Running tests\libtest\libntlmconnect.exe reveals a 1 byte (!) leak in
./lib/curl_ntlm_msgs.c:

perl ..\memanalyze.pl c:memdebug.curl
Leak detected: memory still allocated: 1 bytes
At 9771e8, there's 1 bytes.
allocated by curl_ntlm_msgs.c:399

Snippet from curl_ntlm_msgs.c:
   /* setup ntlm identity's domain and length */
   dup_domain.tchar_ptr = malloc(sizeof(TCHAR) * (domlen + 1));

(my domlen == 0).

'dup_domain.tbyte_ptr' looks to be freed in Curl_ntlm_sspi_cleanup() via
'ntlm->identity.Domain'. But I see no freeing of 'dup_domain.tchar_ptr'.

This bug report properly identified that when doing SMTP and aborting
the transfer with a callback, it must be considered aborted prematurely
by the code to avoid QUIT etc to be attempted as that would cause a
hang.

The new test case 1507 verifies this behavior.

Reported by: Patricia Muscalu
Bug: http://curl.haxx.se/bug/view.cgi?id=1184

It turns out that Leopard (OS X 10.5) doesn't have constants for the ECDH
ciphers in its headers, so the cases for them have been taken out of the
build when building under Leopard. Also added a standard function for
getting a string description of a SecCertificateRef.

Modified imap_authenticate() to add support for sending the initial
response with the AUTHENTICATE command, as per RFC4959.

Changed the SMTP_AUTH_PASSWD state constant to SMTP_AUTH_LOGIN_PASSWD to
better describe the state as the second part of an AUTH LOGIN command,
as well as for consistency with the imap and pop3 modules.

Introduced detection of the SASL-IR capability, in order to add support
for sending the initial response with the AUTHENTICATE command, as per
RFC4959.

This reverts commit 0e66d587.

Updated the automatic response tag generation to follow the examples
given in RC3501, which list a 4 character string such as A001, A002,
etc.

As a unique identifier should be generated for each command the string
generation is based on the connection id and the incrementing command
id.

This is untested, but ought to be enough to still allow it
to work automatically when the entire curl source tree is
dropped into a full Android source tree.

VC6 is _very_ old and we provide working makefiles even for that
compiler. Users who build with the IDE never use that method and project
file anyway and it was just lingering in the root dir.

Renamed a couple of variables and updated some comments in
preparation for upcoming command id / response tag change.

In an attempt to clear up misc files from the root dir

We don't need them and I doubt many people used them. We also don't have
any configs for other editors and we wouldn't want that.

To help maintain the readability of the code in imap.c, pop3.c and
smtp.c moved the starttls code into state_starttls() functions.

Added IDN and HTTP data compression as they were left out of the
document until now.

Added notes for qssl, schannel and Secure Transport supporting SSLv2,
Secure Transport supports NTLM, and axTLS does not support SSLv3.

There was also a typo; "AUTH TSL" should be "AUTH TLS".

When negotiating SASL DIGEST-MD5 authentication, the function
Curl_sasl_create_digest_md5_message() uses the data provided from the
server without doing the proper length checks and that data is then
appended to a local fixed-size buffer on the stack.

This vulnerability can be exploited by someone who is in control of a
server that a libcurl based program is accessing with POP3, SMTP or
IMAP. For applications that accept user provided URLs, it is also
thinkable that a malicious user would feed an application with a URL to
a server hosting code targetting this flaw.

Bug: http://curl.haxx.se/docs/adv_20130206.html

Removed whitespace introduced in commit 5f8f20f5 that caused
formatting issues when generating the website docs.

- rename post-config-vms.h to setup-vms.h
- move its inclusion into proper location in curl_setup.h

- remove multiple declarations of vms_show and add comments

- remove header inclusion already done in curl_setup_once.h

Added SMTP, SMTPS, POP3, POP3S, IMAP and IMAPS features.

Corrected comment clarification made in commit 167717b8.

8 more bug fixes mentioned

config_h.com is a new file that generates a config.h file based on the
curl_config.h.in file and a quick scan of the configure script.  This is
actually a generic procedure that is shared with other VMS packages.

The existing pre-built config-vms.h had over 100 entries that were not
correct and in some cases conflicted with the build options available in
the build_vms.com.

generate_config_vms_h_curl.com is a helper procedure to the
config_h.com.  It covers the cases that the generic config_h.com is not
able to figure out, and accepts input from the build_vms.com procedure.

build_curlbuild_h.com is a new file to generate the curlbuild.h file
that Curl is now using when it is using a curl_config.h file.

post-config-vms.h is a new file that is needed to provide VMS specific
definitions, and most of them need to be set before the system header
files are included.

The VMS build procedure is fixed:

   1. Fixed to link in the correct HP ssl library.
   2. Fixed to detect if HP Kerberos is installed.
   3. Fixed to detect if HP LDAP is installed.
   4. Fixed to detect if gnv$libzshr is installed.
   5. Simplified the input parameter parsing to not use a loop.
   6. Warn that 64 bit pointer option support is not complete
      in comments.
   7. Default to IEEE floating if platform supports it so
      resulting libcurl will be compatible with other
      open source projects on VMS.
   8. Default to LARGEFILE if platform supports it.
   9. Default to enable SSL, LDAP, Kerberos, libz
      if the libraries are present.
   10. Build with exact case global symbols for libcurl.
   11. Generate linker option file needed.
   12. Compiler list option only commonly needed items.
   13. fulllist option for those who really want it.
   14. Create debug symbol file on Alpha, IA64.

By doing this unconditionally, we infer a simpler and more defined
behavior. This also has the upside that test 1021 no longer fails for me
even if I run with valgrind.

Also fixed some wrong comments.

Tidied up the comments in the endofresp() functions to be more
meaningful prior to release.

According KB975858 this flag may cause problems on Windows 7 and
Windows Server 2008 R2 systems. Extended error information is not
currently used by libcurl and therefore not a requirement.

The flag may improve the SSL-connection shutdown in case of an
error. This means it might be a good improvement in the future.

Fixes bug/issue #1187 - thanks for the report

The 56b7c87c change left a case where a good sockfd was not copied to
*sockp before returning with CURLE_OK

pointed out by clang-analyzer