Original commit message from CVS:
* check/Makefile.am:
* check/gst/gstevents.c: (GST_START_TEST), (event_probe),
(test_event), (timediff), (gstevents_suite), (main):
I wrote a test!
* docs/design/part-seeking.txt:
Spelling correction
* docs/gst/tmpl/gstevent.sgml:
Docs updates.
* gst/base/gstbasesink.c: (gst_base_sink_handle_object):
Treat a buffer-without-newsegment the same as a receiving
a newsegment not in time format, and disable syncing to the clock
with a warning.
* gst/gstbus.c: (gst_bus_set_sync_handler):
Assert if anyone tries to replace the existing sync_handler for bus,
as only the owner should be setting it.
* gst/gstevent.h:
Have a fixed set of custom event enums with events identified by
their structure name (as in 0.8), rather than a free-for-all
allowing collisions between enum values from different plugins.
* gst/gstpad.c: (gst_pad_class_init):
Docs change.
* gst/gstqueue.c: (gst_queue_handle_sink_event):
Handle out-of-band downstream events from the sending thread.
Original commit message from CVS:
* docs/design/part-seeking.txt:
Some small additions.
* gst/base/gstbasesink.c: (gst_base_sink_handle_object),
(gst_base_sink_get_times), (gst_base_sink_do_sync),
(gst_base_sink_activate_push), (gst_base_sink_activate_pull):
* gst/base/gstbasesink.h:
discont values are gint64, handle the math correctly.
* gst/base/gstbasesrc.c: (gst_base_src_loop):
Make the basesrc report error if the source pad is not linked.
* gst/gstqueue.c: (gst_queue_link_src), (gst_queue_chain),
(gst_queue_loop), (gst_queue_handle_src_query),
(gst_queue_src_activate_push):
Make queue collect data even if the srcpad is not linked.
Start pushing out data as soon as it is linked.
* gst/gstutils.c: (gst_element_unlink), (gst_flow_get_name):
* gst/gstutils.h:
Added gst_flow_get_name() to ease error reporting.
Original commit message from CVS:
* gst/base/gstbasesrc.c: (gst_base_src_init),
(gst_base_src_do_seek), (gst_base_src_loop), (gst_base_src_start):
* gst/base/gstbasesrc.h:
Add a gboolean to decide when to push out a discont.
* gst/gstqueue.c: (gst_queue_handle_sink_event), (gst_queue_chain),
(gst_queue_loop), (gst_queue_handle_src_query),
(gst_queue_sink_activate_push), (gst_queue_src_activate_push),
(gst_queue_set_property), (gst_queue_get_property):
Some cleanups.
* tests/threadstate/threadstate1.c: (main):
Make a thread test compile and run... very silly..
Original commit message from CVS:
* gst/gstqueue.c: (gst_queue_init), (gst_queue_handle_sink_event),
(gst_queue_chain), (gst_queue_loop), (gst_queue_handle_src_event),
(gst_queue_handle_src_query), (gst_queue_sink_activate_push),
(gst_queue_src_activate_push), (gst_queue_change_state),
(gst_queue_get_property):
* gst/gstqueue.h:
Propagate GstFlowReturn more intelligently upstream and output
an ERROR/EOS when streaming stopped due to fatal error.
Original commit message from CVS:
2005-06-27 Andy Wingo <wingo@pobox.com>
* gst/base/gsttypefindhelper.c (gst_type_find_helper): Unref any
remaining buffer.
* gst/gsttrace.c (gst_alloc_trace_list_sorted): New helper,
returns a sorted copy of the trace list.
(gst_alloc_trace_print_live): New API, only prints traces with
live objects. Sort the list.
(gst_alloc_trace_print_all): Sort the list.
(gst_alloc_trace_print): Align columns.
* gst/elements/gstttypefindelement.c:
* gst/elements/gsttee.c:
* gst/base/gstbasesrc.c:
* gst/base/gstbasesink.c:
* gst/base/gstbasetransform.c:
* gst/gstqueue.c: Adapt for pad activation changes.
* gst/gstpipeline.c (gst_pipeline_init): Unref after parenting
sched.
(gst_pipeline_dispose): Drop ref on sched.
* gst/gstpad.c (gst_pad_init): Set the default activate func.
(gst_pad_activate_default): Push mode by default.
(pre_activate_switch, post_activate_switch): New stubs, things to
do before and after switching activation modes on pads.
(gst_pad_set_active): Take a boolean and not a mode, dispatch to
the pad's activate function to choose which mode to activate.
Shortcut on deactivation and call the right function directly.
(gst_pad_activate_pull): New API, (de)activates a pad in pull
mode.
(gst_pad_activate_push): New API, same for push mode.
(gst_pad_set_activate_function)
(gst_pad_set_activatepull_function)
(gst_pad_set_activatepush_function): Setters for new API.
* gst/gstminiobject.c (gst_mini_object_new, gst_mini_object_free):
Trace all miniobjects.
(gst_mini_object_make_writable): Unref the arg if we copy, like
gst_caps_make_writable.
* gst/gstmessage.c (_gst_message_initialize): No trace init.
* gst/gstghostpad.c (gst_proxy_pad_do_activate)
(gst_proxy_pad_do_activatepull, gst_proxy_pad_do_activatepush):
Adapt for new pad API.
* gst/gstevent.c (_gst_event_initialize): Don't initialize trace.
* gst/gstelement.h:
* gst/gstelement.c (gst_element_iterate_src_pads)
(gst_element_iterate_sink_pads): New API functions.
* gst/gstelement.c (iterator_fold_with_resync): New utility,
should fold into gstiterator.c in some form.
(gst_element_pads_activate): Simplified via use of fold and
delegation of decisions to gstpad->activate.
* gst/gstbus.c (gst_bus_source_finalize): Set the bus to NULL,
help in debugging.
* gst/gstbuffer.c (_gst_buffer_initialize): Ref the buffer type
class once in init, like gstmessage. Didn't run into this issue
but it seems correct. Don't initialize a trace, gstminiobject does
that.
* check/pipelines/simple_launch_lines.c (test_stop_from_app): New
test, runs fakesrc ! fakesink, stopping on ::handoff via a message
to the bus.
(assert_live_count): New util function, uses alloc traces to check
cleanup.
* check/gst/gstghostpad.c (test_ghost_pads): More refcount checks.
To be modified when unlink drops the internal pad.
Original commit message from CVS:
2005-06-08 Andy Wingo <wingo@pobox.com>
* gst/gstutils.c: RPAD fixes all around.
(gst_element_link_pads): Refcounting fixes.
* tools/gst-inspect.c:
* tools/gst-xmlinspect.c:
* parse/grammar.y:
* gst/base/gsttypefindhelper.c:
* gst/base/gstbasesink.c:
* gst/gstqueue.c: RPAD fixes.
* gst/gstghostpad.h:
* gst/gstghostpad.c: New ghost pad implementation as full proxy
pads. The tricky thing is they provide both source and sink
interfaces, since they proxy the internal pad for the external
pad, and vice versa. Implement with lower-level ProxyPad objects,
with the interior proxy pad as a child of the exterior ghost pad.
Should write a doc on this.
* gst/gstpad.h: s/RPAD/PAD/, s/RealPad/Pad/.
(gst_pad_set_name, gst_pad_set_parent): Macros removed, use
gst_object API.
* gst/gstpad.c: Big changes. No more stub base GstPad, now all
pads are real pads. No ghost pads in this file. Not documenting
the myriad s/RPAD/PAD/ and REALIZE fixes.
(gst_pad_class_init): Add properties for "direction" and
"template". Both are construct-only, so they can't change during
the life of the pad. Fixes properly deriving from GstPad.
(gst_pad_custom_new, gst_pad_custom_new_from_template): Gone. For
derived objects, just set properties when creating the objects via
g_object_new.
(gst_pad_get_parent): Implement as a function, return NULL if the
parent is not an element.
(gst_pad_get_real_parent, gst_pad_add_ghost_pad)
(gst_pad_remove_ghost_pad, gst_pad_realize): Removed.
* gst/gstobject.c (gst_object_class_init): Make name a construct
property. Don't set it in the object init.
* gst/gstelement.c (gst_element_add_pad): Don't allow adding pads
with UNKNOWN direction.
(gst_element_add_ghost_pad): Remove non-orthogonal API. Replace
with gst_element_add_pad (e, gst_ghost_pad_new (name, pad)).
(gst_element_remove_pad): Remove ghost-pad special cases.
(gst_element_pads_activate): Remove rpad cruft.
* gst/gstbin.c (gst_bin_change_state): Use gst_pad_get_parent to
catch the pad's-parent-not-an-element case.
* gst/gst.h: Include gstghostpad.h.
* gst/gst.c (init_post): No more real, ghost pads.
* gst/Makefile.am: Add gstghostpad.[ch].
* check/Makefile.am:
* check/gst/gstbin.c:
* check/gst/gstghostpad.c (test_ghost_pads): Check that linking
into a bin creates ghost pads, and that the refcounts are right.
Partly moved from gstbin.c.
Original commit message from CVS:
* gst/gstqueue.c: (gst_queue_handle_sink_event), (gst_queue_chain),
(gst_queue_loop), (gst_queue_handle_src_event):
Remove old pieces of code
Flushing the queue in an upstream event is a very bad idea.
Original commit message from CVS:
* gst/gstbin.c: (bin_element_is_sink), (has_ancestor),
(bin_element_is_semi_sink), (append_child), (gst_bin_change_state):
* gst/gstpad.c: (gst_pad_set_active), (gst_pad_link_prepare),
(gst_pad_link), (gst_pad_accept_caps), (gst_pad_query),
(gst_pad_send_event), (gst_pad_start_task):
* gst/gstqueue.c: (gst_queue_init), (gst_queue_locked_flush),
(gst_queue_handle_sink_event), (gst_queue_chain), (gst_queue_loop),
(gst_queue_sink_activate), (gst_queue_src_activate),
(gst_queue_change_state):
* gst/gstqueue.h:
Fix state changes for non sinks. We now change sinks, then elements
with unconnected srcpads, then the rest.
More efficient queue unlocking in flush and state changes.
Set the pad activate mode even if it does not have an activate
function.
Original commit message from CVS:
* gst/base/gstbasesink.c: (gst_basesink_finish_preroll),
(gst_basesink_chain_unlocked), (gst_basesink_activate):
* gst/gstminiobject.c: (gst_mini_object_get_type),
(gst_mini_object_free):
* gst/gstpad.c: (gst_pad_accept_caps), (gst_pad_query),
(gst_pad_push), (gst_pad_push_event):
* gst/gstqueue.c: (gst_queue_change_state):
Don't queue buffers in basesink when we are flushing.
Unref buffer when flushing in basesink.
Flush queue when going to READY
Unref buffer when _push() returns an error.
Don't free MiniObject instance when refcount is incremented
in _finalize() so that we can recover objects.
Original commit message from CVS:
Remove old query functions. Ported old code.
Added position/convert helper functions to gstutils.
Reordered gstpad.c code, grouping relevant things.
Remove gst_message_new(), always need to speficy a specific
message.
Original commit message from CVS:
* gst/base/gstbasetransform.c: (gst_base_transform_proxy_getcaps):
* gst/elements/gstcapsfilter.c: (gst_capsfilter_getcaps):
* gst/gstpad.c: (gst_pad_peer_get_caps):
* gst/gstqueue.c: (gst_queue_init), (gst_queue_getcaps),
(gst_queue_bufferalloc), (gst_queue_handle_sink_event),
(gst_queue_src_activate), (gst_queue_change_state):
* gst/gstqueue.h:
* gst/gstutils.c: (gst_element_get_compatible_pad_template),
(intersect_caps_func):
Fix gst_pad_peer_get_caps(), make it return NULL if no peer.
Always take QUEUE_LOCK after STREAM_LOCK or we might deadlock.
Some fixes for the peer_get_caps() change.
Original commit message from CVS:
Next big merge.
Added GstBus for mainloop integration.
Added GstMessage for sending notifications on the bus.
Added GstTask as an abstraction for pipeline entry points.
Removed GstThread.
Removed Schedulers.
Simplified GstQueue for multithreaded core.
Made _link threadsafe, removed old capsnego.
Added STREAM_LOCK and PREROLL_LOCK in GstPad.
Added pad blocking functions.
Reworked scheduling functions in GstPad to prepare for
scheduling updates soon.
Moved events out of data stream.
Simplified GstEvent types.
Added return values to push/pull.
Removed clocking from GstElement.
Added prototypes for state change function for next merge.
Removed iterate from bins and state change management.
Fixed some elements, disabled others for now.
Fixed -inspect and -launch.
Added check for GstBus.
Original commit message from CVS:
Remove threadsafe properties. Fix elements because GObject
complains when installing a property before declaring a
set/get_property handler.
Rearrange gstelement.h file, use STATE macros for state locks.
Free mutexes in the finalize method instead of dispose.
Original commit message from CVS:
First THREADED backport attempt, focusing on adding locks and
making sure the API is threadsafe. Needs more work. More docs
follow this week.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_remove_func):
Explicitely make an element release locks in a group when being
remove from a bin.
* gst/gstqueue.c: (gst_queue_chain), (gst_queue_get):
If there's no scheduler, always return immediately (similar to
gst_element_interrupt).
Original commit message from CVS:
* docs/gst/gstreamer-sections.txt:
* docs/gst/tmpl/gstevent.sgml:
* gst/gstevent.c: (gst_event_new_filler_stamped),
(gst_event_filler_get_duration):
* gst/gstevent.h:
Add two new functions for filler events (which are used to
synchronize streams if one of them is not having any data
for a while) without interrupting the actual data-stream.
Basically a no-op.
* gst/gstqueue.c: (gst_queue_init), (gst_queue_getcaps),
(gst_queue_link_sink), (gst_queue_link_src),
(gst_queue_change_state):
Allow for renegotiation while filled. Required for stream
switching while playing.
Original commit message from CVS:
* gst/gstqueue.c: (gst_queue_init), (gst_queue_getcaps),
(gst_queue_link), (gst_queue_handle_src_query):
Reverted to 1.110 until this makes the testsuite and various
apps work.
Original commit message from CVS:
* gst/gstcaps.c: (gst_caps_is_fixed_foreach):
* gst/gstpad.c: (_gst_pad_default_fixate_value),
(_gst_pad_default_fixate_foreach):
* gst/gstvalue.c: (gst_type_is_fixed), (gst_value_is_fixed):
* gst/gstvalue.h:
Deprecate _type_is_fixed, use _value_is_fixed instead, since
in some cases (arrays), the fixedness depends on the content.
* gst/gstqueue.c: (gst_queue_handle_src_query):
Check for availability before doing something.
Original commit message from CVS:
2004-11-24 Martin Soto <martinsoto@users.sourceforge.net>
* gst/gstqueue.c (gst_queue_link_sink): Grab the lock only when
it is necessary to wait.
Original commit message from CVS:
2004-11-23 Martin Soto <martinsoto@users.sourceforge.net>
* gst/gstqueue.c (gst_queue_init, gst_queue_link_sink)
(gst_queue_link_src): Allow for renegotiating the caps of the sink
pad. The queue will now wait until it is empty and forward the new
caps to the source.
* gst/gstbin.c (gst_bin_set_element_sched)
(gst_bin_unset_element_sched): Make sure that all elements and
links are registered and unregistered with the scheduler exactly
once. This elaborates on a fix by Benjamin Otte, but
guarantees that decoupled elements are also registered.
Original commit message from CVS:
* gst/gstqueue.c: (gst_queue_init), (gst_queue_handle_src_query):
Subtract current queue contents from position queries.
Original commit message from CVS:
* gst/gstqueue.c: (gst_queue_dispose), (gst_queue_change_state):
fix memory leaks (fixes#142795). Initial patch by Sebastien Cote
* libs/gst/bytestream/adapter.c:
* libs/gst/bytestream/adapter.h:
fix copyright in header and typo in debugging category name
Original commit message from CVS:
* gst/elements/gstfilesrc.c: (gst_filesrc_get_mmap):
compute mapsize correctly
* gst/elements/gstfilesrc.c: (gst_filesrc_get_mmap):
use correct datatypes when calling a varargs function
* gst/elements/gsttypefindelement.c: (stop_typefinding):
push a DISCONT event as first thing
* gst/gst_private.h:
* gst/gstinfo.c: (_gst_debug_init):
remove GST_DATAFLOW debugging category
* gst/gstbin.c: (gst_bin_iterate):
use GST_SCHEDULING category
* gst/gstpad.c: (gst_pad_get_type), (_invent_event),
(gst_pad_push), (gst_pad_pull), (gst_pad_call_chain_function),
(gst_pad_call_get_function):
add GST_DATAFLOW to easily track flow of buffers or events.
* gst/gstqueue.c: (gst_queue_get_type),
(gst_queue_handle_pending_events), (gst_queue_chain),
(gst_queue_get), (gst_queue_handle_src_event):
use own static debugging category GST_DATAFLOW for dataflow,
use DEBUG category for showing which path events go, use LOG
category for buffers.
Original commit message from CVS:
* gst/gstqueue.c: (gst_queue_init):
Reset default max. values in queues. Reason is simply to avoid
braindead use. If you want wider values, use the properties. The
default is supposed to always work. Wider values would make this
beast a memory hog by default (250 full-PAL RGB32 video frames?
That's 440 MB! No thank you).
Original commit message from CVS:
2004-02-15 Julien MOUTTE <julien@moutte.net>
* gst/gstbin.c: (gst_bin_change_state), (gst_bin_iterate): Fix a big
bug that was breaking pipelines like sinesrc ! { queue ! osssink } when
an error was thrown by osssink. Basically a state change failure for
an element in a different scheduling group was considered as successfull , which means that caps nego was going on and weird stuff happened. Like i wrote in the comment there, if someone want to revert that please
drop me a mail explaining why because i really see no point in keeping
that broken behaviour there.
* gst/gstqueue.c: (gst_queue_get): Add a safety check as the queue CAN
be empty, we then return NULL which will trigger a nice error when
pulling from the pad.
Original commit message from CVS:
2004-02-09 Julien MOUTTE <julien@moutte.net>
* gst/gstdata.c: (gst_data_ref): Adding a categorized debug on data_ref
because data_unref has one and i prefer the debug to be symetric.
* gst/gstqueue.c: (gst_queue_locked_flush): Fix a huge memleak. Buffers
were refed when added to the queue and unrefed only once when the queue
was flushed. Now the flush handler unref the buffers two times : first
unref for the ref added when pushing in the queue's tail and second
unref to destroy the flushed buffer.
