morning Mar 14 08:43:50 mornin Mar 14 08:44:05 moin Mar 14 09:03:18 morning all Mar 14 09:11:48 jo ant Mar 14 09:16:29 good morning Mar 14 09:20:22 hi ant_work, woglinde Mar 14 09:20:40 hi bluelightning Mar 14 09:25:29 03Eric Bénard 07org.openembedded.dev * r0cc657811b 10openembedded.git/conf/distro/include/preferred-slugos-versions.inc: Mar 14 09:25:29 slugos: prefer busybox 1.18.4 Mar 14 09:25:29 Signed-off-by: Eric Bénard Mar 14 09:25:29 Acked-by: Mike Westerhof Mar 14 09:25:32 03Eric Bénard 07org.openembedded.dev * r21780b0df1 10openembedded.git/recipes/busybox/ (10 files in 3 dirs): Mar 14 09:25:32 busybox: remove 1.18.3 Mar 14 09:25:32 now that 1.18.4 is default for angstrom and slugos Mar 14 09:25:32 Signed-off-by: Eric Bénard Mar 14 09:25:32 Acked-by: Khem Raj Mar 14 09:25:53 *sigh* busybox again Mar 14 09:28:36 hi woglinde Mar 14 09:31:03 good morning Mar 14 09:32:04 hi ericben_ Mar 14 09:32:16 hi ericben Mar 14 09:32:23 -> 10:23 < woglinde> *sigh* busybox again Mar 14 09:32:25 *g* Mar 14 09:32:26 ericben_: does your latest apache recipe build with libtool 2.4 ? Mar 14 09:32:51 eFfeM_work: I tested it from scratch using angstrom 2010 and minimal distro so yes it should build Mar 14 09:33:12 cool, will give it a try tonight (time permitting) Mar 14 09:33:15 I'm actually preparing a board to test all that Mar 14 09:33:48 I need that for a training tomorrow ... and it took my quite some time to get it compiling this weekend Mar 14 09:34:03 training? Mar 14 09:34:09 ericben_: btw: as hrw pointed on ML why not even drop P_V from distro configs and use latest busybox with positive D_P? Mar 14 09:34:14 hi jama Mar 14 09:34:19 hillct: woglinde Mar 14 09:34:26 hi :) Mar 14 09:35:12 apache is a PITA to cross compile Mar 14 09:36:18 03Martin Jansa 07master * r2da8e9ff10 10openembedded.git/recipes/shr/ (4 files): Mar 14 09:36:18 shr: bump libshr-glib, shr-specs, libphone-ui, phonefsod SRCREV Mar 14 09:36:18 Signed-off-by: Martin Jansa Mar 14 09:36:31 effem yes Mar 14 09:36:34 I give up Mar 14 09:36:37 on it Mar 14 09:37:07 woglinde: I think I also spend a day or so on it in 2009 and also gave up and moved to lighty Mar 14 09:37:25 yeah lighty is much nicer Mar 14 09:38:19 but the current thing I'm working in might require apache, haven't really looked at what it would take to move to lighty, but it requires quite some apache modules (but fortunately no modphp or modsql) Mar 14 09:45:15 * woglinde wonders why khem needs nfsboot for qemu Mar 14 09:48:53 hi, quick question, in OE, how do I figure out which recipe should provide /etc/bluetooth/default or /etc/init.d/bluetooth? Mar 14 09:49:39 grep recipes Mar 14 09:49:58 JaMa: I just answered to the list, if distro whant that behaviour they can remove this line in the inc file Mar 14 09:50:09 i am trying that, find recipes | xargs egrep /etc/bluetooth/default and not complete yet but no hits Mar 14 09:50:18 it isn't bluez-utils, so am stuck a bit Mar 14 09:51:33 htns: do you have a device where this file is installed? then just do opkg search /etc/bluetooth/default Mar 14 09:51:33 htns, i think it is /etc/default/bluetooth Mar 14 09:51:46 that is in bluez-utils Mar 14 09:52:07 and for init.d: Mar 14 09:52:08 bluez/bluez-utils_3.36.bb: ${sysconfdir}/init.d/bluetooth \ Mar 14 09:52:19 bluez/bluez-utils_3.36.bb: ${sysconfdir}/default/bluetooth \ Mar 14 09:52:26 the 2nd one for default Mar 14 09:53:58 hi, what do I have to do, to build autoconf213? Build complains about missing Makefile.am . I hoped that updating openembedded would help, but now I get some more errors earlier. E.g. "Error parsing ..../bootstrap-image.bb: global name 'runtime_mapping_rename' is not defined" Mar 14 09:54:13 hi GNUtoo|laptop Mar 14 09:55:11 Cyberian: http://patches.openembedded.org/patch/1215/ may help Mar 14 09:56:44 that led me to update my oe git repository Mar 14 09:56:58 hi Mar 14 09:57:07 hi JaMa Mar 14 09:57:17 hi eFfeM_work Mar 14 09:57:37 hi gnutoo Mar 14 09:57:38 JaMa: (don't now if you got his one) I just answered to the list, if distro whant that behaviour they can remove this line in the inc file Mar 14 09:57:39 + ls ${S} | grep -q '\.in$' Mar 14 09:57:46 it seem that I made an error Mar 14 09:57:48 hi woglinde Mar 14 09:58:22 shouldn't that be ${S}/bin/ Mar 14 09:59:47 ericben_: agreed.. just last few busybox upgrade were always ACKed soon by distro maintainers.. so I guess it would make your busybox upgrades easier to send just patch to upgrade busybox and remove old Mar 14 10:00:06 the same with QT :) Mar 14 10:00:51 JaMa: on the other hand that's the problem in mainaining a recipe to the latest stable this creates changes ;-) Mar 14 10:01:13 btw JaMa for phone applications, there is now gsm mux in the kernel Mar 14 10:01:24 it works fine here using 2.6.38+ kernel Mar 14 10:01:36 I know and I've already asked mickey|bbl about it a while back Mar 14 10:02:01 and I just wrote a doc for this in git://git.kernel.org/pub/scm/linux/kernel/git/gregkh/tty-2.6.git Mar 14 10:02:12 in the tty-next branch Documentation/serial/n_gsm.txt Mar 14 10:02:15 thx eFfeM_work , odd, my build is picking bluez-4-4.59 and this doesn't seem to have /etc/init.d/bluetooth Mar 14 10:02:19 this can be of interest for shr Mar 14 10:02:42 mickey|bbl: this could be interesting for you ^^ Mar 14 10:02:44 eFfeM_work which distro is using bluez3? Mar 14 10:03:20 htns: bluetoothd is started by udev on systems with udev.. Mar 14 10:04:46 JaMa, ok, now it is starting to make sense to me. have to go figure out what should start hciattach Mar 14 10:09:26 was afk Mar 14 10:09:56 woglinde: no idea, we still have the recipe, grep gave it as answer Mar 14 10:10:25 effem hm ah Mar 14 10:11:59 actually the init.d is also listed in bluez.inc, but that file is included by no-one in the bluez dir (haven't checked other dirs) Mar 14 10:12:45 hi likewise Mar 14 10:13:13 too much junk :-( Mar 14 10:13:32 8 recipes for bluez4, 7 for obexd Mar 14 10:13:36 ericben in the tty tree isnt that file Mar 14 10:13:43 but I'm not going to clean it .... Mar 14 10:13:46 woglinde: check the tty-next brancg Mar 14 10:14:36 eFfeM_work: bluez is on my task list, I'll need it in a board I should receive by the end of the month Mar 14 10:16:06 htns: sorry no further idea on the init.d thing, only thing I know that there are quite some changes wrt bt when moving from 3 to 4 Mar 14 10:16:13 in that link, from which directory do i need to run bitbake linux-elphel -c clean this kind of clean command? in this link http://www.xora.org.uk/2009/12/10/openembeddedangstrom-kernel-workflow/ Mar 14 10:16:35 woglinde: http://git.kernel.org/?p=linux/kernel/git/gregkh/tty-2.6.git;a=blob;f=Documentation/serial/n_gsm.txt;h=397f41a1f153f4494f7dc5466934aa06c85bcc87;hb=refs/heads/tty-next Mar 14 10:17:08 ericben thanks but now I would find it myself Mar 14 10:19:10 woglinde: you may need https://lkml.org/lkml/2011/3/9/300 + https://lkml.org/lkml/2011/3/11/141 to get it working Mar 14 10:19:46 woglinde, any help for me? in that link, from which directory do i need to run bitbake linux-elphel -c clean this kind of clean command? in this link http://www.xora.org.uk/2009/12/10/openembeddedangstrom-kernel-workflow/ Mar 14 10:19:55 ericben hm I dont actual need it Mar 14 10:20:04 woglinde: so you may not need the patches :-D Mar 14 10:20:13 I only want to look at it Mar 14 10:20:30 I mostly have devices which has some more tty's Mar 14 10:20:36 be default Mar 14 10:20:39 on usb link ? Mar 14 10:20:45 yes Mar 14 10:21:05 in one project we a sierra modems which have 7 devices Mar 14 10:21:07 *g* Mar 14 10:21:14 in that case there is no problem of course, but for low cost gprs modems whihc only have a serial port this is quite useful Mar 14 10:21:21 NotTooDumb3: you do everything from your oe top directory Mar 14 10:21:45 oh..od top directory? ok thanks Mar 14 10:22:34 on my board we have a footprint for mounting gprs to 3.5G modems, but gprs ones only have a serial port Mar 14 10:22:37 ericben woulfdnt it usefull to write an udev example too? Mar 14 10:22:38 dcordes, i still get error..make_angstrom/setup-scripts/sources/openembedded] $ bitbake beagleboard -c clean bash: bitbake: command not found Mar 14 10:22:40 NotTooDumb3: th one that has recipes in it Mar 14 10:23:36 NotTooDumb3: did you setup the environment variables correctly ? is the bitbake binary in your PATH ? Mar 14 10:23:43 anybody using jffs2? I see issues with flash_eraseall -j with last recipe versions (even with teh mot-in-oe 1.4.2). Before adding the new one I'd like to see what's broken :/ Mar 14 10:23:43 woglinde: maybe, will see once I finalize the project here Mar 14 10:23:49 dcordes, no Mar 14 10:24:01 dcordes, which path has to be in PATH? Mar 14 10:24:13 NotTooDumb3: the one that has the bitbake binary Mar 14 10:24:42 ant hm Mar 14 10:24:51 didnt use jffs2 for years Mar 14 10:24:55 he.. mtd-utils: change flash_eraseall to use libmtd-wrapped ioctls probably broke it Mar 14 10:25:55 dcordes, in which directory do i find bitbake binary? Mar 14 10:26:27 NotTooDumb3: which bitbake Mar 14 10:26:48 NotTooDumb3: I don't have a crystal ball. here it is ~/linux/openembedded/bitbake/bin Mar 14 10:27:26 NotTooDumb3: the problem with that magic script you use is, you have no clue what's going on at the point running the commands Mar 14 10:27:46 NotTooDumb3: where is the guide you followed to set everything up ? Mar 14 10:28:33 I would like the _old_ http://www.angstrom-distribution.org/building-angstrom back ! Mar 14 10:28:36 any clues why I got following error?: ERROR: Error parsing .../openembedded/recipes/images/bootstrap-image.bb: global name 'runtime_mapping_rename' is not defined-:-- Mar 14 10:29:14 dcordes, http://www.angstrom-distribution.org/building-angstrom Mar 14 10:29:38 ericben_, which bitbake /usr/bin/which: no bitbake in (/usr/lib/qt-3.3/bin:/usr/kerberos/sbin:/usr/kerberos/bin:/usr/lib/ccache:/usr/local/bin:/usr/bin:/bin:/usr/local/sbin:/usr/sbin:/sbin:/home/vayavya/CodeSourcery/Sourcery_G++_Lite/bin/) Mar 14 10:31:26 NotTooDumb3: then you missed something in the doc Mar 14 10:31:31 NotTooDumb3: it's oebb.sh that sets your PATH. Mar 14 10:31:35 (in your case) Mar 14 10:32:19 ericben_, i manually set the pATH, still ti does not work Mar 14 10:32:42 ok i have to assigne to BBPATH, i did to just PATH let me see Mar 14 10:32:49 woglinde: about that patch for using linmtd, well onecomment says.. " ERASE/ERASE64 were tested on 2.6.18 and 2.6.31. OOB is untested. " Mar 14 10:32:58 woglinde: I guess OOB is failing :/ Mar 14 10:32:59 ant want= Mar 14 10:33:01 ? Mar 14 10:33:16 MTD writeoob failure Mar 14 10:33:18 error 22 (Invalid argument) Mar 14 10:33:53 NotTooDumb3: when you do "MACHINE=beagleboard ./oebb.sh bitbake virtual/kernel", it's the "./oebb.sh" that sets up you environment/PATH. read that script. Mar 14 10:37:12 do i need to run my clean command also like this? MACHINE=beagleboard ./oebb.sh bitbake beagleboard -c clean Mar 14 10:37:27 this seems to be working Mar 14 10:38:21 can i edit the kernel source code directly in ls sources/downloads/linux-2.6.31? Mar 14 10:40:43 NotTooDumb3 why you want to use .31 ? Mar 14 10:41:24 woglinde, i just got that version while following building angstrom from beagle board Mar 14 10:41:38 ? Mar 14 10:41:44 you didnt something wronh Mar 14 10:41:58 for 2008 angstroem the beagle kernel is ompa-psp .32 Mar 14 10:42:08 for 2010 anfgstroem is ins omap .37 Mar 14 10:44:38 anything wrong in this? http://pastebin.com/hP8jDFuW Mar 14 10:44:55 woglinde, sources/openembedded/conf/machine/beagleboard.conf does this file tell me which is going to be used? Mar 14 10:45:21 yes Mar 14 10:45:32 MACHINE=beagleboard bitbake -c clean virtual/kernel Mar 14 10:45:38 MACHINE=beagleboard bitbake -c patch virtual/kernel Mar 14 10:46:59 woglinde, i get this error MACHINE=beagleboard bitbake -c clean virtual/kernel ERROR: Unable to parse conf/bitbake.conf: file conf/bitbake.conf not found in make_angstrom/setup-scripts/ Mar 14 10:47:24 yeah so you didnt something fundamentaly wrong Mar 14 10:47:31 go to the oe manual Mar 14 10:47:34 and read it Mar 14 10:49:26 do i had to run initial bitbatke virtual/kernel in oe directory? Mar 14 10:49:38 ? Mar 14 10:50:53 did i had to run MACHINE=beagleboard ./oebb.sh bitbake virtual/kernel in oe or setup_scripts directory? Mar 14 10:52:36 NotTooDumb3: are you a student or a contractor? Mar 14 10:53:01 i am doing my project work Mar 14 10:53:17 XorA he writes his master thesis or so Mar 14 10:54:02 NotTooDumb3: I would suggest you purchase a copy of Linux for Dummies and work from there, you dont seem to have basic grasp of Linux yet so just arent getting how OE/Angstrom builds or works Mar 14 10:54:03 good luck then :p Mar 14 10:54:13 ynezz, thank you Mar 14 10:54:48 but seems the thesis will fail Mar 14 10:54:56 because he is unwilling to read Mar 14 10:55:05 would be better to ask for another project Mar 14 10:58:03 ynezz, i will learn linux..learning bit by bit with this project Mar 14 11:00:29 it seems that you haven't learned anything in the past five days :p Mar 14 11:02:14 WARNING OE IS A HARSH MISTRESS Mar 14 11:02:54 I wonder what should be easier then to run 3 easy commands Mar 14 11:03:06 ynezz seems he hasnt learned reading the past 10 years Mar 14 11:03:47 NotTooDumb3: its not that we dont like newbies, but you will find probably 99% of the guys here are busy earning money and dont have time for basic linux questions so you need basic linux expertise to build on Mar 14 11:04:00 ynezz, those commands i executed got a uImage which boots too Mar 14 11:04:17 XorA, True Mar 14 11:04:28 amen Mar 14 11:05:47 I still have no clue why bitbake does not find runtime_mapping_rename. i see that it is defined in package.bbclass and called in image.bbclass but I can't figure out why it is not found. It worked till I updated my openembedded git repository Mar 14 11:06:28 Cyberian: out of date bitbake Mar 14 11:06:30 ? Mar 14 11:08:17 bitbake 1.8.18 too old already? Mar 14 11:08:34 ls Mar 14 11:08:39 oops ^^ Mar 14 11:08:39 depends Mar 14 11:09:28 oh wait, that's not my active bitbake Mar 14 11:16:15 woglinde, ~/.oe/environment-2008 what does this do? Mar 14 11:17:01 Cyberian: fixed it? Mar 14 11:18:23 still searching, what was wrong. I followed http://www.headnut.org/2009/08/08/building-oe-workspace but have probably used another bb version than the howto suggests Mar 14 11:18:37 NotTooDumb3: http://docs.openembedded.org/usermanual/usermanual.html Mar 14 11:19:14 NotTooDumb3: you should really take a look. it will answer many questions Mar 14 11:19:22 does bb shows its version somwhow? Mar 14 11:19:26 sure..going through it Mar 14 11:19:34 yes for me Mar 14 11:19:40 Cyberian: bitbake --version? Mar 14 11:20:27 ./bitbake --version Mar 14 11:20:27 Traceback (most recent call last): Mar 14 11:20:28 File "./bitbake", line 35, in Mar 14 11:20:28 import bb Mar 14 11:20:28 ImportError: No module named bb Mar 14 11:21:36 seems like your bitbake installation is hosed somehow. how did you install it? Mar 14 11:22:03 ok figured it out. I have bb 1.12.0 Mar 14 11:22:50 as I said I followed the guide from link posted earlier but manually used a newer bb version Mar 14 11:23:15 usually the makfile gets bitbake via svn from berlios Mar 14 11:25:33 well, i have to go. I'll try again later Mar 14 11:36:37 Hi guys Mar 14 11:37:07 hm did someone update the wiki= Mar 14 11:37:23 to use bitbake 1.12 release Mar 14 12:40:13 Can we use virtual/kernel with PREFERRED_VERSION like PREFERRED_VERSION_virtual/kernel? Mar 14 12:40:26 no Mar 14 12:40:39 Noor: virtual/kernel is virtual Mar 14 12:40:47 PREFERRED_VERSION is for real things Mar 14 12:41:17 PREFERRED_PROVIDER_virtual/kernel = "linux-umbaumba" + PREFERRED_VERSION_umbaumba = "2.6.66" Mar 14 12:42:06 yeah I was having problem so the I moved to PREFERRED_VERSION_linux Mar 14 12:42:09 and it worked Mar 14 12:42:42 so came here just to confirm that we can use virtual/kernel or not Mar 14 12:42:47 Thanks hrw Mar 14 12:47:53 hm whats the bestway now in oe to apply patches only when arch is not x86 or amd64 Mar 14 13:06:12 woglinde: EXTRA_PATCH = "patch.patch" EXTRA_PATCH_amd64 = "" SRC_URI = ${EXTRA_PATCHES}" Mar 14 13:06:32 woglinde: without the 8 million typos Mar 14 13:07:24 xora hm Mar 14 13:07:26 yeah Mar 14 13:07:32 thats sounds reasonable Mar 14 13:07:34 thanks Mar 14 13:08:11 we havent persaded kergoth to do the _! override yet :-D Mar 14 13:52:33 dear guys .. i want to recompile a external kernel module to a different kernel version .. i build the kernel but the module recipe still builds it against the "older" version Mar 14 13:53:41 ? Mar 14 13:53:46 ok Mar 14 13:53:58 external kernel modul is okay Mar 14 13:54:05 we have some recipes in oe Mar 14 13:54:07 i got a 2.6.39 before and build the ti modules for it Mar 14 13:54:15 39, neat Mar 14 13:54:18 what? Mar 14 13:54:19 now i switched to 2.6.35 Mar 14 13:54:25 .39 for beagle? Mar 14 13:54:34 now i switched to 2.6.35 .10 and want to recompile the ti modules Mar 14 13:54:36 there's even no .38 yet Mar 14 13:54:40 no no sorry Mar 14 13:54:42 :-) Mar 14 13:54:55 hm? Mar 14 13:55:04 which ti modules? Mar 14 13:55:07 but the ti modules still get compiled against the 2.6.35.9 Mar 14 13:57:44 ah f** it .. i clean the old kernel Mar 14 13:57:49 then it should work Mar 14 13:57:53 yes Mar 14 13:58:04 didnt you? Mar 14 13:58:09 no Mar 14 13:59:16 i thought the do_setscene will do it for me after i recompile a .10 kernel Mar 14 13:59:41 hm dont count on it Mar 14 13:59:50 roger Mar 14 15:22:55 03Koen Kooi 07org.openembedded.dev * ra96b9ad752 10openembedded.git/recipes/ffmpeg/ffmpeg_git.bb: Mar 14 15:22:55 ffmpeg: switch to libav, that's where the sane people are at Mar 14 15:22:55 Signed-off-by: Koen Kooi Mar 14 15:23:17 :) Mar 14 15:23:40 *g* Mar 14 15:24:49 looks like it's not just Japan shaked by that earthquake :p Mar 14 16:30:49 anyone used a srctree based recipe lately? Mar 14 16:31:46 * kergoth hasn't in a while Mar 14 16:32:51 not working and no good messgae :( Mar 14 16:33:36 http://pastebin.com/92DzhMwd Mar 14 16:33:52 clean fails also Mar 14 16:34:45 http://pastebin.com/dnFUbamP Mar 14 16:35:54 Hi all, I have a silly/newbie question regarding OpenEmbedded version dependency. Lets say I have recipes /foo/foo-1.0.bb foo/foo-1.1.bb foo/foo-1.2.bb and bar/bar-1.0.bb bar/bar-1.1.bb bar/bar-1.2.bb. And I am building a package named foobar. Where in the heck in all of these config and bb files etc does it get specified to build foobar using foo-xx.bb and bar-yy.bb??? 11:32 I see these PREFERRED_VERSION macros…but I cannot seem to Mar 14 16:36:24 that there are many recipes for gtk+. And it relies on many other libraries…all of which have many recipes for different versions. Where, if anywhere, does a person specify that I want gtk-versionX and that it should pull in cairo-versionY and glib-versionZ? Mar 14 16:36:33 "cannot seem to" what? Mar 14 16:36:49 PREFERRED_VERSION_ = Mar 14 16:37:20 which config file its set in is up to you. if you have your own distro, it might make sense to place it there, otherwise local.conf is just fine. bitbake doesn't care, as long as its set in the configuration files somewhere Mar 14 16:37:50 currently version specific dependency via DEPENDS doesn't work, all version preferences have to be specified in the configuration metadata Mar 14 16:38:03 so, a grep for PREFERRED_VERSION_gtk+ should point me to what I'm looking for? Mar 14 16:38:25 or at least a solid working example of this is done Mar 14 16:38:28 correct? Mar 14 16:39:11 03Tom Rini 07master * re787cbcbc3 10openembedded.git/recipes/php/ (php_5.2.13.bb php_5.3.5.bb): Mar 14 16:39:11 php: Add findutils-native to DEPENDS Mar 14 16:39:11 find -readable / -writeable is recent to find. Mar 14 16:39:11 Signed-off-by: Tom Rini Mar 14 16:41:16 @kergoth thanks a ton...I found what I was looking for. I really appreciate it. Mar 14 16:42:15 np Mar 14 16:42:50 ideally, recipes would be able to express things like "i require gtk+ >= M.N", and the like, and it would check the version preferences against those constraints, but we're not there yet Mar 14 16:57:32 kergoth, did you see mmy second pastebin with the srctree clean failure? Mar 14 17:00:28 the /dev/null message is fixed in both 1.12 and master, but not in a release Mar 14 17:00:57 the clean error is a general one, not specific to srctree -- i think its choking when the lockfile dir is missing, or ${S} is, haven't gotten it to be reproducable enough to fix Mar 14 17:02:01 ah, right, i bet its T that doesnt exist due to its removal in the middle of the do_clean -- clean removes WORKDIR, yet running functions writes to T which is in WORKDIR.. Mar 14 17:21:24 hi mickeyl Mar 14 17:21:59 hey woglinde, how are you? Mar 14 17:22:12 hm fine Mar 14 17:22:20 doing java stuff again Mar 14 17:22:21 for oe Mar 14 17:23:28 sounds good Mar 14 17:23:30 family alright? Mar 14 17:23:45 yes family is alright Mar 14 17:27:15 kergoth, anythoughts on how to fix it? Mar 14 17:28:05 Crofton: no, i'd suggest upgrading your bitbake to a version without the dev/null problem so we see more info about the error Mar 14 17:28:30 k Mar 14 17:29:18 hmm, that is not from the 1.12 branch then Mar 14 17:29:34 ? Mar 14 17:29:52 I was running bb from the 1.12 branch Mar 14 17:29:56 ahh Mar 14 17:30:00 right, i guess that hasn't been backported yet :\ Mar 14 17:30:04 thought it had Mar 14 17:30:05 yeah Mar 14 17:30:21 I was only missing the pythong version one Mar 14 17:32:46 hi risca Mar 14 17:32:53 ah Mar 14 17:32:58 Hi Mar 14 17:33:02 srctree may be working Mar 14 17:33:05 except for celan Mar 14 17:33:14 broken clean is still bad ... Mar 14 17:34:38 I don't know if launching xchat is such a great idea when I should be studying... Mar 14 17:35:05 no Mar 14 17:35:09 I can kick you Mar 14 17:35:12 when you want Mar 14 17:35:33 No, that's alright. I needed a break anyway Mar 14 17:35:53 risca whats the subject? Mar 14 17:36:25 I'm writing a review of LTE and LTE-A (commonly knowned as 3G and 4G) Mar 14 17:36:34 ah Mar 14 17:36:47 so you read the full gsm spec Mar 14 17:36:50 haha Mar 14 17:37:17 I'm getting there. Right now I'm summarizing the standard leading up to LTE Mar 14 17:37:25 Wanna read the work so far? Mar 14 17:37:52 not really Mar 14 17:38:07 Hehe I guessed as much. Mar 14 17:38:21 * woglinde wonders when khem will update to qemu 0.14 Mar 14 17:43:01 Crofton: agreed, i'm looking into it. it seems to be somewhat random, haven't narrowed to a particular set of reproduction steps yet Mar 14 17:45:35 thanks Mar 14 17:45:53 it seems ok for a cmake based recipe, but not for an autotools one Mar 14 17:49:33 ericben_: which device do you have that would benefit from the in-kernel 07.10 module? Mar 14 17:55:15 kergoth, looks like the only real problem I have is clean for an autotools based srctree recipe Mar 14 18:00:56 Crofton: weird. okay, i'll try testing that combination. thanks Mar 14 18:15:08 if my build is only creating rootfs tarballs and not filesystem images (the appropriate options are set in local.conf), would it be a build environment issue? Mar 14 18:16:12 Are you sure it's set right in local.conf? Mar 14 18:16:25 bitbake -e whatever-image | grep ^IMAGE_FSTYPES Mar 14 18:16:37 Soopaman1: what do you mean? the tarball is no different from what you consider an "image", they use the *exact* same mechanism. you just aren't configuring it to emit the image type you want Mar 14 18:17:02 that's what i figured Mar 14 18:17:16 ihave jffs2 tar and ext3 set in my local.conf Mar 14 18:17:23 but only the tar is built Mar 14 18:17:57 right Mar 14 18:18:00 so i'm not sure if my dev machine lacks the necessary apt-packages to create those fs Mar 14 18:18:04 so, do what I said and check the output Mar 14 18:18:09 k Mar 14 18:18:11 If it lists jffs2 and such, there's a problem Mar 14 18:18:18 if not, thers's a different problem :) Mar 14 18:26:28 Tartarus: IMAGE_FSTYPES="tar.gz" Mar 14 18:27:06 but in my local.conf, i have IMAGE_FSTYPES = "jffs2 tar ext3" Mar 14 18:27:33 Soopaman1: so a config file that loads after local.conf (DISTRO, MACHINE) is overriding it. use an override. IMAGE_FSTYPES_local = "jffs2 tar ext3" Mar 14 18:28:02 ahhh, probably machine Mar 14 18:28:11 k, thanks Mar 14 18:28:20 yeah most likely Mar 14 18:28:21 np Mar 14 18:28:27 it should probably be using += or ?= Mar 14 18:28:32 but the override will get it working for you Mar 14 18:28:52 i'll probably add the += to be safe Mar 14 18:29:01 until i can track down the actual overide Mar 14 18:29:05 i was referring to the machine, not you Mar 14 18:29:15 use _local in local.conf, or += or ?= in the machine .conf Mar 14 18:29:21 ahhh, k Mar 14 18:29:27 heheh, that would have been ugly Mar 14 18:29:50 if the machine uses ?=, then it only sets it if it wasn't already set. if it used +=, it would just add it to the existing list of ones to emit Mar 14 18:33:32 gm Mar 14 18:34:18 gm Mar 14 18:38:22 kergoth, Tartarus, thanks again for all of the help. Is there a place I can upload a polished version of my OE quickstart script for Ubuntu 10.10? Mar 14 18:50:25 Soopaman1: if it is open for registration again, the OE wiki seems like a fine place (without having seen the script ofc) Mar 14 18:52:57 https://gist.github.com/869616 WARNING: DO NOT use if you aren't a developer, but for devs it may be worth using to save time. i put it in ~/.oe and ensure ~/.oe is in BBPATH everywhere, just after the build dir Mar 14 19:05:47 kergoth: looks nice. For versions I think you can run few more hit and trials like if --version does not work try -v or Mar 14 19:12:01 good idea Mar 14 19:12:35 all this will go into OE right Mar 14 19:12:58 some of that of course is stuff that might not be safe to assume, but i try it until something breaks, which is part of why the average user shouldn't use it. the other part is of course reproducability, having the build change from box to box is not ideal for autobuilds, etc. for a dev's day-to-day use, though, it can save time Mar 14 19:13:34 I could see putting assume.py and assume.bbclass in, but the current assumptions i'm not so sure about, since its fuzzy, don't *know* that its always safe to assume those things Mar 14 19:13:47 its just "working now for the things i've tried to build" Mar 14 19:13:50