Commit graph

178 commits

Author SHA1 Message Date
Jimmy Ohn
4ca3a22b6b hlsdemux: Simplify logic in process_manifest
Simplify logic in process_manifest and remove a TODO item.

https://bugzilla.gnome.org/show_bug.cgi?id=749328
2015-05-18 10:52:35 +03:00
Tim-Philipp Müller
699452ef31 Remove obsolete Android build cruft
This is not needed any longer.
2015-04-26 18:59:32 +01:00
Sebastian Dröge
7c2a5da125 hlsdemux: Use the downloader of the base class instead of creating our own
The one of the base class is completely unused because we override all
the downloading here, so let's just use that one instead.
2015-04-23 17:47:18 +02:00
Sebastian Dröge
5fd2fc3050 hlsdemux: Don't error out if we can't match variant playlists after updating
It's better to just select some random variant playlist instead of stopping,
chances are that it's still continuing to work and we might just have to
select a different variant again later.
2015-04-23 17:47:18 +02:00
Sebastian Dröge
bb36ffb633 hlsdemux: Fix how the playlists are refreshed
We should only refresh the currently selected variant playlist (if any,
otherwise the main playlist), not the main playlist. And only try to
refresh the main playlist if updating the variant playlist fails.

Some servers (Wowza) use the request of the main playlist to create a
"session", which is then part of the URI of the variant playlist and
also the fragments. Refreshing the main playlist would generate a new
session, and the server rate limits that usually. And after a few retries
the server just kicks us out.

Also as a side effect we now use the same downloader for all playlists, so
that we only have 2 instead of 3 connections to the server. And also
previously we just ignored the downloaded data from the main playlist that
the base class gave to us.
2015-04-23 17:47:18 +02:00
Sebastian Dröge
0cd3938345 adaptivedemux: Allow subclasses to override how a new manifest would be downloaded 2015-04-23 17:47:18 +02:00
Thiago Santos
86a503a29b hlsdemux: handle too short segments
When the segment is very short it might be the case that the
typefinding fails and when finishing the segment hlsdemux would
consider the remaining data (pending_buffer) as an encryption
leftover.

This patch fixes it and makes sure an error is properly posted
if typefind failed by refactoring buffer handling to a function
and using it from the data_received and finish_fragment functions.
2015-04-22 20:21:02 -03:00
Thiago Santos
6c513a9d1e hlsdemux: use correct variable type
gst_buffer_resize needs gssize and not gsize. This makes gdb
print it correctly when debugging.
2015-04-22 19:04:43 -03:00
Sebastian Dröge
6c968ed3da hlsdemux: Fix seeking
We also have to update the current_file GList pointer in the M3U playlist
client, otherwise we are just continuing playback from the current position
instead of seeking.
2015-04-21 18:30:33 +02:00
Sebastian Dröge
1e1e73a0f2 hlsdemux: Don't call unlock() just to call lock() immediately afterwards 2015-04-21 18:30:33 +02:00
Thiago Santos
75c9a5f129 hlssink: write the playlist end marker on EOS
Catch EOS from the multifilesink and add the endlist marker to
the playlist when it happens

https://bugzilla.gnome.org/show_bug.cgi?id=747319
2015-04-04 11:00:10 -03:00
Thiago Santos
f00aec48fa hlssink: implement a chain_list to request key unit for segments
upstream might send buffer lists instead of buffers and hlssink's
probe won't get called and a new segment won't be created when needed.

This patch fixes it by adding a chain_list function to the sink pad
that will just pass through the whole bufferlist if no segment needs
to be requested at the moment or convert the list into buffers to
check the proper timestamp to request the next key-unit that will
start the segment.

https://bugzilla.gnome.org/show_bug.cgi?id=746906
2015-03-30 12:31:08 -03:00
Thiago Santos
87f8d7890c hlsdemux: set timestamp on discont buffers
Make sure discont buffers have its timestamp set to allow downstream
to resync if needed
2015-03-04 08:00:48 -03:00
Alex Ashley
a08dd85608 hlsdemux: select correct starting position for live streams
When playing live HLS streams, the media playback starts from the
beginning of the media playlist. When playing a live HLS stream,
media playback should start from 3 fragments from the end of the
playlist.

See section 6.3.3. of the HLS draft [1]

This commit changes the logic to select 3 fragments from the end when
playing a live stream.

[1] http://tools.ietf.org/html/draft-pantos-http-live-streaming-12#page-29

https://bugzilla.gnome.org/show_bug.cgi?id=727742
2015-02-20 19:00:01 -03:00
Guillaume Seguin
b75f64d72e hlssink: reduce playlist's EXT-X-VERSION to 3
We are not using any features greater than the 3rd version,
so reduce to what we actually use to make it compatible with
more players

https://bugzilla.gnome.org/show_bug.cgi?id=744689
2015-02-19 12:17:21 -03:00
Thiago Santos
f79d9522c5 hlsdemux: remove unused define 2015-02-17 11:53:49 -03:00
Thiago Santos
f7a1649c77 adaptivedemux: add bitrate-limit property
Move the property from subclasses to adaptivedemux, it allows
selecing the percentage of the measured bitrate to be used when
selecting stream bitrates
2015-02-17 11:12:42 -03:00
Thiago Santos
00bdac51a3 hlssink: allow creation of 'infinite' playlist
Allow the playlist-length to accept '0' as a value, indicating
that no segment should be removed from the playlist. This allows
generating playlists to be used as VOD when complete.
2015-02-17 11:12:42 -03:00
Thiago Santos
6ec5904d46 hlssink: remove unused attribute
the GFile attribute is never used
2015-02-17 11:12:41 -03:00
Thiago Santos
e9ab79dc36 adaptivedemux: add connection-speed property
Allows to set a bitrate directly instead of measuring it internally
based on the received chunks. The connection-speed was removed from
mssdemux and hlsdemux as it is now in the base class
2015-02-17 11:12:41 -03:00
Thiago Santos
229a15b393 adaptivedemux: refactor chunk downloading flow
Add more power to the chunk_received function (renamed to data_received)
and also to the fragment_finish function.

The data_received function must parse/decrypt the data if necessary and
also push it using the new push_buffer function that is exposed now. The
default implementation gets data from the stream adapter (all available)
and pushes it.

The fragment_finish function must also advance the fragment. The default
implementation only advances the fragment.

This allows the subsegment handling in dashdemux to continuously download
the same file from the server instead of stopping at every subsegment
boundary and starting a new request
2015-01-19 15:30:04 -03:00
Vincent Penquerc'h
d1e8b88eaa hlsdemux: remove unused fragments_cache field 2015-01-16 11:43:22 +00:00
Vincent Penquerc'h
b5860af143 hlsdemux: a few leak fixes 2015-01-16 11:43:22 +00:00
Thiago Santos
d310c0c460 hlsdemux: cache current file position in the list
Avoids iterating the list everytime to look for the next segment
to be used (or to advance to the next one).
2015-01-08 17:55:33 -03:00
Thiago Santos
308b2e5aaf hlsdemux: simplify next segment checking functions
Optimize loop by moving condition outside of it and reuse the
find_next_fragment function to check if there is next instead of
replicating the same loop
2015-01-08 17:55:33 -03:00
Thiago Santos
e21abb62e0 hlsdemux: cache total duration to avoid iterating at every query
Duration queries can be done a few times per second and would cause
the segment list to be traversed for every one. Caching the duration
prevents that.
2015-01-08 17:55:33 -03:00
Luis de Bethencourt
9dcf650a17 hlsdemux: g_return_if_fail in function with return type
Need to use g_return_val_if_fail() when the function returns a type, in this
case a gboolean
2015-01-08 15:54:18 +00:00
Thiago Santos
dd7cb8f632 hlsdemux: implement _has_next_fragment to avoid busy looping
It will allow the demuxer to wait for a fragment to be available instead
of busy looping polling the playlist for a new fragment
2015-01-08 09:59:01 -03:00
Thiago Santos
3a6c2b6b67 hlsdemux: remove duplicate call to uri_join 2015-01-07 10:23:31 -03:00
Thiago Santos
3e9b89111a hlsdemux: skip checking '#EXT-X-' token for all entries
Put this common check before to avoid repeating it for all possible
entries to save some cycles
2015-01-07 09:39:56 -03:00
Thiago Santos
c79e8a78ac hlsdemux: avoid using g_list_append for creating segments list
Use g_list_prepend and reverse it at the end to skip traversing the
same list for every new segment
2015-01-07 09:30:40 -03:00
Thomas Bluemel
cd6069f5af hlsdemux: Don't use approximate duration for fragment buffer pts
The duration values in playlists are approximate only, and for
playlist versions 2 and older they are only rounded integer values.
They cannot be used to timestamp buffers.  This resulted in playback
gaps and skips because the actual duration of fragments is slightly
different.  The solution is to only set the pts of the very first
buffer processed, not for each fragment.
2015-01-07 09:30:40 -03:00
Alex Ashley
50b5d94b2a hlsdemux: Implement live seeking
hlsdemux assumes that seeking is not allowed for live streams,
however seek is possible if there are sufficient fragments in the
manifest. For example the BBC have live streams that contain 2 hours
of fragments.

The seek code for both live and on-demand is common code. The
difference between them is that an offset has to be calculated
for the timecode of the first fragment in the live playlist.

When hlsdemux starts to play a live stream, the possible seek range
is between 0 and A seconds. After some time has passed, the beginning of
the stream will no longer be available in the playlist and the seek
range is between B and C seconds.

Seek range:
start          0 ........... A
later               B ........... C

This commit adds code to keep a note of the B and C values
and the highest sequence number it has seen. Every time it updates the
media playlist, it walks the list of fragments, seeing if there is a
fragment with sequence number > highest_seen_sequence. If so, the values
of B and C are updated. The value of B is used when timestamping
buffers.

It also makes sure the seek range is never closer than three fragments
from the end of the playlist - see 6.3.3. "Playing the Playlist file"
of the HLS draft.

https://bugzilla.gnome.org/show_bug.cgi?id=725435
2015-01-05 17:59:08 -03:00
Thiago Santos
104bd7cf02 hlsdemux: only typefind when we have a minimum amount of data
For small amounts some data might be mistyped and it would cause
the pipeline to fail. For example if you have AAC inside mpegts,
for small amounts, the AAC samples would cause the typefinder to
think it is AAC and not mpegts.

https://bugzilla.gnome.org/show_bug.cgi?id=736061
2014-12-30 08:35:21 -03:00
Vijay Jayaraman
68e77265a0 hlsdemux: typefind might fail if first buffer is too short
If typefind fails, check to see if the buffer is too short for typefind. If this is the case,
prepend the decrypted buffer to the pending buffer and try again the next time around.

https://bugzilla.gnome.org/show_bug.cgi?id=740458
2014-12-23 10:08:57 -03:00
Thiago Santos
1e9ce11efd hlsdemux: port to adaptive base class
Conflicts:
	ext/hls/gsthlsdemux.c
	ext/hls/gsthlsdemux.h
2014-12-23 10:08:57 -03:00
Flávio Ribeiro
c306e0dfa7 hlssink: remove unnecessary title on EXTINF tag
According to the HLS spec the remainder of the line following
the comma on EXTINF tag is not required. This patch removes
the fake title and saves some bytes on the playlist.

https://bugzilla.gnome.org/show_bug.cgi?id=741096
2014-12-04 17:52:08 +01:00
Tim-Philipp Müller
f216b7bb11 Sprinkle some G_PARAM_DEPRECATED and #ifndef GST_REMOVE_DEPRECATED 2014-11-02 17:19:34 +00:00
Tim-Philipp Müller
69ee564d5c hls: fix indentation 2014-11-02 17:17:46 +00:00
Matthieu Bouron
7fb584b087 hlsdemux: reset end_of_playlist attribute when we receive a seek
https://bugzilla.gnome.org/show_bug.cgi?id=738696
2014-10-21 12:55:42 +02:00
Thomas Bluemel
c87835a79f hlsdemux: Fix accessing invalidated memory
In gst_hls_demux_get_next_fragment() the next fragment URI gets
stored in next_fragment_uri, but the gst_hls_demux_updates_loop()
can at any time update the playlist, rendering this string invalid.
Therefore, any data (like key, iv, URIs) that is taken from a
GstM3U8Client needs to be copied. In addition, accessing the
internals of a GstM3U8Client requires locking.

https://bugzilla.gnome.org/show_bug.cgi?id=737793
2014-10-07 15:22:27 +03:00
Philippe Normand
6d67b5263f hlsdemux: lock client mutex before entering the retry_failover block
This is consistent with the case where the block execution is
triggered by the goto invoked after the current_variant update.

https://bugzilla.gnome.org/show_bug.cgi?id=736919
2014-09-18 23:28:15 +01:00
George Kiagiadakis
f4546b64ea hlsdemux: synchronize with the download loop thread to signal it to continue
If EOS or ERROR happens before the download loop thread has reached its
g_cond_wait() call, then the g_cond_signal doesn't have any effect and
the download loop thread stucks later.

https://bugzilla.gnome.org/show_bug.cgi?id=735663
2014-09-18 13:38:27 -03:00
Thiago Santos
07b59c93c2 hlsdemux: fix clearing of eos state in pads
The internal pad still keeps its EOS flag and event as it can be assigned
after the flush-start/stop pair is sent. The EOS is assigned from the streaming
thread so this is racy.

To be sure to clear it, it has to be done after setting the source to READY to
be sure that its streaming thread isn't running.

https://bugzilla.gnome.org/show_bug.cgi?id=736012
2014-09-18 12:14:30 -03:00
Sebastian Dröge
775c5600c9 hlsdemux: Also refetch the playlist after the first fragment failure
Previously we only refetched the playlist if downloading a fragment
has failed once. We should also do that if it failed a second or third time,
chances are that the playlist was updated now and contains new URIs.
2014-09-15 13:33:45 +03:00
Sebastian Dröge
d4a53a16dd hls: Actually retry 3 times as advertised instead of 2 2014-09-15 11:59:46 +03:00
Sebastian Dröge
f6c6d9c280 hlsdemux: Don't send flush events to deactivated pads
https://bugzilla.gnome.org/show_bug.cgi?id=736012
2014-09-04 18:20:58 +03:00
Sebastian Dröge
46b0310acd hlsdemux: Properly assign offsets to the files if we accumulate them instead of reading from the playlist 2014-09-04 17:49:23 +03:00
Thomas Bluemel
04ca723461 hlsdemux: Support OpenSSL for AES decryption of HLS fragments
https://bugzilla.gnome.org//show_bug.cgi?id=735248
2014-08-28 10:34:55 +03:00
Thibault Saunier
16201cec34 hlsdemux: Do not switch playlist on trick modes
Instead always use the low bandwith playlist making things go smoother
as the current heuristic is rather set for normal playback, and
currently it does not behave properly.

https://bugzilla.gnome.org/show_bug.cgi?id=734445
2014-08-13 17:49:12 +02:00