**** BEGIN LOGGING AT Thu Dec 08 02:59:57 2011 Dec 08 10:41:03 morning all Dec 08 15:45:58 nitink: I sent the dosfstools patch we talked about to the mailing list, please ack it if you can Dec 08 16:50:05 otavio: sure I will take a look Dec 08 16:50:51 nitink: great. Dec 08 19:25:23 mebrown: ping Dec 08 20:15:17 zeddii: you around? Dec 08 20:15:46 yep. was about to bolt, but am here. what's up ? Dec 08 20:15:49 zeddii: I am trying your new kernel, and on multi machines getting a not a vaild commit id during the fetch Dec 08 20:16:08 | ERROR a611ba462ac51b2bfdac1c5c6538dac95ad9b4dd is not a valid commit ID. Dec 08 20:16:12 checking. Dec 08 20:16:44 I hope our mirroring is not failing again. Dec 08 20:17:08 or tree is rebased Dec 08 20:17:35 sgw, I have that pushed out. could very well be the mirroring. Dec 08 20:18:26 should there be a 3.0.12 tag? Dec 08 20:18:28 hmm. checking now. I'm forcing out the tag Dec 08 20:18:31 beat me to it .. Dec 08 20:19:05 interesting. the tag seems to be there now. but wasn't before. I always explicitly push tags. how long does the re-mirror take ? Dec 08 20:19:34 I just pulled from git://git.yoctoproject.org/linux-yocto-3.0 to a local clone and it has it now, did not before! Dec 08 20:19:42 let me try a build quickly again Dec 08 20:19:50 * zeddii crosses his fingers :) Dec 08 20:21:46 sgw: https://github.com/OSSystems/oe-core/commit/83266d1b4a3d524453fdc87830dbf63bf6f0651b <= connman 0.78 Dec 08 20:21:59 had to force the fetch but it seems to be kernel_configme so it should be OK Dec 08 20:22:12 sgw: mind to give it a look and see if you spot anything too wrong? otherwise I'll send a pull request with it Dec 08 20:23:47 sgw: cool and phew. Dec 08 20:26:38 otavio: looking Dec 08 20:39:18 sgw: good Dec 08 20:39:36 sgw: I need to go out now, please leave a message or send me an e-mail with comments ... Dec 08 20:40:01 sgw: I have a test at University later tomorrow and we'll do a group learning for each in half hour Dec 08 20:40:05 sgw: so I gotta run Dec 08 20:48:57 otavio: later Dec 08 22:36:29 JaMa: ping Dec 08 22:40:38 pong Dec 08 22:41:32 JaMa: what do you think about make_zimage_symlink_relative () in image.bbclass? Dec 08 22:41:43 not used in meta-oe and oe-core Dec 08 22:41:57 only in meta-handheld Dec 08 22:41:59 andrea@mizar /oe/meta-handheld $ grep -R make_zimage . Dec 08 22:42:00 ./conf/machine/include/LAB-settings.inc:#ROOTFS_POSTPROCESS_COMMAND += "make_zimage_symlink_relative; " Dec 08 22:42:00 ./conf/machine/include/zaurus.inc:ROOTFS_POSTPROCESS_COMMAND += "make_zimage_symlink_relative; " Dec 08 22:42:14 now, the logic should be reversed imho Dec 08 22:42:33 ->if [ ! -L ${IMAGE_ROOTFS}/boot/zImage ]; then Dec 08 22:43:03 (the underlying issue is the missing symlink in /boot) Dec 08 22:43:37 well it doesn't work very well with oe-core (I have to update link after every kernel upgrade) Dec 08 22:43:50 with that addition seems ok Dec 08 22:43:57 but that cannot be fixed in rootfs time as we need it in postinst too Dec 08 22:44:54 (I only had to set ROOTFS_POSTPROCESS_COMMAND = "" in the initramfs image) Dec 08 22:45:33 KERNEL_IMAGE_SYMLINK_NAME is better Dec 08 22:46:01 and I'm off Dec 08 22:46:04 yes, then we should rename the function (uImage) Dec 08 23:29:18 RP__: how would you handle this? kernel-image only contains one file, the kernel. For bootloader purposes we need to point to a generic symlink e.g. /boot/zImage Dec 08 23:29:52 ant____: postinstall updating the symlink? Dec 08 23:30:06 well, somehow it was created in oe-classic... Dec 08 23:30:19 then I founs the stale code in image.bbclass Dec 08 23:30:33 make_zimage_symlink_relative () Dec 08 23:31:20 but as JaMa said we should take care of the kernel upgrades too Dec 08 23:32:40 RP__: maybe adding the check to pkg_postinst_kernel-image () { Dec 09 01:45:29 incandescant: I pushed the changes you asked. https://github.com/OSSystems/oe-core/commit/cd54d3b1ce8da2b2612719e23785cf57632581e3 Dec 09 01:45:40 incandescant: please check if I can keep the ack now Dec 09 01:52:21 otavio: I assume the xtables patch is no longer required as the change is upstream? Can you mention that and why you added ppp to DEPENDS in the commit log? What does ppp buy us? Can it be optional? Dec 09 01:52:28 otavio: sorry I didn't raise these the first round Dec 09 01:54:16 otavio: I see that ppp is a dynamic plugin, so it's not a concern but it would be nice to mention it in the log as it's a new change (to oe-core) Dec 09 01:55:51 incandescant: https://github.com/OSSystems/oe-core/commit/deef6780659b0b0094cc95478a21a1a517e872d8 Dec 09 01:55:59 incandescant: looks good now? Dec 09 01:58:26 otavio: how about change "dbusperms.patch: used by git recipe, that is being dropped" to "dbusperms.patch: used by git recipe, that is being dropped in favour of using sed to make the change in do_compile_append" - I know that's obvious but it makes the review easier by knowing what changes to expect Dec 09 02:02:18 incandescant: https://github.com/OSSystems/oe-core/commit/781495b5d75af610565ce494dc03e57521a63ec5 Dec 09 02:02:21 incandescant: now? Dec 09 02:03:25 thanks otavio, feel free to use my ack Dec 09 02:03:48 incandescant: awesome Dec 09 02:04:02 otavio: how much run time testing of connman do OSSystems do? I heard some people having troubles with it recently. Does it work well for you? Dec 09 02:04:21 incandescant: I will wait until the other two patches were merge and send it Dec 09 02:04:36 incandescant: we are finishing a graphical tool for it Dec 09 02:04:39 incandescant: wait Dec 09 02:04:50 incandescant: https://github.com/OSSystems/qconnman Dec 09 02:05:07 incandescant: but it requires some not merged patches on connman Dec 09 02:05:12 otavio: awesome, I was wondering about writing such a tool Dec 09 02:05:16 incandescant: for hidden wifi networking Dec 09 02:05:23 * incandescant wrote the moblin/meego GUI for ConnMan Dec 09 02:05:39 incandescant: so take a look and contribute if you want Dec 09 02:06:01 otavio: I'll definitely take a look, thanks Dec 09 02:07:14 incandescant: connman will be our default manager from now on Dec 09 02:07:22 incandescant: so we will do a lot of testing/work on it Dec 09 02:07:33 incandescant: we were using nm up to now but it is a nightmare Dec 09 02:08:34 otavio: I liked connman once I figured it out, most stuff just works Dec 09 02:08:48 otavio: will be nice to know more folks are using/testing it in OE-Core :-) Dec 09 02:09:07 incandescant: when we were starting the network management stuff here it was horrible Dec 09 02:09:13 incandescant: too incomplete and unstable Dec 09 02:09:30 incandescant: but now it has most nm features and dbus api is much more sane and easy to use Dec 09 02:10:06 otavio: how long have you been using it? are you using the services api? Dec 09 02:10:21 incandescant: for two weeks Dec 09 02:11:05 incandescant: we had a issue report with nm and it started to make us feel bad; so we needed to decide what to do and we decided to get rid of nm from our os Dec 09 02:11:22 incandescant: it is going to be used in three products Dec 09 02:11:30 otavio: wow, sounds great! Dec 09 02:11:41 otavio: can you tell us what? :-) Dec 09 02:11:42 incandescant: one, the installer is being in final tests Dec 09 02:12:07 incandescant: another will start moving in 2012 Q1 Dec 09 02:12:24 incandescant: and the third is going to use qconnman Dec 09 02:12:35 otavio: sounds like OS Systems are doing well! Dec 09 02:13:16 incandescant: we are still a startup but we do want 2012 to be our best year ever Dec 09 02:13:24 incandescant: so, a lot of work Dec 09 02:13:57 otavio: best of luck to you! sounds like you're setting yourselves up for a good start to the year Dec 09 02:14:15 incandescant: :) for all of us :-D Dec 09 02:15:12 incandescant: please take a look on qconnman and report anything you think we can improve or do different Dec 09 02:15:24 incandescant: for now it supports wifi and ethernet only Dec 09 02:15:41 incandescant: it will be send tomorrow to meta-oe, btw Dec 09 02:16:42 Well, time to go to bed. Dec 09 02:17:33 see ya Dec 09 02:17:37 otavio: I'll add to my list and try and check it out Dec 09 02:17:40 night otavio **** ENDING LOGGING AT Fri Dec 09 02:59:58 2011