Tim-Philipp Müller [Thu, 11 Jan 2018 19:32:08 +0000 (19:32 +0000)]
tests: filesrc: more Windows fixes
Fix typo in newly-added windows uri test.
Tim-Philipp Müller [Thu, 11 Jan 2018 18:44:50 +0000 (18:44 +0000)]
utils: use g_get_monotonic_time() as fallback in gst_utils_get_timestamp()
This is a better fit given that the function docs say this
should (only) be used for interval measurements, but also
this seems to give much better granularity on Windows
systems, where before this change there would often be
10-20 lines of debug log with the same timestamp up front.
Tim-Philipp Müller [Thu, 11 Jan 2018 18:39:50 +0000 (18:39 +0000)]
tests: systemclock: scale stress test threads according to number of cpus
Scale the number of threads used in the stress tests according to
the number of cores/cpus. We want some contention, but we also
don't want too much contention, as some operating systems are
better at handling 100 threads running wild on a single core
than others.
Tim-Philipp Müller [Thu, 11 Jan 2018 17:10:45 +0000 (17:10 +0000)]
plugin: plugin_load() must return a ref even if it was loaded already
Fix refcounting issue when plugin was loaded already.
gst_plugin_load() is supposed to return a ref, so it
must always return a ref.
This also fixes the gstplugin unit test on windows where
fork is not available and where test_load_coreelements()
would unref a plugin ref it didn't get and then mess up
the internal registry plugin list state for the next test,
in case where the test registry does not exist yet.
Tim-Philipp Müller [Thu, 11 Jan 2018 14:56:42 +0000 (14:56 +0000)]
tests: filesrc: fix for windows
Location paths have backslashes on windows when converted from URI.
Tim-Philipp Müller [Thu, 11 Jan 2018 12:27:18 +0000 (12:27 +0000)]
libs: check: print stacktrace on unexpected criticals
Tim-Philipp Müller [Thu, 11 Jan 2018 12:02:47 +0000 (12:02 +0000)]
tests: abi: fix abi test on 64-bit Windows
Add header with structure sizes for 64-bit windows as well.
They're almost the same as on Linux, but it looks like things
like padding unions get aligned slightly differently so there
are a handful of differences:
sizeof(GstGhostPad) is 528, expected 536
sizeof(GstPad) is 512, expected 520
sizeof(GstPadProbeInfo) is 64, expected 72
sizeof(GstProxyPad) is 520, expected 528
Tim-Philipp Müller [Thu, 11 Jan 2018 11:38:53 +0000 (11:38 +0000)]
tests: info: fix post init cat reg test on windows and with CK_FORK=no
The test checks that categories not covered by the pattern in the
GST_DEBUG string have debug level GST_LEVEL_DEFAULT set, but previous
tests mess with the default threshold, which made this test fail on
Windows or when run with CK_FORK=no. Fix this by resetting everything
at the beginning, and then also do a sanity check afterwards.
Tim-Philipp Müller [Thu, 11 Jan 2018 11:36:53 +0000 (11:36 +0000)]
info: reset default threshold to LEVEL_DEFAULT not 0
in set_threshold_from_string().
Håvard Graff [Wed, 8 Nov 2017 19:05:03 +0000 (20:05 +0100)]
meson: make check and tests build on Windows with msvc
Edward Hervey [Tue, 2 Jan 2018 09:02:45 +0000 (10:02 +0100)]
gstharness: Remove double free
Sebastian Dröge [Tue, 26 Dec 2017 17:08:31 +0000 (18:08 +0100)]
streams: GstStreamType/GstStream are available since 1.10
Annotate them as such.
Tim-Philipp Müller [Tue, 26 Dec 2017 12:46:20 +0000 (13:46 +0100)]
meson: skip translations if gettext is not available
Stefan Sauer [Tue, 26 Dec 2017 11:51:22 +0000 (12:51 +0100)]
aggregator: remove DEBUG_FUNCPTR
The new gst_element_do_foreach_pad() does not print the functions anymore.
Stefan Sauer [Tue, 26 Dec 2017 11:17:53 +0000 (12:17 +0100)]
inspect: add comment for how to improve tracer support
Stefan Sauer [Tue, 26 Dec 2017 10:29:39 +0000 (11:29 +0100)]
segment: add a FIXME-2.0 for the format parameters
Capture the somewhat not ordinary use of the extra format parameter in a
comment.
See https://bugzilla.gnome.org/show_bug.cgi?id=788979
Tim-Philipp Müller [Sun, 24 Dec 2017 15:21:38 +0000 (16:21 +0100)]
Skip gst_deinit() if gstreamer was not initialized properly
Can happen if an error occurs during option parsing, for example.
https://bugzilla.gnome.org/show_bug.cgi?id=781914
Jan Alexander Steffens (heftig) [Sat, 23 Dec 2017 22:43:33 +0000 (23:43 +0100)]
tests: multiqueue: Replace large test macro with function
Just a bit of cleanup.
https://bugzilla.gnome.org/show_bug.cgi?id=756867
Jan Alexander Steffens (heftig) [Fri, 15 Dec 2017 08:43:40 +0000 (09:43 +0100)]
tests: multiqueue: Check we get CREATE+ENTER stream-statuses when adding pads
https://bugzilla.gnome.org/show_bug.cgi?id=756867
Jan Alexander Steffens (heftig) [Fri, 15 Dec 2017 08:14:57 +0000 (09:14 +0100)]
multiqueue: Don't start new pads until parented
Start task on new source pads added at runtime after they
have been added to the element, not during activation.
This ensures the pads can post their CREATE stream-status
messages and the application can set thread priorities.
https://bugzilla.gnome.org/show_bug.cgi?id=756867
Jan Alexander Steffens (heftig) [Fri, 15 Dec 2017 08:14:07 +0000 (09:14 +0100)]
multiqueue: Split task handling from gst_single_queue_flush
https://bugzilla.gnome.org/show_bug.cgi?id=756867
Jan Alexander Steffens (heftig) [Sat, 23 Dec 2017 22:25:58 +0000 (23:25 +0100)]
aggregator: Avoid a maybe-uninitialized warning
Arch Linux x86_64, gcc 7.2.1-2, -Og -g3
Sebastian Dröge [Thu, 21 Dec 2017 11:47:52 +0000 (13:47 +0200)]
downloadbuffer: Don't hold the mutex while posint the download-complete message
Something might handle it from a sync message handler and call back into
downloadbuffer, causing a deadlock.
Sebastian Dröge [Wed, 20 Dec 2017 16:56:23 +0000 (18:56 +0200)]
systemclock: set_default() clock parameter can be NULL
Sebastian Dröge [Wed, 20 Dec 2017 16:11:48 +0000 (18:11 +0200)]
element: Annotate set_clock() clock parameter with allow-none
Sebastian Dröge [Wed, 20 Dec 2017 16:09:28 +0000 (18:09 +0200)]
element: Annotate set_bus() bus parameter as allow-none
It's possible to replace the bus with NULL/None
Tim-Philipp Müller [Sun, 10 Dec 2017 22:50:05 +0000 (22:50 +0000)]
gst-uninstalled: update for gl lib move from bad to base
Nicolas Dufresne [Tue, 12 Dec 2017 01:58:16 +0000 (20:58 -0500)]
basetransform: Allow going passthrough inside decide_allocation
Sub-class may want to decide to go passthrough/in-place by inspecting
the support meta APIs. This patch duplicates the check for this mode,
so we still don't do uneeded allocation query while we allow sub-classes
to switch the behaviour during it's own decide_allocation call.
Notice that such sub-class need to reset the class to non-passthrough in
set_caps() in order for decide_allocation to be called again. This is
needed otherwise we'd be doing an allocation query in element in which
it make no sense (notably capsfilter).
https://bugzilla.gnome.org/show_bug.cgi?id=791453
Sebastian Dröge [Sun, 17 Dec 2017 12:18:38 +0000 (14:18 +0200)]
plugin: Annotate add_dependency() arguments as NULL-terminated arrays
Umang Jain [Wed, 13 Dec 2017 18:33:04 +0000 (00:03 +0530)]
docs: GstBus: Provide more information for ref/unref during bus watch.
https://bugzilla.gnome.org/show_bug.cgi?id=791588
Matthew Waters [Thu, 14 Dec 2017 05:05:00 +0000 (16:05 +1100)]
update win32 defs for tracer API addition
Matthew Waters [Thu, 14 Dec 2017 03:48:47 +0000 (14:48 +1100)]
Automatic update of common submodule
From e8c7a71 to 3fa2c9e
Matthew Waters [Tue, 5 Dec 2017 10:36:34 +0000 (21:36 +1100)]
docs: include tracers in the documentation
Requires exposing the tracer GType from the GstTracerFactory in order
to link the plugin with the tracer in the documentation.
https://bugzilla.gnome.org/show_bug.cgi?id=791253
Matthew Waters [Tue, 5 Dec 2017 09:56:09 +0000 (20:56 +1100)]
check/harness: fix transfer annotations on buffer passing functions
Tim-Philipp Müller [Wed, 30 Aug 2017 12:03:28 +0000 (13:03 +0100)]
basesrc: add buffer list support
Add a gst_base_src_submit_buffer_list() function that allows subclasses
to produce a bufferlist containing multiple buffers in the ::create()
function. The buffers in the buffer list will then also be pushed out
in one go as a GstBufferList. This can reduce push overhead
significantly for sources with packetised inputs (such as udpsrc)
in high-throughput scenarios.
The _submit_buffer_list() approach was chosen because it is fairly
straight-forward, backwards-compatible, bindings-friendly (as opposed
to e.g. making the create function return a mini object instead),
and it allows the subclass maximum control: the subclass can decide
dynamically at runtime whether to return a list or a single buffer
(which would be messier if we added a create_list virtual method).
https://bugzilla.gnome.org/show_bug.cgi?id=750241
Tim-Philipp Müller [Thu, 31 Aug 2017 00:18:28 +0000 (01:18 +0100)]
basesrc: minor code readability improvement
Tim-Philipp Müller [Thu, 7 Dec 2017 12:05:23 +0000 (12:05 +0000)]
docs: Fix a few gtk-doc warnings
Broken links mostly.
Tim-Philipp Müller [Wed, 6 Dec 2017 20:58:42 +0000 (20:58 +0000)]
tests: aggregator: fix caps leak in unit test
Edward Hervey [Wed, 6 Dec 2017 16:07:29 +0000 (17:07 +0100)]
gstpad: Handle GST_PAD_PROBE_HANDLED on sticky event push
When actually pushing an event, if we get GST_FLOW_CUSTOM_SUCCESS_1
(which is the conversion of GST_PAD_PROBE_HANDLED return value),
don't consider the stick event push as ignored, but as handled
Sebastian Dröge [Wed, 6 Dec 2017 11:40:46 +0000 (13:40 +0200)]
event/query/message: Annotate get_structure() return value as nullable
Sebastian Dröge [Wed, 6 Dec 2017 11:36:30 +0000 (13:36 +0200)]
query: Add an empty structure in writable_structure() if there is none yet
This is consistent with how it works for GstEvent already.
Edward Hervey [Tue, 5 Dec 2017 17:21:00 +0000 (18:21 +0100)]
docs: Misc addition/fixes
And also add the "Since" API sections for 1.12 and 1.14
Edward Hervey [Tue, 5 Dec 2017 17:20:34 +0000 (18:20 +0100)]
docs: Add documentation for GST_SEQNUM_INVALID
And link to it
Edward Hervey [Tue, 5 Dec 2017 16:28:55 +0000 (17:28 +0100)]
utils: Never return a group_id of 0, add GST_GROUP_ID_INVALID
Various plugins use special values (0 or G_MAXUINT32) as an
invalid/unset group_id, but nothing guarantees a groupid won't have
that value.
Instead define a value which group_id will never have and make
gst_group_id_next() always return a value different from that.
API: GST_GROUP_ID_INVALID
Tim-Philipp Müller [Tue, 5 Dec 2017 16:42:57 +0000 (16:42 +0000)]
harness: make bindings use the GBytes variant for _take_all_data()
Havard Graff [Wed, 23 Nov 2016 12:12:36 +0000 (13:12 +0100)]
harness: use new take_all_data() function in _dump_to_file().
Tim-Philipp Müller [Tue, 5 Dec 2017 15:28:43 +0000 (15:28 +0000)]
harness: add gst_harness_take_all_data() + _take_all_data_as_{bytes,buffer}()
Convenience function to just grab all pending data
from the harness, e.g. if we just want to check if
it matches what we expect and we don't care about
the chunking or buffer metadata.
Based on patch by: Havard Graff <havard.graff@gmail.com>
Tim-Philipp Müller [Tue, 5 Dec 2017 15:16:36 +0000 (15:16 +0000)]
buffer: document that _extract_dup() will return NULL for 0-sized buf
And make it explicit, and don't call _extract() on NULL data buffer.
Nirbheek Chauhan [Tue, 5 Dec 2017 06:57:18 +0000 (12:27 +0530)]
meson: Use array syntax instead of .get() in tests
Nirbheek Chauhan [Thu, 23 Nov 2017 21:09:43 +0000 (02:39 +0530)]
meson: Use new find_program fallback syntax
We use this syntax in libs/gst/helpers/meson.build already.
Sebastian Dröge [Mon, 4 Dec 2017 16:08:39 +0000 (18:08 +0200)]
gst: gst_element_remove_pad() is transfer none for the pad
While the refcount of the pad is decreased, it's the refcount that is
owned by the parent (i.e. the element) and not the one passed in by the
caller.
Fixes a memory leak in bindings.
Tim-Philipp Müller [Mon, 4 Dec 2017 11:24:47 +0000 (11:24 +0000)]
aggregator: add finish_buffer() vfunc
So subclasses can override the finish behaviour
and/or decorate or modify buffers before they
get pushed out.
https://bugzilla.gnome.org/show_bug.cgi?id=760981
Tim-Philipp Müller [Mon, 4 Dec 2017 12:29:05 +0000 (12:29 +0000)]
aggregator: disable tag merging and forwarding for now
Subclasses should handle this for now.
Jan Alexander Steffens (heftig) [Mon, 6 Nov 2017 19:23:12 +0000 (20:23 +0100)]
devicemonitor: Avoid maybe-uninitialized compiler warning
On Arch Linux x86_64, gcc 7.2.0-3, -Og -g3:
gstdevicemonitor.c: In function ‘bus_sync_message’:
gstdevicemonitor.c:276:8: error: ‘matches’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
This commit also simplifies the code a bit.
https://bugzilla.gnome.org/show_bug.cgi?id=789983
Sebastian Dröge [Sun, 3 Dec 2017 12:48:22 +0000 (14:48 +0200)]
gst: Annotate various strings as type filename if they represent a path/filename
Tim-Philipp Müller [Sat, 2 Dec 2017 15:44:48 +0000 (15:44 +0000)]
aggregator: hook up to docs
Tim-Philipp Müller [Sat, 2 Dec 2017 15:24:22 +0000 (15:24 +0000)]
aggregator: hook up to build system
https://bugzilla.gnome.org/show_bug.cgi?id=739010
Tim-Philipp Müller [Sat, 2 Dec 2017 15:12:25 +0000 (15:12 +0000)]
Move GstAggregator from -bad to core
Merge branch 'aggregator-move'
https://bugzilla.gnome.org/show_bug.cgi?id=739010
Mathieu Duponchelle [Mon, 6 Nov 2017 20:07:51 +0000 (21:07 +0100)]
aggregator: Remove klass->sinkpads_type
This posed problems for the python bindings (and possibly others).
Instead, subclasses now use add_pad_template_with_gtype.
https://bugzilla.gnome.org/show_bug.cgi?id=789986
Tim-Philipp Müller [Thu, 2 Nov 2017 18:32:55 +0000 (18:32 +0000)]
aggregator: add doc blurb for gst_aggregator_pad_is_eos()
Tim-Philipp Müller [Thu, 2 Nov 2017 16:05:12 +0000 (16:05 +0000)]
aggregator: also remove now-unused PadForeachFunc declaration
https://bugzilla.gnome.org/show_bug.cgi?id=785679
Olivier Crête [Wed, 2 Aug 2017 16:08:26 +0000 (12:08 -0400)]
aggregator: Remove pad iterator function
Use new gst_element_foreach_sink_pad() from core instead.
https://bugzilla.gnome.org/show_bug.cgi?id=785679
Tim-Philipp Müller [Thu, 2 Nov 2017 12:46:26 +0000 (12:46 +0000)]
aggregator: use new gst_element_foreach_sink_pad()
Instead of gst_aggregator_iterate_sinkpads() which will
soon be removed.
https://bugzilla.gnome.org/show_bug.cgi?id=785679
Stefan Sauer [Wed, 1 Nov 2017 14:18:08 +0000 (15:18 +0100)]
aggregator: add more comments
Stefan Sauer [Mon, 23 Oct 2017 09:52:38 +0000 (11:52 +0200)]
tests: comment and logging cleanups for audiomixer and aggregator
Remove some references to 'collectpads'. Logs pads through the object variants.
Add some more comments. Remove a left over comment.
Stefan Sauer [Sun, 22 Oct 2017 17:43:17 +0000 (19:43 +0200)]
aggregator: fix type for latency property (int64 -> GStClockTime)
The value is used as GstClockTiem in the code. Adapt the hack^H^H^H^Hcode
in live-adder.
Olivier Crête [Thu, 13 Jul 2017 23:03:19 +0000 (19:03 -0400)]
aggregator: Don't take flush lock from output thread
Instead just take it in the chain function.
https://bugzilla.gnome.org/show_bug.cgi?id=784911
Olivier Crête [Thu, 13 Jul 2017 22:38:34 +0000 (18:38 -0400)]
aggregator: Don't block if adding to the tail of the queue
If we're adding to the tail of the queue, it's because we're converting
a gap event, so don't block there it means we're calling from the output
thread.
https://bugzilla.gnome.org/show_bug.cgi?id=784911
Stefan Sauer [Tue, 17 Oct 2017 06:03:02 +0000 (08:03 +0200)]
aggregator: review code related to time level
Add a comment for when the state matters. Use a local var for priv in
update_time_level() to improve readability. Move the our_latency local
var below the query results checks.
Stefan Sauer [Tue, 17 Oct 2017 05:51:51 +0000 (07:51 +0200)]
aggregator: init latency values with 0 instead of FALSE
Stefan Sauer [Sun, 15 Oct 2017 18:46:09 +0000 (20:46 +0200)]
aggregator: code cleanup for event and query func
Only look up klass for non serialized events/queries. For events remove
superfluous assignment for the return value in the flushing case.
Stefan Sauer [Sun, 15 Oct 2017 15:46:45 +0000 (17:46 +0200)]
aggregator: simplify pad_event_func for FLUSH_STOP events
We want to skip serialization for FLUSH_STOP events (apparently). We can
simplify the code to add it to the top-level conditions. There was nothing
done in the first code path if the event was FLUSH_STOP.
Stefan Sauer [Sun, 15 Oct 2017 14:57:13 +0000 (16:57 +0200)]
aggregator: drop special casing for eos
Just queue it like any other serialized event. This way we don't need to
check if there still are buffers in the queue.
Validated with the tests and gst-launch-1.0 pipelines.
Stefan Sauer [Sun, 15 Oct 2017 14:51:21 +0000 (16:51 +0200)]
aggregator: add a doc-blob for the event_func
Stefan Sauer [Sun, 15 Oct 2017 14:48:21 +0000 (16:48 +0200)]
aggregator: rename a local variable
The variable tracks wheter the queue is not empty, but num_buffers==0. That
means we have events or queries to process. Rename accordingly.
Stefan Sauer [Sun, 15 Oct 2017 10:17:42 +0000 (12:17 +0200)]
aggregator: remove commented code
The SEGMENT_DONE event does not require any special treatment. This is
commented out in
6efc106a67.
Stefan Sauer [Sun, 15 Oct 2017 10:14:28 +0000 (12:14 +0200)]
aggregator: move the comment for the locks to the lock macros
Looks like some code was inserted afterwards.
Stefan Sauer [Sun, 15 Oct 2017 08:44:44 +0000 (10:44 +0200)]
aggregator: improve section docs
Mention how data ends up in the queues. Document the relation of the pad
functions and the class vmethods to get events and queries.
Stefan Sauer [Sat, 14 Oct 2017 16:18:44 +0000 (18:18 +0200)]
aggregator: add two more tests for a sequence of data
This verifies that we handle events and queries at the head of the queue and
then buffers.
Stefan Sauer [Sat, 14 Oct 2017 11:26:02 +0000 (13:26 +0200)]
aggregator: refactor the test helper
Make the test helpers use a queue. This lets us also test sequences of events,
queries and data.
Stefan Sauer [Sat, 14 Oct 2017 10:08:19 +0000 (12:08 +0200)]
aggregator: test cleanup
Remove gst_init() from a few tests. Use _OBJECT variants in logging. Remove
arbitrary extra blank lines. Make push_event() more like push_buffer() - set
the event to NULL and add cleanup to _chain_data_clear().
Stefan Sauer [Tue, 3 Oct 2017 10:36:10 +0000 (12:36 +0200)]
aggregator: cleanup event forwarding
Don't copy the whole event struct. Set the input params when we call the
forwarding helper. Initialize the internal fields and return values in the
helper.
Stefan Sauer [Tue, 3 Oct 2017 10:08:42 +0000 (12:08 +0200)]
aggregator: simplify src_event
Avoid extra ref/unref, we have a ref and do_seek unrefs. Just return the result
as we have. This lets us remove the local var plus the label.
Stefan Sauer [Sun, 17 Sep 2017 19:37:03 +0000 (12:37 -0700)]
aggregator: register func for do_events_and_queries
This fixes logging the func ptr from _iterate_sinkpads().
Stefan Sauer [Sun, 17 Sep 2017 19:30:37 +0000 (12:30 -0700)]
aggregator: only set clipped_buffer to NULL if needed
Stefan Sauer [Sun, 17 Sep 2017 19:25:37 +0000 (12:25 -0700)]
aggregator: rename check_events
This function also handles queries. Update the code to loop until all events and
queuries are handled.
Stefan Sauer [Sun, 17 Sep 2017 19:24:54 +0000 (12:24 -0700)]
aggregator: add a few more comments to PadPrivate struct
Stefan Sauer [Sun, 17 Sep 2017 18:39:12 +0000 (11:39 -0700)]
aggregator: rename buffers field to data
The queue stores buffers, events and queries.
Stefan Sauer [Sun, 17 Sep 2017 17:18:56 +0000 (10:18 -0700)]
aggregator: documentaion fixes
Fix typos and remove params docs, where the param was moved.
Edward Hervey [Tue, 5 Sep 2017 12:26:52 +0000 (14:26 +0200)]
check: Fix usage of dual probes
Using two (or more) probes on the same pad where one of the probe
returns HANDLED or DROP is tricky since the other probes might
not be called.
Instead use regular probes and a proper pad (the sinkpad already existed,
it only required to be activated and have a dummy chain function for
the events/buffers to be received/handled properly)
Stefan Sauer [Sun, 30 Jul 2017 10:17:57 +0000 (12:17 +0200)]
aggregator: log all events
We already log a few events explicitly, just log them all with more detail.
Tim-Philipp Müller [Sat, 29 Jul 2017 15:54:38 +0000 (16:54 +0100)]
aggregator: fix header formatting
Sebastian Dröge [Mon, 24 Jul 2017 15:38:57 +0000 (18:38 +0300)]
aggregator: Remove the GAP event from the queue before queueing up the GAP buffer
Otherwise check_events() will not remove the GAP event (as the queue
tail is not the event anymore but the GAP buffer), then the GAP buffer
is handled, then the GAP event is handled again, ... forever.
Tim-Philipp Müller [Mon, 17 Jul 2017 23:30:51 +0000 (00:30 +0100)]
aggregator: mark symbols explicitly for export with GST_EXPORT
Stefan Sauer [Thu, 13 Jul 2017 20:00:58 +0000 (22:00 +0200)]
aggregator: remove duplicated code fragment
This code already runs above when (event || query).
Stefan Sauer [Thu, 13 Jul 2017 19:55:55 +0000 (21:55 +0200)]
aggregator: code cleanups
Fix comment typos, some copy'n'paste in logging. Add more doc comments.
Mathieu Duponchelle [Thu, 13 Apr 2017 20:11:55 +0000 (22:11 +0200)]
aggregator: Invalidate pad's tail position ...
when dequeuing a segment event.
https://bugzilla.gnome.org/show_bug.cgi?id=784593
Stefan Sauer [Sat, 1 Jul 2017 18:23:25 +0000 (20:23 +0200)]
aggregator: fix "'aggclass' may be used uninitialized in this function"
Olivier Crête [Mon, 22 May 2017 22:53:57 +0000 (00:53 +0200)]
aggregator: Process serialized queries through the queue
This ensures that they really get processed in order with
buffers. Just waiting for the queue to be empty is sometimes not
enough as the buffers are dropped from the pad before the result is
pushed to the next element, sometimes resulting in surprising
re-ordering.
Olivier Crête [Mon, 22 May 2017 22:53:23 +0000 (00:53 +0200)]
aggregator: Set flow to FLUSHING on pad stop
Fixes a rare race where the pad is being stopped while doing a query.