gstreamer/gst-libs/gst/play/play.c

1277 lines
41 KiB
C
Raw Normal View History

/* GStreamer
* Copyright (C) 2003 Julien Moutte <julien@moutte.net>
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Library General Public
* License as published by the Free Software Foundation; either
* version 2 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Library General Public License for more details.
*
* You should have received a copy of the GNU Library General Public
* License along with this library; if not, write to the
* Free Software Foundation, Inc., 59 Temple Place - Suite 330,
* Boston, MA 02111-1307, USA.
*/
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#include <string.h>
#include "play.h"
#define TICK_INTERVAL_MSEC 200
enum
{
TIME_TICK,
STREAM_LENGTH,
HAVE_VIDEO_SIZE,
LAST_SIGNAL
};
struct _GstPlayPrivate {
char *location;
GHashTable *elements;
gint64 time_nanos;
gint64 length_nanos;
gint get_length_attempt;
gint tick_unblock_remaining; /* how many msecs left
to unblock due to seeking */
guint tick_id;
guint length_id;
gulong handoff_hid;
/* error/debug handling */
GError *error;
gchar *debug;
};
static guint gst_play_signals[LAST_SIGNAL] = { 0 };
static GstPipelineClass *parent_class = NULL;
/* ======================================================= */
/* */
/* Private Methods */
/* */
/* ======================================================= */
static GstCaps * gst_play_video_fixate (GstPad *pad, const GstCaps *caps, gpointer user_data);
static GstCaps * gst_play_audio_fixate (GstPad *pad, const GstCaps *caps, gpointer user_data);
static GQuark
gst_play_error_quark (void)
{
static GQuark quark = 0;
if (quark == 0)
quark = g_quark_from_static_string ("gst-play-error-quark");
return quark;
}
/* General GError creation */
static void
gst_play_error_create (GError ** error, const gchar *message)
{
/* check if caller wanted an error reported */
if (error == NULL)
return;
*error = g_error_new (GST_PLAY_ERROR, 0, message);
return;
}
/* GError creation when plugin is missing */
/* FIXME: what if multiple elements could have been used and they're all
* missing ? varargs ? */
static void
gst_play_error_plugin (const gchar *element, GError ** error)
{
gchar *message;
message = g_strdup_printf ("The %s element could not be found. "
"This element is essential for playback. "
"Please install the right plug-in and verify "
"that it works by running 'gst-inspect %s'",
element, element);
gst_play_error_create (error, message);
g_free (message);
return;
}
#define GST_PLAY_MAKE_OR_ERROR(el, factory, name, error) \
G_STMT_START { \
el = gst_element_factory_make (factory, name); \
if (!GST_IS_ELEMENT (el)) \
{ \
gst_play_error_plugin (factory, error); \
return FALSE; \
} \
} G_STMT_END
#define GST_PLAY_ERROR_RETURN(error, message) \
G_STMT_START { \
gst_play_error_create (error, message); \
return FALSE; \
} G_STMT_END
static gboolean
gst_play_pipeline_setup (GstPlay *play, GError **error)
{
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Threads */
GstElement *work_thread, *audio_thread, *video_thread;
/* Main Thread elements */
GstElement *source, *autoplugger, *audioconvert, *volume, *tee, *identity;
GstElement *identity_cs;
/* Visualization bin */
GstElement *vis_bin, *vis_queue, *vis_element, *vis_cs;
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Video Thread elements */
GstElement *video_queue, *video_switch, *video_cs, *video_balance;
GstElement *balance_cs, *video_scaler, *video_sink;
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Audio Thread elements */
GstElement *audio_queue, *audio_sink;
/* Some useful pads */
GstPad *tee_pad1, *tee_pad2;
g_return_val_if_fail (play != NULL, FALSE);
g_return_val_if_fail (GST_IS_PLAY (play), FALSE);
/* Creating main thread and its elements */
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
{
GST_PLAY_MAKE_OR_ERROR (work_thread, "thread", "work_thread", error);
g_hash_table_insert (play->priv->elements, "work_thread", work_thread);
gst_bin_add (GST_BIN (play), work_thread);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Placeholder for datasrc */
GST_PLAY_MAKE_OR_ERROR (source, "fakesrc", "source", error);
g_hash_table_insert (play->priv->elements, "source", source);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Autoplugger */
GST_PLAY_MAKE_OR_ERROR (autoplugger, "spider", "autoplugger", error);
g_hash_table_insert (play->priv->elements, "autoplugger", autoplugger);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Make sure we convert audio to the needed format */
GST_PLAY_MAKE_OR_ERROR (audioconvert, "audioconvert", "audioconvert", error);
g_hash_table_insert (play->priv->elements, "audioconvert", audioconvert);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Duplicate audio signal to audio sink and visualization thread */
GST_PLAY_MAKE_OR_ERROR (tee, "tee", "tee", error);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
tee_pad1 = gst_element_get_request_pad (tee, "src%d");
tee_pad2 = gst_element_get_request_pad (tee, "src%d");
g_hash_table_insert (play->priv->elements, "tee_pad1", tee_pad1);
g_hash_table_insert (play->priv->elements, "tee_pad2", tee_pad2);
g_hash_table_insert (play->priv->elements, "tee", tee);
gst_bin_add_many (GST_BIN (work_thread), source, autoplugger, audioconvert, tee, NULL);
if (!gst_element_link_many (source, autoplugger, audioconvert, tee, NULL))
GST_PLAY_ERROR_RETURN (error, "Could not link source thread elements");
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* identity ! colorspace ! switch */
GST_PLAY_MAKE_OR_ERROR (identity, "identity", "identity", error);
g_hash_table_insert (play->priv->elements, "identity", identity);
identity_cs = gst_element_factory_make ("ffcolorspace", "identity_cs");
if (!GST_IS_ELEMENT (identity_cs)) {
identity_cs = gst_element_factory_make ("colorspace", "identity_cs");
if (!GST_IS_ELEMENT (identity_cs))
{
gst_play_error_plugin ("colorspace", error);
return FALSE;
}
}
g_hash_table_insert (play->priv->elements, "identity_cs", identity_cs);
gst_bin_add_many (GST_BIN (work_thread), identity, identity_cs, NULL);
if (!gst_element_link_many (autoplugger, identity, identity_cs, NULL))
GST_PLAY_ERROR_RETURN (error, "Could not link work thread elements");
}
/* Visualization bin (note: it s not added to the pipeline yet) */
{
vis_bin = gst_bin_new ("vis_bin");
if (!GST_IS_ELEMENT (vis_bin))
{
gst_play_error_plugin ("bin", error);
return FALSE;
}
g_hash_table_insert (play->priv->elements, "vis_bin", vis_bin);
/* Buffer queue for video data */
GST_PLAY_MAKE_OR_ERROR (vis_queue, "queue", "vis_queue", error);
g_hash_table_insert (play->priv->elements, "vis_queue", vis_queue);
/* Visualization element placeholder */
GST_PLAY_MAKE_OR_ERROR (vis_element, "identity", "vis_element", error);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
g_hash_table_insert (play->priv->elements, "vis_element", vis_element);
/* Colorspace conversion */
vis_cs = gst_element_factory_make ("ffcolorspace", "vis_cs");
if (!GST_IS_ELEMENT (vis_cs)) {
vis_cs = gst_element_factory_make ("colorspace", "vis_cs");
if (!GST_IS_ELEMENT (vis_cs))
{
gst_play_error_plugin ("colorspace", error);
return FALSE;
}
}
g_hash_table_insert (play->priv->elements, "vis_cs", vis_cs);
gst_bin_add_many (GST_BIN (vis_bin), vis_queue, vis_element, vis_cs, NULL);
if (!gst_element_link_many (vis_queue, vis_element, vis_cs, NULL))
GST_PLAY_ERROR_RETURN (error, "Could not link visualisation thread elements");
gst_element_add_ghost_pad (vis_bin,
gst_element_get_pad (vis_cs, "src"), "src");
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
}
/* Creating our video output bin */
{
GST_PLAY_MAKE_OR_ERROR (video_thread, "thread", "video_thread", error);
g_hash_table_insert (play->priv->elements, "video_thread", video_thread);
gst_bin_add (GST_BIN (work_thread), video_thread);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Buffer queue for video data */
GST_PLAY_MAKE_OR_ERROR (video_queue, "queue", "video_queue", error);
g_hash_table_insert (play->priv->elements, "video_queue", video_queue);
GST_PLAY_MAKE_OR_ERROR (video_switch, "switch", "video_switch", error);
g_hash_table_insert (play->priv->elements, "video_switch", video_switch);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Colorspace conversion */
video_cs = gst_element_factory_make ("ffcolorspace", "video_cs");
if (!GST_IS_ELEMENT (video_cs)) {
video_cs = gst_element_factory_make ("colorspace", "video_cs");
if (!GST_IS_ELEMENT (video_cs))
{
gst_play_error_plugin ("colorspace", error);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
return FALSE;
}
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
}
g_hash_table_insert (play->priv->elements, "video_cs", video_cs);
/* Software colorbalance */
GST_PLAY_MAKE_OR_ERROR (video_balance, "videobalance", "video_balance", error);
g_hash_table_insert (play->priv->elements, "video_balance", video_balance);
/* Colorspace conversion */
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
balance_cs = gst_element_factory_make ("ffcolorspace", "balance_cs");
if (!GST_IS_ELEMENT (balance_cs)) {
balance_cs = gst_element_factory_make ("colorspace", "balance_cs");
if (!GST_IS_ELEMENT (balance_cs))
{
gst_play_error_plugin ("colorspace", error);
return FALSE;
}
}
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
g_hash_table_insert (play->priv->elements, "balance_cs", balance_cs);
/* Software scaling of video stream */
GST_PLAY_MAKE_OR_ERROR (video_scaler, "videoscale", "video_scaler", error);
g_hash_table_insert (play->priv->elements, "video_scaler", video_scaler);
g_signal_connect (gst_element_get_pad(video_scaler, "src"), "fixate",
G_CALLBACK (gst_play_video_fixate), play);
/* Placeholder for future video sink bin */
GST_PLAY_MAKE_OR_ERROR (video_sink, "fakesink", "video_sink", error);
g_hash_table_insert (play->priv->elements, "video_sink", video_sink);
gst_bin_add_many (GST_BIN (video_thread), video_queue, video_switch, video_cs,
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
video_balance, balance_cs, video_scaler, video_sink, NULL);
/* break down linking so we can figure out what might be failing */
if (!gst_element_link (video_queue, video_switch))
GST_PLAY_ERROR_RETURN (error, "Could not link video output thread (queue and switch)");
if (!gst_element_link (video_switch, video_cs))
GST_PLAY_ERROR_RETURN (error, "Could not link video output thread (switch and cs)");
if (!gst_element_link (video_cs, video_balance))
GST_PLAY_ERROR_RETURN (error, "Could not link video output thread (cs and balance)");
if (!gst_element_link (video_balance, balance_cs))
GST_PLAY_ERROR_RETURN (error, "Could not link video output thread (balance and balance_cs)");
if (!gst_element_link (balance_cs, video_scaler))
GST_PLAY_ERROR_RETURN (error, "Could not link video output thread (balance_cs and scaler)");
if (!gst_element_link (video_scaler, video_sink))
GST_PLAY_ERROR_RETURN (error, "Could not link video output thread (balance_cs and scaler)");
gst_element_add_ghost_pad (video_thread,
gst_element_get_pad (video_queue, "sink"),
"sink");
if (!gst_element_link (identity_cs, video_thread))
GST_PLAY_ERROR_RETURN (error, "Could not link video output thread elements");
}
/* Creating our audio output bin
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
{ queue ! fakesink } */
{
GST_PLAY_MAKE_OR_ERROR (audio_thread, "thread", "audio_thread", error);
g_hash_table_insert (play->priv->elements, "audio_thread", audio_thread);
gst_bin_add (GST_BIN (work_thread), audio_thread);
/* Buffer queue for our audio thread */
GST_PLAY_MAKE_OR_ERROR (audio_queue, "queue", "audio_queue", error);
g_hash_table_insert (play->priv->elements, "audio_queue", audio_queue);
/* Volume control */
GST_PLAY_MAKE_OR_ERROR (volume, "volume", "volume", error);
g_hash_table_insert (play->priv->elements, "volume", volume);
g_signal_connect (gst_element_get_pad(volume, "src"), "fixate",
G_CALLBACK (gst_play_audio_fixate), play);
/* Placeholder for future audio sink bin */
GST_PLAY_MAKE_OR_ERROR (audio_sink, "fakesink", "audio_sink", error);
g_hash_table_insert (play->priv->elements, "audio_sink", audio_sink);
gst_bin_add_many (GST_BIN (audio_thread), audio_queue, volume, audio_sink, NULL);
if (!gst_element_link_many (audio_queue, volume, audio_sink, NULL))
GST_PLAY_ERROR_RETURN (error, "Could not link audio output thread elements");
gst_element_add_ghost_pad (audio_thread,
gst_element_get_pad (audio_queue, "sink"),
"sink");
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
gst_pad_link (tee_pad2, gst_element_get_pad (audio_queue, "sink"));
}
return TRUE;
}
static void
gst_play_have_video_size (GstElement *element, gint width,
gint height, GstPlay *play)
{
g_return_if_fail (play != NULL);
g_return_if_fail (GST_IS_PLAY (play));
g_signal_emit (G_OBJECT (play), gst_play_signals[HAVE_VIDEO_SIZE],
0, width, height);
}
static gboolean
gst_play_tick_callback (GstPlay *play)
{
GstFormat format = GST_FORMAT_TIME;
gboolean q = FALSE;
GstElement *audio_sink_element = NULL;
g_return_val_if_fail (play != NULL, FALSE);
/* just return without updating the UI when we are in the middle of seeking */
if (play->priv->tick_unblock_remaining > 0)
{
play->priv->tick_unblock_remaining -= TICK_INTERVAL_MSEC;
return TRUE;
}
if (!GST_IS_PLAY (play)) {
play->priv->tick_id = 0;
return FALSE;
}
audio_sink_element = g_hash_table_lookup (play->priv->elements,
"audio_sink_element");
if (!GST_IS_ELEMENT (audio_sink_element)) {
play->priv->tick_id = 0;
return FALSE;
}
q = gst_element_query (audio_sink_element, GST_QUERY_POSITION, &format,
&(play->priv->time_nanos));
if (q)
g_signal_emit (G_OBJECT (play), gst_play_signals[TIME_TICK],
0,play->priv->time_nanos);
if (GST_STATE (GST_ELEMENT (play)) == GST_STATE_PLAYING)
return TRUE;
else {
play->priv->tick_id = 0;
return FALSE;
}
}
static gboolean
gst_play_get_length_callback (GstPlay *play)
{
GstElement *audio_sink_element, *video_sink_element;
GstFormat format = GST_FORMAT_TIME;
gint64 value;
gboolean q = FALSE;
g_return_val_if_fail (play != NULL, FALSE);
g_return_val_if_fail (GST_IS_PLAY (play), FALSE);
/* We try to get length from all real sink elements */
audio_sink_element = g_hash_table_lookup (play->priv->elements,
"audio_sink_element");
video_sink_element = g_hash_table_lookup (play->priv->elements,
"video_sink_element");
if (!GST_IS_ELEMENT (audio_sink_element) &&
!GST_IS_ELEMENT (video_sink_element)) {
play->priv->length_id = 0;
return FALSE;
}
/* Audio first and then Video */
if (GST_IS_ELEMENT (audio_sink_element))
q = gst_element_query (audio_sink_element, GST_QUERY_TOTAL, &format,
&value);
if ( (!q) && (GST_IS_ELEMENT (video_sink_element)) )
q = gst_element_query (video_sink_element, GST_QUERY_TOTAL, &format,
&value);
if (q) {
play->priv->length_nanos = value;
g_signal_emit (G_OBJECT (play), gst_play_signals[STREAM_LENGTH],
0,play->priv->length_nanos);
play->priv->length_id = 0;
return FALSE;
}
play->priv->get_length_attempt++;
/* We try 16 times */
if (play->priv->get_length_attempt > 15) {
play->priv->length_id = 0;
return FALSE;
}
else
return TRUE;
}
static GstCaps *
gst_play_video_fixate (GstPad *pad, const GstCaps *caps, gpointer user_data)
{
GstStructure *structure;
GstCaps *newcaps;
GST_DEBUG ("video fixate %p %" GST_PTR_FORMAT, pad, caps);
if (gst_caps_get_size (caps) > 1) return NULL;
newcaps = gst_caps_copy (caps);
structure = gst_caps_get_structure (newcaps, 0);
if (gst_caps_structure_fixate_field_nearest_int (structure, "width", 320)) {
return newcaps;
}
if (gst_caps_structure_fixate_field_nearest_int (structure, "height", 240)) {
return newcaps;
}
if (gst_caps_structure_fixate_field_nearest_double (structure, "framerate",
30.0)) {
return newcaps;
}
/* failed to fixate */
gst_caps_free (newcaps);
return NULL;
}
static GstCaps *
gst_play_audio_fixate (GstPad *pad, const GstCaps *caps, gpointer user_data)
{
GstCaps *newcaps;
GstStructure *structure;
GST_DEBUG ("audio fixate %p %" GST_PTR_FORMAT, pad, caps);
newcaps = gst_caps_new_full (gst_structure_copy (
gst_caps_get_structure (caps, 0)), NULL);
structure = gst_caps_get_structure (newcaps, 0);
if (gst_caps_structure_fixate_field_nearest_int (structure, "rate", 44100)) {
return newcaps;
}
if (gst_caps_structure_fixate_field_nearest_int (structure, "depth", 16)) {
return newcaps;
}
if (gst_caps_structure_fixate_field_nearest_int (structure, "width", 16)) {
return newcaps;
}
if (gst_caps_structure_fixate_field_nearest_int (structure, "channels", 2)) {
return newcaps;
}
gst_caps_free (newcaps);
return NULL;
}
static void
gst_play_state_change (GstElement *element, GstElementState old,
GstElementState state)
{
GstPlay *play;
g_return_if_fail (element != NULL);
g_return_if_fail (GST_IS_PLAY (element));
play = GST_PLAY (element);
if (state == GST_STATE_PLAYING) {
if (play->priv->tick_id) {
g_source_remove (play->priv->tick_id);
play->priv->tick_id = 0;
}
play->priv->tick_id = g_timeout_add (TICK_INTERVAL_MSEC,
(GSourceFunc) gst_play_tick_callback,
play);
play->priv->get_length_attempt = 0;
if (play->priv->length_id) {
g_source_remove (play->priv->length_id);
play->priv->length_id = 0;
}
play->priv->length_id = g_timeout_add (TICK_INTERVAL_MSEC,
(GSourceFunc) gst_play_get_length_callback,
play);
}
else {
if (play->priv->tick_id) {
g_source_remove (play->priv->tick_id);
play->priv->tick_id = 0;
}
if (play->priv->length_id) {
g_source_remove (play->priv->length_id);
play->priv->length_id = 0;
}
}
if (GST_ELEMENT_CLASS (parent_class)->state_change)
GST_ELEMENT_CLASS (parent_class)->state_change (element, old, state);
}
static void
gst_play_identity_handoff (GstElement *identity, GstBuffer *buf, GstPlay *play)
{
g_signal_handler_disconnect (G_OBJECT (identity), play->priv->handoff_hid);
play->priv->handoff_hid = 0;
gst_play_connect_visualization (play, FALSE);
}
/* =========================================== */
/* */
/* Init & Dispose & Class init */
/* */
/* =========================================== */
static void
gst_play_dispose (GObject *object)
{
GstPlay *play;
g_return_if_fail (object != NULL);
g_return_if_fail (GST_IS_PLAY (object));
play = GST_PLAY (object);
if (play->priv->length_id) {
g_source_remove (play->priv->length_id);
play->priv->length_id = 0;
}
if (play->priv->tick_id) {
g_source_remove (play->priv->tick_id);
play->priv->tick_id = 0;
}
if (play->priv->location) {
g_free (play->priv->location);
play->priv->location = NULL;
}
if (play->priv->elements) {
g_hash_table_destroy (play->priv->elements);
play->priv->elements = NULL;
}
G_OBJECT_CLASS (parent_class)->dispose (object);
}
static void
gst_play_init (GstPlay *play)
{
play->priv = g_new0 (GstPlayPrivate, 1);
play->priv->location = NULL;
play->priv->length_nanos = 0;
play->priv->time_nanos = 0;
play->priv->elements = g_hash_table_new (g_str_hash, g_str_equal);
play->priv->error = NULL;
play->priv->debug = NULL;
if (!gst_play_pipeline_setup (play, &play->priv->error))
{
g_warning ("libgstplay: failed initializing pipeline, error: %s",
play->priv->error->message);
}
}
static void
gst_play_class_init (GstPlayClass *klass)
{
GObjectClass *gobject_class = G_OBJECT_CLASS (klass);
GstElementClass *element_class = GST_ELEMENT_CLASS (klass);
parent_class = g_type_class_peek_parent (klass);
gobject_class->dispose = gst_play_dispose;
element_class->state_change = gst_play_state_change;
gst_play_signals[TIME_TICK] =
g_signal_new ("time-tick", G_TYPE_FROM_CLASS (klass), G_SIGNAL_RUN_FIRST,
G_STRUCT_OFFSET (GstPlayClass, time_tick), NULL, NULL,
gst_marshal_VOID__INT64, G_TYPE_NONE, 1, G_TYPE_INT64);
gst_play_signals[STREAM_LENGTH] =
g_signal_new ("stream-length", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_FIRST,
G_STRUCT_OFFSET (GstPlayClass, stream_length), NULL, NULL,
gst_marshal_VOID__INT64, G_TYPE_NONE, 1, G_TYPE_INT64);
gst_play_signals[HAVE_VIDEO_SIZE] =
g_signal_new ("have-video-size", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_FIRST,
G_STRUCT_OFFSET (GstPlayClass, have_video_size), NULL, NULL,
gst_marshal_VOID__INT_INT, G_TYPE_NONE, 2,
G_TYPE_INT, G_TYPE_INT);
}
/* ======================================================= */
/* */
/* Public Methods */
/* */
/* ======================================================= */
/**
* gst_play_set_location:
* @play: a #GstPlay.
* @location: a const #char* indicating location to play
*
* Set location of @play to @location.
*
* Returns: TRUE if location was set successfully.
*/
gboolean
gst_play_set_location (GstPlay *play, const char *location)
{
GstElement *work_thread, *source, *autoplugger;
GstElement *audioconvert, *identity;
g_return_val_if_fail (play != NULL, FALSE);
g_return_val_if_fail (GST_IS_PLAY (play), FALSE);
if (play->priv->location)
g_free (play->priv->location);
play->priv->location = g_strdup (location);
if (GST_STATE (GST_ELEMENT (play)) != GST_STATE_READY)
gst_element_set_state (GST_ELEMENT (play), GST_STATE_READY);
work_thread = g_hash_table_lookup (play->priv->elements, "work_thread");
if (!GST_IS_ELEMENT (work_thread))
return FALSE;
source = g_hash_table_lookup (play->priv->elements, "source");
if (!GST_IS_ELEMENT (source))
return FALSE;
autoplugger = g_hash_table_lookup (play->priv->elements, "autoplugger");
if (!GST_IS_ELEMENT (autoplugger))
return FALSE;
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
audioconvert = g_hash_table_lookup (play->priv->elements, "audioconvert");
if (!GST_IS_ELEMENT (audioconvert))
return FALSE;
identity = g_hash_table_lookup (play->priv->elements, "identity");
if (!GST_IS_ELEMENT (identity))
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
return FALSE;
/* Spider can autoplugg only once. We remove the actual one and put a new
autoplugger */
gst_element_unlink (source, autoplugger);
gst_element_unlink (autoplugger, identity);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
gst_element_unlink (autoplugger, audioconvert);
gst_bin_remove (GST_BIN (work_thread), autoplugger);
autoplugger = gst_element_factory_make ("spider", "autoplugger");
if (!GST_IS_ELEMENT (autoplugger))
return FALSE;
gst_bin_add (GST_BIN (work_thread), autoplugger);
gst_element_link (source, autoplugger);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
gst_element_link (autoplugger, audioconvert);
gst_element_link (autoplugger, identity);
g_hash_table_replace (play->priv->elements, "autoplugger", autoplugger);
/* FIXME: Why don't we have an interface to do that kind of stuff ? */
g_object_set (G_OBJECT (source), "location", play->priv->location, NULL);
play->priv->length_nanos = 0LL;
play->priv->time_nanos = 0LL;
g_signal_emit (G_OBJECT (play), gst_play_signals[STREAM_LENGTH], 0, 0LL);
g_signal_emit (G_OBJECT (play), gst_play_signals[TIME_TICK], 0, 0LL);
return TRUE;
}
/**
* gst_play_get_location:
* @play: a #GstPlay.
*
* Get current location of @play.
*
* Returns: a const #char* pointer to current location.
*/
char *
gst_play_get_location (GstPlay *play)
{
g_return_val_if_fail (play != NULL, NULL);
g_return_val_if_fail (GST_IS_PLAY (play), NULL);
return g_strdup (play->priv->location);
}
/**
* gst_play_seek_to_time:
* @play: a #GstPlay.
* @time_nanos: a #gint64 indicating a time position.
*
* Performs a seek on @play until @time_nanos.
*/
gboolean
gst_play_seek_to_time (GstPlay * play, gint64 time_nanos)
{
GstElement *audio_seek_element, *video_seek_element, *audio_sink_element;
g_return_val_if_fail (play != NULL, FALSE);
g_return_val_if_fail (GST_IS_PLAY (play), FALSE);
if (time_nanos < 0LL)
time_nanos = 0LL;
audio_seek_element = g_hash_table_lookup (play->priv->elements,
"audioconvert");
audio_sink_element = g_hash_table_lookup (play->priv->elements,
"audio_sink_element");
video_seek_element = g_hash_table_lookup (play->priv->elements,
"identity");
if (GST_IS_ELEMENT (audio_seek_element) &&
GST_IS_ELEMENT (video_seek_element) &&
GST_IS_ELEMENT (audio_sink_element)) {
gboolean s = FALSE;
/* HACK: block tick signal from idler for 500 msec */
/* GStreamer can't currently report when seeking is finished,
so we just chose a .5 sec default block time */
play->priv->tick_unblock_remaining = 500;
s = gst_element_seek (video_seek_element, GST_FORMAT_TIME |
GST_SEEK_METHOD_SET | GST_SEEK_FLAG_FLUSH,
time_nanos);
if (!s) {
s = gst_element_seek (audio_seek_element, GST_FORMAT_TIME |
GST_SEEK_METHOD_SET | GST_SEEK_FLAG_FLUSH,
time_nanos);
}
if (s) {
GstFormat format = GST_FORMAT_TIME;
gboolean q = FALSE;
q = gst_element_query (audio_sink_element, GST_QUERY_POSITION, &format,
&(play->priv->time_nanos));
if (q)
g_signal_emit (G_OBJECT (play), gst_play_signals[TIME_TICK],
0,play->priv->time_nanos);
}
}
return TRUE;
}
/**
* gst_play_set_data_src:
* @play: a #GstPlay.
* @data_src: a #GstElement.
*
* Set @data_src as the source element of @play.
*
* Returns: TRUE if call succeeded.
*/
gboolean
gst_play_set_data_src (GstPlay *play, GstElement *data_src)
{
GstElement *work_thread, *old_data_src, *autoplugger;
g_return_val_if_fail (play != NULL, FALSE);
g_return_val_if_fail (GST_IS_PLAY (play), FALSE);
/* We bring back the pipeline to READY */
if (GST_STATE (GST_ELEMENT (play)) != GST_STATE_READY)
gst_element_set_state (GST_ELEMENT (play), GST_STATE_READY);
/* Getting needed objects */
work_thread = g_hash_table_lookup (play->priv->elements, "work_thread");
if (!GST_IS_ELEMENT (work_thread))
return FALSE;
old_data_src = g_hash_table_lookup (play->priv->elements, "source");
if (!GST_IS_ELEMENT (old_data_src))
return FALSE;
autoplugger = g_hash_table_lookup (play->priv->elements, "autoplugger");
if (!GST_IS_ELEMENT (autoplugger))
return FALSE;
/* Unlinking old source from autoplugger, removing it from pipeline, adding
the new one and connecting it to autoplugger FIXME: we should put a new
autoplugger here as spider can autoplugg only once */
gst_element_unlink (old_data_src, autoplugger);
gst_bin_remove (GST_BIN (work_thread), old_data_src);
gst_bin_add (GST_BIN (work_thread), data_src);
gst_element_link (data_src, autoplugger);
g_hash_table_replace (play->priv->elements, "source", data_src);
return TRUE;
}
/**
* gst_play_set_video_sink:
* @play: a #GstPlay.
* @video_sink: a #GstElement.
*
* Set @video_sink as the video sink element of @play.
*
* Returns: TRUE if call succeeded.
*/
gboolean
gst_play_set_video_sink (GstPlay *play, GstElement *video_sink)
{
GstElement *video_thread, *old_video_sink, *video_scaler, *video_sink_element;
g_return_val_if_fail (play != NULL, FALSE);
g_return_val_if_fail (GST_IS_PLAY (play), FALSE);
g_return_val_if_fail (video_sink != NULL, FALSE);
g_return_val_if_fail (GST_IS_ELEMENT (video_sink), FALSE);
/* We bring back the pipeline to READY */
if (GST_STATE (GST_ELEMENT (play)) != GST_STATE_READY)
gst_element_set_state (GST_ELEMENT (play), GST_STATE_READY);
/* Getting needed objects */
video_thread = g_hash_table_lookup (play->priv->elements, "video_thread");
if (!GST_IS_ELEMENT (video_thread))
return FALSE;
old_video_sink = g_hash_table_lookup (play->priv->elements, "video_sink");
if (!GST_IS_ELEMENT (old_video_sink))
return FALSE;
video_scaler = g_hash_table_lookup (play->priv->elements, "video_scaler");
if (!GST_IS_ELEMENT (video_scaler))
return FALSE;
/* Unlinking old video sink from video scaler, removing it from pipeline,
adding the new one and linking it */
gst_element_unlink (video_scaler, old_video_sink);
gst_bin_remove (GST_BIN (video_thread), old_video_sink);
gst_bin_add (GST_BIN (video_thread), video_sink);
gst_element_link (video_scaler, video_sink);
g_hash_table_replace (play->priv->elements, "video_sink", video_sink);
video_sink_element = gst_play_get_sink_element (play, video_sink,
GST_PLAY_SINK_TYPE_VIDEO);
if (GST_IS_ELEMENT (video_sink_element)) {
g_hash_table_replace (play->priv->elements, "video_sink_element",
video_sink_element);
if (GST_IS_X_OVERLAY (video_sink_element)) {
g_signal_connect (G_OBJECT (video_sink_element),
"desired_size_changed",
G_CALLBACK (gst_play_have_video_size), play);
}
}
gst_element_set_state (video_sink, GST_STATE (GST_ELEMENT(play)));
return TRUE;
}
/**
* gst_play_set_audio_sink:
* @play: a #GstPlay.
* @audio_sink: a #GstElement.
*
* Set @audio_sink as the audio sink element of @play.
*
* Returns: TRUE if call succeeded.
*/
gboolean
gst_play_set_audio_sink (GstPlay *play, GstElement *audio_sink)
{
GstElement *old_audio_sink, *audio_thread, *volume, *audio_sink_element;
g_return_val_if_fail (play != NULL, FALSE);
g_return_val_if_fail (GST_IS_PLAY (play), FALSE);
g_return_val_if_fail (audio_sink != NULL, FALSE);
g_return_val_if_fail (GST_IS_ELEMENT (audio_sink), FALSE);
/* We bring back the pipeline to READY */
if (GST_STATE (GST_ELEMENT (play)) != GST_STATE_READY)
gst_element_set_state (GST_ELEMENT (play), GST_STATE_READY);
/* Getting needed objects */
old_audio_sink = g_hash_table_lookup (play->priv->elements, "audio_sink");
if (!GST_IS_ELEMENT (old_audio_sink))
return FALSE;
audio_thread = g_hash_table_lookup (play->priv->elements, "audio_thread");
if (!GST_IS_ELEMENT (audio_thread))
return FALSE;
volume = g_hash_table_lookup (play->priv->elements, "volume");
if (!GST_IS_ELEMENT (volume))
return FALSE;
/* Unlinking old audiosink, removing it from pipeline, putting the new one
and linking it */
gst_element_unlink (volume, old_audio_sink);
gst_bin_remove (GST_BIN (audio_thread), old_audio_sink);
gst_bin_add (GST_BIN (audio_thread), audio_sink);
gst_element_link (volume, audio_sink);
g_hash_table_replace (play->priv->elements, "audio_sink", audio_sink);
audio_sink_element = gst_play_get_sink_element (play, audio_sink,
GST_PLAY_SINK_TYPE_AUDIO);
if (GST_IS_ELEMENT (audio_sink_element)) {
g_hash_table_replace (play->priv->elements, "audio_sink_element",
audio_sink_element);
}
gst_element_set_state (audio_sink, GST_STATE (GST_ELEMENT(play)));
return TRUE;
}
/**
* gst_play_set_visualization:
* @play: a #GstPlay.
* @element: a #GstElement.
*
* Set @video_sink as the video sink element of @play.
*
* Returns: TRUE if call succeeded.
*/
gboolean
gst_play_set_visualization (GstPlay *play, GstElement *vis_element)
{
GstElement *vis_bin, *vis_queue, *old_vis_element, *vis_cs;
gboolean was_playing = FALSE;
g_return_val_if_fail (play != NULL, FALSE);
g_return_val_if_fail (GST_IS_PLAY (play), FALSE);
g_return_val_if_fail (vis_element != NULL, FALSE);
g_return_val_if_fail (GST_IS_ELEMENT (vis_element), FALSE);
/* Getting needed objects */
vis_bin = g_hash_table_lookup (play->priv->elements, "vis_bin");
if (!GST_IS_ELEMENT (vis_bin))
return FALSE;
vis_queue = g_hash_table_lookup (play->priv->elements, "vis_queue");
if (!GST_IS_ELEMENT (vis_queue))
return FALSE;
old_vis_element = g_hash_table_lookup (play->priv->elements,
"vis_element");
if (!GST_IS_ELEMENT (old_vis_element))
return FALSE;
vis_cs = g_hash_table_lookup (play->priv->elements, "vis_cs");
if (!GST_IS_ELEMENT (vis_cs))
return FALSE;
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* We bring back the pipeline to PAUSED */
if (GST_STATE (GST_ELEMENT (play)) == GST_STATE_PLAYING) {
gst_element_set_state (GST_ELEMENT (play), GST_STATE_PAUSED);
was_playing = TRUE;
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
}
gst_element_unlink_many (vis_queue, old_vis_element, vis_cs, NULL);
gst_bin_remove (GST_BIN (vis_bin), old_vis_element);
gst_bin_add (GST_BIN (vis_bin), vis_element);
gst_element_link_many (vis_queue, vis_element, vis_cs, NULL);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
g_hash_table_replace (play->priv->elements, "vis_element", vis_element);
if (was_playing)
gst_element_set_state (GST_ELEMENT (play), GST_STATE_PLAYING);
return TRUE;
}
/**
* gst_play_connect_visualization:
* @play: a #GstPlay.
* @connect: a #gboolean indicating wether or not
* visualization should be connected.
*
* Connect or disconnect visualization bin in @play.
*
* Returns: TRUE if call succeeded.
*/
gboolean
gst_play_connect_visualization (GstPlay * play, gboolean connect)
{
GstElement *video_thread, *vis_queue, *vis_bin, *video_switch, *identity;
GstPad *tee_pad1, *vis_queue_pad, *vis_bin_pad, *switch_pad;
gboolean was_playing = FALSE;
g_return_val_if_fail (play != NULL, FALSE);
g_return_val_if_fail (GST_IS_PLAY (play), FALSE);
/* Until i fix the switch */
return TRUE;
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Getting needed objects */
video_thread = g_hash_table_lookup (play->priv->elements, "video_thread");
if (!GST_IS_ELEMENT (video_thread))
return FALSE;
vis_bin = g_hash_table_lookup (play->priv->elements, "vis_bin");
if (!GST_IS_ELEMENT (vis_bin))
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
return FALSE;
vis_queue = g_hash_table_lookup (play->priv->elements, "vis_queue");
if (!GST_IS_ELEMENT (vis_queue))
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
return FALSE;
video_switch = g_hash_table_lookup (play->priv->elements, "video_switch");
if (!GST_IS_ELEMENT (video_switch))
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
return FALSE;
identity = g_hash_table_lookup (play->priv->elements, "identity");
if (!GST_IS_ELEMENT (identity))
return FALSE;
tee_pad1 = g_hash_table_lookup (play->priv->elements, "tee_pad1");
if (!GST_IS_PAD (tee_pad1))
return FALSE;
vis_queue_pad = gst_element_get_pad (vis_queue, "sink");
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Check if the vis element is in the pipeline. That means visualization is
connected already */
if (gst_element_get_managing_bin (vis_bin)) {
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* If we are supposed to connect then nothing to do we return */
if (connect) {
return TRUE;
}
/* Disconnecting visualization */
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* We bring back the pipeline to PAUSED */
if (GST_STATE (GST_ELEMENT (play)) == GST_STATE_PLAYING) {
gst_element_set_state (GST_ELEMENT (play), GST_STATE_PAUSED);
was_playing = TRUE;
}
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Unlinking, removing */
gst_pad_unlink (tee_pad1, vis_queue_pad);
vis_bin_pad = gst_element_get_pad (vis_bin, "src");
switch_pad = gst_pad_get_peer (vis_bin_pad);
gst_pad_unlink (vis_bin_pad, switch_pad);
gst_element_release_request_pad (video_switch, switch_pad);
gst_object_ref (GST_OBJECT (vis_bin));
gst_bin_remove (GST_BIN (video_thread), vis_bin);
}
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
else {
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* If we are supposed to disconnect then nothing to do we return */
if (!connect) {
return TRUE;
}
/* Connecting visualization */
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* We bring back the pipeline to PAUSED */
if (GST_STATE (GST_ELEMENT (play)) == GST_STATE_PLAYING) {
gst_element_set_state (GST_ELEMENT (play), GST_STATE_PAUSED);
was_playing = TRUE;
}
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
/* Adding, linking */
play->priv->handoff_hid = g_signal_connect (G_OBJECT (identity),
"handoff",
G_CALLBACK (gst_play_identity_handoff), play);
gst_bin_add (GST_BIN (video_thread), vis_bin);
gst_pad_link (tee_pad1, vis_queue_pad);
gst_element_link (vis_bin, video_switch);
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
}
if (was_playing)
gst_element_set_state (GST_ELEMENT (play), GST_STATE_PLAYING);
return TRUE;
}
/**
* gst_play_get_sink_element:
* @play: a #GstPlay.
* @element: a #GstElement.
* @sink_type: a #GstPlaySinkType.
*
* Searches recursively for a sink #GstElement with
* type @sink_type in @element which is supposed to be a #GstBin.
*
* Returns: the sink #GstElement of @element.
*/
GstElement *
gst_play_get_sink_element (GstPlay *play,
gst-libs/gst/play/gstplay.c: Reworked the pipeline from scratch. Visualization is back and switch went out as i reali... Original commit message from CVS: 2004-01-23 Julien MOUTTE <julien@moutte.net> * gst-libs/gst/play/gstplay.c: (gst_play_pipeline_setup), (gst_play_set_location), (gst_play_seek_to_time), (gst_play_set_audio_sink), (gst_play_set_visualization), (gst_play_connect_visualization), (gst_play_get_sink_element): Reworked the pipeline from scratch. Visualization is back and switch went out as i realized it was not possible to use the way i wanted. * sys/ximage/ximagesink.c: (gst_ximagesink_imagepool_clear), (gst_ximagesink_change_state), (gst_ximagesink_dispose): Move xcontext clearing in state change from READY to NULL. So that one can clean the X ressources keeping the element. * sys/xvimage/xvimagesink.c: (gst_xvimagesink_xcontext_get), (gst_xvimagesink_imagepool_clear), (gst_xvimagesink_change_state), (gst_xvimagesink_colorbalance_set_value), (gst_xvimagesink_colorbalance_get_value), (gst_xvimagesink_set_property), (gst_xvimagesink_dispose), (gst_xvimagesink_init): Same xcontext cleaning than ximagesink in state change from READY to NULL and fixed some stupid bugs in colorbalance get/set values. Also added the following feature : when nobody tries to set some values to the colorbalance levels before the xcontext is grabbed, then when creating channels list from Xv attributes we set the internal values to the Xv defaults. This way we handle buggy Xv drivers that set default hue values far from the middle of the range (Thanks to Jon Trowbridge for pointing that issue). * sys/xvimage/xvimagesink.h: Adding a cb_changed boolean to know if colorbalance levels have been set before xcontext is grabbed.
2004-01-22 23:54:34 +00:00
GstElement *element, GstPlaySinkType sink_type)
{
GList *elements = NULL;
const GList *pads = NULL;
gboolean has_src, has_correct_type;
g_return_val_if_fail (play != NULL, NULL);
g_return_val_if_fail (element != NULL, NULL);
g_return_val_if_fail (GST_IS_PLAY (play), NULL);
g_return_val_if_fail (GST_IS_ELEMENT (element), NULL);
if (!GST_IS_BIN (element)) {
/* since its not a bin, we'll presume this
* element is a sink element */
return element;
}
elements = (GList *) gst_bin_get_list (GST_BIN (element));
/* traverse all elements looking for a src pad */
while (elements) {
element = GST_ELEMENT (elements->data);
/* Recursivity :) */
if (GST_IS_BIN (element)) {
element = gst_play_get_sink_element (play, element, sink_type);
if (GST_IS_ELEMENT (element))
return element;
}
else {
pads = gst_element_get_pad_list (element);
has_src = FALSE;
has_correct_type = FALSE;
while (pads) {
/* check for src pad */
if (GST_PAD_DIRECTION (GST_PAD (pads->data)) == GST_PAD_SRC) {
has_src = TRUE;
break;
}
else {
/* If not a src pad checking caps */
GstCaps *caps;
GstStructure *structure;
gboolean has_video_cap = FALSE;
gboolean has_audio_cap = FALSE;
caps = gst_pad_get_caps (GST_PAD (pads->data));
structure = gst_caps_get_structure (caps, 0);
if (strcmp (gst_structure_get_name (structure),
"audio/x-raw-int") == 0) {
has_audio_cap = TRUE;
}
if (strcmp (gst_structure_get_name (structure),
"video/x-raw-yuv") == 0 ||
strcmp (gst_structure_get_name (structure),
"video/x-raw-rgb") == 0) {
has_video_cap = TRUE;
}
gst_caps_free (caps);
switch (sink_type) {
case GST_PLAY_SINK_TYPE_AUDIO:
if (has_audio_cap)
has_correct_type = TRUE;
break;
case GST_PLAY_SINK_TYPE_VIDEO:
if (has_video_cap)
has_correct_type = TRUE;
break;
case GST_PLAY_SINK_TYPE_ANY:
if ((has_video_cap) || (has_audio_cap))
has_correct_type = TRUE;
break;
default:
has_correct_type = FALSE;
}
}
pads = g_list_next (pads);
}
if ((!has_src) && (has_correct_type))
return element;
}
elements = g_list_next (elements);
}
/* we didn't find a sink element */
return NULL;
}
GstPlay *
gst_play_new (GError **error)
{
GstPlay *play = g_object_new (GST_TYPE_PLAY, NULL);
if (play->priv->error)
{
if (error)
{
*error = play->priv->error;
play->priv->error = NULL;
}
else
{
g_warning ("Error creating GstPlay object.\n%s", play->priv->error->message);
g_error_free (play->priv->error);
}
}
return play;
}
/* =========================================== */
/* */
/* Object typing & Creation */
/* */
/* =========================================== */
GType
gst_play_get_type (void)
{
static GType play_type = 0;
if (!play_type) {
static const GTypeInfo play_info = {
sizeof (GstPlayClass),
NULL,
NULL,
(GClassInitFunc) gst_play_class_init,
NULL,
NULL,
sizeof (GstPlay),
0,
(GInstanceInitFunc) gst_play_init,
NULL
};
play_type = g_type_register_static (GST_TYPE_PIPELINE, "GstPlay",
&play_info, 0);
}
return play_type;
}