The Callbacks interface notifies new packets. But in addition, the
screen encoder will need to write headers on start.
We could add a function onStart(), but for simplicity, just remove the
interface, which brings no value, and call the streamer directly.
Refs 87972e2022
On any error, all previously opened sockets must be closed.
Handle these errors in a single catch-block. Currently, there are only 2
sockets, but this will simplify even more with more sockets.
Note: this commit is better displayed with --ignore-space-change (-b).
The recorder opened the target file from the packet sink open()
callback, called by the demuxer. Only then the recorder thread was
started.
One golden rule for the recorder is to never block the demuxer for I/O,
because it would impact mirroring. This rule is respected on recording
packets, but not for the initial recorder opening.
Therefore, start the recorder thread from sc_recorder_init(), open the
file immediately from the recorder thread, then make it wait for the
stream to start (on packet sink open()).
Now that the recorder can report errors directly (rather than making the
demuxer call fail), it is possible to report file opening error even
before the packet sink is open.
The recorder has two initialization phases: one to initialize the
concrete recorder object, and one to open its packet_sink trait.
Initialize mutex and condvar as part of the object initialization.
If there were several packet_sink traits, the mutex and condvar would
still be initialized only once.
Stop scrcpy on recorder errors.
It was previously indirectly stopped by the demuxer, which failed to
push packets to a recorder in error. Report it directly instead:
- it avoids to wait for the next demuxer call;
- it will allow to open the target file from a separate thread and stop
immediately on any I/O error.
User-friendly error messages are printed on specific configuration
exceptions. In that case, do not print the stacktrace.
Also handle the user-friendly error message directly where the error
occurs, and print multiline messages in a single log call, to avoid
confusing interleaving.
DesktopConnection implements Closeable, so it is implicitly closed after
its try-with-resources block. Closing the DesktopConnection shutdowns
the sockets, so it is necessary in particular to wake up blocking read()
calls from the controller.
But the controller thread was joined before the DesktopConnection was
closed, causing a deadlock. To fix the problem, join the controller
thread only after the DesktopConnection is closed.
Refs 400a1c69b1
When a call to a packet or frame sink fails, do not log the error on the
caller side: either the "failure" is expected (explicitly stopped) or it
must be logged by the packet or frame sink implementation.
On close, the client closes the socket. This wakes up socket blocking
calls on the server-side, by throwing an exception. Since this exception
is expected, it was not logged.
However, other IOExceptions might occur, which must not be ignored. For
that purpose, log only IOException when they are not caused by an EPIPE
error.
The PTS received from MediaCodec are expressed relative to an arbitrary
clock origin. We consider the PTS of the first frame to be 0, and the
PTS of every other frame is relative to this first PTS (note that the
PTS is only used for recording, it is ignored for mirroring).
For simplicity, this relative PTS was computed on the server-side.
To prepare support for multiple stream (video and audio), send the
packet with its original PTS, and handle the PTS offset on the
client-side (by the recorder).
Since we can't know in advance which stream will produce the first
packet with the lowest PTS (a packet received later on one stream may
have a PTS lower than a packet received earlier on another stream),
computing the PTS on the server-side would require unnecessary waiting.
MediaCodec may fail spuriously, typically when stopping an encoding and
starting a new one immediately (for example on device rotation).
In that case, retry a few times, in many cases it should work.
Refs #3693 <https://github.com/Genymobile/scrcpy/issues/3693>
For the initial connection between the device and the computer, an adb
tunnel is established (with "adb reverse" or "adb forward").
The device-side of the tunnel is a local socket having the hard-coded
name "scrcpy". This may cause issues when several scrcpy instances are
started in a few seconds for the same device, since they will try to
bind the same name.
To avoid conflicts, make the client generate a random UID, and append
this UID to the local socket name ("scrcpy_01234567").
Use av_packet_ref() to reference the packet without copy.
This also simplifies the logic, by making the "offset" variable and the
memcpy() call local to the if-block.
This fixes the following warning:
> WARNING: Running the setup command as `meson [options]` instead of
> `meson setup [options]` is ambiguous and deprecated.
If the current adb port is not 5555 (typically 0 because it is not in
TCP/IP mode), --tcpip automatically executes (among other commands):
adb tcpip 5555
In case adb was already listening on another port, this command forced
to listen on 5555, and the connection should still succeed.
But this reconfiguration might be inconvenient for the user. If adb is
already in TCP/IP mode, use the current enabled port without
reconfiguration.
Fixes#3591 <https://github.com/Genymobile/scrcpy/issues/3591>
Mandatory arguments may be passed in either of these two forms:
1. --key value
2. --key=value
Optional argument may only be passed in the second form.
For consistency, always document using --key=value.
Refs f76fe2c0d4
Romain Vimont
