**** BEGIN LOGGING AT Thu Oct 29 02:59:57 2009 Oct 29 04:42:38 zhenhua: It might be a good idea to leave that CLCC in, since the network is free to normalize the number according to its own rules Oct 29 04:42:59 zhenhua: So if CLCC is supported, don't bother saving 1 extra CLCC Oct 29 04:46:18 zhenhua: I pushed your patches, but refactored them somewhat Oct 29 04:46:54 zhenhua: Maybe I broke something, send fixes if that is the case. Also, I hate the voicecall_data:mpty_call and call variables Oct 29 04:47:17 zhenhua: They're just hacks Oct 29 04:48:17 zhenhua: And actually useless since you don't implement any commands that would allow a multi-call situation to happen Oct 29 04:53:41 denkenz: ok. i will use one extra CLCC for dialing Oct 29 04:54:36 denkenz: for voicecall_data:mpty_call, i may remove it Oct 29 04:55:05 denkenz: but for voicecall_data: call, we could leave it there to save one search from GSlist *calls. Oct 29 04:55:41 Right now you can only have 1 call anyway Oct 29 04:55:54 So calls is useless Oct 29 04:56:10 denkenz: yeah, but you need it for multicall Oct 29 04:56:16 once you have multiple calls, then call is useless Oct 29 04:57:01 yes Oct 29 04:57:15 but call always point to the current active call Oct 29 04:57:30 Which is useless in multicall ;) Oct 29 04:57:40 in single call, it saves one more search from gslist, why not Oct 29 04:58:01 heh, list travelsal of size 1 is just 1 Oct 29 04:58:04 It saves you nothing Oct 29 04:58:34 but i hate to search it Oct 29 04:58:46 you can assign it and it's more clear that search Oct 29 04:59:21 anyway, if you really don't like it, i am fine Oct 29 04:59:31 I know why you're doing it ;) Oct 29 04:59:52 Anyway I'm letting it slide since I know it'll change, but in general don't look too far ahead Oct 29 04:59:52 keep an id and keep a pointer are the same Oct 29 05:00:30 Submit the code that does what is intended with bare minimum Oct 29 05:00:38 okay Oct 29 05:00:42 You've bunch of extraneous stuff in that patch Oct 29 05:01:22 i know that, many are duplicated from atmodem so that they are for multi-call, like release_id Oct 29 05:01:38 id_list and local_release Oct 29 05:02:30 I know, don't worry too much, its more education for future patches Oct 29 05:02:32 i am testing my multi call, i cann't guarantee it works for all phone but i am testing it. Oct 29 05:02:41 I'm much easier to get stuff past than others ;) Oct 29 05:02:43 yes, i know Oct 29 05:02:51 ;) Oct 29 05:04:39 denkenz: again, about hfp_dial, if AG doesn't support CLCC, i still need to pass phone number to atd_cb() Oct 29 05:04:53 i think it should be fine. Oct 29 05:04:59 if it doesn't, that's fine Oct 29 05:05:03 we do the best we can Oct 29 05:05:06 okay Oct 29 18:46:23 denkenz: I'm able to get netreg properties, Is there any error here http://pastebin.com/d766cc023 ? Oct 29 18:47:26 looks fine to me, you might wanna print them Oct 29 18:48:47 yes, but getproperties is not working. I got that: dbus.exceptions.DBusException: org.freedesktop.DBus.Error.UnknownMethod: Method "GetProperties" with signature "" on interface "org.ofono.NetworkRegistration" doesn't exist Oct 29 18:49:02 denkenz: any idea? Oct 29 19:07:23 do you actually register a netreg atom? Oct 29 19:08:56 padovan: It sounds like you don't netreg_create / netreg_register Oct 29 19:12:10 denkenz: I do netreg_create at pre_sim. Oct 29 19:14:03 padovan: Then in your netreg driver, do you do netreg_register? Oct 29 19:14:26 No, I just checked that. Oct 29 19:14:29 denkenz: thanks. Oct 29 20:20:40 denkenz: Do I need to query registration status and operator name into netreg_probe? Or we left that only for the user query whenever it wants? Oct 29 20:21:03 padovan: The core should do that for you at startup Oct 29 20:21:53 Right. Oct 29 20:23:10 denkenz: I put ofono_netreg_register into netreg_probe, but I needed to change src/network.c to keep hfp working: http://pastebin.com/d6c6b4a5 Oct 29 20:23:15 take a look Oct 29 20:23:57 why are you doing that? Oct 29 20:24:25 Ok, I see, do what every other driver does and use g_idle_add to register Oct 29 20:24:40 denkenz: I'm getting netreg->driver == NULL inside netreg_register Oct 29 20:25:08 denkenz: Ok. Oct 29 20:25:22 padovan: Check drivers/atmodem/devinfo.c probe/register for details Oct 29 20:25:56 You can't register within probe, I assumed all drivers needed to do _some_ initialization Oct 30 01:10:18 denkenz: ping Oct 30 01:36:51 zhenhua: pong Oct 30 01:38:59 there's some inconsistent about call id type. Oct 30 01:39:20 in types.h, we have struct ofono_call { unsigned id; } Oct 30 01:39:43 but in atmodem, like at_release_specific(int id). we use int Oct 30 01:40:01 which one is correct Oct 30 01:46:07 denkenz: and I believe CHUP only release one active call in HFP. not the whole mpty call Oct 30 01:46:56 signed / unsigned doesn't matter, its always greater than 0 Oct 30 01:47:49 right, but compiler has warning when comparing, so we could align them Oct 30 01:48:04 Send a patch Oct 30 01:49:25 will do later Oct 30 01:50:45 change to int? Oct 30 01:52:56 unsigned, unless we -1 it somewhere Oct 30 01:55:16 okay Oct 30 01:56:35 Yikes, if you wanna see a really scary implementation, look in BlueZ's HFP AG one Oct 30 01:56:48 It does all sorts of stuff not according to spec Oct 30 01:57:02 like what? any example? Oct 30 01:57:06 Bet they certified it though :) Oct 30 01:57:26 gateway.c? Oct 30 01:58:46 telephony-maemo.c Oct 30 01:59:00 okay, i will take a look at it. Oct 30 01:59:04 e.g. CHLD0 is just plain wrong for instance Oct 30 01:59:42 btw, CHLD0 can be only tested in multicalls, and i still not test it yet. Oct 30 02:00:34 i am pretty sure that CHUP only release active call, not the whole Oct 30 02:00:48 because i have tested it on iPhone and LG phone. :) Oct 30 02:00:57 Which one is 'active' though Oct 30 02:01:11 Just try joining two calls and sending a CHUP Oct 30 02:01:15 Tell me which one is released Oct 30 02:01:52 Either way, it won't work for oFono Oct 30 02:02:08 The UI will tell to release a specific call Oct 30 02:02:19 If that is not the call that is released, the user will be pissed Oct 30 02:07:19 if two calls are joined, and use CHUP, what will happen? two calls are all released right? Oct 30 02:07:42 that's what I think _should_ happen Oct 30 02:08:21 i will test it Oct 30 02:23:33 when we use CHLD=2, actually I found the held call is in waiting status... Oct 30 02:36:37 sorry? Oct 30 02:49:57 let me do more testing. Oct 30 02:50:02 but it's tricky **** ENDING LOGGING AT Fri Oct 30 02:59:58 2009