**** BEGIN LOGGING AT Tue Jan 27 05:11:47 2009 **** BEGIN LOGGING AT Tue Jan 27 05:27:26 2009 **** BEGIN LOGGING AT Tue Jan 27 05:37:20 2009 Jan 27 07:37:55 morning Jan 27 07:47:47 moin Jan 27 07:56:06 morning Jan 27 08:12:49 03Michael 'Mickey' Lauer 07fso/milestone5 * r408fe9df0c 10openembedded.git/conf/distro/include/fso-autorev.inc: fso-autorev.inc: add correct name of fso-monitor Jan 27 08:12:59 03Michael 'Mickey' Lauer 07fso/milestone5 * rdf756c1061 10openembedded.git/ (2 files in 2 dirs): fso-gsm0710muxd: bump to 0.9.3.1 Jan 27 09:17:43 <_boto> good morning Jan 27 09:30:20 _boto: hi Jan 27 09:44:32 <_boto> how are you koan? Jan 27 09:44:49 <_boto> i love angsrom on beagleboard :-) Jan 27 09:45:20 <_boto> mckoan: you are the distri expert, right? Jan 27 09:46:06 <_boto> trying to build a fresh setup package i ran into a problem Jan 27 09:46:18 <_boto> the file includes Jan 27 09:46:36 <_boto> and linux/delay.h includes which cannot be found Jan 27 09:47:05 <_boto> i searched for the missing delay.h, but i could not find ayn Jan 27 09:47:07 <_boto> -any Jan 27 09:47:22 <_boto> any idea what's going wrong? Jan 27 09:47:45 <_boto> jerm^: thanks again for your help yesterday Jan 27 09:48:06 <_boto> the drivers can be built now :-) Jan 27 09:48:36 <_boto> the next step will be to include them into the kernel Jan 27 09:49:37 <_boto> DjWillis had suggested to use the defconf patching mechanism Jan 27 09:49:44 <_boto> so i will try that Jan 27 09:51:03 <_boto> but first i have to figure out what is going on with missing header Jan 27 09:57:41 _boto: I'm newbie, not expert :-) Jan 27 09:58:25 <_boto> mckoan: then welcome to the club of noobs ;-) Jan 27 09:58:37 <_boto> i am a newbie too Jan 27 09:59:11 _boto: the file includes in into your kernel Jan 27 10:04:30 <_boto> you are right Jan 27 10:04:34 <_boto> i found it, thanks Jan 27 10:04:45 <_boto> what i searched was asm/delay Jan 27 10:04:56 <_boto> it is in arch/arm folder Jan 27 10:05:18 <_boto> i have set the proper -I in compiler options Jan 27 10:15:00 03Stefan Schmidt 07fso/milestone5 * r3f298b0d33 10openembedded.git/ (58 files in 30 dirs): Jan 27 10:15:00 Revert "Merge commit 'origin/john_lee/fastboot-devel' into org.openembedded.dev" Jan 27 10:15:00 This reverts commit 366bf6043abfbbbfa98168548f42aa9aaf6c28b9. Jan 27 10:15:00 This did broke all image for !om-gta02. Reverted and PRs bumped. Jan 27 10:41:59 Anyone built Unionfs-modules from GIT recently? Jan 27 10:52:26 what is the best way to add a package to an image? Jan 27 10:52:38 if I want to add grub to x11-image Jan 27 10:53:07 is better to clone x11.image.bb and customize or working on original one? Jan 27 10:53:39 maybe there is the possibility to add it from command line? Jan 27 10:54:42 mckoan: I would add it to a device description. Jan 27 10:55:11 mckoan: Its a bootloader and these depend on the device... Jan 27 10:57:35 florian: do you mean a machine? Jan 27 11:04:48 mckoan: yes Jan 27 11:05:32 mckoan: In this way you do not need to change the image description but get what you need to boot your device. Jan 27 11:06:22 florian: and generally speaking of a package which is not directly related to a machine, what is the best way to add a package to an image? Jan 27 11:07:01 IMAGE_INSTALL into /packages/images/* ? Jan 27 11:07:38 mickey|zzZZzz: FWIW I may well be going to FOSDEM now Jan 27 11:08:28 mckoan: I would create a derived image description. It might be a good idea to create a task package as well if you want to change more than just on package. Jan 27 11:10:39 mickey|zzZZzz: thanks for adding the coherence recipe; it builds nicely for me (haven't tested it yet) Jan 27 11:17:32 morning #oe Jan 27 11:18:21 anybody played with grub trying to get an image with /boot/grub ? Jan 27 11:18:34 i'm trying (and failing) to make bitbake include a .bb file from inside the sources it has just checked out Jan 27 11:19:26 morning Jan 27 11:20:03 mckoan: do you mean your grub package isn't installing to the right directory? Jan 27 11:20:43 mattface: grub_0.97 isn't installing at all Jan 27 11:20:48 ah Jan 27 11:21:28 I added it, and I expected to see its files into /boot/grub in the target image, but there is nothing Jan 27 11:21:33 mckoan: grub --root-directory=/a/place/where/my/boot/is Jan 27 11:22:23 03Stefan Schmidt 07fso/milestone5 * r6d3445928a 10openembedded.git/ (2 files in 2 dirs): angstrom.inc: Catch up with openmoko kernel recipe renames. Jan 27 11:22:37 hrw: that is the installation from command line Jan 27 11:23:01 I expected that grub recipe did that Jan 27 11:23:24 mckoan: best to give the bb file a read Jan 27 11:24:00 I did and I am going to do it again later :-) thx Jan 27 11:26:41 using a fresh git & bitbake svn, I get that when I try a "bitbake nano" to try the config (smdk2440) : Error, Your PACKAGE_ARCHS field contains duplicates. Perhaps you set EXTRA_PACKAGE_ARCHS twice accidently through some tune file Jan 27 11:27:11 anyone has a clue where to look ? I followed the FAQ up to there Jan 27 11:30:28 hi mattface Jan 27 11:30:40 BusError: remove EXTRA_PACKAGE_ARCHS from machine config Jan 27 11:30:45 hi florian Jan 27 11:31:03 hi pb_ Jan 27 11:31:06 hi hrw Jan 27 11:31:32 morning pb_ Jan 27 11:33:39 hrw ahah that worked :D Jan 27 11:35:24 pb_: we've been considering keeping a bb file inside the sources for each package, so we dont end up with hundreds of bb recipes in everyones copy of oe Jan 27 11:36:30 mattface: doesn't that give you a bit of a bootstrapping problem? Jan 27 11:36:47 if the .bb file is in the sources then you can't build it unless you've already fetched the source code Jan 27 11:40:00 indeed, i've been toying with some kind of wrapper recipe to fetch the source and then include the bb inside it, but bb doesn't seem to like including on-the-fly Jan 27 11:40:52 mm. to be honest, it sounds to me like you are trying to solve a non-problem. Jan 27 11:41:39 if you're worried about the proliferation of .bb files for different versions of the same package then you could factor those out: have a single .bb file and set the version elsewhere (in some mattface-versions.conf for example). Jan 27 11:43:28 or just order some more disks and live with the extra files :-} Jan 27 11:43:44 :) Jan 27 11:45:06 ok, time to push one change which will fix some stuff and can make me new enemys Jan 27 11:47:47 03Marcin Juszkiewicz 07org.openembedded.dev * r950e06a7db 10openembedded.git/conf/machine/ (60 files in 2 dirs): Jan 27 11:47:47 conf/machine: EXTRA_PACKAGES_ARCHS cleanup Jan 27 11:47:47 All duplicates in EXTRA_PACKAGE_ARCHS are removed. Jan 27 11:47:47 Moved EXTRA_PACKAGES_ARCHS to include/tune-* files (PPC and x86 Jan 27 11:47:47 targets), dropped it for ARM targets (as tune-arm* files have them Jan 27 11:47:50 defined properly). Jan 27 11:53:50 Dear OE devs. May I ask, why did you include bluetooth into the task-base? Jan 27 11:54:37 In absence of any additional hints, it pulls x11, gtk, gstramers and hell knows what else with it. Jan 27 12:00:37 xl0: you'd need to take that up with your DISTRO maintainers. task packages are a question of distro policy. Jan 27 12:01:13 xl0: your machine has bt or usb or pci in MACHINE_FEATURES? Jan 27 12:03:38 ah, I see why Jan 27 12:04:14 hmm qemu-native faisl to build. there is no error in the log, it seems to be going to the link part, but the build is reported as failed... Jan 27 12:04:35 no visible error that I can see there, or in the log Jan 27 12:05:49 hrw: pci. Jan 27 12:06:07 pb_: I've created a distro based on minimal.conf. Jan 27 12:06:52 And bluetooth is inconditionalyl enabled in task-base.bb. Jan 27 12:06:55 pb_: the problem is with task-base nevermind of distro Jan 27 12:07:15 bt/wifi are enabled always and bluez req alsa, gstreamer... Jan 27 12:07:17 If I get the package format correctly. Jan 27 12:07:43 xl0: if you are your own distro maintainer then you should feel free to bash the task packages any way you want. Jan 27 12:08:34 every distro will have different requirements there (and some might not use task-*.bb at all). the stuff in packages/task/* is probably best regarded just as an example. Jan 27 12:09:13 So I should define my own task-base rule? Jan 27 12:09:28 Still I'd expect to see a sane default one. Jan 27 12:09:59 Not pulling bluetooth or wifi stuff into _base_. Jan 27 12:10:57 I'd expect to define my own requirements, and then just be able to include the default task-base for the reall base stuff that everyone needs. Jan 27 12:12:23 xl0: use task-base isntead of task-base in images so you will not get bt/wifi Jan 27 12:12:33 xl0: but they still will get built Jan 27 12:13:08 Is there a reason to do so? Jan 27 12:13:55 For myself, I have just removed the thing from task-base.bb Jan 27 12:15:27 xl0: thats history Jan 27 12:15:48 I think that better way can be done for it Jan 27 12:18:43 xl0: well, the problem with that in general is that there is very little consensus on what would be the "real base stuff". Jan 27 12:19:01 more or less any non-trivial task base that you could think of will be inapplicable for at least somebody. Jan 27 12:19:42 admittedly, including bluetooth in it does seem fairly perverse, but if you assume that every distro is probably going to have to create its own task packages anyway then to be honest it doesn't matter much what is in the "default" ones. Jan 27 12:20:05 X11, gtk+ and gstreamer are not "real base stuff" and only a moron could argue this. Jan 27 12:20:42 sure, it is easy to point at what is not "real base stuff", but it is much harder to produce an intelligent list of packages that are, in fact, "real base". Jan 27 12:21:49 though, for that matter, there are plenty of applications where the three packages that you mentioned probably would be considered "base". if you are building a graphical system then you won't get far without X11. Jan 27 12:21:51 May we just apply this patch then? Jan 27 12:21:52 http://dumpz.org/5075/ Jan 27 12:22:27 in your own tree? sure, you can do what you like. Jan 27 12:22:49 pb_: Are you an OE developer? Jan 27 12:23:54 depends what exactly you mean. I'm not a core team member. Jan 27 12:24:07 Good. Jan 27 12:25:18 03Stefan Schmidt 07fso/milestone5 * r7e59b710b1 10openembedded.git/packages/images/fso-image.bb: fso-image: Fix ll alias Jan 27 12:30:45 <_boto> what do i have to do in order to build the angstrom X11 image against kernel version 2.6.18 ? Jan 27 12:31:15 <_boto> i am searching for the proper bb, but no success until now Jan 27 12:31:33 <_boto> or do i have to change a config? Jan 27 12:41:29 _boto: did I not read that your a Beagle user, why would you want such an ancient kernel? Jan 27 12:43:05 http://pl.youtube.com/watch?v=gR8ISMJdPFg Jan 27 12:52:14 probably a MV kernel distributed with a dev baord :) Jan 27 13:13:49 nick eFfeM Jan 27 13:14:14 03Philip Balister 07org.openembedded.dev * rbd10e25779 10openembedded.git/packages/gnuradio/gnuradio_svn.bb: gnuradio : Add dependency on gsl (GNU Scientific Library). Jan 27 13:14:19 mattface: openembedded/packages/grub/grub_0.97.bb doesn't do any install Jan 27 13:21:51 mckoan: it does install Jan 27 13:22:16 mckoan: the one in classes/autotools.bbclass Jan 27 13:24:17 hrw: I missed classes/* sorry Jan 27 13:24:46 BTW why my image does not contain grub files? Jan 27 13:24:49 hi, is it possible for oe to build a image which contains a kernel and an initrd? Jan 27 13:25:30 <_boto> i have dma access problems: "DMA misaligned error with device 0" Jan 27 13:25:53 <_boto> it seems to be the reason for kernel panics Jan 27 13:26:13 <_boto> it may have to do with the usb interface Jan 27 13:26:28 <_boto> i have read that there is a patch solving this problem Jan 27 13:26:38 <_boto> any idea which patch it is? Jan 27 13:27:30 mckoan: does your image recipe install grub? or does your machine has it in MACHINE_EXTRA_RDEPENDS/RRECOMMENDS? Jan 27 13:27:33 <_boto> another question: when will kernel 2.6.18 be available for beagleboard? Jan 27 13:27:49 mckoan: did you rebuilt task-boot/task-base after changing machine config? Jan 27 13:27:58 _boto: someone use 2.6.18 on bb?? Jan 27 13:28:09 <_boto> aha Jan 27 13:28:32 * balister_ hearts git-stash Jan 27 13:28:32 <_boto> hrw: you mean i can already use it building angstrom x11 image? Jan 27 13:28:40 _boto: beagleboard is evolving fast - I would not try to run older kernel then OE latest for it Jan 27 13:29:00 _boto: bb is one of nicer supported devices in OE iirc Jan 27 13:29:03 <_boto> hrw: older? i need the newest ;-) Jan 27 13:29:06 * hrw wants bb revC Jan 27 13:29:11 s/older/newest Jan 27 13:29:13 sorry Jan 27 13:29:28 <_boto> i have b7 Jan 27 13:29:43 <_boto> ah ok Jan 27 13:30:03 <_boto> you mean i should wait until 2.6.18 is officially supported for bb? Jan 27 13:30:20 _boto: 2.6.18 or 2.6.28? Jan 27 13:30:32 2.6.18 is year old? or even more... Jan 27 13:30:39 <_boto> it was a typo, i was meaning x.28 Jan 27 13:30:48 much more. Jan 27 13:30:52 <_boto> hehe Jan 27 13:31:17 <_boto> ... i was a little child in that days, lol Jan 27 13:31:42 <_boto> well, the crashes make me the life really hard atm Jan 27 13:32:02 <_boto> i have to power cycle bb every 10 minutes Jan 27 13:32:10 <_boto> i want to get rid of if Jan 27 13:32:33 <_boto> someone mentioned that the newest angrom image contains already the usb patch Jan 27 13:32:49 <_boto> so i wonder how i get my git to check for updates on all image packages Jan 27 13:56:12 hrw: I'm rebuilding task-base Jan 27 13:57:21 _boto: you just need to update OE and do build Jan 27 13:58:52 _boto: sorry I havent gotten back to you yet! Jan 27 13:59:01 _boto: I'm testing your package right now... Jan 27 14:05:56 ERROR: IO Error: file 'conf/bitbake.conf' not found - what's the matter with it? Jan 27 14:07:12 booxter: that's pretty serious, have you set your OEROOT and BBPATH environment? Jan 27 14:07:43 OEROOT? no I don't. I try to follow GettingStarted guide Jan 27 14:07:55 there is no info on OEROOT Jan 27 14:08:01 ignore OEROOT Jan 27 14:08:02 is devshell working ? Jan 27 14:08:05 It fails here Jan 27 14:08:06 booxter: your BBPATH is wrong Jan 27 14:08:32 yeah, do what he says ^ :) Jan 27 14:08:51 echo $BBPATH Jan 27 14:08:52 /home/booxter/winbox/oe/build:/home/booxter/winbox/oe/openembedded Jan 27 14:09:31 that's what GettingStarted doc recommends Jan 27 14:09:32 _boto: just sent you an email...let me know if you have any questions... Jan 27 14:10:10 booxter: and does /home/booxter/winbox/oe/build/conf/bitbake.conf exist? Jan 27 14:10:39 mattface: no Jan 27 14:11:06 should I create it? what's the contents? GettingStarted doc doesn't mention it in any way Jan 27 14:11:17 you need to find it Jan 27 14:11:25 <_boto> jerm^, thanks, checking my mails now ... Jan 27 14:12:01 should I copy ./bitbake/conf/bitbake.conf into ./build/conf/bitbake.conf? Jan 27 14:12:33 no add that bitbake directory to your BBPATH Jan 27 14:14:05 03Otavio Salvador 07org.openembedded.dev * rafdc8d908b 10openembedded.git/ (conf/checksums.ini packages/atftp/atftp_0.7.bb): atftp: use 0.7-6 Debian diff since 0.7-3 has been droped Jan 27 14:15:41 booxter: NO and NOI Jan 27 14:15:42 booxter: NO and NO Jan 27 14:15:50 it seems that bitbake tries to get the file from: ./build Jan 27 14:15:57 and /usr/share/... Jan 27 14:16:00 BBPATH=path-to-build-dir:path-to-oe Jan 27 14:16:01 and then it fails Jan 27 14:16:51 hrw: if not, what should I do? Jan 27 14:22:36 bitbake should use bbpath always Jan 27 14:23:14 >? Jan 27 14:27:19 <_boto> jerm^, thank you for the recipe Jan 27 14:27:25 <_boto> it works fine Jan 27 14:31:20 hrw: so what? it should use, yes. But it doesn't :) Jan 27 14:32:21 <_boto> when i fetch the newest state out of git, should i clean before building in bitbake? Jan 27 14:39:56 _boto: should not be necessary...revision bumps take care of rebuilds Jan 27 14:42:28 <_boto> ok Jan 27 14:43:11 <_boto> i had observed that if modify the makefile (usb driver) then try to build it then the new makefile was not copied into tmp folder Jan 27 14:43:20 <_boto> i had to clean and then build Jan 27 14:45:13 <_boto> i hope that after the update the crashes disappear Jan 27 14:45:41 <_boto> as otherwise i cannot test the new usb drivers Jan 27 14:52:04 _bobo: either bump the PR in the recipe or bitbake -crebuild package Jan 27 14:52:22 _boto: bitbake only understands .bb recipes, it does not monitor timestamps on any other files Jan 27 14:52:25 <_boto> ah, good hint Jan 27 15:01:57 03Jan Luebbe 07fso/milestone5 * r5dee5a7b3d 10openembedded.git/packages/ (8 files in 8 dirs): Make PR increase over the revert in 3f298b0d3342d1fbbbc1830528698fb4dd640d34 Jan 27 15:05:26 Question, can anybody tell me what I am doing wrong with http://rafb.net/p/FXLcZo43.html. Been meaning to sort this for a while and I can't see what's wrong. Builds an IPK that installs fine but won't populate the ROOTFS when built as a machine depenency (or task for that matter). The do_stage stuff in there is just something I was hacking with trying to make sure my understanding was correct. It really should be dead simple Jan 27 15:05:26 , pick up the binaries and drop them into /lib/firmware on the resulting image. A working installable IPK is a bonus in this case ;-). Jan 27 15:07:23 <_boto> DjWillis, you did include this package into your image? Jan 27 15:07:51 03Jan Lübbe 07fso/milestone5 * r03d2c4cbbd 10openembedded.git/packages/netbase/ (3 files in 3 dirs): Jan 27 15:07:51 netbase: for om-gta01/02 use allow-hotplug instead of auto Jan 27 15:07:51 This fixes the case where usb net is loaded after the network is brought up. Jan 27 15:08:38 _boto: yep ;-). Jan 27 15:09:05 <_boto> DJWillis: how about trying ROOTFS_POSTPROCESS_COMMAND ? Jan 27 15:10:03 _boto: MACHINE_EXTRA_RDEPENDS = " omap3-sgx-modules kernel-modules pandora-firmware pandora-wifi pandora-wifi-tools " Jan 27 15:10:03 in the machine file, also tried puting it into one of my tasks that my images use. I know it is something really silly but it's driving me nuts ;-). Jan 27 15:10:41 DJWillis: style note, please use RRECOMMANDS_${PN} Jan 27 15:10:43 <_boto> hehe, i know this feeling Jan 27 15:10:54 <_boto> had enough such feelings last days :-) Jan 27 15:10:56 _boto: I would think postprocess would be a bit of a crude hack around ;-) Jan 27 15:11:27 <_boto> aha, i have seen that in some packages and used this mechanism to customize the rootfs Jan 27 15:12:14 XorA: Noted Jan 27 15:12:32 DJWillis: otherwise -dev -dev -dbg all RRECOMEND as well Jan 27 15:12:47 -doc -dev -dbg Jan 27 15:12:57 why does locally fetched bitbake tries to access config in /usr/share...? Jan 27 15:13:02 XorA: ahhh, now that makes sense ;-) Jan 27 15:13:04 strace shows: access("/usr/share/bitbake/conf/bitbake.conf", R_OK) = -1 ENOENT (No such file or directory) Jan 27 15:13:32 DJWillis: is it just that you need to rebuild the task that uses MACHINE_EXTRA_RDEPENDS? Jan 27 15:13:47 booxter: looks like you have a distro level bitbake installed? Jan 27 15:14:07 DJWillis: sounds like the .ipk for the task doesnt know the new depends because it didnt get rebuilt since you changed them Jan 27 15:14:14 DJWillis: The program 'bitbake' is currently not installed. You can install it by typing: Jan 27 15:14:15 sudo apt-get install bitbake Jan 27 15:14:28 XorA: that is what I thought but several -crebuilds later it's still not playing ;-) Jan 27 15:14:49 DJWillis: check the dependency actually in the .ipk using dpkg-dev Jan 27 15:14:52 dpkg-deb Jan 27 15:15:06 I think its dpkg-deb -I blah.ipk Jan 27 15:16:04 XorA: will do, also, is the do_stage really needed in this case as the do install will still popualte the rootfs will it not? Jan 27 15:16:05 <_boto> DJWillis: sorry, my line was droped Jan 27 15:17:32 DJWillis: you dont need do_stage Jan 27 15:17:51 DJWillis: you can also shorten the FILE_${PN} to /lib/firmware if you feel like Jan 27 15:18:00 and dont need those mkdir -p Jan 27 15:18:30 XorA: still learning ;-) Jan 27 15:19:51 install -d and mkdir -p are pretty much the same operation :-) Jan 27 15:22:13 XorA: I thought as much but saw a lot of other BB's that did both so figured there was a reason ;-). http://rafb.net/p/UK9Y0U33.html revised one, style comments apprecated. Jan 27 15:24:43 RRECOMMANDS == RRECOMMENDS Jan 27 15:25:08 you want to run it thorugh oe-stylize though, to get the fields in the right order Jan 27 15:25:38 hm, I tried to use OE with system-wide bitbake. Now I get: ERROR: Please set the 'CACHE' variable. Jan 27 15:25:41 wtf? Jan 27 15:26:12 XorA: oe-stylize? hmmm, now there is something I did not know existed. Handy Jan 27 15:26:29 DJWillis: very handy :-) Jan 27 15:26:50 DJWillis: means I can write initial .bb as a total mess then clean it later Jan 27 15:27:05 XorA: now I have to go and process the complete local overlay Jan 27 15:28:45 DJWillis: sorry :-D Jan 27 15:29:17 just to jump in here, oe-stylize? Jan 27 15:29:25 is it a package? Jan 27 15:29:36 mattface: in contrib directory, tidies up .bb files Jan 27 15:29:56 aha, thx Jan 27 15:31:11 pipe bb files through it, gotcha Jan 27 15:31:17 very nifty Jan 27 15:34:59 XorA: daft but what did you mean by RRECOMMANDS == RRECOMMENDS? Just that? Jan 27 15:35:58 typo Jan 27 15:37:17 Doh!, told you I am having a bad day, I did not even spot that when you prompted ;-). Jan 27 15:43:01 DJWillis: its obviously beer o clock Jan 27 15:51:24 XorA: I wish, gives me an idea, drive home and go to pub and just hope problems go away ;-) Jan 27 15:53:44 sounds like a wise plan Jan 27 16:14:53 mickeyl: good morning Jan 27 16:14:59 good morning pb_ Jan 27 16:18:36 hi mickeyl Jan 27 16:18:54 g'day kergoth Jan 27 16:19:18 yo hrw Jan 27 16:19:22 how's family? Jan 27 16:19:27 fine, thx Jan 27 16:19:59 hey Jan 27 16:21:33 I did it git pull --rebase and got a merge conflict in checksums.ini Jan 27 16:21:40 how do I resolve this? Jan 27 16:21:42 happens Jan 27 16:21:45 normally Jan 27 16:22:04 it says the merge failed Jan 27 16:24:36 git-mergetool Jan 27 16:24:44 git rebase --continue Jan 27 16:26:50 I'm having trouble finding the problem Jan 27 16:26:57 checksums.ini is the issue Jan 27 16:27:14 my git fires up meld to do the merges Jan 27 16:27:29 I ended up with emerge Jan 27 16:27:57 ah Jan 27 16:28:16 wiggle ftw! Jan 27 16:28:36 anyway I have to go Jan 27 16:47:09 03Philip Balister 07org.openembedded.dev * r4a581fc991 10openembedded.git/packages/gsl/ (gsl.inc gsl_1.10.bb): gsl : Convert recipe to use include file. Jan 27 16:47:18 cross your fingers that commit is ok :) Jan 27 16:47:19 03Philip Balister 07org.openembedded.dev * rdaf60a6a27 10openembedded.git/ (conf/checksums.ini packages/gsl/gsl_1.12.bb): gsl : Add recipe for gsl-1.12. Jan 27 17:07:39 bye Jan 27 17:12:22 hi all Jan 27 17:52:18 03ghost 07org.openembedded.dreambox * r8251747cad 10openembedded.git/packages/libdreamdvd/libdreamdvd.bb: libdreamdvd.bb: update SRCDATE, enable hardware lpcm decode for dm800/dm8000 Jan 27 18:23:16 this might not be an issue in newer version of OE/bitbake, but I typically work with gumstix an a recurring issue is having the kernel freeze at 'freeing init memory' which 99% of the time means the build machine has /bin/sh linked to dash (ubuntu does this) Jan 27 18:23:35 this patch adds a check to sanity.bbclass to check for that http://pastebin.com/m37c47d8a Jan 27 18:23:46 how can i select which sf mirror is used? currently i get puzzle.dl.sourceforge.net, but that is not too responsive Jan 27 18:24:01 I realize this isn't against the main tree but mabye it can be useful? Jan 27 18:34:03 eFfeM: seen my commit? Jan 27 18:35:48 mickeyl: yes, thanks alot Jan 27 18:36:09 haven't had time to try it yet though, coherence does not seem to like my clients Jan 27 18:40:39 somehow I messed up, did a get push, ended up with a merge conflict, maybe resolved it improper due to lack of knowledge, did not push my changes, and want to undo my commit locally. what is the magic incantation for that Jan 27 18:40:52 s/get/git/ Jan 27 18:44:14 check the manpage for git reset Jan 27 18:44:28 in particular, git reset --hard will yank your changes Jan 27 18:47:34 03Michael 'Mickey' Lauer 07org.openembedded.dev * r45e86d87f7 10openembedded.git/packages/tasks/task-openmoko-feed.bb: task-openmoko-feed: s/tichy/paroli/ Jan 27 18:52:45 03Cliff Brake 07org.openembedded.dev * rc1b8471c36 10openembedded.git/classes/module.bbclass: Jan 27 18:52:45 module.bbclass: add common make options to module_do_install() Jan 27 18:52:45 the problem was that KERNELDIR was not getting passed to Jan 27 18:52:45 "make install", so the host kernel src dir got used, and Jan 27 18:52:45 the wrong kernel version was then extracted. Jan 27 18:52:47 03Cliff Brake 07org.openembedded.dev * rfed025d090 10openembedded.git/packages/ (12 files in 10 dirs): Jan 27 18:52:52 misc kernel module recipes: add note suggesting not using kernel source from work dir Jan 27 18:52:54 try to discourage people from using KERNEL_SOURCE when writing Jan 27 18:52:56 kernel module recipes. KERNEL_SOURCE is located in the work dir Jan 27 18:52:58 and is not available if rm_work class is used. Jan 27 18:56:53 mickeyl: thanks Jan 27 18:58:30 mickeyl: git reset --hard still leaves my commit at the top of the log Jan 27 18:58:54 and git pull gives me a merge conflict :-( Jan 27 18:59:23 git reset --hard HEAD^ ? Jan 27 18:59:54 Tartarus: tnx that one seems to do it Jan 27 19:02:23 03Frans Meulenbroeks 07org.openembedded.dev * rca1ecb8465 10openembedded.git/conf/checksums.ini: checksums.ini: removed some junk at the end Jan 27 19:08:28 heh Jan 27 19:08:47 obviously, my git-mergetool skillz are not l33t enough Jan 27 19:21:25 re Jan 27 19:26:07 * * OE Bug 5000 has been created by roman(AT)khimov.ru Jan 27 19:26:09 * * glibc 2.7 fails to build Jan 27 19:26:11 * * http://bugs.openembedded.net/show_bug.cgi?id=5000 Jan 27 19:27:43 Crofton|work: extract your patches with "git format-patch origin/org.openembedded.dev", then checkout clean .dev, update it to latest and use 'git am' to apply your changes? Jan 27 19:28:23 hrw, I had updated checksums.ini and in between my update and commit, another checksum update occured Jan 27 19:28:51 I had to merge, which almost went ok, except I left some cruft at the end of checksum.ini Jan 27 19:28:52 but Jan 27 19:29:09 at least I added the checksum with the recipe this time :) Jan 27 19:29:13 Crofton|work: next time fix such cruft and use 'git commit --amend'? Jan 27 19:29:18 heh Jan 27 19:29:25 someone noticed b4 me :) Jan 27 19:30:26 03Frans Meulenbroeks 07org.openembedded.dev * r55475f550a 10openembedded.git/ (3 files in 2 dirs): gphoto2: updated to 2.4.4 Jan 27 19:31:32 03Marcin Juszkiewicz 07org.openembedded.dev * rc5ffbe60b4 10openembedded.git/conf/checksums.ini: checksums.ini: added 2.6.25.20 patch Jan 27 19:31:32 03Frans Meulenbroeks 07org.openembedded.dev * r55475f550a 10openembedded.git/ (3 files in 2 dirs): gphoto2: updated to 2.4.4 Jan 27 19:34:58 03Frans Meulenbroeks 07org.openembedded.dev * r6d9b4cb0e7 10openembedded.git/ (conf/checksums.ini packages/libexif/libexif_0.6.17.bb): libexif: updated to 0.6.17 Jan 27 19:34:59 03Frans Meulenbroeks 07org.openembedded.dev * rfa4cbea0e1 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Jan 27 19:39:23 mickeyl, http://www.fosdem.org/2009/free-bus-train-station Jan 27 19:40:23 nice Jan 27 19:40:26 and very needed :) Jan 27 19:41:01 you had mentioned you planned to leave Sunday Jan 27 19:43:15 i'm afraid i won't be there at all Jan 27 19:43:31 with all the problems getting a meeting going, i changed plans Jan 27 19:43:37 bummer Jan 27 19:44:00 I have an extra day to be a tourist I guess Jan 27 19:44:08 drink some more Leffe Bruine Jan 27 19:44:19 will do Jan 27 19:49:42 if I want to add python to console-image, how do I do that? I tried to add ANGSTROM_EXTRA_DEPENDS = "python" to my local.conf but nothing got added to the rootfs. Jan 27 19:50:00 add in your task.bb Jan 27 19:51:45 woglinde: the console-image.bb in images/ ? Jan 27 19:51:52 nope Jan 27 19:52:06 the task- in packages/task Jan 27 19:52:17 hm or was it in the image Jan 27 19:52:21 hm right sorry Jan 27 19:52:23 in the image Jan 27 19:52:55 woglinde: add to DEPENDS or ? Jan 27 19:53:28 IMAGE_INSTALL Jan 27 19:53:32 is in your friend Jan 27 19:54:05 ~curse angstrom_* variables in metadata Jan 27 19:54:06 May you be reincarnated as a Windows XP administrator, angstrom_* variables in metadata ! Jan 27 19:54:30 heh Jan 27 19:54:45 woglinde: ok, and how do i find the package name or is it just 'python'? Jan 27 19:54:46 I keep meaning to change it to IMAGE_EXTRA_INSTALL .... Jan 27 19:55:59 hi woglinde Jan 27 19:56:27 pb__: hey Jan 27 19:56:33 hi khem Jan 27 19:56:43 pb__: regarding the alsa-utils Jan 27 19:57:12 I thought that uclibc provided the strl* functions only when __USE_BSD is set Jan 27 19:57:20 so it will be a complex check Jan 27 19:57:30 hi pb and hi khem Jan 27 19:57:44 hm maybe alsa should switch to another function Jan 27 19:57:46 thats why I chose to expose the definitions in alsa-utils Jan 27 19:58:22 khem: can't you just add -D_BSD_SOURCE to the alsa-utils cppflags in that case? Jan 27 19:58:22 and this check will become uglier if dietlibc or some such comes into picture Jan 27 19:58:30 kehm is strl for stringlength? Jan 27 19:58:35 or whatever the appropriate rune is to get __USE_BSD defined with uclibc Jan 27 19:59:14 pb__: we could but if that option is not set during uclibc kconfig then these functions will be absent in libc.so Jan 27 19:59:23 and hence linking will fail for alsautils Jan 27 19:59:33 oh, I see. crazy uclibc. Jan 27 20:00:04 yeah configurability comes at a price Jan 27 20:00:42 in that case, yeah, short of writing an actual autoconf test to see whether the uclibc dice have landed in a position that means those functions exist, I guess your patch is probably the best thing to do. Jan 27 20:01:13 pb__ hm *g* I bet there is already one Jan 27 20:01:23 I guess that might be too much like hard work for the sake of the couple of hundred bytes. Jan 27 20:01:38 woglinde: ah yes, possibly. Jan 27 20:01:45 ok, so i did bitbake console-image -crebuild but it didn't add anything, how do I rebuild it so that it adds the new package? Jan 27 20:01:58 I can write a autoconf test too let me try that today after work Jan 27 20:02:03 but the .m4 macro should be fairly easy Jan 27 20:02:32 yeah, I guess it's basically just AC_CHECK_FUNC(whatever) Jan 27 20:02:39 *g* right Jan 27 20:02:48 yeah should be easy IMO Jan 27 20:03:05 bbiab Jan 27 20:03:09 * pb__ heads off to hammer some nails Jan 27 20:03:13 hehe Jan 27 20:03:15 have fun Jan 27 20:03:45 woglinde: did you try 4.3.3 yet ? Jan 27 20:03:53 nope Jan 27 20:04:04 are there some backports from 4.4 Jan 27 20:04:06 ? Jan 27 20:04:11 like ebai for armv4? Jan 27 20:04:18 or the thumb fixes= Jan 27 20:04:23 eabi Jan 27 20:05:10 not yet Jan 27 20:05:25 but there are some quirks Jan 27 20:05:33 actually I think we should wait till 4.4. Jan 27 20:05:43 hm what did they fixing? Jan 27 20:05:44 hm okay Jan 27 20:05:45 and jump straight Jan 27 20:05:50 will read the relase note myself Jan 27 20:06:21 woglinde: http://gcc.gnu.org/bugzilla/buglist.cgi?bug_status=RESOLVED&resolution=FIXED&target_milestone=4.3.3 Jan 27 20:07:52 hm Jan 27 20:07:54 ICE with member function in inline asm Jan 27 20:10:35 03Koen Kooi 07org.openembedded.dev * r9cf4a577a0 10openembedded.git/packages/linux/linux-omap_2.6.28.bb: linux-omap: make 2.6.28 the default for beagle, this makes both EHCI and OTG work Jan 27 20:10:38 03Koen Kooi 07org.openembedded.dev * rf0d2a9c5f9 10openembedded.git/: Merge branch 'org.openembedded.dev' of git@git.openembedded.net:openembedded into org.openembedded.dev Jan 27 20:24:35 morning pplz :) Jan 27 20:40:22 hmmmm Jan 27 20:40:25 # cdparanoia OE build file Jan 27 20:40:25 # Copyright (C) 2005, Advanced Micro Devices, Inc. All Rights Reserved Jan 27 20:40:25 # Released under the MIT license (see packages/COPYING) Jan 27 20:40:25 PR ="r0" Jan 27 20:40:25 LICENSE="GPL" Jan 27 20:40:55