Add the pad mode to the activate function so that we can reuse the same function
for all activation modes. This makes the core logic smaller and allows for some
elements to make their activation code easier. It would allow us to add more
scheduling modes later without having to add more activate functions.
Remove the getcaps function on the pad and use the CAPS query for
the same effect.
Add PROXY_CAPS to the pad flags. This instructs the default caps event and query
handlers to pass on the CAPS related queries and events. This simplifies a lot
of elements that passtrough caps negotiation.
Make two utility functions to proxy caps queries and aggregate the result. Needs
to use the pad forward function instead later.
Make the _query_peer_ utility functions use the gst_pad_peer_query() function to
make sure the probes are emited properly.
This reverts commit cf4fbc005c.
This change did not improve the situation for bindings because
queries are usually created, then directly passed to a function
and not stored elsewhere, and the writability problem with
miniobjects usually happens with buffers or caps instead.
Improve GstSegment, rename some fields. The idea is to have the GstSegment
structure represent the timing structure of the buffers as they are generated by
the source or demuxer element.
gst_segment_set_seek() -> gst_segment_do_seek()
Rename the NEWSEGMENT event to SEGMENT.
Make parsing of the SEGMENT event into a GstSegment structure.
Pass a GstSegment structure when making a new SEGMENT event. This allows us to
pass the timing info directly to the next element. No accumulation is needed in
the receiving element, all the info is inside the element.
Remove gst_segment_set_newsegment(): This function as used to accumulate
segments received from upstream, which is now not needed anymore because the
segment event contains the complete timing information.
This reverts commit 9ef1346b1f.
Way to much for one commit and I'm not sure we want to get rid of the pad caps
just like that. It's nice to have the buffer and its type in onw nice bundle
without having to drag the complete context with it.
Remove pad_alloc and all references. This can now be done more efficiently and
more flexible with the ALLOCATION query and the bufferpool objects. There is no
reverse negotiation yet but that will be done with an event later.
1) We need to lock and get a strong ref to the parent, if still there.
2) If it has gone away, we need to handle that gracefully.
This is necessary in order to safely modify a running pipeline. Has been
observed when a streaming thread is doing a buffer_alloc() while an
application thread sends an event on a pad further downstream, and from
within a pad probe (holding STREAM_LOCK) carries out the pipeline plumbing
while the streaming thread has its buffer_alloc() in progress.
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.
Split gst_queue_locked_enqueue() into variant for buffer and event to get rid of
the if() and make the code more readable (constant boolean parameters are never
nice). Removes the if (item) checks as we dereference the pointer before anyway.
Also apply the same idea of reusing the previous knowledge in
gst_queue_locked_dequeue to remove more type checks.
We know whether we have a buffer or an event, use that instead of going
trough the expensive GLib typecheck.
The overall instruction fetch reduction introduced by this commit and the
2 previous commits:
* receiving a buffer (_chain) by 20%
* popping a buffer (_loop) by 14%
Numbers acquired through callgrind passing 100000 buffers through queue.
If downstream returns error and upstream has already delivered
everything (including EOS) and will no longer be around to find
out that we paused (and why), post error message. Fixes#589991.
The compiler suggests to add some () to indicate if the && or the || takes
priority, so reflow code a bit so we don't have to add yet another layer
of (). Hopefully this was the intended meaning of the code.
When min-threshold is set on a queue, it is possible that one of
the minima remains unsatisfied while one of the maxima is already
reached. Therefore, always consider the queue non-empty if it is full.
Fixes#585433.
Original commit message from CVS:
* plugins/elements/gstqueue.c:
Allow through queries when we don't know how
to adjust them (not TIME or BYTES), as otherwise it's
not possible to query the current position in order
to seek in other formats at all.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_init),
(gst_queue_acceptcaps):
Add and use a custom acceptcaps function instead of falling back to the
potentially less optimized default implementation.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_leak_downstream):
Since we're not called only from the chain function any longer,
we can't assume that there's always data in the queue, so move
the is_full check to the beginning of the loop (otherwise we'd
hit the assert when changing the limit properties while the
queue is empty or not running yet).
Also, only set a discont if items were actually removed from
the queue.
* tests/check/elements/queue.c: (test_leaky_downstream):
Test case for the above.
Original commit message from CVS:
Patch by: Jonas Holmberg <jonas dot holmberg at axis dot com>
* plugins/elements/gstqueue.c: (gst_queue_leak_downstream),
(gst_queue_chain), (queue_capacity_change),
(gst_queue_set_property):
When changing thr max capacity of a leaky queue, immediatly drop buffers
instead of waiting for a push on the sinkpad. Fixes#530637.
Original commit message from CVS:
Patch by: Olivier Crete <tester at tester dot ca>
* plugins/elements/gstqueue.c: (gst_queue_handle_src_query):
Add the min-threshold to the min latency if possible. Fixes#529148.
Original commit message from CVS:
* plugins/elements/gstqueue.c:
* plugins/elements/gstqueue.h:
When dropping buffers in leaky modes, mark next buffers we sent as
DISCONT.
Original commit message from CVS:
* gst/gstpreset.c:
Add some comments and TODOs.
* gst/gstpreset.h:
Add padding for future changes.
* plugins/elements/gstqueue.c:
Implement the iface.
Original commit message from CVS:
* plugins/elements/gstmultiqueue.c: (gst_single_queue_push_one),
(gst_single_queue_new):
* plugins/elements/gstqueue.c: (gst_queue_init),
(gst_queue_push_one):
Fix queue negotiation. If acceptcaps unconditionally returns TRUE,
upstream is tricked into thinking it can suggest a format downstream
while downstream does not support that format. The real problem is that
core calls acceptcaps when pushing a buffer with new caps, for which we
do a little workaround by setting the caps on the srcpad ourselves
before pushing the buffer (until this is figured out). Fixes#486758.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_locked_enqueue),
(gst_queue_locked_dequeue):
Comment the segment-related code... in the PROPER function.
See #482147 and my commit from yesterday.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_locked_dequeue):
Disable code that's breaking the current-time-level reporting.
See #482147
Original commit message from CVS:
Patch by: Felipe Contreras <felipe dot contreras at gmail dot com>
* plugins/elements/gstqueue.c: (gst_queue_handle_sink_event):
First patch of code cleanups, use the macros and right arguments in the
macros to signal and lock the queue. See #480858.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (apply_buffer),
(gst_queue_locked_enqueue), (gst_queue_locked_dequeue):
Measure queue level based on the diff between head and tail timestamps
even when pushing the first buffer.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_locked_enqueue),
(gst_queue_handle_sink_event), (gst_queue_chain),
(gst_queue_push_one), (gst_queue_handle_src_query),
(gst_queue_sink_activate_push), (gst_queue_src_activate_push):
* plugins/elements/gstqueue.h:
When downstream returns UNEXPECTED from pushing a buffer, don't try to
push more buffers but allow pushing of EOS and NEWSEGMENT.
Add some more debug info here and there. Fixes#476514.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_handle_src_query):
Queue can latency to the pipeline up to the configured max size in time.
Report this fact in the latency query.
Original commit message from CVS:
* libs/gst/base/gstdataqueue.c: (gst_data_queue_cleanup),
(gst_data_queue_finalize), (gst_data_queue_locked_is_empty),
(gst_data_queue_set_flushing), (gst_data_queue_push),
(gst_data_queue_pop), (gst_data_queue_drop_head),
(gst_data_queue_limits_changed), (gst_data_queue_get_level):
* libs/gst/base/gstdataqueue.h:
Various cleanups.
Added methods to get the current levels and to inform the queue that the
'full' limits changed.
* plugins/elements/gstmultiqueue.c: (gst_multi_queue_init),
(gst_multi_queue_finalize), (gst_multi_queue_set_property),
(gst_single_queue_flush), (update_time_level), (apply_segment),
(apply_buffer), (gst_single_queue_push_one),
(gst_multi_queue_item_steal_object),
(gst_multi_queue_item_destroy), (gst_multi_queue_item_new),
(gst_multi_queue_loop), (gst_multi_queue_chain),
(gst_multi_queue_sink_activate_push), (gst_multi_queue_sink_event),
(gst_multi_queue_getcaps), (gst_multi_queue_src_activate_push),
(gst_multi_queue_src_query), (single_queue_overrun_cb),
(single_queue_underrun_cb), (single_queue_check_full),
(gst_single_queue_new):
Keep track of time in the queue by measuring the difference between
running_time on input and output. This gives more accurate results and
can compensate for segments correctly.
Make a queue by default only 5 buffers deep. We will now increase the
buffer size depending on the filledness of the other queues.
Factor out commong flush code.
Make sure we don't add additional refcounts to buffers when we can avoid
it.
Propagate GstFlowReturn differently.
Use GSlice for intermediate GstMultiQueueItems.
Keep track of EOS.
Resize queues on over and underruns based on filled level of other
queues.
When checking if the queue is filled, prefer to measure in time if we
can and fall back to bytes when no time is known.
* plugins/elements/gstqueue.c:
Fix return value.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_loop):
Hello, I am Mr Taymans' personal debugger. Today I will introduce a fix
for the subtle art of warning a potentially blocking thread that it
should check the source pad return value, and relay the information
upstream.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_init),
(gst_queue_finalize), (update_time_level), (apply_segment),
(apply_buffer), (gst_queue_locked_flush),
(gst_queue_locked_enqueue), (gst_queue_locked_dequeue),
(gst_queue_handle_sink_event), (gst_queue_chain),
(gst_queue_push_one), (gst_queue_loop):
* plugins/elements/gstqueue.h:
Refactor an cleanup queue a bit.
Do better time level calculations that also work when the srcpad is not
yet running.
Remove some unneeded debug lines.
* tests/check/elements/queue.c: (GST_START_TEST), (queue_suite):
Added testcase for time level measurement.
Try to make some stuff more racefree.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_class_init),
(update_time_level), (gst_queue_locked_flush),
(gst_queue_handle_sink_event), (gst_queue_chain),
(gst_queue_push_one), (gst_queue_loop):
* plugins/elements/gstqueue.h:
Be smarter when calculating the current amount of data in the queue by
measuring the difference between start and end timestamps (in running
time) inside the queue. Fixes#432876.
API: GstQueue::pushing to notify elements that we are pushing data again
since the running signal is rather broken for this purpose.
Original commit message from CVS:
* plugins/elements/gstqueue.c (_do_init, gst_queue_signals,
gst_queue_base_init, gst_queue_init):
use GST_BOILERPLATE
Original commit message from CVS:
* gst/gst.c:
Docs typo fix.
* plugins/elements/gstqueue.c: (gst_queue_class_init),
(gst_queue_init):
Fix incorrect documentation and flesh it out a bit more.
Set default values for the max properties on the GParamSpec as well,
so it shows up correctly in gst-inspect.
Original commit message from CVS:
Patch by: Sjoerd Simons <sjoerd at luon dot net>
* plugins/elements/gstqueue.c: (gst_queue_chain):
Recheck queue filledness after signalling the overrun when we're about
to leak downstream because we released the lock when emitting the signal
and the queue could be empty again. Fixes#352345.
Original commit message from CVS:
* gst/gstpad.c: (gst_pad_set_caps):
Short-circuit gst_pad_set_caps if setting the existing
caps pointer again, and avoid printing debug and
reffing/unreffing the caps.
* plugins/elements/gstqueue.c: (gst_queue_push_one):
There's actually no need to set the caps before pushing -
the acceptcaps method will handle it anyway.
Original commit message from CVS:
* plugins/elements/gstcapsfilter.c: (gst_capsfilter_prepare_buf):
The caps intersection algorithm can produce multiple copies of the
caps. Until that is fixed, we need to simplify the result to be
sure whether the allowed caps are fixed or not.
* plugins/elements/gstqueue.c: (gst_queue_init),
(gst_queue_bufferalloc), (gst_queue_acceptcaps),
(gst_queue_push_one):
Proxied buffer alloc should not set the caps on the source pad.
When pushing buffers, we always accept the caps change that triggers.
This prevents negotiation errors caused by caps changing mid-stream
and then being refused on our source pad (because upstream is now
refusing those caps).
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_push_one):
The queue is not responsible for pushing an EOS when receiving a fatal
flow error. It's up to the real element driving the pipeline to do that.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_push_one):
The queue was posting a non-needed GST_MESSAGE_ERROR when pushing a
buffer returned a fatal error. It should just send an EOS and stop
it's task.
Upstream elements will then properly receive the GST_FLOW_UNEXPECTED
when pushing buffers on the queue and will be able to handle the event.
Original commit message from CVS:
* plugins/elements/gstqueue.c: (gst_queue_chain), (gst_queue_loop),
(gst_queue_set_property):
Don't forget to signal the _chain or _loop function
when the queue size or thresholds change since that might
cause them to make progres again.
Original commit message from CVS:
Patch by Tommi Myöhänen <ext-tommi dot myohanen at nokia dot com>
* plugins/elements/gstqueue.c: (gst_queue_init),
(gst_queue_locked_flush), (gst_queue_handle_sink_event),
(gst_queue_set_property):
* plugins/elements/gstqueue.h:
In queue, when EOS is received, if minimum threshold > max_size -
current_level, there is chance that queue blocks forever in conditional item
del wait. This is because the queue is not emptied completely due to minimum
threshold.
Here is another approach. Instead of setting cur_levels to max in EOS, just
zero all minimum threshold levels. This should make sure that queue gives out
all data. When going to READY (stop) state, just reset the original minimum
threshold levels.
Fixes#336336.
