**** BEGIN LOGGING AT Tue Sep 09 02:59:57 2008 Sep 09 04:32:41 hi all. I just did a "make openmoko-devel-image" with all default settings. The only change was IMAGE_FSTYPES = "jffs2 tar" in local.conf. Now, the rootfs created is of qtopia. Isn't it supposed to be an OM image by default? qtopia has its own make target, right? Sep 09 04:36:15 hil__: http://wiki.openmoko.org/wiki/MokoMakefile#Environment Sep 09 04:36:29 Not sure which is the default now Sep 09 04:39:40 hnn... i thought OM was default. but lemme check Sep 09 04:43:06 ok. it seems like om2008.8 uses qtopia Sep 09 04:43:14 so, i guess that's not a problem. Sep 09 04:43:59 (the fact remains that my X-server deosn't start on this filesystem ... but I'll get back on that ) Sep 09 09:08:35 hi Sep 09 09:09:37 Hey there. Sep 09 09:11:27 i've been quite busy lately, any shr news? Sep 09 09:11:44 Don't know, I was off for a time. Sep 09 09:18:30 Ainulindale: i'm assigning shr tasks in trac as they are in the wiki (so we can remove them), you might receive some mail :-) Sep 09 09:19:09 I can't access my mail right now.? Sep 09 09:19:14 -? Sep 09 09:19:41 :-) Sep 09 09:29:03 roh: thanks for the trac updates Sep 09 09:45:17 Ainulindale: Good to hear your still alive Sep 09 09:46:10 Ainulindale: I sent you a mail regarding how to get stuff into lfg Sep 09 09:48:25 AndreasD: hey. do you have a trac user? Sep 09 09:51:12 aleix: on docs.openmoko.org yes, andreas.dalsgaard Sep 09 09:53:16 AndreasD: great. i'm assigning the shr tasks, so we can clean the wiki and use the trac instead Sep 09 09:53:35 aleix: nice :-) Sep 09 09:54:19 AndreasD: someone has to do it :-) Sep 09 09:54:47 my moko build is of latest version (2008.08). when im trying to build a sample helloworld application using bitbake method, im getting a error like " sqlite3.OperationalError: attempt to write a readonly database " . can anyone help me in solving this error?? Sep 09 09:55:19 the error is in the URL: http://rafb.net/p/reOK6A92.html Sep 09 09:57:09 morning Sep 09 10:01:03 AndreasD: did you receive two emails from two tasks? Sep 09 10:08:07 Hey Sep 09 10:08:31 wurp2|working, what's the state of the dialer3? :) Sep 09 10:10:09 Ainulindale, what about your ncurses dialer? Sep 09 10:21:16 aleix: Yes Sep 09 12:15:43 atm fso with zhone is unable to accept incoming calls, right? Sep 09 12:26:32 Hi. What are the reasons for choosing the "whole" Debian instead of an Emdebian flavour ? Sep 09 12:26:44 OdyX: have you read the mailing lists? ;) Sep 09 12:26:53 welll parts of Sep 09 12:27:05 its a pretty big bunch of english text Sep 09 12:27:20 yes, grep helps Sep 09 12:27:23 but if you say the info relies somewhere inside, I'll dive into them Sep 09 12:27:25 ;) Sep 09 12:28:35 OdyX: yes, in the mailing lists ;) Sep 09 12:29:10 thanks Sep 09 12:29:11 ;) Sep 09 12:34:11 lack of perl ? Sep 09 12:34:13 pfff Sep 09 12:36:00 python.. Sep 09 12:53:04 hi Sup3rkiddo Sep 09 13:05:33 playya, yo Sep 09 13:09:24 how are you Sup3rkiddo Sep 09 13:09:44 playya, doing good, thanks...et vous? Sep 09 13:11:11 had a problem wtih the new ubuntu kernel last week so i upload my tarbal tonight :( Sep 09 13:12:01 but a week without pc was fine too Sep 09 13:12:08 ooooo Sep 09 13:12:40 looks like mine is the the smallest :D, goes to show who was slacking off the entire summer ;) Sep 09 13:13:38 my first one was about 5MB Sep 09 13:13:57 but i missed to delete .c .h .o files :P Sep 09 13:14:09 hehe Sep 09 13:30:24 mickeyl, evening, i fixed the CPU usage issue and all is well.. zhone just works. (automatic screen lock et all..) can i go ahead and do a 0.0.1 release after milestone3 Sep 09 13:31:03 amazing! yes, absolutely. trying to get ms3 out today or tomorrow, then you can tag 0.0.1 in git and release. :) Sep 09 13:31:50 yay, thanks Sep 09 13:34:39 Sup3rkiddo, releasing what with tag 0.0.1? :) Sep 09 13:35:20 quickdev, my gsoc code. odeviced Sep 09 13:37:27 Sup3rkiddo, it replaces the frameworkd device part? Sep 09 13:39:11 quickdev, na, doesnt replace it..just another implementation in C/vala... Sep 09 13:40:18 quickdev, although i have been using it for sometime now as a replacement of the python subsystem.... Sep 09 13:41:44 quickdev, http://git.freesmartphone.org/?p=openmoko-gsoc2008.git;a=summary if you are curious or want to have a few laughs ;o Sep 09 13:42:48 Sup3rkiddo, I know the repository, already browsed through it :) Sep 09 13:43:07 watched the commits ;) Sep 09 13:43:47 quickdev, oh, ok..i am just a novice, so there is bound to be crappy code..if you find anything, please ping me Sep 09 13:46:55 Sup3rkiddo, I assume it's faster than the python implementation? Sep 09 13:47:46 quickdev, i think so, but i do not have any solid numbers though Sep 09 13:49:09 it provides the same functionality as the frameworkd part odeviced? So I could disable the odeviced in my frameworkd and run your daemon, right? Sep 09 13:49:11 yes, odeviced should be a bit faster Sep 09 13:49:22 however don't expect order of magnitudes Sep 09 13:49:33 it's merely reading/writing from file descriptors Sep 09 13:49:39 plan is twofold Sep 09 13:49:41 yeah Sep 09 13:49:47 make oframeworkd load c plugins Sep 09 13:49:53 make framework-ng load py-plugins Sep 09 13:49:57 both should be very interesting Sep 09 13:50:12 i think much of the time is spent in actually setting up everything up-front Sep 09 13:50:22 (in python that is) Sep 09 13:50:26 yeah Sep 09 13:50:45 although we gain something regarding power consumption w/ the vala implementation Sep 09 13:50:47 after that, both are almost equally fast...except for less memory usage of course Sep 09 13:50:56 that would mean a side effect is a saving of boot time? ;) Sep 09 13:51:04 boot time.. not reallly Sep 09 13:51:09 time is wasted elsewhere there Sep 09 13:51:14 yeah maybe 100 nanosecs :D Sep 09 13:51:17 need to stop booting like a desktop Sep 09 13:51:21 start booting like a mobile device Sep 09 13:51:28 but apparantly hardly anyone interested in that Sep 09 13:51:29 *sigh* Sep 09 13:51:56 hmm...looks like it is quite hard Sep 09 13:52:09 depending on the amount of backwards compatiblity, no Sep 09 13:52:18 someone needs to write a small linuxrc Sep 09 13:52:30 that sets up devices, mounts filesystems, launches dbus Sep 09 13:52:34 then boot into X Sep 09 13:52:36 _done_ Sep 09 13:52:40 mickeyl, i might be way off, but how about openrc? Sep 09 13:52:43 it uses C Sep 09 13:52:49 instead of shell scripts Sep 09 13:52:49 interesting, never heard about that one Sep 09 13:52:59 it powers gentoo Sep 09 13:54:58 "It's fast - 29 seconds to boot my laptop running Gentoo/Linux" Sep 09 13:55:22 maybe we could strip out some bloat Sep 09 13:55:23 a big advantage is that processes are handled simulatenously Sep 09 13:55:39 hald and friends take a significant time here in my lappy Sep 09 13:56:26 but shuts down awesomely fast, 5-7 seconds and poof Sep 09 13:57:16 I'd be curious how difficult it would be to set openrc up with our images Sep 09 13:58:05 quickdev, theres also OE to take care of. it might need some patches to work with openrc as well? Sep 09 13:58:21 lots of, probably Sep 09 13:59:10 How is that working? Submitting to their tracker and they apply it? Sep 09 14:00:13 "# Works with POSIX shell scripts and userland... # Works with existing init" Sep 09 14:02:08 http://www.gentoo.org/doc/en/openrc-migration.xml Sep 09 14:31:28 runit and minit seems to be an alternative, too Sep 09 14:35:52 mickeyl, you think giving a different init system a try isn't too difficult? Sep 09 14:51:12 not if you're going the simple way Sep 09 14:51:21 if you want to have a new one w/ compat support, you're doomed Sep 09 14:51:23 that's a lot of work Sep 09 14:51:31 but if you're going a simple init script Sep 09 14:51:48 w/ the possibility to just use sysvinit, if some power users really want Sep 09 14:51:54 then this should be pretty simple Sep 09 14:52:11 i.e. you could use update-alternatives for init Sep 09 14:52:19 and have the kernel start your simpleinit Sep 09 14:52:38 then you just do the things necessary to get dbus and X up Sep 09 14:52:49 and leave the rest to someone else Sep 09 15:28:58 freesmartphone.org: 03sudharsh 07openmoko-gsoc2008 * ra4426e45879b 10/odeviced/src/plugins/idlenotifier/idlenotifier.plugin: do not ignore touchscreen events Sep 09 15:28:58 freesmartphone.org: 03sudharsh 07openmoko-gsoc2008 * r468da20ada3b 10/opimd/ (10 files in 2 dirs): Merge branch 'master' of git@git.freesmartphone.org:openmoko-gsoc2008 Sep 09 15:46:20 mickeyl, do you know how to pass the init= variable to my kernel in uboot? Sep 09 15:46:52 yeah Sep 09 15:47:00 you need to go into the u-boot command line mode Sep 09 15:47:03 and then change the bootargs Sep 09 15:48:20 freesmartphone.org: 03mickey 07zhone * raa4871b45c2c 10/src/zhone: improve sending SMS stability, we now regularly try to send unsent messages Sep 09 15:49:33 mickeyl, bootargs? I did (obviously something wrong). Sep 09 15:50:33 how does your environment look like? Sep 09 15:50:55 the best way may be to permanently add something like ${XTRA_BOOT} to bootargs_base Sep 09 15:51:04 which usually is empty Sep 09 15:51:14 and when you want to test you can set it to something meaningful Sep 09 15:51:16 and call boot Sep 09 15:51:47 it's working, yes, I modified bootargs_base Sep 09 15:53:14 now I'm trying to set up runit Sep 09 15:55:11 cool Sep 09 15:55:13 good speed! Sep 09 15:55:17 keep us posted Sep 09 16:20:45 freesmartphone.org: 03jluebbe 07zhone * rd0780d613df2 10/src/zhone: show incoming sms Sep 09 16:27:44 freesmartphone.org: 03jluebbe 07framework * r6544f1741f8b 10/etc/freesmartphone/oevents/rules.yaml: oeventsd: switch to gsmhandset while the state is outgoing Sep 09 16:41:55 alphaone, is it possible to accept an incoming call in fso somehow? Sep 09 16:49:55 quickdev: zhone does it? Sep 09 17:00:08 freesmartphone.org: 03sudharsh 07openmoko-gsoc2008 * r12da14f4e01d 10/odeviced/src/plugins/idlenotifier/ (idlenotifier.plugin idlenotifier.vala): Ensure that the device does not fall into SUSPEND state except when the DBus interface is used Sep 09 17:24:17 mickey|bbl, do you know where sshd is started? it's quite hidden Sep 09 17:27:44 quickdev, its dropbear in FSO and other iirc Sep 09 17:27:55 aah :) Sep 09 17:44:45 When starting it seems frameworkd and its python process is using most resources.. Sep 09 18:44:11 mickey|bbl, saveenv is quitting after "Erasing nand..." and changes aren't applied, you know that issue? Sep 09 18:49:10 hi Sep 09 18:49:26 unable to call a cellular phone with fso-testing Sep 09 18:49:39 in fact it works at recipient side Sep 09 18:49:52 but i can't hearing anything on MY side Sep 09 18:50:12 (like ring tonality and even voice) Sep 09 18:50:48 common issue already pointed out ? Sep 09 18:50:52 unable to call a *cellular* phone? Sep 09 18:50:58 as in, calling a landline works? Sep 09 18:51:25 works well on landline network Sep 09 18:51:32 lambda: Did you try using alsaconf to load the gsmhandset sound config manually? Sep 09 18:51:44 wurp2|working, no Sep 09 18:51:57 I didn't realize there was any difference at all on the FR end between calling a landline phone and calling a cell phone Sep 09 18:52:08 My guess would be an intermittent problem setting the alsa conf Sep 09 18:52:19 i am little bit confused and baffled Sep 09 18:52:50 thank for suggestion Sep 09 18:52:52 well, it is fso-testing Sep 09 18:53:01 not a MS release :-/ Sep 09 18:53:04 yw Sep 09 18:53:27 i have to open a bug ? Sep 09 18:53:33 on fso tracks Sep 09 18:55:05 lambda: I'm not sure they want bugs on things in fso-testing... maybe so, I guess it's not head Sep 09 18:55:31 I'm not sure what the policy is. I know mickey doesn't claim everything will work if you grab from head at any given time Sep 09 18:55:46 (and it would be foolish to claim that on any project, really) Sep 09 18:56:35 But the name of fso-testing implies that maybe they want bug reports Sep 09 19:00:28 wurp2|working, may I ask you about the status of the dialer? Sep 09 19:19:55 quickdev: Same as last time, I think :-( Sep 09 19:20:16 pin code entry works, I think making the call works, but then you don't get the 'talking' dialog, or it's incomplete Sep 09 19:20:17 wurp2|working, I thought about an efl dialer, what would you think about that? Sep 09 19:20:36 At any rate, you can't end the call or enter tones Sep 09 19:20:44 quickdev: I think that would be great Sep 09 19:20:54 But I think it would be easier to do it as a fork of zhone Sep 09 19:21:30 IMO, the biggest thing is being able to dial from contacts Sep 09 19:21:36 (the biggest thing missing from zhone) Sep 09 19:23:15 yeah.. Sep 09 19:23:54 I made some C testing code for a dialer user interface, coding with efl is quite straightforward Sep 09 19:24:17 wurp2|working, what about libframeworkd-glib - is it ready for use? Sep 09 19:24:47 quickdev: yep Sep 09 19:25:11 afaik it is complete... someone should do some heavy duty testing, though. Sep 09 19:25:35 coding with it would be heavy duty testing ;) Sep 09 19:25:49 quickdev: Hmm, you could use the existing libphonegui as example code Sep 09 19:26:01 The part that's done there is the pin code entry & dialing, iirc Sep 09 19:26:35 wurp2|working, where do I find that lib? Sep 09 19:27:40 let me get you the right name Sep 09 19:27:48 It's in the same svn repo as lfg Sep 09 19:28:19 libframeworkd-phonegui-gtk is the right name Sep 09 19:28:25 it's a sibling directory to lfg Sep 09 19:29:55 yeah, but where is the repository? ;) Sep 09 19:30:40 You would also need to look at ophonekitd, to see how to handle incoming dbus events Sep 09 19:30:48 lfg is all asynchronous Sep 09 19:31:01 it's on the SHR wiki page or the SHR Development one Sep 09 19:31:08 ok Sep 09 19:31:11 daria.fortytwo or something.. sec Sep 09 19:31:34 svn://daria.forty-two.fr/shr/trunk/shr Sep 09 19:34:52 thanks for the hints Sep 09 19:35:28 yw Sep 09 19:35:39 Thanks for working on an FSO dialer :-) Sep 09 19:55:58 quickdev: if you are going to use libframeworkd-glib you should use the CallStatus handle to listen for when to bring up the GUI Sep 09 19:56:37 however I could not get CallStatus handler to work, I've got a fix for it Sep 09 19:56:38 fwHandler.callCallStatus = NULL, yes Sep 09 19:56:40 but it really should go in the git repo not in shr-svn Sep 09 19:57:07 well the problem is in the marshalling Sep 09 19:57:51 yes, take a look at ophonekitd Sep 09 20:00:02 quickdev: are you going to try out the code now? Sep 09 20:00:17 I can commit the in svn? then I'll ask Ainulindale to sync later Sep 09 20:01:03 will start hacking the call part the next days..would be fine if the bug is fixed then ;) Sep 09 20:08:54 quickdev: It's in svn now Sep 09 20:09:38 AndreasD, and it will be synced to git soon? Sep 09 20:11:23 I wrote a mail to Ainulindale last week, but for now just use the svn repo, there isn't anything new in git Sep 09 20:12:16 I have to be going now, see you tomorrow Sep 09 20:12:32 freesmartphone.org: 03jluebbe 07zhone * rbe3d72c6e5f5 10/src/zhone: Sep 09 20:12:32 freesmartphone.org: fix ordering in cbCallStatus and add dbus_error logging helper Sep 09 20:12:32 freesmartphone.org: SetStatus failed, but the exception was ignored because cbCallStatus is Sep 09 20:12:32 freesmartphone.org: called from a dbus signal handler. Sep 09 20:23:32 freesmartphone.org: 03jluebbe 07framework * r248637c12646 10/framework/subsystems/odeviced/idlenotifier.py: odeviced: fix SetState Sep 09 20:46:09 quickdev: Do you know which git repo AndreasD was talking about? Sep 09 20:47:00 He talked about the git repository of libframeworkd-glib on git.freesmartphone.org and svn://daria.forty-two.fr/shr/trunk Sep 09 20:47:51 I wasn't aware there was a git repo for lfg Sep 09 20:47:58 I am way too out of the loop lately Sep 09 20:50:58 wurp2|working, it was moved recently afaik Sep 09 20:53:43 wurp2|working, quickdev: FSO team offered this to keep it closer to the rest of the FSO software. We hope to follow API changes in the framework faster this way. Sep 09 20:53:50 wurp2|working: I'm with you on the "out of the loop" one Sep 09 20:53:53 We will see if it works out. :) Sep 09 20:55:14 abraxa_: Join he loop by integrating pimd into fso :D Sep 09 20:56:23 I would if you could magically make my days 32 hours long Sep 09 20:56:24 ~slap YHA australia site for using flash for their booking site Sep 09 20:56:43 abraxa_: Well, just change your sleep cycle :) Sep 09 20:57:05 hmm, you guys have no bot here Sep 09 21:02:21 stefan_schmidt: excellent - I much prefer git anyway (at least intellectually, given that I have almost 0 experience with it) Sep 09 21:02:34 wurp2|working: heh Sep 09 21:02:58 I have used darcs a fair amount and arch a bit, so I know the advantages of distributed repositories Sep 09 21:04:20 ok, never used them. Directly from svn to git Sep 09 21:05:19 well, a bit mtn for OE in between. But I used that one more in svn-mode. Nothing more then mtn pull && up, diff, commit and commit && push Sep 09 21:06:34 I actually like darcs better than git from the little bit I know about git, but that may just be lack of knowledge of git, and I certainly understand that with Linus behind git, it's gonna win regardless of merits :-/ Sep 09 21:08:31 wurp2|working: http://git.or.cz/gitwiki/GitCheatSheet might come in handy then Sep 09 21:08:39 abraxa_: awesome, thanks! Sep 09 21:09:48 Very nice; bookmarked Sep 09 21:10:24 You're welcome Sep 09 21:28:21 ooooh, git cheatsheet, just what i needed, thx :] Sep 09 22:42:32 i want to get the full contents of the console (i.e. all of the kernel startup messages). is NeoCon the tool i need for that? (assuming it's even possible without a debug board) Sep 09 22:42:59 i'm having wifi problems and i suspect a kernel issue, but i think the relevant messages are scrolling off the top and aren't available in 'dmesg' later Sep 09 22:45:35 Weiss: Quick'n'dirty solution: put a "dmesg > /log" in one of the init.d scripts that is started rather early Sep 09 22:47:40 Weiss: dmesg captures everything except very early (decompressing the kernel) messages -- at least those issued by a "printk" in the kernel. The dmesg buffer should be large enough if you read it soon enough after boot. Sep 09 22:48:57 thanks abraxa_ and mwester. however, there are a few screenfuls of errors which go past, which i think fill it up. hang on - i'll paste Sep 09 22:50:51 here's what's in dmesg immediately after booting: http://paste.ubuntu.com/45065/ Sep 09 22:51:45 i suspect the crashiness at [ 10.260000] and the messages above are from the wifi driver, but i'm ready to be told otherwise. wifi doesn't work at all for me - eth0 just doesn't exist Sep 09 22:52:01 (this is 2008-update+opkg upgrade) Sep 09 22:53:28 everything else is working nicely :) - calls, sms (even while suspended), resume, GPS, .. Sep 09 22:54:36 Yep, it's broken. Sep 09 22:55:40 Check the bug database, and if you can't find something that matches your symptoms, open a new bug. But be sure to note exactly which kernel and firmware you are using, and how someone at Om can setup a device to duplicate the problem. Sep 09 22:56:41 (although there's no guarantee that this kernel problem is the cause of your wifi problems; just that it seems likely) Sep 09 22:59:36 sure, just wondering if there's anything i can do (sans debug board) to get more information Sep 09 23:04:01 There's nothing more to get, even with a debug board. Sep 09 23:04:43 You can try to track the null pointer dereference in the stack dump back to the kernel source, but that isn't done on the device in any way... :( Sep 09 23:05:44 i mean... if i could get hold of the kernel messages from before those TIMEOUT messages begin? Sep 09 23:10:50 Hmm... that might find something else wrong, and I guess it's possible that might provide additional hints to debug the driver, but I think the dmesg captured the debug info. In any case, in order to capture that, you'll need to rebuild the kerne with a larger buffer for dmesg data, or use a debug board -- and probably not necessary in order to file a bug report. Sep 09 23:12:12 Is that output from the initial boot of the device, or from a resume after suspend? Sep 09 23:12:33 mwester: Can the message buffer be increased using a kernel parameter? Just wondering if you knew Sep 09 23:13:10 Don't know offhand -- possible, but I don't ever recall seeing that documented (although I've not looked for it either) Sep 09 23:13:21 mwester: that's from a fresh boot, before any suspending whatsoever Sep 09 23:13:47 The loglevel argument can control the verbosity, but I didn't see much in that log that could be suprressed easily. Sep 09 23:14:01 Weiss: I'm surprised nobody else has reported that then. Sep 09 23:15:20 yes, i was surprised as well... Sep 10 00:10:35 morning Sep 10 00:11:37 moo! Sep 10 00:11:46 man i hate compiler bugs Sep 10 00:11:53 sdcc: SDCCopt.c:680: convilong: Assertion `0' failed. Sep 10 00:11:53 Caught signal 6: SIGABRT Sep 10 00:15:46 brr Sep 10 00:15:57 Is this this sdcc http://sdcc.sourceforge.net/? Sep 10 00:16:05 * stefan_schmidt wonders what raster tries to compile Sep 10 00:16:38 yeah Sep 10 00:19:56 oh, already after 2am. Time to sleep. Later guys. Sep 10 00:23:27 nite! **** ENDING LOGGING AT Wed Sep 10 02:59:57 2008