CHANGES

                                  _   _ ____  _
                              ___| | | |  _ \| |
                             / __| | | | |_) | |
                            | (__| |_| |  _ <| |___
                             \___|\___/|_| \_\_____|

                                  Changelog

Daniel Stenberg (12 Aug 2009)
- Carsten Lange reported a bug and provided a patch for TFTP upload and the
  sending of the TSIZE option. I don't like fixing bugs just hours before
  a release, but since it was broken and the patch fixes this for him I decided
  to get it in anyway.

Daniel Stenberg (11 Aug 2009)
- Peter Sylvester made the HTTPS test server use specific certificates for
  each test, so that the test suite can now be used to actually test the
  verification of cert names etc. This made an error show up in the OpenSSL-
  specific code where it would attempt to match the CN field even if a
  subjectAltName exists that doesn't match. This is now fixed and verified
  in test 311.

- Benbuck Nason posted the bug report #2835196
  (http://curl.haxx.se/bug/view.cgi?id=2835196), fixing a few compiler
  warnings when mixing ints and bools.

Daniel Fandrich (10 Aug 2009)
- Fixed a memory leak in the FTP code and an off-by-one heap buffer overflow.

Daniel Fandrich (9 Aug 2009)
- Fixed some memory leaks in the command-line tool that caused most of the
  torture tests to fail.

Daniel Stenberg (2 Aug 2009)
- Curt Bogmine reported a problem with SNI enabled on a particular server. We
  should introduce an option to disable SNI, but as we're in feature freeze
  now I've addressed the obvious bug here (pointed out by Peter Sylvester): we
  shouldn't try to enable SNI when SSLv2 or SSLv3 is explicitly selected.
  Code for OpenSSL and GnuTLS was fixed. NSS doesn't seem to have a particular
  option for SNI, or are we simply not using it?

Daniel Stenberg (1 Aug 2009)
- Scott Cantor posted the bug report #2829955
  (http://curl.haxx.se/bug/view.cgi?id=2829955) mentioning the recent SSL cert
  verification flaw found and exploited by Moxie Marlinspike. The presentation
  he did at Black Hat is available here:
  https://www.blackhat.com/html/bh-usa-09/bh-usa-09-archives.html#Marlinspike

  Apparently at least one CA allowed a subjectAltName or CN that contain a
  zero byte, and thus clients that assumed they would never have zero bytes
  were exploited to OK a certificate that didn't actually match the site. Like
  if the name in the cert was "example.com\0theatualsite.com", libcurl would
  happily verify that cert for example.com.

  libcurl now better uses the length of the extracted name, not using the zero
  termination for getting the string length.

  This fixing only made and needed in OpenSSL interfacing code.

- Tanguy Fautre pointed out that OpenSSL's function RAND_screen() (present
  only in some OpenSSL installs - like on Windows) isn't thread-safe and we
  agreed that moving it to the global_init() function is a decent way to deal
  with this situation.

- Alexander Beedie provided the patch for a noproxy problem: If I have set
  CURLOPT_NOPROXY to "*", or to a host that should not use a proxy, I actually
  could still end up using a proxy if a proxy environment variable was set.

Daniel Stenberg (27 Jul 2009)
- All the quote options (CURLOPT_QUOTE, CURLOPT_POSTQUOTE and
  CURLOPT_PREQUOTE) now accept a preceeding asterisk before the command to
  send when using FTP, as a sign that libcurl shall simply ignore the response
  from the server instead of treating it as an error. Not treating a 400+ FTP
  response code as an error means that failed commands will not abort the
  chain of commands, nor will they cause the connection to get disconnected.

Daniel Stenberg (26 Jul 2009)
- Johan van Selst posted bug report #2825989
  (http://curl.haxx.se/bug/view.cgi?id=2825989) pointing out that
  OpenSSL-powered libcurl didn't support the SHA-2 digest algorithm, and
  provided the solution too: to use OpenSSL_add_all_algorithms() in addition
  to the older SSLeay_* alternative. OpenSSL_add_all_algorithms was added in
  OpenSSL 0.9.5

Daniel Stenberg (23 Jul 2009)
- Added CURLOPT_SSH_KNOWNHOSTS, CURLOPT_SSH_KEYFUNCTION, CURLOPT_SSH_KEYDATA.
  They introduce known_host support for SSH keys to libcurl. See docs for
  details. Note that this feature depends on a new enough libssh2 version, to
  be supported in libssh2 1.2 and later (or current git repo at this time).

Michal Marek (22 Jul 2009)
- David Binderman found a memory and fd leak in lib/gtls.c:load_file()
  (https://bugzilla.novell.com/523919). When looking at the code, I found that
  also the ptr pointer can leak.

Kamil Dudka (20 Jul 2009)
- Claes Jakobsson improved the support for client certificates handling in
  NSS-powered libcurl. Now the client certificates can be selected
  automatically by a NSS built-in hook. Additionally pre-login to all PKCS11
  slots is no more performed. It used to cause problems with HW tokens.

- Fixed reference counting for NSS client certificates. Now the PEM reader
  module should be always properly unloaded on Curl_nss_cleanup(). If the
  unload fails though, libcurl will try to reuse the already loaded instance.

Daniel Fandrich (15 Jul 2009)
- Added nonblock.c to the non-automake makefiles (note that the dependencies
  in the Watcom makefiles aren't quite correct).

Michal Marek (15 Jul 2009)
- Changed the description of CURLINFO_OS_ERRNO to make it clear that the
  errno is not reset on success.

Guenter Knauf (14 Jul 2009)
- renamed generated config.h to curl_config.h to avoid any future clashes
  with config.h from other projects.
  
Daniel Stenberg (9 Jul 2009)
- Eric Wong introduced curlx_nonblock() that the curl tool now (re-)uses for
  setting a file descriptor non-blocking. Used by the functionality Eric
  himself brough on June 15th.

Daniel Stenberg (8 Jul 2009)
- Constantine Sapuntzakis posted bug report #2813123
  (http://curl.haxx.se/bug/view.cgi?id=2813123) and an a patch that fixes the
  problem:

  Url A is accessed using auth. Url A redirects to Url B (on a different
  server0. Url B reuses a persistent connection. Url B has auth, even though
  it's on a different server.

  Note: if Url B does not reuse a persistent connection, auth is not sent.

  reason:

  data->state.first_host is not initialized becuase Curl_http_connect is not
  called when a connection is reused.

  Solution:

  move initialization of data->state.first_host to Curl_http. No code before
  Curl_http uses data->state.first_host anyway.

Guenter Knauf (4 Jul 2009)
- Markus Koetter provided a patch to avoid getnameinfo() usage which broke a
  couple of both IPv4 and IPv6 autobuilds.

Daniel Stenberg (29 Jun 2009)
- Markus Koetter made CURLOPT_FTPPORT (and curl's -P/--ftpport) support a port
  range if given colon-separated after the host name/address part. Like
  "192.168.0.1:2000-10000"

- Modified the separators used for CURLOPT_CERTINFO in multi-part outputs. I
  don't know how they got wrong in the first place, but using this output
  format makes it possible to quite easily separate the string into an array
  of multiple items.

Daniel Fandrich (16 June 2009)
- Added a few more compiler warning options for gcc.

Daniel Stenberg (16 Jun 2009)
- Reuven Wachtfogel made curl -o - properly produce a binary output on windows
  (no newline translations). Use -B/--use-ascii if you rather get the ascii
  approach.

Michal Marek (16 Jun 2009)
- When doing non-anonymous ftp via http proxies and the password is not
  provided in the url, add it there (squid needs this).

Daniel Stenberg (15 Jun 2009)
- Eric Wong's patch:

  This allows curl(1) to be used as a client-side tunnel for arbitrary stream
  protocols by abusing chunked transfer encoding in both the HTTP request and
  HTTP response.  This requires server support for sending a response while a
  request is still being read, of course.

  If attempting to read from stdin returns EAGAIN, then we pause our sender.
  This leaves curl to attempt to read from the socket while reading from stdin
  (and thus sending) is paused.

  This change was needed to allow successfully tunneling the git protocol over
  HTTP (--no-buffer is needed, as well).

Patrick Monnerat (15 Jun 2009)
- Replaced use of standard C library rand()/srand() by our own pseudo-random
  number generator.

Yang Tse (11 Jun 2009)
- I adapted testcurl script to allow building test harness programs when
  cross-compiling for a *-*-mingw* host.

Daniel Stenberg (10 Jun 2009)
- Fabian Keil ran clang on the (lib)curl code, found a bunch of warnings and
  contributed a range of patches to fix them.

Yang Tse (10 Jun 2009)
- I introduced configure script option --enable-curldebug which now allows
  the decoupled enabling or disabling of the curl debug memory tracking
  feature from the --enable-debug option which no longer controls this.

  curl --version will list 'Debug' feature for debug enabled builds, and
  will list 'TrackMemory' feature for curl debug memory tracking capable
  builds. These features are independent and can be controlled when running
  the configure script. When --enable-debug is given both features will be
  enabled, unless some restriction prevents memory tracking from being used.

  Internally, definition of preprocessor symbol DEBUGBUILD restricts code
  which is only compiled for debug enabled builds. And symbol CURLDEBUG is
  used to differentiate code which is _only_ used for memory tracking.

Yang Tse (9 Jun 2009)
- Daniel Steinberg pointed out that Curl_FormInit() in formdata.c was not
  initializing the fread callback pointer and this triggered a compiler
  warning, also provided a friendly suggestion on how to fix it.

Daniel Stenberg (8 Jun 2009)
- Claes Jakobsson provided a patch for libcurl-NSS that fixed a bad refcount
  issue with client certs that caused issues like segfaults.
  http://curl.haxx.se/mail/lib-2009-05/0316.html

- Triggered by bug report #2798852 and the patch in there, I fixed configure
  to detect gnutls build options with pkg-config only and not libgnutls-config
  anymore since GnuTLS has stopped distributing that tool. If an explicit path
  is given to configure, we will instead guess on how to link and use that
  lib. I did not use the patch from the bug report.

Yang Tse (8 Jun 2009)
- Igor Novoseltsev adjusted Makefile.vxworks to get sources and headers
  included from Makefile.inc, and provided docs\INSTALL VxWorks section.

- I removed buildconf.bat from release and daily snapshot archives. This
  file is only for CVS tree checkout builds.

Daniel Stenberg (8 Jun 2009)
- Eric Wong fixed --no-buffer to actually switch off output buffering. Been
  broken since 7.19.0

Bill Hoffman (6 Jun 2009)
- Added some cmake docs and fixed socklen_t in the build.

Yang Tse (5 Jun 2009)
- John E. Malmberg provided VMS specific patch: "This fixes an existing bug
  in urlglob.c where it was not converting the Curl Unix exit code to a VMS
  DCL compatible exit code.  This fix required the enhancement described next.
  This also adds an enhancement to main.c so that when curl is run under a
  Unix shell like Bash on VMS, it will return the standard Unix exit codes
  and messages." And another patch for docs/examples.

  I introduced os-specific.c and os-specific.h for use in curl tool code
  and adjusted John E. Malmberg's patch placement to use these new files
  as an effort to prevent main.c from growing ad infinitum. Code already
  existing in main.c which is OS specific should be moved into these files.

Daniel Stenberg (4 June 2009)
- Setting the Content-Length: header from your app when you do a POST or PUT
  is almost always a VERY BAD IDEA. Yet there are still apps out there doing
  this, and now recently it triggered a bug/side-effect in libcurl as when
  libcurl sends a POST or PUT with NTLM, it sends an empty post first when it
  knows it will just get a 401/407 back. If the app then replaced the
  Content-Length header, it caused the server to wait for input that libcurl
  wouldn't send. Aaron Oneal reported this problem in bug report #2799008
  (http://curl.haxx.se/bug/view.cgi?id=2799008) and helped us verify the fix.

Yang Tse (4 Jun 2009)
- Igor Novoseltsev provided patches and information, that after some
  adjustments to better fit curl's way of doing things, have resulted
  in the posibility of building libcurl for VxWorks.

Daniel Fandrich (2 June 2009)
- Checked in a Google Android make file. To use it, you must first
  create a config.h file by running configure in the Android environment,
  which doesn't seem to be easy to do. If no easy way can be found, a
  static config-android.h may need to be created and checked in to the
  libcurl source tree.

Daniel Stenberg (1 June 2009)
- Claes Jakobsson fixed the configure script to better find and use NSS
  without pkg-config.

Yang Tse (1 Jun 2009)
- John E. Malmberg provided a VMS specific clean-up for curl.h, and pointed
  out that the configure script was failing to detect the timeval struct on
  VMS when building with _XOPEN_SOURCE_EXTENDED undefined due to definition
  taking place in socket.h instead of time.h.  I have adjusted configure
  script to also include this header when checking struct timeval.

Daniel Stenberg (27 May 2009)
- Frank McGeough provided a small OpenSSL #include fix to make libcurl compile
  fine with Nokia 5th edition 1.0 SDK for Symbian.

- Andre Guibert de Bruet found a call to a OpenSSL function that didn't check
  for a failure properly.

- Mike Crowe pointed out that setting CURLOPT_USERPWD to NULL used to clear
  the auth credentials back in 7.19.0 and earlier while now you have to set ""
  to get the same effect. His patch brings back the ability to use NULL.

- Claes Jakobsson fixed libcurl-NSS to build fine even without the
  PK11_CreateGenericObject() function.

Daniel Stenberg (25 May 2009)
- bug report #2796358 (http://curl.haxx.se/bug/view.cgi?id=2796358) pointed
  out that the cookie parser would leak memory when it parses cookies that are
  received with domain, path etc set multiple times in the same header. While
  such a cookie is questionable, they occur in the wild and libcurl no longer
  leaks memory for them. I added such a header to test case 8.

Daniel Fandrich (22 May 2009)
- Removed some obsolete digest code that caused a valgrind error in test 551.

Daniel Fandrich (20 May 2009)
- Added "non-existing host" test keywords to make it easy to skip those
  tests on machines that have broken DNS configurations (such as
  those configured to use OpenDNS).

Daniel Stenberg (19 May 2009)
- Kamil Dudka brought the patch from the Redhat bug entry
  https://bugzilla.redhat.com/show_bug.cgi?id=427966 which was libcurl closing
  a bad file descriptor when closing down the FTP data connection.  Caolan
  McNamara seems to be the original author of it.

Version 7.19.5 (18 May 2009)

Daniel Stenberg (17 May 2009)
- James Bursa posted a patch to the mailing list that fixed a problem with
  no_proxy which made it not skip the proxy if the URL entered contained a
  user name. I added test case 1101 to verify.

Daniel Stenberg (11 May 2009)
- Balint Szilakszi reported a memory leak when libcurl did gzip decompression
  of streams that had some parts (legitimately) missing. We now provide and use
  a proper cleanup function for the content encoding submodule.
  http://curl.haxx.se/mail/lib-2009-05/0092.html

- Kamil Dudka provided a fix for libcurl-NSS reported by Michael Cronenworth
  at https://bugzilla.redhat.com/show_bug.cgi?id=453612#c12

  If an incorrect password is given while loading a private key, libcurl ends
  up in an infinite loop consuming memory. The bug is critical.

- I fixed the problem with doing NTLM, POST and then following a 302 redirect,
  as reported by Ebenezer Ikonne (on curl-users) and Laurent Rabret (on
  curl-library). The transfer was mistakenly marked to get more data to send
  but since it didn't actually have that, it just hung there...

Daniel Stenberg (10 May 2009)
- Andre Guibert de Bruet correctly pointed out an over-alloc with one wasted
  byte in the digest code.

Yang Tse (9 May 2009)
- Removed DOS and TPF package's subdirectory Makefile.am, it was only used
  to include some files in the distribution tarball serving no other purpose.
  Files from the DOS and TPF subdirectories are now included in the EXTRA_DIST
  of the Makefile in the parent subdirectory.

Yang Tse (8 May 2009)
- Changed host name literal in several tests to one under the haxx.se domain.

- Renamed vc6 workspace and project files to avoid filename clash when used
  for conversion to later VS versions.

Daniel Stenberg (8 May 2009)
- Constantine Sapuntzakis fixed bug report #2784055
  (http://curl.haxx.se/bug/view.cgi?id=2784055) identifying a problem to
  connect to SOCKS proxies when using the multi interface. It turned out to
  almost not work at all previously. We need to wait for the TCP connect to
  be properly verified before doing the SOCKS magic.

  There's still a flaw in the FTP code for this.

Daniel Stenberg (7 May 2009)
- Made the SO_SNDBUF setting for the data connection socket for ftp uploads as
  well. See change 28 Apr 2009.

Yang Tse (7 May 2009)
- Fixed an issue affecting FTP transfers, introduced with the transfer.c
  patch committed May 4.

Daniel Stenberg (7 May 2009)
- Man page *roff problems fixed thanks to input from Colin Watson. Problems
  reported in the Debian package.

- Vijay G filed bug report #2723236
  (http://curl.haxx.se/bug/view.cgi?id=2723236) identifying a problem with
  libcurl's TFTP code and its lack of dealing with the OACK packet.

Yang Tse (5 May 2009)
- Fixed the --ftp-port address of test #251 to the CLIENTIP address, and
  reverted the change affecting test suite harness committed 4 May.

Daniel Stenberg (5 May 2009)
- Inspired by Michael Smith's session id fix for OpenSSL, I did the
  corresponding fix in the GnuTLS code: make sure to store the new session id
  in case the previous re-used one is rejected.

Daniel Stenberg (4 May 2009)
- Michael Smith posted bug report #2786255
  (http://curl.haxx.se/bug/view.cgi?id=2786255) with a patch, identifying how
  libcurl did not deal with SSL session ids properly if the server rejected a
  re-use of one. Starting now, it will forget the rejected one and remember
  the new. This change was for OpenSSL only, it is likely that other SSL lib
  code needs similar fixes.

Yang Tse (4 May 2009)
- Applied David McCreedy's "transfer.c fixes for CURL_DO_LINEEND_CONV and
  non-ASCII platform HTTP requests" patch addressing two HTTP PUT problems:
  1) On non-ASCII platforms not all of the protocol portions of the PUT are
  being translated to ASCII.  2) On all platforms the line endings of part of
  the protocol portions are mangled from CRLF to CRCRLF if data->set.crlf or
  data->set.prefer_ascii are set (depending on CURL_DO_LINEEND_CONV).

- Applied David McCreedy's patch to fix test suite harness to allow test FTP
  server and client on different machines, providing FTP client address when
  running the FTP test server.

Daniel Fandrich (3 May 2009)
- Added and disabled test case 563 which shows KNOWN_BUGS #59.  The bug
  report failed to mention that a proxy must be used to reproduce it.

Yang Tse (2 May 2009)
- Use a build-time configured curl_socklen_t data type instead of socklen_t.

Yang Tse (1 May 2009)
- Applied David McCreedy's patches "TPF-platform specific changes to various
  files" and "http.c fix to Curl_proxyCONNECT for non-ASCII platforms", the
  former with minor edits.

Daniel Stenberg (30 Apr 2009)
- I was going to fix issue #59 in KNOWN_BUGS

  If the CURLOPT_PORT option is used on an FTP URL like
  "ftp://example.com/file;type=A" the ";type=A" is stripped off.

  I added test case 562 to verify, only to find out that I couldn't repeat
  this bug so I hereby consider it not a bug anymore!

Daniel Stenberg (29 Apr 2009)
- Based on bug report #2723219 (http://curl.haxx.se/bug/view.cgi?id=2723219)
  I've now made TFTP "connections" not being kept for re-use within libcurl.
  TFTP is UDP-based so the benefit was really low (if even existing) to begin
  with so instead of tracking down to fix this problem we instead removed the
  re-use. I also enabled test case 1099 that I wrote a few days ago to verify
  that this change fixes the reported problem.

Daniel Stenberg (28 Apr 2009)
- Constantine Sapuntzakis filed bug report #2783090
  (http://curl.haxx.se/bug/view.cgi?id=2783090) pointing out that on windows
  we need to grow the SO_SNDBUF buffer somewhat to get really good upload
  speeds. http://support.microsoft.com/kb/823764 has the details. Friends
  confirmed that simply adding 32 to CURL_MAX_WRITE_SIZE is enough.

- Bug report #2709004 (http://curl.haxx.se/bug/view.cgi?id=2709004) by Tim
  Chen pointed out how curl couldn't upload with resume when reading from a
  pipe.

  This ended up with the introduction of a new return code for the
  CURLOPT_SEEKFUNCTION callback that basically says that the seek failed but
  that libcurl may try to resolve the situation anyway. In our case this means
  libcurl will attempt to instead read that much data from the stream instead
  of seeking and that way curl can now upload with resume when data is read
  from a stream!

Daniel Stenberg (26 Apr 2009)
- Bug report #2779733 (http://curl.haxx.se/bug/view.cgi?id=2779733) by Sven
  Wegener pointed out that CURLINFO_APPCONNECT_TIME didn't work with the multi
  interface and provided a patch that fixed the problem!

Daniel Stenberg (24 Apr 2009)
- Kamil Dudka fixed another NSS-related leak when client certs were used.

- Bug report #2779245 (http://curl.haxx.se/bug/view.cgi?id=2779245) by Rainer
  Koenig pointed out that the man page didn't tell that the *_proxy
  environment variables can be specified lower case or UPPER CASE and the
  lower case takes precedence,

Daniel Fandrich (21 Apr 2009)
- Added new libcurl source files to Amiga, RiscOS and VC6 build files.

Yang Tse (21 Apr 2009)
- Moved potential inclusion of system's malloc.h and memory.h header files to
  setup_once.h.  Inclusion of each header file is based on the definition of
  NEED_MALLOC_H and NEED_MEMORY_H respectively.

  Renamed libcurl's memory.h to curl_memory.h

Daniel Stenberg (20 Apr 2009)
- Leanic Lefever reported a crash and did some detailed research on why and
  how it occurs (http://curl.haxx.se/mail/lib-2009-04/0289.html). The
  conclusion was that if an error is detected and Curl_done() is called for
  the connection, ftp_done() could at times return another error code that
  then would take precedence and that new code confused existing logic that
  works for the first error code (CURLE_SEND_ERROR) only.

- Gisle Vanem noticed that --libtool would produce bogus strings at times for
  OBJECTPOINT options. Now we've introduced a new function - my_setopt_str -
  within the app for setting plain string options to avoid the risk of this
  mistake happening.

Daniel Stenberg (17 Apr 2009)
- Pramod Sharma reported and tracked down a bug when doing FTP over a HTTP
  proxy. libcurl would then wrongly close the connection after each
  request. In his case it had the weird side-effect that it killed NTLM auth
  for the proxy causing an inifinite loop!

  I added test case 1098 to verify this fix. The test case does however not
  properly verify that the transfers are done persistently - as I couldn't
  think of a clever way to achieve it right now - but you need to read the
  stderr output after a test run to see that it truly did the right thing.

Daniel Stenberg (13 Apr 2009)
- bug report #2727981 (http://curl.haxx.se/bug/view.cgi?id=2727981) by Martin
  Storsjö pointed out how setting CURLOPT_NOBODY to 0 could be downright
  confusing as it set the method to either GET or HEAD. The example he showed
  looked like:

   curl_easy_setopt(curl, CURLOPT_PUT, 1);
   curl_easy_setopt(curl, CURLOPT_NOBODY, 0);

  The new way doesn't alter the method until the request is about to start. If
  CURLOPT_NOBODY is then 1 the HTTP request will be HEAD. If CURLOPT_NOBODY is
  0 and the request happens to have been set to HEAD, it will then instead be
  set to GET. I believe this will be less surprising to users, and hopefully
  not hit any existing users badly.

- Toshio Kuratomi reported a memory leak problem with libcurl+NSS that turned
  out to be leaking cacerts. Kamil Dudka helped me complete the fix. The issue
  is found in Redhat's bug tracker:
  https://bugzilla.redhat.com/show_bug.cgi?id=453612

  There are still memory leaks present, but they seem to have other reasons.

Daniel Fandrich (11 Apr 2009)
- Added new libcurl source files to Symbian OS build files.
- Improved Symbian support for SSL.

Yang Tse (10 Apr 2009)
- Daniel Johnson improved the MacOSX-Framework shell script to now perform all
  the steps required to build a Mac OS X four way fat ppc/i386/ppc64/x86_64
  libcurl.framework.  Four way fat framework requires OS X 10.5 SDK or later.

Yang Tse (8 Apr 2009)
- Removed Sun compilers preprocessor block from curlbuild.h.dist, this also
  removes it from the curlbuild.h file originally distributed by the cURL
  project as this file is intended for systems not capable of running the
  configure script.  For those who have been building curl out of the source
  code curl distribution tarball provided by curl.haxx.se the change implies
  nothing.  Previous change in this area committed 2 Apr becomes irrelevant.

Daniel Stenberg (6 Apr 2009)
- I clarified in the docs that CURLOPT_SEEKFUNCTION should return 0 on success
  and 1 on fatal errors. Previously it only mentioned non-zero on fatal
  errors. This is a slight change in meaning, but it follows what we've done
  elsewhere before and it opens up for LOTS of more useful return codes
  whenever we can think of them...

Yang Tse (2 Apr 2009)
- Fix curl_off_t definition for builds done using Sun compilers and a
  non-configured libcurl. In this case curl_off_t data type was gated
  to the off_t data type which depends on the _FILE_OFFSET_BITS. This
  configuration is exactly the unwanted configuration for our curl_off_t
  data type which must not depend on such setting. This breaks ABI for
  libcurl libraries built with Sun compilers which were built without
  having run the configure script with _FILE_OFFSET_BITS different than
  64 and using the ILP32 data model.

Daniel Stenberg (1 Apr 2009)
- Andre Guibert de Bruet fixed a NULL pointer use in an infof() call if a
  strdup() call failed. 

Daniel Fandrich (31 Mar 2009)
- Properly return an error code in curl_easy_recv (reported by Jim Freeman).

Daniel Stenberg (18 Mar 2009)
- Kamil Dudka brought a patch that enables 6 additional crypto algorithms when
  NSS is used. These ciphers were added in NSS 3.4 and require to be enabled
  explicitly.
 
Daniel Stenberg (13 Mar 2009)
- Use libssh2_version() to present the libssh2 version in case the libssh2
  library is found to support it.

Yang Tse (12 Mar 2009)
- Added missing Curl_read() return code checking in TELNET transfers.

- Pierre Brico found and fixed TELNET transfers not being aborted upon
  a write callback failure.

Daniel Stenberg (11 Mar 2009)
- Kamil Dudka made the curl tool properly call curl_global_init() before any
  other libcurl function.

Yang Tse (11 Mar 2009)
- Added missing TELNET timeout support for Windows builds. This issue was
  reported by Pierre Brico.

Daniel Stenberg (9 Mar 2009)
- Frank Hempel found out a bug and provided the fix: 
  
  curl_easy_duphandle did not necessarily duplicate the CURLOPT_COOKIEFILE
  option. It only enabled the cookie engine in the destination handle if
  data->cookies is not NULL (where data is the source handle). In case of a
  newly initialized handle which just had the cookie support enabled by a
  curl_easy_setopt(handle, CURL_COOKIEFILE, "")-call, handle->cookies was
  still NULL because the setopt-call only appends the value to
  data->change.cookielist, hence duplicating this handle would not have the
  cookie engine switched on.

  We also concluded that the slist-functionality would be suitable for being
  put in its own module rather than simply hanging out in lib/sendf.c so I
  created lib/slist.[ch] for them.

- Andreas Farber made the 'buildconf' script check for the presence of m4
  scripts to make it detect a bad checkout earlier. People with older
  checkouts who don't do cvs update with the -d option won't get the new dirs
  and then will get funny outputs that can be a bit hard to understand and
  fix.
  
Daniel Stenberg (8 Mar 2009)
- Andre Guibert de Bruet found and fixed a code segment in ssluse.c where the
  allocation of the memory BIO was not being properly checked.

- Andre Guibert de Bruet fixed the gnutls-using code: There are a few places
  in the gnutls code where we were checking for negative values for errors,
  when the man pages state that GNUTLS_E_SUCCESS is returned on success and
  other values indicate error conditions.

- Bill Egert pointed out (http://curl.haxx.se/bug/view.cgi?id=2671602) that
  curl didn't use sprintf() in a way that is documented to work in POSIX but
  since we use our own printf() code (from libcurl) that shouldn't be a
  problem. Nonetheless I modified the code to not rely on such particular
  features and to not cause further raised eyebrowse with no good reason.

Daniel Fandrich (5 Mar 2009)
- Expanded the security section of the libcurl-tutorial man page to cover
  more issues for authors to consider when writing robust libcurl-using
  applications.

Yang Tse (5 Mar 2009)
- Fixed NTLM authentication memory leak on SSPI enabled Windows builds. This
  issue was noticed by Chris Deidun.

Daniel Fandrich (4 Mar 2009)
- Fixed a problem with m4 quoting in the OpenSSL configure check reported
  by Daniel Johnson.

Daniel Stenberg (3 Mar 2009)
- David James brought a patch that make libcurl close (all) dead connections
  whenever you attempt to open a new connection.

  1. After cleaning up a dead connection, "continue" instead of
     returning FALSE. This ensures that we clean up all dead connections,
     rather than just cleaning up the first dead connection.
  2. Move up the cleanup for dead connections so that it occurs for
     all connections, rather than just the connections which have the same
     preferences as our current new connection.

Version 7.19.4 (3 March 2009)

Daniel Stenberg (3 Mar 2009)
- David Kierznowski notified us about a security flaw
  (http://curl.haxx.se/docs/adv_20090303.html also known as CVE-2009-0037) in
  which previous libcurl versions (by design) can be tricked to access an
  arbitrary local/different file instead of a remote one when
  CURLOPT_FOLLOWLOCATION is enabled. This flaw is now fixed in this release
  together this the addition of two new setopt options for controlling this
  new behavior:

  o CURLOPT_REDIR_PROTOCOLS controls what protocols libcurl is allowed to
  follow to when CURLOPT_FOLLOWLOCATION is enabled. By default, this option
  excludes the FILE and SCP protocols and thus you nee to explicitly allow
  them in your app if you really want that behavior.

  o CURLOPT_PROTOCOLS controls what protocol(s) libcurl is allowed to fetch
  using the primary URL option. This is useful if you want to allow a user or
  other outsiders control what URL to pass to libcurl and yet not allow all
  protocols libcurl may have been built to support.

Daniel Stenberg (27 Feb 2009)
- Senthil Raja Velu reported a problem when CURLOPT_INTERFACE and
  CURLOPT_LOCALPORT were used together (the local port bind failed), and
  Markus Koetter provided the fix!

Daniel Stenberg (25 Feb 2009)
- As Daniel Fandrich figured out, we must do the GnuTLS initing in the
  curl_global_init() function to properly maintain the performing functions
  thread-safe. We've previously (28 April 2007) moved the init to a later time
  just to avoid it to fail very early when libgcrypt dislikes the situation,
  but that move was bad and the fix should rather be in libgcrypt or
  elsewhere.
  
Daniel Stenberg (24 Feb 2009)
- Brian J. Murrell found out that Negotiate proxy authentication didn't work.
  It happened because the code used the struct for server-based auth all the
  time for both proxy and server auth which of course was wrong.

Daniel Stenberg (23 Feb 2009)
- After a bug reported by James Cheng I've made curl_easy_getinfo() for
  CURLINFO_CONTENT_LENGTH_DOWNLOAD and CURLINFO_CONTENT_LENGTH_UPLOAD return
  -1 if the sizes aren't know. Previously these returned 0, make it impossible
  to detect the difference between actually zero and unknown.

Yang Tse (23 Feb 2009)
- Daniel Johnson provided a shell script that will perform all the steps needed
  to build a Mac OS X fat ppc/i386 or ppc64/x86_64 libcurl.framework

Daniel Stenberg (23 Feb 2009)
- I renamed everything in the windows builds files that used the name 'curllib'
  to the proper 'libcurl' as clearly this caused confusion.

Yang Tse (20 Feb 2009)
- Do not halt compilation when using VS2008 to build a Windows 2000 target.

Daniel Stenberg (20 Feb 2009)
- Linus Nielsen Feltzing reported and helped me repeat and fix a problem with
  FTP with the multi interface: when a transfer fails, like when aborted by a
  write callback, the control connection was wrongly closed and thus not
  re-used properly.

  This change is also an attempt to cleanup the code somewhat in this area, as
  now the FTP code attempts to keep (better) track on pending responses
  necessary to get read in ftp_done().

Daniel Stenberg (19 Feb 2009)
- Patrik Thunstrom reported a problem and helped me repeat it. It turned out 
  libcurl did a superfluous 1000ms wait when doing SFTP downloads!

  We read data with libssh2 while doing the "DO" operation for SFTP and then
  when we were about to start getting data for the actual file part, the
  "TRANSFER" part, we waited for socket action (in 1000ms) before doing a
  libssh2-read. But in this case libssh2 had already read and buffered the
  data so we ended up always just waiting 1000ms before we get working on the
  data!

Patrick Monnerat (18 Feb 2009)
- FTP downloads (i.e.: RETR) ending with code 550 now return error
  CURLE_REMOTE_FILE_NOT_FOUND instead of CURLE_FTP_COULDNT_RETR_FILE.

Daniel Stenberg (17 Feb 2009)
- Kamil Dudka made NSS-powered builds compile and run again!

- A second follow-up change by Andre Guibert de Bruet to fix a related memory
  leak like that fixed on the 14th. When zlib returns failure, we need to
  cleanup properly before returning error.

- CURLOPT_FTP_CREATE_MISSING_DIRS can now be set to 2 in addition to 1 for
  plain FTP connections, and it will then allow MKD to fail once and retry the
  CWD afterwards. This is especially useful if you're doing many simultanoes
  connections against the same server and they all have this option enabled,
  as then CWD may first fail but then another connection does MKD before this
  connection and thus MKD fails but trying CWD works! The numbers can
  (should?) now be set with the convenience enums now called
  CURLFTP_CREATE_DIR and CURLFTP_CREATE_DIR_RETRY.

  Tests has proven that if you're making an application that uploads a set of
  files to an ftp server, you will get a noticable gain in speed if you're
  using multiple connections and this option will be then be very useful.

Daniel Stenberg (14 Feb 2009)
- Andre Guibert de Bruet found and fixed a memory leak in the content encoding
  code, which could happen on libz errors.

Daniel Fandrich (12 Feb 2009)
- Added support for Digest and NTLM authentication using GnuTLS.

Daniel Stenberg (11 Feb 2009)
- CURLINFO_CONDITION_UNMET was added to allow an application to get to know if
  the condition in the previous request was unmet. This is typically a time
  condition set with CURLOPT_TIMECONDITION and was previously not possible to
  reliably figure out. From bug report #2565128
  (http://curl.haxx.se/bug/view.cgi?id=2565128) filed by Jocelyn Jaubert.

Daniel Fandrich (4 Feb 2009)
- Don't add the standard /usr/lib or /usr/include paths to LDFLAGS and CPPFLAGS
  (respectively) when --with-ssl=/usr is used (patch based on FreeBSD).

- Added an explicit buffer limit check in msdosify() (patch based on FreeBSD).
  This couldn't ever overflow in curl, but might if the code were used
  elsewhere or under different conditions.

Daniel Stenberg (3 Feb 2009)
- Hidemoto Nakada provided a small fix that makes it possible to get the
  CURLINFO_CONTENT_LENGTH_DOWNLOAD size from file:// "transfers" with
  CURLOPT_NOBODY set true.

Daniel Stenberg (2 Feb 2009)
- Patrick Scott found a rather large memory leak when using the multi
  interface and setting CURLMOPT_MAXCONNECTS to something less than the number
  of handles you add to the multi handle. All the connections that didn't fit
  in the cache would not be properly disconnected nor freed!

- Craig A West brought us: libcurl now defaults to do CONNECT with HTTP
  version 1.1 instead of 1.0 like before. This change also introduces the new
  proxy type for libcurl called 'CURLPROXY_HTTP_1_0' that then allows apps to
  switch (back) to CONNECT 1.0 requests. The curl tool also got a --proxy1.0
  option that works exactly like --proxy but sets CURLPROXY_HTTP_1_0.

  I updated all test cases cases that use CONNECT and I tried to do some using
  --proxy1.0 and some updated to do CONNECT 1.1 to get both versions run.

Daniel Stenberg (31 Jan 2009)
- When building with c-ares 1.6.1 (not yet released) or later and IPv6 support
  enabled, we can now take advantage of its brand new AF_UNSPEC support in
  ares_gethostbyname(). This makes test case 241 finally run fine for me with
  this setup since it now parses the "::1 ip6-localhost" line fine in my
  /etc/hosts file!

Daniel Stenberg (30 Jan 2009)
- Scott Cantor filed bug report #2550061
  (http://curl.haxx.se/bug/view.cgi?id=2550061) mentioning that I failed to
  properly make sure that the VC9 makefiles got included in the latest
  release. I've now fixed the release script and verified it so next release
  will hopefully include them properly!

Daniel Fandrich (30 Jan 2009)
- Fixed --disable-proxy for FTP and SOCKS. Thanks to Daniel Egger for
  reporting.

Yang Tse (29 Jan 2009)
- Introduced curl_sspi.c and curl_sspi.h for the implementation of functions
  Curl_sspi_global_init() and Curl_sspi_global_cleanup() which previously were
  named Curl_ntlm_global_init() and Curl_ntlm_global_cleanup() in http_ntlm.c
  Also adjusted socks_sspi.c to remove the link-time dependency on the Windows
  SSPI library using it now in the same way as it was done in http_ntlm.c.

Daniel Stenberg (28 Jan 2009)
- Markus Moeller introduced two new options to libcurl:
  CURLOPT_SOCKS5_GSSAPI_SERVICE and CURLOPT_SOCKS5_GSSAPI_NEC to allow libcurl
  to do GSS-style authentication with SOCKS5 proxies. The curl tool got the
  options called --socks5-gssapi-service and --socks5-gssapi-nec to enable
  these.

Daniel Stenberg (26 Jan 2009)
- Chad Monroe provided the new CURLOPT_TFTP_BLKSIZE option that allows an app
  to set desired block size to use for TFTP transfers instead of the default
  512 bytes.

- The "-no_ticket" option was introduced in Openssl0.9.8j. It's a flag to
  disable "rfc4507bis session ticket support".  rfc4507bis was later turned
  into the proper RFC5077 it seems: http://tools.ietf.org/html/rfc5077

  The enabled extension concerns the session management. I wonder how often
  libcurl stops a connection and then resumes a TLS session. also, sending the
  session data is some overhead. .I suggest that you just use your proposed
  patch (which explicitly disables TICKET).

  If someone writes an application with libcurl and openssl who wants to
  enable the feature, one can do this in the SSL callback.

  Sharad Gupta brought this to my attention. Peter Sylvester helped me decide
  on the proper action.

- Alexey Borzov filed bug report #2535504
  (http://curl.haxx.se/bug/view.cgi?id=2535504) pointing out that realms with
  quoted quotation marks in HTTP Digest headers didn't work. I've now added 
  test case 1095 that verifies my fix.

- Craig A West brought CURLOPT_NOPROXY and the corresponding --noproxy option.
  They basically offer the same thing the NO_PROXY environment variable only
  offered previously: list a set of host names that shall not use the proxy
  even if one is specified.

Daniel Fandrich (20 Jan 2009)
- Call setlocale() for libtest tests to test the effects of locale-induced
  libc changes on libcurl.

- Fixed a couple more locale-dependent toupper conversions, mainly for
  clarity.  This does fix one problem that causes ;type=i FTP URLs
  to fail in the Turkish locale when CURLOPT_PROXY_TRANSFER_MODE is
  used (test case 561)

- Added tests 561 and 1091 through 1094 to test various combinations
  of ;type= and ;mode= URLs that could potentially fail in the Turkish
  locale.

Daniel Stenberg (20 Jan 2009)
- Lisa Xu pointed out that the ssh.obj file was missing from the
  lib/Makefile.vc6 file (and thus from the vc8 and vc9 ones too).

Version 7.19.3 (19 January 2009)

Daniel Stenberg (16 Jan 2009)
- Andrew de los Reyes fixed curlbuild.h for "generic" gcc builds on PPC, both
  32 bit and 64 bit.

Daniel Stenberg (15 Jan 2009)
- Tim Ansell fixed a compiler warning in lib/cookie.c

Daniel Stenberg (14 Jan 2009)
- Grant Erickson fixed timeouts for TFTP such that specifying a
  connect-timeout, a max-time or both options work correctly and as expected
  by passing the correct boolean value to Curl_timeleft via the
  'duringconnect' parameter.

  With this small change, curl TFTP now behaves as expected (and likely as
  originally-designed):

  1) For non-existent or unreachable dotted IP addresses:

   a) With no options, follows the default curl 300s timeout...
   b) With --connect-timeout only, follows that value...
   c) With --max-time only, follows that value...
   d) With both --connect-timeout and --max-time, follows the smaller value...

   and times out with a "curl: (7) Couldn't connect to server" error.

  2) For transfers to/from a valid host:

   a) With no options, follows default curl 300s timeout for the
      first XRQ/DATA/ACK transaction and the default TFTP 3600s
      timeout for the remainder of the transfer...

   b) With --connect-time only, follows that value for the
      first XRQ/DATA/ACK transaction and the default TFTP 3600s
      timeout for the remainder of the transfer...

   c) With --max-time only, follows that value for the first
      XRQ/DATA/ACK transaction and for the remainder of the
      transfer...

   d) With both --connect-timeout and --max-time, follows the former
      for the first XRQ/DATA/ACK transaction and the latter for the
      remainder of the transfer...

   and times out with a "curl: (28) Timeout was reached" error as
   appropriate.

Daniel Stenberg (13 Jan 2009)
- Michael Wallner fixed a NULL pointer deref when calling
  curl_easy_setup(curl, CURLOPT_COOKIELIST, "SESS") on a CURL handle with no
  cookies data.

- Stefan Teleman brought a patch to fix the default curlbuild.h file for the
  SunPro compilers.

Daniel Stenberg (12 Jan 2009)
- Based on bug report #2498665 (http://curl.haxx.se/bug/view.cgi?id=2498665)
  by Daniel Black, I've now added magic to the configure script that makes it
  use pkg-config to detect gnutls details as well if the existing method
  (using libgnutls-config) fails. While doing this, I cleaned up and unified
  the pkg-config usage when detecting openssl and nss as well.

Daniel Stenberg (11 Jan 2009)
- Karl Moerder brought the patch that creates vc9 Makefiles, and I made
  'maketgz' now use the actual makefile targets to do the VC8 and VC9
  makefiles.

Daniel Stenberg (10 Jan 2009)
- Emil Romanus fixed:

  When using the multi interface over HTTP and the server returns a Location
  header, the running easy handle will get stuck in the CURLM_STATE_PERFORM
  state, leaving the external event loop stuck waiting for data from the
  ingoing socket (when using the curl_multi_socket_action stuff). While this
  bug was pretty hard to find, it seems to require only a one-line fix. The
  break statement on line 1374 in multi.c caused the function to skip the call
  to multistate().

  How to reproduce this bug? Well, that's another question.  evhiperfifo.c in
  the examples directory chokes on this bug only _sometimes_, probably
  depending on how fast the URLs are added. One way of testing the bug out is
  writing to hiper.fifo from more than one source at the same time.

Daniel Fandrich (7 Jan 2009)
- Unified much of the SessionHandle initialization done in Curl_open() and
  curl_easy_reset() by creating Curl_init_userdefined(). This had the side
  effect of fixing curl_easy_reset() so it now also resets
  CURLOPT_FTP_FILEMETHOD and CURLOPT_SSL_SESSIONID_CACHE

Daniel Stenberg (7 Jan 2009)
- Rob Crittenden did once again provide an NSS update:

  I have to jump through a few hoops now with the NSS library initialization
  since another part of an application may have already initialized NSS by the
  time Curl gets invoked. This patch is more careful to only shutdown the NSS
  library if Curl did the initialization.

  It also adds in a bit of code to set the default ciphers if the app that
  call NSS_Init* did not call NSS_SetDomesticPolicy() or set specific
  ciphers. One might argue that this lets other application developers get
  lazy and/or they aren't using the NSS API correctly, and you'd be right.
  But still, this will avoid terribly difficult-to-trace crashes and is
  generally helpful.

Daniel Stenberg (1 Jan 2009)
- 'reconf' is removed since we rather have users use 'buildconf'

Daniel Stenberg (31 Dec 2008)
- Bas Mevissen reported http://curl.haxx.se/bug/view.cgi?id=2479030 pointing
  out that 'reconf' didn't properly point out the m4 subdirectory when running
  aclocal.

Daniel Stenberg (29 Dec 2008)
 - Phil Lisiecki filed bug report #2413067
  (http://curl.haxx.se/bug/view.cgi?id=2413067) that identified a problem that
  would cause libcurl to mark a DNS cache entry "in use" eternally if the
  subsequence TCP connect failed. It would thus never get pruned and refreshed
  as it should've been.

  Phil provided his own patch to this problem that while it seemed to work