Pierre Chapuis [Thu, 10 May 2012 16:50:51 +0000 (18:50 +0200)]
autoconf: improve handling of versioned symbols
It checks whether versioned symbols should be enabled before checking
whether it is possible (i.e. the linker supports --version-script) or
not. This avoids a useless warning when building cURL on a platform that
does not use GNU ld.
Moreover, it fixes broken indentation of this chunk of code.
Roman Mamedov spotted (in
http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=670126) that curl would
not complain when given a URL with an IPv6 numerical address without
brackets. It would simply cut off the last ":[hex]" part and thus not
work correctly.
That's a URL using an illegal syntax and now libcurl will instead return
a clear error code and error message detailing the error.
The above mentioned bug report claims this to be a regression but
libcurl does not guarantee functionality when given URLs that aren't
following the URL spec (RFC3986 mostly). I consider the fact that it
used to handle this differently a mere coincidence.
Daniel Stenberg [Fri, 20 Apr 2012 18:40:55 +0000 (20:40 +0200)]
HTTP: empty chunked POST ended up in two zero size chunks
When doing a chunked-encoded POST with -d (CURLOPT_POSTFIELDS) and the
size of the POST was zero length, it made libcurl first send a zero
chunk and then the terminating one. This could confuse a receiver and it
should rather just send the terminating chunk as it does with this fix.
Daniel Stenberg [Thu, 19 Apr 2012 21:31:50 +0000 (23:31 +0200)]
singleipconnect: return OK even when Curl_socket() fails
Commit 9109cdec11ee5a brought this regression (shipped since 7.24.0).
The singleipconnect() function must not return an error if Curl_socket()
returns an error. It should then simply return OK and pass a SOCKET_BAD
back simply because that is how the user of this function expects it to
work and something else is not fine.
Yang Tse [Wed, 18 Apr 2012 19:03:20 +0000 (21:03 +0200)]
Some explicit conversion to 'long' of curl_easy_setopt() third argument
Explicit conversion to 'long' of curl_easy_setopt() third argument for options
CURLOPT_HTTPAUTH and CURLOPT_PROXYAUTH given that this is how its bitmasks are
docummented to be used.
Daniel Stenberg [Mon, 16 Apr 2012 22:45:20 +0000 (00:45 +0200)]
-# progress meter: avoid superfluous updates and duplicate lines
By comparing if a different "progress point" is reached or not since the
previous update, the progress function callback for this now avoids many
superfluous screen updates. This has the nice side-effect that it fixes
a problem that causes a second progress meter line.
The second line output happened because when we use the -# progress
meter, we force a newline output after the transfer in the main loop in
curl, but when libcurl calls the progress callback from
curl_easy_cleanup() it would then output the progress display
again. Possibly the naive newline output is wrong but this optimization
was suitable anyway...
Reported by: Daniel Theron
Bug: http://curl.haxx.se/bug/view.cgi?id=3517418
Yang Tse [Thu, 12 Apr 2012 17:13:23 +0000 (19:13 +0200)]
OS400/initscript.sh: fix db2_name() module name generation
Allow repeatable file name length reduction on file names with underscore or
dash characters. This is done in order to better support libcurl's existing
source file names and allow OS/400 package to build out of the box again.
Yang Tse [Wed, 11 Apr 2012 17:33:54 +0000 (19:33 +0200)]
build adjustments: CURL_HIDDEN_SYMBOLS no longer defined in config files
configure script now provides conditional definitions for Makefile.am
that result in CURL_HIDDEN_SYMBOLS being defined by resulting makefiles
when appropriate.
Additionally, configure script option for symbol hiding control is now
named --enable-symbol-hiding --disable-symbol-hiding. While still valid,
old option name --enable-hidden-symbols --disable-hidden-symbols will
be deprecated in some future release.
Yang Tse [Mon, 9 Apr 2012 19:24:16 +0000 (21:24 +0200)]
configure: Windows cross-compilation fixes
BUILDING_LIBCURL and CURL_STATICLIB are no longer defined in curl_config.h,
configure will generate appropriate conditionals so that mentioned symbols
get defined and used in Makefiles at compilation time
Yang Tse [Fri, 6 Apr 2012 21:35:15 +0000 (23:35 +0200)]
curl tool: use configuration files from lib directory
Configuration files such as curl_config.h and all config-*.h no longer exist
nor are generated/copied into 'src' directory, now these only exist in 'lib'
directory from where curl tool sources uses them.
Additionally old src/setup.h has been refactored into src/tool_setup.h which
now pulls lib/setup.h
The possibility of a makefile needing an include path adjustment exists.
Daniel Stenberg [Tue, 3 Apr 2012 22:07:26 +0000 (00:07 +0200)]
runtests: yassl and polarssl are not openssl
Don't set the "has_openssl" variable if yassl or polarssl is found as
they will simply not work as 100% drop-in replacements for some of the
stuff the "OpenSSL" feature is used for.
I spotted this problem when doing test runs with PolarSSL builds.
connect.c: return changed to CURLE_COULDNT_CONNECT when opensocket fails
Curl_socket returns CURLE_COULDNT_CONNECT when the opensocket callback
returns CURL_SOCKET_BAD. Previous return value CURLE_FAILED_INIT
conveys incorrect information to the user.
Steve Holme [Mon, 2 Apr 2012 22:24:00 +0000 (23:24 +0100)]
pop3: Reworked the command sending and handling
Reworked the command sending from two specific LIST and RETR command
functions into a single command based function as well as the two
associated response handlers into a generic command handler.
Dave Reisner [Sun, 4 Mar 2012 23:07:03 +0000 (23:07 +0000)]
curl tool: add filename_effective token for --write-out
By modifying the parameter list for ourWriteOut() and passing the
OutStruct that collects data in tool_operate, we get access to the
remote name that we're writing to. Shell scripters should find this
useful when used in conjuntion with the --remote-header-name option.
If an empty string is passed to CURLOPT_SSH_PUBLIC_KEYFILE, libcurl will
pass no public key to libssh2 which then tries to compute it from the
private key. This is known to work when libssh2 1.4.0+ is linked against
OpenSSL.
OpenSSL: Made cert hostname check conform to RFC 6125
This change replaces RFC 2818 based hostname check in OpenSSL build with
RFC 6125 [1] based one.
The hostname check in RFC 2818 is ambiguous and each project implements
it in the their own way and they are slightly different. I check curl,
gnutls, Firefox and Chrome and they are all different.
I don't think there is a bug in current implementation of hostname
check. But it is not as strict as the modern browsers do. Currently,
curl allows multiple wildcard character '*' and it matches '.'. (as
described in the comment in ssluse.c).
Firefox implementation is also based on RFC 2818 but it only allows at
most one wildcard character and it must be in the left-most label in the
pattern and the wildcard must not be followed by any character in the
label.[2] Chromium implementation is based on RFC 6125 as my patch does.
Firefox and Chromium both require wildcard in the left-most label in the
presented identifier.
This patch is more strict than the current implementation, so there may
be some cases where old curl works but new one does not. But at the same
time I think it is good practice to follow the modern browsers do and
follow the newer RFC.
Daniel Stenberg [Fri, 23 Mar 2012 22:42:37 +0000 (23:42 +0100)]
HTTP: reset expected DL/UL sizes on redirects
With FOLLOWLOCATION enabled. When a 3xx page is downloaded and the
download size was known (like with a Content-Length header), but the
subsequent URL (transfered after the 3xx page) was chunked encoded, then
the previous "known download size" would linger and cause the progress
meter to get incorrect information, ie the former value would remain
being sent in. This could easily result in downloads that were WAY
larger than "expected" and would cause >100% outputs with the curl
command line tool.
Test case 599 was created and it was used to repeat the bug and then
verify the fix.
Bug: http://curl.haxx.se/bug/view.cgi?id=3510057
Reported by: Michael Wallner
Gökhan Şengün [Fri, 30 Mar 2012 20:09:47 +0000 (23:09 +0300)]
md5: Add support for calculating the md5 sum of buffers incrementally
It is now possible to calculate the md5 sum as the stream of buffers
becomes known where as previously it was only possible to calculate the
md5 sum of a pre-prepared buffer.