Hi Johan.
On Tue, Mar 27, 2012 at 7:50 AM, Johan Hedberg <johan.hedberg@xxxxxxxxx> wrote:
>
> Hi Jefferson,
>
> On Fri, Mar 23, 2012, Jefferson Delfes wrote:
> > In interactive mode, when connection is lost, the prompt used to remain
> > in "connected" state. This patch fixes that case, by always showing the
> > actual connection state.
> > ---
> > attrib/interactive.c | 11 +++++++++++
> > 1 file changed, 11 insertions(+)
> >
> > diff --git a/attrib/interactive.c b/attrib/interactive.c
> > index be81424..47ddcf6 100644
> > --- a/attrib/interactive.c
> > +++ b/attrib/interactive.c
> > @@ -56,6 +56,7 @@ struct characteristic_data {
> > };
> >
> > static void cmd_help(int argcp, char **argvp);
> > +static void cmd_disconnect(int argcp, char **argvp);
>
> We should really try to avoid forward declarations when possible and the
> cmd_* functions should be reserved for real user entered commands and
> not as generic helpers as you're trying to do. So I'd factor out code
> from cmd_disconnect into a static helper function somewhere above both
> cmd_disconnect and cmd_connect, call it e.g. disconnect_io().
Ok, I'll do these changes.
> > +static gboolean channel_watcher(GIOChannel *chan, GIOCondition cond,
> > + gpointer user_data)
> > +{
> > + cmd_disconnect(0, NULL);
> > +
> > + return TRUE;
> > +}
>
> Shouldn't this be returning FALSE instead of TRUE so that the watch
> doesn't get left hanging around GLibs watch lists wasting memory?
Oops! You are right. I didn't know about this. I made some tests in
valgrind now and I did see some still reachable extra bytes.
> Johan
I'll send a replace for this patch.
Thanks.
--
To unsubscribe from this list: send the line "unsubscribe linux-bluetooth" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
