gst/gstsystemclock.c: Add some more docs to explain why a FIXME was wrongly added.

Original commit message from CVS:
* gst/gstsystemclock.c: (gst_system_clock_id_wait_jitter_unlocked):
Add some more docs to explain why a FIXME was wrongly added.
This commit is contained in:
Wim Taymans 2007-12-24 14:35:24 +00:00
parent 43377772a6
commit 20b951efc1
2 changed files with 13 additions and 6 deletions

View file

@ -1,3 +1,8 @@
2007-12-24 Wim Taymans <wim.taymans@collabora.co.uk>
* gst/gstsystemclock.c: (gst_system_clock_id_wait_jitter_unlocked):
Add some more docs to explain why a FIXME was wrongly added.
2007-12-22 Sebastian Dröge <slomo@circular-chaos.org>
* gst/gstobject.c:

View file

@ -363,7 +363,8 @@ gst_system_clock_id_wait_jitter_unlocked (GstClock * clock,
GstClockTime entryt, real, now, target;
GstClockTimeDiff diff;
/* need to call the overridden method */
/* need to call the overridden method because we want to sync against the time
* of the clock, whatever the subclass uses as a clock. */
real = GST_CLOCK_GET_CLASS (clock)->get_internal_time (clock);
entryt = GST_CLOCK_ENTRY_TIME (entry);
@ -371,12 +372,13 @@ gst_system_clock_id_wait_jitter_unlocked (GstClock * clock,
if (jitter) {
*jitter = GST_CLOCK_DIFF (entryt, now);
}
/* the diff of the entry with the clock is the amount of time we have to
* wait */
diff = entryt - now;
/* FIXME: should that be not just:
* target = GST_CLOCK_GET_CLASS (clock)->get_internal_time (clock) + diff;
* or even
* target = real + diff;
*/
/* Our GCond implementation expects an absolute time against the system clock
* as a timeout value. We use our internal time to get the system time and add
* the expected timeout to it, this gives us the absolute time of the
* timeout. */
target = gst_system_clock_get_internal_time (clock) + diff;
GST_CAT_DEBUG (GST_CAT_CLOCK, "entry %p"