new files to contain layout code split off from EDayView an EWeekView, so

2001-05-28  Damon Chaplin  <damon@ximian.com>

	* gui/e-week-view-layout.[hc]:
	* gui/e-day-view-layout.[hc]: new files to contain layout code split
	off from EDayView an EWeekView, so we can use it for printing.

	* gui/print.c: rewritten to use the same layout code as the EDayView
	and EWeekView widgets.

	* gui/gnome-cal.c (gnome_calendar_get_task_pad_cal_client): added
	function so we can get the CalClient used for the TaskPad for printing.

	* gui/Makefile.am (evolution_calendar_SOURCES): added
	e-day-view-layout.[hc] amd e-week-view-layout.[hc].

	* cal-util/timeutil.c (time_week_begin):
	(time_week_end): added week_start_day argument.

	* cal-util/cal-recur.c: added comments describing problems in it.

svn path=/trunk/; revision=10037

1 files changed, 18 insertions, 0 deletions

diff --git a/calendar/cal-util/cal-recur.c b/calendar/cal-util/cal-recur.c
index 0b4f114484..8ac6f771d9 100644
--- a/calendar/cal-util/cal-recur.c
+++ b/calendar/cal-util/cal-recur.c
@@ -84,6 +84,24 @@
  *   the BYSETPOS property is used to select which of the occurrences are
  *   finally output. If BYSETPOS is not specified then all the occurrences are
  *   output.
+ *
+ *
+ * FIXME: I think there are a few errors in this code:
+ *
+ *  1) I'm not sure it should be generating events in parallel like it says
+ *     above. That needs to be checked.
+ *
+ *  2) I didn't think about timezone changes when implementing this. I just
+ *     assumed all the occurrences of the event would be in local time.
+ *     But when clocks go back or forwards due to daylight-saving time, some
+ *     special handling may be needed, especially for the shorter frequencies.
+ *     e.g. for a MINUTELY frequency it should probably iterate over all the
+ *     minutes before and after clocks go back (i.e. some may be the same local
+ *     time but have different UTC offsets). For longer frequencies, if an
+ *     occurrence lands on the overlapping or non-existant time when clocks
+ *     go back/forward, then it may need to choose which of the times to use
+ *     or move the time forward or something. I'm not sure this is clear in the
+ *     spec.
  */
 
 /* Define this for some debugging output. */