**** BEGIN LOGGING AT Fri Feb 12 02:59:56 2010
Feb 12 21:00:12 <denkenz> padovan: I'm going to push some changes to hfp in about an hour, would you be available to do a test then?
Feb 12 21:00:24 <denkenz> padovan: I wanna have a release this weekend
Feb 12 21:03:06 <Ryback_> yay, great :-)
Feb 12 21:19:54 <padovan> denkenz: yes, I can test it.
Feb 12 21:20:11 <padovan> denkenz: did you take a look at my patches?
Feb 12 21:23:42 <denkenz> padovan: Yeah I've integrated them all already
Feb 12 21:24:39 <padovan> Nice.
Feb 12 21:35:46 <padovan> denkenz, holtmann: It would be good if we have a new release on bluez togheter.
Feb 12 21:47:20 <denkenz> padovan: I have pushed my main changes, the only thing left to solve now is hfp modem device naming
Feb 12 21:47:42 <denkenz> padovan: would be trivial if not for the uuids property changed handling
Feb 12 21:52:15 <holtmann> padovan: BlueZ release is tagged. Just happens that kernel.org sync is slow today.
Feb 12 21:56:59 <denkenz> bleah, the UUIDs thing is a bit nasty to get right
Feb 12 21:57:11 <padovan> denkenz: I'll test it here once I fetch kernel.org ;-)
Feb 12 21:58:04 <denkenz> if you don't get it in next 15 minutes or so I just git format-patch and send them to you
Feb 12 21:58:50 <padovan> Ok
Feb 12 22:15:19 <padovan> ofono and bluez fetched :)
Feb 12 22:52:43 <padovan> denkenz: I found a issue, we are calling Disconnect after receive a Release(). It's on my code, not yours.
Feb 12 22:52:52 <padovan> ofonod[24539]: plugins/hfp.c:hfp_disable() 0x6a7040
Feb 12 22:52:54 <padovan> ofonod[24539]: plugins/hfp.c:hfp_power_down() Disconnect reply: Method "Disconnect" with signature "" on interface "org.bluez.HandsfreeGateway" doesn't exist
Feb 12 22:53:25 <padovan> this happens when I remove the dongle with hfpmodem powered up.
Feb 12 22:58:07 <padovan> denkenz: That is the fix: http://pastebin.com/d4f500536, If you agree with me I'll send you a patch.
Feb 12 23:24:35 <denkenz> hmm, why does disable get called when we don't have an agent registered?
Feb 12 23:25:18 <denkenz> padovan: Shouldn't we have already ofono_modem_set_powered by now?
Feb 12 23:26:41 <padovan> denkenz: agent is registered, bluez calls Release()
Feb 12 23:27:24 <padovan> but when it call Disconnect the dongle was already removed.
Feb 12 23:27:41 <denkenz> Ah, because we use ofono_modem_remove
Feb 12 23:27:42 <denkenz> gotcha
Feb 12 23:27:47 <denkenz> Yes that patch is fine
Feb 12 23:32:49 <padovan> denkenz: patch sent
Feb 12 23:58:07 <denkenz> padovan: ok, I should have a bunch more changes shortly, and hopefully that'll be it
Feb 12 23:59:32 <padovan> Ok.
Feb 13 00:43:32 <denkenz> padovan: OK, I think we're good to go, do another sanity check and test device aliases
Feb 13 00:43:51 <denkenz> padovan: It works for me, but double check, if no issues we're ready for a release
Feb 13 01:01:56 <padovan> denkenz: Would be good add hfp on the path too?
Feb 13 01:05:54 <padovan> and what about do something like bluez on the obj path? for example: /hfp/adp_00_1A_6B_F7_A1_0E/dev_00_1A_6B_F7_A1_0E
Feb 13 01:12:13 <denkenz> [Thursday 28 January 2010] [10:56:55] <holtmann>        padovan: Use /hfp/0011FF..._0011FF as path. No extra objects in between.
Feb 13 01:12:15 <denkenz> [Thursday 28 January 2010] [10:57:16] <holtmann>        And the the Name use the BlueZ alias.
Feb 13 01:12:32 <denkenz> I suppose we can add /hfp/ prefix
Feb 13 01:12:42 <padovan> ok.
Feb 13 01:13:06 <denkenz> was that the only issue? :)
Feb 13 01:13:25 <padovan> denkenz: still checking..
Feb 13 01:18:22 <denkenz> -       create_path(dev_addr, adapter_addr, buf, sizeof(buf));
Feb 13 01:18:23 <denkenz> +       strcpy(buf, "/hfp/");
Feb 13 01:18:26 <denkenz> +       create_path(dev_addr, adapter_addr, buf + 5, sizeof(buf) - 5);
Feb 13 01:24:05 <padovan> hum, I get that on bluez:
Feb 13 01:24:07 <padovan> bluetoothd[9285]: Agent reply: Method "NewConnection" with signature "h" on interface "org.bluez.HandsfreeAgent" doesn't exist 
Feb 13 01:24:32 <padovan> let me rebase my bluez tree and check again
Feb 13 01:28:29 <denkenz> I should not have been messing with agent registration
Feb 13 01:28:43 <denkenz> at least I don't think I did
Feb 13 01:28:58 <padovan> Yes, I still don't know what caused that.
Feb 13 01:30:22 <padovan> denkenz: your change to add /hfp/ is not working
Feb 13 01:30:48 <denkenz> hmm, does modem_create barf?
Feb 13 01:32:04 <padovan> debuging
Feb 13 01:33:02 <denkenz> Feb 12 14:33:42 localhost dbus-daemon: Rejected send message, 4 matched rules; type="method_return", sender=":1.5" (uid=0 pid=6299 comm="/usr/sbin/bluetoothd) interface="(unset)" member="(unset)" error name="(unset)" requested_reply=0 destination=":1.22" (uid=1001 pid=13258 comm="src/ofonod))
Feb 13 01:33:20 <denkenz> I'm also getting that, and have no idea which dbus policy prevents method_returns...
Feb 13 01:34:29 <padovan> it is failing here __ofono_dbus_valid_object_path
Feb 13 01:34:52 <denkenz> ah, we can't use / with hfp
Feb 13 01:35:32 <denkenz> because the daemon actually adds the slash in front
Feb 13 01:35:41 <padovan> lets do hfp_000..
Feb 13 01:36:10 <padovan> Ah..gotcha
Feb 13 01:36:15 <denkenz> -       create_path(dev_addr, adapter_addr, buf, sizeof(buf));
Feb 13 01:36:17 <denkenz> +       strcpy(buf, "hfp/");
Feb 13 01:36:18 <denkenz> +       create_path(dev_addr, adapter_addr, buf + 4, sizeof(buf) - 4);
Feb 13 01:37:02 <padovan> nice, it works now
Feb 13 01:39:22 <padovan> I figured out why bluez is reporting a error:
Feb 13 01:39:38 <padovan> 4c0790da4fe55084f82b5df4b9d7fe25f1558b0a
Feb 13 01:39:51 <denkenz> holtmann: You need to add allow send_interface="org.bluez.HandsfreeAgent"
Feb 13 01:39:52 <padovan> we can't unregister on disable =(
Feb 13 01:40:13 <holtmann> denkenz: Send patch for it.
Feb 13 01:41:34 <padovan> denkenz: my patch is wrong, maybe we need a flag to no call Disconnect under Release
Feb 13 01:41:54 <denkenz> Yes you can't unregister on disable :P
Feb 13 01:42:31 <denkenz> what was the original problem again?
Feb 13 01:43:35 <denkenz> should we do something like if (data->handsfree_path) send_disconnect?
Feb 13 01:43:44 <padovan> 20:52    padovan> ofonod[24539]: plugins/hfp.c:hfp_disable() 0x6a7040
Feb 13 01:43:45 <padovan> 20:52    padovan> ofonod[24539]: plugins/hfp.c:hfp_power_down() Disconnect reply: Method "Disconnect" with signature "" on interface "org.bluez.HandsfreeGateway" doesn't exist
Feb 13 01:43:48 <padovan> 20:53    padovan> this happens when I remove the dongle with hfpmodem powered up.
Feb 13 01:46:08 <padovan> no, data->hf_path is freed only on hfp_remove
Feb 13 01:46:34 <denkenz> yes we need a flag, I don't think we have anything else we can use
Feb 13 01:46:43 <denkenz> unless there's a dbus_is_registered or something
Feb 13 01:51:37 <padovan> I don't know if it is worth, but can create a g_dbus_is_registered
Feb 13 01:52:59 <padovan> denkenz: function find_interface() on gdbus/object.c
Feb 13 01:54:42 <denkenz> Yes we could do that, holtmann: comments?
Feb 13 01:55:50 <denkenz> otherwise, http://pastebin.ca/1794957
Feb 13 01:56:25 <holtmann> Do we care about the reply? If not, we could just send the method with NO_REPLY flag.
Feb 13 01:56:54 <denkenz> we don't, but we shouldn't call it anyway for efficiency reasons
Feb 13 01:59:20 <padovan> denkenz: patch looks good but the uneeded gboolean send_disconnect var.
Feb 13 01:59:46 <denkenz> why unneeded? we clear the data
Feb 13 01:59:47 <padovan> But I would prefer use some dbus_is_registered method.
Feb 13 02:00:18 <denkenz> so would I, but holtmann needs to review it
Feb 13 02:00:50 <holtmann> We wanna check if a path exists, right?
Feb 13 02:01:00 <padovan> denkenz: wo don't clear it. only hfp_remove free it
Feb 13 02:01:35 <holtmann> padovan: Send a proposal to the oFono mailing list.
Feb 13 02:01:45 <denkenz> ah yes, you're right
Feb 13 02:02:34 <denkenz> its getting late, so lets go with this fix and work on g_dbus_is_registered next week
Feb 13 02:02:46 <denkenz> then we can review it properly
Feb 13 02:03:11 <denkenz> Make sure it works so I can push :)
Feb 13 02:09:48 <padovan> denkenz: let's do the release, it works.
Feb 13 02:11:42 <denkenz> padovan, holtmann: Ok pushed, lets release
Feb 13 02:11:59 <denkenz> padovan: Work on the g_dbus changes next week, and hopefully we get more feedback anyway :)
Feb 13 02:12:09 <denkenz> Won't be the last issue we fix
Feb 13 02:13:18 <padovan> denkenz: Ok.
Feb 13 02:15:07 <padovan> Now I'm off. Going to the carnival. It's carnival here in Brazil! =)
Feb 13 02:15:35 <holtmann> padovan: Have fun.
Feb 13 02:16:38 <padovan> holtmann: thanks :)
**** ENDING LOGGING AT Sat Feb 13 02:59:58 2010