Preparing the protocol spec for multi-player support.

Fixes a potential deadlock introduced by commit 94528735

Using variable-size objects is not worth the trouble here.  Let's drop
this and use existing and simpler code.

Fixes another buffer overflow: if the stream has a very long title or
URL, resulting in a metadata string of more than 2 kB, icy_string[0]
is a negative value, which gets casted to size_t - ouch!

 https://bugs.musicpd.org/view.php?id=4652

Fixes a buffer overflow due to the bad formula rounding the buffer
size up.  At the same time, remove the "+1" from the meta_length
calculation, which takes the padding into account and at the same time
implements proper rounding.

Why did this cast exist??

With this commit, multi-player support becomes possible... it's just
not wired to the frontend yet.

This is based on massive amounts of refactoring work I did over the
past 9 years.

The first step to multi-player support.  Not much, just a dummy
command.

During UnlockActivate() while the mutex is unlocked, the IOThread can
set a new error condition, and will never again wake up the
OutputThread.  This race condition can cause a deadlock in the
OutputThread.

Since we switched to C++ exceptions, there is no code path which
returns false.

We don't need to access Thread::handle early inside the EventThread,
so we don't need this trick anymore.

Eliminate global variables.