platform/upstream/curl.git
15 years ago- Karl M brought the patch that creates vc9 Makefiles, and I made 'maketgz'
Daniel Stenberg [Sun, 11 Jan 2009 13:34:05 +0000 (13:34 +0000)]
- Karl M brought the patch that creates vc9 Makefiles, and I made 'maketgz'
  now use the actual makefile targets to do the VC8 and VC9 makefiles.

15 years ago- Emil Romanus fixed:
Daniel Stenberg [Sat, 10 Jan 2009 22:10:57 +0000 (22:10 +0000)]
- 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.

15 years agoOnly care about the first line of libtool --version as otherwise this script
Daniel Stenberg [Thu, 8 Jan 2009 22:53:37 +0000 (22:53 +0000)]
Only care about the first line of libtool --version as otherwise this script
get fooled by libtool 2, and at least libtool 1.5.2x do output the interesting
version string in the first line.

15 years ago"204 - Rob C's NSS fix" is done and since we enter feature freeze now I moved
Daniel Stenberg [Thu, 8 Jan 2009 10:32:54 +0000 (10:32 +0000)]
"204 - Rob C's NSS fix" is done and since we enter feature freeze now I moved
the remaining three issues to next release since they are about new things and
not just fixing bugs

15 years agoAdded a few more issues
Dan Fandrich [Thu, 8 Jan 2009 01:09:29 +0000 (01:09 +0000)]
Added a few more issues

15 years agoUnified much of the SessionHandle initialization done in Curl_open() and
Dan Fandrich [Thu, 8 Jan 2009 00:31:49 +0000 (00:31 +0000)]
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

15 years agoAlways use nocheck="yes" for consistency
Dan Fandrich [Wed, 7 Jan 2009 21:57:26 +0000 (21:57 +0000)]
Always use nocheck="yes" for consistency

15 years agoCreated a CURLMIN macro to match CURLMAX
Dan Fandrich [Wed, 7 Jan 2009 19:39:35 +0000 (19:39 +0000)]
Created a CURLMIN macro to match CURLMAX

15 years agofix compiler warnings
Daniel Stenberg [Wed, 7 Jan 2009 14:12:01 +0000 (14:12 +0000)]
fix compiler warnings

15 years ago- Rob Crittenden did once again provide an NSS update:
Daniel Stenberg [Wed, 7 Jan 2009 14:10:35 +0000 (14:10 +0000)]
- 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.

15 years agofour fresh items to deal with!
Daniel Stenberg [Tue, 6 Jan 2009 22:30:40 +0000 (22:30 +0000)]
four fresh items to deal with!

15 years ago- 'reconf' is removed since we rather have users use 'buildconf'
Daniel Stenberg [Thu, 1 Jan 2009 20:50:59 +0000 (20:50 +0000)]
- 'reconf' is removed since we rather have users use 'buildconf'

15 years agocredit Bas
Daniel Stenberg [Wed, 31 Dec 2008 20:36:08 +0000 (20:36 +0000)]
credit Bas

15 years ago- Bas Mevissen reported http://curl.haxx.se/bug/view.cgi?id=2479030 pointing
Daniel Stenberg [Wed, 31 Dec 2008 20:35:46 +0000 (20:35 +0000)]
- Bas Mevissen reported curl.haxx.se/bug/view.cgi?id=2479030 pointing
  out that 'reconf' didn't properly point out the m4 subdirectory when running
  aclocal.

15 years agosync'd address with other tests.
Gunter Knauf [Tue, 30 Dec 2008 15:33:09 +0000 (15:33 +0000)]
sync'd address with other tests.

15 years agotrial with an address which has a valid domain part but invalid host to avoid dns...
Gunter Knauf [Tue, 30 Dec 2008 09:16:38 +0000 (09:16 +0000)]
trial with an address which has a valid domain part but invalid host to avoid dns redirections.

15 years agoadded HAVE_SOCKADDR_IN6_SIN6_SCOPE_ID to ares Makefile.netware and sync'd with other...
Gunter Knauf [Tue, 30 Dec 2008 08:16:24 +0000 (08:16 +0000)]
added HAVE_SOCKADDR_IN6_SIN6_SCOPE_ID to ares Makefile.netware and sync'd with other Makefile.netware.

15 years agochanged HAVE_SIN6_SCOPE_ID define to HAVE_SOCKADDR_IN6_SIN6_SCOPE_ID since just found...
Gunter Knauf [Tue, 30 Dec 2008 08:05:38 +0000 (08:05 +0000)]
changed HAVE_SIN6_SCOPE_ID define to HAVE_SOCKADDR_IN6_SIN6_SCOPE_ID since just found that ares already uses this define.

15 years agoadded HAVE_SIN6_SCOPE_ID define to all non-configure platforms which seem to be IPv6...
Gunter Knauf [Tue, 30 Dec 2008 07:31:04 +0000 (07:31 +0000)]
added HAVE_SIN6_SCOPE_ID define to all non-configure platforms which seem to be IPv6-aware.

15 years ago - Phil Lisiecki filed bug report #2413067
Daniel Stenberg [Mon, 29 Dec 2008 22:25:50 +0000 (22:25 +0000)]
 - 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.

15 years agofurther clarifcation based on input from Anthony Bryan
Daniel Stenberg [Mon, 29 Dec 2008 21:26:11 +0000 (21:26 +0000)]
further clarifcation based on input from Anthony Bryan

15 years ago- Peter Korsgaard fixed building libcurl with "configure --with-ssl
Daniel Stenberg [Sun, 28 Dec 2008 22:00:36 +0000 (22:00 +0000)]
- Peter Korsgaard fixed building libcurl with "configure --with-ssl
  --disable-verbose".

15 years agoAnthony Bryan's man page cleanup in language and spelling
Daniel Stenberg [Sun, 28 Dec 2008 21:56:56 +0000 (21:56 +0000)]
Anthony Bryan's man page cleanup in language and spelling

15 years agoAnthony Bryan reported quirks, I updated
Daniel Stenberg [Sun, 28 Dec 2008 21:43:34 +0000 (21:43 +0000)]
Anthony Bryan reported quirks, I updated

15 years agoAnthony Bryan reported and I corrected two typos.
Daniel Stenberg [Sun, 28 Dec 2008 21:35:53 +0000 (21:35 +0000)]
Anthony Bryan reported and I corrected two typos.

15 years agoThe CURLOPT_KRBLEVEL description wasn't properly formatted and thus was
Daniel Stenberg [Sun, 28 Dec 2008 21:31:55 +0000 (21:31 +0000)]
The CURLOPT_KRBLEVEL description wasn't properly formatted and thus was
corrupted in ouputs. Another report from Anthony Bryan.

15 years agoAnthony Bryan reported this outputs wrong in the PDF and I've now tried to
Daniel Stenberg [Sun, 28 Dec 2008 21:29:13 +0000 (21:29 +0000)]
Anthony Bryan reported this outputs wrong in the PDF and I've now tried to
escape these letters what I think is the correct way.

15 years agosilent 'unused' warnings.
Gunter Knauf [Sun, 28 Dec 2008 05:49:39 +0000 (05:49 +0000)]
silent 'unused' warnings.

15 years agolibssh2 0.19 became 1.0 instead
Daniel Stenberg [Sat, 27 Dec 2008 23:25:31 +0000 (23:25 +0000)]
libssh2 0.19 became 1.0 instead

15 years agoadded check for sin6_scope_id member in struct sockaddr_in6.
Gunter Knauf [Sat, 27 Dec 2008 04:30:36 +0000 (04:30 +0000)]
added check for sin6_scope_id member in struct sockaddr_in6.

15 years ago193 removed, no work has been started on this and I personally don't care that
Daniel Stenberg [Thu, 25 Dec 2008 15:09:18 +0000 (15:09 +0000)]
193 removed, no work has been started on this and I personally don't care that
much about it

15 years agoRemove trailing #undef value. Typo?
Gisle Vanem [Mon, 22 Dec 2008 18:46:12 +0000 (18:46 +0000)]
Remove trailing #undef value. Typo?

15 years agolibssh2_sftp_seek2 was just renamed to libssh2_sftp_seek64 ...
Daniel Stenberg [Mon, 22 Dec 2008 13:21:23 +0000 (13:21 +0000)]
libssh2_sftp_seek2 was just renamed to libssh2_sftp_seek64 ...

15 years ago- Given a recent enough libssh2, libcurl can now seek/resume with SFTP even
Daniel Stenberg [Mon, 22 Dec 2008 13:12:36 +0000 (13:12 +0000)]
- Given a recent enough libssh2, libcurl can now seek/resume with SFTP even
  on file indexes beyond 2 or 4GB.

15 years ago- Anthony Bryan provided a set of patches that cleaned up manual language,
Daniel Stenberg [Mon, 22 Dec 2008 13:07:13 +0000 (13:07 +0000)]
- Anthony Bryan provided a set of patches that cleaned up manual language,
  corrected spellings and more.

15 years agomalloc+memset => calloc
Daniel Stenberg [Sat, 20 Dec 2008 22:51:57 +0000 (22:51 +0000)]
malloc+memset => calloc

15 years agosame procedure, simpler code
Daniel Stenberg [Sat, 20 Dec 2008 22:47:49 +0000 (22:47 +0000)]
same procedure, simpler code

15 years ago197 - IIS-bug in Digest
Daniel Stenberg [Sat, 20 Dec 2008 22:10:42 +0000 (22:10 +0000)]
197 - IIS-bug in Digest
 The curl tool parts are postponed to a later time

201 - "bug: header data output to the body callback function after set header"
 Was probably not a bug, I asked about it but I didn't get any response.

202 - "hangs up of application above libcurl" - problems with the multi_socket
 Fixes from Igor have been committed and there's currently no pending ones.

15 years ago- Igor Novoseltsev fixed a bad situation for the multi_socket() API when doing
Daniel Stenberg [Sat, 20 Dec 2008 22:03:22 +0000 (22:03 +0000)]
- Igor Novoseltsev fixed a bad situation for the multi_socket() API when doing
  pipelining, as libcurl could then easily get confused and A) work on the
  handle that was not "first in queue" on a pipeline, or even B) tell the app
  to REMOVE a socket while it was in use by a second handle in a pipeline. Both
  errors caused hanging or stalling applications.

15 years agomake the debug/helper function output to stderr as that makes it play nicer
Daniel Stenberg [Sat, 20 Dec 2008 21:48:34 +0000 (21:48 +0000)]
make the debug/helper function output to stderr as that makes it play nicer
in combination with infof() calls

15 years agoremove outdated stuff
Daniel Stenberg [Sat, 20 Dec 2008 17:16:45 +0000 (17:16 +0000)]
remove outdated stuff

15 years ago- curl_multi_timeout() could return a timeout value of 0 even though nothing
Daniel Stenberg [Fri, 19 Dec 2008 22:58:22 +0000 (22:58 +0000)]
- curl_multi_timeout() could return a timeout value of 0 even though nothing
  was actually ready to get done, as the internal time resolution is higher
  than the returned millisecond timer. Therefore it could cause applications
  running on fast processors to do short bursts of busy-loops.
  curl_multi_timeout() will now only return 0 if the timeout is actually
  alreay triggered.

15 years agooops, removing debug output that wasn't supposed to be there
Daniel Stenberg [Fri, 19 Dec 2008 22:27:11 +0000 (22:27 +0000)]
oops, removing debug output that wasn't supposed to be there

15 years ago- Using the libssh2 0.19 function libssh2_session_block_directions(), libcurl
Daniel Stenberg [Fri, 19 Dec 2008 21:14:52 +0000 (21:14 +0000)]
- Using the libssh2 0.19 function libssh2_session_block_directions(), libcurl
  now has an improved ability to do right when the multi interface (both
  "regular" and multi_socket) is used for SCP and SFTP transfers. This should
  result in (much) less busy-loop situations and thus less CPU usage with no
  speed loss.

15 years agoexpand the CURLOPT_POSTREDIR explanation
Daniel Stenberg [Fri, 19 Dec 2008 19:03:55 +0000 (19:03 +0000)]
expand the CURLOPT_POSTREDIR explanation

15 years ago- SCP and SFTP with the multi interface had the same flaw: the 'DONE'
Daniel Stenberg [Wed, 17 Dec 2008 12:32:41 +0000 (12:32 +0000)]
- SCP and SFTP with the multi interface had the same flaw: the 'DONE'
  operation didn't complete properly if the EAGAIN equivalent was returned but
  libcurl would simply continue with a half-completed close operation
  performed. This ruined persistent connection re-use and cause some
  SSH-protocol errors in general. The correction is unfortunately adding a
  blocking function - doing it entirely non-blocking should be considered for
  a better fix.

15 years agoUpdated dependencies based on "gcc -MM".
Gisle Vanem [Tue, 16 Dec 2008 09:12:46 +0000 (09:12 +0000)]
Updated dependencies based on "gcc -MM".

15 years agoAdded the use of Watt-32 tcp/ip stack for Win32 target.
Gisle Vanem [Tue, 16 Dec 2008 09:03:58 +0000 (09:03 +0000)]
Added the use of Watt-32 tcp/ip stack for Win32 target.
If USE_WATT32=1 one needs to use stack-based calls (-3s).
So to keep the makefile nice and clean, specify -3s for
Winsock target too (there's hardly any speed-gain using -3r).

15 years agoAdded the use of Watt-32 tcp/ip stack for Win32 targets.
Gisle Vanem [Tue, 16 Dec 2008 08:59:45 +0000 (08:59 +0000)]
Added the use of Watt-32 tcp/ip stack for Win32 targets.

15 years agoAdded the use of Watt-32 tcp/ip stack for Win32 targets.
Gisle Vanem [Tue, 16 Dec 2008 08:48:44 +0000 (08:48 +0000)]
Added the use of Watt-32 tcp/ip stack for Win32 targets.
Added USE_WATT32 unconditionally for MSDOS targets since
it's the only option. Adjusted the text for '--wdebug'.

15 years agoAdded the use of Watt-32 tcp/ip stack for Win32 targets.
Gisle Vanem [Tue, 16 Dec 2008 08:25:55 +0000 (08:25 +0000)]
Added the use of Watt-32 tcp/ip stack for Win32 targets.

15 years ago- libssh2_sftp_last_error() was wrongly used at some places in libcurl which
Daniel Stenberg [Mon, 15 Dec 2008 23:04:51 +0000 (23:04 +0000)]
- libssh2_sftp_last_error() was wrongly used at some places in libcurl which
  made libcurl sometimes not properly abort problematic SFTP transfers.

15 years ago196 is gone
Daniel Stenberg [Fri, 12 Dec 2008 12:38:47 +0000 (12:38 +0000)]
196 is gone

15 years ago- More work with Igor Novoseltsev to first fix the remaining stuff for
Daniel Stenberg [Fri, 12 Dec 2008 12:21:11 +0000 (12:21 +0000)]
- More work with Igor Novoseltsev to first fix the remaining stuff for
  removing easy handles from multi handles when the easy handle is/was within
  a HTTP pipeline. His bug report #2351653
  (http://curl.haxx.se/bug/view.cgi?id=2351653) was also related and was
  eventually fixed by a patch by Igor himself.

15 years agoMake the getoff all pipelines function only set _inuse to FALSE if the
Daniel Stenberg [Fri, 12 Dec 2008 12:04:10 +0000 (12:04 +0000)]
Make the getoff all pipelines function only set _inuse to FALSE if the
specified data pointer was head.

15 years agocredit Mark Karpeles for his report and work
Daniel Stenberg [Fri, 12 Dec 2008 08:36:56 +0000 (08:36 +0000)]
credit Mark Karpeles for his report and work

15 years agoPatrick Monnerat fixed a build regression, introduced in 7.19.2, affecting
Yang Tse [Fri, 12 Dec 2008 03:24:59 +0000 (03:24 +0000)]
Patrick Monnerat fixed a build regression, introduced in 7.19.2, affecting
OS/400 compilations with IPv6 enabled.

15 years agoRemoved 200 as it wasn't a bug after all
Daniel Stenberg [Thu, 11 Dec 2008 23:55:20 +0000 (23:55 +0000)]
Removed 200 as it wasn't a bug after all
Added 203 "dns cache memory leak and TTL failure after failed conn"

15 years ago- Bug report #2416182 titled "crash in ConnectionExists when using
Daniel Stenberg [Thu, 11 Dec 2008 23:52:56 +0000 (23:52 +0000)]
- Bug report #2416182 titled "crash in ConnectionExists when using
  duphandle+curl_mutli" (http://curl.haxx.se/bug/view.cgi?id=2416182) showed
  that curl_easy_duphandle() wrongly also copied the pointer to the connection
  cache, which was plain wrong and caused a segfault if the handle would be
  used in a different multi handle than the handle it was duplicated from.

15 years agoDan F pointed out that we don't need to scan IPv6 addresses for '%'-letters
Daniel Stenberg [Thu, 11 Dec 2008 23:27:22 +0000 (23:27 +0000)]
Dan F pointed out that we don't need to scan IPv6 addresses for '%'-letters
in the parse_remote_port() function as the scope id has already been stripped
from the string.

15 years ago- Keshav Krity found out that libcurl failed to deal with dotted IPv6
Daniel Stenberg [Thu, 11 Dec 2008 22:22:46 +0000 (22:22 +0000)]
- Keshav Krity found out that libcurl failed to deal with dotted IPv6
  addresses if they were very long (>39 letters) due to a too strict address
  validity parser. It now accepts addresses up to 45 bytes long.

15 years ago_ Use getaddrinfo/getnameinfo ascii wrappers on OS400.
Patrick Monnerat [Thu, 11 Dec 2008 19:20:03 +0000 (19:20 +0000)]
_ Use getaddrinfo/getnameinfo ascii wrappers on OS400.
_ Adjust OS400 make script for non-CVS distributions.
_ Upgrade ILE/RPG binding.
_ Define CURL_HIDDEN_SYMBOLS on OS400, since only CURL_EXTERN-marked symbols are exported.

15 years agoMention what makes this test special
Dan Fandrich [Thu, 11 Dec 2008 03:01:38 +0000 (03:01 +0000)]
Mention what makes this test special

15 years ago- Internet Explorer had a broken HTTP digest authentication before v7 and
Daniel Stenberg [Wed, 10 Dec 2008 23:13:31 +0000 (23:13 +0000)]
- Internet Explorer had a broken HTTP digest authentication before v7 and
  there are servers "out there" that relies on the client doing this broken
  Digest authentication. Apache even comes with an option to work with such
  broken clients.

  The difference is only for URLs that contain a query-part (a '?'-letter and
  text to the right of it).

  libcurl now supports this quirk, and you enable it by setting the
  CURLAUTH_DIGEST_IE bit in the bitmask you pass to the CURLOPT_HTTPAUTH or
  CURLOPT_PROXYAUTH options. They are thus individually controlled to server
  and proxy.

15 years agouse the new URL
Daniel Stenberg [Tue, 9 Dec 2008 22:49:51 +0000 (22:49 +0000)]
use the new URL

15 years agostart over on the 1.6.1 release...
Daniel Stenberg [Tue, 9 Dec 2008 22:41:36 +0000 (22:41 +0000)]
start over on the 1.6.1 release...

15 years agoadd space
Daniel Stenberg [Tue, 9 Dec 2008 22:30:59 +0000 (22:30 +0000)]
add space

15 years agoDocumented CURLOPT_CONNECT_ONLY as being useful only on HTTP URLs.
Dan Fandrich [Tue, 9 Dec 2008 22:00:18 +0000 (22:00 +0000)]
Documented CURLOPT_CONNECT_ONLY as being useful only on HTTP URLs.

15 years agoAdded test cases 1089 and 1090 to test --write-out after a redirect to
Dan Fandrich [Tue, 9 Dec 2008 21:59:57 +0000 (21:59 +0000)]
Added test cases 1089 and 1090 to test --write-out after a redirect to
test a report that the size didn't work, but these test cases pass.

15 years ago- Ken Hirsch simplified how libcurl does FTPS: now it doesn't assume any
Daniel Stenberg [Tue, 9 Dec 2008 15:02:37 +0000 (15:02 +0000)]
- Ken Hirsch simplified how libcurl does FTPS: now it doesn't assume any
  particular state for the control connection like it did before for implicit
  FTPS (libcurl assumed such control connections to be encrypted while some
  FTPS servers such as FileZilla assumes such connections to be clear
  mode). Use the CURLOPT_USE_SSL option to set your desired level.

15 years agoFix for Win32 targets using Watt-32.
Gisle Vanem [Tue, 9 Dec 2008 14:39:08 +0000 (14:39 +0000)]
Fix for Win32 targets using Watt-32.

15 years agoC89 compilers (like Minix' ACK) only need to handle 31 functions arguments
Dan Fandrich [Tue, 9 Dec 2008 01:02:28 +0000 (01:02 +0000)]
C89 compilers (like Minix' ACK) only need to handle 31 functions arguments
so split a long sprintf into two calls to get below that number.

15 years ago195 - SIZE returning 550 must not abort the transfer
Daniel Stenberg [Mon, 8 Dec 2008 20:21:36 +0000 (20:21 +0000)]
195 - SIZE returning 550 must not abort the transfer
Fixed!

15 years ago- Fred Machado posted about a weird FTP problem on the curl-users list and when
Daniel Stenberg [Mon, 8 Dec 2008 20:20:51 +0000 (20:20 +0000)]
- Fred Machado posted about a weird FTP problem on the curl-users list and when
  researching it, it turned out he got a 550 response back from a SIZE command
  and then I fell over the text in RFC3659 that says:

   The presence of the 550 error response to a SIZE command MUST NOT be taken
   by the client as an indication that the file cannot be transferred in the
   current MODE and TYPE.

  In other words: the change I did on September 30th 2008 and that has been
  included in the last two releases were a regression and a bad idea. We MUST
  NOT take a 550 response from SIZE as a hint that the file doesn't exist.

15 years agoAdded needed defines for Watt-32 on Windows.
Gisle Vanem [Mon, 8 Dec 2008 16:12:11 +0000 (16:12 +0000)]
Added needed defines for Watt-32 on Windows.

15 years agoUndefine 'optarg', 'optind' and 'opterr' when using Watt-32
Gisle Vanem [Mon, 8 Dec 2008 16:11:16 +0000 (16:11 +0000)]
Undefine 'optarg', 'optind' and 'opterr' when using Watt-32
(to get correct linkage on Windows).

15 years agoares_writev() shall not be exported when using Watt-32 (has writev).
Gisle Vanem [Mon, 8 Dec 2008 16:09:21 +0000 (16:09 +0000)]
ares_writev() shall not be exported when using Watt-32 (has writev).
Added _USE_32BIT_TIME_T to avoid runtime warning. Applies to
VC-2008+ only.

15 years agoRemoved unneeded defines HAVE_SIGNAL_H, HAVE_SIG_ATOMIC_T,
Gisle Vanem [Mon, 8 Dec 2008 14:58:56 +0000 (14:58 +0000)]
Removed unneeded defines HAVE_SIGNAL_H, HAVE_SIG_ATOMIC_T,
RETSIGTYPE and HAVE_PROCESS_H.

15 years agoone more
Daniel Stenberg [Mon, 8 Dec 2008 14:26:28 +0000 (14:26 +0000)]
one more

15 years ago192 - "infinite loop during GSS authentication" bug #2221237
Daniel Stenberg [Mon, 8 Dec 2008 14:24:12 +0000 (14:24 +0000)]
192 - "infinite loop during GSS authentication" bug #2221237
Fixed!

15 years ago- Christian Krause filed bug #2221237
Daniel Stenberg [Mon, 8 Dec 2008 13:52:20 +0000 (13:52 +0000)]
- Christian Krause filed bug #2221237
  (http://curl.haxx.se/bug/view.cgi?id=2221237) that identified an infinite
  loop during GSS authentication given some specific conditions. With his
  patience and great feedback I managed to narrow down the problem and
  eventually fix it although I can't test any of this myself!

15 years agothe initial version of the ares_set_socket_callback man page
Daniel Stenberg [Thu, 4 Dec 2008 12:54:43 +0000 (12:54 +0000)]
the initial version of the ares_set_socket_callback man page

15 years agoGregor Jasny provided the patch that introduces ares_set_socket_callback(),
Daniel Stenberg [Thu, 4 Dec 2008 12:53:03 +0000 (12:53 +0000)]
Gregor Jasny provided the patch that introduces ares_set_socket_callback(),
and I edited it to also get duped by ares_dup().

15 years agoBring the sys/include.h include test in line with curl's.
Dan Fandrich [Thu, 4 Dec 2008 07:18:13 +0000 (07:18 +0000)]
Bring the sys/include.h include test in line with curl's.

15 years agoFixed the getifaddrs version of Curl_if2ip to work on systems without IPv6
Dan Fandrich [Thu, 4 Dec 2008 06:24:00 +0000 (06:24 +0000)]
Fixed the getifaddrs version of Curl_if2ip to work on systems without IPv6
support (e.g. Minix)

15 years ago- Igor Novoseltsev filed bug #2351645
Daniel Stenberg [Wed, 3 Dec 2008 15:20:27 +0000 (15:20 +0000)]
- Igor Novoseltsev filed bug #2351645
  (http://curl.haxx.se/bug/view.cgi?id=2351645) that identified a problem with
  the multi interface that occured if you removed an easy handle while in
  progress and the handle was used in a HTTP pipeline.

15 years agofilled in SONAME number bump info, after some CVS digging
Daniel Stenberg [Wed, 3 Dec 2008 15:20:06 +0000 (15:20 +0000)]
filled in SONAME number bump info, after some CVS digging

15 years ago- Pawel Kierski pointed out a mistake in the cookie code that could lead to a
Daniel Stenberg [Wed, 3 Dec 2008 15:08:09 +0000 (15:08 +0000)]
- Pawel Kierski pointed out a mistake in the cookie code that could lead to a
  bad fclose() after a fatal error had occured.
  (http://curl.haxx.se/bug/view.cgi?id=2382219)

15 years agoLet's not call ares_save_options() deprecated just yet
Daniel Stenberg [Wed, 3 Dec 2008 10:03:07 +0000 (10:03 +0000)]
Let's not call ares_save_options() deprecated just yet

15 years agoIntroduce ares_dup(3) and new thoughts about API/ABI and how to move forwards.
Daniel Stenberg [Wed, 3 Dec 2008 09:59:50 +0000 (09:59 +0000)]
Introduce ares_dup(3) and new thoughts about API/ABI and how to move forwards.
Also discussed on the ml.

15 years agoexplain the two *channel_inuse fields somewhat better
Daniel Stenberg [Tue, 2 Dec 2008 23:00:10 +0000 (23:00 +0000)]
explain the two *channel_inuse fields somewhat better

15 years agoMake sure sys/socket.h is included before netinet/in.h (required by
Dan Fandrich [Tue, 2 Dec 2008 02:58:04 +0000 (02:58 +0000)]
Make sure sys/socket.h is included before netinet/in.h (required by
OpenWatcom C, and condoned by SUS)

15 years agominor indent fix
Daniel Stenberg [Mon, 1 Dec 2008 22:36:39 +0000 (22:36 +0000)]
minor indent fix

15 years agoConvert the public config struct to the same binary size/construct as in the
Daniel Stenberg [Mon, 1 Dec 2008 19:06:24 +0000 (19:06 +0000)]
Convert the public config struct to the same binary size/construct as in the
latest releases to remain ABI compatible.

15 years agoadded some more entries for stuff done the last ~2 years
Daniel Stenberg [Mon, 1 Dec 2008 18:46:31 +0000 (18:46 +0000)]
added some more entries for stuff done the last ~2 years

15 years agoPeople have successfully ran libcurl on iphone
Daniel Stenberg [Sun, 30 Nov 2008 22:51:54 +0000 (22:51 +0000)]
People have successfully ran libcurl on iphone

15 years agoFurther adjust a libssh2 preprocessor function-symbol definition check
Yang Tse [Sat, 29 Nov 2008 17:45:16 +0000 (17:45 +0000)]
Further adjust a libssh2 preprocessor function-symbol definition check

15 years agoAdjust some libssh2 preprocessor symbol definition checks
Yang Tse [Sat, 29 Nov 2008 16:39:56 +0000 (16:39 +0000)]
Adjust some libssh2 preprocessor symbol definition checks

15 years agoAdded '-DHAVE_GETHOSTNAME'.
Gisle Vanem [Sat, 29 Nov 2008 15:19:46 +0000 (15:19 +0000)]
Added '-DHAVE_GETHOSTNAME'.

15 years agoMake sure sys/socket.h is included before netinet/in.h (required by
Dan Fandrich [Sat, 29 Nov 2008 00:26:07 +0000 (00:26 +0000)]
Make sure sys/socket.h is included before netinet/in.h (required by
OpenWatcom C)