[maemo-developers] [maemo-developers] btcond / d-bus problems

[From: Andrew Ramsay]

On Tue, 27 Dec 2005, Johan Hedberg wrote:

> Hi Andrew,
>
> On Tue, Dec 27, 2005, Andrew Ramsay wrote:
>> I started off with the basic example program which just connects to the
>> session bus, lists the services available, then exits. That works fine
>> with both session and system buses. I then tried to modify it slightly to
>> obtain a proxy for the "com.nokia.btcond" service, which didn't work until
>> I discovered you had to connect to the system rather than the session bus.
>> Once I'd got the proxy, I tried to call the rfcomm_bind method as
>> suggested in the wiki. However, all I get back is an error saying that
>> this method doesn't exist on the com.nokia.btcond.request interface.
>> Calling any of the other methods results in the error
>> "com.nokia.btcond.invalid_dev", which is supposed to indicate an invalid
>> BDA from what I can tell, although the address I'm using is definitely
>> correct.
>
> You didn't mention which object path you're using, but it should be
> "/com/nokia/btcond/request". You should also be giving the rfcomm_bind
> method call two arguments: the BT address (e.g. "11:22:33:44:55:66")
> and the profile name (e.g. "SPP").
I'm sure the device address is correct, although I wasn't certain that was 
the correct format for it, and I was using SPP for the profile name. For 
the path/interface names, I was using the strings #defined in bt-dbus.h, 
so they should be OK too.

>> Can anyone help me out here? As I said, I have no previous experience with
>> D-BUS programming, so I may well be doing something very wrong and just
>> haven't noticed yet. I'm using the Glib bindings for D-BUS since that was
>> what the example code used. If anyone has actually done this already I'd
>> be very interested to hear how you did it.
>
> It is quite possible that the GLib bindings are buggy in the D-BUS
> version that the current 770 SW images use. If you are still
> experiencing problems I'd recommend trying out the low level C-bindings.
>
> Johan
>
I've tried converting the code I had to the basic C bindings, but I'm 
still seeing exactly the same problems (rfcomm_bind apparently doesn't 
exist, other methods return com.nokia.btcond.error.invalid_dev).

Here are the important bits of the code I'm using..

#define BDA "00:02:76:C0:56:A4"

DBusError error;
DBusConnection* connection;
DBusMessage* message, *reply;
char* ret1, *ret2;
int reply_timeout, i;

dbus_error_init(&error);

connection = dbus_bus_get(DBUS_BUS_SYSTEM, &error);
if(connection == NULL) {
 	fprintf(stderr, "Failed to open bus connection: %s\n", 
error.message);
 	return 1;
}

message = dbus_message_new_method_call(BTCOND_SERVICE,
 			BTCOND_REQ_PATH,
 			BTCOND_REQ_INTERFACE,
 			BTCOND_RFCOMM_BIND_REQ);

dbus_message_append_args(message, DBUS_TYPE_STRING, BDA,
 				DBUS_TYPE_STRING, "SPP",
 				DBUS_TYPE_INVALID);

reply_timeout = -1;
reply = dbus_connection_send_with_reply_and_block(connection,
 						message,
 						reply_timeout,
 						&error);

if(dbus_error_is_set(&error)) {
 	fprintf(stderr, "Error: %s\n", error.message);
 	return 1;
}

if(!dbus_message_get_args(reply, &error, DBUS_TYPE_STRING, ret1,
 					DBUS_TYPE_INVALID)) {
 	fprintf(stderr, "get_args failed: %s\n", error.message);
 	return 1;
}

printf("returned: %s\n", ret1);

dbus_message_unref(message);
dbus_message_unref(reply);
...

Is there anything obviously wrong with that?

Andrew