The compiler test used a variable before it was assigned when it tried
to see how it acts on a mismatching prototype, which could cause a false
positive.
Petr Písař [Tue, 20 Aug 2013 15:02:53 +0000 (17:02 +0200)]
Pass password to OpenSSL engine by user interface
Recent OpenSSL uses user interface abstraction to negotiate access to
private keys in the cryprographical engines. An OpenSSL application is
expected to implement the user interface. Otherwise a default one
provided by OpenSSL (interactive standard I/O) will be used and the
aplication will have no way how to pass a password to the engine.
Gisle Vanem [Fri, 6 Sep 2013 11:56:35 +0000 (13:56 +0200)]
http2: use correct include for snprintf
Using the first little merge of nghttp2 into libcurl, I stumbeled on the
missing 'snprintf' in MSVCRT. Isn't this how we do it for other libcurl
files? I.e. use 'curl_msnprintf' and not 'snprintf' directly:
When an error occurs parsing an LDAP URL, The ludp->lud_attrs[i] entries
could be freed even though they sometimes point to data within an
allocated area.
This change introduces a lud_attrs_dup[] array for the duplicated string
pointers, and it removes the unused lud_exts array.
Steve Holme [Wed, 4 Sep 2013 20:27:01 +0000 (21:27 +0100)]
imap/smtp: Fixed incorrect SASL mechanism selection with XOAUTH2 servers
XOAUTH2 would be selected in preference to LOGIN and PLAIN if the IMAP
or SMTP server advertised support for it even though a user's password
was supplied but bearer token wasn't.
Modified the selection logic so that XOAUTH2 will only be selected if
the server supports it and A) The curl user/libcurl programmer has
specifically asked for XOAUTH via the ;AUTH=XOAUTH login option or 2)
The bearer token is specified. Obviously if XOAUTH is asked for via
the login option but no token is specified the user will receive a
authentication failure which makes more sense than no known
authentication mechanisms supported!
Daniel Stenberg [Tue, 3 Sep 2013 22:01:03 +0000 (00:01 +0200)]
acinclude: fix --without-ca-path when cross-compiling
The commit 7b074a460b64811 to CURL_CHECK_CA_BUNDLE in 7.31 (don't check
for paths when cross-compiling) causes --without-ca-path to no longer
works when cross-compiling, since ca and capath only ever get set to
"no" when not cross-compiling, I attach a patch that works for me. Also
in the cross-compilation case, no ca-path seems to be a better default
(IMVHO) than empty ca-path.
Bug: http://curl.haxx.se/bug/view.cgi?id=1273 Patch-by: Stefan Neis
Daniel Stenberg [Thu, 29 Aug 2013 07:57:14 +0000 (09:57 +0200)]
easy: rename struct monitor to socketmonitor
'struct monitor', introduced in 6cf8413e, already exists in an IRIX
header file (sys/mon.h) which gets included via various standard headers
by lib/easy.c
cc-1101 cc: ERROR File = ../../curl/lib/easy.c, Line = 458
"monitor" has already been declared in the current scope.
Steve Holme [Wed, 28 Aug 2013 21:58:33 +0000 (22:58 +0100)]
ftpserver.pl: Corrected accidental move of logmsg() call
Corrected the call to logmsg() in the IMAP SEARCH handler from commit 4ae7b7ea691497 as it should have been outputting the what argument and
not the test number.
When waiting for a 100-continue response from the server, the
Curl_readwrite() will refuse to run if called until the timeout has been
reached.
We timeout code in multi_socket() allows code to run slightly before the
actual timeout time, so for test 154 it could lead to the function being
executed but refused in Curl_readwrite() and then the application would
just sit idling forever.
This was detected with runtests.pl -e on test 154.
Steve Holme [Sun, 25 Aug 2013 21:32:34 +0000 (22:32 +0100)]
curl: Moved check for password out of get parameter loop
Moved the calls to checkpasswd() out of the getparameter() function
which allows for any related arguments to be specified on the command
line before or after --user (and --proxy-user).
For example: --bearer doesn't need to be specified before --user to
prevent curl from asking for an unnecessary password as is the case
with commit e7dcc454c67a2f.
Daniel Stenberg [Mon, 26 Aug 2013 09:51:18 +0000 (11:51 +0200)]
security.h: rename to curl_sec.h to avoid name collision
I brought back security.h in commit bb5529331334e. As we actually
already found out back in 2005 in commit 62970da675249, the file name
security.h causes problems so I renamed it curl_sec.h instead.
Daniel Stenberg [Mon, 26 Aug 2013 07:17:55 +0000 (09:17 +0200)]
runtests.pl: allow -vc point to a separate curl binary to verify with
The specified curl binary will then be used to verify the running
server(s) instead of the development version. This is very useful in
some cases when the development version fails to verify correctly as
then the test case may not run at all.
The actual test will still be run with the "normal" curl executable
(unless the test case specifies something differently).
Kyle L. Huff [Sun, 25 Aug 2013 17:17:58 +0000 (13:17 -0400)]
smtp: added basic SASL XOAUTH2 support
Added the ability to use an XOAUTH2 bearer token [RFC6750] with SMTP for
authentication using RFC6749 "OAuth 2.0 Authorization Framework".
The bearer token is expected to be valid for the user specified in
conn->user. If CURLOPT_XOAUTH2_BEARER is defined and the connection has
an advertised auth mechanism of "XOAUTH2", the user and access token are
formatted as a base64 encoded string and sent to the server as
"AUTH XOAUTH2 <bearer token>".
Kyle L. Huff [Sun, 25 Aug 2013 17:17:35 +0000 (13:17 -0400)]
imap: added basic SASL XOAUTH2 support
Added the ability to use an XOAUTH2 bearer token [RFC6750] with IMAP for
authentication using RFC6749 "OAuth 2.0 Authorization Framework".
The bearer token is expected to be valid for the user specified in
conn->user. If CURLOPT_XOAUTH2_BEARER is defined and the connection has
an advertised auth mechanism of "XOAUTH2", the user and access token are
formatted as a base64 encoded string and sent to the server as
"A001 AUTHENTICATE XOAUTH2 <bearer token>".
Kyle L. Huff [Sun, 25 Aug 2013 17:17:20 +0000 (13:17 -0400)]
options: added basic SASL XOAUTH2 support
Added the ability to specify an XOAUTH2 bearer token [RFC6750] via the
option CURLOPT_XOAUTH2_BEARER for authentication using RFC6749 "OAuth
2.0 Authorization Framework".
Kyle L. Huff [Sun, 25 Aug 2013 17:17:01 +0000 (13:17 -0400)]
sasl: added basic SASL XOAUTH2 support
Added the ability to generated a base64 encoded XOAUTH2 token
containing: "user=<username>^Aauth=Bearer <bearer token>^A^A"
as per RFC6749 "OAuth 2.0 Authorization Framework".
Daniel Stenberg [Fri, 23 Aug 2013 14:16:42 +0000 (16:16 +0200)]
FTP: remove krb4 support
We've announced this pending removal for a long time and we've
repeatedly asked if anyone would care or if anyone objects. Nobody has
objected. It has probably not even been working for a good while since
nobody has tested/used this code recently.
The stuff in krb4.h that was generic enough to be used by other sources
is now present in security.h
Daniel Stenberg [Wed, 21 Aug 2013 21:08:38 +0000 (23:08 +0200)]
multi: move on from STATE_DONE faster
Make sure we always return CURLM_CALL_MULTI_PERFORM when we reach
CURLM_STATE_DONE since the state is transient and it can very well
continue executing as there is nothing to wait for.
Bug: http://curl.haxx.se/mail/lib-2013-08/0211.html Reported-by: Yi Huang
Gisle Vanem [Wed, 21 Aug 2013 20:19:52 +0000 (22:19 +0200)]
curl_easy_perform_ev: make it CURL_EXTERN
I build curl.exe (using MingW) with '-DCURLDEBUG' and by importing from
libcurl.dll. Which means the new curl_easy_perform_ev() must be
exported from libcurl.dll.
Daniel Stenberg [Tue, 20 Aug 2013 10:27:50 +0000 (12:27 +0200)]
CURLM_ADDED_ALREADY: new error code
Doing curl_multi_add_handle() on an easy handle that is already added to
a multi handle now returns this error code. It previously returned
CURLM_BAD_EASY_HANDLE for this condition.
Daniel Stenberg [Tue, 20 Aug 2013 12:48:08 +0000 (14:48 +0200)]
multi_init: moved init code here from add_handle
The closure_handle is "owned" by the multi handle and it is
unconditional so the setting up of it should be in the Curl_multi_handle
function rather than curl_multi_add_handle.
Daniel Stenberg [Sat, 10 Aug 2013 20:55:59 +0000 (22:55 +0200)]
curl_easy_perform_ev: debug/test function
This function is meant to work *exactly* as curl_easy_perform() but will
use the event-based libcurl API internally instead of
curl_multi_perform(). To avoid relying on an actual event-based library
and to not use non-portable functions (like epoll or similar), there's a
rather inefficient emulation layer implemented on top of Curl_poll()
instead.
There's currently some convenience logging done in curl_easy_perform_ev
which helps when tracking down problems. They may be suitable to remove
or change once things seem to be fine enough.
curl has a new --test-event option when built with debug enabled that
then uses curl_easy_perform_ev() instead of curl_easy_perform(). If
built without debug, using --test-event will only output a warning
message.
NOTE: curl_easy_perform_ev() is not part if the public API on purpose.
It is only present in debug builds of libcurl and MUST NOT be considered
stable even then. Use it for libcurl-testing purposes only.
runtests.pl now features an -e command line option that makes it use
--test-event for all curl command line tests. The man page is updated.
Jonathan Nieder [Mon, 19 Aug 2013 08:36:46 +0000 (01:36 -0700)]
url: handle arbitrary-length username and password before '@'
libcurl quietly truncates usernames, passwords, and options from
before an '@' sign in a URL to 255 (= MAX_CURL_PASSWORD_LENGTH - 1)
characters to fit in fixed-size buffers on the stack. Allocate a
buffer large enough to fit the parsed fields on the fly instead to
support longer passwords.
After this change, there are no more uses of MAX_CURL_OPTIONS_LENGTH
left, so stop defining that constant while at it. The hardcoded max
username and password length constants, on the other hand, are still
used in HTTP proxy credential handling (which this patch doesn't
touch).
Jonathan Nieder [Mon, 19 Aug 2013 08:01:26 +0000 (01:01 -0700)]
url: handle exceptional cases first in parse_url_login()
Instead of nesting "if(success)" blocks and leaving the reader in
suspense about what happens in the !success case, deal with failure
cases early, usually with a simple goto to clean up and return from
the function.
No functional change intended. The main effect is to decrease the
indentation of this function slightly.
Jonathan Nieder [Mon, 19 Aug 2013 07:57:54 +0000 (00:57 -0700)]
Curl_setopt: handle arbitrary-length username and password
libcurl truncates usernames, passwords, and options set with
curl_easy_setopt to 255 (= MAX_CURL_PASSWORD_LENGTH - 1) characters.
This doesn't affect the return value from curl_easy_setopt(), so from
the caller's point of view, there is no sign anything strange has
happened, except that authentication fails.
For example:
# Prepare a long (300-char) password.
s=0123456789; s=$s$s$s$s$s$s$s$s$s$s; s=$s$s$s;
# Start a server.
nc -l -p 8888 | tee out & pid=$!
# Tell curl to pass the password to the server.
curl --user me:$s http://localhost:8888 & sleep 1; kill $pid
# Extract the password.
userpass=$(
awk '/Authorization: Basic/ {print $3}' <out |
tr -d '\r' |
base64 -d
)
password=${userpass#me:}
echo ${#password}
Expected result: 300
Actual result: 255
The fix is simple: allocate appropriately sized buffers on the heap
instead of trying to squeeze the provided values into fixed-size
on-stack buffers.
Jonathan Nieder [Mon, 19 Aug 2013 07:48:24 +0000 (00:48 -0700)]
netrc: handle longer username and password
libcurl truncates usernames and passwords it reads from .netrc to
LOGINSIZE and PASSWORDSIZE (64) characters without any indication to
the user, to ensure the values returned from Curl_parsenetrc fit in a
caller-provided buffer.
Fix the interface by passing back dynamically allocated buffers
allocated to fit the user's input. The parser still relies on a
256-character buffer to read each line, though.
So now you can include an ~246-character password in your .netrc,
instead of the previous limit of 63 characters.
Jonathan Nieder [Mon, 19 Aug 2013 07:38:08 +0000 (00:38 -0700)]
url: use goto in create_conn() for exception handling
Instead of remembering before each "return" statement which temporary
allocations, if any, need to be freed, take care to set pointers to
NULL when no longer needed and use a goto to a common block to exit
the function and free all temporaries.
No functional change intended. Currently the only temporary buffer in
this function is "proxy" which is already correctly freed when
appropriate, but there will be more soon.