Having the includes in src/node_buffer.h outside of the include guard
is not really harmful but it's inconsistent with other header files.

Conflicts:
	AUTHORS
	ChangeLog
	lib/dgram.js
	lib/http.js
	src/node_crypto.cc
	src/node_version.h

Add range checks for the offset, length and port arguments to
dgram.Socket#send().  Fixes the following assertion:

    node: ../../src/udp_wrap.cc:264: static v8::Handle<v8::Value>
    node::UDPWrap::DoSend(const v8::Arguments&, int): Assertion
    `offset < Buffer::Length(buffer_obj)' failed.

And:

    node: ../../src/udp_wrap.cc:265: static v8::Handle<v8::Value>
    node::UDPWrap::DoSend(const v8::Arguments&, int): Assertion
    `length <= Buffer::Length(buffer_obj) - offset' failed.

Interestingly enough, a negative port number was accepted until now but
silently ignored.  (In other words, it would send the datagram to a
random port.)

This commit exposed a bug in the simple/test-dgram-close test which
has also been fixed.

This is a back-port of commit 41ec6d05 from the master branch.

Fixes #6025.

On windows, libuv will immediately make a `ReadConsole` call (in the
thread pool) when a 'flowing' `uv_tty_t` handle is switched to
line-buffered mode. That causes an immediate issue for some users,
since libuv can't cancel the `ReadConsole` operation on Windows 8 /
Server 2012 and up if the program switches back to raw mode later.

But even if this will be fixed in libuv at some point, it's better to
avoid the overhead of starting work in the thread pool and immediately
cancelling it afther that.

See also f34f1e30, where the same change is made for the opposite
flow, e.g. move `resume()` after `_setRawMode(true)`.

Fixes #5927

This is a backport of dfb0461c (see #5930) to the v0.10 branch.

On windows, libuv will immediately make a `ReadConsole` call (in the
thread pool) when a 'flowing' `uv_tty_t` handle is switched to
line-buffered mode. That causes an immediate issue for some users,
since libuv can't cancel the `ReadConsole` operation on Windows 8 /
Server 2012 and up if the program switches back to raw mode later.

But even if this will be fixed in libuv at some point, it's better to
avoid the overhead of starting work in the thread pool and immediately
cancelling it afther that.

See also f34f1e30, where the same change is made for the opposite
flow, e.g. move `resume()` after `_setRawMode(true)`.

Fixes #5927
Closes #5930

For several encodings the byte length is simple arithmetic. Don't call
into C++ in those cases.

* v8: back-port fix for CVE-2013-2882

* npm: Upgrade to 1.3.8

* crypto: fix assert() on malformed hex input (Ben Noordhuis)

* crypto: fix memory leak in randomBytes() error path (Ben Noordhuis)

* events: fix memory leak, don't leak event names (Ben Noordhuis)

* http: Handle hex/base64 encodings properly (isaacs)

* http: improve chunked res.write(buf) performance (Ben Noordhuis)

* stream: Fix double pipe error emit (Eran Hammer)

This is the conceptual back-port of commit ec548734 from the master
branch.

RandomBytes() allocated memory with new[] which was then handed off to
Buffer::Use() which eventually releases it again with free().

Mixing the two is technically a violation of the spec and besides, it's
generally frowned upon.

It's faster, because it doesn't have to check that each char is in the
ASCII plane.

In other Writable streams, the 'finish' event means that all of the data
was written, and flushed to the underlying system.

The 'prefinish' event means that end() was called, and all of the data
was processed, but not necessarily completely flushed.

This change brings the http OutgoingMessage classes more in sync with
the other Writable classes throughout Node.

Unfortunately, this change highlights an issue with http
IncomingMessages, where the _dump() method will not actually pull the
data off the wire.  This is a minor issue that is typically only
relevant in test cases, and will be addressed in the next commit.

This removes a dubious performance "optimization" where strings body
chunks were concatenated to one another (and to the headers) without any
regard for their encoding.

This is a backport of 6d3d60aced39d59eaa5e705b7d822c227d0d3dae for
v0.10.

The format is [key,value,key,value,...] because that seems to have the
lowest overhead.

Close #4844

Close #5839

Commit 0aa13357 changes the way timeout events are dispatched. Update
two tests that still used the old way.

These checks are now done in C++ and don't need to also be checked in
JS. Also remove a couple unused variables.

Achieve a minor speed-up by looking up the timeout callback on the timer
object by using an array index rather than a named property.

Gives a performance boost of about 1% on the misc/timers benchmarks.

Internal helper function for dispatching by array index rather than
named property.

Don't check for the key first before retrieving it. Just fetch it and
check that it has the type we expect.

And build/include, build/include_alpha and readability/streams
probably too, though those are currently unused and therefore untested.

Seems to have escaped the Eye of Sauron^WFedor in commit 8e29ce9f.

Fix sudden uncatchable ECONNRESETs, when using https server.

Use array indices rather than named properties to store callbacks on
the HTTPParser object.  Speeds up the http benchmarks by a few percent.

Don't lazy-init the oncomplete symbol string in the After() function,
it has already been created at module init time.

Move the 'free FSReqWrap data?' logic into the class itself.