Use GST_PAD_PROBE_PASS to pass through all events other than EOS instead of
blocking on the first non-EOS event forever. Also fix a typo in a comment in
that function.
Thanks to David Jaggard for reporting this on the mailing list.
'gst_element_post_message' takes the ownership of the message, so it
shall unref it when there is no post_message implementation. Otherwise
message is leaked.
https://bugzilla.gnome.org/show_bug.cgi?id=759300
This lock seems to exist only to prevent elements from changing states while
events are being processed. However events are going to be processed
nonetheless in those elements if sent directly via pads, so protection must
already be implemented inside the elements for event handling if it is needed.
As such having the lock here is not very useful and is actually causing
various deadlocks in different situations as described in
https://bugzilla.gnome.org/show_bug.cgi?id=744040
segment.position is meant for internal usage only, but the various
GST_EVENT_SEGMENT creationg/parsing functions won't clear that field.
Use the appropriate segment boundary as an initial value instead
Otherwise each bin might have a different latency in the end, causing
synchronization problems.
The bin will still first handle latency internally as before, but gives the
overall pipeline the opportunity to update the latency of the whole pipeline
afterwards.
https://bugzilla.gnome.org/show_bug.cgi?id=759125
Watching videos with variant bitrate is common to have delta
more than 10 kbps, resulting in tag list spam.
Instead of relying on fixed 10 kpbs delta, it is better to
calculale the difference in percentage and update tag list
only when bitrate changes more than 2%.
https://bugzilla.gnome.org/show_bug.cgi?id=759055
The pad could be activated but flushing because of a FLUSH_START event. That's
not what we're looking for here, we want to check for activated pads.
https://bugzilla.gnome.org/show_bug.cgi?id=758928
Changing states up and down while buffers are being pushed is not
a valid use case. If a pad is deactivated and reactivated during
a buffer push it is racy with the check of pushed sticky events
and the actual chainfunction call. As it might call the chain
without noticing the peer pad lost its previous sticky events.
https://bugzilla.gnome.org/show_bug.cgi?id=758340
My previous fix for #758029 wasn't quite right and simply made the race rarer.
Some of the files are installed by install-exec and others by install-exec, so
the hooks need to be split too.
https://bugzilla.gnome.org/show_bug.cgi?id=758029
When synchronizing the output by time, there are some use-cases (like
allowing gapless playback downstream) where we want the unlinked streams
to stay slightly behind the linked streams.
The "unlinked-cache-time" property allows the user to specify by how
much time the unlinked streams should wait before pushing again.
Multiqueue should only be used to cope with:
* decoupling upstream and dowstream threading (i.e. having separate threads
for elementary streams).
* Ensuring individual queues have enough space to cope with upstream interleave
(distance in stream time between co-located samples). This is to guarantee
that we have enough room in each individual queues to provide new data in
each, without being blocked.
* Limit the queue sizes to that interleave distance (and an extra minimal
buffering size). This is to ensure we don't consume too much memory.
Based on that, multiqueue now continuously calculates the input interleave
(per incoming streaming thread). Based on that, it calculates a target
interleave (currently 1.5 x real_interleave + 250ms padding).
If the target interleave is greater than the current max_size.time, it will
update it accordingly (to allow enough margin to not block).
If the target interleave goes down by more than 50%, we re-adjust it once
we know we have gone past a safe distance (2 x current max_size.time).
This mode can only be used for incoming streams that are guaranteed to be
properly timestamped.
Furthermore, we ignore sparse streams when calculating interleave and maximum
size of queues.
For the simplest of use-cases (single stream), multiqueue acts as a single
queue with a time limit of 250ms.
If there are multiple inputs, but each come from a different streaming thread,
the maximum time limit will also end up being 250ms.
On regular files (more than one input stream from the same upstream streaming
thread), it can reduce the total memory used as much as 10x, ending up with
max_size.time around 500ms.
Due to the adaptive nature, it can also cope with changing interleave (which
can happen commonly on some files at startup/pre-roll time)
This will mean a much lower delay before a subtitles track changes take
effect. Also avoids excessive memory usage in many cases.
This will also consider sparse streams as (individually) never full, so
as to avoid blocking all playback due to one sparse stream.
https://bugzilla.gnome.org/show_bug.cgi?id=600648
* Avoid the computation completely if we know we don't need it (not in
sync time mode)
* Make sure we don't override highest time with GST_CLOCK_TIME_NONE on
unlinked pads
* Ensure the high_time gets properly updated if all pads are not linked
* Fix the comparision in the loop whether the target high time is the same
as the current time
* Split wake_up_next_non_linked method to avoid useless calculation
https://bugzilla.gnome.org/show_bug.cgi?id=757353
When preparing a buffering message, don't report 0% if there
is any bytes left in the queue at all. We still have something
to push, so don't tell the app to start buffering - maybe
we'll get more data before actually running dry.
This is useful for feature that are produced after probing a specific
node. You want to reload this plugin if the specific node(s) have been
removed, added, or reloaded.
https://bugzilla.gnome.org/show_bug.cgi?id=758080
In plugin is responsible for calculating a hash of the dependencies
in order to determine if the cache should be invalidated or not.
Currently, the hash combining method removes a bit of the original
have before combining with an addition. As we use 32bits for our hash
and shift 1 bit for each file and directory, that resulting hash only
account for the last 32 files. And is more affected by the last file.
Rotating technique (shifting, and adding back the ending bit), can be
use to make the addition non-commutative. In a way that different order
gives different hashes. In this case, I don't preserve this behaviour
because the order in which the files are provided by the OS is
irrelevant.
In most cases, the XOR operation is used to combine hashes. In this
code we use the addition. I decided to preserve the addition because
we make use of non-random hash ((guint) -1) in the algorithm for
matching files that are not really part of the hash (symlinks, special
files). Doing successive XOR on this value, will simply switch from
full ones, to full zero. The XOR used with whitelist has been preserved
as it's based on a fairly randomized hash (g_str_hash).
https://bugzilla.gnome.org/show_bug.cgi?id=758078
