hostcache_timestamp_remove() should remove old *unused* entries from the
host cache, but it never checked whether the entry was actually in
use. This complements commit 030a2b8c.

Bug: http://curl.haxx.se/bug/view.cgi?id=1327

tftp_done() can get called with its TFTP state pointer still being NULL
on an early time-out, which caused a segfault when dereferenced.

Reported-by: Glenn Sheridan
Bug: http://curl.haxx.se/mail/lib-2014-01/0246.html

Make it possible to call
curl_easy_getinfo(curl, CURLINFO_CONTENT_LENGTH_DOWNLOAD, &filesize)
and related functions on remote sftp:// files, without downloading them.

Reported-by: Yingwei Liu
Bug: http://curl.haxx.se/mail/lib-2014-01/0139.html

This prevents sending a `Content-Length: -1` header, e.g this ocurred
with the following combination:

* standard HTTP POST (no chunked encoding),
* user-defined read function set,
* `CURLOPT_POSTFIELDSIZE(_LARGE)` NOT set.

With this fix it now behaves like HTTP PUT.

Make GnuTLS old and new consistent, specify the desired protocol, cipher
and certificate type in always in both modes. Disable insecure ciphers
as reported by howsmyssl.com. Honor not only --sslv3, but also the
--tlsv1[.N] switches.

Related Bug: http://curl.haxx.se/bug/view.cgi?id=1323

conversion from 'curl_off_t' to 'size_t', possible loss of data

Where curl_off_t is a 64-bit word and size_t is 32-bit - for example
with 32-bit Windows builds.

A follow up patch to commit d2671340 as _WIN32_WINNT_WIN2K and
_WIN32_WINNT_WIN2K may not be defined on all systems.

Following some auto build failures after commit c7a76bb0 changed
the preprocessor check to use _WIN32_WINNT.

1 - allow >31 bit max-age values

2 - don't overflow on extremely large max-age values when we add the
value to the current time

3 - make sure max-age takes precedence over expires as dictated by
RFC6265

Bug: http://curl.haxx.se/mail/lib-2014-01/0130.html
Reported-by: Chen Prog

Allow for chunked-encoding data to get parsed with only LF line endings.
This is allowed by browsers.

The now argument is unused by curl_getdate()

Starting with Visual Studio 2013 (VC12) and Windows 8.1 the
GetVersionInfoEx() function has been marked as deprecated and it's
return value atered. Updated connect.c and curl_sspi.c to use
VerifyVersionInfo() where possible, which has been available since
Windows 2000.

A transfer timeout could result in an error message such as "Operation
timed out after 3000 milliseconds with 19 bytes of -1 received".  This
patch removes the non-sensical "of -1" when the size of the transfer
is unknown, mirroring the logic in lib/transfer.c

By default even recent versions of OpenSSL support and accept both
"export strength" ciphers, small-bitsize ciphers as well as downright
deprecated ones.

This change sets a default cipher set that avoids the worst ciphers, and
subsequently makes https://www.howsmyssl.com/a/check no longer grade
curl/OpenSSL connects as 'Bad'.

Bug: http://curl.haxx.se/bug/view.cgi?id=1323
Reported-by: Jeff Hodges

With the recently added timeout "reminder" functionality, there's no
reason left for us to execute timeout code before the time is
ripe. Simplifies the handling too.

This will make the *TIMEOUT and *CONNECTTIMEOUT options more accurate
again, which probably is most important when the *_MS versions are used.

In multi_socket, make sure to update 'now' after having handled activity
on a socket.

BACKGROUND:

We have learned that on some systems timeout timers are inaccurate and
might occasionally fire off too early. To make the multi_socket API work
with this, we made libcurl execute timeout actions a bit early too if
they are within our MULTI_TIMEOUT_INACCURACY. (added in commit
2c72732e, present since 7.21.0)

Switching everything to the multi API made this inaccuracy problem
slightly more notable as now everyone can be affected.

Recently (commit 21091549) we tweaked that inaccuracy value to make
timeouts more accurate and made it platform specific. We also figured
out that we have code at places that check for fixed timeout values so
they MUST NOT run too early as then they will not trigger at all (see
commit be28223f and a691e044) - so there are definitately problems
with running timeouts before they're supposed to run. (We've handled
that so far by adding the inaccuracy margin to those specific timeouts.)

The libcurl multi_socket API tells the application with a callback that
a timeout expires in N milliseconds (and it explicitly will not tell it
again for the same timeout), and the application is then supposed to
call libcurl when that timeout expires. When libcurl subsequently gets
called with curl_multi_socket_action(...CURL_SOCKET_TIMEOUT...), it
knows that the application thinks the timeout expired - and alas, if it
is within the inaccuracy level libcurl will run code handling that
handle.

If the application says CURL_SOCKET_TIMEOUT to libcurl and _isn't_
within the inaccuracy level, libcurl will not consider the timeout
expired and it will not tell the application again since the timeout
value is still the same.

NOW:

This change introduces a modified behavior here. If the application says
CURL_SOCKET_TIMEOUT and libcurl finds no timeout code to run, it will
inform the application about the timeout value - *again* even if it is
the same timeout that it already told about before (although libcurl
will of course tell it the updated time so that it'll still get the
correct remaining time). This way, we will not risk that the application
believes it has done its job and libcurl thinks the time hasn't come yet
to run any code and both just sit waiting. This also allows us to
decrease the MULTI_TIMEOUT_INACCURACY margin, but that will be handled
in a separate commit.

A repeated timeout update to the application risk that the timeout will
then fire again immediately and we have what basically is a busy-loop
until the time is fine even for libcurl. If that becomes a problem, we
need to address it.

The net effect of this bug as it appeared to users, would be that
libcurl would timeout in the connect phase.

When disabling IPv6 use but still using getaddrinfo, libcurl would
wrongly not init the "hints" struct field in init_thread_sync() which
would subsequently lead to a getaddrinfo() invoke with a zeroed hints
with ai_socktype set to 0 instead of SOCK_STREAM. This would lead to
different behaviors on different platforms but basically incorrect
output.

This code was introduced in 483ff1ca, released in curl 7.20.0.

This bug became a problem now due to the happy eyeballs code and how
libcurl now traverses the getaddrinfo() results differently.

Bug: http://curl.haxx.se/mail/lib-2014-01/0061.html
Reported-by: Fabian Frank
Debugged-by: Fabian Frank

It turns out errSecDecode wasn't defined in Leopard's headers. So
we use the enum's value instead.

Bug: http://curl.haxx.se/mail/lib-2013-12/0150.html
Reported by: Abram Pousada

getpeername() doesn't work for UDP sockets since they're not connected

Reported-by: Priyanka Shah
Bug: http://curl.haxx.se/mail/archive-2014-01/0016.html

Removed some of the infof() calls that were added with the recent
pipeline improvements but they're not useful to the vast majority of
readers and the pipelining seems to fundamentaly work - the debugging
outputs can easily be added there if debugging these functions is needed
again.

AF_INET6 may not exist then

Patched-by: Iida Yosiaki
Bug: http://curl.haxx.se/bug/view.cgi?id=1322

When the requested authentication bitmask includes NTLM, we cannot
re-use a connection for another username/password as we then risk
re-using NTLM (connection-based auth).

This has the unfortunate downside that if you include NTLM as a possible
auth, you cannot re-use connections for other usernames/passwords even
if NTLM doesn't end up the auth type used.

Reported-by: Paras S
Patched-by: Paras S
Bug: http://curl.haxx.se/mail/lib-2014-01/0046.html

A regression introduced in 7f3b87d8 (present in the 7.21.4 release)
broke the total parser. Now skip the whitespace and the digits.

Reported-by: Justin Maggard
Bug: http://curl.haxx.se/mail/lib-2014-01/0019.html