Olivier Crête authored Jul 11, 2017 and GStreamer Marge Bot committed Sep 18, 2020

This flag always causes problems as it prevents subbuffering,
instead one should create a custom GstAllocator to pool the GstMemory objects
and not rely on the lifetime of the GstBuffer object they were originally
attached to.

https://bugzilla.gnome.org/show_bug.cgi?id=757254

Part-of: <gstreamer/gstreamer!614>

Mathieu Duponchelle authored Sep 16, 2020 and GStreamer Marge Bot committed Sep 18, 2020

Subsequent lookups in the hashtable are probably better done
on memory we're confident is allocated to us :)

It was easy to trigger invalid reads by calling gst_meta_register
with dynamically allocated memory, freeing that memory, then
calling gst_meta_get_info()

Part-of: <gstreamer/gstreamer!628>

Olivier Crête authored Jul 06, 2020 and GStreamer Marge Bot committed Sep 18, 2020

This should make the API usage more consistent. Also document that the subclasses
should just return the devices as floating.

Part-of: <gstreamer/gstreamer!558>

Matthew Waters authored Jul 31, 2020 and GStreamer Marge Bot committed Sep 18, 2020

If one pad returns not-negotiated from a caps event, then all other sink
pads were returning not-negotiated.

In our case, we can't reliably easily fail at all so just remove that
code.

Part-of: <gstreamer/gstreamer!623>

Matthew Waters authored Jul 31, 2020 and GStreamer Marge Bot committed Sep 18, 2020

e.g. h264parse ! video/x-h264,stream-format=avc receives the following:
- caps: video/x-raw,stream-format=byte-stream
- gap event: baseparse tries to choose some default caps but would
  override the downstream chosen caps field with upstreams value.

Part-of: <gstreamer/gstreamer!581>

Sebastian Dröge authored Sep 18, 2020 and GStreamer Marge Bot committed Sep 18, 2020

Otherwise the clock id we access might not be a valid pointer anymore.

Part-of: <gstreamer/gstreamer!630>

In case that no available stack tracer, leak tracer will set
null string value for `trace` field. I would likely happen
on Windows.

Part-of: <gstreamer/gstreamer!625>

g_thread_yield() doesn't ensure thread switching actually.
It would result to adding so many pads. Depending on system,
timeout might happen then.

Part-of: <gstreamer/gstreamer!625>

Expected segment-done message might not be seen within expected
time if system is not powerful enough.

Part-of: <gstreamer/gstreamer!625>

gstvalve.c(285) : warning C4715: 'gst_valve_event_needs_dropping':
not all control paths return a value

Part-of: <gstreamer/gstreamer!624>

Sebastian Dröge authored Jul 01, 2020 and Sebastian Dröge committed Sep 10, 2020

This allows to signal the reason for the gap, for example missing data
like packet loss.

Based on a patch by Mikhail Fludkov <misha@pexip.com>

Part-of: <gstreamer/gstreamer!551>

By default each harness returns is_live=TRUE in latency queries. This is
often not desired and can now be overridden.

Part-of: <gstreamer/gstreamer!599>

Some base classes like videoaggregator try retrieving the latency during
construction, which causes the latency values to be set already until
reconfiguration happens.

By resetting them the same way as in stop() we ensure that we always
start cleanly.

Part-of: <gstreamer/gstreamer!599>

Tom Schoonjans authored Sep 08, 2020 and GStreamer Marge Bot committed Sep 09, 2020

Part-of: <gstreamer/gstreamer!620>

Xi Ruoyao authored Sep 09, 2020 and GStreamer Marge Bot committed Sep 09, 2020

Part-of: <gstreamer/gstreamer!621>

Sebastian Dröge authored Aug 20, 2020 and Sebastian Dröge committed Sep 09, 2020

When going to PLAYING we will now have a clock and can stop waiting on
the condition variable and instead start waiting on the clock if
necessary for the current configuration.

In the other direction when going to PAUSED the clock might have
disappeared and we might need to wait on the condition variable again
instead.

Part-of: <gstreamer/gstreamer!601>

Sebastian Dröge authored Aug 24, 2020 and GStreamer Marge Bot committed Sep 09, 2020

Otherwise deactivating them will cause a deadlock as they're blocking
inside the streaming thread.

Fixes gstreamer/gstreamer#601

Part-of: <gstreamer/gstreamer!606>

Sebastian Dröge authored Aug 14, 2020 and GStreamer Marge Bot committed Sep 09, 2020

Fixes gstreamer/gstreamer#587
Fixes gstreamer/gstreamer#84

Part-of: <gstreamer/gstreamer!594>

Mathieu Duponchelle authored Sep 01, 2020 and GStreamer Marge Bot committed Sep 09, 2020

Enforce that the last buffer that was peeked (or had its existence
checked) on a pad is the one that gets popped / dropped, resetting
at the end of each aggregation cycle.

Fixes gstreamer/gstreamer#603

Part-of: <gstreamer/gstreamer!608>

Sebastian Dröge authored Sep 08, 2020 and GStreamer Marge Bot committed Sep 08, 2020

Part-of: <gstreamer/gstreamer!619>

Part-of: <gstreamer/gstreamer!618>

Part-of: <!607>

Sebastian Dröge authored Aug 20, 2020 and GStreamer Marge Bot committed Aug 24, 2020

On the first buffer the base class would update the segment position
based on the start-time-selection. If the subclass provides its own
segment this will caused unexpected behaviour and override segment
information that was explicitly set by the subclass.

Part-of: <gstreamer/gstreamer!600>

GI expects the instance parameter to be documented, omitting it
leads to a msismatched output in the gir.

Part-of: <gstreamer/gstreamer!592>

Subclasses can use this to provide more information, for example
audioaggregator could provide the offset into the output buffer where
the next data is going to be filled.

See gst-plugins-base#805

Part-of: <!590>

Part-of: <gstreamer/gstreamer!588>

Xavier Claessens authored Aug 04, 2020 and GStreamer Marge Bot committed Aug 04, 2020

Part-of: <gstreamer/gstreamer!586>

Jordan Petridіs authored Aug 03, 2020 and GStreamer Marge Bot committed Aug 04, 2020

the typesystem checks in g_atomic_pointer_compare_and_exchange
seem to trigger some false positives with clang 10

Part-of: <gstreamer/gstreamer!584>

Subclasses use the pad segment to determine whether a buffer
should be skipped, we thus don't want to check if a buffer
needs to be skipped before processing the segment it's part
of.

Part-of: <gstreamer/gstreamer!585>

Nirbheek Chauhan authored Jul 30, 2020 and GStreamer Marge Bot committed Aug 03, 2020

New API: gst_uri_from_string_escaped()

Identical to gst_uri_from_string() except that the userinfo and
fragment components of the URI will not be unescaped while parsing.

This is needed for correctly parsing usernames or passwords with `:`
in them such as reported at:
gstreamer/gst-plugins-bad#831

Part-of: <gstreamer/gstreamer!583>

Nirbheek Chauhan authored Aug 01, 2020 and GStreamer Marge Bot committed Aug 03, 2020

Add tests that exercise unescaping of userinfo and fragments.

Also convert to a modular macro-based definition so that we can reuse
the list of tests in the next commit.

Part-of: <gstreamer/gstreamer!583>

Thibault Saunier authored Jul 30, 2020 and GStreamer Marge Bot committed Aug 03, 2020

Part-of: <gstreamer/gstreamer!582>

Mathieu Duponchelle authored Jun 30, 2020 and Sebastian Dröge committed Jul 31, 2020

See gstreamer/gst-plugins-base#771
for context.

This exposes new API that subclasses must call from their
aggregate() implementation to signal that they have selected
the next samples they will aggregate: gst_aggregator_selected_samples()

GstAggregator will emit a new signal there, `samples-selected`,
handlers can then look up samples per pad with the newly-added
gst_aggregator_peek_next_sample.

In addition, a new FIXME is logged when subclasses haven't actually
called `selected_samples` from their aggregate() implementation.

Part-of: <gstreamer/gstreamer!549>

Hosang Lee authored Jul 28, 2020 and GStreamer Marge Bot committed Jul 29, 2020

Don't use percent-encoding for '=' in http queries.
'=' in the following kind of http query should be maintained.
example:
?token=exp=123~acl=/QualityLevels(*~hmac=0cb ...

Part-of: <gstreamer/gstreamer!580>

Camilo Celis Guzman authored Jul 24, 2020 and GStreamer Marge Bot committed Jul 28, 2020

Part-of: <gstreamer/gstreamer!575>

Olivier Crête authored Jul 24, 2020 and GStreamer Marge Bot committed Jul 28, 2020

In 1.0, there is no concept of segment update, so don't push new
identical segments.

Part-of: <gstreamer/gstreamer!578>