gstreamer/gst/gstpad.c

2939 lines
78 KiB
C
Raw Normal View History

/* GStreamer
* Copyright (C) 1999,2000 Erik Walthinsen <omega@cse.ogi.edu>
* 2000 Wim Taymans <wtay@chello.be>
*
* gstpad.c: Pads for linking elements together
*
* 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.
*/
#include "gst_private.h"
#include "gstpad.h"
#include "gstmarshal.h"
#include "gstutils.h"
#include "gstelement.h"
#include "gstbin.h"
#include "gstscheduler.h"
#include "gstevent.h"
#include "gstinfo.h"
#include "gsterror.h"
#include "gstvalue.h"
GST_DEBUG_CATEGORY_STATIC (debug_dataflow);
#define DEBUG_DATA(obj,data,notice) G_STMT_START{\
if (!data) { \
GST_CAT_DEBUG_OBJECT (debug_dataflow, obj, "NULL data value"); \
} else if (GST_IS_EVENT (data)) { \
GST_CAT_DEBUG_OBJECT (debug_dataflow, obj, "%s event %p (type %d, refcount %d)", notice, data, \
GST_EVENT_TYPE (data), GST_DATA_REFCOUNT_VALUE (data)); \
} else { \
GST_CAT_LOG_OBJECT (debug_dataflow, obj, "%s buffer %p (size %u, refcount %d)", notice, data, \
GST_BUFFER_SIZE (data), GST_BUFFER_REFCOUNT_VALUE (data)); \
} \
}G_STMT_END
#define GST_CAT_DEFAULT GST_CAT_PADS
enum
{
TEMPL_PAD_CREATED,
/* FILL ME */
TEMPL_LAST_SIGNAL
};
static GstObject *padtemplate_parent_class = NULL;
static guint gst_pad_template_signals[TEMPL_LAST_SIGNAL] = { 0 };
GType _gst_pad_type = 0;
/***** Start with the base GstPad class *****/
static void gst_pad_class_init (GstPadClass * klass);
static void gst_pad_init (GstPad * pad);
static void gst_pad_dispose (GObject * object);
static void gst_pad_set_pad_template (GstPad * pad, GstPadTemplate * templ);
#ifndef GST_DISABLE_LOADSAVE
static xmlNodePtr gst_pad_save_thyself (GstObject * object, xmlNodePtr parent);
#endif
static GstObject *pad_parent_class = NULL;
GType
gst_pad_get_type (void)
{
if (!_gst_pad_type) {
static const GTypeInfo pad_info = {
sizeof (GstPadClass), NULL, NULL,
(GClassInitFunc) gst_pad_class_init, NULL, NULL,
sizeof (GstPad),
0,
(GInstanceInitFunc) gst_pad_init, NULL
};
_gst_pad_type = g_type_register_static (GST_TYPE_OBJECT, "GstPad",
&pad_info, 0);
GST_DEBUG_CATEGORY_INIT (debug_dataflow, "GST_DATAFLOW",
GST_DEBUG_BOLD | GST_DEBUG_FG_GREEN, "dataflow inside pads");
}
return _gst_pad_type;
}
static void
gst_pad_class_init (GstPadClass * klass)
{
GObjectClass *gobject_class;
gobject_class = (GObjectClass *) klass;
pad_parent_class = g_type_class_ref (GST_TYPE_OBJECT);
gobject_class->dispose = GST_DEBUG_FUNCPTR (gst_pad_dispose);
}
static void
gst_pad_init (GstPad * pad)
{
/* all structs are initialized to NULL by glib */
}
static void
gst_pad_dispose (GObject * object)
{
GstPad *pad = GST_PAD (object);
gst_pad_set_pad_template (pad, NULL);
G_OBJECT_CLASS (pad_parent_class)->dispose (object);
}
/***** Then do the Real Pad *****/
/* Pad signals and args */
enum
{
REAL_LINKED,
REAL_UNLINKED,
REAL_REQUEST_LINK,
/* FILL ME */
REAL_LAST_SIGNAL
};
enum
{
REAL_ARG_0,
REAL_ARG_CAPS,
REAL_ARG_ACTIVE
/* FILL ME */
};
static void gst_real_pad_class_init (GstRealPadClass * klass);
static void gst_real_pad_init (GstRealPad * pad);
static void gst_real_pad_dispose (GObject * object);
static void gst_real_pad_set_property (GObject * object, guint prop_id,
const GValue * value, GParamSpec * pspec);
static void gst_real_pad_get_property (GObject * object, guint prop_id,
GValue * value, GParamSpec * pspec);
GType _gst_real_pad_type = 0;
static GstPad *real_pad_parent_class = NULL;
static guint gst_real_pad_signals[REAL_LAST_SIGNAL] = { 0 };
GType
gst_real_pad_get_type (void)
{
if (!_gst_real_pad_type) {
static const GTypeInfo pad_info = {
sizeof (GstRealPadClass), NULL, NULL,
(GClassInitFunc) gst_real_pad_class_init, NULL, NULL,
sizeof (GstRealPad),
0,
(GInstanceInitFunc) gst_real_pad_init, NULL
};
_gst_real_pad_type = g_type_register_static (GST_TYPE_PAD, "GstRealPad",
&pad_info, 0);
}
return _gst_real_pad_type;
}
static void
gst_real_pad_class_init (GstRealPadClass * klass)
{
GObjectClass *gobject_class;
GstObjectClass *gstobject_class;
gobject_class = (GObjectClass *) klass;
gstobject_class = (GstObjectClass *) klass;
real_pad_parent_class = g_type_class_ref (GST_TYPE_PAD);
gobject_class->dispose = GST_DEBUG_FUNCPTR (gst_real_pad_dispose);
gobject_class->set_property = GST_DEBUG_FUNCPTR (gst_real_pad_set_property);
gobject_class->get_property = GST_DEBUG_FUNCPTR (gst_real_pad_get_property);
gst_real_pad_signals[REAL_LINKED] =
g_signal_new ("linked", G_TYPE_FROM_CLASS (klass), G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (GstRealPadClass, linked), NULL, NULL,
gst_marshal_VOID__OBJECT, G_TYPE_NONE, 1, GST_TYPE_PAD);
gst_real_pad_signals[REAL_UNLINKED] =
g_signal_new ("unlinked", G_TYPE_FROM_CLASS (klass), G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (GstRealPadClass, unlinked), NULL, NULL,
gst_marshal_VOID__OBJECT, G_TYPE_NONE, 1, GST_TYPE_PAD);
gst_real_pad_signals[REAL_REQUEST_LINK] =
g_signal_new ("request_link", G_TYPE_FROM_CLASS (klass),
G_SIGNAL_RUN_LAST, G_STRUCT_OFFSET (GstRealPadClass, request_link), NULL,
NULL, gst_marshal_VOID__OBJECT, G_TYPE_NONE, 0);
g_object_class_install_property (G_OBJECT_CLASS (klass), REAL_ARG_ACTIVE,
g_param_spec_boolean ("active", "Active", "Whether the pad is active.",
TRUE, G_PARAM_READWRITE));
g_object_class_install_property (G_OBJECT_CLASS (klass), REAL_ARG_CAPS,
g_param_spec_boxed ("caps", "Caps", "The capabilities of the pad",
GST_TYPE_CAPS, G_PARAM_READABLE));
#ifndef GST_DISABLE_LOADSAVE
gstobject_class->save_thyself = GST_DEBUG_FUNCPTR (gst_pad_save_thyself);
#endif
gstobject_class->path_string_separator = ".";
}
static void
gst_real_pad_init (GstRealPad * pad)
{
pad->peer = NULL;
WARNING: Don't grab this updated unless you're really, REALLY sure. Original commit message from CVS: WARNING: Don't grab this updated unless you're really, REALLY sure. WARNING: Wait for the next one. Whole lotta changes here, including a few random bits: examples/*/Makefile: updated to use `libtool gcc`, not just `gcc` gst/ gstbuffer.h: updated to new flag style gst.c, gstdebug.h: added new debugging for function ptrs gstpipeline.c: set type of parent_class to the class, not the object gstthread.c: ditto plugins/ cdparanoia/cdparanoia.c: added an argument type, updated some defaults cobin/spindentity.c: updated to new do/while loopfunction style mp3encode/lame/gstlame.c: argument types, whole lotta lame options tests/: various changes Now, for the big changes: Once again, the scheduling system has changed. And once again, it broke a whole bunch of things. The gist of the change is that there is now a function pointer for gst_pad_push and gst_pad_pull, instead of a hard-wired function. Well, currently they are functions, but that's for debugging purposes only, they just call the function pointer after spewing lots of DEBUG(). This changed the GstPad structure a bit, and the GstPad API as well. Where elements used to provide chain() and pull() functions, they provide chain() and get() functions. gst_pad_set_pull[region]_function has been changed to get_pad_set_get[region]_function. This means all the elements out there that used to have pull functions need to be updated. The calls to that function have been changed in the normal elements, but the names of the functions passed is still _pull[region](), which is an aesthetic issue more than anything. As for what doesn't work yet, just about anything dealing with Connections is hosed, meaning threaded stuff won't work. This will be fixed about 12 hours from now, after I've slept, etc. The simplefake.c test works in both cothreaded and chained cases, but not much else will work due to the Connection problem. Needless to say, don't grab this unless you *need* these features *now*, else wait to update this stuff until tomorrow. I'm going to sleep now.
2000-12-16 10:18:09 +00:00
pad->chainfunc = NULL;
WARNING: Don't grab this updated unless you're really, REALLY sure. Original commit message from CVS: WARNING: Don't grab this updated unless you're really, REALLY sure. WARNING: Wait for the next one. Whole lotta changes here, including a few random bits: examples/*/Makefile: updated to use `libtool gcc`, not just `gcc` gst/ gstbuffer.h: updated to new flag style gst.c, gstdebug.h: added new debugging for function ptrs gstpipeline.c: set type of parent_class to the class, not the object gstthread.c: ditto plugins/ cdparanoia/cdparanoia.c: added an argument type, updated some defaults cobin/spindentity.c: updated to new do/while loopfunction style mp3encode/lame/gstlame.c: argument types, whole lotta lame options tests/: various changes Now, for the big changes: Once again, the scheduling system has changed. And once again, it broke a whole bunch of things. The gist of the change is that there is now a function pointer for gst_pad_push and gst_pad_pull, instead of a hard-wired function. Well, currently they are functions, but that's for debugging purposes only, they just call the function pointer after spewing lots of DEBUG(). This changed the GstPad structure a bit, and the GstPad API as well. Where elements used to provide chain() and pull() functions, they provide chain() and get() functions. gst_pad_set_pull[region]_function has been changed to get_pad_set_get[region]_function. This means all the elements out there that used to have pull functions need to be updated. The calls to that function have been changed in the normal elements, but the names of the functions passed is still _pull[region](), which is an aesthetic issue more than anything. As for what doesn't work yet, just about anything dealing with Connections is hosed, meaning threaded stuff won't work. This will be fixed about 12 hours from now, after I've slept, etc. The simplefake.c test works in both cothreaded and chained cases, but not much else will work due to the Connection problem. Needless to say, don't grab this unless you *need* these features *now*, else wait to update this stuff until tomorrow. I'm going to sleep now.
2000-12-16 10:18:09 +00:00
pad->getfunc = NULL;
pad->ghostpads = NULL;
pad->caps = NULL;
pad->linkfunc = NULL;
pad->getcapsfunc = NULL;
pad->eventfunc = gst_pad_event_default;
pad->convertfunc = gst_pad_convert_default;
pad->queryfunc = gst_pad_query_default;
pad->intlinkfunc = gst_pad_get_internal_links_default;
pad->eventmaskfunc = gst_pad_get_event_masks_default;
pad->formatsfunc = gst_pad_get_formats_default;
pad->querytypefunc = gst_pad_get_query_types_default;
GST_FLAG_SET (pad, GST_PAD_DISABLED);
GST_FLAG_UNSET (pad, GST_PAD_NEGOTIATING);
pad->stream_lock = g_mutex_new ();
pad->stream_cond = g_cond_new ();
pad->block_cond = g_cond_new ();
gst_probe_dispatcher_init (&pad->probedisp);
}
static void
gst_real_pad_set_property (GObject * object, guint prop_id,
const GValue * value, GParamSpec * pspec)
{
g_return_if_fail (GST_IS_PAD (object));
switch (prop_id) {
case REAL_ARG_ACTIVE:
gst_pad_set_active (GST_PAD (object), g_value_get_boolean (value));
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
static void
gst_real_pad_get_property (GObject * object, guint prop_id,
GValue * value, GParamSpec * pspec)
{
g_return_if_fail (GST_IS_PAD (object));
switch (prop_id) {
case REAL_ARG_ACTIVE:
g_value_set_boolean (value, !GST_FLAG_IS_SET (object, GST_PAD_DISABLED));
break;
case REAL_ARG_CAPS:
g_value_set_boxed (value, GST_PAD_CAPS (GST_REAL_PAD (object)));
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
/* FIXME-0.9: Replace these custom functions with proper inheritance via _init
functions and object properties */
/**
* gst_pad_custom_new:
* @type: the #Gtype of the pad.
* @name: the name of the new pad.
* @direction: the #GstPadDirection of the pad.
*
* Creates a new pad with the given name and type in the given direction.
* If name is NULL, a guaranteed unique name (across all pads)
* will be assigned.
*
* Returns: a new #GstPad, or NULL in case of an error.
*/
GstPad *
gst_pad_custom_new (GType type, const gchar * name, GstPadDirection direction)
{
GstRealPad *pad;
pad = g_object_new (type, NULL);
gst_object_set_name (GST_OBJECT (pad), name);
GST_RPAD_DIRECTION (pad) = direction;
return GST_PAD (pad);
}
/**
* gst_pad_new:
* @name: the name of the new pad.
* @direction: the #GstPadDirection of the pad.
*
* Creates a new real pad with the given name in the given direction.
* If name is NULL, a guaranteed unique name (across all pads)
* will be assigned.
*
* Returns: a new #GstPad, or NULL in case of an error.
*/
GstPad *
gst_pad_new (const gchar * name, GstPadDirection direction)
{
return gst_pad_custom_new (gst_real_pad_get_type (), name, direction);
}
/**
* gst_pad_custom_new_from_template:
* @type: the custom #GType of the pad.
* @templ: the #GstPadTemplate to instantiate from.
* @name: the name of the new pad.
*
* Creates a new custom pad with the given name from the given template.
* If name is NULL, a guaranteed unique name (across all pads)
* will be assigned.
*
* Returns: a new #GstPad, or NULL in case of an error.
*/
GstPad *
gst_pad_custom_new_from_template (GType type, GstPadTemplate * templ,
const gchar * name)
{
GstPad *pad;
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (GST_IS_PAD_TEMPLATE (templ), NULL);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
pad = gst_pad_custom_new (type, name, templ->direction);
gst_pad_set_pad_template (pad, templ);
return pad;
}
/**
* gst_pad_new_from_template:
* @templ: the pad template to use
* @name: the name of the element
*
* Creates a new real pad with the given name from the given template.
* If name is NULL, a guaranteed unique name (across all pads)
* will be assigned.
*
* Returns: a new #GstPad, or NULL in case of an error.
*/
GstPad *
gst_pad_new_from_template (GstPadTemplate * templ, const gchar * name)
{
return gst_pad_custom_new_from_template (gst_real_pad_get_type (),
templ, name);
}
/**
* gst_pad_get_direction:
* @pad: a #GstPad to get the direction of.
*
* Gets the direction of the pad.
*
* Returns: the #GstPadDirection of the pad.
*/
GstPadDirection
gst_pad_get_direction (GstPad * pad)
{
GstPadDirection result;
g_return_val_if_fail (GST_IS_PAD (pad), GST_PAD_UNKNOWN);
/* since the direction cannot change at runtime we can
* safely read without locking. */
result = GST_PAD_DIRECTION (pad);
return result;
}
/**
* gst_pad_set_active:
* @pad: the #GstPad to activate or deactivate.
* @active: TRUE to activate the pad.
*
* Activates or deactivates the given pad.
*
* Returns: TRUE if the operation was successfull.
*/
gboolean
gst_pad_set_active (GstPad * pad, gboolean active)
{
GstRealPad *realpad;
gboolean old;
gboolean result = FALSE;
GstPadActivateFunction activatefunc;
g_return_val_if_fail (GST_IS_PAD (pad), result);
GST_LOCK (pad);
old = GST_PAD_IS_ACTIVE (pad);
if (old == active)
goto done;
realpad = GST_PAD_REALIZE (pad);
/* make sure data is disallowed when going inactive */
if (!active) {
GST_CAT_DEBUG (GST_CAT_PADS, "de-activating pad %s:%s",
GST_DEBUG_PAD_NAME (realpad));
GST_FLAG_SET (realpad, GST_PAD_DISABLED);
/* unlock blocked pads so element can resume and stop */
GST_PAD_BLOCK_SIGNAL (pad);
}
activatefunc = realpad->activatefunc;
if (activatefunc) {
/* unlock so element can sync */
GST_UNLOCK (pad);
result = activatefunc (pad, active);
/* and lock again */
GST_LOCK (pad);
} else {
result = TRUE;
}
/* when going to active alow data passing now */
if (active && result == TRUE) {
GST_CAT_DEBUG (GST_CAT_PADS, "activating pad %s:%s",
GST_DEBUG_PAD_NAME (realpad));
GST_FLAG_UNSET (realpad, GST_PAD_DISABLED);
}
done:
GST_UNLOCK (pad);
return result;
}
/**
* gst_pad_is_active:
* @pad: the #GstPad to query
*
* Query if a pad is active
*
* Returns: TRUE if the pad is active.
*/
gboolean
gst_pad_is_active (GstPad * pad)
{
gboolean result = FALSE;
g_return_val_if_fail (GST_IS_PAD (pad), result);
GST_LOCK (pad);
result = !GST_FLAG_IS_SET (pad, GST_PAD_DISABLED);
GST_UNLOCK (pad);
return result;
}
/**
* gst_pad_set_blocked_async:
* @pad: the #GstPad to block or unblock
* @blocked: boolean indicating we should block or unblock
* @callback: #GstPadBlockCallback that will be called when the
* operation succeeds.
* @user_data: user data passed to the callback
*
* Blocks or unblocks the dataflow on a pad. The provided callback
* is called when the operation succeeds. This can take a while as
* the pad can only become blocked when real dataflow is happening.
* When the pipeline is stalled, for example in PAUSED, this can
* take an indeterminate amount of time.
* You can pass NULL as the callback to make this call block. Be
* carefull with this blocking call as it might not return for
* reasons stated above.
*
* Returns: TRUE if the pad could be blocked. This function can fail
* wrong parameters were passed or the pad was already in the
* requested state.
*/
gboolean
gst_pad_set_blocked_async (GstPad * pad, gboolean blocked,
GstPadBlockCallback callback, gpointer user_data)
{
gboolean result = TRUE;
gboolean was_blocked;
GstRealPad *realpad;
/* only for real pads for now, need to transfer the lock
* to the realized pad */
g_return_val_if_fail (GST_IS_REAL_PAD (pad), FALSE);
GST_LOCK (pad);
realpad = GST_PAD_REALIZE (pad);
/* beware for turning flags into booleans */
was_blocked = !!GST_RPAD_IS_BLOCKED (realpad);
if (was_blocked == blocked) {
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad, "pad %s:%s was in right state",
GST_DEBUG_PAD_NAME (pad));
result = FALSE;
goto done;
}
if (blocked) {
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad, "blocking pad %s:%s",
GST_DEBUG_PAD_NAME (pad));
GST_FLAG_SET (realpad, GST_PAD_BLOCKED);
realpad->block_callback = callback;
realpad->block_data = user_data;
if (!callback) {
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad, "waiting for block");
GST_PAD_BLOCK_WAIT (pad);
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad, "blocked");
}
} else {
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad, "unblocking pad %s:%s",
GST_DEBUG_PAD_NAME (pad));
GST_FLAG_UNSET (realpad, GST_PAD_BLOCKED);
realpad->block_callback = callback;
realpad->block_data = user_data;
if (callback) {
GST_PAD_BLOCK_SIGNAL (pad);
} else {
GST_PAD_BLOCK_SIGNAL (pad);
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad, "waiting for unblock");
GST_PAD_BLOCK_WAIT (pad);
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad, "unblocked");
}
}
done:
GST_UNLOCK (pad);
return result;
}
/**
* gst_pad_set_blocked:
* @pad: the #GstPad to block or unblock
* @blocked: boolean indicating we should block or unblock
*
* Blocks or unblocks the dataflow on a pad. This function is
* a shortcut for @gst_pad_set_blocked_async() with a NULL
* callback.
*
* Returns: TRUE if the pad could be blocked. This function can fail
* wrong parameters were passed or the pad was already in the
* requested state.
*/
gboolean
gst_pad_set_blocked (GstPad * pad, gboolean blocked)
{
return gst_pad_set_blocked_async (pad, blocked, NULL, NULL);
}
/**
* gst_pad_is_blocked:
* @pad: the #GstPad to query
*
* Checks if the pad is blocked or not. This function returns the
* last requested state of the pad. It is not certain that the pad
* is actually blocked at this point.
*
* Returns: TRUE if the pad is blocked.
*/
gboolean
gst_pad_is_blocked (GstPad * pad)
{
gboolean result = FALSE;
g_return_val_if_fail (GST_IS_PAD (pad), result);
GST_LOCK (pad);
result = GST_FLAG_IS_SET (pad, GST_PAD_BLOCKED);
GST_UNLOCK (pad);
return result;
}
/**
* gst_pad_set_activate_function:
* @pad: a real sink #GstPad.
* @chain: the #GstPadActivateFunction to set.
*
* Sets the given activate function for the pad. The activate function is called to
* start or stop dataflow on a pad.
*/
void
gst_pad_set_activate_function (GstPad * pad, GstPadActivateFunction activate)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
GST_RPAD_ACTIVATEFUNC (pad) = activate;
GST_CAT_DEBUG (GST_CAT_PADS, "activatefunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (activate));
}
/**
WARNING: Don't grab this updated unless you're really, REALLY sure. Original commit message from CVS: WARNING: Don't grab this updated unless you're really, REALLY sure. WARNING: Wait for the next one. Whole lotta changes here, including a few random bits: examples/*/Makefile: updated to use `libtool gcc`, not just `gcc` gst/ gstbuffer.h: updated to new flag style gst.c, gstdebug.h: added new debugging for function ptrs gstpipeline.c: set type of parent_class to the class, not the object gstthread.c: ditto plugins/ cdparanoia/cdparanoia.c: added an argument type, updated some defaults cobin/spindentity.c: updated to new do/while loopfunction style mp3encode/lame/gstlame.c: argument types, whole lotta lame options tests/: various changes Now, for the big changes: Once again, the scheduling system has changed. And once again, it broke a whole bunch of things. The gist of the change is that there is now a function pointer for gst_pad_push and gst_pad_pull, instead of a hard-wired function. Well, currently they are functions, but that's for debugging purposes only, they just call the function pointer after spewing lots of DEBUG(). This changed the GstPad structure a bit, and the GstPad API as well. Where elements used to provide chain() and pull() functions, they provide chain() and get() functions. gst_pad_set_pull[region]_function has been changed to get_pad_set_get[region]_function. This means all the elements out there that used to have pull functions need to be updated. The calls to that function have been changed in the normal elements, but the names of the functions passed is still _pull[region](), which is an aesthetic issue more than anything. As for what doesn't work yet, just about anything dealing with Connections is hosed, meaning threaded stuff won't work. This will be fixed about 12 hours from now, after I've slept, etc. The simplefake.c test works in both cothreaded and chained cases, but not much else will work due to the Connection problem. Needless to say, don't grab this unless you *need* these features *now*, else wait to update this stuff until tomorrow. I'm going to sleep now.
2000-12-16 10:18:09 +00:00
* gst_pad_set_chain_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real sink #GstPad.
* @chain: the #GstPadChainFunction to set.
*
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* Sets the given chain function for the pad. The chain function is called to
* process a #GstData input buffer.
*/
void
gst_pad_set_chain_function (GstPad * pad, GstPadChainFunction chain)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
g_return_if_fail (GST_RPAD_DIRECTION (pad) == GST_PAD_SINK);
GST_RPAD_CHAINFUNC (pad) = chain;
GST_CAT_DEBUG (GST_CAT_PADS, "chainfunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (chain));
}
/**
* gst_pad_set_get_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real source #GstPad.
* @get: the #GstPadGetFunction to set.
*
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* Sets the given get function for the pad. The get function is called to
* produce a new #GstData to start the processing pipeline. Get functions cannot
* return %NULL.
*/
void
gst_pad_set_get_function (GstPad * pad, GstPadGetFunction get)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
g_return_if_fail (GST_RPAD_DIRECTION (pad) == GST_PAD_SRC);
GST_RPAD_GETFUNC (pad) = get;
GST_CAT_DEBUG (GST_CAT_PADS, "getfunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (get));
}
/**
* gst_pad_set_event_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real source #GstPad.
* @event: the #GstPadEventFunction to set.
*
* Sets the given event handler for the pad.
*/
void
gst_pad_set_event_function (GstPad * pad, GstPadEventFunction event)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
GST_RPAD_EVENTFUNC (pad) = event;
GST_CAT_DEBUG (GST_CAT_PADS, "eventfunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (event));
}
/**
* gst_pad_set_event_mask_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real #GstPad of either direction.
* @mask_func: the #GstPadEventMaskFunction to set.
*
* Sets the given event mask function for the pad.
*/
void
gst_pad_set_event_mask_function (GstPad * pad,
GstPadEventMaskFunction mask_func)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
GST_RPAD_EVENTMASKFUNC (pad) = mask_func;
GST_CAT_DEBUG (GST_CAT_PADS, "eventmaskfunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (mask_func));
}
/**
* gst_pad_get_event_masks:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a #GstPad.
*
* Gets the array of eventmasks from the given pad.
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Returns: a zero-terminated array of #GstEventMask, or NULL if the pad does
* not have an event mask function.
*/
const GstEventMask *
gst_pad_get_event_masks (GstPad * pad)
{
GstRealPad *rpad;
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
rpad = GST_PAD_REALIZE (pad);
g_return_val_if_fail (rpad, NULL);
if (GST_RPAD_EVENTMASKFUNC (rpad))
return GST_RPAD_EVENTMASKFUNC (rpad) (GST_PAD (pad));
return NULL;
}
static gboolean
gst_pad_get_event_masks_dispatcher (GstPad * pad, const GstEventMask ** data)
{
*data = gst_pad_get_event_masks (pad);
return TRUE;
}
/**
* gst_pad_get_event_masks_default:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a #GstPad.
*
* Invokes the default event masks dispatcher on the pad.
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Returns: a zero-terminated array of #GstEventMask, or NULL if none of the
* internally-linked pads have an event mask function.
*/
const GstEventMask *
gst_pad_get_event_masks_default (GstPad * pad)
{
GstEventMask *result = NULL;
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
gst_pad_dispatcher (pad, (GstPadDispatcherFunction)
gst_pad_get_event_masks_dispatcher, &result);
return result;
}
/**
* gst_pad_set_convert_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real #GstPad of either direction.
* @convert: the #GstPadConvertFunction to set.
*
* Sets the given convert function for the pad.
*/
void
gst_pad_set_convert_function (GstPad * pad, GstPadConvertFunction convert)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
GST_RPAD_CONVERTFUNC (pad) = convert;
GST_CAT_DEBUG (GST_CAT_PADS, "convertfunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (convert));
}
/**
* gst_pad_set_query_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real #GstPad of either direction.
* @query: the #GstPadQueryFunction to set.
*
* Set the given query function for the pad.
*/
void
gst_pad_set_query_function (GstPad * pad, GstPadQueryFunction query)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
WARNING: Don't grab this updated unless you're really, REALLY sure. Original commit message from CVS: WARNING: Don't grab this updated unless you're really, REALLY sure. WARNING: Wait for the next one. Whole lotta changes here, including a few random bits: examples/*/Makefile: updated to use `libtool gcc`, not just `gcc` gst/ gstbuffer.h: updated to new flag style gst.c, gstdebug.h: added new debugging for function ptrs gstpipeline.c: set type of parent_class to the class, not the object gstthread.c: ditto plugins/ cdparanoia/cdparanoia.c: added an argument type, updated some defaults cobin/spindentity.c: updated to new do/while loopfunction style mp3encode/lame/gstlame.c: argument types, whole lotta lame options tests/: various changes Now, for the big changes: Once again, the scheduling system has changed. And once again, it broke a whole bunch of things. The gist of the change is that there is now a function pointer for gst_pad_push and gst_pad_pull, instead of a hard-wired function. Well, currently they are functions, but that's for debugging purposes only, they just call the function pointer after spewing lots of DEBUG(). This changed the GstPad structure a bit, and the GstPad API as well. Where elements used to provide chain() and pull() functions, they provide chain() and get() functions. gst_pad_set_pull[region]_function has been changed to get_pad_set_get[region]_function. This means all the elements out there that used to have pull functions need to be updated. The calls to that function have been changed in the normal elements, but the names of the functions passed is still _pull[region](), which is an aesthetic issue more than anything. As for what doesn't work yet, just about anything dealing with Connections is hosed, meaning threaded stuff won't work. This will be fixed about 12 hours from now, after I've slept, etc. The simplefake.c test works in both cothreaded and chained cases, but not much else will work due to the Connection problem. Needless to say, don't grab this unless you *need* these features *now*, else wait to update this stuff until tomorrow. I'm going to sleep now.
2000-12-16 10:18:09 +00:00
GST_RPAD_QUERYFUNC (pad) = query;
GST_CAT_DEBUG (GST_CAT_PADS, "queryfunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (query));
}
/**
* gst_pad_set_query_type_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real #GstPad of either direction.
* @type_func: the #GstPadQueryTypeFunction to set.
*
* Set the given query type function for the pad.
*/
void
gst_pad_set_query_type_function (GstPad * pad,
GstPadQueryTypeFunction type_func)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
GST_RPAD_QUERYTYPEFUNC (pad) = type_func;
GST_CAT_DEBUG (GST_CAT_PADS, "querytypefunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (type_func));
}
/**
* gst_pad_get_query_types:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a #GstPad.
*
* Get an array of supported queries that can be performed
* on this pad.
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Returns: a zero-terminated array of #GstQueryType.
*/
const GstQueryType *
gst_pad_get_query_types (GstPad * pad)
{
GstRealPad *rpad;
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
rpad = GST_PAD_REALIZE (pad);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (rpad, NULL);
if (GST_RPAD_QUERYTYPEFUNC (rpad))
return GST_RPAD_QUERYTYPEFUNC (rpad) (GST_PAD (pad));
return NULL;
}
static gboolean
gst_pad_get_query_types_dispatcher (GstPad * pad, const GstQueryType ** data)
{
*data = gst_pad_get_query_types (pad);
return TRUE;
}
/**
* gst_pad_get_query_types_default:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a #GstPad.
*
* Invoke the default dispatcher for the query types on
* the pad.
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Returns: an zero-terminated array of #GstQueryType, or NULL if none of the
* internally-linked pads has a query types function.
*/
const GstQueryType *
gst_pad_get_query_types_default (GstPad * pad)
{
GstQueryType *result = NULL;
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
gst_pad_dispatcher (pad, (GstPadDispatcherFunction)
gst_pad_get_query_types_dispatcher, &result);
return result;
}
/**
* gst_pad_set_internal_link_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real #GstPad of either direction.
* @intlink: the #GstPadIntLinkFunction to set.
*
* Sets the given internal link function for the pad.
*/
void
gst_pad_set_internal_link_function (GstPad * pad, GstPadIntLinkFunction intlink)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
GST_RPAD_INTLINKFUNC (pad) = intlink;
GST_CAT_DEBUG (GST_CAT_PADS, "internal link for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (intlink));
}
/**
* gst_pad_set_formats_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real #GstPad of either direction.
* @formats: the #GstPadFormatsFunction to set.
*
* Sets the given formats function for the pad.
*/
void
gst_pad_set_formats_function (GstPad * pad, GstPadFormatsFunction formats)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
GST_RPAD_FORMATSFUNC (pad) = formats;
GST_CAT_DEBUG (GST_CAT_PADS, "formats function for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (formats));
}
/**
* gst_pad_set_link_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real #GstPad.
* @link: the #GstPadLinkFunction to set.
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
*
* Sets the given link function for the pad. It will be called when the pad is
* linked or relinked with caps. The caps passed to the link function is
* the filtered caps for the connnection. It can contain a non fixed caps.
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
*
* The return value GST_PAD_LINK_OK should be used when the connection can be
* made.
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
*
* The return value GST_PAD_LINK_REFUSED should be used when the connection
* cannot be made for some reason.
*/
void
gst_pad_set_link_function (GstPad * pad, GstPadLinkFunction link)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
GST_RPAD_LINKFUNC (pad) = link;
GST_CAT_DEBUG (GST_CAT_PADS, "linkfunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (link));
}
/**
* gst_pad_set_unlink_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real #GstPad.
* @unlink: the #GstPadUnlinkFunction to set.
*
* Sets the given unlink function for the pad. It will be called
* when the pad is unlinked.
*/
void
gst_pad_set_unlink_function (GstPad * pad, GstPadUnlinkFunction unlink)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
GST_RPAD_UNLINKFUNC (pad) = unlink;
GST_CAT_DEBUG (GST_CAT_PADS, "unlinkfunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (unlink));
}
/**
* gst_pad_set_getcaps_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real #GstPad.
* @getcaps: the #GstPadGetCapsFunction to set.
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
*
* Sets the given getcaps function for the pad. @getcaps should return the
* allowable caps for a pad in the context of the element's state, its link to
* other elements, and the devices or files it has opened. These caps must be a
* subset of the pad template caps. In the NULL state with no links, @getcaps
* should ideally return the same caps as the pad template. In rare
* circumstances, an object property can affect the caps returned by @getcaps,
* but this is discouraged.
*
* You do not need to call this function if @pad's allowed caps are always the
* same as the pad template caps.
*
* For most filters, the caps returned by @getcaps is directly affected by the
* allowed caps on other pads. For demuxers and decoders, the caps returned by
* the srcpad's getcaps function is directly related to the stream data. Again,
* @getcaps should return the most specific caps it reasonably can, since this
* helps with autoplugging.
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
*
* Note that the return value from @getcaps is owned by the caller.
*/
void
gst_pad_set_getcaps_function (GstPad * pad, GstPadGetCapsFunction getcaps)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
GST_RPAD_GETCAPSFUNC (pad) = getcaps;
GST_CAT_DEBUG (GST_CAT_PADS, "getcapsfunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (getcaps));
}
/**
* gst_pad_set_bufferalloc_function:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real sink #GstPad.
* @bufalloc: the #GstPadBufferAllocFunction to set.
*
* Sets the given bufferalloc function for the pad. Note that the
* bufferalloc function can only be set on sinkpads.
*/
void
gst_pad_set_bufferalloc_function (GstPad * pad,
GstPadBufferAllocFunction bufalloc)
{
g_return_if_fail (GST_IS_REAL_PAD (pad));
g_return_if_fail (GST_PAD_IS_SINK (pad));
GST_RPAD_BUFFERALLOCFUNC (pad) = bufalloc;
GST_CAT_DEBUG (GST_CAT_PADS, "bufferallocfunc for %s:%s set to %s",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_FUNCPTR_NAME (bufalloc));
}
/**
* gst_pad_unlink:
* @srcpad: the source #GstPad to unlink.
* @sinkpad: the sink #GstPad to unlink.
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Unlinks the source pad from the sink pad. Will emit the "unlinked" signal on
* both pads.
*
* Returns: TRUE if the pads were unlinked. This function returns FALSE if
* the pads were not linked together.
*
* MT safe.
*/
gboolean
gst_pad_unlink (GstPad * srcpad, GstPad * sinkpad)
{
GstRealPad *realsrc, *realsink;
g_return_val_if_fail (GST_IS_PAD (srcpad), FALSE);
g_return_val_if_fail (GST_IS_PAD (sinkpad), FALSE);
GST_CAT_INFO (GST_CAT_ELEMENT_PADS, "unlinking %s:%s(%p) and %s:%s(%p)",
GST_DEBUG_PAD_NAME (srcpad), srcpad,
GST_DEBUG_PAD_NAME (sinkpad), sinkpad);
GST_LOCK (srcpad);
realsrc = GST_PAD_REALIZE (srcpad);
if (srcpad != GST_PAD_CAST (realsrc)) {
GST_LOCK (realsrc);
/* we don't care if the ghostpad goes away now */
GST_UNLOCK (srcpad);
}
if (G_UNLIKELY (GST_RPAD_DIRECTION (realsrc) != GST_PAD_SRC))
goto not_srcpad;
if (G_UNLIKELY (GST_RPAD_PEER (realsrc) == NULL))
goto was_unlinked;
GST_LOCK (sinkpad);
realsink = GST_PAD_REALIZE (sinkpad);
if (sinkpad != GST_PAD_CAST (realsink)) {
GST_LOCK (realsink);
/* we don't care if the ghostpad goes away now */
GST_UNLOCK (sinkpad);
}
if (G_UNLIKELY (GST_RPAD_DIRECTION (realsink) != GST_PAD_SINK))
goto not_sinkpad;
if (G_UNLIKELY (GST_RPAD_PEER (realsrc) != realsink))
goto not_linked_together;
if (GST_RPAD_UNLINKFUNC (realsrc)) {
GST_RPAD_UNLINKFUNC (realsrc) (GST_PAD (realsrc));
}
if (GST_RPAD_UNLINKFUNC (realsink)) {
GST_RPAD_UNLINKFUNC (realsink) (GST_PAD (realsink));
}
/* first clear peers */
GST_RPAD_PEER (realsrc) = NULL;
GST_RPAD_PEER (realsink) = NULL;
/* clear filter, note that we leave the pad caps as they are */
gst_caps_replace (&GST_RPAD_APPFILTER (realsrc), NULL);
gst_caps_replace (&GST_RPAD_APPFILTER (realsink), NULL);
GST_UNLOCK (realsrc);
GST_UNLOCK (realsink);
/* fire off a signal to each of the pads telling them
* that they've been unlinked */
g_signal_emit (G_OBJECT (realsrc), gst_real_pad_signals[REAL_UNLINKED],
0, realsink);
g_signal_emit (G_OBJECT (realsink), gst_real_pad_signals[REAL_UNLINKED],
0, realsrc);
GST_CAT_INFO (GST_CAT_ELEMENT_PADS, "unlinked %s:%s and %s:%s",
GST_DEBUG_PAD_NAME (srcpad), GST_DEBUG_PAD_NAME (sinkpad));
return TRUE;
not_srcpad:
{
g_critical ("pad %s is not a source pad", GST_PAD_NAME (realsrc));
GST_UNLOCK (realsrc);
return FALSE;
}
not_sinkpad:
{
g_critical ("pad %s is not a sink pad", GST_PAD_NAME (realsink));
GST_UNLOCK (realsrc);
GST_UNLOCK (realsink);
return FALSE;
}
was_unlinked:
{
/* we do not emit a warning in this case because unlinking cannot
* be made MT safe.*/
GST_UNLOCK (realsrc);
return FALSE;
}
not_linked_together:
{
/* we do not emit a warning in this case because unlinking cannot
* be made MT safe.*/
GST_UNLOCK (realsrc);
GST_UNLOCK (realsink);
return FALSE;
}
}
/**
* gst_pad_is_linked:
* @pad: pad to check
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Checks if a @pad is linked to another pad or not.
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Returns: TRUE if the pad is linked, FALSE otherwise.
*
* MT safe.
*/
gboolean
gst_pad_is_linked (GstPad * pad)
{
gboolean result;
g_return_val_if_fail (GST_IS_PAD (pad), FALSE);
GST_LOCK (pad);
result = (GST_PAD_PEER (pad) != NULL);
GST_UNLOCK (pad);
return result;
}
/**
* gst_pad_link_filtered:
* @srcpad: the source #GstPad to link.
* @sinkpad: the sink #GstPad to link.
* @filtercaps: the filter #GstCaps.
*
* Links the source pad and the sink pad, constrained
* by the given filter caps.
*
* The filtercaps will be copied and refcounted, so you should unref
* it yourself after using this function.
*
* Returns: A result code indicating if the connection worked or
* what went wrong.
*/
GstPadLinkReturn
gst_pad_link_filtered (GstPad * srcpad, GstPad * sinkpad,
const GstCaps * filtercaps)
{
GstRealPad *realsrc, *realsink;
GstPadLinkReturn result;
/* generic checks */
g_return_val_if_fail (GST_IS_PAD (srcpad), GST_PAD_LINK_REFUSED);
g_return_val_if_fail (GST_IS_PAD (sinkpad), GST_PAD_LINK_REFUSED);
GST_CAT_INFO (GST_CAT_PADS, "trying to link %s:%s and %s:%s",
GST_DEBUG_PAD_NAME (srcpad), GST_DEBUG_PAD_NAME (sinkpad));
/* now we need to deal with the real/ghost stuff */
realsrc = GST_PAD_REALIZE (srcpad);
realsink = GST_PAD_REALIZE (sinkpad);
if ((GST_PAD (realsrc) != srcpad) || (GST_PAD (realsink) != sinkpad)) {
GST_CAT_INFO (GST_CAT_PADS, "*actually* linking %s:%s and %s:%s",
GST_DEBUG_PAD_NAME (realsrc), GST_DEBUG_PAD_NAME (realsink));
}
g_return_val_if_fail (realsrc != NULL, GST_PAD_LINK_REFUSED);
g_return_val_if_fail (realsink != NULL, GST_PAD_LINK_REFUSED);
GST_LOCK (realsrc);
GST_LOCK (realsink);
/* FIXME: shouldn't we convert this to g_return_val_if_fail? */
if (GST_RPAD_PEER (realsrc) != NULL) {
GST_CAT_INFO (GST_CAT_PADS, "Real source pad %s:%s has a peer, failed",
GST_DEBUG_PAD_NAME (realsrc));
result = GST_PAD_LINK_REFUSED;
goto error;
}
if (GST_RPAD_PEER (realsink) != NULL) {
GST_CAT_INFO (GST_CAT_PADS, "Real sink pad %s:%s has a peer, failed",
GST_DEBUG_PAD_NAME (realsink));
result = GST_PAD_LINK_REFUSED;
goto error;
}
if (GST_RPAD_DIRECTION (realsrc) != GST_PAD_SRC) {
GST_CAT_INFO (GST_CAT_PADS,
"Real src pad %s:%s is not a source pad, failed",
GST_DEBUG_PAD_NAME (realsrc));
result = GST_PAD_LINK_REFUSED;
goto error;
}
if (GST_RPAD_DIRECTION (realsink) != GST_PAD_SINK) {
GST_CAT_INFO (GST_CAT_PADS, "Real sink pad %s:%s is not a sink pad, failed",
GST_DEBUG_PAD_NAME (realsink));
result = GST_PAD_LINK_REFUSED;
goto error;
}
/* FIXME check pad caps for non-empty intersection */
/* update filter before calling the link functions */
if (filtercaps) {
GstCaps *filtercopy = gst_caps_copy (filtercaps);
filtercopy = gst_caps_ref (filtercopy);
GST_RPAD_APPFILTER (realsrc) = filtercopy;
GST_RPAD_APPFILTER (realsink) = filtercopy;
} else {
GST_RPAD_APPFILTER (realsrc) = NULL;
GST_RPAD_APPFILTER (realsink) = NULL;
}
GST_UNLOCK (realsrc);
GST_UNLOCK (realsink);
if (GST_RPAD_LINKFUNC (realsrc)) {
/* this one will call the peer link function */
result =
GST_RPAD_LINKFUNC (realsrc) (GST_PAD (realsrc), GST_PAD (realsink));
} else if (GST_RPAD_LINKFUNC (realsink)) {
/* if no source link function, we need to call the sink link
* function ourselves. */
result =
GST_RPAD_LINKFUNC (realsink) (GST_PAD (realsink), GST_PAD (realsrc));
} else {
result = GST_PAD_LINK_OK;
}
GST_LOCK (realsrc);
GST_LOCK (realsink);
if (result == GST_PAD_LINK_OK) {
GST_RPAD_PEER (realsrc) = GST_REAL_PAD (realsink);
GST_RPAD_PEER (realsink) = GST_REAL_PAD (realsrc);
GST_UNLOCK (realsrc);
GST_UNLOCK (realsink);
/* fire off a signal to each of the pads telling them
* that they've been linked */
g_signal_emit (G_OBJECT (realsrc), gst_real_pad_signals[REAL_LINKED],
0, realsink);
g_signal_emit (G_OBJECT (realsink), gst_real_pad_signals[REAL_LINKED],
0, realsrc);
GST_CAT_INFO (GST_CAT_PADS, "linked %s:%s and %s:%s, successful",
GST_DEBUG_PAD_NAME (realsrc), GST_DEBUG_PAD_NAME (realsink));
} else {
GST_CAT_INFO (GST_CAT_PADS, "link between %s:%s and %s:%s failed",
GST_DEBUG_PAD_NAME (realsrc), GST_DEBUG_PAD_NAME (realsink));
/* remove the filter again */
if (filtercaps) {
gst_caps_replace (&GST_RPAD_APPFILTER (realsrc), NULL);
gst_caps_replace (&GST_RPAD_APPFILTER (realsink), NULL);
}
GST_UNLOCK (realsrc);
GST_UNLOCK (realsink);
}
return result;
error:
GST_UNLOCK (realsrc);
GST_UNLOCK (realsink);
return result;
}
/**
* gst_pad_link:
* @srcpad: the source #GstPad to link.
* @sinkpad: the sink #GstPad to link.
*
* Links the source pad to the sink pad.
*
* Returns: A result code indicating if the connection worked or
* what went wrong.
*/
GstPadLinkReturn
gst_pad_link (GstPad * srcpad, GstPad * sinkpad)
{
return gst_pad_link_filtered (srcpad, sinkpad, NULL);
}
static void
gst_pad_set_pad_template (GstPad * pad, GstPadTemplate * templ)
{
/* this function would need checks if it weren't static */
gst_object_replace ((GstObject **) & pad->padtemplate, (GstObject *) templ);
if (templ) {
gst_object_sink (GST_OBJECT (templ));
g_signal_emit (G_OBJECT (templ),
gst_pad_template_signals[TEMPL_PAD_CREATED], 0, pad);
}
}
/**
* gst_pad_get_pad_template:
* @pad: a #GstPad.
*
* Gets the template for @pad.
*
* Returns: the #GstPadTemplate from which this pad was instantiated, or %NULL
* if this pad has no template.
*/
GstPadTemplate *
gst_pad_get_pad_template (GstPad * pad)
{
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
return GST_PAD_PAD_TEMPLATE (pad);
}
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
/**
* gst_pad_get_real_parent:
* @pad: a #GstPad to get the real parent of.
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
*
* Gets the real parent object of this pad. If the pad
* is a ghost pad, the actual owner of the real pad is
* returned, as opposed to #gst_pad_get_parent().
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
*
* Returns: the parent #GstElement.
*/
GstElement *
gst_pad_get_real_parent (GstPad * pad)
{
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
return GST_PAD_PARENT (GST_PAD (GST_PAD_REALIZE (pad)));
}
/* FIXME not MT safe */
static void
gst_pad_add_ghost_pad (GstPad * pad, GstPad * ghostpad)
{
GstRealPad *realpad;
/* if we're ghosting a ghost pad, drill down to find the real pad */
realpad = (GstRealPad *) pad;
while (GST_IS_GHOST_PAD (realpad))
realpad = GST_GPAD_REALPAD (realpad);
g_return_if_fail (GST_IS_REAL_PAD (realpad));
/* will ref the pad template */
GST_GPAD_REALPAD (ghostpad) = realpad;
realpad->ghostpads = g_list_prepend (realpad->ghostpads, ghostpad);
gst_pad_set_pad_template (GST_PAD (ghostpad), GST_PAD_PAD_TEMPLATE (pad));
}
static void
gst_pad_remove_ghost_pad (GstPad * pad, GstPad * ghostpad)
{
GstRealPad *realpad;
realpad = GST_PAD_REALIZE (pad);
g_return_if_fail (GST_GPAD_REALPAD (ghostpad) == realpad);
gst_pad_set_pad_template (GST_PAD (ghostpad), NULL);
realpad->ghostpads = g_list_remove (realpad->ghostpads, ghostpad);
GST_GPAD_REALPAD (ghostpad) = NULL;
}
/**
* gst_pad_relink_filtered:
* @srcpad: the source #GstPad to relink.
* @sinkpad: the sink #GstPad to relink.
* @filtercaps: the #GstPad to use as a filter in the relink.
*
* Relinks the given source and sink pad, constrained by the given
* capabilities. If the relink fails, the pads are unlinked
* and an error code is returned.
*
* Returns: The result code of the operation.
*
* MT safe
*/
GstPadLinkReturn
gst_pad_relink_filtered (GstPad * srcpad, GstPad * sinkpad,
const GstCaps * filtercaps)
{
GstCaps *filtercopy;
/* create copy with two refs */
filtercopy = gst_caps_copy (filtercaps);
filtercopy = gst_caps_ref (filtercopy);
/* probably check if intersection with new filter is ok */
GST_LOCK (srcpad);
GST_LOCK (sinkpad);
gst_caps_replace (&GST_PAD_CAPS (srcpad), NULL);
gst_caps_replace (&GST_PAD_APPFILTER (srcpad), filtercopy);
gst_caps_replace (&GST_PAD_CAPS (sinkpad), NULL);
gst_caps_replace (&GST_PAD_APPFILTER (sinkpad), filtercopy);
GST_UNLOCK (sinkpad);
GST_UNLOCK (srcpad);
return GST_PAD_LINK_OK;
}
/**
* gst_pad_get_caps:
* @pad: a #GstPad to get the capabilities of.
*
* Gets the capabilities of this pad.
*
* Returns: the #GstCaps of this pad. This function returns a new caps, so use
* gst_caps_unref to get rid of it.
*/
GstCaps *
gst_pad_get_caps (GstPad * pad)
{
GstRealPad *realpad;
GstCaps *result = NULL, *filter;
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
realpad = GST_PAD_REALIZE (pad);
GST_CAT_DEBUG (GST_CAT_CAPS, "get pad caps of %s:%s (%p)",
GST_DEBUG_PAD_NAME (realpad), realpad);
if (GST_PAD_IS_DISPATCHING (realpad)) {
GST_CAT_DEBUG (GST_CAT_CAPS,
"pad %s:%s is already dispatching!", GST_DEBUG_PAD_NAME (realpad));
g_warning ("pad %s:%s recursively called getcaps!",
GST_DEBUG_PAD_NAME (realpad));
return NULL;
}
if (GST_RPAD_GETCAPSFUNC (realpad)) {
GST_CAT_DEBUG (GST_CAT_CAPS, "dispatching to pad getcaps function");
GST_FLAG_SET (realpad, GST_PAD_DISPATCHING);
result = GST_RPAD_GETCAPSFUNC (realpad) (GST_PAD (realpad));
GST_FLAG_UNSET (realpad, GST_PAD_DISPATCHING);
if (result == NULL) {
g_critical ("pad %s:%s returned NULL caps from getcaps function\n",
GST_DEBUG_PAD_NAME (realpad));
} else {
#ifndef G_DISABLE_ASSERT
/* check that the returned caps are a real subset of the template caps */
if (GST_PAD_PAD_TEMPLATE (realpad)) {
const GstCaps *templ_caps =
GST_PAD_TEMPLATE_CAPS (GST_PAD_PAD_TEMPLATE (realpad));
if (!gst_caps_is_subset (result, templ_caps)) {
GstCaps *temp;
GST_CAT_ERROR_OBJECT (GST_CAT_CAPS, pad,
"pad returned caps %" GST_PTR_FORMAT
" which are not a real subset of its template caps %"
GST_PTR_FORMAT, result, templ_caps);
g_warning
("pad %s:%s returned caps that are not a real subset of its template caps",
GST_DEBUG_PAD_NAME (realpad));
temp = gst_caps_intersect (templ_caps, result);
gst_caps_unref (result);
result = temp;
}
}
#endif
goto done;
}
}
if (GST_PAD_PAD_TEMPLATE (realpad)) {
GstPadTemplate *templ = GST_PAD_PAD_TEMPLATE (realpad);
result = GST_PAD_TEMPLATE_CAPS (templ);
GST_CAT_DEBUG (GST_CAT_CAPS,
"using pad template %p with caps %p %" GST_PTR_FORMAT, templ, result,
result);
result = gst_caps_copy (result);
goto done;
}
GST_CAT_DEBUG (GST_CAT_CAPS, "pad has no caps");
result = gst_caps_new_any ();
done:
filter = GST_RPAD_APPFILTER (realpad);
if (filter) {
GST_CAT_DEBUG (GST_CAT_CAPS,
"app filter %p %" GST_PTR_FORMAT, filter, filter);
result = gst_caps_intersect (result, filter);
GST_CAT_DEBUG (GST_CAT_CAPS,
"caps after intersection with app filter %p %" GST_PTR_FORMAT, result,
result);
}
return result;
}
/**
* gst_pad_set_caps:
* @pad: a #GstPad to set the capabilities of.
* @caps: a #GstCaps to set.
*
* Sets the capabilities of this pad. The caps must be fixed. Any previous
* caps on the pad will be unreffed. This function refs the caps so you should
* unref if as soon as you don't need it anymore.
* It is possible to set NULL caps, which will make the pad unnegotiated
* again.
*
* Returns: TRUE if the caps could be set. FALSE if the caps were not fixed
* or bad parameters were provided to this function.
*
* MT safe.
*/
gboolean
gst_pad_set_caps (GstPad * pad, GstCaps * caps)
{
g_return_val_if_fail (GST_IS_REAL_PAD (pad), FALSE);
GST_LOCK (pad);
if (GST_PAD_CAPS (pad))
gst_caps_unref (GST_PAD_CAPS (pad));
if (caps)
caps = gst_caps_ref (caps);
GST_PAD_CAPS (pad) = caps;
GST_UNLOCK (pad);
g_object_notify (G_OBJECT (pad), "caps");
return TRUE;
}
/**
* gst_pad_get_pad_template_caps:
* @pad: a #GstPad to get the template capabilities from.
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Gets the capabilities for @pad's template.
*
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
* Returns: the #GstCaps of this pad template. If you intend to keep a reference
* on the caps, make a copy (see gst_caps_copy ()).
*/
const GstCaps *
gst_pad_get_pad_template_caps (GstPad * pad)
{
static GstStaticCaps anycaps = GST_STATIC_CAPS ("ANY");
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
if (GST_PAD_PAD_TEMPLATE (pad))
return GST_PAD_TEMPLATE_CAPS (GST_PAD_PAD_TEMPLATE (pad));
return gst_static_caps_get (&anycaps);
}
/**
* gst_pad_get_peer:
* @pad: a #GstPad to get the peer of.
*
* Gets the peer of @pad. This function refs the peer pad so
* you need to unref it after use.
*
* Returns: the peer #GstPad. Unref after usage.
*
* MT safe.
*/
GstPad *
gst_pad_get_peer (GstPad * pad)
{
GstRealPad *realpad;
GstRealPad *result;
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
GST_LOCK (pad);
realpad = GST_PAD_REALIZE (pad);
if (pad != GST_PAD_CAST (realpad)) {
GST_LOCK (realpad);
GST_UNLOCK (pad);
}
result = GST_RPAD_PEER (realpad);
gst_object_ref (GST_OBJECT (result));
GST_UNLOCK (realpad);
return GST_PAD_CAST (result);
}
/**
* gst_pad_get_allowed_caps:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a real #GstPad.
*
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* Gets the capabilities of the allowed media types that can flow through @pad.
* The caller must free the resulting caps.
*
* Returns: the allowed #GstCaps of the pad link. Free the caps when
* you no longer need it.
*/
GstCaps *
gst_pad_get_allowed_caps (GstPad * pad)
{
const GstCaps *mycaps;
GstCaps *caps;
GstCaps *peercaps;
g_return_val_if_fail (GST_IS_REAL_PAD (pad), NULL);
GST_CAT_DEBUG (GST_CAT_PROPERTIES, "%s:%s: getting allowed caps",
GST_DEBUG_PAD_NAME (pad));
mycaps = gst_pad_get_caps (pad);
if (GST_RPAD_PEER (pad) == NULL) {
GST_CAT_DEBUG (GST_CAT_PROPERTIES, "%s:%s: no peer, returning own caps",
GST_DEBUG_PAD_NAME (pad));
return gst_caps_copy (mycaps);
}
peercaps = gst_pad_get_caps (GST_PAD_PEER (pad));
caps = gst_caps_intersect (mycaps, peercaps);
gst_caps_unref (peercaps);
GST_CAT_DEBUG (GST_CAT_CAPS, "allowed caps %" GST_PTR_FORMAT, caps);
return caps;
}
/**
* gst_pad_alloc_buffer:
* @pad: a source #GstPad
* @offset: the offset of the new buffer in the stream
* @size: the size of the new buffer
* @caps: the caps of the new buffer
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Allocates a new, empty buffer optimized to push to pad @pad. This
* function only works if @pad is a source pad. You need to check the
* caps of the buffer after performing this function and renegotiate
* to the format if needed.
*
* Returns: a new, empty #GstBuffer, or NULL if there is an error
*/
GstBuffer *
gst_pad_alloc_buffer (GstPad * pad, guint64 offset, gint size)
{
GstRealPad *peer;
GstBuffer *result = NULL;
GstPadBufferAllocFunction bufferallocfunc;
g_return_val_if_fail (GST_IS_REAL_PAD (pad), NULL);
g_return_val_if_fail (GST_PAD_IS_SRC (pad), NULL);
if ((peer = GST_RPAD_PEER (pad)) == NULL)
goto fallback;
if ((bufferallocfunc = peer->bufferallocfunc) == NULL)
goto fallback;
GST_CAT_DEBUG (GST_CAT_PADS,
"calling bufferallocfunc &%s (@%p) of peer pad %s:%s",
GST_DEBUG_FUNCPTR_NAME (bufferallocfunc),
&bufferallocfunc, GST_DEBUG_PAD_NAME (((GstPad *) peer)));
result = bufferallocfunc (GST_PAD (peer), offset, size, GST_PAD_CAPS (pad));
if (result == NULL)
goto fallback;
return result;
/* fallback case */
fallback:
result = gst_buffer_new_and_alloc (size);
gst_buffer_set_caps (result, GST_PAD_CAPS (pad));
return result;
}
static void
gst_real_pad_dispose (GObject * object)
{
GstPad *pad = GST_PAD (object);
/* No linked pad can ever be disposed.
* It has to have a parent to be linked
* and a parent would hold a reference */
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
/* FIXME: what about if g_object_dispose is explicitly called on the pad? Is
that legal? otherwise we could assert GST_OBJECT_PARENT (pad) == NULL as
well... */
g_assert (GST_PAD_PEER (pad) == NULL);
GST_CAT_DEBUG (GST_CAT_REFCOUNTING, "dispose %s:%s",
GST_DEBUG_PAD_NAME (pad));
/* we destroy the ghostpads, because they are nothing without the real pad */
if (GST_REAL_PAD (pad)->ghostpads) {
GList *orig, *ghostpads;
orig = ghostpads = g_list_copy (GST_REAL_PAD (pad)->ghostpads);
while (ghostpads) {
GstPad *ghostpad = GST_PAD (ghostpads->data);
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
if (GST_IS_ELEMENT (GST_OBJECT_PARENT (ghostpad))) {
GstElement *parent = GST_ELEMENT (GST_OBJECT_PARENT (ghostpad));
GST_CAT_DEBUG (GST_CAT_REFCOUNTING,
"removing ghost pad from element '%s'", GST_OBJECT_NAME (parent));
gst_element_remove_pad (parent, ghostpad);
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
} else {
/* handle the case where we have some floating ghost pad that was never
added to an element */
g_object_set (ghostpad, "real-pad", NULL, NULL);
}
ghostpads = g_list_next (ghostpads);
}
g_list_free (orig);
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
/* as the ghost pads are removed, they remove themselves from ->ghostpads.
So it should be empty now. Let's assert that. */
g_assert (GST_REAL_PAD (pad)->ghostpads == NULL);
}
/* clear the caps */
gst_caps_replace (&GST_RPAD_CAPS (pad), NULL);
gst_caps_replace (&GST_RPAD_APPFILTER (pad), NULL);
if (GST_IS_ELEMENT (GST_OBJECT_PARENT (pad))) {
GST_CAT_DEBUG (GST_CAT_REFCOUNTING, "removing pad from element '%s'",
GST_OBJECT_NAME (GST_OBJECT (GST_ELEMENT (GST_OBJECT_PARENT (pad)))));
gst_element_remove_pad (GST_ELEMENT (GST_OBJECT_PARENT (pad)), pad);
}
G_OBJECT_CLASS (real_pad_parent_class)->dispose (object);
}
#ifndef GST_DISABLE_LOADSAVE
/* FIXME: why isn't this on a GstElement ? */
/**
* gst_pad_load_and_link:
* @self: an #xmlNodePtr to read the description from.
* @parent: the #GstObject element that owns the pad.
*
* Reads the pad definition from the XML node and links the given pad
* in the element to a pad of an element up in the hierarchy.
*/
void
gst_pad_load_and_link (xmlNodePtr self, GstObject * parent)
{
xmlNodePtr field = self->xmlChildrenNode;
GstPad *pad = NULL, *targetpad;
gchar *peer = NULL;
gchar **split;
GstElement *target;
GstObject *grandparent;
gchar *name = NULL;
while (field) {
if (!strcmp (field->name, "name")) {
name = xmlNodeGetContent (field);
pad = gst_element_get_pad (GST_ELEMENT (parent), name);
g_free (name);
} else if (!strcmp (field->name, "peer")) {
peer = xmlNodeGetContent (field);
}
field = field->next;
}
g_return_if_fail (pad != NULL);
if (peer == NULL)
return;
split = g_strsplit (peer, ".", 2);
if (split[0] == NULL || split[1] == NULL) {
GST_CAT_DEBUG (GST_CAT_XML,
"Could not parse peer '%s' for pad %s:%s, leaving unlinked",
peer, GST_DEBUG_PAD_NAME (pad));
g_free (peer);
return;
}
g_free (peer);
g_return_if_fail (split[0] != NULL);
g_return_if_fail (split[1] != NULL);
grandparent = gst_object_get_parent (parent);
if (grandparent && GST_IS_BIN (grandparent)) {
target = gst_bin_get_by_name_recurse_up (GST_BIN (grandparent), split[0]);
} else
goto cleanup;
if (target == NULL)
goto cleanup;
targetpad = gst_element_get_pad (target, split[1]);
if (targetpad == NULL)
goto cleanup;
gst_pad_link (pad, targetpad);
cleanup:
g_strfreev (split);
}
/**
* gst_pad_save_thyself:
* @pad: a #GstPad to save.
* @parent: the parent #xmlNodePtr to save the description in.
*
* Saves the pad into an xml representation.
*
* Returns: the #xmlNodePtr representation of the pad.
*/
static xmlNodePtr
gst_pad_save_thyself (GstObject * object, xmlNodePtr parent)
{
GstRealPad *realpad;
GstPad *peer;
g_return_val_if_fail (GST_IS_REAL_PAD (object), NULL);
realpad = GST_REAL_PAD (object);
xmlNewChild (parent, NULL, "name", GST_PAD_NAME (realpad));
if (GST_RPAD_PEER (realpad) != NULL) {
gchar *content;
peer = GST_PAD (GST_RPAD_PEER (realpad));
/* first check to see if the peer's parent's parent is the same */
/* we just save it off */
content = g_strdup_printf ("%s.%s",
GST_OBJECT_NAME (GST_PAD_PARENT (peer)), GST_PAD_NAME (peer));
xmlNewChild (parent, NULL, "peer", content);
g_free (content);
} else
xmlNewChild (parent, NULL, "peer", "");
return parent;
}
/**
* gst_ghost_pad_save_thyself:
* @pad: a ghost #GstPad to save.
* @parent: the parent #xmlNodePtr to save the description in.
*
* Saves the ghost pad into an xml representation.
*
* Returns: the #xmlNodePtr representation of the pad.
*/
xmlNodePtr
gst_ghost_pad_save_thyself (GstPad * pad, xmlNodePtr parent)
{
xmlNodePtr self;
g_return_val_if_fail (GST_IS_GHOST_PAD (pad), NULL);
self = xmlNewChild (parent, NULL, "ghostpad", NULL);
xmlNewChild (self, NULL, "name", GST_PAD_NAME (pad));
xmlNewChild (self, NULL, "parent", GST_OBJECT_NAME (GST_PAD_PARENT (pad)));
/* FIXME FIXME FIXME! */
return self;
}
#endif /* GST_DISABLE_LOADSAVE */
/* should be called with pad lock held */
static void
handle_pad_block (GstPad * pad)
{
GstPadBlockCallback callback;
gpointer user_data;
GstRealPad *realpad;
realpad = GST_REAL_PAD (pad);
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad,
"signal block taken on pad %s:%s", GST_DEBUG_PAD_NAME (pad));
callback = realpad->block_callback;
if (callback) {
user_data = realpad->block_data;
GST_UNLOCK (pad);
callback (pad, TRUE, user_data);
GST_LOCK (pad);
} else {
GST_PAD_BLOCK_SIGNAL (pad);
}
while (GST_RPAD_IS_BLOCKED (pad))
GST_PAD_BLOCK_WAIT (pad);
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad, "got unblocked");
callback = realpad->block_callback;
if (callback) {
user_data = realpad->block_data;
GST_UNLOCK (pad);
callback (pad, FALSE, user_data);
GST_LOCK (pad);
} else {
GST_PAD_BLOCK_SIGNAL (pad);
}
}
/**
* gst_pad_push:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a source #GstPad.
* @buffer: the #GstBuffer to push.
*
* Pushes a buffer to the peer of @pad. @pad must be linked. May
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* only be called by @pad's parent.
*
* Returns: a #GstFlowReturn from the peer pad.
*/
GstFlowReturn
gst_pad_push (GstPad * pad, GstBuffer * buffer)
{
GstRealPad *peer;
GstFlowReturn ret;
GstPadChainFunction chainfunc;
g_return_val_if_fail (GST_IS_REAL_PAD (pad), GST_FLOW_ERROR);
g_return_val_if_fail (GST_PAD_DIRECTION (pad) == GST_PAD_SRC, GST_FLOW_ERROR);
g_return_val_if_fail (buffer != NULL, GST_FLOW_ERROR);
g_return_val_if_fail (GST_IS_BUFFER (buffer), GST_FLOW_ERROR);
This is an attempt at not segfaulting on errors but reporting some usefull info instead. Original commit message from CVS: This is an attempt at not segfaulting on errors but reporting some usefull info instead. - bin changes so errors can propagate. - changed the _FAST macros to _CAST because that is what they do. - removed all references to cothreads out of the core, they are really a scheduler issue, handler with a sched_private gpointer. - added a live buffer count, for debugging buffer leaks. - added error checking in gst_scheduler_state_transition this solves the "out of cothreads" problem. - GST_ELEMENT_NO_ENTRY == GST_ELEMENT_INFINITE_LOOP - added 2 private element flasg for use by the scheduler (_COTHREAD_STOPPING) is now - added scheduler entry points: - _yield : to create possible scheduling points. - _interrupt: to stop execution of an element. - _error: to signal en error condition to the scheduler. - improved error messages for pads. - signal gst_element_error where appropriate. - added the a new bin to the parent before entering it so one can reference its children. - queue memleak fixes on dispose. - added possible deadlock detection in queue (turned off be default) - GstBasicScheduler is a real class of its own now, hiding its internal variables. - GST_ELEMENT_IS_COTHREAD_STOPPING is gone. either call explicit _yield operations, or make a sane loop. - Better state change handling in filesrc. Better error reporting/recovery too. - updated core plugins. - detect non decoupled elements on scheduler boundries and error.
2001-12-22 21:18:17 +00:00
GST_LOCK (pad);
while (G_UNLIKELY (GST_RPAD_IS_BLOCKED (pad)))
handle_pad_block (pad);
if (G_UNLIKELY ((peer = GST_RPAD_PEER (pad)) == NULL))
goto not_linked;
if (G_UNLIKELY (!GST_RPAD_IS_ACTIVE (peer)))
goto not_active;
if (G_UNLIKELY ((chainfunc = peer->chainfunc) == NULL))
goto no_function;
GST_UNLOCK (pad);
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad,
"calling chainfunction &%s of peer pad %s:%s",
GST_DEBUG_FUNCPTR_NAME (chainfunc), GST_DEBUG_PAD_NAME (peer));
ret = chainfunc (GST_PAD_CAST (peer), buffer);
return ret;
/* ERROR recovery here */
not_linked:
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad,
"pushing, but it was not linked");
GST_UNLOCK_RETURN (pad, GST_FLOW_NOT_CONNECTED);
not_active:
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad, "pushing, but it was inactive");
GST_UNLOCK_RETURN (pad, GST_FLOW_WRONG_STATE);
no_function:
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad, "pushing, but not chainhandler");
g_warning ("push on pad %s:%s but it has no chainhandler, file a bug.",
GST_DEBUG_PAD_NAME (peer));
GST_UNLOCK_RETURN (pad, GST_FLOW_ERROR);
}
/**
* gst_pad_pull:
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* @pad: a sink #GstPad.
* @buffer: a pointer to hold the #GstBuffer.
*
* Pulls a buffer from the peer pad. May only be called by @pad's
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* parent.
*
* Returns: a #GstFlowReturn from the peer pad.
*/
GstFlowReturn
gst_pad_pull (GstPad * pad, GstBuffer ** buffer)
{
GstRealPad *peer;
GstFlowReturn ret;
GstPadGetFunction getfunc;
g_return_val_if_fail (GST_IS_REAL_PAD (pad), GST_FLOW_ERROR);
g_return_val_if_fail (GST_PAD_DIRECTION (pad) == GST_PAD_SINK,
GST_FLOW_ERROR);
g_return_val_if_fail (buffer != NULL, GST_FLOW_ERROR);
GST_LOCK (pad);
while (G_UNLIKELY (GST_RPAD_IS_BLOCKED (pad)))
handle_pad_block (pad);
if (G_UNLIKELY ((peer = GST_RPAD_PEER (pad)) == NULL))
goto not_connected;
2004-04-04 02:36:12 +00:00
if (G_UNLIKELY ((getfunc = peer->getfunc) == NULL))
goto no_function;
2004-04-04 02:36:12 +00:00
GST_UNLOCK (pad);
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad,
"calling getfunc %s of peer pad %s:%s",
GST_DEBUG_FUNCPTR_NAME (getfunc), GST_DEBUG_PAD_NAME (peer));
ret = getfunc (GST_PAD_CAST (peer), buffer);
return ret;
/* ERROR recovery here */
not_connected:
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad,
"pulling, but it was not linked");
GST_UNLOCK_RETURN (pad, GST_FLOW_NOT_CONNECTED);
no_function:
GST_ELEMENT_ERROR (GST_PAD_PARENT (pad), CORE, PAD, (NULL),
("pull on pad %s:%s but the peer pad %s:%s has no getfunc",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_PAD_NAME (peer)));
GST_UNLOCK_RETURN (pad, GST_FLOW_ERROR);
}
2004-04-04 02:36:12 +00:00
/**
* gst_pad_pull_range:
* @pad: a sink #GstPad.
* @buffer: a pointer to hold the #GstBuffer.
* @offset: The start offset of the buffer
* @length: The length of the buffer
2004-04-04 02:36:12 +00:00
*
* Pulls a buffer from the peer pad. May only be called by @pad's
* parent.
2004-04-04 02:36:12 +00:00
*
* Returns: a #GstFlowReturn from the peer pad.
2004-04-04 02:36:12 +00:00
*/
GstFlowReturn
gst_pad_pull_range (GstPad * pad, guint64 offset, guint size,
GstBuffer ** buffer)
2004-04-04 02:36:12 +00:00
{
GstRealPad *peer;
GstFlowReturn ret;
GstPadGetRangeFunction getrangefunc;
2004-04-04 02:36:12 +00:00
g_return_val_if_fail (GST_IS_REAL_PAD (pad), GST_FLOW_ERROR);
g_return_val_if_fail (GST_PAD_DIRECTION (pad) == GST_PAD_SINK,
GST_FLOW_ERROR);
g_return_val_if_fail (buffer != NULL, GST_FLOW_ERROR);
2004-04-04 02:36:12 +00:00
GST_LOCK (pad);
2004-04-04 02:36:12 +00:00
while (G_UNLIKELY (GST_RPAD_IS_BLOCKED (pad)))
handle_pad_block (pad);
2004-04-04 02:36:12 +00:00
if (G_UNLIKELY ((peer = GST_RPAD_PEER (pad)) == NULL))
goto not_connected;
if (G_UNLIKELY ((getrangefunc = peer->getrangefunc) == NULL))
goto no_function;
GST_UNLOCK (pad);
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad,
"calling getrangefunc %s of peer pad %s:%s",
GST_DEBUG_FUNCPTR_NAME (getrangefunc), GST_DEBUG_PAD_NAME (peer));
ret = getrangefunc (GST_PAD_CAST (peer), offset, size, buffer);
2004-04-04 02:36:12 +00:00
return ret;
2004-04-04 02:36:12 +00:00
/* ERROR recovery here */
not_connected:
GST_CAT_LOG_OBJECT (GST_CAT_SCHEDULING, pad,
"pulling range, but it was not linked");
GST_UNLOCK_RETURN (pad, GST_FLOW_NOT_CONNECTED);
2004-04-04 02:36:12 +00:00
no_function:
GST_ELEMENT_ERROR (GST_PAD_PARENT (pad), CORE, PAD, (NULL),
("pullrange on pad %s:%s but the peer pad %s:%s has no getrangefunction",
GST_DEBUG_PAD_NAME (pad), GST_DEBUG_PAD_NAME (peer)));
GST_UNLOCK_RETURN (pad, GST_FLOW_ERROR);
2004-04-04 02:36:12 +00:00
}
/************************************************************************
*
* templates
*
*/
static void gst_pad_template_class_init (GstPadTemplateClass * klass);
static void gst_pad_template_init (GstPadTemplate * templ);
static void gst_pad_template_dispose (GObject * object);
GType
gst_pad_template_get_type (void)
{
static GType padtemplate_type = 0;
if (!padtemplate_type) {
static const GTypeInfo padtemplate_info = {
sizeof (GstPadTemplateClass), NULL, NULL,
(GClassInitFunc) gst_pad_template_class_init, NULL, NULL,
sizeof (GstPadTemplate),
0,
(GInstanceInitFunc) gst_pad_template_init, NULL
};
padtemplate_type =
g_type_register_static (GST_TYPE_OBJECT, "GstPadTemplate",
&padtemplate_info, 0);
}
return padtemplate_type;
}
static void
gst_pad_template_class_init (GstPadTemplateClass * klass)
{
GObjectClass *gobject_class;
GstObjectClass *gstobject_class;
gobject_class = (GObjectClass *) klass;
gstobject_class = (GstObjectClass *) klass;
padtemplate_parent_class = g_type_class_ref (GST_TYPE_OBJECT);
gst_pad_template_signals[TEMPL_PAD_CREATED] =
g_signal_new ("pad-created", G_TYPE_FROM_CLASS (klass), G_SIGNAL_RUN_LAST,
G_STRUCT_OFFSET (GstPadTemplateClass, pad_created),
NULL, NULL, gst_marshal_VOID__OBJECT, G_TYPE_NONE, 1, GST_TYPE_PAD);
gobject_class->dispose = gst_pad_template_dispose;
gstobject_class->path_string_separator = "*";
}
static void
gst_pad_template_init (GstPadTemplate * templ)
{
}
static void
gst_pad_template_dispose (GObject * object)
{
GstPadTemplate *templ = GST_PAD_TEMPLATE (object);
g_free (GST_PAD_TEMPLATE_NAME_TEMPLATE (templ));
if (GST_PAD_TEMPLATE_CAPS (templ)) {
gst_caps_unref (GST_PAD_TEMPLATE_CAPS (templ));
}
G_OBJECT_CLASS (padtemplate_parent_class)->dispose (object);
}
/* ALWAYS padtemplates cannot have conversion specifications, it doesn't make
* sense.
* SOMETIMES padtemplates can do whatever they want, they are provided by the
* element.
* REQUEST padtemplates can be reverse-parsed (the user asks for 'sink1', the
* 'sink%d' template is automatically selected), so we need to restrict their
* naming.
*/
static gboolean
name_is_valid (const gchar * name, GstPadPresence presence)
{
const gchar *str;
if (presence == GST_PAD_ALWAYS) {
if (strchr (name, '%')) {
g_warning ("invalid name template %s: conversion specifications are not"
" allowed for GST_PAD_ALWAYS padtemplates", name);
return FALSE;
}
} else if (presence == GST_PAD_REQUEST) {
if ((str = strchr (name, '%')) && strchr (str + 1, '%')) {
g_warning ("invalid name template %s: only one conversion specification"
" allowed in GST_PAD_REQUEST padtemplate", name);
return FALSE;
}
if (str && (*(str + 1) != 's' && *(str + 1) != 'd')) {
g_warning ("invalid name template %s: conversion specification must be of"
" type '%%d' or '%%s' for GST_PAD_REQUEST padtemplate", name);
return FALSE;
}
if (str && (*(str + 2) != '\0')) {
g_warning ("invalid name template %s: conversion specification must"
" appear at the end of the GST_PAD_REQUEST padtemplate name", name);
return FALSE;
}
}
return TRUE;
}
/**
* gst_static_pad_template_get:
* @pad_template: the static pad template
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Converts a #GstStaticPadTemplate into a #GstPadTemplate.
*
* Returns: a new #GstPadTemplate.
*/
GstPadTemplate *
gst_static_pad_template_get (GstStaticPadTemplate * pad_template)
{
GstPadTemplate *new;
if (!name_is_valid (pad_template->name_template, pad_template->presence))
return NULL;
new = g_object_new (gst_pad_template_get_type (),
"name", pad_template->name_template, NULL);
GST_PAD_TEMPLATE_NAME_TEMPLATE (new) = g_strdup (pad_template->name_template);
GST_PAD_TEMPLATE_DIRECTION (new) = pad_template->direction;
GST_PAD_TEMPLATE_PRESENCE (new) = pad_template->presence;
GST_PAD_TEMPLATE_CAPS (new) =
gst_caps_copy (gst_static_caps_get (&pad_template->static_caps));
return new;
}
/**
* gst_pad_template_new:
* @name_template: the name template.
* @direction: the #GstPadDirection of the template.
* @presence: the #GstPadPresence of the pad.
* @caps: a #GstCaps set for the template. The caps are taken ownership of.
*
* Creates a new pad template with a name according to the given template
* and with the given arguments. This functions takes ownership of the provided
* caps, so be sure to not use them afterwards.
*
* Returns: a new #GstPadTemplate.
*/
GstPadTemplate *
gst_pad_template_new (const gchar * name_template,
GstPadDirection direction, GstPadPresence presence, GstCaps * caps)
{
GstPadTemplate *new;
g_return_val_if_fail (name_template != NULL, NULL);
g_return_val_if_fail (caps != NULL, NULL);
g_return_val_if_fail (direction == GST_PAD_SRC
|| direction == GST_PAD_SINK, NULL);
g_return_val_if_fail (presence == GST_PAD_ALWAYS
|| presence == GST_PAD_SOMETIMES || presence == GST_PAD_REQUEST, NULL);
if (!name_is_valid (name_template, presence))
return NULL;
new = g_object_new (gst_pad_template_get_type (),
"name", name_template, NULL);
GST_PAD_TEMPLATE_NAME_TEMPLATE (new) = g_strdup (name_template);
GST_PAD_TEMPLATE_DIRECTION (new) = direction;
GST_PAD_TEMPLATE_PRESENCE (new) = presence;
GST_PAD_TEMPLATE_CAPS (new) = caps;
return new;
}
/**
* gst_pad_template_get_caps:
* @templ: a #GstPadTemplate to get capabilities of.
*
* Gets the capabilities of the pad template.
*
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
* Returns: the #GstCaps of the pad template. If you need to keep a reference to
* the caps, make a copy (see gst_caps_copy ()).
*/
const GstCaps *
gst_pad_template_get_caps (GstPadTemplate * templ)
{
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (GST_IS_PAD_TEMPLATE (templ), NULL);
return GST_PAD_TEMPLATE_CAPS (templ);
}
/**
* gst_pad_set_element_private:
* @pad: the #GstPad to set the private data of.
* @priv: The private data to attach to the pad.
*
* Set the given private data gpointer on the pad.
* This function can only be used by the element that owns the pad.
*/
void
gst_pad_set_element_private (GstPad * pad, gpointer priv)
{
pad->element_private = priv;
}
/**
* gst_pad_get_element_private:
* @pad: the #GstPad to get the private data of.
*
* Gets the private data of a pad.
*
* Returns: a #gpointer to the private data.
*/
gpointer
gst_pad_get_element_private (GstPad * pad)
{
return pad->element_private;
}
/***** ghost pads *****/
GType _gst_ghost_pad_type = 0;
static void gst_ghost_pad_class_init (GstGhostPadClass * klass);
static void gst_ghost_pad_init (GstGhostPad * pad);
static void gst_ghost_pad_dispose (GObject * object);
static void gst_ghost_pad_get_property (GObject * object, guint prop_id,
GValue * value, GParamSpec * pspec);
static void gst_ghost_pad_set_property (GObject * object, guint prop_id,
const GValue * value, GParamSpec * pspec);
static GstPad *ghost_pad_parent_class = NULL;
/* static guint gst_ghost_pad_signals[LAST_SIGNAL] = { 0 }; */
enum
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
{
GPAD_ARG_0,
GPAD_ARG_REAL_PAD
/* fill me */
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
};
GType
gst_ghost_pad_get_type (void)
{
if (!_gst_ghost_pad_type) {
static const GTypeInfo pad_info = {
sizeof (GstGhostPadClass), NULL, NULL,
(GClassInitFunc) gst_ghost_pad_class_init, NULL, NULL,
sizeof (GstGhostPad),
0,
(GInstanceInitFunc) gst_ghost_pad_init,
NULL
};
_gst_ghost_pad_type = g_type_register_static (GST_TYPE_PAD, "GstGhostPad",
&pad_info, 0);
}
return _gst_ghost_pad_type;
}
static void
gst_ghost_pad_class_init (GstGhostPadClass * klass)
{
GObjectClass *gobject_class;
gobject_class = (GObjectClass *) klass;
ghost_pad_parent_class = g_type_class_ref (GST_TYPE_PAD);
gobject_class->dispose = GST_DEBUG_FUNCPTR (gst_ghost_pad_dispose);
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
gobject_class->set_property = GST_DEBUG_FUNCPTR (gst_ghost_pad_set_property);
gobject_class->get_property = GST_DEBUG_FUNCPTR (gst_ghost_pad_get_property);
g_object_class_install_property (gobject_class, GPAD_ARG_REAL_PAD,
g_param_spec_object ("real-pad", "Real pad",
"The real pad for the ghost pad", GST_TYPE_PAD, G_PARAM_READWRITE));
}
static void
gst_ghost_pad_init (GstGhostPad * pad)
{
/* zeroed by glib */
}
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
static void
gst_ghost_pad_dispose (GObject * object)
{
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
g_object_set (object, "real-pad", NULL, NULL);
G_OBJECT_CLASS (ghost_pad_parent_class)->dispose (object);
}
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
static void
gst_ghost_pad_set_property (GObject * object, guint prop_id,
const GValue * value, GParamSpec * pspec)
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
{
GstPad *ghostpad = (GstPad *) object;
GstPad *oldrealpad = (GstPad *) GST_GPAD_REALPAD (ghostpad);
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
GstPad *realpad = NULL;
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
switch (prop_id) {
case GPAD_ARG_REAL_PAD:
realpad = g_value_get_object (value);
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
if (oldrealpad) {
if (realpad == oldrealpad)
return;
else
gst_pad_remove_ghost_pad (oldrealpad, ghostpad);
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
}
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
if (realpad)
gst_pad_add_ghost_pad (realpad, ghostpad);
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
static void
gst_ghost_pad_get_property (GObject * object, guint prop_id,
GValue * value, GParamSpec * pspec)
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
{
switch (prop_id) {
case GPAD_ARG_REAL_PAD:
g_value_set_object (value, GST_GPAD_REALPAD (object));
break;
default:
G_OBJECT_WARN_INVALID_PROPERTY_ID (object, prop_id, pspec);
break;
}
}
/**
* gst_ghost_pad_new:
* @name: the name of the new ghost pad.
* @pad: the #GstPad to create a ghost pad for.
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Creates a new ghost pad associated with @pad, and named @name. If @name is
* %NULL, a guaranteed unique name (across all ghost pads) will be assigned.
*
gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. Original commit message from CVS: 2004-02-10 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (gst_pad_custom_new): Add a FIXME, this is a hacky way to do inheritance. (gst_pad_get_event_masks, gst_pad_get_event_masks_default) (gst_pad_get_query_types, gst_pad_get_query_types_default): Routine docs. (gst_pad_set_link_function, gst_pad_set_fixate_function) (gst_pad_set_getcaps_function): Doc from Dave's negotation random doc. (gst_pad_unlink, gst_pad_is_linked): Docs. (gst_pad_renegotiate): A brief description of capsnego. (gst_pad_try_set_caps): Document. (gst_pad_try_set_caps_nonfixed): Document. (gst_pad_can_link_filtered, gst_pad_link_filtered): Doc fixes. (gst_pad_set_parent): Deprecated (although not out of the API). (gst_pad_get_parent): Deprecated, although many plugins use this. (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): Doc that these are private and will go away in 0.9. (gst_pad_perform_negotiate): Doc. (gst_pad_link_unnegotiate): I think this is meant to be static. (gst_pad_get_negotiated_caps, gst_pad_get_pad_template_caps) (gst_pad_template_get_caps_by_name, gst_pad_check_compatibility) (gst_pad_get_peer): Doc updates. (gst_pad_caps_change_notify): Doc. (gst_pad_alloc_buffer, gst_pad_push, gst_static_pad_template_get) (gst_ghost_pad_new): Doc fixes. * gst/gstobject.c (gst_object_get_parent, gst_object_unparent) (gst_object_check_uniqueness): * gst/gstelement.c (gst_element_add_pad) (gst_element_add_ghost_pad, gst_element_remove_pad) (gst_element_remove_ghost_pad, gst_element_get_pad) (gst_element_get_static_pad, gst_element_get_pad_list) (gst_element_class_get_pad_template_list) (gst_element_class_get_pad_template): Work on the docs. (gst_element_get_pad_template_list): Uses the class method. (gst_element_get_compatible_pad_template): Docs, and consolidate some test conditions. (gst_element_get_pad_from_template): New static function. (gst_element_request_compatible_pad): Docs, and work with non-request compatible templates. (gst_element_get_compatible_pad_filtered): Docs and remove redundant checks. (gst_element_get_compatible_pad, gst_element_link_pads_filtered) (gst_element_link_filtered, gst_element_link_many) (gst_element_link, gst_element_link_pads) (gst_element_unlink_many): Docs. 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-11 13:26:04 +00:00
* Returns: a new ghost #GstPad, or %NULL in case of an error.
*/
GstPad *
gst_ghost_pad_new (const gchar * name, GstPad * pad)
{
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
GstPad *gpad;
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
gpad = g_object_new (GST_TYPE_GHOST_PAD, "name", name, "real-pad", pad, NULL);
GST_CAT_DEBUG (GST_CAT_PADS, "created ghost pad \"%s\" for pad %s:%s",
GST_OBJECT_NAME (gpad), GST_DEBUG_PAD_NAME (pad));
gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. Original commit message from CVS: 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstpad.c (_gst_real_pad_fixate_accumulator): s/pointer/boxed/. * gst/gstmarshal.list (VOID:BOXED, BOXED:BOXED): New marshallers. * gst/gstpad.c (gst_real_pad_class_init): Use a BOXED:BOXED marshaller for ::fixate, and VOID:BOXED for ::caps-nego-failed, with the type=GST_TYPE_CAPS. This allows language bindings to know what kind of data they're dealing with. * gst/gstcaps.c (_gst_caps_value_init): GBoxed values initialize to NULL when g_value_init is called. GstCaps, which rolls its own type implementation, now does the same instead of allocating empty caps. (_gst_caps_initialize, _gst_caps_collect_value, _gst_caps_lcopy_value): Provide collect_value and lcopy_value type table methods. This allows G_VALUE_COLLECT to work. 2004-02-05 Andy Wingo <wingo@pobox.com> * configure.ac: * testsuite/Makefile.am (SUBDIRS): * testsuite/ghostpads/Makefile.am: * testsuite/ghostpads/ghostpads.c: A new test for ghost pads. * gst/gstpad.c (gst_pad_add_ghost_pad, gst_pad_remove_ghost_pad): These two routines are the only ones that set GST_GPAD_REALPAD(gpad), the ghost pad list, and the ghost pad's pad template. They should be made static, depending on ABI needs. (gst_real_pad_dispose): Handle the case of ghost pads without a parent. Assert after dealing with ghost pads that the ghost pad list is empty. (gst_ghost_pad_class_init): New property added, ::real-pad. Can be set after creation. (gst_ghost_pad_dispose): Set ::real-pad to NULL. (gst_ghost_pad_set_property, gst_ghost_pad_get_property): New functions. set_property will call add_ghost_pad/remove_ghost_pad as appropriate. (gst_ghost_pad_new): All the work is offloaded to g_object_new. * gst/gstelement.c (gst_element_add_pad): Handle ghost pads as well. (gst_element_add_ghost_pad): Remove code duplicated from _add_pad. (gst_element_remove_pad): Handle ghost pads as well. (gst_element_remove_ghost_pad): Deprecated (could be removed, depending on API-stability needs). 2004-02-05 Andy Wingo <wingo@pobox.com> * gst/gstbin.[ch]: (gst_bin_get_by_interface): GTypes are scalars, of course they're const
2004-02-05 13:51:26 +00:00
return gpad;
}
/**
* gst_pad_get_internal_links_default:
* @pad: the #GstPad to get the internal links of.
*
* Gets a list of pads to which the given pad is linked to
* inside of the parent element.
* This is the default handler, and thus returns a list of all of the
* pads inside the parent element with opposite direction.
* The caller must free this list after use.
*
* Returns: a newly allocated #GList of pads.
*/
GList *
gst_pad_get_internal_links_default (GstPad * pad)
{
GList *res = NULL;
GstElement *parent;
GList *parent_pads;
GstPadDirection direction;
GstRealPad *rpad;
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
rpad = GST_PAD_REALIZE (pad);
direction = rpad->direction;
parent = GST_PAD_PARENT (rpad);
parent_pads = parent->pads;
while (parent_pads) {
GstRealPad *parent_pad = GST_PAD_REALIZE (parent_pads->data);
if (parent_pad->direction != direction) {
res = g_list_prepend (res, parent_pad);
}
parent_pads = g_list_next (parent_pads);
}
return res;
}
/**
* gst_pad_get_internal_links:
* @pad: the #GstPad to get the internal links of.
*
* Gets a list of pads to which the given pad is linked to
* inside of the parent element.
* The caller must free this list after use.
*
* Returns: a newly allocated #GList of pads.
*/
GList *
gst_pad_get_internal_links (GstPad * pad)
{
GList *res = NULL;
GstRealPad *rpad;
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
rpad = GST_PAD_REALIZE (pad);
if (GST_RPAD_INTLINKFUNC (rpad))
res = GST_RPAD_INTLINKFUNC (rpad) (GST_PAD_CAST (rpad));
return res;
}
static gboolean
gst_pad_event_default_dispatch (GstPad * pad, GstEvent * event)
{
GList *orig, *pads;
gboolean result;
GST_INFO_OBJECT (pad, "Sending event %p to all internally linked pads",
event);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
result = (GST_PAD_DIRECTION (pad) == GST_PAD_SINK);
orig = pads = gst_pad_get_internal_links (pad);
while (pads) {
GstPad *eventpad = GST_PAD (pads->data);
pads = g_list_next (pads);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
/* for all of the internally-linked pads that are actually linked */
if (GST_PAD_IS_LINKED (eventpad)) {
if (GST_PAD_DIRECTION (eventpad) == GST_PAD_SRC) {
/* for each pad we send to, we should ref the event; it's up
* to downstream to unref again when handled. */
GST_LOG_OBJECT (pad, "Reffing and sending event %p to %s:%s", event,
GST_DEBUG_PAD_NAME (eventpad));
gst_event_ref (event);
gst_pad_push_event (eventpad, event);
} else {
/* we only send the event on one pad, multi-sinkpad elements
* should implement a handler */
GST_LOG_OBJECT (pad, "sending event %p to one sink pad %s:%s", event,
GST_DEBUG_PAD_NAME (eventpad));
result = gst_pad_push_event (eventpad, event);
goto done;
}
}
}
/* we handled the incoming event so we unref once */
GST_LOG_OBJECT (pad, "handled event %p, unreffing", event);
gst_event_unref (event);
done:
g_list_free (orig);
return result;
}
/**
* gst_pad_event_default:
* @pad: a #GstPad to call the default event handler on.
* @event: the #GstEvent to handle.
*
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
* Invokes the default event handler for the given pad. End-of-stream and
* discontinuity events are handled specially, and then the event is sent to all
* pads internally linked to @pad. Note that if there are many possible sink
* pads that are internally linked to @pad, only one will be sent an event.
* Multi-sinkpad elements should implement custom event handlers.
*
* Returns: TRUE if the event was sent succesfully.
*/
gboolean
gst_pad_event_default (GstPad * pad, GstEvent * event)
{
g_return_val_if_fail (GST_IS_PAD (pad), FALSE);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (event != NULL, FALSE);
return gst_pad_event_default_dispatch (pad, event);
}
/**
* gst_pad_dispatcher:
* @pad: a #GstPad to dispatch.
* @dispatch: the #GstDispatcherFunction to call.
* @data: gpointer user data passed to the dispatcher function.
*
* Invokes the given dispatcher function on all pads that are
* internally linked to the given pad.
* The GstPadDispatcherFunction should return TRUE when no further pads
* need to be processed.
*
* Returns: TRUE if one of the dispatcher functions returned TRUE.
*/
gboolean
gst_pad_dispatcher (GstPad * pad, GstPadDispatcherFunction dispatch,
gpointer data)
{
gboolean res = FALSE;
GList *int_pads, *orig;
g_return_val_if_fail (GST_IS_PAD (pad), FALSE);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (dispatch != NULL, FALSE);
orig = int_pads = gst_pad_get_internal_links (pad);
while (int_pads) {
GstRealPad *int_rpad = GST_PAD_REALIZE (int_pads->data);
GstRealPad *int_peer = GST_RPAD_PEER (int_rpad);
if (int_peer) {
res = dispatch (GST_PAD (int_peer), data);
if (res)
break;
}
int_pads = g_list_next (int_pads);
}
g_list_free (orig);
return res;
}
/**
* gst_pad_push_event:
* @pad: a #GstPad to push the event to.
* @event: the #GstEvent to send to the pad.
*
* Sends the event to the peer of the given pad.
*
* Returns: TRUE if the event was handled.
*/
gboolean
gst_pad_push_event (GstPad * pad, GstEvent * event)
{
GstRealPad *peerpad;
gboolean result;
g_return_val_if_fail (GST_IS_REAL_PAD (pad), FALSE);
g_return_val_if_fail (event != NULL, FALSE);
GST_LOCK (pad);
peerpad = GST_RPAD_PEER (pad);
if (peerpad == NULL)
goto not_linked;
GST_UNLOCK (pad);
result = gst_pad_send_event (GST_PAD_CAST (peerpad), event);
return result;
/* ERROR handling */
not_linked:
result = FALSE;
goto error;
error:
GST_UNLOCK (pad);
return result;
}
/**
* gst_pad_send_event:
* @pad: a #GstPad to send the event to.
* @event: the #GstEvent to send to the pad.
*
* Sends the event to the pad. This function can be used
* by applications to send events in the pipeline.
*
* Returns: TRUE if the event was handled.
*/
gboolean
gst_pad_send_event (GstPad * pad, GstEvent * event)
{
gboolean result = FALSE;
GstRealPad *rpad;
GstPadEventFunction eventfunc;
g_return_val_if_fail (GST_IS_PAD (pad), FALSE);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (event != NULL, FALSE);
rpad = GST_PAD_REALIZE (pad);
if (GST_EVENT_SRC (event) == NULL)
GST_EVENT_SRC (event) = gst_object_ref (GST_OBJECT (rpad));
GST_CAT_DEBUG (GST_CAT_EVENT, "have event type %d on pad %s:%s",
GST_EVENT_TYPE (event), GST_DEBUG_PAD_NAME (rpad));
if ((eventfunc = GST_RPAD_EVENTFUNC (rpad)) == NULL)
goto no_function;
result = eventfunc (GST_PAD_CAST (rpad), event);
return result;
/* ERROR handling */
no_function:
g_warning ("pad %s:%s has no event handler, file a bug.",
GST_DEBUG_PAD_NAME (rpad));
result = FALSE;
gst_event_unref (event);
return result;
}
typedef struct
{
GstFormat src_format;
gint64 src_value;
GstFormat *dest_format;
gint64 *dest_value;
}
GstPadConvertData;
static gboolean
gst_pad_convert_dispatcher (GstPad * pad, GstPadConvertData * data)
{
return gst_pad_convert (pad, data->src_format, data->src_value,
data->dest_format, data->dest_value);
}
/**
* gst_pad_convert_default:
* @pad: a #GstPad to invoke the default converter on.
* @src_format: the source #GstFormat.
* @src_value: the source value.
* @dest_format: a pointer to the destination #GstFormat.
* @dest_value: a pointer to the destination value.
*
* Invokes the default converter on a pad.
* This will forward the call to the pad obtained
* using the internal link of
* the element.
*
* Returns: TRUE if the conversion could be performed.
*/
gboolean
gst_pad_convert_default (GstPad * pad,
GstFormat src_format, gint64 src_value,
GstFormat * dest_format, gint64 * dest_value)
{
GstPadConvertData data;
g_return_val_if_fail (GST_IS_PAD (pad), FALSE);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (dest_format != NULL, FALSE);
g_return_val_if_fail (dest_value != NULL, FALSE);
data.src_format = src_format;
data.src_value = src_value;
data.dest_format = dest_format;
data.dest_value = dest_value;
return gst_pad_dispatcher (pad, (GstPadDispatcherFunction)
gst_pad_convert_dispatcher, &data);
}
/**
* gst_pad_convert:
* @pad: a #GstPad to invoke the default converter on.
* @src_format: the source #GstFormat.
* @src_value: the source value.
* @dest_format: a pointer to the destination #GstFormat.
* @dest_value: a pointer to the destination value.
*
* Invokes a conversion on the pad.
*
* Returns: TRUE if the conversion could be performed.
*/
gboolean
gst_pad_convert (GstPad * pad,
GstFormat src_format, gint64 src_value,
GstFormat * dest_format, gint64 * dest_value)
{
GstRealPad *rpad;
g_return_val_if_fail (GST_IS_PAD (pad), FALSE);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (dest_format != NULL, FALSE);
g_return_val_if_fail (dest_value != NULL, FALSE);
if (src_format == *dest_format) {
*dest_value = src_value;
return TRUE;
}
rpad = GST_PAD_REALIZE (pad);
if (GST_RPAD_CONVERTFUNC (rpad)) {
return GST_RPAD_CONVERTFUNC (rpad) (GST_PAD (rpad), src_format,
src_value, dest_format, dest_value);
}
return FALSE;
}
typedef struct
{
GstQueryType type;
GstFormat *format;
gint64 *value;
}
GstPadQueryData;
static gboolean
gst_pad_query_dispatcher (GstPad * pad, GstPadQueryData * data)
{
return gst_pad_query (pad, data->type, data->format, data->value);
}
/**
* gst_pad_query_default:
* @pad: a #GstPad to invoke the default query on.
* @type: the #GstQueryType of the query to perform.
* @format: a pointer to the #GstFormat of the result.
* @value: a pointer to the result.
*
* Invokes the default query function on a pad.
*
* Returns: TRUE if the query could be performed.
*/
gboolean
gst_pad_query_default (GstPad * pad, GstQueryType type,
GstFormat * format, gint64 * value)
{
GstPadQueryData data;
g_return_val_if_fail (GST_IS_PAD (pad), FALSE);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (format != NULL, FALSE);
g_return_val_if_fail (value != NULL, FALSE);
data.type = type;
data.format = format;
data.value = value;
return gst_pad_dispatcher (pad, (GstPadDispatcherFunction)
gst_pad_query_dispatcher, &data);
}
/**
* gst_pad_query:
* @pad: a #GstPad to invoke the default query on.
* @type: the #GstQueryType of the query to perform.
* @format: a pointer to the #GstFormat asked for.
* On return contains the #GstFormat used.
* @value: a pointer to the result.
*
* Queries a pad for one of the available properties. The format will be
* adjusted to the actual format used when specifying formats such as
* GST_FORMAT_DEFAULT.
* FIXME: Tell if the format can be adjusted when specifying a definite format.
*
* Returns: TRUE if the query could be performed.
*/
gboolean
gst_pad_query (GstPad * pad, GstQueryType type,
GstFormat * format, gint64 * value)
{
GstRealPad *rpad;
g_return_val_if_fail (GST_IS_PAD (pad), FALSE);
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (format != NULL, FALSE);
g_return_val_if_fail (value != NULL, FALSE);
rpad = GST_PAD_REALIZE (pad);
g_return_val_if_fail (rpad, FALSE);
if (GST_RPAD_QUERYFUNC (rpad))
return GST_RPAD_QUERYFUNC (rpad) (GST_PAD (pad), type, format, value);
return FALSE;
}
static gboolean
gst_pad_get_formats_dispatcher (GstPad * pad, const GstFormat ** data)
{
*data = gst_pad_get_formats (pad);
return TRUE;
}
/**
* gst_pad_get_formats_default:
* @pad: a #GstPad to query
*
* Invoke the default format dispatcher for the pad.
*
* Returns: An array of GstFormats ended with a 0 value.
*/
const GstFormat *
gst_pad_get_formats_default (GstPad * pad)
{
GstFormat *result = NULL;
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
gst_pad_dispatcher (pad, (GstPadDispatcherFunction)
gst_pad_get_formats_dispatcher, &result);
return result;
}
/**
* gst_pad_get_formats:
* @pad: a #GstPad to query
*
* Gets the list of supported formats from the pad.
*
* Returns: An array of GstFormats ended with a 0 value.
*/
const GstFormat *
gst_pad_get_formats (GstPad * pad)
{
GstRealPad *rpad;
gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. Original commit message from CVS: 2004-02-24 Andy Wingo <wingo@pobox.com> * gst/gstelement.c (gst_element_dispose): Protect against multiple invocations. * gst/schedulers/gstoptimalscheduler.c I added a mess of prototypes at the top of the file by way of documentation. Some of the operations on chains and groups were re-organized. (create_group): Added a type argument so if the group is enabled, the setup_group_scheduler knows what to do. (group_elements): Added a type argument here, too, to be passed on to create_group. (group_element_set_enabled): If an unlinked PLAYING element is added to a bin, we have to create a new group to hold the element, and this function will be called before the group is added to the chain. Thus we have a valid case for group->chain==NULL. Instead of calling chain_group_set_enabled, just set the flag on the group (the chain's status will be set when the group is added to it). (gst_opt_scheduler_state_transition, chain_group_set_enabled): Setup the group scheduler when the group is enabled, not specifically when an element goes PAUSED->PLAYING. This means PLAYING elements can be added, linked, and scheduled into a PLAYING pipeline, as was intended. (add_to_group): Don't ref the group twice. I don't know when this double-ref got in here. Removing it has the potential to cause segfaults if other parts of the scheduler are buggy. If you find that the scheduler is segfaulting for you, put in an extra ref here and see if that hacks over the underlying issue. Of course, then find out what code is unreffing a group it doesn't own... (create_group): Make the extra refcount floating, and remove it after adding the element. This means that... (unref_group): Destroy when the refcount reaches 0, not 1, like every other refcounted object in the known universe. (remove_from_group): When a group becomes empty, set it to be not active, and remove it from its chain. Don't unref it again, there's no floating reference any more. (destroy_group): We have to remove the group from the chain in remove_from_group (rather than here) to break refcounting cycles (the chain always has a ref on the group). So assert that group->chain==NULL. (ref_group_by_count): Removed, it was commented out anyway. (merge_chains): Use the remove_from_chain and add_to_chain primitives to do the reparenting, instead of rolling our own implementation. (add_to_chain): The first non-disabled group in the chain's group list will be the entry point for the chain. Because buffers can accumulate in loop elements' peer bufpens, we preferentially schedule loop groups before get groups to avoid unnecessary execution of get-based groups when the bufpens are already full. (gst_opt_scheduler_schedule_run_queue): Debug fixes. (get_group_schedule_function): Ditto. (loop_group_schedule_function): Ditto. (gst_opt_scheduler_loop_wrapper): Ditto. (gst_opt_scheduler_iterate): Ditto. I understand the opt scheduler now, yippee! * gst/gstpad.c: All throughout, added FIXMEs to look at for 0.9. (gst_pad_get_name, gst_pad_set_chain_function) (gst_pad_set_get_function, gst_pad_set_event_function) (gst_pad_set_event_mask_function, gst_pad_get_event_masks) (gst_pad_get_event_masks_default, gst_pad_set_convert_function) (gst_pad_set_query_function, gst_pad_get_query_types) (gst_pad_get_query_types_default) (gst_pad_set_internal_link_function) (gst_pad_set_formats_function, gst_pad_set_link_function) (gst_pad_set_fixate_function, gst_pad_set_getcaps_function) (gst_pad_set_bufferalloc_function, gst_pad_unlink) (gst_pad_renegotiate, gst_pad_set_parent, gst_pad_get_parent) (gst_pad_add_ghost_pad, gst_pad_proxy_getcaps) (gst_pad_proxy_pad_link, gst_pad_proxy_fixate) (gst_pad_get_pad_template_caps, gst_pad_check_compatibility) (gst_pad_get_peer, gst_pad_get_allowed_caps) (gst_pad_alloc_buffer, gst_pad_push, gst_pad_pull) (gst_pad_selectv, gst_pad_select, gst_pad_template_get_caps) (gst_pad_event_default_dispatch, gst_pad_event_default) (gst_pad_dispatcher, gst_pad_send_event, gst_pad_convert_default) (gst_pad_convert, gst_pad_query_default, gst_pad_query) (gst_pad_get_formats_default, gst_pad_get_formats): Better argument checks, and some doc fixes. (gst_pad_custom_new_from_template): Um, does anyone use these functions? Actually make a custom pad instead of a normal one. (gst_pad_try_set_caps): Transpose some checks. (gst_pad_try_set_caps_nonfixed): Same, and use a macro to check if the pad is in negotiation. (gst_pad_try_relink_filtered): Use pad_link_prepare. * gst/gstelement.c: Remove prototypes also defined in gstclock.h. * gst/gstelement.h: * gst/gstclock.h: Un-deprecate the old clocking API, as discussed on the list.
2004-02-25 13:16:12 +00:00
g_return_val_if_fail (GST_IS_PAD (pad), NULL);
rpad = GST_PAD_REALIZE (pad);
if (GST_RPAD_FORMATSFUNC (rpad))
return GST_RPAD_FORMATSFUNC (rpad) (GST_PAD (pad));
return NULL;
}