Daniel Stenberg [Tue, 22 Jul 2003 09:58:18 +0000 (09:58 +0000)]
Added a separate struct for the proxyntlm data, as it will/can be different
than the remote server's. That is, both the server and the proxy can in
fact require NTLM auth.
Daniel Stenberg [Tue, 22 Jul 2003 09:57:09 +0000 (09:57 +0000)]
CURLDEBUG is the symbol now
Daniel Stenberg [Tue, 22 Jul 2003 08:23:16 +0000 (08:23 +0000)]
Don't depend on the TIME_WITH_SYS_TIME define. win32 doesn't have sys/time.h
and I don't think we need it.
Daniel Stenberg [Mon, 21 Jul 2003 13:16:30 +0000 (13:16 +0000)]
moved the proxyuser and proxypasswd fields from the sessionhandle to the
connectdata to work as expected
Daniel Stenberg [Mon, 21 Jul 2003 13:16:01 +0000 (13:16 +0000)]
adjusted to support NTLM for proxies
Daniel Stenberg [Mon, 21 Jul 2003 09:19:48 +0000 (09:19 +0000)]
krb4-fixes for the moved user+password fields within the structs
Daniel Stenberg [Mon, 21 Jul 2003 08:25:31 +0000 (08:25 +0000)]
pre4-commit
Daniel Stenberg [Mon, 21 Jul 2003 08:25:21 +0000 (08:25 +0000)]
pre4
Daniel Stenberg [Mon, 21 Jul 2003 07:54:20 +0000 (07:54 +0000)]
the CWD-null bug fix
Daniel Stenberg [Sun, 20 Jul 2003 00:19:44 +0000 (00:19 +0000)]
the fixed skip-blanks in the FTP CWD code called for this adjustment
Daniel Stenberg [Sun, 20 Jul 2003 00:18:11 +0000 (00:18 +0000)]
David Gardner pointed out in bug report 770755 that using the FTP command CWD
with a blank argument is a bad idea. Now skip blanks.
Daniel Stenberg [Sun, 20 Jul 2003 00:02:47 +0000 (00:02 +0000)]
If NTLM is requested, only re-use connections that have the exact same
credentials.
Daniel Stenberg [Sat, 19 Jul 2003 23:58:21 +0000 (23:58 +0000)]
explains my fixes just committed
Daniel Stenberg [Sat, 19 Jul 2003 23:57:08 +0000 (23:57 +0000)]
modified to work fine with the new persistant connection working test suite
HTTP server
Daniel Stenberg [Sat, 19 Jul 2003 23:56:44 +0000 (23:56 +0000)]
swsclose added
Daniel Stenberg [Sat, 19 Jul 2003 23:56:33 +0000 (23:56 +0000)]
Access the user and passwd fields from the connectdata struct now instead
of the sessionhandle struct, as that was not good.
Daniel Stenberg [Sat, 19 Jul 2003 23:55:15 +0000 (23:55 +0000)]
No longer stores user+password in the sessionhandle, now doing that in the
connectdata struct instead. Each being an allocated pointer.
The passwdgiven field was turned into a local variable in the only
function it was being used.
Daniel Stenberg [Sat, 19 Jul 2003 23:54:15 +0000 (23:54 +0000)]
fixed the CONNECT thing again
Daniel Stenberg [Sat, 19 Jul 2003 23:44:22 +0000 (23:44 +0000)]
If the data contents contains the word 'swsclose', then this server will
disconnect the client after the response have been sent. This also happens
if the respons is zero byte long.
In all other cases (unless an error happens), it will now maintain the
connection to allow proper persistant connection testing. This was required
for the NTLM testing to work so I finally had to fix this. Of course most of
the existing HTTP tests will be adjusted to work with this new rule of test
file syntax for HTTP tests.
Also fixed the log function to deal with varargs for better logging.
Daniel Stenberg [Wed, 16 Jul 2003 00:04:45 +0000 (00:04 +0000)]
recent changes
Daniel Stenberg [Tue, 15 Jul 2003 23:47:25 +0000 (23:47 +0000)]
more fixes from Doug Kaufman for DJGPP builds for DOS
Daniel Stenberg [Tue, 15 Jul 2003 23:38:06 +0000 (23:38 +0000)]
updated to work with Dan Winship's NTLM domain stuff fix
Daniel Stenberg [Tue, 15 Jul 2003 23:36:50 +0000 (23:36 +0000)]
Moved the NTLM credentials to the connectdata struct instead, as NTLM
authenticates connections and not single requests. This should make it work
better when we mix requests from multiple hosts. Problem pointed out by
Cris Bailiff.
Daniel Stenberg [Tue, 15 Jul 2003 23:06:02 +0000 (23:06 +0000)]
Fix to the endless loop of bad Basic authentication as reported in Cris
Bailiff's bug report 768275.
Daniel Stenberg [Tue, 15 Jul 2003 22:58:36 +0000 (22:58 +0000)]
Dan Winship's patch added that makes use of DOMAIN\USER or DOMAIN/USER
for the user field. I changed it slightly to stay with strchr() only instead
of strpbrk() for portability reasons.
Daniel Stenberg [Tue, 15 Jul 2003 22:46:01 +0000 (22:46 +0000)]
Dan Winship's fix to make the new auth stuff such as NTLM to work with
the multi interface
Daniel Stenberg [Tue, 15 Jul 2003 22:44:48 +0000 (22:44 +0000)]
Dan Winship pointed out this flaw
Daniel Stenberg [Sat, 5 Jul 2003 13:27:02 +0000 (13:27 +0000)]
re-use existing variable instead of declaring a new local one
Daniel Stenberg [Sat, 5 Jul 2003 13:13:49 +0000 (13:13 +0000)]
Some of Doug Kaufman's changes for the DOS port
Daniel Stenberg [Fri, 4 Jul 2003 18:18:17 +0000 (18:18 +0000)]
the latest changes
Daniel Stenberg [Fri, 4 Jul 2003 18:17:58 +0000 (18:17 +0000)]
HAVE_SETVBUF removed, no longer used
Daniel Stenberg [Fri, 4 Jul 2003 18:15:53 +0000 (18:15 +0000)]
removed
Daniel Stenberg [Fri, 4 Jul 2003 18:15:25 +0000 (18:15 +0000)]
Dan Grayson pointed out that we set the CURL_CA_BUNDLE variable wrongly in
the configure script. We set it differently now and generate the
lib/ca-bundle.h file entirely.
Daniel Stenberg [Fri, 4 Jul 2003 17:16:34 +0000 (17:16 +0000)]
remove the usage of setvbuf() and use fflush() instead if no buffering should
be done on the output
Daniel Stenberg [Fri, 4 Jul 2003 16:37:16 +0000 (16:37 +0000)]
CURLDEBUG not MALLOCDEBUG anymore
Daniel Stenberg [Fri, 4 Jul 2003 16:36:31 +0000 (16:36 +0000)]
adjusted to the NTLM updates
Daniel Stenberg [Fri, 4 Jul 2003 16:29:23 +0000 (16:29 +0000)]
Peter Sylvester's patch was applied that introduces the following:
CURLOPT_SSL_CTX_FUNCTION to set a callback that gets called with the
OpenSSL's ssl_ctx pointer passed in and allow a callback to act on it. If
anything but CURLE_OK is returned, that will also be returned by libcurl
all the way back. If this function changes the CURLOPT_URL, libcurl will
detect this and instead go use the new URL.
CURLOPT_SSL_CTX_DATA is a pointer you set to get passed to the callback set
with CURLOPT_SSL_CTX_FUNCTION.
Daniel Stenberg [Tue, 1 Jul 2003 15:21:42 +0000 (15:21 +0000)]
David Byron's patch that allows a client to make the server quit with a
magic url.
Daniel Stenberg [Tue, 1 Jul 2003 12:12:10 +0000 (12:12 +0000)]
new CVS info
Daniel Stenberg [Tue, 1 Jul 2003 10:12:52 +0000 (10:12 +0000)]
Gisle Vanem found a lib handle leak in the ldap code
Daniel Stenberg [Fri, 27 Jun 2003 14:37:38 +0000 (14:37 +0000)]
When I introduced the DIST_SUBDIRS usage, I broken the 'make install' for
include files and docs, so now I've added a custom install hook to run
make install for docs and install when data is installed at the top-level.
Sterling Hughes [Thu, 26 Jun 2003 21:30:48 +0000 (21:30 +0000)]
revert out my bogus commit. ;-)
Sterling Hughes [Thu, 26 Jun 2003 21:17:29 +0000 (21:17 +0000)]
test commit
Daniel Stenberg [Thu, 26 Jun 2003 11:45:04 +0000 (11:45 +0000)]
new tests
Daniel Stenberg [Thu, 26 Jun 2003 11:44:01 +0000 (11:44 +0000)]
produce a skip-report at the end of all tests, and thus record and count
them properly
Daniel Stenberg [Thu, 26 Jun 2003 11:42:54 +0000 (11:42 +0000)]
beautified and added comments all over
Daniel Stenberg [Thu, 26 Jun 2003 11:41:24 +0000 (11:41 +0000)]
mention the new flag bits we support
Daniel Stenberg [Thu, 26 Jun 2003 11:41:08 +0000 (11:41 +0000)]
mention that it copies the string you add
Daniel Stenberg [Thu, 26 Jun 2003 11:40:44 +0000 (11:40 +0000)]
added lots, mostly the new auth-related option(s)
Daniel Stenberg [Thu, 26 Jun 2003 11:40:04 +0000 (11:40 +0000)]
added lots of auth stuff and updated other things too
Daniel Stenberg [Thu, 26 Jun 2003 11:38:53 +0000 (11:38 +0000)]
mention COOKIES, removed added entries, corrected the FPL-SSL link/reference
Daniel Stenberg [Thu, 26 Jun 2003 11:37:13 +0000 (11:37 +0000)]
mention the other formats the docs come in
Daniel Stenberg [Thu, 26 Jun 2003 11:36:32 +0000 (11:36 +0000)]
adjusted to recent changes
Daniel Stenberg [Thu, 26 Jun 2003 11:35:48 +0000 (11:35 +0000)]
we do support NTLM now...
Daniel Stenberg [Thu, 26 Jun 2003 11:34:36 +0000 (11:34 +0000)]
CURLDEBUG is the symbol to use, no longer MALLOCDEBUG
Daniel Stenberg [Thu, 26 Jun 2003 11:34:07 +0000 (11:34 +0000)]
adjusted the compressed generation to be more helpful in comments etc
Daniel Stenberg [Thu, 26 Jun 2003 11:33:29 +0000 (11:33 +0000)]
support for the new auth stuff
more output on --version/-V
mention --manual on the help output text
Daniel Stenberg [Thu, 26 Jun 2003 11:31:50 +0000 (11:31 +0000)]
new httpauth support, changed filetime variable kind
Daniel Stenberg [Thu, 26 Jun 2003 11:30:59 +0000 (11:30 +0000)]
added CURLOPT_HTTPAUTH support
Daniel Stenberg [Thu, 26 Jun 2003 11:30:26 +0000 (11:30 +0000)]
Adjusted to work properly with the new authentication stuff
Added code to deal with white spaces in relocation headers.
Daniel Stenberg [Thu, 26 Jun 2003 11:28:26 +0000 (11:28 +0000)]
use CURLDEBUG instead of MALLOCDEBUG
Daniel Stenberg [Thu, 26 Jun 2003 11:27:38 +0000 (11:27 +0000)]
use CURLDEBUG
Daniel Stenberg [Thu, 26 Jun 2003 11:27:22 +0000 (11:27 +0000)]
adjusted to use the same API as the OpenSSL version of the MD5 functions
Daniel Stenberg [Thu, 26 Jun 2003 11:26:50 +0000 (11:26 +0000)]
added ntlm flag bits
Daniel Stenberg [Thu, 26 Jun 2003 11:26:26 +0000 (11:26 +0000)]
Many fixes, most of them based on comments by Eric Glass
Daniel Stenberg [Thu, 26 Jun 2003 11:25:42 +0000 (11:25 +0000)]
new proto for Curl_input_negotiate
Daniel Stenberg [Thu, 26 Jun 2003 11:25:23 +0000 (11:25 +0000)]
kill warnings
Daniel Stenberg [Thu, 26 Jun 2003 11:24:55 +0000 (11:24 +0000)]
major adjustments to the new authentication support
Daniel Stenberg [Thu, 26 Jun 2003 11:22:48 +0000 (11:22 +0000)]
include GSS in the debug string if available, support a few new flag
booleans
Daniel Stenberg [Thu, 26 Jun 2003 11:22:12 +0000 (11:22 +0000)]
use CURLDEBUG instead of MALLOCDEBUG for preprocessor conditions
Daniel Stenberg [Thu, 26 Jun 2003 11:16:37 +0000 (11:16 +0000)]
use CURLDEBUG instead of MALLOCDEBUG
Daniel Stenberg [Thu, 26 Jun 2003 06:52:48 +0000 (06:52 +0000)]
one typecast less for the localtime(), use CURLDEBUG instead of MALLOCDEBUG
Daniel Stenberg [Thu, 26 Jun 2003 06:50:32 +0000 (06:50 +0000)]
use CURLDEBUG instead of MALLOCDEBUG
Daniel Stenberg [Thu, 26 Jun 2003 06:47:20 +0000 (06:47 +0000)]
CURLDEBUG instead of MALLOCDEBUG
Daniel Stenberg [Thu, 26 Jun 2003 06:45:15 +0000 (06:45 +0000)]
We noe use CURLDEBUG instead of MALLOCDEBUG
Daniel Stenberg [Thu, 26 Jun 2003 06:21:29 +0000 (06:21 +0000)]
moved from former CVS
Daniel Stenberg [Thu, 26 Jun 2003 06:19:37 +0000 (06:19 +0000)]
Added time_t
Daniel Stenberg [Wed, 25 Jun 2003 23:40:48 +0000 (23:40 +0000)]
up to date with the actual situation
Daniel Stenberg [Fri, 13 Jun 2003 10:15:55 +0000 (10:15 +0000)]
Cris Bailiff's patch that should make us do NTLM correctly. When we've
authenticated our connection, we can continue without any Authorization:
headers as long as our connection is maintained.
Daniel Stenberg [Fri, 13 Jun 2003 09:09:04 +0000 (09:09 +0000)]
stdout is good enough
Daniel Stenberg [Fri, 13 Jun 2003 09:04:08 +0000 (09:04 +0000)]
work more on pids, less on pidfiles to be able to do better kills at the
end of the test where the pidfiles aren't found, but "our" server is running
Daniel Stenberg [Fri, 13 Jun 2003 08:03:45 +0000 (08:03 +0000)]
fixed NTLM test 67, added test 68 for bad NTLM name/password
Daniel Stenberg [Fri, 13 Jun 2003 07:56:38 +0000 (07:56 +0000)]
Cris Bailiff's bugfix
Daniel Stenberg [Fri, 13 Jun 2003 07:14:46 +0000 (07:14 +0000)]
use more curlish strings, these should be able to change...
Daniel Stenberg [Fri, 13 Jun 2003 06:48:04 +0000 (06:48 +0000)]
Marty Kuhrt's #include fixes for VMS
Daniel Stenberg [Thu, 12 Jun 2003 23:05:12 +0000 (23:05 +0000)]
get and use only the first line of the curl --version output
Daniel Stenberg [Thu, 12 Jun 2003 23:03:08 +0000 (23:03 +0000)]
modified
Daniel Stenberg [Thu, 12 Jun 2003 23:02:36 +0000 (23:02 +0000)]
Make the HTTP auth stuff work, Dan Fandrich made --version output a list
of all supported protocols.
Daniel Stenberg [Thu, 12 Jun 2003 19:17:08 +0000 (19:17 +0000)]
remove the dumpit file after use
Daniel Stenberg [Thu, 12 Jun 2003 17:40:56 +0000 (17:40 +0000)]
corrected a comment
Daniel Stenberg [Thu, 12 Jun 2003 17:34:27 +0000 (17:34 +0000)]
CURLHTTP* renamed to CURLAUTH* and NEGOTIATE is now GSSNEGOTIATE as there's
a "plain" Negotiate as well.
Daniel Stenberg [Thu, 12 Jun 2003 16:39:35 +0000 (16:39 +0000)]
NTLM test case
Daniel Stenberg [Thu, 12 Jun 2003 16:38:14 +0000 (16:38 +0000)]
require the netrc_debug feature the same way we now can require SSL
present client-side
Daniel Stenberg [Thu, 12 Jun 2003 16:22:52 +0000 (16:22 +0000)]
now test cases can be set to be dependent on the presence of "SSL" in the
client/library
Daniel Stenberg [Thu, 12 Jun 2003 13:55:40 +0000 (13:55 +0000)]
make it build with older OpenSSL
Daniel Stenberg [Thu, 12 Jun 2003 13:18:10 +0000 (13:18 +0000)]
attempt to make older OpenSSL versions work with the DES stuff
Daniel Stenberg [Thu, 12 Jun 2003 12:54:34 +0000 (12:54 +0000)]
Based on Dan Fandrich's patch and gzip unpack function, we now compress
the 'hugehelp' text if libz and gzip are available at build time.
Daniel Stenberg [Thu, 12 Jun 2003 12:53:38 +0000 (12:53 +0000)]
store HAVE_LIBZ as an automake conditional
Daniel Stenberg [Wed, 11 Jun 2003 16:14:45 +0000 (16:14 +0000)]
fixing details for NTLM
Daniel Stenberg [Wed, 11 Jun 2003 15:33:09 +0000 (15:33 +0000)]
more how I envision it _should_ work, but it still doesn't...