The normal I/O event thread can have a large latency, e.g. when
libgnutls loads all TLS CA certificates for a https connect.  This
makes it unreliable for the ALSA I/O notifications, and causes ring
buffer xruns.  To avoid interfering with high latency events such as
CURL's, we move the ALSA I/O events to a separate I/O thread which
also obtains real-time scheduling (if possible).

Closes #221

release v0.20.17

Due to rounding errors, a slightly negative value can be passed to
set_normalized_volume(), which will make the log10() call fail.
Actually, volume 0 is already failing because log10(0) is illegal.  So
let's fix this by implementing two corner cases: <=0 and >=100.

Closes #212

Closes #213.

Apparently, clang defaults to NEON when ARMv7 is used.  Not all ARMv7
CPUs we target have NEON, so we need to disable that.

Not needed on Android, and the implementation uses getpwuid_r() which
is unavailable on old Android versions.

Requires Android API 21, but we want to support older versions as
well.

For Android pre-5.0 compatibility (#213).

For Android pre-5.0 compatibility (#213).

For version numbers such as OpenSSH's, e.g.: "7.2p2"

Closes #217

Fixes valgrind warning.

This was a buffer overflow bug which could cause MPD crahes when
playing back 8 channels with the ALSA output plugin.

Closes #216

There is a POSIX definition for sched_setscheduler(), but Linux does
not implement that; instead of changing the process's scheduler, it
only affects one thread.  This has caused some confusion among
application developers and C library developers.

While glibc implements Linux semantics, Musl has made their
sched_setscheduler() function an always-failing no-op, causing the
error message "sched_setscheduler failed: Function not implemented".

 http://git.musl-libc.org/cgit/musl/commit/src/sched/sched_setscheduler.c?id=1e21e78bf7a5c24c217446d8760be7b7188711c2

Instead of relying on the C library which may be unreliable here, we
now roll our own system call wrapper.

Closes #218

Juse in cas glibc gets a wrapper for the system call which would then
conflict with ours.

Error message sent to client was "basic_string::_M_construct null not
valid" due to passing nullptr to the std::string constructor.

Regression caused by commit 386688b8

release v0.20.16

Closes #173

Fixes deadlock bug.  Closes #210

The PulseOutput needs to be "enabled" before WaitConnection() may be
called.

Closes #207

This commit also puts an instance of RemoteTagScanner into the
Instance class, and hooks it into the "add" and "addid" commands.