**** BEGIN LOGGING AT Sun Jan 24 02:59:57 2010 Jan 24 10:42:43 Laibsch: hi, yes I have R/W acces, why? Jan 24 10:46:19 18:19:46 < Laibsch> mckoan|away: I think you have RW now, don't you? Jan 24 10:46:20 18:20:00 < Laibsch> There is still at least one patch in patchwork (something ronetix-related) Jan 24 10:47:00 ynezz: thx :_) Jan 24 10:47:33 ynezz, Laibsch: do you know patch number please? Jan 24 10:48:46 koen kooi in irc ? Jan 24 10:49:55 mlip: not in this channel, but I think he hangs out in the beagleboard channels sometimes Jan 24 10:50:19 pb__: hi Jan 24 10:50:29 hi mckoan Jan 24 10:51:23 thx, actually I just wanted to ask how to deal with a program that can have multiple (different) hardware (nodes) to be controlled Jan 24 10:55:56 pb__: hey there Jan 24 10:58:13 hi ant__ Jan 24 10:58:59 have you discovered that sat/sunday are 'daddy-days' ? Jan 24 10:59:17 ant__: LOL Jan 24 10:59:51 pb__: morning and ping Jan 24 10:59:55 Crofton: morning and ping Jan 24 11:00:25 zecke: hail zecke Jan 24 11:02:15 ant__: heh. yeah, often, though in fact Emma is going to her grandparent's house today. Jan 24 11:02:51 pb__: http://patchwork.openembedded.org/patch/691/, we still have this patch in the queue. Do you plan to update it? Jan 24 11:03:45 mickeyl: there are a couple of things I noticed about 'minimal' distro: Jan 24 11:04:02 morning Jan 24 11:04:09 1) needs PREFERRED_PROVIDER_virtual/psplash ?= "psplash" Jan 24 11:04:29 2) why not NHERIT += "testlab" Jan 24 11:05:05 I have a script comparing the size of the packages Jan 24 11:05:15 I'm doing a distro-shootout Jan 24 11:05:27 atm uclibc < eglibc as expected Jan 24 11:05:51 but if I try to compare the packages I have a small issue: Jan 24 11:06:17 see: Jan 24 11:06:21 1456<-->KiB /armv5te/libc6_2.9-r35.3.5_armv5te.ipk Jan 24 11:06:48 1456<-->KiB /armv5te/libc6_2.10-r8.6_armv5te.ipk Jan 24 11:06:59 one has DISTRO_PR or ABI ? Jan 24 11:07:04 that .5 Jan 24 11:07:26 only Angstrom has it Jan 24 11:07:59 tbh, i never understood what's DISTRO_PR good for Jan 24 11:08:04 hence minimal does not set it Jan 24 11:08:35 well, in the above example there is ver mismatch, but you got the point :) Jan 24 11:15:14 mickeyl: DISTRO_PR. You change global TARGET_CFLAGS. E.g. at Openmoko we started to use GNU HASHSTYLE Jan 24 11:15:28 mickeyl: and we wanted to upgrade all installed packages... -> DISTRO_PR Jan 24 11:15:44 ah, makes sense Jan 24 11:15:48 zecke: I'll take a look in about an hour, I have to go over to my mother in law now Jan 24 11:15:56 yes, minimal can set it to 0 then Jan 24 11:16:12 pb__: take care! Jan 24 11:16:14 btw allows us to notify users when the format of TMPDIR changes in Jan 24 11:16:17 * zecke play patch monkey Jan 24 11:16:22 OELAYOUT_ABI Jan 24 11:16:33 is the same thing? Jan 24 11:17:32 ant__: no Jan 24 11:17:59 I mean, practically you have to rebuild Jan 24 11:17:59 ant__: e.g. if we rename staging, to zecke-rocks, we bump the OELAYOUT_ABI Jan 24 11:18:21 ant__: DISTRO_PR is something that will end up in package versions, OELAYOUT_ABI is just a local thing Jan 24 11:19:02 hmm.. Jan 24 11:19:04 ./packaged-staging.bbclass:PSTAGE_EXTRAPATH ?= "/${OELAYOUT_ABI}/${DISTRO_PR}/" Jan 24 11:20:01 it makes sense Jan 24 11:20:20 the content is controled by distro_pr, and the place by oelayout_abi Jan 24 11:21:02 mickeyl: any idea what we should do with patches that are not well formed but show an issue and the dev is MIA? Jan 24 11:22:23 zecke: I see, it's another level Jan 24 11:22:31 still, Jan 24 11:22:33 ./distro/include/sane-toolchain.inc:PSTAGE_EXTRAPATH = "${LIBC}" Jan 24 11:22:41 ./distro/include/angstrom.inc:PSTAGE_EXTRAPATH = "${LIBC}" Jan 24 11:22:56 so it ends up unused Jan 24 11:23:08 zecke: change media... e.g. stick them in bugtracker and wait for the developer to reappear or someone to claim responsibility Jan 24 11:29:58 I have a problem with git sources fetching... Jan 24 11:29:59 rsync: read error: Connection reset by peer (104) Jan 24 11:30:07 but fetching it by hand works ok Jan 24 11:30:43 filip: and it is using rsync too? Jan 24 11:31:06 dunno why Jan 24 11:31:14 I have an idea though Jan 24 11:31:41 filip: and you don't provide much context. You want to say, from a recipe you wrote yourself it does not work? Jan 24 11:31:52 filip: and when you type git clone git://foo; it is working? Jan 24 11:31:54 zecke: yes, I didn't add protocol=git Jan 24 11:31:59 it works now Jan 24 11:33:54 *sigh* Jan 24 11:34:11 ;) Jan 24 11:34:30 will the kernel class separate kernel image into separate module packages? Jan 24 11:36:57 yes, modules are automatically seperated Jan 24 11:38:01 cool :) Jan 24 11:38:03 thanks Jan 24 12:05:56 can I reflect this: "git-checkout --track -b v2.6.17-hpc origin/v2.6.17-hpc" in a git src_uri? using branch=v2.6.17-hpc results in: git read-tree 1 failed with signal 128, output: Jan 24 12:05:59 fatal: Not a valid object name 1 Jan 24 13:47:09 Should it be~ possible to define modules_autoload in image/task? Or only in distro/machine settings? Jan 24 13:48:22 I would like to avoid modifying distro/machine just to autoload wifi specific kernel module Jan 24 13:49:12 I've module_autoload_rt3070sta = "rt3070sta" in my image recipe, but it seems, that it doesn't work Jan 24 13:52:00 nm, just found it in kernel.bb class so it seems, that I've to rebuild kernel to get it working Jan 24 14:00:28 is there an udev script which can ipkg install modules before modprobing them? Jan 24 14:18:06 zecke, pong Jan 24 14:48:07 Crofton: I replied to your libtool question. It is the only of your patches in the queue. Jan 24 14:52:56 re Jan 24 14:58:27 my libtool question? Jan 24 14:58:54 I am behing on email, we helped a friend move yesterday Jan 24 14:59:15 Crofton: it is older, you wanted to add libtool to the SDK task? Jan 24 14:59:42 ah Jan 24 14:59:43 ok Jan 24 15:00:18 I know I was sloppy on the patch :) Jan 24 15:00:39 did you check that it worked, I haven't had a chance to check it :) Jan 24 15:03:47 Crofton: haha, it doesn't even apply. Jan 24 15:03:57 as your mailer wrapped the lines ;) Jan 24 15:04:10 ok Jan 24 15:04:18 I am going to push it, it builds Jan 24 15:04:36 I jsut haven't gone back and checked the package would build without more ipks installed Jan 24 15:04:42 a guy was building gphoto Jan 24 15:04:49 naively Jan 24 15:05:51 I thought you might have tried the using the native sdk and also run across the need for libtool Jan 24 15:06:51 done Jan 24 15:06:56 thanks for reminding me Jan 24 15:12:17 ynezz: yes.. autoload/module options are set as part of postinst in kernel-module-sth Jan 24 15:12:35 Crofton: I use it for qt, which comes with this qmake stuff. Jan 24 15:12:42 anyway, I need to sleep. take care. Jan 24 15:14:48 gn Jan 24 15:14:57 shold be pushed Jan 24 15:15:06 not sure why the bot has not reported it Jan 24 15:16:20 do somehow have also 2.2.6b version? Jan 24 15:16:46 seems like transmission is failing with 2.2.6a.. so I guess someone has.. Jan 24 15:21:39 mckoan|away: it took me something like 30 seconds to search for http://patchwork.openembedded.org/patch/773/ and http://patchwork.openembedded.org/patch/770/ Doesn't take an Übereffort to find. Both patches seem to be needing some work (or just close them if they are obsolete by now) Jan 24 15:59:39 JaMa: ok, so this is just for in kernel modules, not for the modules built outside... Jan 24 16:03:19 ynezz: yes.. we had few echo module_name /etc/modprobe.d/module_name in image*.bb for that (instead modules_autoload) Jan 24 16:27:52 how can I select a remote git branch with SRC_URI? Jan 24 16:38:36 I think you are confusing something Jan 24 16:38:58 look for examples Jan 24 16:39:11 find recipes -name *_git.bb Jan 24 17:17:35 Laibsch: I mean a line that would cause: git-checkout --track -b v.2.6.30-hpc origin/v2.6.30-hpc Jan 24 17:17:54 find recipes -name *_git.bb Jan 24 17:18:01 Laibsch: branch=v2.6.30-hpc doesn't do this Jan 24 17:18:02 gm Jan 24 17:18:02 I did Jan 24 17:18:12 but... it wasn't of any help for me Jan 24 17:18:52 for now I did like in linux/linux-efika_2.6.21+git.bb (own do_fetch), but it's dirty Jan 24 17:19:03 Laibsch: I already saw those old patches, of course, I was seeking for something new Jan 24 17:19:29 mckoan: If they don't need to be committed anymore, then please be so kind to close the tickets Jan 24 17:19:30 thanks Jan 24 17:19:41 it'll help zecke in his work Jan 24 17:20:09 Laibsch: yep :-) Jan 24 17:21:14 Laibsch: done Jan 24 17:21:21 great, thanks Jan 24 17:21:29 BTW, you are an assurer for cacert? Jan 24 17:21:45 Laibsch: yep Jan 24 17:35:30 mickey, khem: Can you please ACK my RFC about distutils-common-base? Jan 24 17:35:54 hmm Jan 24 17:36:08 i'm not sure about the effects this change has Jan 24 17:36:28 I'd like to see what happens to task-python-everything before and after that change Jan 24 17:51:54 btw, do anybody recieved touchbook already? Jan 24 17:57:07 Laibsch: about .dbg, see pam recipe Jan 24 17:57:09 NOTE: the following files were installed but not shipped in any package: Jan 24 17:57:17 all libpam-xy.so Jan 24 17:57:43 err..I mean NOTE: /usr/lib/security/.debug/pam_xyz.so Jan 24 18:05:52 Jay7: sure Koen has a TB, not sure if any other OE devs do. Jan 24 18:07:03 DJWillis: do you know are they shipping TB already? Jan 24 18:07:10 or still pre-ordering? Jan 24 18:08:02 Jay7: I understand orders made now will ship mid feb but don't quote me. Jan 24 18:08:22 Jay7: they have shipped a fair few units in previous batches. Jan 24 18:08:47 DJWillis: thanks :) Jan 24 18:14:18 mickeyl: I think it was just an omission. the current behvaiour is certainly broken Jan 24 18:14:37 ant__: what do you mean? pam has the same problem? is pam inheriting that class? Jan 24 18:14:53 no, I see there is some python, though Jan 24 19:33:03 ant__: Can kexecboot from CF? I can't seem to do that Jan 24 19:34:01 brb Jan 24 21:05:29 ant__: does kexecboot support booting from Compact flash? Jan 24 21:11:04 Laibsch: it should Jan 24 21:11:18 did you try successfully? Jan 24 21:11:23 morning all Jan 24 21:11:30 I tried at least with collie and it didn't work Jan 24 21:11:30 I do on tosa some time ago Jan 24 21:11:33 hey RP Jan 24 21:11:38 OK Jan 24 21:11:46 hi RP :) Jan 24 21:11:49 thank you for letting me know, Jay7 Jan 24 21:11:49 Laibsch: try to do rescan Jan 24 21:11:54 or reboot Jan 24 21:11:56 I did already Jan 24 21:12:22 we have some problems with CF but that was wrong kernel config Jan 24 21:12:36 no CF storage was detected Jan 24 21:12:59 that may have been the problem Jan 24 21:13:08 I was never shown anything from the CF Jan 24 21:13:24 I'll keep an eye on it Jan 24 21:13:42 thanks for testing :) Jan 24 21:13:49 RP: hey Jan 24 21:14:13 RP: can you explain a bit about lzma-compressed kernel booting scheme? Jan 24 21:14:23 RP: should loader have special support? Jan 24 21:14:32 or kernel is self-decompressing? Jan 24 21:14:55 ant__ have tried to boot lzma-compressed kernel on his Zaurus but without success Jan 24 21:58:51 Laibsch: iirc thesing and XorA tested on collie. The old images here http://www.angstrom-distribution.org/demo/kexecboot/ were not yet enabled for boot from CF (on clamshells) Jan 24 22:02:43 Jay7: I'm pretty sure the kernel should decompress itself. Jan 24 22:03:10 It does when launched by kexec. It's just another bootloader legacy-issue :/ Jan 24 22:11:40 he... recipes using do_stage have strange behaviors... Jan 24 22:11:51 NOTE: Running task 2120 of 4025 (ID: 3715, /oe/org.openembedded.dev/recipes/tcltk/tcl-native_8.4.19.bb, do_package) Jan 24 22:11:51 cp: cannot stat `/oe/build/tmp/work/i686-linux/tcl-native-8.4.19-r4/image/*': No such file or directory Jan 24 22:11:51 NOTE: Running task 2121 of 4025 (ID: 1029, /oe/org.openembedded.dev/recipes/avahi/avahi_0.6.25.bb, do_populate_staging) Jan 24 22:12:01 several recipes... Jan 24 22:12:24 cp: cannot stat .... No such file or directory Jan 24 22:41:18 I see RP reworked these recipes in Poky... Jan 24 23:08:15 Jay7: I really have no idea about that **** ENDING LOGGING AT Mon Jan 25 02:59:56 2010