and CURLOPT_NEW_DIRECTORY_PERMS.  These control the premissions for files
and directories created on the remote server.  CURLOPT_NEW_FILE_PERMS
defaults to 0644 and CURLOPT_NEW_DIRECTORY_PERMS defaults to 0755

hash function for different hashes, and also expanded the default size for
the socket hash table used in multi handles to greatly enhance speed when
very many connections are added and the socket API is used.

chunked encoding (that also lacks "Connection: close"). It now simply
assumes that the connection WILL be closed to signal the end, as that is how
RFC2616 section 4.4 point #5 says we should behave.

http://curl.haxx.se/mail/lib-2007-06/0238.html, libcurl didn't properly do
no-body requests on FTP files on re-used connections properly, or at least
it didn't provide the info back in the header callback properly in the
subsequent requests.

tool reports and it was indeed a legitimate one and it is one fixed. It was
a use of a share without doing the proper locking first.

other authentication methods.  Terminate with a memory error.

from the #1739100 bug report

allocated when needed

server!

Caught by the auto-builds

HTTP CONNECT over a proxy

only affected sftp_sendquote() for the "chgrp/chmod/chown" commands.

This also fixed failure of test 614 on a system that previously failed.

done if the sys/poll.h file is missing, as we have seen machines with poll()
present but without the header file and machines that don't get HAVE_POLL
defined but that do have the sys/poll.h header file...

with the latest libssh2, the listing would be truncated at the symlink.
Fix by looping on LIBSSH2_ERROR_EAGAIN, like the rest of the calls.

uses the --ftp-create-dirs flag to control if cURL will try and create
directories that are specified in an upload path, but don't exist.

ie 0.14, that don't know about newer constants used in the state machine.

complicated work-around for 64bit HPUX compiles. We do the fix using inline
static functions to make them follow the header file properly and thus get
used fine in the test suite too etc.

  libssh2_sftp_shutdown() and libssh2_session_free() can now return
  LIBSSH2_ERROR_EAGAIN.

* Fix the _send() and _recv() return values so non-blocking works

suite, transfer fails in the real world.  So after connect set to blocking
as full non-blocking is migrated out.

LIBSSH2_APINO >= 200706012030.  More to come...

bug #1733955

the whole 'tsd' structure on exit in destroy_thread_sync_data().

memory leak, and it is the right thing to do.

* As of (LIBSSH2_APINO >= 200706012030) there are not *nb() functions
* As of (LIBSSH2_APINO >= 200706012030) most libssh2_*() functions
  can return LIBSSH2_ERROR_EAGAIN to indicate that the call would block.

To make the code work as previously, blocking, all the code has been
updated so that when (LIBSSH2_APINO >= 200706012030) it loops simulating
blocking.  This allows the existing code to function and not hold up
the upcoming release.