**** BEGIN LOGGING AT Thu Dec 23 02:59:58 2010 Dec 23 06:34:45 mickey|zzZZzz: ohh, great thanks Dec 23 06:34:52 was already starting to debug why it does not work :-) Dec 23 07:16:33 moin Dec 23 07:28:36 moin Dec 23 07:51:48 mrmoku: ericben from #oe pointed me to 2 threads related to new gcc-4.5.1 issues with u-boot, can you try new binary in half an hour or so? Dec 23 07:59:41 JaMa: yup Dec 23 08:04:07 mrmoku: http://build.shr-project.org/tests/jama/u-boot/u-boot-nokia900-2010.06+gitr0+bd2313078114c4b44c4a5ce149af43bcb7fc8854-r68.bin Dec 23 08:06:36 ohh, that was fast :-) Dec 23 08:08:02 :) Dec 23 08:10:23 JaMa: it does not boot maemo :-/ Dec 23 08:10:36 and shr I do have on the first partition Dec 23 08:10:52 but maemo is hanging at Starting kernel as usual Dec 23 08:10:56 still hanging on "Starting kernel"? Dec 23 08:10:58 ah ok Dec 23 08:11:14 so this is not enough http://patchwork.ozlabs.org/patch/76386/ :/ Dec 23 08:11:58 looks like :/ Dec 23 08:12:32 thanks for test Dec 23 08:12:37 np Dec 23 08:55:22 mrmoku: I have binary from master with that patch too, if you feel lucky today :) Dec 23 09:11:30 JaMa: I feel always lucky... though I often feel wrong :P Dec 23 09:11:36 let me try anyway :) Dec 23 09:25:54 mrmoku: http://build.shr-project.org/tests/jama/u-boot/u-boot.4.2.master.debug.bin Dec 23 09:26:42 mrmoku: you'll probably just see nolo splash (nokia logo) and no sign of u-boot starting Dec 23 09:44:03 JaMa: hmm... does nothing Dec 23 09:44:15 just shows me the white screen with blue NOKIA Dec 23 09:47:58 ok, so the same as my previous git master experiments :/, thanks Dec 23 09:51:51 I'll try http://lists.denx.de/pipermail/u-boot/2010-December/083817.html later Dec 23 09:52:03 ok Dec 23 09:53:56 actually I can try to be dirty and compile whole u-boot except clock.c with gcc-4.5.1 and only clock.c with gcc-4.2 to confirm that's the issue we have Dec 23 10:25:55 mrmoku: dirty one http://build.shr-project.org/tests/jama/u-boot/u-boot-nokia900-2010.06+gitr0+bd2313078114c4b44c4a5ce149af43bcb7fc8854-r67.bin.4.5.3.clock Dec 23 10:31:57 * mrmoku trying Dec 23 10:32:35 JaMa: this one shows me the u-boot screen again, but hangs at Starting kernel Dec 23 10:33:46 ok, so it's not the same issue (or at least not cause for hangs), thx again Dec 23 10:33:54 again np :) Dec 23 10:52:30 * mrmoku desperately searching the vala equivalent for g_io_channel_unix_new Dec 23 10:53:29 ahh Dec 23 10:53:33 nvm. found it :) Dec 23 10:58:33 consider using the ReactorChannel Dec 23 10:58:48 hmm Dec 23 10:58:52 which is a wrapper of that Dec 23 10:58:58 but we can do it later Dec 23 10:59:03 ok Dec 23 10:59:09 (not too many variables at once) Dec 23 11:03:24 Hi there :) I wonder if it's normal for a minimal angstrom oe build to take up > 40Gb of disk space. Dec 23 11:05:39 I even added INHERIT+="rm_work" to my local.conf but i guess that's the wrong place or the wrong word. Dec 23 11:16:12 40G sounds a bit on the fat side Dec 23 11:16:23 is that including sources? Dec 23 11:17:50 my shr-lite-image build takes over 50Gb if i build everything from scratch at once Dec 23 11:18:10 because rm_work comes at last Dec 23 11:20:28 mickeyl: how do I do that: http://shr.pastebin.com/vf2e9frR ? Dec 23 11:20:55 uah :D Dec 23 11:21:26 it won't be so elegant Dec 23 11:21:53 I don't care ;) Dec 23 11:22:05 I want it to see working before I leave tomorrow morning :/ Dec 23 11:22:51 ok, so basically you can create the structs and then use component assignment Dec 23 11:23:29 i'd start with defining the outer struct first Dec 23 11:23:36 mickeyl: yes, including sources Dec 23 11:23:44 ok Dec 23 11:23:46 struct assignment i'm not sure, perhaps you have to resort to memcpy for tahat Dec 23 11:23:54 (i.e. assigning one struct to another) Dec 23 11:24:13 PaulFertser: ok, then it's not so surprising. the sources even for a small image are huge, since they contain all the tools Dec 23 11:24:33 rm_work though should make it a bit smaller than what you see Dec 23 11:24:56 mickeyl: i know but it still feels a bit over the top. I mean i have a separate 40Gb partition for oe and it's not enough. Dec 23 11:25:11 At least openwrt takes much less ;) Dec 23 11:25:17 mickeyl: lunch first Dec 23 11:25:24 mrmoku: feel free to commit what you have already Dec 23 11:25:38 you can add the stuff left to convert in /* */ Dec 23 11:25:41 then i can give you a hand Dec 23 11:26:16 PaulFertser: well... openwrt _is_ less ;) Dec 23 11:26:27 no seriously, it was much smaller before Dec 23 11:26:31 i don't know what they did to OE in 2010 Dec 23 11:26:53 mickeyl: yes, i remember compiling shr a year ago and it took reasonable space. Dec 23 11:27:00 with the yocto merge and separating OE into layers it might get better Dec 23 11:27:12 i remember something about 10G without sources Dec 23 11:27:38 Yeah Dec 23 11:29:46 Ok, thanks everybody :) Dec 23 11:33:07 PaulFertser: to keep tmpdir size resonable I'm building images in "stages" like "bitbake -k gcc-cross; bitbake -k virtual/kernel; bitbake -k shr-lite-image; bitbake -k shr-image" Dec 23 11:33:35 PaulFertser: so rm_work cleans tmpdir between each stage (and ~ 8G is enough for tmpdir itself) Dec 23 11:34:44 bbiab, groceries... Dec 23 11:38:44 JaMa: a rather decent advice again! :) Dec 23 11:41:07 PaulFertser: :) Dec 23 11:42:15 PaulFertser: btw I feel quite sad that I've persuaded you to work on gta01 support in 2.6.34 and nobody provided decent feedback after it :/ where are all those people who were complaining about state of SHR on gta01 :/ Dec 23 12:29:37 mickeyl: ok Dec 23 12:30:33 mickeyl: NLMSG_DATA and IFA_PAYLOAD are missing too Dec 23 12:30:50 or maybe are done otherwise... dunno Dec 23 12:36:01 mickeyl: and ifinfomsg from rtnetlink.h too Dec 23 12:40:16 mickeyl: and ifaddrmsg and rtattr Dec 23 12:44:08 JaMa: tbh i still feel ashamed i didn't actually test it on my device Dec 23 12:54:54 freesmartphone.org: 03mok 07cornucopia * r4a5fd0905f52 10/fsogsmd/conf/nokia_n900/fsogsmd.conf: Dec 23 12:54:54 freesmartphone.org: fsogsmd/conf/nokia900: default to DEBUG for now Dec 23 12:54:54 freesmartphone.org: Signed-off-by: Klaus Kurzmann Dec 23 12:54:58 freesmartphone.org: 03mok 07cornucopia * rcaa07b5a3f0b 10/fsogsmd/src/plugins/lowlevel_nokia900/ (nokia-gpio.c nokia-gpio.h plugin.vala): Dec 23 12:54:58 freesmartphone.org: fsogsmd/lowlevel_nokia900: implement nokia-gpio in the plugin itself - WIP Dec 23 12:54:58 freesmartphone.org: Signed-off-by: Klaus Kurzmann Dec 23 12:55:26 mickeyl: I pushed it... it won't build though without updated vala Dec 23 12:55:41 and then one has to change PosixExtra to Posix Dec 23 12:55:56 the missing parts are mainly the netlink_... thingies Dec 23 12:56:06 I added the ofono c code as comment there Dec 23 13:01:22 bbl Dec 23 13:02:06 back Dec 23 13:14:38 mrmoku: I'll take care about netlink_getlin know Dec 23 13:55:00 hmm, grr Dec 23 13:55:07 usr/include/linux/if.h:131: error: redefinition of ‘struct ifmap’ Dec 23 13:56:35 oh the fun mixing userland and kernel headers *sigh* Dec 23 14:16:36 ok, netlink_getlink() done Dec 23 14:17:52 btw., it's 10 points for elegance to convert it to vala, but pragmatically i would have just reused the c file and created some ad-hoc extern() bindings Dec 23 14:18:16 to reuse the netlink stuff Dec 23 14:18:23 but anyways, it's almost done now.. Dec 23 14:19:57 * mickeyl binds some more netlink macros Dec 23 14:39:29 freesmartphone.org: 03mickey 07cornucopia * r4175a613f7cc 10/fsogsmd/src/plugins/lowlevel_nokia900/plugin.vala: fsodeviced: lowlevel_n900: convert more netlink code to vala. still WIP and needs latest Vala GIT to compile Dec 23 14:39:55 we should really consider just embedding the netlink functions Dec 23 14:40:22 it's as simple as adding the c file and writing a bunch of extern declarations in plugin.vala Dec 23 14:40:30 but it's your call Dec 23 14:47:21 freesmartphone.org: 03mickey 07cornucopia * rc332e62e0207 10/misc/vapi/posixextra.vapi: misc: socklen_t is now upstream Dec 23 14:55:41 does anyone knows the problem ERROR: Fetch command git ls-remote http://git.shr-project.org/repo/libphone-ui-shr.git master gave empty output Dec 23 14:55:41 while parsing /media/disk/oe/shr-unstable/openembedded/recipes/shr/libphone-ui-shr_git.bb Dec 23 14:55:41 while building shr-unstable in oe Dec 23 15:00:03 here is the complete error log http://pastebin.com/EbrGJzP4 Dec 23 15:29:00 freesmartphone.org: 03mickey 07gdbus * rd61be0517284 10libfso-glib/README: document branch Dec 23 15:34:51 freesmartphone.org: 03mickey 07libfso-glib * r501d1eaed3c1 10/ (13 files in 2 dirs): remove everything but the notice Dec 23 15:36:02 freesmartphone.org: 03mickey 07gdbus * r858fe67fe578 10libfso-glib/configure.ac: bump Dec 23 15:42:21 dos1, ping Dec 23 15:43:17 pespin: pong Dec 23 15:44:49 dos1, I'd like to do some refactoring of the code of shr-settings, as there is some code which could be shared among all modules and would make things easier to write new modules. These are things like label+entry widgets, popup errors, etc. Dec 23 15:45:16 dos1, is it ok for you? which would be the best way in your opinion? Dec 23 15:45:31 create some shared.py and import it in modules which use it? Dec 23 15:45:39 pespin: yep, i'm ok. Just don't break too much ;) Dec 23 15:45:52 dos1, yeah don't worry, I'll try things before commiting Dec 23 15:46:35 pespin: there is already file with skeleton of class for module Dec 23 15:47:14 pespin: module.py Dec 23 15:47:53 dos1, well but I think not all modules will need it Dec 23 15:48:11 and that one is used for every module Dec 23 15:48:56 pespin: ok, so do some helpers.py or something like that Dec 23 15:49:01 dos1, ok :) Dec 23 15:53:11 freesmartphone.org: 03mickey 07gdbus * r92b2a799270f 10libfso-glib/README: prepare for merging gdbus back to master Dec 23 15:53:13 freesmartphone.org: 03mickey 07libfso-glib * rceed86d1982a 10/README: fix typo Dec 23 16:41:52 mickeyl: great... btw. it started because of log output of the c thing going to stdout instead of fsogsmd.log :P Dec 23 16:43:24 mickeyl: and embedding the netlink functions via extern declarations is fine for me Dec 23 16:43:38 as long as the important stuff like the state machine is in vala Dec 23 16:46:12 TAsn, hey! I've just remembered I had to remind you to look at the phoneui-sms-report code! :P Dec 23 17:11:40 damn... I should have prepared the gdbus merge for phonefsod/phoneuid in a branch too :( Dec 23 18:10:24 mrmoku: ok, great. i'll take a shot at doing that while you're away then Dec 23 18:10:34 gdbus migration has priority for me Dec 23 18:10:37 mickeyl: good Dec 23 18:10:39 i want to start w/ that tomorrow Dec 23 18:10:49 when do you leave? Dec 23 18:10:49 mickeyl: I'm looking into what that means for our stuff right now Dec 23 18:10:53 ok, cool Dec 23 18:10:55 tomorrow morning :/ Dec 23 18:11:01 ah Dec 23 18:11:18 installing vala from git on my laptop right now Dec 23 18:11:26 right Dec 23 18:11:29 let's see how big the changes are... Dec 23 18:12:35 btw. libfso-glib building emits quite some warnings :P Dec 23 18:13:49 sad thing I don't have permanent internet in italy :/ Dec 23 18:14:09 ya, and roaming costs are absurd Dec 23 18:14:09 hopefully they will force providers to not super-charge data roaming someday Dec 23 18:14:13 :) Dec 23 18:14:15 *nod* Dec 23 18:14:37 fortunatelly some of the bars there have free wifi Dec 23 18:14:43 right, that may come in handy Dec 23 18:14:52 btw., the compund struct assignment is equally elegant to that in C Dec 23 18:14:56 compound Dec 23 18:15:10 ohh, nice :) Dec 23 18:15:47 ok, now fsobasics Dec 23 18:15:57 err... libnl-2 first Dec 23 18:16:27 grr Dec 23 18:16:44 * mrmoku has libnl installed for networkmanager :/ Dec 23 18:17:02 mrmoku: does it mean your going on xmas vacation to your wife's parents? Dec 23 18:17:11 PaulFertser: yup Dec 23 18:17:32 * mrmoku would prefer not to :P Dec 23 18:17:43 mrmoku: why? I'd like to visit Italy someday :) Dec 23 18:17:47 but don't tell anybody Dec 23 18:18:22 PaulFertser: sure, but I don't even remember how often I went to italy Dec 23 18:18:25 hehe Dec 23 18:18:29 so it's not very exciting for me Dec 23 18:18:37 brb Dec 23 18:18:39 hi all! Dec 23 18:18:45 :) Dec 23 18:18:46 ok, everyone prepared for the damage? Dec 23 18:18:55 * mickeyl git merge gdbus Dec 23 18:18:56 mickeyl, what damage? gdbus? Dec 23 18:18:57 eeeeeh Dec 23 18:19:01 here we go Dec 23 18:19:03 :D Dec 23 18:19:03 :) Dec 23 18:19:26 mickeyl, those things get you really excited, don't they? :P Dec 23 18:20:03 daniele_athome: i would lie if i'd say no :) Dec 23 18:20:09 hihi Dec 23 18:20:37 well, if i wasn't sure that it's the right thing to do i would have saved all the work Dec 23 18:21:11 freesmartphone.org: 03mickey 07dbus-glib * ra568426435c9 10cornucopia/ (183 files in 94 dirs): Merge branch 'gdbus' into dbus-glib Dec 23 18:21:11 freesmartphone.org: 03mickey 07cornucopia * r83c4da4a1c19 10/ (183 files in 94 dirs): Merge remote branch 'origin/gdbus' Dec 23 18:21:41 aargh Dec 23 18:21:48 aaaaaah Dec 23 18:21:50 lol Dec 23 18:21:58 *sigh* Dec 23 18:22:01 mickeyl, what happened? Dec 23 18:22:07 did the repository explode? :) Dec 23 18:22:11 seriously what is it? Dec 23 18:23:02 that first merge should not have happened Dec 23 18:23:05 fix it with -f Dec 23 18:23:07 err, fixed Dec 23 18:23:28 dbus-glib now contains the last state before the merge Dec 23 18:23:39 SHR-T can continue to use that if they have to apply bugfixes Dec 23 18:23:40 mickeyl, that's the excitement :P Dec 23 18:24:28 hehe, yeah Dec 23 18:24:41 :D Dec 23 18:29:27 Heinervdm: remember to adjust branches for shr-t :) Dec 23 18:29:39 mrmoku: will do :) Dec 23 18:31:23 freesmartphone.org: 03mickey 07gdbus * rc58f310d8cf2 10libfso-glib/configure.ac: gdbus is enabled by default in this branch Dec 23 18:33:47 mickeyl: hmm... there are now libfso-glib.git/gdbus and libfso-gdbus.git/master Dec 23 18:34:17 i will remove libfso-gdbus Dec 23 18:34:19 gcc: unrecognized option '-nl-genl' Dec 23 18:34:28 hmm Dec 23 18:34:30 mrmoku: sorry, that slipped into the release tarball Dec 23 18:34:37 please add -l in .pc file Dec 23 18:34:43 ok Dec 23 18:35:00 mickeyl: ahh... yeah, I remember Dec 23 18:37:21 i should put up a new tarball Dec 23 18:37:33 then again, i don't feel like handpatching all that stuff again Dec 23 18:37:37 who knows what he broke this time Dec 23 18:38:59 mickeyl: there is still now binding tool for gdbus, right? Dec 23 18:39:09 s/now/no/ Dec 23 18:39:11 mrmoku meant: mickeyl: there is still no binding tool for gdbus, right? Dec 23 18:39:46 i don't know. since we reuse vala-dbus-binding-tool also for gdbus, i never looked for C Dec 23 18:40:05 ok Dec 23 18:40:07 * mrmoku googles Dec 23 18:40:33 http://people.freedesktop.org/~david/gdbus-binding-tool-20100518/gdbus-binding-tool.html Dec 23 18:40:36 :) Dec 23 18:43:26 JaMa|Off: we will have to package gdbus-binding-tool :-) Dec 23 18:48:02 ah, cool Dec 23 18:48:35 http://cgit.freedesktop.org/~david/gdbus-binding-tool/ Dec 23 18:50:35 freesmartphone.org: 03mickey 07dbus-glib * ra568426435c9 10cornucopia/ (183 files in 94 dirs): Merge branch 'gdbus' into dbus-glib Dec 23 18:50:36 freesmartphone.org: 03mickey 07cornucopia * r6f55e1ab5dbd 10/fsodeviced/src/plugins/palmpre_quirks/ (5 files): palmpre_quirks: convert to gdbus Dec 23 18:50:39 freesmartphone.org: 03mickey 07cornucopia * r7fd0841da98c 10/fsodeviced/src/plugins/gpio_input/ (Makefile.am plugin.vala): fsodeviced: convert gpio_input to gdbus Dec 23 18:55:21 wrt. n900, do we need a watchdog-keepalive plugin? Dec 23 18:58:02 mickeyl: we had success with watchdog too... but maybe our own plugin would even be better Dec 23 18:58:15 to be able to act on stuff... Dec 23 18:58:17 yeah Dec 23 19:05:37 plugin.vala:754.56-754.70: error: The type name `Posix.socklen_t' could not be found Dec 23 19:08:19 better disable this plugin in this branch Dec 23 19:24:55 ok, that will be a little bit of work Dec 23 19:25:06 don't think I manage to finish it today Dec 23 19:25:23 but I will try to do it within the next days on the road Dec 23 19:35:24 mickeyl: does gdbus libfso-glib make use of properties? Dec 23 19:47:13 mrmoku: u-boot fix for gcc-4.5 is close :) Dec 23 19:48:03 mrmoku: no, nothing of FSO has a property Dec 23 19:48:09 i have been reluctant to use them Dec 23 19:49:53 mickeyl: should I bump vala version to 0.11.2.1 from downloads.fso.org? Dec 23 19:50:13 JaMa|Off: yes, please! Dec 23 19:50:24 please keep the one before though Dec 23 19:50:33 the new one has more strict syntax Dec 23 19:50:38 and won't compile some of the older things Dec 23 19:50:54 JaMa|Off: great Dec 23 19:51:02 mickeyl: good... one thing less too learn :P Dec 23 19:51:35 hehe Dec 23 19:51:47 i found method calls to be sufficient Dec 23 19:51:54 i have yet to see a compelling argument for properties Dec 23 19:56:37 freesmartphone.org: 03mickey 07cornucopia * raeee7f7dacb0 10/fsodeviced/src/plugins/player_gstreamer/plugin.vala: fsodeviced: player_gstreamer: catch up with vala Dec 23 20:03:38 hm, on the openmoko devices, do we want to be able to change the dai mode while a stream is open? Dec 23 20:04:38 DocScrutinizer: ^ Dec 23 20:05:09 Also TAsn might be interested, at least he understands the question :) Dec 23 20:05:28 hm? Dec 23 20:05:44 DocScrutinizer: do you have an opinion wrt changing dai mode while the stream is open? Dec 23 20:06:06 booom Dec 23 20:06:31 I do not see any possible usecase yet. Though probably if one want to have PA constantly running and having the device permanently open etc etc. Dec 23 20:06:38 dinner, ttyl Dec 23 20:07:39 PaulFertser: yep. thats what i though, too. one process that keeps a open stream could be blocking the dai from ever being changed Dec 23 20:10:11 on OM2007.2 there was pulseaudio running and having permanently opened the device Dec 23 20:11:13 so PaulFertser's idea can be the answer ;) Dec 23 20:19:00 the dai mode is what exactly? Dec 23 20:19:53 the routing of bluethooth and SoC audio streams Dec 23 20:20:11 http://wiki.openmoko.org/wiki/Neo_1973_audio_subsystem Dec 23 20:32:29 freesmartphone.org: 03mickey 07cornucopia * rf231d7b8eee3 10/fsogsmd/src/plugins/ (8 files in 2 dirs): fsogsmd: more gdbus conversion Dec 23 20:32:30 freesmartphone.org: 03mickey 07cornucopia * r9dc9db3aaa3e 10/fsotdld/tests/Makefile.am: fsotdld: gdbus Dec 23 20:39:02 PaulFertser, ? Dec 23 20:39:27 TAsn: you had enough experience with wolfson to answer larsc's question. Dec 23 20:40:21 enough experience is quite and overstatement Dec 23 20:40:25 sec, let me read his q Dec 23 20:40:34 freesmartphone.org: 03mickey 07cornucopia * rab89284af760 10/fsodeviced/src/plugins/ (5 files in 5 dirs): fsodeviced: use registerWithPrefix instead of manually keeping track of the counter Dec 23 20:41:53 hm.. thinking about it Dec 23 20:44:18 mrmoku: you can try u-boot from OE :) Dec 23 20:44:34 PaulFertser, after thinking about it. Dec 23 20:44:39 PaulFertser, I came to a conclusion Dec 23 20:44:51 PaulFertser, I don't know enough about the implications Dec 23 20:44:57 of having to close+open the stream Dec 23 20:45:02 because switching live Dec 23 20:45:05 is obviously needed Dec 23 20:45:13 but can't you just close+open it immediately? Dec 23 20:45:34 (obviously needed: lets say if you want to switch to bt, but that also depends on what's connected there and how) Dec 23 20:46:18 or in other words Dec 23 20:46:22 enough experience is quite and overstatement Dec 23 20:46:22 :) Dec 23 20:48:48 you canmt chanfe DAI on the fly. PCM streams will get outa sync Dec 23 20:49:01 dos1: did you build newer cornucopia in shr-u feeds? Dec 23 20:49:32 and keeping a stream open is a gashead idea regarding power conservation Dec 23 20:49:50 JaMa: yup, but only few commits newer Dec 23 20:50:14 dos1: i'm calling make update to build u-boot for n900, so it will revert your uncommited changes Dec 23 20:50:17 there was important commit for gpio-input module on n900 Dec 23 20:50:27 JaMa: i'll send it to ml Dec 23 20:52:20 or we can bump it to gdbus rev later Dec 23 20:52:22 DocScrutinizer51, you are just proving my point :) Dec 23 20:53:02 JaMa: I fear I won't manage to try another u-boot Dec 23 20:53:44 mrmoku: no problem it works here, so I guess it will work everybody Dec 23 20:53:56 JaMa: good news :D Dec 23 20:54:09 do we have an installation doc for u-boot on n900? Dec 23 20:55:24 not realy, it's similar to docs from meego/debian, but I should at least describe env stored in current u-boot binary Dec 23 20:56:03 will do, but not today.. my fingers are freezing here, not good for typing :/ Dec 23 20:56:12 *nodÜ Dec 23 20:56:18 * mickeyl launches an OE build from scratch and goes away for tonite as well Dec 23 20:56:30 JaMa: btw. I disagree with your conf :P Dec 23 20:56:45 JaMa: well, now i don't care, as i already tested what i wanted :) Dec 23 20:56:52 I strongly distaste the thing booting to maemo just because I'm not fast enough to open the keyboard ;) Dec 23 20:57:15 mrmoku: ah, right I forgot that one Dec 23 20:57:41 :) Dec 23 20:57:53 mrmoku: but we should agree also on partition layout etc, because what I pushed is just how I use it Dec 23 20:58:07 JaMa: btw. I started to convert our stuff to gdbus Dec 23 20:58:09 * JaMa doesn't say it's best scenario Dec 23 20:58:24 * mrmoku has shr on first partition and the kernel in /boot there Dec 23 20:58:34 which has the advantage that upgrading kernel just works Dec 23 20:58:53 (gdbus) I won't finish before leaving, but will work on it the next days Dec 23 20:59:01 and push when I'm done Dec 23 20:59:18 but won't work for ie btrfs rootfs with meego Dec 23 20:59:35 do you use swap? Dec 23 20:59:38 no Dec 23 20:59:48 and for meego we can have something different Dec 23 20:59:59 something like: Dec 23 21:00:08 first partition shr including kernel in boot Dec 23 21:00:14 second partition meego rootfs Dec 23 21:00:20 third partition meego kernel Dec 23 21:00:35 and optional boot.scr on 3rd Dec 23 21:00:43 yeah, that would be fine Dec 23 21:00:52 and I think flexible enough Dec 23 21:01:01 1st and 3rd ext2/3? Dec 23 21:01:27 yup Dec 23 21:01:41 ~hail gdbus Dec 23 21:01:41 * apt bows down to gdbus and chants, "I'M NOT WORTHY!!" Dec 23 21:01:47 it is so much nicer to use :D Dec 23 21:01:52 GNUtoo doesn't like vfat but keeping vfat on 3rd would be a bit more compatible with meego layout Dec 23 21:01:53 ok Dec 23 21:02:02 JaMa: fine for me too Dec 23 21:02:10 so how to install u-boot on n900 and boot SHR with our kernel? ;) Dec 23 21:02:25 i think it's time to finally do that ;) Dec 23 21:02:26 ok, then I'll change env tomorrow while preparing docs Dec 23 21:02:38 good Dec 23 21:03:35 dos1: https://elektranox.org/n900/kernel/uboot.html Dec 23 21:05:29 JaMa: thanks :) Dec 23 21:13:09 ~hail gdbus= Dec 23 21:13:13 mrmoku, anything interesting? Dec 23 21:13:38 DocScrutinizer51: so i should only ever change the dai mode if the codec is inactive? Dec 23 21:15:06 larsc, I'm expecting to see a super amoled screen + tegra :P Dec 23 21:16:35 TAsn: yeah, indeed Dec 23 21:16:55 TAsn: fso has switched to gdbus today Dec 23 21:17:03 and that means we will switch too Dec 23 21:17:08 if we want or not :P Dec 23 21:17:14 though we really want Dec 23 21:17:18 SHR: 03seba.dos1 07libphone-ui-shr * rea5bf5746fc7 10/src/widget/elm_keypad.c: elm_keypad: add missing W under 9 Dec 23 21:17:18 much nicer to use Dec 23 21:18:43 TAsn? Dec 23 21:20:48 larsc: i agree with DocScrutinizer51, there's zero reason to keep the stream open all the time, so changing while idle is both reasonable and easily doable. Dec 23 21:23:21 mrmoku, obviously. Dec 23 21:23:27 mrmoku, should be way better iirc Dec 23 21:23:33 especially when it comes to signals Dec 23 21:23:35 larsc: yes. Much the same like you should swap HDD only when no write access and preferrably only after proper umount Dec 23 21:30:00 TAsn, ha! you're here, so you should have read my memo :P Dec 23 21:33:53 mrmoku: http://paste.pocoo.org/show/309068/ Dec 23 21:36:31 for the brave: http://metafoo.de/s3c24xx-openmoko-gta02-2.6.37-uImage Dec 23 21:40:29 * JaMa is brave enough only to build it himself Dec 23 21:41:21 <[Rui]> larsc: what to expect from that new kernel? :) Dec 23 21:42:01 [Rui]: lots of WSODs Dec 23 21:44:21 i've mainly fixed regressions introduced in upstream so far, so theres not really something new featurewise in the kernel Dec 23 21:44:46 <[Rui]> larsc: I was hoping for nilch WS :) Dec 23 21:47:33 JaMa: + "loadbootscript=mmc init; ext2load mmc 0:1 ${loadaddr} boot.scr\0" \ Dec 23 21:47:43 shouldn't that be 0:2 ? Dec 23 21:48:21 mrmoku: debian has it on 1st ext2 Dec 23 21:48:47 ok, fine too Dec 23 21:48:59 * mrmoku ment 0:3 anyway :P Dec 23 21:49:07 but on first is fine too Dec 23 21:49:32 or we can support both like this one http://gitorious.org/u-boot-shr/u-boot/commit/b03d608109e18f10b8b697e281f84cfd76674e0a Dec 23 21:50:10 mmc 2:1 ? Dec 23 21:51:11 pespin, honestly: I missed it. Dec 23 21:51:12 mmnt, reading. Dec 23 21:51:21 pespin, no memo. Dec 23 21:51:51 JaMa: mrmoku: there MUST be a better way for ubot to get his envs than via hardcoded shit in binary that can't ever be changed excetp by new compile run! Dec 23 21:52:17 TAsn, just to remind you to look at phoneui+sms-report ;) Dec 23 21:52:38 lol, you know I honestly forgot :))) Dec 23 21:52:44 I know Dec 23 21:52:45 me too Dec 23 21:52:51 but really Dec 23 21:52:55 won't happen this weekend :| Dec 23 21:53:01 maybe saturday night Dec 23 21:53:03 having to remind you on thursdays is weird for me hehe Dec 23 21:53:05 if you remind me. Dec 23 21:53:22 ok will try to remember myself :P Dec 23 21:53:33 JaMa: mrmoku : I bet you can teach uboot to try and load envs form a file on either VFAT (MyDocs) or NAND rootfs, or even learn to rad out env from CAL partition Dec 23 21:53:40 pespin, sorry :(( Dec 23 21:53:42 read* Dec 23 21:54:52 TAsn, np :) Dec 23 21:54:58 mrmoku: didn't you say you planned to look into uBoot's watchdog kicking feature? Dec 23 21:57:20 DocScrutinizer: that's what that default env does, tries to find boot.scr and only if there is no boot.scr it continues with this default env Dec 23 21:57:26 DocScrutinizer: or when slide is closed Dec 23 21:57:27 honestly, has anybody done the fancy and looked into a raw dump of CAL partition and tried to understand if it maybe is exactly the same structure uBoot uses for his env partition? Dec 23 21:59:24 DocScrutinizer: (watchdog kicking) yup, but that was on hold, because of build problems Dec 23 21:59:29 (which are now solved thanks to JaMa) Dec 23 22:01:04 (CAL) would be a nice double kill with one shot: uBoot regains its ability to saveenv, and same strike we can tweak CAL settings without specialized flasher tools or simlar shit Dec 23 22:02:35 ***WARNING!!*** for experienced developers only. Users stay away, you can *really* brick your device Dec 23 22:04:34 http://talk.maemo.org/showthread.php?t=20465 Dec 23 22:04:58 for background - if you feel there's a bit of truth in my rant Dec 23 22:09:41 DocScrutinizer: I'm leaving for italy tomorrow morning... won't be able to touch u-boot this year Dec 23 22:10:22 I'll survive this bad news :-D Enjoy! Dec 23 22:11:28 Anyway maybe you can read a few bitsa, so you can tell a few new interesting details about uBoot, CAL/config, and watchdogs, after you are back from Italy :-) Dec 23 22:19:02