Re: [PATCH 1/3] Fix A2DP Sink/Source recognition

[Johan Hedberg <johan.hedberg@gmail.com>]

Hi Chanyeol,

On Fri, May 18, 2012, chanyeol.park@samsung.com wrote:
> From: Chan-yeol Park <chanyeol.park@samsung.com>
> 
> ---
>  audio/a2dp.c     |    8 ++++----
>  audio/audio.conf |    2 +-
>  audio/manager.c  |    4 ++--
>  3 files changed, 7 insertions(+), 7 deletions(-)
> 
> diff --git a/audio/a2dp.c b/audio/a2dp.c
> index cc8f139..145dcf6 100644
> --- a/audio/a2dp.c
> +++ b/audio/a2dp.c
> @@ -1406,9 +1406,9 @@ int a2dp_register(DBusConnection *conn, const bdaddr_t *src, GKeyFile *config)
>  		g_clear_error(&err);
>  	} else {
>  		if (strstr(str, "Sink"))
> -			source = TRUE;
> -		if (strstr(str, "Source"))
>  			sink = TRUE;
> +		if (strstr(str, "Source"))
> +			source = TRUE;
>  		if (strstr(str, "Socket"))
>  			socket = TRUE;
>  		g_free(str);
> @@ -1421,9 +1421,9 @@ int a2dp_register(DBusConnection *conn, const bdaddr_t *src, GKeyFile *config)
>  		g_clear_error(&err);
>  	} else {
>  		if (strstr(str, "Sink"))
> -			source = FALSE;
> -		if (strstr(str, "Source"))
>  			sink = FALSE;
> +		if (strstr(str, "Source"))
> +			source = FALSE;
>  		if (strstr(str, "Socket"))
>  			socket = FALSE;
>  		g_free(str);

This doesn't look right to me. The config file values have (for
historical reasons) been referring to the remote device object D-Bus
interface names whereas these gboolean variables are used to determine
what kind of local stream endpoints need to be registered. So having
e.g. support for "Source" on a remote device requires registering a
local sink stream endpoint.

Johan