The type is duplicated otherwise

Found with modernize-use-auto
Signed-off-by: Rosen Penev <rosenp@gmail.com>

Found with readability-const-return-type
Signed-off-by: Rosen Penev <rosenp@gmail.com>

Found with readability-use-anyofallof
Signed-off-by: Rosen Penev <rosenp@gmail.com>

loop is wrong anyway. It's missing a break;
Signed-off-by: Rosen Penev <rosenp@gmail.com>

The latter is easier to read and is the "correct" thing to do.
Signed-off-by: Rosen Penev <rosenp@gmail.com>

Signed-off-by: Rosen Penev <rosenp@gmail.com>

Found with cppcoreguidelines-avoid-non-const-global-variables
Signed-off-by: Rosen Penev <rosenp@gmail.com>

Found with modernize-use-default
Signed-off-by: Rosen Penev <rosenp@gmail.com>

Don't truncate the FILETIME to second resolution to pass it to
std::chrono::system_clock::from_time_t(); instead, calculate the
offset between the FILETIME epoch and the
std::system_clock::time_point epoch, and use that to initialize the
time_point directly.

Allows running the unit tests on Linux.

Got lost in commit 0f56ddb8

This avoids underruns at the start of playback.

This was missing in commit 8a243e6e

If draining was not requested, generate silence instead if there is no
data in the ring buffer.

The problem is that pw_stream_flush() appears to disable the stream
permanently, even though there is no state_changed callback - the
stream state remains at PW_STREAM_STATE_STREAMING, but the stream is
defunct.  I have no idea why and I havn't found any documentation
about it.

Closes https://github.com/MusicPlayerDaemon/MPD/issues/1219

This was missing in commit 4d1ce702

release v0.22.10

This replaces 967af603 with a more
effective workaround.

Closes https://github.com/MusicPlayerDaemon/MPD/issues/822

snd_mixer_handle_events() has already been called by
DispatchSockets().  This way, we can also skip the exception handler.