mirror of
https://gitlab.freedesktop.org/gstreamer/gstreamer.git
synced 2024-12-27 18:50:48 +00:00
libs/gst/base/gstbasesink.c: Rearrange the latency query code. We always want to do the upstream query, even if we ar...
Original commit message from CVS: * libs/gst/base/gstbasesink.c: (gst_base_sink_query_latency), (gst_base_sink_send_event): Rearrange the latency query code. We always want to do the upstream query, even if we are not live so that the upstream elements can get the latency results too. If we fail doing the query and we are live, we return TRUE afterwards.
This commit is contained in:
parent
d9c9c264e2
commit
d8d6ab79cb
2 changed files with 53 additions and 39 deletions
|
@ -1,3 +1,12 @@
|
|||
2008-04-25 Wim Taymans <wim.taymans@collabora.co.uk>
|
||||
|
||||
* libs/gst/base/gstbasesink.c: (gst_base_sink_query_latency),
|
||||
(gst_base_sink_send_event):
|
||||
Rearrange the latency query code. We always want to do the upstream
|
||||
query, even if we are not live so that the upstream elements can get the
|
||||
latency results too. If we fail doing the query and we are live, we
|
||||
return TRUE afterwards.
|
||||
|
||||
2008-04-24 Stefan Kost <ensonic@users.sf.net>
|
||||
|
||||
patch by: Jason Zhao <e3423c@motorola.com>
|
||||
|
|
|
@ -912,19 +912,16 @@ gst_base_sink_query_latency (GstBaseSink * sink, gboolean * live,
|
|||
max = -1;
|
||||
us_live = FALSE;
|
||||
|
||||
/* we are live */
|
||||
if (l) {
|
||||
if (have_latency) {
|
||||
/* we are live and ready for a latency query */
|
||||
query = gst_query_new_latency ();
|
||||
|
||||
/* ask the peer for the latency */
|
||||
if (!(res = gst_base_sink_peer_query (sink, query)))
|
||||
goto query_failed;
|
||||
if (have_latency) {
|
||||
GST_DEBUG_OBJECT (sink, "we are ready for LATENCY query");
|
||||
/* we are ready for a latency query this is when we preroll or when we are
|
||||
* not async. */
|
||||
query = gst_query_new_latency ();
|
||||
|
||||
/* ask the peer for the latency */
|
||||
if ((res = gst_base_sink_peer_query (sink, query))) {
|
||||
/* get upstream min and max latency */
|
||||
gst_query_parse_latency (query, &us_live, &us_min, &us_max);
|
||||
gst_query_unref (query);
|
||||
|
||||
if (us_live) {
|
||||
/* upstream live, use its latency, subclasses should use these
|
||||
|
@ -932,37 +929,38 @@ gst_base_sink_query_latency (GstBaseSink * sink, gboolean * live,
|
|||
min = us_min;
|
||||
max = us_max;
|
||||
}
|
||||
} else {
|
||||
/* we are live but are not yet ready for a latency query */
|
||||
res = FALSE;
|
||||
}
|
||||
} else {
|
||||
/* not live, result is always TRUE */
|
||||
res = TRUE;
|
||||
}
|
||||
|
||||
GST_DEBUG_OBJECT (sink, "latency query: live: %d, have_latency %d,"
|
||||
" upstream: %d, min %" GST_TIME_FORMAT ", max %" GST_TIME_FORMAT, l,
|
||||
have_latency, us_live, GST_TIME_ARGS (min), GST_TIME_ARGS (max));
|
||||
|
||||
if (live)
|
||||
*live = l;
|
||||
if (upstream_live)
|
||||
*upstream_live = us_live;
|
||||
if (min_latency)
|
||||
*min_latency = min;
|
||||
if (max_latency)
|
||||
*max_latency = max;
|
||||
|
||||
return res;
|
||||
|
||||
/* ERRORS */
|
||||
query_failed:
|
||||
{
|
||||
GST_DEBUG_OBJECT (sink, "latency query failed");
|
||||
gst_query_unref (query);
|
||||
return FALSE;
|
||||
} else {
|
||||
GST_DEBUG_OBJECT (sink, "we are not yet ready for LATENCY query");
|
||||
res = FALSE;
|
||||
}
|
||||
|
||||
/* not live, we tried to do the query, if it failed we return TRUE anyway */
|
||||
if (!res) {
|
||||
if (!l) {
|
||||
res = TRUE;
|
||||
GST_DEBUG_OBJECT (sink, "latency query failed but we are not live");
|
||||
} else {
|
||||
GST_DEBUG_OBJECT (sink, "latency query failed and we are live");
|
||||
}
|
||||
}
|
||||
|
||||
if (res) {
|
||||
GST_DEBUG_OBJECT (sink, "latency query: live: %d, have_latency %d,"
|
||||
" upstream: %d, min %" GST_TIME_FORMAT ", max %" GST_TIME_FORMAT, l,
|
||||
have_latency, us_live, GST_TIME_ARGS (min), GST_TIME_ARGS (max));
|
||||
|
||||
if (live)
|
||||
*live = l;
|
||||
if (upstream_live)
|
||||
*upstream_live = us_live;
|
||||
if (min_latency)
|
||||
*min_latency = min;
|
||||
if (max_latency)
|
||||
*max_latency = max;
|
||||
}
|
||||
return res;
|
||||
}
|
||||
|
||||
static void
|
||||
|
@ -2932,13 +2930,20 @@ gst_base_sink_send_event (GstElement * element, GstEvent * event)
|
|||
|
||||
gst_event_parse_latency (event, &latency);
|
||||
|
||||
/* store the latency. We use this to adjust the running_time before syncing
|
||||
* it to the clock. */
|
||||
GST_OBJECT_LOCK (element);
|
||||
basesink->priv->latency = latency;
|
||||
GST_OBJECT_UNLOCK (element);
|
||||
GST_DEBUG_OBJECT (basesink, "latency set to %" GST_TIME_FORMAT,
|
||||
GST_TIME_ARGS (latency));
|
||||
|
||||
/* don't forward, yet */
|
||||
/* don't forward, yet. FIXME. The latency event should likely be forwarded
|
||||
* to upstream element so that they can configure themselves. Each element
|
||||
* would subtract the amount of LATENCY it can maximally compensate for.
|
||||
* It's currently not very useful; even if this sink cannot compensate for
|
||||
* all the latency, upstream will block while this sink waits which will
|
||||
* trigger implicit buffering and latency there. */
|
||||
forward = FALSE;
|
||||
break;
|
||||
}
|
||||
|
|
Loading…
Reference in a new issue