We mostly use classical C strings, but the memory management around them
can get confusing and error prone. Let's use std::string for the cases
where we need to return a newly allocated string.

A blank string might be very confusing, depending on where this will be
used. Let's give something more visible back in the cases where we
cannot get the proper name for the peer.

This mimics how some system functions (like inet_ntop()) work, and
avoids complexity around ownership of the returned string buffer.

The downside is that the string must be consumed directly as it will be
overwritten on the next call, but that is not an issue with the current
usage.

Avoid complicating things by moving things in to a second buffer here as
there is no need for it.

We handled this in the failure scenario, but not in the vastly more
common successful case.

The principle can be used in a more general fashion than just TCP
streams.

Created a new subclass of Exception called GAIException() that will
handle error messages from getaddrinfo() instead of letting Exception()
handle it. GAIException() will make use of gai_strerror() to map the
error code to text. On Windows, gai_strerrorW() must be used if the text
is encoded with UTF-8.

Makes the code more general and not directly tied to specifically
TCP sockets.

This makes it possible to create a derived class from Socket which is
not TCP based, without having VNCSConnectionST.cxx trying to call
setsockopt() on a non-socket.

The previous detection would fail if the socket closed before we
had time to inspect it, which got us stuck in a loop as we would
try (and fail) to do accept() on a non-listening socket.

The previous detection would fail if the socket closed before we
had time to inspect it, which got us stuck in a loop as we would
try (and fail) to do accept() on a non-listening socket.

Parsing a filter might be the first thing we do so we need to
make sure the socket subsystem is running so that we can look
up names.

It is easier to control object life time and avoid magical socket
duplication by having a single TcpListener object to pass around.
We have to be more careful about deleting the object though.

Closing the socket might overwrite errno so we need to save
its value.

The logic was flawed and would treat all connect errors as
if there were no addresses found.

SO_REUSEADDR needs to be set before we call bind() for it to have
any effect. It got shuffled around in e4d9726.

Every reasonably current system support getaddrinfo() so get rid
of the legacy fallbacks.

It doesn't work on sockets, which require a bit more care.

It is needed by WinVNC, but got removed in 892d10a70. Also fix
a couple of issues:

 - Use getnameinfo() in order to be compatible with Windows XP
 - Make it static since it doesn't use a specific socket
 - Respect UseIPv4 and UseIPv6
 - Flags for getaddrinfo() that match binding code
 - Dummy service value for Windows compatibility

The TcpListener constructor now takes a 'struct sockaddr*' instead of
a string, and the createTcpListeners function creates TcpListener
instances for an address based on the results from getaddrinfo().

The XserverDesktop class now takes a list of TcpListener instances for
each of the RFB and HTTP sockets.

The TcpListener::closeFd member variable is not used and has been
removed.

The logic is a bit complex so we're going to have to silence the
warning with this explicit assignment.

They use char* rather than void*, requiring a cast here.

This is the format we support when converting a string to an
address, so it makes sense to use them the other way around.

There might be more protocols in the future, and we might also
misdetect things depending on varying IPv6 functions being missing.

It's either not used, or no longer relevant.

when necessary.


git-svn-id: svn://svn.code.sf.net/p/tigervnc/code/trunk@5018 3789f03b-4d11-0410-bbf8-ca57d06f2519

controlled manner. Unfortunately there is no equivalent for Windows.


git-svn-id: svn://svn.code.sf.net/p/tigervnc/code/trunk@4781 3789f03b-4d11-0410-bbf8-ca57d06f2519