Daniel Stenberg [Tue, 25 Mar 2003 14:23:12 +0000 (14:23 +0000)]
white space and indent fix
Daniel Stenberg [Mon, 24 Mar 2003 23:11:16 +0000 (23:11 +0000)]
ignore getdate.c
Daniel Stenberg [Mon, 24 Mar 2003 23:10:38 +0000 (23:10 +0000)]
Richard Bramante's provided a fix for a handle re-use problem seen when you
change options on an SSL-enabled connection between requests.
Daniel Stenberg [Mon, 24 Mar 2003 11:06:57 +0000 (11:06 +0000)]
Removed the "TC TrustCenter, Germany, Class 0 CA." certificate:
"It is a DEMO certificate and was never intended to be in any list of trusted
CA certificates."
(quote by Götz Babin-Ebell, trustcenter.de)
Daniel Stenberg [Mon, 24 Mar 2003 10:47:06 +0000 (10:47 +0000)]
all those changes
Daniel Stenberg [Fri, 21 Mar 2003 08:09:48 +0000 (08:09 +0000)]
Hopefully this change addresses these two bug reports: 707003 and 706624.
We need to make sure that when we init a 'connectdata' struct and then
afterwards check for and re-use another one, we must be careful so that the
newly set values are transmitted and used in the surviving connectdata struct.
Daniel Stenberg [Thu, 20 Mar 2003 15:12:43 +0000 (15:12 +0000)]
lib/getdate.c.cvs may need a rename if you don't have yacc or bison
Daniel Stenberg [Thu, 20 Mar 2003 14:38:42 +0000 (14:38 +0000)]
if the cvs update fails, don't continue further
Daniel Stenberg [Wed, 19 Mar 2003 21:28:39 +0000 (21:28 +0000)]
make the ENGINE depend on the USE_SSLEAY define too
Daniel Stenberg [Wed, 19 Mar 2003 09:26:29 +0000 (09:26 +0000)]
set binary mode for some file handling and it might work better on some
cygwin installations (using DOS-style files somehow?)
Daniel Stenberg [Wed, 19 Mar 2003 09:16:59 +0000 (09:16 +0000)]
typecast the conversion from const char * to char *
Daniel Stenberg [Wed, 19 Mar 2003 09:09:40 +0000 (09:09 +0000)]
Rename getdate.c to getdate.c.cvs, since the "normal" build procedure do
imply that yacc/bison exists and can generate this file. Those without one
of those tools can then checkout and rename the getdate.c.cvs file.
Daniel Stenberg [Tue, 18 Mar 2003 10:01:51 +0000 (10:01 +0000)]
clarify that 22 can be returned on --fail for all HTTP errors being 400
or above
Daniel Stenberg [Mon, 17 Mar 2003 17:20:26 +0000 (17:20 +0000)]
more defensive checking as platforms differ...
Daniel Stenberg [Mon, 17 Mar 2003 12:38:08 +0000 (12:38 +0000)]
Andy Cedilnik's corrections
Daniel Stenberg [Sun, 16 Mar 2003 18:41:00 +0000 (18:41 +0000)]
it just never ends
Daniel Stenberg [Sun, 16 Mar 2003 16:20:56 +0000 (16:20 +0000)]
regenerated from getdate.y
Daniel Stenberg [Sun, 16 Mar 2003 16:15:24 +0000 (16:15 +0000)]
Juan F. Codagnone pointed out a missing thing from the march 2 fix
Daniel Stenberg [Sun, 16 Mar 2003 10:46:52 +0000 (10:46 +0000)]
typecase getpid() to int to prevent compiler warning
Daniel Stenberg [Sat, 15 Mar 2003 21:04:01 +0000 (21:04 +0000)]
figure out select()'s argument types
Daniel Stenberg [Sat, 15 Mar 2003 21:02:20 +0000 (21:02 +0000)]
Gisle Vanem's fix to get this working nicely on windows
Daniel Stenberg [Sat, 15 Mar 2003 21:00:47 +0000 (21:00 +0000)]
Gisle Vanem fixed a name collision with structure '"CONTEXT" in <winnt.h>
Daniel Stenberg [Sat, 15 Mar 2003 17:26:59 +0000 (17:26 +0000)]
missing newline added
Daniel Stenberg [Sat, 15 Mar 2003 17:12:46 +0000 (17:12 +0000)]
no server needed
Daniel Stenberg [Sat, 15 Mar 2003 17:11:13 +0000 (17:11 +0000)]
allow 'none' as server
Daniel Stenberg [Sat, 15 Mar 2003 16:51:43 +0000 (16:51 +0000)]
Sort out the ENGINE problems people seem to be having. Now we put all ENGINE
related stuff within HAVE_OPENSSL_ENGINE_H and we don't make any private
typedef or similar if the header is missing...
Daniel Stenberg [Sat, 15 Mar 2003 16:43:58 +0000 (16:43 +0000)]
* use the pid returned back from test-servers and kill them before starting
them the first time
* verify that the server we start really comes up fine and works as
expected before continue
* count test cases where the server can't be run (for whatever reason)
* prefix lots of messages with RUN: to make it easier to realize which script
is saying what when running tests verbose
* remove the generic sleep(1) from each test, makes the suite fly! ;-)
I hope these changes will make the tests run somewhat more reliably on more
platforms.
Daniel Stenberg [Sat, 15 Mar 2003 16:39:15 +0000 (16:39 +0000)]
report pid back in the WE ROOLZ message
Daniel Stenberg [Sat, 15 Mar 2003 16:05:47 +0000 (16:05 +0000)]
removed the "banner" when the server is starting
Daniel Stenberg [Sat, 15 Mar 2003 15:08:55 +0000 (15:08 +0000)]
detect lack of perl before running tests
Daniel Stenberg [Sat, 15 Mar 2003 14:47:09 +0000 (14:47 +0000)]
Rick Jones' minor thing to build better on HPUX 11
Daniel Stenberg [Fri, 14 Mar 2003 17:21:06 +0000 (17:21 +0000)]
Nico Baggus little adjustment to build with OpenSSL 0.9.7 (the ENGINE thing)
Daniel Stenberg [Fri, 14 Mar 2003 12:44:56 +0000 (12:44 +0000)]
improved "deeper" check
Daniel Stenberg [Thu, 13 Mar 2003 23:02:33 +0000 (23:02 +0000)]
No longer halts operation if select or socket are missing, since in most
cases this is wrong... and if they're truly missing, we won't succeed to
link later on anyway.
Daniel Stenberg [Thu, 13 Mar 2003 21:41:02 +0000 (21:41 +0000)]
Philippe Raoult needed this to build on FreeBSD
Daniel Stenberg [Thu, 13 Mar 2003 17:06:55 +0000 (17:06 +0000)]
Extra function-find magic for platforms that don't like the way the
default AC_CHECK_FUNCS() work. HPUX 11 is one of them.
Daniel Stenberg [Thu, 13 Mar 2003 15:56:31 +0000 (15:56 +0000)]
output what cvs returned, see if we can make the script bail out when
cvs update fails
Daniel Stenberg [Thu, 13 Mar 2003 15:54:46 +0000 (15:54 +0000)]
use include path from BUILD dir since we want the config.h
Daniel Stenberg [Wed, 12 Mar 2003 14:29:47 +0000 (14:29 +0000)]
Things are moving along...
Daniel Stenberg [Wed, 12 Mar 2003 14:20:16 +0000 (14:20 +0000)]
When we append stuff to the URL, we must make sure the text is properly
URL encoded before. Test case 58 added to verify this.
Daniel Stenberg [Wed, 12 Mar 2003 14:14:52 +0000 (14:14 +0000)]
-m on curl on windows with telnet doesn't work
Daniel Stenberg [Wed, 12 Mar 2003 14:04:23 +0000 (14:04 +0000)]
This verifies that my fix for bug report #700275 works.
Daniel Stenberg [Wed, 12 Mar 2003 14:03:53 +0000 (14:03 +0000)]
test58 added
Daniel Stenberg [Wed, 12 Mar 2003 13:42:07 +0000 (13:42 +0000)]
improved the header checks
--enable-libgcc
check for a sed before using it
Daniel Stenberg [Wed, 12 Mar 2003 08:54:45 +0000 (08:54 +0000)]
Add include files to prevent warnings on some (HPUX) systems.
Daniel Stenberg [Wed, 12 Mar 2003 08:54:11 +0000 (08:54 +0000)]
include "config.h" from the lib's private dir
Daniel Stenberg [Wed, 12 Mar 2003 08:53:44 +0000 (08:53 +0000)]
Setup include path to the lib dir to enable inclusion of "config.h"
Daniel Stenberg [Wed, 12 Mar 2003 08:44:00 +0000 (08:44 +0000)]
Made set_local_option() properly static as reported by Rick Jones
Daniel Stenberg [Wed, 12 Mar 2003 08:40:45 +0000 (08:40 +0000)]
Prefix defines and symbols with CURL_ to reduce the risk of colliding with
various system's other defines.
Daniel Stenberg [Wed, 12 Mar 2003 08:11:24 +0000 (08:11 +0000)]
support a few more common typedefs
Daniel Stenberg [Wed, 12 Mar 2003 08:07:30 +0000 (08:07 +0000)]
Massige use of AC_HELP_STRING() all over makes the --help output so much
nicer!
Daniel Stenberg [Tue, 11 Mar 2003 19:22:10 +0000 (19:22 +0000)]
another week of changes, especially libtool gave us an adventure to remember
Daniel Stenberg [Tue, 11 Mar 2003 19:12:07 +0000 (19:12 +0000)]
syntax error
Daniel Stenberg [Tue, 11 Mar 2003 19:07:41 +0000 (19:07 +0000)]
Christophe Demory fixed the check to work better for non-blocking on HP-UX
systems. Bug report #701749.
Daniel Stenberg [Tue, 11 Mar 2003 18:58:21 +0000 (18:58 +0000)]
Use ssize_t instead of 'int' to make the 64 bit sparc compiler happier.
Fix by Richard Gorton.
Daniel Stenberg [Tue, 11 Mar 2003 18:55:34 +0000 (18:55 +0000)]
Richard Gorton improved the random_the_seed() function for systems where
we don't find/know of a good random source. This way, we get a better
randomness which in turn should make SSL connections more secure.
Daniel Stenberg [Tue, 11 Mar 2003 17:16:02 +0000 (17:16 +0000)]
don't check for netinet/if_ether.h, we don't include it and it causes
configure warnings on many systems
Daniel Stenberg [Tue, 11 Mar 2003 16:28:23 +0000 (16:28 +0000)]
Martin C. Martin's fix to produce an error message in case of failure
in the Curl_is_connected() function.
Daniel Stenberg [Mon, 10 Mar 2003 20:46:54 +0000 (20:46 +0000)]
added things to fix at the next major release/change
Daniel Stenberg [Mon, 10 Mar 2003 20:43:59 +0000 (20:43 +0000)]
clarify
Daniel Stenberg [Mon, 10 Mar 2003 17:01:11 +0000 (17:01 +0000)]
no the data is not freed, this is left for the app to do when needed
Daniel Stenberg [Mon, 10 Mar 2003 14:52:33 +0000 (14:52 +0000)]
AAAARG
libtool 1.4.3 is scary as hell and caused just about every build on all sorts
of platforms to stop working, thanks to the fact that it ruquires a SED
variables somehow set by the configure script. It works fine on my linux
running autoconf 2.57 and automake 1.7 but others seem not to do as fine.
Reverting back to the ltmain.sh we had previously, which I believe is 1.4.2
including handmade patches for FreeBSD.
ALERT ALERT ALERT before we try 1.4.3 or similar versions again, check the
${SED} stuff and similar carefully.
Daniel Stenberg [Mon, 10 Mar 2003 12:25:32 +0000 (12:25 +0000)]
Include sys/types.h as well. Ray DeGennaro reports successful compiling on
AIX when this fix is applied and I cannot see how this will break any
systems.
Daniel Stenberg [Fri, 7 Mar 2003 13:36:32 +0000 (13:36 +0000)]
figure out the path to a 'sed' as otherwise libtool gets crazy
Daniel Stenberg [Fri, 7 Mar 2003 09:03:09 +0000 (09:03 +0000)]
libtoolize 1.4.3 brought these
Daniel Stenberg [Tue, 4 Mar 2003 06:41:50 +0000 (06:41 +0000)]
Removed define, risc os build, POST-GET bug fixed, AIX 4.3 problems solved
and two makefiles fixed.
Daniel Stenberg [Mon, 3 Mar 2003 23:26:39 +0000 (23:26 +0000)]
output the md5sum as the last step
Daniel Stenberg [Mon, 3 Mar 2003 22:39:34 +0000 (22:39 +0000)]
Added share.obj
Daniel Stenberg [Mon, 3 Mar 2003 22:31:58 +0000 (22:31 +0000)]
moved the disable-thread warning to the switch code so that the AIX 4.3
automatic disable won't cause a warning
Daniel Stenberg [Mon, 3 Mar 2003 22:30:25 +0000 (22:30 +0000)]
Detect AIX 4.3 or later, and if found disable the check for the thread-safe
*_r() functions as they're not needed (and if fact mess things up for us).
Brought to our attention by the friendly Troels Walsted Hansen in bug report
#696217.
Daniel Stenberg [Mon, 3 Mar 2003 22:23:48 +0000 (22:23 +0000)]
AIX 4.3 or later should use gethostbyname() and not the *_r() version.
Daniel Stenberg [Mon, 3 Mar 2003 06:45:27 +0000 (06:45 +0000)]
Added typecast to please the MSVC compiler.
Daniel Stenberg [Mon, 3 Mar 2003 06:42:52 +0000 (06:42 +0000)]
another typecast added to please the borland compiler
Daniel Stenberg [Mon, 3 Mar 2003 06:40:36 +0000 (06:40 +0000)]
Add (void) on our uses of the swrite() macro when we don't read the return
code as this makes compiler warnings. We *should* fix the code to deal with
the return codes instead...
Daniel Stenberg [Sun, 2 Mar 2003 17:43:42 +0000 (17:43 +0000)]
Init postdata properly before issuing a request, so that there isn't any
lingering POST-stuff that confuses GET requests. Juan F. Codagnone reported
this problem in bug report #653859.
Daniel Stenberg [Sun, 2 Mar 2003 17:20:59 +0000 (17:20 +0000)]
moved a variable declaration to remove a compiler warnings with the MSVC
compiler, mentioned by Andi Jahja
Daniel Stenberg [Fri, 28 Feb 2003 15:50:05 +0000 (15:50 +0000)]
include the engine stuff
Daniel Stenberg [Fri, 28 Feb 2003 15:49:32 +0000 (15:49 +0000)]
Andres Garcia Garcia updated to build with the most recent OpenSSL and
the recent libcurl changes.
Daniel Stenberg [Fri, 28 Feb 2003 13:11:10 +0000 (13:11 +0000)]
James Bursa made it compile on RISC OS as well.
Daniel Stenberg [Fri, 28 Feb 2003 13:10:54 +0000 (13:10 +0000)]
James Bursa wrote a section about cross-compiling for RISC OS
Daniel Stenberg [Fri, 28 Feb 2003 12:20:08 +0000 (12:20 +0000)]
the strequal and strnequal should now be called with the proper curl_ prefix
Daniel Stenberg [Fri, 28 Feb 2003 12:17:56 +0000 (12:17 +0000)]
Removed the defines for strequal() and strnequal().
Daniel Stenberg [Fri, 28 Feb 2003 08:40:57 +0000 (08:40 +0000)]
recent stuff
Daniel Stenberg [Fri, 28 Feb 2003 07:55:01 +0000 (07:55 +0000)]
mention what happens if size is set to -1
Daniel Stenberg [Fri, 28 Feb 2003 07:53:31 +0000 (07:53 +0000)]
spell out that POSTFIELDS should be url-encoded in most cases
Daniel Stenberg [Thu, 27 Feb 2003 23:10:38 +0000 (23:10 +0000)]
spell better
Daniel Stenberg [Thu, 27 Feb 2003 14:25:54 +0000 (14:25 +0000)]
Updated to better reflect reality. Also displays how the CURLMsg struct
looks like.
Daniel Stenberg [Thu, 27 Feb 2003 12:50:54 +0000 (12:50 +0000)]
It appears that there are FTP-servers that return size 0 for files
when SIZE is used on the file while being in BINARY mode. To work
around that (stupid) behavior, we attempt to parse the RETR response
even if the SIZE returned size zero.
Debugging help from Salvatore Sorrentino on February 26, 2003.
Daniel Stenberg [Wed, 26 Feb 2003 17:09:47 +0000 (17:09 +0000)]
test138 is for RETR without size and without a working SIZE
Daniel Stenberg [Wed, 26 Feb 2003 17:05:36 +0000 (17:05 +0000)]
support <size>-1</size> to completely disable the SIZE command
Daniel Stenberg [Wed, 26 Feb 2003 16:57:00 +0000 (16:57 +0000)]
added support for RETRNOSIZE in the control file to tell RETR to not
include size in the 150-reply
Daniel Stenberg [Wed, 26 Feb 2003 16:56:00 +0000 (16:56 +0000)]
added a test case for RETR that doesn't get the size in the 150-reply
Daniel Stenberg [Wed, 26 Feb 2003 13:46:38 +0000 (13:46 +0000)]
added index.html
Daniel Stenberg [Wed, 26 Feb 2003 13:01:29 +0000 (13:01 +0000)]
random updates
Daniel Stenberg [Wed, 26 Feb 2003 12:42:25 +0000 (12:42 +0000)]
No longer loop to read multiple times before returning back from the transfer
function, as this could easily end up looping for a very long time (more or
less until the whole transfer was done) and no library-using app would want
that.
Found thanks to a report by Kyle Sallee.
Daniel Stenberg [Tue, 25 Feb 2003 08:52:32 +0000 (08:52 +0000)]
updated, now features less mentions about older versions
Daniel Stenberg [Tue, 25 Feb 2003 08:36:29 +0000 (08:36 +0000)]
better sslcerts link
Daniel Stenberg [Mon, 24 Feb 2003 18:14:48 +0000 (18:14 +0000)]
7.10.4-pre2 commit
Daniel Stenberg [Mon, 24 Feb 2003 16:53:53 +0000 (16:53 +0000)]
Fixes to bring back the the "Expect: 100-continue" functionality. If the
header is used, we must wait for a 100-code (or timeout), before we send the
data. The timeout is merely 1000 ms at this point. We may have reason to set
a longer timeout in the future.
Daniel Stenberg [Mon, 24 Feb 2003 14:50:20 +0000 (14:50 +0000)]
Kjetil Jacobsen found out that setting CURLOPT_MAXCONNECTS to a value higher
than 5 could cause a segfault.