platform/upstream/curl.git
16 years ago<postcheck> delay no longer needed for this test.
Yang Tse [Wed, 23 Apr 2008 23:58:00 +0000 (23:58 +0000)]
<postcheck> delay no longer needed for this test.

http://curl.haxx.se/mail/lib-2008-04/0392.html

16 years agoimprove synchronization between test harness runtests.pl script
Yang Tse [Wed, 23 Apr 2008 23:55:34 +0000 (23:55 +0000)]
improve synchronization between test harness runtests.pl script
and test harness servers to minimize risk of false test failures.

http://curl.haxx.se/mail/lib-2008-04/0392.html

16 years agoSymbian OS is a.k.a. EPOC32
Dan Fandrich [Wed, 23 Apr 2008 18:29:44 +0000 (18:29 +0000)]
Symbian OS is a.k.a. EPOC32

16 years agoMention that P.I.P.S. is needed for Symbian.
Dan Fandrich [Wed, 23 Apr 2008 18:14:25 +0000 (18:14 +0000)]
Mention that P.I.P.S. is needed for Symbian.

16 years agoList extra files individually instead of by directory to avoid including
Dan Fandrich [Wed, 23 Apr 2008 05:14:14 +0000 (05:14 +0000)]
List extra files individually instead of by directory to avoid including
CVS directories.

16 years agoAdded support for running on Symbian OS.
Dan Fandrich [Tue, 22 Apr 2008 22:53:53 +0000 (22:53 +0000)]
Added support for running on Symbian OS.

16 years agofix minor memory leak triggered upon test failure
Yang Tse [Tue, 22 Apr 2008 13:07:27 +0000 (13:07 +0000)]
fix minor memory leak triggered upon test failure

16 years agoRemove fflush() + fsync() previously introduced accelerated writing of
Yang Tse [Tue, 22 Apr 2008 12:40:05 +0000 (12:40 +0000)]
Remove fflush() + fsync() previously introduced accelerated writing of
server input and response request files of the test harness sws server.

Reintroduce, for test # 1001, the <postcheck> small delay. The delay is
needed even with the accelerated writing of server input and response
request files in test harness sws server.

http://curl.haxx.se/mail/lib-2008-04/0385.html

16 years agoRemove previously introduced small delay to verify if it can be avoided
Yang Tse [Tue, 22 Apr 2008 00:23:18 +0000 (00:23 +0000)]
Remove previously introduced small delay to verify if it can be avoided
with the accelerated writing of server input and response request files
in test harness sws server.

16 years agoAllocate the decompression buffer for the --manual option on the heap
Dan Fandrich [Mon, 21 Apr 2008 23:17:20 +0000 (23:17 +0000)]
Allocate the decompression buffer for the --manual option on the heap
instead of the stack.

16 years agoFixed typo in log message
Dan Fandrich [Mon, 21 Apr 2008 23:16:24 +0000 (23:16 +0000)]
Fixed typo in log message

16 years agoadded Haskell binding, unified the formatting somewhat
Daniel Stenberg [Mon, 21 Apr 2008 21:44:03 +0000 (21:44 +0000)]
added Haskell binding, unified the formatting somewhat

16 years agoHAVE_FFLUSH and HAVE_FSYNC symbol definitions for WIN32 systems
Yang Tse [Mon, 21 Apr 2008 19:17:42 +0000 (19:17 +0000)]
HAVE_FFLUSH and HAVE_FSYNC symbol definitions for WIN32 systems

16 years agoIgnore the result of the postcheck command in torture mode
Dan Fandrich [Mon, 21 Apr 2008 17:19:44 +0000 (17:19 +0000)]
Ignore the result of the postcheck command in torture mode

16 years agoMake the test work with nroffs that use special escaping for bold output.
Dan Fandrich [Mon, 21 Apr 2008 16:57:27 +0000 (16:57 +0000)]
Make the test work with nroffs that use special escaping for bold output.

16 years agoaccelerate the writing of server input and response request files
Yang Tse [Sun, 20 Apr 2008 19:15:08 +0000 (19:15 +0000)]
accelerate the writing of server input and response request files
to disk, trying to defeat file and disk write-behind algorithms

16 years agoAdded test cases 1026 and 1027 to do some rudimentary tests on the --manual
Dan Fandrich [Fri, 18 Apr 2008 22:31:52 +0000 (22:31 +0000)]
Added test cases 1026 and 1027 to do some rudimentary tests on the --manual
and --help options.

16 years agotest 1001 needs a small delay between client part execution and test
Yang Tse [Fri, 18 Apr 2008 17:25:54 +0000 (17:25 +0000)]
test 1001 needs a small delay between client part execution and test
result file verifications to allow the test server to completely write
out all files

16 years agotest 1001 needs a small delay between client part execution and test
Yang Tse [Fri, 18 Apr 2008 17:17:14 +0000 (17:17 +0000)]
test 1001 needs a small delay between client part execution and test
result file verifications to allow the test server to completely write
out all files

16 years agofix compiler warning: enumerated type mixed with another type
Yang Tse [Thu, 17 Apr 2008 11:59:46 +0000 (11:59 +0000)]
fix compiler warning: enumerated type mixed with another type

16 years agoSome trivial changes
Dan Fandrich [Thu, 17 Apr 2008 00:45:33 +0000 (00:45 +0000)]
Some trivial changes

16 years agoAdded test case 555, a variation of 547 but using multi interface instead of
Daniel Stenberg [Wed, 16 Apr 2008 21:11:46 +0000 (21:11 +0000)]
Added test case 555, a variation of 547 but using multi interface instead of
easy. This was reported not working by Penugonda Chenna Reddy in
http://curl.haxx.se/mail/lib-2008-04/0046.html but I fail to repeat that
problem.

16 years agoRemoved 129 and 130 due to lack of response on those issues
Daniel Stenberg [Wed, 16 Apr 2008 14:48:01 +0000 (14:48 +0000)]
Removed 129 and 130 due to lack of response on those issues

16 years agoallow disabling the typechecker by defining CURL_DISABLE_TYPECHECK, as
Michal Marek [Mon, 14 Apr 2008 19:01:41 +0000 (19:01 +0000)]
allow disabling the typechecker by defining CURL_DISABLE_TYPECHECK, as
discussed in http://curl.haxx.se/mail/lib-2008-04/0291.html

16 years ago- Stefan Krause reported a case where the OpenSSL handshake phase wasn't
Daniel Stenberg [Mon, 14 Apr 2008 15:26:34 +0000 (15:26 +0000)]
- Stefan Krause reported a case where the OpenSSL handshake phase wasn't
  properly acknowledging the timeout values, like if you pulled the network
  plug in the midst of it.

16 years ago- Andre Guibert de Bruet fixed a second case of not checking the malloc()
Daniel Stenberg [Mon, 14 Apr 2008 15:22:45 +0000 (15:22 +0000)]
- Andre Guibert de Bruet fixed a second case of not checking the malloc()
  return code in the Negotiate code.

16 years agomention Sandor as contributor
Daniel Stenberg [Mon, 14 Apr 2008 15:20:07 +0000 (15:20 +0000)]
mention Sandor as contributor

16 years ago- Sandor Feldi reported bug #1942022
Daniel Stenberg [Mon, 14 Apr 2008 15:19:05 +0000 (15:19 +0000)]
- Sandor Feldi reported bug #1942022
  (http://curl.haxx.se/bug/view.cgi?id=1942022) pointing out a mistake in the
  lib/Makefile.vc[68] makefiles' release-ssl-dll target.

16 years ago- Brock Noland reported that curl behaved differently depending on which order
Daniel Stenberg [Mon, 14 Apr 2008 14:42:06 +0000 (14:42 +0000)]
- Brock Noland reported that curl behaved differently depending on which order
  you used -i and -I.

16 years ago- Andre Guibert de Bruet found and fixed a case where malloc() was called but
Daniel Stenberg [Sat, 12 Apr 2008 11:50:51 +0000 (11:50 +0000)]
- Andre Guibert de Bruet found and fixed a case where malloc() was called but
  was not checked for a NULL return, in the Negotiate code.

16 years agoreturn 0 not -1 at end of data!
Daniel Stenberg [Sat, 12 Apr 2008 08:35:04 +0000 (08:35 +0000)]
return 0 not -1 at end of data!

16 years agoAdded test case 1025 to test a command-line cookie with Location: following
Dan Fandrich [Thu, 10 Apr 2008 18:18:55 +0000 (18:18 +0000)]
Added test case 1025 to test a command-line cookie with Location: following

16 years agomention what happens to the data when a write callback returns pause
Daniel Stenberg [Thu, 10 Apr 2008 09:06:47 +0000 (09:06 +0000)]
mention what happens to the data when a write callback returns pause

16 years agoclarify the COOKIE option a bit
Daniel Stenberg [Thu, 10 Apr 2008 09:03:26 +0000 (09:03 +0000)]
clarify the COOKIE option a bit

16 years agoAdded test case 1024 to test a scenario similar to the one reported
Dan Fandrich [Thu, 10 Apr 2008 04:21:08 +0000 (04:21 +0000)]
Added test case 1024 to test a scenario similar to the one reported
by Ben Combee where libcurl would send the wrong cookie to a redirected
server.  libcurl was doing the right thing in this test case.

16 years ago"tag" the function referals properly
Daniel Stenberg [Wed, 9 Apr 2008 11:27:47 +0000 (11:27 +0000)]
"tag" the function referals properly

16 years agomake previous compiler warning fix more portable
Yang Tse [Mon, 7 Apr 2008 19:12:35 +0000 (19:12 +0000)]
make previous compiler warning fix more portable

16 years agofix compiler warning: argument is incompatible with corresponding format string conve...
Yang Tse [Mon, 7 Apr 2008 15:40:49 +0000 (15:40 +0000)]
fix compiler warning: argument is incompatible with corresponding format string conversion

16 years agofix compiler warning: indirection to slightly different base types
Yang Tse [Mon, 7 Apr 2008 14:37:17 +0000 (14:37 +0000)]
fix compiler warning: indirection to slightly different base types

16 years agofix compiler warning: local variable may be used without having been initialized
Yang Tse [Mon, 7 Apr 2008 14:20:05 +0000 (14:20 +0000)]
fix compiler warning: local variable may be used without having been initialized

16 years agoAdapt OS400 SSL (qssl.h) to V5R4
Patrick Monnerat [Mon, 7 Apr 2008 13:09:17 +0000 (13:09 +0000)]
Adapt OS400 SSL (qssl.h) to V5R4
Fix qssl.c wrong error message
Upgrade OS400 wrappers and makefiles to 7.18.1

16 years agofix compiler warning: unreferenced formal parameter
Yang Tse [Mon, 7 Apr 2008 12:44:47 +0000 (12:44 +0000)]
fix compiler warning: unreferenced formal parameter

16 years agofix compiler warning: assignment within conditional expression
Yang Tse [Mon, 7 Apr 2008 12:40:05 +0000 (12:40 +0000)]
fix compiler warning: assignment within conditional expression

16 years ago- Fix the MIT / Heimdal check for good:
Michal Marek [Mon, 7 Apr 2008 09:26:30 +0000 (09:26 +0000)]
- Fix the MIT / Heimdal check for good:
  Define HAVE_GSSMIT if <gssapi/{gssapi.h,gssapi_generic.h,gssapi_krb5.h}> are
  available, otherwise define HAVE_GSSHEIMDAL if <gssapi.h> is available.

  Only define GSS_C_NT_HOSTBASED_SERVICE to gss_nt_service_name if
  GSS_C_NT_HOSTBASED_SERVICE isn't declared by the gssapi headers. This should
  avoid breakage in case we wrongly recognize Heimdal as MIT again.

16 years ago- Alexey Simak fixed curl_easy_reset() to reset the max redirect limit properly
Daniel Stenberg [Sat, 5 Apr 2008 21:13:44 +0000 (21:13 +0000)]
- Alexey Simak fixed curl_easy_reset() to reset the max redirect limit properly

16 years agoprovide CURL_DEFAULT_PROXY_PORT set to 1080 for the default port libcurl
Daniel Stenberg [Sat, 5 Apr 2008 21:13:31 +0000 (21:13 +0000)]
provide CURL_DEFAULT_PROXY_PORT set to 1080 for the default port libcurl
assumes proxies to use

16 years ago- Based on the Debian bug report #474224 that complained about the FTP error
Daniel Stenberg [Sat, 5 Apr 2008 21:02:38 +0000 (21:02 +0000)]
- Based on the Debian bug report #474224 that complained about the FTP error
  message when libcurl doesn't get a 220 back immediately on connect, I now
  changed it to be more specific on what the problem is. Also worth noticing:
  while the bug report contains an example where the response is:

    421 There are too many connected users, please try again later

  we cannot assume that the error message will always be this readable nor
  that it fits within a particular boundary etc.

16 years ago- Alexey Simak fixed the VC dsp file by adding the missing source file
Daniel Stenberg [Fri, 4 Apr 2008 20:26:06 +0000 (20:26 +0000)]
- Alexey Simak fixed the VC dsp file by adding the missing source file
  ares_expand_string.c

16 years agoAlexey Simak made adig support NAPTR records
Daniel Stenberg [Fri, 4 Apr 2008 20:24:41 +0000 (20:24 +0000)]
Alexey Simak made adig support NAPTR records

16 years agoEino Tuominen improved the code when a file is used to seed the randomizer
Daniel Stenberg [Fri, 4 Apr 2008 20:05:23 +0000 (20:05 +0000)]
Eino Tuominen improved the code when a file is used to seed the randomizer

16 years agoGive a hint as to why a url_fopen failed.
Dan Fandrich [Fri, 4 Apr 2008 18:45:37 +0000 (18:45 +0000)]
Give a hint as to why a url_fopen failed.

16 years agoSunSSH sshd ignores UsePrivilegeSeparation option
Yang Tse [Fri, 4 Apr 2008 14:47:32 +0000 (14:47 +0000)]
SunSSH sshd ignores UsePrivilegeSeparation option

16 years agoSunSSH 1.2 options sync
Yang Tse [Fri, 4 Apr 2008 14:08:36 +0000 (14:08 +0000)]
SunSSH 1.2 options sync

16 years agooops, forgot to add test554
Daniel Stenberg [Fri, 4 Apr 2008 11:39:32 +0000 (11:39 +0000)]
oops, forgot to add test554

16 years agoAdded test627 to test SFTP with CURLOPT_NOBODY
Dan Fandrich [Fri, 4 Apr 2008 02:06:35 +0000 (02:06 +0000)]
Added test627 to test SFTP with CURLOPT_NOBODY

16 years ago- Setting CURLOPT_NOBODY to FALSE will now switch the HTTP request method to
Daniel Stenberg [Thu, 3 Apr 2008 21:44:49 +0000 (21:44 +0000)]
- Setting CURLOPT_NOBODY to FALSE will now switch the HTTP request method to
  GET simply because previously when you set CURLOPT_NOBODY to TRUE first and
  then FALSE you'd end up in a broken state where a HTTP request would do a
  HEAD by still act a lot like for a GET and hang waiting for the content etc.

16 years agoScott Barrett added support for CURLOPT_NOBODY over SFTP
Daniel Stenberg [Thu, 3 Apr 2008 20:56:59 +0000 (20:56 +0000)]
Scott Barrett added support for CURLOPT_NOBODY over SFTP

16 years agoMade sure that curl_global_init is called in all the multithreaded
Dan Fandrich [Thu, 3 Apr 2008 20:28:32 +0000 (20:28 +0000)]
Made sure that curl_global_init is called in all the multithreaded
example programs.

16 years agoremoved unused var.
Gunter Knauf [Wed, 2 Apr 2008 03:11:34 +0000 (03:11 +0000)]
removed unused var.

16 years agoremoved double dependency.
Gunter Knauf [Wed, 2 Apr 2008 03:08:21 +0000 (03:08 +0000)]
removed double dependency.

16 years ago128 - Phil Blundell's ares and ipv6 fix (feedback lacking)
Daniel Stenberg [Tue, 1 Apr 2008 21:49:51 +0000 (21:49 +0000)]
128 - Phil Blundell's ares and ipv6 fix (feedback lacking)
removed

133 - Setting CURLOPT_NOBODY to "false" causes cURL to wait for content if a
      content-length header is read
added

16 years agoEetu contributed back in 2000...
Daniel Stenberg [Tue, 1 Apr 2008 13:54:49 +0000 (13:54 +0000)]
Eetu contributed back in 2000...

16 years ago132 - Xponaut's CURLFORM_STREAM option to curl_formadd()
Daniel Stenberg [Mon, 31 Mar 2008 20:32:31 +0000 (20:32 +0000)]
132 - Xponaut's CURLFORM_STREAM option to curl_formadd()
done

16 years agomoved CURL_CA_BUNDLE define to generated config.h.
Gunter Knauf [Mon, 31 Mar 2008 12:51:44 +0000 (12:51 +0000)]
moved CURL_CA_BUNDLE define to generated config.h.

16 years agoRemoved the generated ca-bundle.h file. The verbatim value of $ca and
Michal Marek [Mon, 31 Mar 2008 12:09:43 +0000 (12:09 +0000)]
Removed the generated ca-bundle.h file. The verbatim value of $ca and
$capath is known to configure, so it can be defined in config.h instead.

16 years agoexpanded a comment around some of the new formpost callback usage
Daniel Stenberg [Mon, 31 Mar 2008 10:16:34 +0000 (10:16 +0000)]
expanded a comment around some of the new formpost callback usage

16 years ago- Added CURLFORM_STREAM as a supported option to curl_formadd() to allow an
Daniel Stenberg [Mon, 31 Mar 2008 10:02:23 +0000 (10:02 +0000)]
- Added CURLFORM_STREAM as a supported option to curl_formadd() to allow an
  application to provide data for a multipart with the read callback. Note
  that the size needs to be provided with CURLFORM_CONTENTSLENGTH when the
  stream option is used. This feature is verified by the new test case
  554. This feature was sponsored by Xponaut.

16 years agoChanged the makefile so the doc/examples/ programs are never built in a
Dan Fandrich [Mon, 31 Mar 2008 03:01:13 +0000 (03:01 +0000)]
Changed the makefile so the doc/examples/ programs are never built in a
normal build/install (only with the 'make check' target), so that a
build failure in the examples isn't fatal.

16 years agoadded people from the 7.18.1 release announcement
Daniel Stenberg [Sun, 30 Mar 2008 09:30:15 +0000 (09:30 +0000)]
added people from the 7.18.1 release announcement

16 years agoback to a blank for 7.18.2
Daniel Stenberg [Sun, 30 Mar 2008 09:22:45 +0000 (09:22 +0000)]
back to a blank for 7.18.2

16 years agostart working on 7.18.2
Daniel Stenberg [Sun, 30 Mar 2008 09:22:22 +0000 (09:22 +0000)]
start working on 7.18.2

16 years ago7.18.1
Daniel Stenberg [Sun, 30 Mar 2008 09:11:35 +0000 (09:11 +0000)]
7.18.1

16 years ago132 - Xponaut's CURLFORM_STREAM option to curl_formadd()
Daniel Stenberg [Sun, 30 Mar 2008 09:08:40 +0000 (09:08 +0000)]
132 - Xponaut's CURLFORM_STREAM option to curl_formadd()

16 years agoI'm officially pushing the two remaining issues to the next release instead,
Daniel Stenberg [Sun, 30 Mar 2008 09:07:08 +0000 (09:07 +0000)]
I'm officially pushing the two remaining issues to the next release instead,
since they're still not clear enough to be to sort about before 7.18.1

16 years agoMade the test work on perl 5.00
Dan Fandrich [Fri, 28 Mar 2008 18:19:11 +0000 (18:19 +0000)]
Made the test work on perl 5.00

16 years ago- Stephen Collyer pointed out that configure --with-libssh2 without a given
Daniel Stenberg [Thu, 27 Mar 2008 23:13:57 +0000 (23:13 +0000)]
- Stephen Collyer pointed out that configure --with-libssh2 without a given
  path didn't work properly but now it does!

16 years agofix --with-libssh2 when given without path
Daniel Stenberg [Thu, 27 Mar 2008 23:10:24 +0000 (23:10 +0000)]
fix --with-libssh2 when given without path

16 years ago- As found out and reported by Dan Petitt, libcurl didn't show progress/call
Daniel Stenberg [Thu, 27 Mar 2008 13:07:12 +0000 (13:07 +0000)]
- As found out and reported by Dan Petitt, libcurl didn't show progress/call
  the progress callback for the first (potentially huge) piece of body data
  sent together with the POST request headers in the initial send().

16 years agospell!
Daniel Stenberg [Tue, 25 Mar 2008 19:23:26 +0000 (19:23 +0000)]
spell!

16 years ago- Made setting the CURLOPT_SSL_CTX_FUNCTION option return a failure in case
Daniel Stenberg [Tue, 25 Mar 2008 19:19:49 +0000 (19:19 +0000)]
- Made setting the CURLOPT_SSL_CTX_FUNCTION option return a failure in case
  libcurl wasn't built to use OpenSSL as that is a prerequisite for this
  option to function!

16 years agospell it out loudly and clearly that CURLOPT_SSL_CTX_FUNCTION is only
Daniel Stenberg [Tue, 25 Mar 2008 19:17:26 +0000 (19:17 +0000)]
spell it out loudly and clearly that CURLOPT_SSL_CTX_FUNCTION is only
functional if libcurl is built against OpenSSL

16 years ago- Fixed the problem with doing a zero byte SCP transfer, verified with test
Daniel Stenberg [Sat, 22 Mar 2008 22:00:21 +0000 (22:00 +0000)]
- Fixed the problem with doing a zero byte SCP transfer, verified with test
  case 617 (which was added by Daniel Fandrich 5 Mar 2008).

16 years agoroadmap ahead
Daniel Stenberg [Fri, 21 Mar 2008 11:53:59 +0000 (11:53 +0000)]
roadmap ahead

16 years agoJes reported the curl-config bug
Dan Fandrich [Thu, 20 Mar 2008 20:16:01 +0000 (20:16 +0000)]
Jes reported the curl-config bug

16 years agoFixed a problem where curl-config --protocols could erroneously show LDAPS
Dan Fandrich [Thu, 20 Mar 2008 20:08:44 +0000 (20:08 +0000)]
Fixed a problem where curl-config --protocols could erroneously show LDAPS
support when curl didn't even have regular LDAP support.  It looks like
this could happen when the --enable-ldaps configure switch is given but
configure couldn't find the LDAP headers or libraries.

16 years agoadded --extvercmd parameter which can be used to specify an external command to displ...
Gunter Knauf [Thu, 20 Mar 2008 16:10:54 +0000 (16:10 +0000)]
added --extvercmd parameter which can be used to specify an external command to display 'curl --version', f.e. with MinW32 crosscompile --extvercmd=/usr/bin/wine can be used.

16 years ago- Added --with-ca-path=DIRECTORY configure option to use an openSSL CApath by
Michal Marek [Thu, 20 Mar 2008 08:09:23 +0000 (08:09 +0000)]
- Added --with-ca-path=DIRECTORY configure option to use an openSSL CApath by
  default instead of a ca bundle. The configure script will also look for a
  ca path if no ca bundle is found and no option given.

- Fixed detection of previously installed curl-ca-bundle.crt

16 years agoFixed an infinite loop when given an invalid SFTP quote command.
Dan Fandrich [Tue, 18 Mar 2008 22:59:04 +0000 (22:59 +0000)]
Fixed an infinite loop when given an invalid SFTP quote command.

16 years agoAdded test 626 to reproduce an infinite loop when given an invalid
Dan Fandrich [Tue, 18 Mar 2008 17:05:29 +0000 (17:05 +0000)]
Added test 626 to reproduce an infinite loop when given an invalid
SFTP quote command reported by Vincent Le Normand, but left it disabled.

16 years ago- Added curl_easy_getinfo typechecker.
Michal Marek [Tue, 18 Mar 2008 08:14:37 +0000 (08:14 +0000)]
- Added curl_easy_getinfo typechecker.

- Added macros for curl_share_setopt and curl_multi_setopt to check at least
  the correct number of arguments.

16 years agoMark the statement expr with __extension__ so that gcc -pedantic doesn't emit
Michal Marek [Mon, 17 Mar 2008 14:22:10 +0000 (14:22 +0000)]
Mark the statement expr with __extension__ so that gcc -pedantic doesn't emit
any hard-to-grasp warnings in curl_easy_setopt() calls in applications.
Also delete superfluous semicolons.

16 years agoSpell the commands right
Dan Fandrich [Sat, 15 Mar 2008 01:03:23 +0000 (01:03 +0000)]
Spell the commands right

16 years agoAdded tests 622-625 to test SFTP/SCP uploads. Test 625 was an attempt to
Dan Fandrich [Thu, 13 Mar 2008 22:51:39 +0000 (22:51 +0000)]
Added tests 622-625 to test SFTP/SCP uploads. Test 625 was an attempt to
reproduce the --ftp-create-dirs problem reported by Brian Ulm, but that
seems to need a call curl_easy_reset() which this test case doesn't do.

16 years ago- Brian Ulm figured out that if you did an SFTP upload with
Daniel Stenberg [Thu, 13 Mar 2008 21:43:39 +0000 (21:43 +0000)]
- Brian Ulm figured out that if you did an SFTP upload with
  CURLOPT_FTP_CREATE_MISSING_DIRS to create a directory, and then re-used the
  handle and uploaded another file to another directory that needed to be
  created, the second upload would fail. Another case of a state variable that
  wasn't properly reset between requests.

- I rewrote the 100-continue code to use a single state variable instead of
  the previous two ones. I think it made the logic somewhat clearer.

16 years ago- Brian Ulm figured out that if you did an SFTP upload with
Daniel Stenberg [Thu, 13 Mar 2008 21:43:30 +0000 (21:43 +0000)]
- Brian Ulm figured out that if you did an SFTP upload with
  CURLOPT_FTP_CREATE_MISSING_DIRS to create a directory, and then re-used the
  handle and uploaded another file to another directory that needed to be
  created, the second upload would fail. Another case of a state variable that
  wasn't properly reset between requests.

16 years agoChange the confusing two variables for the expect 100 continue stuff into
Daniel Stenberg [Thu, 13 Mar 2008 20:56:13 +0000 (20:56 +0000)]
Change the confusing two variables for the expect 100 continue stuff into
a single state variable to make the code easier to follow and understand.

16 years ago--ftp-create-dirs works on SFTP as well
Dan Fandrich [Thu, 13 Mar 2008 20:49:06 +0000 (20:49 +0000)]
--ftp-create-dirs works on SFTP as well

16 years agofix code that is normally #ifdef'ed out
Daniel Stenberg [Thu, 13 Mar 2008 12:36:22 +0000 (12:36 +0000)]
fix code that is normally #ifdef'ed out

16 years agoDone: 125 - Michal Marek's typechecker-gcc work
Daniel Stenberg [Wed, 12 Mar 2008 13:14:09 +0000 (13:14 +0000)]
Done: 125 - Michal Marek's typechecker-gcc work

16 years agoupdated according to the name resolve race condition fix just committed
Daniel Stenberg [Tue, 11 Mar 2008 22:58:52 +0000 (22:58 +0000)]
updated according to the name resolve race condition fix just committed