**** BEGIN LOGGING AT Fri Sep 12 02:59:57 2008 Sep 12 03:39:19 * pjz admits he'll probably try it with a more stable kernel first Sep 12 07:14:39 morning Sep 12 08:32:36 mmontour: mokomakefile/openembedded/packages/u-boot/u-boot-mkimage-openmoko-native/ seems to have the openmoko specific patches for u-boot (as separate patch files) Sep 12 08:42:52 hi Sep 12 09:14:45 rwhitby: ping Sep 12 09:34:32 hi Sep 12 09:34:46 I have just updated my debian-gta02 Sep 12 09:35:01 I have now fso-m3 Sep 12 09:35:11 SMS do not work Sep 12 09:35:28 I have in zhone log : Sep 12 09:35:30 2008-09-12 11:25:19,734 INFO did not receive any unsent messages: org.freesmartphone.GSM.SIM.NotFound: Sep 12 09:35:36 any ideas? Sep 12 09:43:07 TeXitoi_: < gismo> nomeata: OK, Debian is just missing contacts, SMS is not working with upstream FSO Sep 12 09:52:18 lindi-: OK, so FSO bug Sep 12 09:52:27 but contect works here Sep 12 09:52:37 s/contect/contacts/ Sep 12 11:33:12 Hey Sep 12 11:34:44 quickdev: hello Sep 12 11:42:27 MarcOChapeau, do you have a minute for me? Sep 12 11:43:06 quickdev: depends on the question ;-) Sep 12 11:44:37 I'd like to implement a little daemon which starts a gui when the sim code is requested. When the system starts, frameworkd isn't ready immediately. I'd like my daemon to retry every 5 seconds to connect to the dbus interface, but the problem is, that libframeworkd-glib handles some certain errors with exit(). I'd like to catch them ;) Sep 12 11:45:03 ??? Sep 12 11:45:15 MarcOChapeau, I assume you are the programmer of libframeworkd-glib? Sep 12 11:45:26 you are describing whet we call ophonekitd Sep 12 11:45:31 it already exists Sep 12 11:47:17 yes, but it doesn't handle the situation when the dbus service is not available yet Sep 12 11:47:31 ? Sep 12 11:47:40 no dbus, -> no calls Sep 12 11:48:08 on an fso system you can considered the phone to be turned off until dbus is there Sep 12 11:48:51 MarcOChapeau, no calls, yes, but the current implementation of libframeworkd-glib aborts with exit(1) if there's no dbus and that's why I can't proceed and handle the error. Sep 12 11:49:41 you understand what my problem is? I'd like to handle the case when there's no dbus yet Sep 12 11:51:21 libframeworkd-glib isn't used unless ophonekitd is started. ophonekitd is started after dbus Sep 12 11:51:25 you have no problem Sep 12 11:52:20 ophonekitd is started after dbus and frameworkd, right, but frameworkd needs some seconds to initialize Sep 12 11:52:54 and a started frameworkd doesn't mean, that the dbus services are already initialized Sep 12 11:57:35 MarcOChapeau, or am I wrong? Sep 12 11:59:24 MarcOChapeau, zhone is approaching similar, it tries every 10 seconds to get a connection to frameworkd. That's what I'd like to reach with libframeworkd-glib. Sep 12 12:07:25 quickdev: I still think that's a non issue. not much time to look into it right now Sep 12 12:12:00 ok, would be fine if we could talk about it when you've got some time Sep 12 12:15:09 quickdev: try and take a look at dbus_connect_to_interface in frameworkd-glib-dbus.c Sep 12 12:16:28 quickdev: I think what you would like to do it to make lfg retry later if itf==NULL Sep 12 12:16:41 AndreasD, http://git.freesmartphone.org/?p=libframeworkd-glib.git;a=blob;f=src/frameworkd-glib-dbus.c;h=d5c5bde9dedeedb9b18cf2020e826a824b7f41b9;hb=HEAD, line 71 Sep 12 12:20:03 AndreasD, interesting...I think checking if callBus != NULL would be sufficient Sep 12 12:20:31 AndreasD, but that isn't very elegant, is it? ;) Sep 12 12:22:54 quickdev: no, callBus will not be set later and will remain NULL Sep 12 12:23:51 AndreasD, what I miss is a callback function that handles if dbus_connect_to_bus(); fails Sep 12 12:29:19 quickdev: wouldn't it be nicer if dbus_connect_to_bus() blocked until it was connected? Sep 12 12:30:28 AndreasD, for example if dbus is not started, it should somehow return an error, via a return function or a callback but it doesn't do that Sep 12 12:30:39 *return value Sep 12 12:31:52 quickdev: I think I get you, but why would you do that? Sep 12 12:32:46 AndreasD, if dbus_connect_to_bus() fails I'd like to wait some seconds and retry it as zhone does. That's why ;) Sep 12 12:33:50 quickdev: why not do that in lfg? Sep 12 12:34:26 AndreasD, because my application should be able to decide what do to if an error occurs? Sep 12 12:37:14 quickdev: hmm what exactly are your application doing? Sep 12 12:38:24 quickdev: If the only thing it does is to reconnect I think we should put it in lfg, to avoid code duplication Sep 12 12:38:42 AndreasD, it was meant conceptually. I think the application should be able to decide if it likes to retry to connect or to abort the whole thing or display a message for the user or whatever ;) Sep 12 12:41:45 quickdev: I think your are right conceptually :-) but I think the retry code should be in lfg Sep 12 12:42:32 AndreasD, would be ok for me in this case.. Sep 12 12:53:02 AndreasD, what about merging the old repository to the new one? Who was responsible for that? Sep 12 12:55:33 quickdev, I quess we just have to wait for Ainulindale or poke mickey until he give us access to the repo :-) Sep 12 12:56:02 ah, ok Sep 12 14:53:49 When using feeds from fso-testing on FSO-ms3, opkg upgrade segfaults ... Sep 12 14:54:37 OdyX, see if you can get a stack trace Sep 12 14:56:01 thos: I already had this with ms2 and I had gathered a solution in #openmoko (copy libopkg.so.0.0.0 in /usr/lib) so I have no more the problem Sep 12 14:56:12 sorry, it was just to make the devs aware ;=) Sep 12 14:59:26 OdyX: There are no functional feeds for FSO-ms3 Sep 12 14:59:41 mwester: OKay... wild guess on my part... Sep 12 14:59:51 mwester: but opkg still segfaults Sep 12 14:59:53 ;) Sep 12 15:01:26 OdyX, please get a stack trace for opkg, it will help us improve it Sep 12 15:08:28 thos: I'll do that next week.. I have a "real-life week-end" ;) Sep 12 15:09:18 Hi. Does anyone here know who to properly upload a new feature to the zhone git repository? As a feature branch, or boldly onto master? quickdev maybe? Sep 12 15:10:32 ping alphaone Sep 12 15:11:08 alphaone: ^ :-) Sep 12 15:11:38 It’s something that might need a quick review (visual feedback when suspending with the POWER button), so I’m reluctant. Sep 12 15:16:00 pong quickdev Sep 12 15:16:37 quickdev bounces alphaone’s pong to nomeata Sep 12 15:16:41 nomeata: Ah, I always need to lookup this stuff in the git manpages Sep 12 15:16:54 alphaone: I know how to do it, but what am I supposed to do? Sep 12 15:17:10 alphaone: is there an unwritten policy for git.fso.org? Sep 12 15:17:13 nomeata: What kind of new feature is it? Sep 12 15:17:43 If it's invasive and will probably break key features during development I'd use a branch Sep 12 15:17:48 As I did with sms-pdu Sep 12 15:17:49 alphaone: http://lists.openmoko.org/pipermail/community/2008-September/030329.html Sep 12 15:18:28 it touches how suspend workd, but will not break stuff completely, I think :-) Sep 12 15:18:33 *works Sep 12 15:18:51 nomeata: Sounds like there's no need for a branch Sep 12 15:19:03 alphaone: ok, thx Sep 12 15:19:07 np Sep 12 15:19:26 if mickey|zzZZzz screams I'll blame you ;-) Sep 12 15:20:36 no problem, I’m going to be in holidays for a few days from tomorrow on :-P Sep 12 15:22:21 hmm. CIA bot sends out only one of my patches. Sep 12 17:01:26 freesmartphone.org: 03mail 07zhone * r444fb114fb79 10/src/zhone: retab my recent code to follow fso style Sep 12 17:34:45 freesmartphone.org: 03sudharsh 07openmoko-gsoc2008 * r78e5cc381ac6 10/odeviced/README: modify README Sep 12 18:40:41 ehlo Sep 12 18:40:49 can someone point me to the source code of the gtk dialer on 2007.2? Sep 12 18:41:01 looking at the wiki, it only links to openembedded? Sep 12 18:42:53 downloads.openmoko.org/downloads maybe? Sep 12 18:42:59 argh - / sources Sep 12 18:45:19 b0ef: svn.openmoko.org/trunk/src/target/OM-2007.2/ Sep 12 18:47:22 abraxa_: thanks;) Sep 12 20:40:17 hi, i've installed xtightvncviewer on the freerunner .... cool :) but i only can scroll in the right corner of the host screen but not back... i gues this is a buck... now i hope for a fix or an alternativ vnc viewer - any tips? Sep 12 20:56:46 govinda: a buck? Sep 12 22:11:17 having build a uImage from the kernel sources checked out from Git, how do i go about handling the modules? Sep 12 22:12:11 built* Sep 12 22:20:23 "make install" with the correct variable defined to install the modules to your local fs rather than your host's /lib -- then tar up that directory and install it on the FR before you flash the new kernel. Sep 12 22:43:40 mwester: thanks (INSTALL_MOD_PATH=modules.tmp make modules_install) - knew there must be a better way than what i was trying! Sep 12 22:55:57 hmm.. my modules dir takes up about 50Mb compared to the 5Mb in my installed image. is there an obvious debug symbols option somewhere which i could turn off? Sep 12 22:59:03 Weiss: kernel hacking--> kernel debugging Sep 12 23:26:23 celtune1: much better, thanks (once i figured out you have to do ARCH=arm make menuconfig) - disabled CONFIG_DEBUG_INFO Sep 12 23:33:55 yw Sep 12 23:49:25 morning Sep 12 23:57:56 Where do I have to put a link to my application in order to get it listed on the "phone desktop" ? Sep 13 00:13:31 is this the right place to ask about toolchain errors? or toolchain choice in general? Sep 13 00:36:24 TAsn, try Sep 13 02:12:59 * torindel got armv4tl-softfloat-linux-gnueabi and armv4tl-softfloat-linux-uclibcgnueabi toolchains **** ENDING LOGGING AT Sat Sep 13 02:59:56 2008