Code review comment for lp:~charlesk/indicator-datetime/lp-1204532

> * For decrementing the subtask_count you should probably use g_atomic_int_dec_and_test().

Fixed r255

> * In on_appointments_change() you should probably handle the case of a GDateTime being NULL. I think that the indicator functions are fine, but the g_datetime_unref() is not.

get_calendar_date() doesn't return NULL, and the arguments to g_date_time_new_local() all look sane. I don't see the NULL case you're wanting to safeguard against?

> the menu refreshes, but the 5 appointments stay the same.

Oog. in progress.