Hi Philippe,

On 09/05/2012 11:56 AM, Philippe Nunes wrote:
> The "USSD terminated by network" notification is handled to recover
> the idle state in case of USSD transaction error.
> ---
>   src/ussd.c |   12 ++++++++++++
>   1 file changed, 12 insertions(+)
>
> diff --git a/src/ussd.c b/src/ussd.c
> index 74888b2..a35f172 100644
> --- a/src/ussd.c
> +++ b/src/ussd.c
> @@ -414,6 +414,18 @@ void ofono_ussd_notify(struct ofono_ussd *ussd, int status, int dcs,
>   		return;
>   	}
>
> +	if (status == OFONO_USSD_STATUS_TERMINATED&&
> +			(ussd->state == USSD_STATE_IDLE ||

Why do you need this guard?

> +			ussd->state == USSD_STATE_USER_ACTION)) {

Or this one for that matter.  In theory the terminated notification can 
come at any time.

> +		ussd_change_state(ussd, USSD_STATE_IDLE);
> +
> +		if (ussd->pending == NULL)
> +			return;
> +
> +		reply = __ofono_error_failed(ussd->pending);
> +		goto out;
> +	}
> +
>   	if (status == OFONO_USSD_STATUS_NOT_SUPPORTED) {
>   		ussd_change_state(ussd, USSD_STATE_IDLE);
>

Can't we treat it exactly like 'NOT_SUPPORTED' or 'TIMED_OUT' but with a 
different error?

Regards,
-Denis