platform/upstream/curl.git
13 years agoCURLE_TLSAUTH_FAILED: removed
Quinn Slack [Wed, 9 Feb 2011 22:34:30 +0000 (23:34 +0100)]
CURLE_TLSAUTH_FAILED: removed

On second thought, I think CURLE_TLSAUTH_FAILED should be eliminated. It
was only being raised when an internal error occurred while allocating
or setting the GnuTLS SRP client credentials struct. For TLS
authentication failures, the general CURLE_SSL_CONNECT_ERROR seems
appropriate; its error string already includes "passwords" as a possible
cause. Having a separate TLS auth error code might also cause people to
think that a TLS auth failure means the wrong username or password was
entered, when it could also be a sign of a man-in-the-middle attack.

13 years agoTLS-SRP: new options documented
Quinn Slack [Wed, 9 Feb 2011 22:33:06 +0000 (23:33 +0100)]
TLS-SRP: new options documented

13 years agoCURLOPT_SOCKOPTFUNCTION: return proper error code
Daniel Stenberg [Wed, 9 Feb 2011 14:36:36 +0000 (15:36 +0100)]
CURLOPT_SOCKOPTFUNCTION: return proper error code

When the callback returns an error, this function must make sure to return
CURLE_ABORTED_BY_CALLBACK properly and not CURLE_OK as before to allow the
callback to properly abort the operation.

13 years agocurl.1: typo in -v description
Daniel Stenberg [Tue, 8 Feb 2011 21:39:04 +0000 (22:39 +0100)]
curl.1: typo in -v description

Reported by: Ian D Allen
Bug: https://bugs.launchpad.net/ubuntu/+source/curl/+bug/714895

Forwarded to us by:

Reported by: Andreas Olsson
Bug: http://curl.haxx.se/bug/view.cgi?id=3175422

13 years agonetrc: Removed dead code.
Julien Chaffraix [Fri, 28 Jan 2011 16:34:17 +0000 (08:34 -0800)]
netrc: Removed dead code.

The main has not been updated from some time and is out of sync with
the code. The code is now tested by several test cases so no need for
a seperate code path.

13 years agonetrc: Tightened up the type checks.
Julien Chaffraix [Fri, 28 Jan 2011 16:20:37 +0000 (08:20 -0800)]
netrc: Tightened up the type checks.

The state should not be anonymous so that we can check if the values
are fine. Added 2 unreachables states to the switch as a result of this
change.

13 years agoimap: Fixed typo in a comment.
Julien Chaffraix [Thu, 27 Jan 2011 15:48:19 +0000 (07:48 -0800)]
imap: Fixed typo in a comment.

13 years agoCurl_gmtime: avoid future mistakes
Daniel Stenberg [Mon, 7 Feb 2011 14:09:24 +0000 (15:09 +0100)]
Curl_gmtime: avoid future mistakes

Document Curl_gmtime() and define away the old functions so that they
won't be used internally again by mistake.

13 years agoCurl_gmtime: added a portable gmtime
Daniel Stenberg [Mon, 7 Feb 2011 14:00:48 +0000 (15:00 +0100)]
Curl_gmtime: added a portable gmtime

Instead of polluting many places with #ifdefs, we create a single place
for this function, and also check return code properly so that a NULL
pointer returned won't cause problems.

13 years agomk-ca-bundle.vbs: use new cacert url
Guenter Knauf [Thu, 3 Feb 2011 08:59:00 +0000 (09:59 +0100)]
mk-ca-bundle.vbs: use new cacert url

The official Mozilla page at http://www.mozilla.org/projects/security/certs/
points out a new place as the "proper" place to get Mozilla's CA certs from
so this script is now updated to use that instead.

Reported by: Daniel Mentz

13 years agomk-ca-bundle.pl: use new cacert url
Daniel Stenberg [Wed, 2 Feb 2011 21:22:15 +0000 (22:22 +0100)]
mk-ca-bundle.pl: use new cacert url

The official Mozilla page at
http://www.mozilla.org/projects/security/certs/ points out a new place
as the "proper" place to get Mozilla's CA certs from so this script is
now updated to use that instead.

Reported by: Daniel Mentz

13 years agossluse: improved error message on SSL_CTX_new failures
Bjoern Sikora [Wed, 2 Feb 2011 17:25:57 +0000 (18:25 +0100)]
ssluse: improved error message on SSL_CTX_new failures

"SSL: couldn't create a context" really isn't that helpful, now it'll
also extract an explanation from OpenSSL and append to the right.

13 years agomulti: fix CURLM_STATE_TOOFAST for multi_socket
Nicholas Maniscalco [Wed, 2 Feb 2011 12:41:22 +0000 (13:41 +0100)]
multi: fix CURLM_STATE_TOOFAST for multi_socket

The code in the toofast state needs to first recalculate the values
before it uses them again since it may have been a while since it last
did it when it reaches this point.

13 years agounit1300: code style cleanup
Daniel Stenberg [Wed, 2 Feb 2011 12:35:55 +0000 (13:35 +0100)]
unit1300: code style cleanup

13 years agoadding unit tests for Curl_llist_remove
Amr Shahin [Sat, 29 Jan 2011 15:33:02 +0000 (17:33 +0200)]
adding unit tests for Curl_llist_remove

13 years agoInclude TLSAUTH options in ILE/RPG binding.
Patrick Monnerat [Mon, 31 Jan 2011 16:07:18 +0000 (17:07 +0100)]
Include TLSAUTH options in ILE/RPG binding.

13 years agofile: add support for CURLOPT_TIMECONDITION
Dave Reisner [Sun, 30 Jan 2011 03:41:07 +0000 (22:41 -0500)]
file: add support for CURLOPT_TIMECONDITION

13 years agotransfer: add Curl_meets_timecondition()
Dave Reisner [Sun, 30 Jan 2011 03:12:33 +0000 (22:12 -0500)]
transfer: add Curl_meets_timecondition()

This will be used by file_do() and Curl_readwrite() as a unified method
of checking to see if a remote document meets the supplied
CURLOPT_TIMEVAL and CURLOPT_TIMECONDITION.

Signed-off-by: Dave Reisner <d@falconindy.com>
13 years agoFAQ: clarified/expanded 6.7 a bit
Daniel Stenberg [Sat, 29 Jan 2011 22:52:40 +0000 (23:52 +0100)]
FAQ: clarified/expanded 6.7 a bit

"6.7 What are my obligations when using libcurl in my commercial apps?"
got the piece about what exactly "in all copies" mean to a user of the
code.

This interpretation is based on what other MIT-like licenses have made
more explicit.

13 years agoCOPYING: update the year to 2011
Daniel Stenberg [Sat, 29 Jan 2011 22:41:15 +0000 (23:41 +0100)]
COPYING: update the year to 2011

The generic copyright year range now includes 2011

13 years agoWindows build: alternative makefile
Pierre Joye [Fri, 28 Jan 2011 21:19:44 +0000 (22:19 +0100)]
Windows build: alternative makefile

This is a separate makefile for MSVC builds. It is deliberately put in
another dir than src/ and lib/ to allow a different build experience
than the previous - at least during a period. Eventually we should
unify.

13 years agotest: add test 580 to the dist
Daniel Stenberg [Thu, 27 Jan 2011 23:23:48 +0000 (00:23 +0100)]
test: add test 580 to the dist

13 years agoSome minor edits including updates to function names
Dan Fandrich [Thu, 27 Jan 2011 22:37:16 +0000 (14:37 -0800)]
Some minor edits including updates to function names

13 years agoares: memory leak fix
Daniel Stenberg [Thu, 27 Jan 2011 13:37:25 +0000 (14:37 +0100)]
ares: memory leak fix

The double name resolve trick used with c-ares could leave allocated
memory in 'temp_ai' if the operation was aborted in the middle.

13 years agoares_query_completed_cb: don't touch invalid data
Daniel Stenberg [Thu, 27 Jan 2011 12:59:11 +0000 (13:59 +0100)]
ares_query_completed_cb: don't touch invalid data

When this callback is called due to the destruction of the ares handle,
the connection pointer passed in as an argument may no longer pointing
to valid data and this function doesn't need to do anything with it
anyway so we make sure it doesn't.

Bug: http://curl.haxx.se/mail/lib-2011-01/0333.html
Reported by: Vsevolod Novikov

13 years agonss: avoid memory leaks and failure of NSS shutdown
Kamil Dudka [Thu, 27 Jan 2011 09:55:02 +0000 (10:55 +0100)]
nss: avoid memory leaks and failure of NSS shutdown

... in case more than one CA is loaded.

Bug: https://bugzilla.redhat.com/670802

13 years agoMention that sftp quote commands can be quoted
Dan Fandrich [Thu, 27 Jan 2011 01:02:33 +0000 (17:02 -0800)]
Mention that sftp quote commands can be quoted

13 years agotests: more multiple headers checks
Daniel Stenberg [Tue, 25 Jan 2011 17:09:54 +0000 (18:09 +0100)]
tests: more multiple headers checks

13 years agoHTTP: memory leak on multiple Location:
Daniel Stenberg [Tue, 25 Jan 2011 11:06:50 +0000 (12:06 +0100)]
HTTP: memory leak on multiple Location:

The HTTP parser allocated memory on each received Location: header
without properly freeing old data. Starting now, the code only considers
the first Location: header and will blissfully ignore subsequent ones.

Bug: http://curl.haxx.se/bug/view.cgi?id=3165129
Reported by: Martin Lemke

13 years agoFAQ: clarify 5.13 how to stop a transfer
Daniel Stenberg [Mon, 24 Jan 2011 21:18:43 +0000 (22:18 +0100)]
FAQ: clarify 5.13 how to stop a transfer

13 years agoFixed C++ style comment not allowed in ISO C90.
Guenter Knauf [Sat, 22 Jan 2011 08:52:09 +0000 (09:52 +0100)]
Fixed C++ style comment not allowed in ISO C90.

13 years agoMention axTLS in some more documentation
Dan Fandrich [Fri, 21 Jan 2011 22:27:10 +0000 (14:27 -0800)]
Mention axTLS in some more documentation

13 years agoruntests.pl: make -s not show skipped tests
Daniel Stenberg [Thu, 20 Jan 2011 22:17:22 +0000 (23:17 +0100)]
runtests.pl: make -s not show skipped tests

13 years agounittest: add 3 tests to test1300
Amr Shahin [Tue, 18 Jan 2011 22:39:47 +0000 (00:39 +0200)]
unittest: add 3 tests to test1300

Testing Curl_llist_insert_next

13 years agoFixed compile using OpenSSL versions < 0.9.4a
Darshan Mody [Thu, 20 Jan 2011 20:11:22 +0000 (12:11 -0800)]
Fixed compile using OpenSSL versions < 0.9.4a

13 years agomain: make the tlsauth options always present
Daniel Stenberg [Wed, 19 Jan 2011 22:14:55 +0000 (23:14 +0100)]
main: make the tlsauth options always present

... to not make the connection between the tool and the libcurl used
tighter than necessary, the tlsauth options are now always present but
if the used libcurl doesn't have TLSAUTH support it will return failure.

Also, replaced strncmp() with strequal to get case insensitive matching.

13 years agosymbols-in-versions: add the new TLSAUTH-SRP symbols
Daniel Stenberg [Wed, 19 Jan 2011 21:45:02 +0000 (22:45 +0100)]
symbols-in-versions: add the new TLSAUTH-SRP symbols

13 years agoconfigure: TLS-SRP wasn't added as a feature
Daniel Stenberg [Wed, 19 Jan 2011 20:27:06 +0000 (21:27 +0100)]
configure: TLS-SRP wasn't added as a feature

Test case 1014 failed since TLS-SRP was correctly set to the features
variable so curl-config --features didn't output it.

13 years agodist: add new certs to tarball
Daniel Stenberg [Wed, 19 Jan 2011 20:02:54 +0000 (21:02 +0100)]
dist: add new certs to tarball

13 years agoTLS-SRP: support added when using GnuTLS
Quinn Slack [Wed, 19 Jan 2011 19:35:02 +0000 (20:35 +0100)]
TLS-SRP: support added when using GnuTLS

13 years agoCURLOPT_SSL_VERIFYPEER: more clarifications
Daniel Stenberg [Wed, 19 Jan 2011 12:19:44 +0000 (13:19 +0100)]
CURLOPT_SSL_VERIFYPEER: more clarifications

The default value is 1.

curl _uses_ a default CA bundle, it doesn't install one.

Drop the references to 7.10 as that is now >8 years old!

13 years agoTheArtOfHttpScripting: extended
Daniel Stenberg [Wed, 19 Jan 2011 12:06:36 +0000 (13:06 +0100)]
TheArtOfHttpScripting: extended

Extended the intial HTTP protcol part and added a mention of --trace and
--trace-ascii.

Replaced most URLs in the text to use example.com instead of all the
made up strange names.

Shortened a bunch of lines.

13 years agocurl_easy_setopt.3: clarify VERIFYHOST/PEER
Daniel Stenberg [Wed, 19 Jan 2011 11:54:17 +0000 (12:54 +0100)]
curl_easy_setopt.3: clarify VERIFYHOST/PEER

Extended the descriptions somewhat and made the options get listed next
to each other.

13 years agoRELEASE-NOTES: mention contributors
Daniel Stenberg [Tue, 18 Jan 2011 22:41:57 +0000 (23:41 +0100)]
RELEASE-NOTES: mention contributors

13 years agoRELEASE-NOTES: synced with 7fcbdd68b9e
Daniel Stenberg [Tue, 18 Jan 2011 22:40:48 +0000 (23:40 +0100)]
RELEASE-NOTES: synced with 7fcbdd68b9e

13 years agoFixed configure define for Win32.
Guenter Knauf [Tue, 18 Jan 2011 14:02:45 +0000 (15:02 +0100)]
Fixed configure define for Win32.

Submitted by Vincent Torri.

13 years agonss: fix a bug in handling of CURLOPT_CAPATH
Kamil Dudka [Tue, 18 Jan 2011 12:53:43 +0000 (13:53 +0100)]
nss: fix a bug in handling of CURLOPT_CAPATH

... and update the curl.1 and curl_easy_setopt.3 man pages such that
they do not suggest to use an OpenSSL utility if curl is not built
against OpenSSL.

Bug: https://bugzilla.redhat.com/669702

13 years agoAvoid redefines.
Guenter Knauf [Mon, 17 Jan 2011 00:27:57 +0000 (01:27 +0100)]
Avoid redefines.

13 years agocurl.1: fix spelling
Daniel Stenberg [Sat, 15 Jan 2011 21:46:03 +0000 (22:46 +0100)]
curl.1: fix spelling

Bug: http://curl.haxx.se/bug/view.cgi?id=3157232
Reported by: John Bradshaw

13 years agoAdded casts to silent gcc warnings.
Guenter Knauf [Fri, 14 Jan 2011 11:39:54 +0000 (12:39 +0100)]
Added casts to silent gcc warnings.

13 years agobuild: BCC - makefile.b32 tweak
Yang Tse [Thu, 13 Jan 2011 14:54:14 +0000 (15:54 +0100)]
build: BCC - makefile.b32 tweak

Get rid of stdout redirection to NUL and move stderr redirection
into RM and RMDIR macros.

13 years agobuild: BCC - makefile.b32 tweak
Yang Tse [Thu, 13 Jan 2011 14:33:34 +0000 (15:33 +0100)]
build: BCC - makefile.b32 tweak

Check for BCCDIR environment var done now as other checks.

13 years agoUse env var for PSDK instead of hardcoded path.
Guenter Knauf [Thu, 13 Jan 2011 04:55:09 +0000 (05:55 +0100)]
Use env var for PSDK instead of hardcoded path.

13 years agoEnabled SSPI support by default.
Guenter Knauf [Thu, 13 Jan 2011 01:55:26 +0000 (02:55 +0100)]
Enabled SSPI support by default.

13 years agoUse MAKE macro with BCC targets.
Guenter Knauf [Thu, 13 Jan 2011 01:48:11 +0000 (02:48 +0100)]
Use MAKE macro with BCC targets.

13 years agodocs - update BCC INSTALL section
Yang Tse [Wed, 12 Jan 2011 22:28:19 +0000 (23:28 +0100)]
docs - update BCC INSTALL section

13 years agobuild: BCC - require Borlands's MAKE for Makefile.b32 processing.
Yang Tse [Wed, 12 Jan 2011 19:53:28 +0000 (20:53 +0100)]
build: BCC - require Borlands's MAKE for Makefile.b32 processing.

13 years agobuild: BCC - recover lost functionality from commit 3d813204260b37289411
Yang Tse [Wed, 12 Jan 2011 17:07:04 +0000 (18:07 +0100)]
build: BCC - recover lost functionality from commit 3d813204260b37289411

Borland's $(MAKEDIR) expands to the path where make.exe is located,
use this feature to define BCCDIR when user has not defined BCCDIR.

13 years agobuild: use external preprocessor cpp32 when building with Borland C
Yang Tse [Wed, 12 Jan 2011 01:35:14 +0000 (02:35 +0100)]
build: use external preprocessor cpp32 when building with Borland C

13 years agobuild: allow usage of Borland 5.5.1 external preprocessor cpp32
Yang Tse [Wed, 12 Jan 2011 01:33:39 +0000 (02:33 +0100)]
build: allow usage of Borland 5.5.1 external preprocessor cpp32

13 years agoDisable LDAP support since BCC headers are insufficient.
Guenter Knauf [Tue, 11 Jan 2011 17:32:38 +0000 (18:32 +0100)]
Disable LDAP support since BCC headers are insufficient.

13 years agoInstead of exiting with error lets set BCCDIR self.
Guenter Knauf [Tue, 11 Jan 2011 17:22:42 +0000 (18:22 +0100)]
Instead of exiting with error lets set BCCDIR self.

13 years agoSome Borland C++ makefile tweaks.
Guenter Knauf [Tue, 11 Jan 2011 16:51:14 +0000 (17:51 +0100)]
Some Borland C++ makefile tweaks.

13 years agoBorland C++ doesnt have struct sockaddr_storage.
Guenter Knauf [Tue, 11 Jan 2011 16:40:13 +0000 (17:40 +0100)]
Borland C++ doesnt have struct sockaddr_storage.

13 years agoconnect: use UDP correctly
Daniel Stenberg [Mon, 10 Jan 2011 23:14:36 +0000 (00:14 +0100)]
connect: use UDP correctly

The idea that the protocol and socktype is part of name resolving in the
libc functions is nuts. We keep the name resolver functions assume
TCP/STREAM and we make sure that when we want to connect to a UDP
service we use the correct UDP/DGRAM set instead. This bug was because
the ->protocol field was not always set correctly.

This bug was only affecting ipv6-disabled non-cares non-threaded builds.

Bug: http://curl.haxx.se/bug/view.cgi?id=3154436
Reported by: "dperham"

13 years agoSSL: fix memory leak
Quinn Slack [Mon, 10 Jan 2011 12:53:01 +0000 (13:53 +0100)]
SSL: fix memory leak

In OOM situation. Follow-up fix to commit a9cd4f4ed49e1a0.

13 years agogtls: fix memory leak
Daniel Stenberg [Wed, 5 Jan 2011 23:47:37 +0000 (00:47 +0100)]
gtls: fix memory leak

Bug: http://curl.haxx.se/mail/lib-2011-01/0079.html
Reported by: Quinn Slack

13 years agoHTTP: HTTP Negotiate authentication using SSPI
Marcel Roelofs [Wed, 5 Jan 2011 16:01:07 +0000 (17:01 +0100)]
HTTP: HTTP Negotiate authentication using SSPI

Only under Windows

13 years agocmake: removed two files
Daniel Stenberg [Thu, 6 Jan 2011 22:41:01 +0000 (23:41 +0100)]
cmake: removed two files

CMake/CheckTypeSize.c.in and CMake/CheckTypeSize.cmake were removed in
the previous cmake commit

13 years agoMade unit_setup() return an error code to abort the test early
Dan Fandrich [Thu, 6 Jan 2011 07:53:24 +0000 (23:53 -0800)]
Made unit_setup() return an error code to abort the test early

This makes it possible to skip the call to unit_stop() in such
cases.  Also use Curl_safefree() in unit test 1302 so it will
pass the memory torture test.

13 years agoSSH: speedcheck clobbered existing error
Daniel Stenberg [Wed, 5 Jan 2011 23:19:17 +0000 (00:19 +0100)]
SSH: speedcheck clobbered existing error

The just added speedcheck must not ruin the error code if already set
due to a problem.

13 years agoCMake: Use upstream CheckTypeSize module
Brad King [Wed, 5 Jan 2011 16:32:41 +0000 (17:32 +0100)]
CMake: Use upstream CheckTypeSize module

The CheckTypeSize module that comes with CMake 2.6.2 and above does
everything we need and also supports cross-compiling.  Avoid duplicating
an older version of it here.  This also fixes a cross-compiling error
because the old line

  include ("${CMAKE_MODULE_PATH}/CheckTypeSize.cmake")

failed because CMAKE_MODULE_PATH is a search path and not a directory.

Signed-off-by: Brad King <brad.king@kitware.com>
13 years agoSSH: acknowledge speedcheck
Daniel Stenberg [Wed, 5 Jan 2011 13:09:53 +0000 (14:09 +0100)]
SSH: acknowledge speedcheck

Check for speedcheck limits during the state machine traversals

13 years agotests: <info> tag corrections
Daniel Stenberg [Tue, 4 Jan 2011 22:12:25 +0000 (23:12 +0100)]
tests: <info> tag corrections

13 years agounittest: 1303 tests Curl_timeleft
Daniel Stenberg [Tue, 4 Jan 2011 22:10:45 +0000 (23:10 +0100)]
unittest: 1303 tests Curl_timeleft

I came up with 33 different ways to call it and verify that it returns the
correct return code.

13 years agocurlcheck.h: add fail()
Daniel Stenberg [Tue, 4 Jan 2011 22:09:19 +0000 (23:09 +0100)]
curlcheck.h: add fail()

fail is a new function/macro that a test case can use to indicate a test
failure for cases when the standard macros are not sufficient.

13 years agoCurl_timeleft: s/conn/data in first argument
Daniel Stenberg [Tue, 4 Jan 2011 22:07:58 +0000 (23:07 +0100)]
Curl_timeleft: s/conn/data in first argument

As the function doesn't really use the connectdata struct but only the
SessionHanadle struct I modified what argument it wants.

13 years agoFixed path to allow out-of-tree builds
Dan Fandrich [Tue, 4 Jan 2011 20:59:44 +0000 (12:59 -0800)]
Fixed path to allow out-of-tree builds

13 years agonss: avoid CURLE_OUT_OF_MEMORY given a file name without any slash
Kamil Dudka [Tue, 4 Jan 2011 12:52:54 +0000 (13:52 +0100)]
nss: avoid CURLE_OUT_OF_MEMORY given a file name without any slash

Bug: https://bugzilla.redhat.com/623663

13 years agoignore: all executable unit test cases
Daniel Stenberg [Tue, 4 Jan 2011 15:51:41 +0000 (16:51 +0100)]
ignore: all executable unit test cases

13 years agotests: add 1302 to the package
Daniel Stenberg [Tue, 4 Jan 2011 15:50:59 +0000 (16:50 +0100)]
tests: add 1302 to the package

13 years agounittest: test base64 encode/decode
Daniel Stenberg [Tue, 4 Jan 2011 15:42:31 +0000 (16:42 +0100)]
unittest: test base64 encode/decode

13 years agocurlcheck.h: avoid using NULL pointers
Daniel Stenberg [Tue, 4 Jan 2011 15:41:42 +0000 (16:41 +0100)]
curlcheck.h: avoid using NULL pointers

13 years agocurlcheck.h: add verify_memory
Daniel Stenberg [Tue, 4 Jan 2011 15:31:54 +0000 (16:31 +0100)]
curlcheck.h: add verify_memory

This check verifies that a pointer contains the correct data.

13 years agocurlcheck.h: add newlines in error messages
Daniel Stenberg [Tue, 4 Jan 2011 15:14:23 +0000 (16:14 +0100)]
curlcheck.h: add newlines in error messages

13 years agounittest: verify curl_strequal
Daniel Stenberg [Tue, 4 Jan 2011 15:13:58 +0000 (16:13 +0100)]
unittest: verify curl_strequal

13 years agoget_cert_chain: support larger data sets
Daniel Stenberg [Tue, 4 Jan 2011 09:20:28 +0000 (10:20 +0100)]
get_cert_chain: support larger data sets

512 bytes turned out too short for some data, so now we allocate a
larger buffer instead

Bug: http://curl.haxx.se/mail/archive-2011-01/0002.html

13 years agoRELEASE-NOTES: synced with 83e9fb21aabbec2
Daniel Stenberg [Mon, 3 Jan 2011 23:07:30 +0000 (00:07 +0100)]
RELEASE-NOTES: synced with 83e9fb21aabbec2

13 years agocurlcheck.h: add fail_if() fix code
Daniel Stenberg [Mon, 3 Jan 2011 22:47:34 +0000 (23:47 +0100)]
curlcheck.h: add fail_if() fix code

The UNITTEST_START and UNITTEST_STOP defines needed to do a new brace
level so that test cases can declare variables fine and still remain
fine C89 code.

13 years agounittests: basic docs
Daniel Stenberg [Mon, 3 Jan 2011 22:47:13 +0000 (23:47 +0100)]
unittests: basic docs

13 years agoignore: unit test files
Daniel Stenberg [Mon, 3 Jan 2011 21:52:50 +0000 (22:52 +0100)]
ignore: unit test files

13 years agounittests: a dedicated feature in tests
Daniel Stenberg [Mon, 3 Jan 2011 21:42:46 +0000 (22:42 +0100)]
unittests: a dedicated feature in tests

The test runner script now knows if unittests can run and the unit test
setup file says it is one. I also made runtests.pl deal with no
<command> tag set, so that the description file can get even simpler.

13 years agounittesting: build a separate static lib
Daniel Stenberg [Mon, 3 Jan 2011 21:16:16 +0000 (22:16 +0100)]
unittesting: build a separate static lib

When configure --enable-debug has been used, all files in lib/ are now
built twice and a separate static library crafted for unit-testing will
be linked. The unit tests in the tests/unit subdir will use that
library.

13 years agounittest: framework for unit-testing
Daniel Stenberg [Sat, 1 Jan 2011 16:33:42 +0000 (17:33 +0100)]
unittest: framework for unit-testing

This is the first approach at doing fairly clean and easy to write and
debug unit tests.

13 years agoSSH: avoid PATH_MAX with alloc
Daniel Stenberg [Mon, 3 Jan 2011 12:34:13 +0000 (13:34 +0100)]
SSH: avoid PATH_MAX with alloc

We cannot assume that PATH_MAX will be enough for the remote path name
so allocating room for it is the only sensible approach.

13 years agoTODO: get rid of PATH_MAX
Daniel Stenberg [Mon, 3 Jan 2011 08:40:40 +0000 (09:40 +0100)]
TODO: get rid of PATH_MAX

13 years agoCurl_nss_connect: avoid PATH_MAX
Daniel Stenberg [Sun, 2 Jan 2011 22:41:49 +0000 (23:41 +0100)]
Curl_nss_connect: avoid PATH_MAX

Since some systems don't have PATH_MAX and it isn't that clever to
assume a fixed maximum path length, the code now allocates buffer space
instead of using stack.

Reported by: Samuel Thibault
Bug: http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=608521

13 years agoSFTP: make pwd output result to header callback
Daniel Stenberg [Sat, 1 Jan 2011 14:35:53 +0000 (15:35 +0100)]
SFTP: make pwd output result to header callback

Sending "pwd" as a QUOTE command only sent the reply to the
DEBUGFUNCTION. Now it also sends an FTP-like header to the header
callback to allow similar operations as with FTP, and apps can re-use
the same parser.

13 years agopubkey_show: allocate buffer to fit any-size result
Daniel Stenberg [Sat, 1 Jan 2011 14:33:57 +0000 (15:33 +0100)]
pubkey_show: allocate buffer to fit any-size result

The loop condition was wrong so keys larger than 340 bits would overflow
the local stack-based buffer.

13 years agoCURLINFO_FTP_ENTRY_PATH: sftp support
Daniel Stenberg [Thu, 30 Dec 2010 22:49:03 +0000 (23:49 +0100)]
CURLINFO_FTP_ENTRY_PATH: sftp support