Commit graph

1602 commits

Author SHA1 Message Date
Sebastian Dröge f5260f9b36 fallbackswitch: Explictly set priority of pads in the test 2022-04-14 08:14:26 +00:00
Sebastian Dröge 0ca0d485a0 fallbackswitch: Log affected pad when scheduling buffer waiting for it 2022-04-14 08:14:26 +00:00
Sebastian Dröge d73bce2985 fallbackswitch: Only update the output running time if it's increasing 2022-04-14 08:14:26 +00:00
Sebastian Dröge 930cdca750 fallbackswitch: Only drop raw audio/video buffers if they're trailing
Otherwise this can cause decoding errors further downstream.
2022-04-14 08:14:26 +00:00
Sebastian Dröge bbb7ced95a fallbackswitch: Improve some debug outputs
Don't print "buffer buffer ...".
2022-04-14 08:14:26 +00:00
Sebastian Dröge ca7cf7dee7 fallbackswitch: Fix clipping of buffers against the output running time
To handle buffers outside the segment correctly.

Also fix debug output to print the correct value.
2022-04-14 08:14:26 +00:00
Sebastian Dröge 617a2ef49e fallbackswitch: Clip buffers against the segment
Also clip raw audio/video buffers according to the caps if they have the
relevant information in the caps, and drop raw audio samples if they're
outside the segment too.

In addition also set durations on raw audio/video buffers based on the
caps if no duration is set.
2022-04-14 08:14:26 +00:00
Sebastian Dröge 6ce523a7a8 fallbackswitch: Improve debug output of clock times and simplify some code 2022-04-14 08:14:26 +00:00
Sebastian Dröge 9e3f713aa9 Update to m3u8-rs 4.0 2022-04-14 07:41:18 +00:00
Thibault Saunier 94b7677318 meson: Fix building when git isn't available
Fixes https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/issues/178
2022-04-13 17:01:34 -04:00
Guillaume Desmottes 5004479c6f uriplaylistbin: disable more tests
See https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/issues/194
2022-04-13 07:55:27 +00:00
Sebastian Dröge a4fd144647 deny: Remove now unneeded miniz_oxide override 2022-04-13 10:40:29 +03:00
Seungha Yang 930cc41aaa fallbacksrc: Don't forward manual flush events to downstream
fallbackswitch might forward flush event if it's for the currently
active pad. But forwarded flush event will be problematic in various
reasons and that's not a behavior we expected.
2022-04-12 23:21:32 +09:00
Sebastian Dröge 5788837fb6 fallbacksrc: Unflush downstream pads of the source after shutting down everything
Otherwise the pads might still be flushing on the next try and the
source would never recover.
2022-04-12 14:40:08 +03:00
Sebastian Dröge 7233d6936c fallbacksrc: Remove imagefreeze if it becomes unneeded 2022-04-12 14:37:29 +03:00
Sebastian Dröge 874063668b Update license text in the README.md 2022-04-12 12:24:55 +03:00
Sebastian Dröge 28dbbe33db Update README.md with all the new plugins/elements 2022-04-12 12:21:00 +03:00
Vivia Nikolaidou de936f42e9 fallbackswitch: Minor cleanup in get_sync_time 2022-04-11 15:57:20 +03:00
Sebastian Dröge 3de317b3c9 fallbacksrc: Don't panic when shutting down the source and there are source pad blocks but no actual source pads
This can happen if the source removed the actual source pad already but
our downstream pad that is blocked still exists.
2022-04-11 11:01:27 +00:00
Sebastian Dröge 84402f39ef fallbacksrc: Reset state in fallbacksrc custom source also if changing the state to READY before failed
Otherwise it will panic on the next try because always pads are already
linked to the ghostpads.
2022-04-11 11:01:27 +00:00
Vivia Nikolaidou 0ba6ebb10f fallbackswitch: Notify active-pad without state lock held
Can otherwise deadlock.
2022-04-11 10:48:28 +00:00
Vivia Nikolaidou b5a3a99825 m3u8-rs: Depend on version exactly 3.0.0
https://github.com/rutgersc/m3u8-rs/pull/46#issuecomment-1094867533
2022-04-11 13:22:44 +03:00
Sebastian Dröge 5af52f94a8 threadshare: Remove glib::SendUnique usage
It's being removed from the GLib bindings because it does not add much
value.
2022-04-09 08:41:58 +00:00
Seungha Yang 558656deb5 transcriberbin: Protect runtime "passthrough=false" change
There can be small race where transcription-bin is linked with
tee but state change of the transcription-bin is not finished.
And at the same time, upstream pushes event/buffer to the
transcription-bin. Do state change first then link to avoid
the condition

Part-of: <https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/merge_requests/716>
2022-04-08 21:55:54 +00:00
Sebastian Dröge 08379ab389 fallbacksrc: Request the main pad first from fallbackswitch
By default the first created pad is the active one before the
fallbackswitch is actually processing data.
2022-04-08 20:24:03 +03:00
Sebastian Dröge 96c28a5728 fallbackswitch: Only notify active-pad property on state changes if it actually changed 2022-04-08 20:23:28 +03:00
Sebastian Dröge 452ea76a69 fallbackswitch: Rename debug category from fallback-switch to fallbackswitch
The element name is without hyphen so the debug category should ideally
too to reduce confusion.
2022-04-08 20:07:01 +03:00
Sebastian Dröge d3d98c73ca fallbacksrc: Connect to notify::active-pad after requesting pads
Requesting the first pad will emit the property because the first pad is
then selected. That will cause the callback to be called, which tries to
take the same mutex that is already locked during element setup and
causes a deadlock.
2022-04-08 19:53:38 +03:00
Guillaume Desmottes 7daab76f17 uriplaylistbin: skip missing_http test
It's not reliable.
2022-04-08 16:18:03 +02:00
Guillaume Desmottes 00b3199727 Revert "Revert "uriplaylistbin: disable test deadlocking""
This reverts commit 821ec857e1.

Looks like the test is still unreliable. :(
2022-04-08 16:16:05 +02:00
Guillaume Desmottes cde5fdf202 uriplaylistbin: break reference cycle
Passing ownership of item to the probe callback was introducing a reference cycle as the item is owning the sinkpad.
2022-04-08 15:50:48 +02:00
Guillaume Desmottes ffdcc8167c uriplaylistbin: remove useless uridecodebin clone 2022-04-08 15:50:48 +02:00
Guillaume Desmottes 821ec857e1 Revert "uriplaylistbin: disable test deadlocking"
This reverts commit d9f3e8e9e7.

The test seems to work reliably so hopefully it's been fixed with the
latest dead lock fixes.

Fix #191
2022-04-08 12:58:13 +02:00
Guillaume Desmottes 69ceaa3a5e uriplaylistbin: add Status::ShuttingDown
Fix a race where we were trying to process new decodedin pads while the
element was being shutdown, resulting in a dead lock.
2022-04-08 12:58:13 +02:00
Guillaume Desmottes 0ed72a360d uriplaylistbin: replace errored with Status enum 2022-04-08 12:58:13 +02:00
Guillaume Desmottes 97dba9046b uridecodebin: stop using crossbeam_channel
I give up on crossbeam_channel. For some reasons some receivers are not
always unblocked and I was not able to reproduce using simpler test
cases.
Use with mpsc channels instead which are more reliable.
2022-04-08 12:58:13 +02:00
Sebastian Dröge e8f15eb1e8 ci: Rebuild image and move coverage job to the stable image 2022-04-07 19:48:57 +03:00
Sebastian Dröge 803e452889 Update minimum supported GStreamer version to 1.14 2022-04-07 12:41:54 +03:00
Jan Alexander Steffens (heftig) 0a3e9c81f9
fallbackswitch: Add missing requires_private 2022-04-06 11:11:25 +02:00
Sebastian Dröge 584daeaa81 deny: Update for tokio-util 2022-04-06 11:46:09 +03:00
Vivia Nikolaidou f8d162d909 rustfmt.toml: Add file
This will automatically run `cargo fmt` when a file is saved.
2022-04-05 19:13:59 +03:00
Jan Schmidt bd2ff494c7 fallbackswitch: Replace with priorityswitch
fallbackswitch now supports multiple sink pads, and on a timeout of the
active pad, it will automatically switch to the next lowest priority pad
that has data available.

fallbackswitch sink pads follow the `sink_%u` template and have
`priority` as a pad property.

Co-authored-by: Vivia Nikolaidou <vivia.nikolaidou@ltnglobal.com>
2022-04-05 18:52:31 +03:00
Guillaume Desmottes bf14939b9b uriplaylistbin: remove pad probe when shutting down
The previous patch was unblocking the receiver in the pad probe but
it could be blocked again as the item streaming status was not changed.
2022-04-01 11:58:50 +02:00
Vivia Nikolaidou d48d732038 .gitlab-ci.yml: Fix deprecated -Zinstrument-coverage switch 2022-03-30 16:16:25 +03:00
Guillaume Desmottes 5dab4bc502 uriplaylistbin: handle waiting items no longer being set
ab96219c19 broke some assertions as
waiting items are now unset when shutting down the element.
2022-03-28 13:02:40 +02:00
Sebastian Dröge 518e43fc5f cea608tott: Don't zero-pad SRT indices but zero-pad hours to at least two digits
Zero-padding is not specified for the indices but all time components
need to be zero-padded (3 digits for fractional seconds, 2 digits for
everything else).
2022-03-28 12:17:34 +03:00
François Laignel 59ca466081 ts: log max throttling when creating Context 2022-03-28 08:47:32 +00:00
François Laignel 1ef9ae6398 ts/jitterbuffer: don't wake up immediately...
... when next wakeup delay is shorter than the max throttling duration.

See https://gitlab.freedesktop.org/gstreamer/gst-plugins-rs/-/issues/169
2022-03-28 08:47:32 +00:00
François Laignel 8eb8ea0e7d ts/rt/Task: use light weight executor blocking on ack or join handle
Previous version used the Context::block_on_or_add_sub_task which
spawns a full-fledged executor with timer and io Reactor for no
reason when we just need to wait for a Receiver or JoinHandle.
2022-03-28 08:47:32 +00:00
François Laignel c1615d01e6 ts/rt/Task: awake the iteration loop when it needs to be aborted
When the iteration loop is throttling, the call to `abort` on the
`loop_abort_handle` returns immediately, but the actual `Future`
for the iteration loop is aborted only when the scheduler throttling
completes. State transitions which requires the loop to be aborted &
which are serialized at the pipeline level can incur long delays.

This commit makes sure the Task Context's scheduler is awaken as soon
as the task loop is aborted.
2022-03-28 08:47:32 +00:00