| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
2000-05-11 Federico Mena Quintero <federico@helixcode.com>
* gui/gnome-cal.c (gnome_calendar_update_all): Removed unused
arguments. Load the initial alarms here.
(load_alarms): New function to load a day's worth of alarms.
(gnome_calendar_class_init): Eeeek! This was taking in an
incorrect argument type.
(gnome_calendar_init): Now the calendar keeps a hash table of
UIDs->queued alarms. Create the hash table here.
(gnome_calendar_destroy): Destroy the alarms hash table.
(gnome_calendar_object_updated_cb): Remove the alarms for the
object and regenerate them.
(gnome_calendar_object_removed_cb): Remove the alarms for the
object.
* gui/alarm.c (alarm_add): Do not take in a CalendarAlarm, just
the trigger time, the callback and the closure data. Return an
opaque identifier for the alarm so that it can be removed by the
client code if needed. Use the queue_alarm() helper function.
(queue_alarm): Helper function to actually queue the alarm and set
up the itimer. Deal with a nonzero return value from
setitimer().
(alarm_remove): New function to remove an alarm based on its ID.
(pop_alarm): New helper function; pops the first alarm of the
queue and resets the timer as appropriate.
(alarm_ready): Simplified a lot by using pop_alarm().
* idl/evolution-calendar.idl (Cal): Added get_alarms_in_range().
* pcs/cal.c (build_instance_seq): New function to build a CORBA
sequence from the internal list of instances.
(Cal_get_events_in_range): Use build_instance_seq().
(Cal_get_alarms_in_range): Implemented new method.
* pcs/cal-backend.c (cal_backend_get_alarms_in_range): New
function with the get_alarms_in_range() engine.
* pcs/cal-backend-imc.c (cal_backend_imc_get_alarms_in_range):
Implemented the get_alarms_in_range() method.
* cal-client/cal-client.c (cal_client_get_alarms_in_range): New
client-side function for getting the alarms.
(build_instance_list): New helper function to build the
CalObjInstance list from the CORBA sequence.
(cal_client_get_events_in_range): Use build_instance_list().
* gui/calendar-commands.h: #include <cal-util/calobj.h>. #include
"gnome-cal.h".
* gui/e-week-view.c: #include "calendar-commands.h" instead of
main.h; the latter is an obsolete file and will be killed.
* gui/evolution-calendar-control.c (main): Call init_bonobo()
before anything else. We need the GTK+ object system initialized.
* gui/Makefile.am (evolution_calendar_SOURCES): Do not use main.h.
* cal-util/cal-util.c (cal_alarm_instance_list_free): New function.
svn path=/trunk/; revision=2987
|
|
|
2000-05-08 Federico Mena Quintero <federico@helixcode.com>
* pcs/cal-backend.h (CalBackendClass): CalBackendClass now is just
an interface for calendar backends; this is an abstract class.
Put in the vtable for the backend methods.
* pcs/cal-backend.c (cal_backend_new): Removed function, since
CalBackend is not just an abstract class.
Removed implementation-specific functions and made public
functions call the virtual methods instead.
* pcs/cal-backend-imc.[ch]: New files with the CalBackendIMC
implementation; this implements a backend for iCalendar and
vCalendar files. Moved the implementation-specific stuff from
cal-backend.[ch] to here.
* pcs/cal-backend-imc.c (CalendarFormat): Moved enumeration to
here. Added a CAL_UNKNOWN value for when the backend is not
loaded yet.
(cal_backend_imc_init): Initialize priv->format as CAL_UNKNOWN.
(save_to_vcal): Use the same VCProdIdProp value as in
cal-util/calobj.c. Use "1.0" as the VCVersionProp as per the
vCalendar spec.
(ensure_uid): Return nothing, since the result value need not be
used anymore.
(add_object): Since we mark the calendar as dirty anyways, we do
not need to check the result value of ensure_uid() anymore.
(remove_object): Asssert that we know how to handle the object's
type. We do this in add_object() anyways.
* pcs/Makefile.am (libpcs_a_SOURCES): Added cal-backend-imc.[ch].
* gui/gnome-cal.c: Replaced debugging printf()s with g_message()
so that we can see the line number where they occur.
* gui/gnome-cal.c (gnome_calendar_load_cb): Sort of handle the
LOAD_METHOD_NOT_SUPPORTED result code, and added a default for the
switch.
* cal-client/cal-listener.h (CalListenerLoadStatus): Removed
enumeration; it is stupid to translate all values for the
CalClient when it is going to translate them again.
(CalListenerClass::cal_loaded): This signal now passes the
LoadStatus directly from the CORBA side.
* cal-client/cal-listener.c (Listener_cal_loaded): Do not
translate the status value.
* cal-client/cal-client.h (CalClientLoadStatus): Added the
CAL_CLIENT_LOAD_METHOD_NOT_SUPPORTED error code.
* cal-client/cal-client.c (cal_loaded_cb): Translate the CORBA
version of the LoadStatus result code.
* pcs/cal-factory.c (CalFactoryPrivate): New methods field for the
hash table from method strings to the GtkTypes for backend class
types.
(cal_factory_init): Create the priv->methods hash table.
(cal_factory_destroy): Free the priv->methods hash table.
(cal_factory_register_method): New function to register a backend
class for a particular URI method.
(launch_backend_for_uri): New function to launch a backend for a
particular URI's method.
(load_backend): Use launch_backend_for_uri(). Move the error
notification code from load_fn() to here.
(create_backend): Use launch_backend_for_uri(). Move the error
notification code form create_fn() to here; it is #ifdefed out
since currently cal_backend_create() does not have any error
reporting capabilities.
* idl/evolution-calendar.idl (Listener::LoadStatus): Added a
PROTOCOL_NOT_SUPPORTED error code.
* pcs/cal-factory.c (cal_factory_load cal_factory_create): Removed
functions, since they were supposed to be internal only.
(CalFactory_load): Call queue_load_create_job() directly.
(CalFactory_create): Likewise.
svn path=/trunk/; revision=2921
|
