- Adds new dgram module, for all data-gram type transports
- Supports both UDP client and servers
- Supports Unix Daemon sockets in DGRAM mode too (think syslog)
- Uses a shared Buffer and slices that as needed to be reasonably
  performant.
- One supplied test program so far, test-dgram-pingpong
- Passes test cases on osx 10.6 and ubuntu 9.10u

This fixes #169 by putting the require members onto the cwdRequire function
which is exposed to the repl scope.

The deprecation warning was broken, so clearly no one is using this thing.

- reference to OpenSSL
- note about http-parser

This is needed in case the provided socket is not the default 'tcp4' type
(i.e. and needs different read/write/etc methods). With this patch, one can
call listenFD(sock, 'unix') to bind to existing UNIX domain sockets.

Setting this environmental variable to a non-zero integer will start all
REPL interfaces without readline. For use with rlwrap.

For eventual large character support.

Now that FD passing is in master, it'd be great to be able to use a received
socket (which has already had bind(2) and listen(2) called on it) to fire up a
new net.Server instance. This patch adds a net.Server.listenFD() method which
will start up the accept watcher on the provided FD.

Original commit: 75f0cf47

The current node has a bug where it will fail to pass the option argument
immediately preceding the first non-option argument to V8. That is the
--perf flag will be ignored by V8 when running 'node --perf script.js'.

The fs.readFile bug was hiding another bug that was causing this test
to pass, even so it was broken:

require.async("../fixtures/throws_error1") in test-module-loading.js

This patch fixes the original test by running _compile within a
try..catch block for _loadScript.

_loadScriptSync also had some useless (deprecated?) code for dealing
with module entry point exceptions. This code was also removed for
greater clarity.

fs.readFile was executing a callback in a try..catch context, which is
a problem in itself. To make matters worse, it would re-execute the
same callback if there was an execution.

This patch fixes both of these problems.