Drop in old GstBus code for the release to play it safe, since
regressions that are apparently hard to track down and reproduce
have been reported (on windows/OSX mostly) against the lockfree
version, and more time is needed to fix them.
This reverts commit 03391a8970.
This reverts commit 43cdbc17e6.
This reverts commit 80eb160e0f.
This reverts commit c41b0ade28.
This reverts commit 874d60e589.
This reverts commit 79370d4b17.
This reverts commit 2cb3e52351.
This reverts commit bd1c400114.
This reverts commit 4bf8f1524f.
This reverts commit 14d7db1b52.
https://bugzilla.gnome.org/show_bug.cgi?id=647493
This allows to only create the socketpair when it is really required instead
of always creating it and immediately destroying it again for child buses.
https://bugzilla.gnome.org/show_bug.cgi?id=647005
This is used by GstBin to create a child bus without
a socketpair because child buses will always work
synchronous. Otherwise too many sockets could be
created and the limit of file descriptors for the
process could be reached.
Fixes bug #646624.
basesrc's default event handler returns TRUE regardless of whether the
event is handled or not. This fixes the handler to conform with the
expected behaviour (which is to only return TRUE when the event has
actually benn handled). gst_bin_do_latency_func() depended on this
(incorrect) behaviour, and is now modified as well.
(Remaining 1-liner change in gstbasesrc.c is to keep gst-indent happy)
Track elements tagged with the IS_SOURCE flag in a similar way we track the sink
elements. This allows us to efficiently dispatch downstream events to the right
elements.
Add an option to forward all the internal messages that would otherwise be
filtered such as EOS, SEGMENT and ASYNC messages.
This allows the application to, for example, detect that a partial pipeline is
prerolled or reached eos.
The original messages are wrapped inside an element message because the parent
bins are not supposed to see those internal messages escape.
When there is a sink inside a bin, the SINK flag is set on the bin. When we are
trying to iterate the source elements, also include the bins with the SINK flag
because they could also contain source elements, in which case they are also a
source.
This solves the case where sending an EOS to a pipeline didn't get dispatched to
all source elements.
See #625597
Since everything GstXML related has been deprecated, we can now skip the
libxml includes from the public headers when GST_DISABLE_DEPRECATED is
defined.
See #463435.
Pipeline serialisation to and from XML is horribly broken for all
but the most simple use cases, and will likely never be fixed.
Make sure everyone playing around with these tools is aware of
this, to avoid frustration. See countless bug reports in bugzilla.
Fixes bug #622685.
When an error message is received on the bus, mark the bin as being in the error
state and unlock all current _get_state() calls with an error.
Fixes#505770
When an element is removed from a bin because it caused a state change error,
don't unref the child twice.
Add some more debug info.
Add a unit test for this error.
Fixes#615756
Right now deleyed set would only try for first set of children. We need to keep
trying to support arbitrary deep hierarchies (like in playbin2 with auto*sinks).
Also GstBin would need to actualy emit the child-added/removed signal as it
implements the iface. Fixes#613215.
Never skip the state change to playing, even if the element is already in the
right state. We need this because we also distribute the base_time while doing
the state change and skipping this step would leave some elements without a new
base_time.
Fixes#600313
There's not much point in using GST_DEBUG_FUNCPTR with GObject
virtual functions such as get_property, set_propery, finalize and
dispose, since they'll never be used by anyone anyway. Saves a
few bytes and possibly a tenth of a polar bear.
Post the structure change messages on the sinkpads of the elements. This allows
us to catch unlinked pads earlier without ending up with inconsistent element
degrees.
When we detect a pad unlink in progress, we will not be updating the degree of
the parent element. This can cause false loop detected warnings because the
degree counter is invalid. Handle this case by marking the iterator as 'dirty'
when we detect a pad unlink and avoid emiting the warning in this case. We have
to continue our state change as good as we can, we will eventually resync when
the pad unlink completed.
When an element is added to the bin, only set the index if we have a
cached index, rather than setting a NULL index on elements that might
have a default index object of their own.
Cache the last index that was set with _set_index() and return this in the
_get_index() call.
Set the cached index on newly added elements.
Fixes#566881
When we get an ASYNC_DONE message when a state change was busy, set the
pending_async_done flag so that after the state change completes, the bin can
check if all async elements are finished. Don't only do this for the bin itself
but for all elements.
This fixes some bins in bins that simulate async state changes by posting ASYNC
messages (such as sdpparse in uridecodebin/playbin2).
Bin collects segment-start messages and segent-done messages. it posts a
segment-done message to its parent, once it has received a segment-done for
each segment-start. Imho it should also send a segment-start if it receives the
first segment start and if parent is !=NULL. This is needed for bins in bins,
so that also higher order bins can group segment-starts and segment-dones.
Right now higher order bins will post a segment-done for each segment-done
received.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_set_index_func), (gst_bin_set_clock_func),
(gst_bin_change_state_func):
Use an iterator to set the clock and the index so that we can release
the object lock appropriately. Fixes#566393.
Original commit message from CVS:
* gst/gstbin.c: (_gst_boolean_accumulator), (gst_bin_class_init),
(gst_bin_recalculate_latency), (gst_bin_do_latency_func),
(gst_bin_change_state_func):
* gst/gstbin.h:
Add do-latency signal with the old default fallback implementation. This
allows for custom latency calculations for when the default is not
sufficient.
API: GstBin::do-latency signal.
Original commit message from CVS:
* docs/gst/gstreamer-sections.txt:
* gst/gstbin.c: (gst_bin_recalculate_latency),
(gst_bin_change_state_func):
* gst/gstbin.h:
Add method to recalculate and redistribute the latency on a bin.
API: gst_bin_recalculate_latency().
Original commit message from CVS:
* gst/gstbin.c: (bin_handle_async_start),
(gst_bin_handle_message_func), (gst_bin_query):
* libs/gst/base/gstbasesink.c: (gst_base_sink_render_object),
(gst_base_sink_event), (gst_base_sink_change_state):
* libs/gst/base/gstbasesrc.c: (gst_base_src_perform_seek),
(gst_base_src_loop), (gst_base_src_change_state):
Copy seqnums from events to messages so that they can all be related
back to eachother.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_remove_func), (update_degree),
(gst_bin_handle_message_func):
The message src can be NULL, don't try to print the object names in that
case.
* libs/gst/base/gstbasesink.c: (gst_base_sink_pad_activate):
Add some more debug info.
* tests/check/pipelines/simple-launch-lines.c: (run_pipeline),
(GST_START_TEST):
Add some debug.
Fix the test, pull based sinks go ASYNC to PAUSED, just like other
scheduling modes.
Original commit message from CVS:
Base on Patch by: Olivier Crete <tester at tester dot ca>
* gst/gstbin.c: (gst_bin_init), (gst_bin_add_func),
(gst_bin_remove_func), (update_degree),
(gst_bin_sort_iterator_new), (gst_bin_handle_message_func):
Keep track of pads that are being linked/unlinked and resync the state
changes.
* gst/gstpad.c: (gst_pad_get_direction),
(gst_pad_set_chain_function), (gst_pad_set_getrange_function),
(gst_pad_set_checkgetrange_function), (gst_pad_unlink),
(gst_pad_link_prepare), (gst_pad_link),
(gst_pad_event_default_dispatch), (gst_pad_chain), (gst_pad_push),
(gst_pad_check_pull_range), (gst_pad_get_range),
(gst_pad_pull_range):
Some code cleanups, use macros to check pad direction.
Don't need to take the lock on the pad direction.
Post structure change when pads are linked/unlinked.
Change some checks into _return_if_fail().
* tests/check/gst/gstbin.c:
(test_link_structure_change_state_changed_sync_cb),
(GST_START_TEST), (gst_bin_suite):
Add testcase for pad link/unlinke resync during a state change.
Fixes#510354.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_element_set_state):
Take the (recursive) state lock between getting the locked state of an
element and changing the element state. This allows the application to
lock an element's state and then change its state without races.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_element_set_state):
When an element is in the locked state we still want to update the
base_time of the element.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_add_func), (gst_bin_remove_func),
(gst_bin_change_state_func), (bin_handle_async_done),
(gst_bin_handle_message_func):
Fix race for bins that simulate ASYNC state changes by inserting
ASYNC_START and ASYNC_DONE messages in their bus. We need to check for
pending ASYNC messages even when the bin does not have ASYNC children.
We note detect this behaviour because we will receive an ASYNC message
that is originating from the bin itself.
Fixes races with decodebin2 state changes.
* tests/check/gst/gstbin.c: (GST_START_TEST):
Add some more debug.
Original commit message from CVS:
* configure.ac:
* docs/gst/gstreamer-sections.txt:
* docs/gst/gstreamer.types:
* docs/gst/gstreamer.types.in:
* gst/Makefile.am:
* gst/gst.c:
* gst/gstbin.c: (gst_bin_class_init), (gst_bin_set_index_func):
* gst/gstconfig.h.in:
* gst/gstelement.c: (gst_element_get_index):
* gst/gstregistrybinary.c: (gst_registry_binary_save_feature),
(gst_registry_binary_load_feature),
(gst_registry_binary_read_cache):
* gst/gstregistryxml.c: (load_feature),
(gst_registry_xml_read_cache), (gst_registry_xml_save_feature):
* plugins/Makefile.am:
* tools/gst-indent:
* tools/gst-inspect.c: (print_index_info), (print_element_list),
(print_plugin_features), (print_element_features):
* tools/gst-xmlinspect.c: (print_event_masks),
(print_element_info):
* win32/common/gstconfig.h:
Remove GST_DISABLE_(ENUMTYPES|INDEX|URI) everywhere.
Disabling the indexers and URI handler code will only reduce the
required amount of memory by a very small amount but on the other hand
requires much more maintaince work. Apart from that many places of
code are broken when disabling them.
Disabling the enum types doesn't reduce the required amount of memory
by more than a few bytes and makes it hard to fix bugs like #539772,
i.e. use the enums as GObject properties.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_class_init), (gst_bin_init),
(gst_bin_dispose):
Use the GLib stuff to create a private structure.
Add some locking around some dispose methods to make them a little
safer, see #529723. Patch by: Antoine Tremblay <hexa00 at gmail dot com>
Original commit message from CVS:
* docs/gst/gstreamer-sections.txt:
Move GParamSepc macros to standart section.
* gst/gstbin.c:
Dn't document _get_type - its in private section in docs anyway and
this doc-blob was incomplete.
* gst/gstclock.h:
Fix wrong symbol names in docs.
* gst/gstmacros.h:
Add once doc sentence.
* tests/check/gst/.cvsignore:
Ignore more.
Original commit message from CVS:
* docs/random/ensonic/profiling.txt:
More ideas.
* gst/gstbin.c:
Fix typo and give better log output.
* gst/gstdebugutils.c:
* gst/gstdebugutils.h:
More ideas, make graphs a bit smaller and fix param name in macro.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_class_init), (gst_bin_init),
(gst_bin_add_func), (gst_bin_remove_func),
(gst_bin_change_state_func), (gst_bin_continue_func):
Change email, cleanups add some more debug and comments.
Also set bus and clock on new elements when the pipeline was in error.
Original commit message from CVS:
* gst/gstbin.c: (bin_handle_async_done):
Correctly set the next state if all of our async children commited their
state. This makes sure we can actually cancel the state change in
progress. Fixes a regression in Rhythmbox when seeking.
Original commit message from CVS:
* gst/gstbin.c: (remove_from_queue):
Work around a problem with pipelines containing (semi)loops until a
proper, more complicated solution is ready. See #475455.
Original commit message from CVS:
* gst/gstbin.c: (bin_query_latency_fold), (gst_bin_query):
A latency query fails when one of the sinks fail.
* gst/gstelement.c: (gst_element_set_base_time):
Improve debugging.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_class_init), (clear_queue),
(do_bin_latency), (gst_bin_change_state_func):
* gst/gstpipeline.c: (gst_pipeline_change_state):
Move latency query from GstPipeline to GstBin so that we can also
use it when async-handling is enabled on bins.
Original commit message from CVS:
* gst/gstbin.c: (is_eos), (gst_bin_add_func),
(bin_handle_async_start), (gst_bin_handle_message_func):
Improve debugging.
When adding elements, insert messages into the bus of the newly added
element and make sure the element is the source of the message. This
allows the parent bin to intercept the message and do the
right thing. It also avoids us posting ASYNC_START and CLOCK_PROVIDE
messages to the app (which is not allowed).
Update some docs.
* tests/check/gst/gstghostpad.c: (GST_START_TEST):
Fix testsuite so that is does not work around messages that should not
have been posted in the first place.
Original commit message from CVS:
* gst/gstbin.c: (add_to_queue), (remove_from_queue), (clear_queue),
(update_degree), (gst_bin_sort_iterator_next):
Fix annoying bug in the sorted iterator where a sink that is not really
a sink (when it has downstream links) screwed up the iterator.
* tests/check/gst/gstbin.c: (GST_START_TEST), (gst_bin_suite):
Unit test to verify the fix.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_add_func), (gst_bin_element_set_state),
(bin_handle_async_start), (gst_bin_handle_message_func):
Move ASYNC_START message posting to where it belongs, similar to
async_done.
Don't post ASYNC_START when we are in error.
Post ASYNC_START when we added an async element to a bin.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_element_set_state):
Always change the state of a NO_PREROLL element even if it has ASYNC
elements inside (in case of a bin).
* tests/check/generic/sinks.c: (GST_START_TEST), (gst_sinks_suite):
Unit test for this case.
Original commit message from CVS:
* docs/design/part-TODO.txt:
Add some more TODO items
* gst/gstbin.c: (find_message), (gst_bin_change_state_func):
Improve debugging.
* gst/gstcaps.c: (gst_caps_intersect):
Optimize trivial intersection case between identical caps pointers.
* gst/gstelement.c: (gst_element_continue_state),
(gst_element_set_state_func):
* gst/gstpad.c:
Fix spelling and grammar mistakes.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_change_state_func),
(bin_handle_async_done), (gst_bin_handle_message_func):
Don't forget about the fact that some element went ASYNC even after a
resync. This makes us post the ASYNC_DONE message correctly.
Fixes#462558.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_continue_func):
Check that the state_cookie hasn't changed since the continue_func
was scheduled. Avoids problems where the state changes back to
something it shouldn't be because it was changed in the meantime.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_add_func), (gst_bin_remove_func),
(gst_bin_change_state_func), (bin_push_state_continue),
(bin_handle_async_start), (bin_handle_async_done),
(gst_bin_handle_message_func):
Move the common code for posting state-change messages into
one function.
Broadcast the state signal after we posted the messages.
Mark the bin as busy when it's doing a state-change.
Make sure async-start/done messages don't interfere with the bin's
state when it's busy.
After the state change, let the bin check which elements completed the
state change while it was busy so that it can update its state.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_init), (gst_bin_add_func),
(gst_bin_remove_func), (gst_bin_get_state_func),
(gst_bin_element_set_state), (gst_bin_continue_func),
(bin_push_state_continue), (bin_handle_async_start),
(bin_handle_async_done), (gst_bin_handle_message_func):
* gst/gstbin.h:
Immediatly commit the toplevel bin state when receiving an async-done
message. This enables us to avoid spawning a thread to commit the state
in some common cases and it also avoids some races.
Avoid spawning a state thread when adding/removing async elements to a
toplevel bin. Instead we immediatly update the bin state.
Get rid of iterating all the children when getting the state in the bin
because it is now always up-to-date.
Fix bug where locked elements would always return _SUCCESS even it they
returned NO_PREROLL before being locked.
Fix the order of the state_change, async-start/done messages that was
sometimes incorrect.
Mark the state_dirty field as deprecated, we don't need it anymore as we
are always up-to-date.
* gst/gstelement.c: (gst_element_get_state_func),
(gst_element_continue_state):
Small debug inprovements.
Return the previous element state return when nothing is pending instead
of blindly returning SUCCESS.
* tests/check/generic/sinks.c: (GST_START_TEST), (pad_blocked_cb),
(gst_sinks_suite):
Add a whole bunch of new testcases.
Original commit message from CVS:
* gst/gstbin.c: (find_message), (bin_replace_message), (is_eos),
(gst_bin_remove_func), (gst_bin_element_set_state),
(bin_handle_async_start), (bin_handle_async_done),
(gst_bin_handle_message_func):
Add helper function to find messages.
Generate the async-done messages together with the state change
messages.
Small cleanups in handling toplevel bins.
Original commit message from CVS:
* gst/gstbin.c: (bin_handle_async_done):
Make sure that the child bin stops after completing the async state
change so that the parent can continue the state change to PLAYING.
Fixes#441159.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_class_init), (gst_bin_get_state_func),
(gst_bin_continue_func):
Enable latency for next release.
Restore STATE_LOCK around recalc_state that was left out during the
rewrite and could result in racy behaviour when _get_state and
recalc_state are run concurrently. See #440463.
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_init), (gst_bin_dispose),
(gst_bin_set_property), (gst_bin_get_property),
(gst_bin_remove_func), (gst_bin_handle_message_func):
* gst/gstbin.h:
'private' is a c++ keyword, let's not use that in header files,
otherwise c++ compilers will throw a tantrum.
Original commit message from CVS:
* docs/gst/gstreamer-sections.txt:
* gst/gstbin.c: (gst_bin_class_init), (gst_bin_init),
(gst_bin_dispose), (gst_bin_set_property), (gst_bin_get_property),
(gst_bin_remove_func), (gst_bin_handle_message_func):
* gst/gstbin.h:
Add a property for bins that handle the state change of their childs.
Fixes#435880
Original commit message from CVS:
* gst/gstbin.c: (gst_bin_get_type), (gst_bin_element_set_state):
Prepare for release where we warn against possible app breakage in the
case of live pipelines along with an env var to enable/disable live
preroll mode (GST_COMPAT=[no-]live-preroll).
Original commit message from CVS:
* docs/gst/gstreamer-sections.txt:
Add new element field and method.
* gst/gstbin.c: (gst_bin_class_init), (gst_bin_init),
(bin_remove_messages), (gst_bin_add_func), (gst_bin_remove_func),
(gst_bin_recalc_state), (gst_bin_get_state_func),
(gst_bin_element_set_state), (gst_bin_change_state_func),
(gst_bin_continue_func), (bin_bus_handler),
(bin_push_state_continue), (bin_handle_async_start),
(bin_handle_async_done), (gst_bin_handle_message_func):
Make async state changes a bit smarter by using new ASYNC_START and
ASYNC_DONE messages. This reduces the number of times we run the state
recalculation thread.
Don't change state of element with a pending ASYNC_START message.
Deprecate STATE_DIRTY messages.
* gst/gstelement.c: (gst_element_init), (gst_element_send_event),
(gst_element_get_state_func), (gst_element_continue_state),
(gst_element_lost_state), (gst_element_set_state_func),
(gst_element_change_state):
* gst/gstelement.h:
Keep the state that was last set by the app in a new element field.
Don't allow state changes when handling an element event.
Post ASYNC_START and ASYNC_DONE messages.
Change lost_state so that we go to PAUSED and wait for the parent to set
us to PLAYING again (so latency calculation can be performed)
Export gst_element_change_state() method so that subclasses can use it.
API: gst_element_change_state()
API: GST_STATE_TARGET
* gst/gstpipeline.c: (gst_pipeline_class_init),
(reset_stream_time), (gst_pipeline_change_state),
(gst_pipeline_handle_message), (gst_pipeline_set_new_stream_time):
Using the new ASYNC_START message we can reset the base_time when
needed. This can then be used to implement base_time redistribution in
flushing seeks so that we can remove the explicit seek handling.
Perform latency query and configuration when going to PLAYING.
* libs/gst/base/gstbasesink.c: (gst_base_sink_commit_state),
(gst_base_sink_query), (gst_base_sink_change_state):
Post new ASYNC_START/ASYNC_DONE messages.
* tests/check/generic/sinks.c: (GST_START_TEST):
Fix test because the bin will not set the async element to PLAYING right
away.
* tests/check/gst/gstbin.c: (pop_async_done), (GST_START_TEST):
Make the message check a little stronger.
Handle ASYNC messages.
* tests/check/pipelines/cleanup.c: (GST_START_TEST):
* tests/check/pipelines/simple-launch-lines.c: (GST_START_TEST):
Expect ASYNC_DONE messages.
Original commit message from CVS:
* gst/gstbin.c:(gst_bin_add):
Use GST_STR_NULL to prevent NULL pointer to be passed to GST_CAT_DEBUG.
* win32/common/libgstbase.def:
* win32/common/libgstreamer.def:
Add new exported functions.
