Introduced in commit 2a4ee0d2 sending of data via the FILE
protocol would always return CURLE_WRITE_ERROR regardless of whether
CURL_WRITEFUNC_PAUSE was returned from the callback function or not.

Make sure that we detect such attempts and return a proper error code
instead of silently handling this in problematic ways.

Updated the documentation to mention this limitation.

Bug: http://curl.haxx.se/bug/view.cgi?id=1286

Previously this memdebug free() replacement didn't properly work with a
NULL argument which has made us write code that avoids calling
free(NULL) - which causes some extra nuisance and unnecessary code.
Starting now, we should allow free(NULL) even when built with the
memdebug system enabled.

free(NULL) is permitted by POSIX

free() itself allows a NULL input but our memory debug system requires
Curl_safefree() to be used instead when a "legitimate" NULL may be freed. Like
in the code here.

Pointed-out-by: Steve Holme

... since pthread_t may be non-scalar and/or may represent a real thread
with scalar 0.

Bug: http://curl.haxx.se/bug/view.cgi?id=1314

If a user indicated they preferred to authenticate using a SASL
mechanism, but SASL authentication wasn't supported by the server, curl
would always fall back to clear text when CAPABILITY wasn't supported,
even though the user didn't want to use this.

If a user indicated they preferred to authenticate using APOP or a SASL
mechanism, but neither were supported by the server, curl would always
fall back to clear text when CAPA wasn't supported, even though the
user didn't want to use this.

This also fixes the auto build failure caused by commit 6f2d5f05.

Fixes commit 1deac31e

This commit replaces that of 9f260b5d because according to RFC-2449,
section 6, there is no APOP capability "...even though APOP is an
optional command in [POP3].  Clients discover server support of APOP by
the presence in the greeting banner of an initial challenge enclosed in
angle brackets."

SASL downgrade tests: 833, 835, 879, 881, 935 and 937 would fail as
they contained a minus sign in their authentication mechanism and this
would be missed by the custom reply parser.

The FILE:// code doesn't support this option - and it doesn't make sense
to support it as long as it works as it does since then it'd only block
even longer.

But: setting CURLOPT_MAX_RECV_SPEED_LARGE would make the transfer first
get done and then libcurl would wait until the average speed would get
low enough. This happened because the transfer happens completely in the
DO state for FILE:// but then it would still unconditionally continue in
to the PERFORM state where the speed check is made.

Starting now, the code will skip from DO_DONE to DONE immediately if no
socket is set to be recv()ed or send()ed to.

Bug: http://curl.haxx.se/bug/view.cgi?id=1312
Reported-by: Mohammad AlSaleh

Use of uninitialized value $FTPARG in concatenation (.) or string at
line 3255.

Added SASL CANCELLATION keywords to differentiate these tests from the
upcoming SASL downgrade tests.

The comment in the code mentions the zero terminating after having
copied data, but it mistakingly zero terminated the source data and not
the destination! This caused the test 864 problem discussed on the list:

http://curl.haxx.se/mail/lib-2013-12/0113.html


Signed-off-by: Daniel Stenberg <daniel@haxx.se>

This reverts commit 727d798d.