Yang Tse [Tue, 20 Mar 2007 20:00:40 +0000 (20:00 +0000)]
Fixed: When a signal was caught awaiting for an event using Curl_select()
or Curl_poll() with a non-zero timeout both functions would restart the
specified timeout. This could even lead to the extreme case that if a
signal arrived with a frecuency lower to the specified timeout neither
function would ever exit.
Added experimental symbol definition check CURL_ACKNOWLEDGE_EINTR in
Curl_select() and Curl_poll(). When compiled with CURL_ACKNOWLEDGE_EINTR
defined both functions will return as soon as a signal is caught. Use it
at your own risk, all calls to these functions in the library should be
revisited and checked before fully supporting this feature.
Yang Tse [Mon, 19 Mar 2007 12:02:33 +0000 (12:02 +0000)]
Bryan Henderson fixed the progress function so that it can get called
more frequently allowing same calling frecuency for the client progress
callback, while keeping the once a second frecuency for speed calculations
and internal display of the transfer progress.
Yang Tse [Sun, 18 Mar 2007 04:51:40 +0000 (04:51 +0000)]
Code refactoring, extracting a new function wait_ms() from Curl_select and
Curl_poll() which is called whenever not a single valid file descriptor is
passed to these functions.
Improve readibility using a poll() macro to replace WSApoll().
Dan Fandrich [Mon, 12 Mar 2007 20:50:16 +0000 (20:50 +0000)]
libcurl supplies its own crypto hash functions when SSL is disabled, so
'crypto' tests aren't dependent on SSL. Compiling with
--disable-crypto-auth will cause test failures, however.
Daniel Stenberg [Sat, 10 Mar 2007 12:11:21 +0000 (12:11 +0000)]
- Bryan Henderson introduces two things:
1) the progress callback gets called more frequently (at times)
2) libcurl *might* call the callback when it receives a signal
Dan Fandrich [Fri, 9 Mar 2007 23:39:42 +0000 (23:39 +0000)]
Updated the test harness to add a new "crypto" feature check and updated the
appropriate test case to use it. For now, this is treated the same as the
"SSL" feature because curl doesn't list it separately.
Daniel Stenberg [Fri, 9 Mar 2007 21:51:38 +0000 (21:51 +0000)]
- Justin Fletcher fixed a file descriptor leak in the curl tool when trying to
upload a file it couldn't open. Bug #1676581
(http://curl.haxx.se/bug/view.cgi?id=1676581)
Dan Fandrich [Fri, 9 Mar 2007 21:01:39 +0000 (21:01 +0000)]
Updated the test harness to check for protocol support before running each
test, fixing KNOWN_BUGS #11. Fixed some tests to more accurately specify
their required servers and features.
Yang Tse [Thu, 8 Mar 2007 12:04:30 +0000 (12:04 +0000)]
remove code superceeded by the new method used to force
libtool to skip C++ and Fortran checks in patchset:
http://cool.haxx.se/cvs.cgi/curl/ares/configure.ac.diff?r1=1.60&r2=1.64
Dan Fandrich [Thu, 8 Mar 2007 02:38:49 +0000 (02:38 +0000)]
Added test infrastructure to support basic FTPS tests. This currently
supports only ftps:// URLs with --ftp-ssl-control specified, which
implicitly encrypts the control channel but not the data channels. That
allows stunnel to be used with an unmodified ftp server in exactly the
same way that the test https server is set up.
Added test case 400 as a basic FTPS test.
Yang Tse [Wed, 7 Mar 2007 01:13:03 +0000 (01:13 +0000)]
Autoconf redefines the M4 builtin macro 'm4_undefine' in such a way that
it fails if the macro that is being undefined is not already defined. To
make this work under all cases and be sure that at a certain point some
specific macro isn't defined we must use the following style in configure:
Daniel Stenberg [Fri, 2 Mar 2007 22:42:43 +0000 (22:42 +0000)]
- Robert A. Monat and Shmulik Regev helped out to fix the new */Makefile.vc8
makefiles that are included in the source release archives, generated from
the Makefile.vc6 files by the maketgz script. I also modified the root
Makefile to have a VC variable that defaults to vc6 but can be overridden to
allow it to be used for vc8 as well. Like this:
Yang Tse [Thu, 1 Mar 2007 12:02:17 +0000 (12:02 +0000)]
Do not remove CURLM_STATE_WAITPROXYCONNECT from the CURLMstate enum
in builds with HTTP support disabled to keep consistent enum values
for CURLMstate in all kind of builds.
Daniel Stenberg [Tue, 27 Feb 2007 22:12:15 +0000 (22:12 +0000)]
- Hang Kin Lau found and fixed: When I use libcurl to connect to an https
server through a proxy and have the remote https server port set using the
CURLOPT_PORT option, protocol gets reset to http from https after the first
request.
User defined URL was modified internally by libcurl and subsequent reuse of
the easy handle may lead to connection using a different protocol (if not
originally http).
I found that libcurl hardcoded the protocol to "http" when it tries to
regenerate the URL if CURLOPT_PORT is set. I tried to fix the problem as
follows and it's working fine so far
Daniel Stenberg [Tue, 27 Feb 2007 12:44:14 +0000 (12:44 +0000)]
Somewhat updated, changes include: I tried to be more agnostic about the
specific SSL library that might be used, and I cut out the closepolicy stuff
that we no longer support
Daniel Stenberg [Sun, 25 Feb 2007 11:38:13 +0000 (11:38 +0000)]
- Adam D. Moss made the HTTP CONNECT procedure less blocking when used from
the multi interface. Note that it still does a part of the connection in a
blocking manner.