platform/upstream/curl.git
18 years agoAdded CURLOPT_LOCALPORT and CURLOPT_LOCALPORTRANGE to libcurl. Set with the
Daniel Stenberg [Mon, 30 Jan 2006 08:24:07 +0000 (08:24 +0000)]
Added CURLOPT_LOCALPORT and CURLOPT_LOCALPORTRANGE to libcurl. Set with the
curl tool with --local-port. Plain and simply set the range of ports to bind
the local end of connections to. Implemented on to popular demand.

Not extensively tested. Please let me know how it works.

18 years agoBased on an error report by Philippe Vaucher, we no longer count a retried
Daniel Stenberg [Mon, 30 Jan 2006 08:20:52 +0000 (08:20 +0000)]
Based on an error report by Philippe Vaucher, we no longer count a retried
connection setup as a follow-redirect. It turns out 1) this fails when a FTP
connection is re-setup and 2) it does make the max-redirs counter behave
wrong. This fix was not verified since the reporter vanished, but I believe
this is the right fix nonetheless.

18 years agomore mirrors
Daniel Stenberg [Sun, 29 Jan 2006 13:13:45 +0000 (13:13 +0000)]
more mirrors

18 years agowe should fix the system includes in the public headers to be based on checks
Daniel Stenberg [Sat, 28 Jan 2006 13:14:38 +0000 (13:14 +0000)]
we should fix the system includes in the public headers to be based on checks
of the system instead of depending on what particular systems we think need
various headers

18 years agoinclude sys/select.h on NetBSD as well
Daniel Stenberg [Sat, 28 Jan 2006 13:13:58 +0000 (13:13 +0000)]
include sys/select.h on NetBSD as well

18 years agotypo pointed out by Mike Griffiths
Daniel Stenberg [Fri, 27 Jan 2006 21:23:04 +0000 (21:23 +0000)]
typo pointed out by Mike Griffiths

18 years agoCyrill Osterwalder pointed out that sending "" as data in a header is in
Daniel Stenberg [Fri, 27 Jan 2006 15:01:10 +0000 (15:01 +0000)]
Cyrill Osterwalder pointed out that sending "" as data in a header is in
fact equal to a blank one according to the spec.

18 years agoupdated source header
Daniel Stenberg [Thu, 26 Jan 2006 10:39:25 +0000 (10:39 +0000)]
updated source header

18 years agoMichal Marek provided a patch for FTP that makes libcurl continue to try PASV
Daniel Stenberg [Tue, 24 Jan 2006 14:40:43 +0000 (14:40 +0000)]
Michal Marek provided a patch for FTP that makes libcurl continue to try PASV
even after EPSV returned a positive response code, if libcurl failed to
connect to the port number the EPSV response said. Obviously some people are
going through protocol-sensitive firewalls (or similar) that don't understand
EPSV and then they don't allow the second connection unless PASV was
used. This also called for a minor fix of test case 238.

18 years agoFixed some statements about handling multiple occurrences of options.
Dan Fandrich [Fri, 20 Jan 2006 18:56:27 +0000 (18:56 +0000)]
Fixed some statements about handling multiple occurrences of options.
Tried to make some of the wording a bit more consistent.

18 years agothe second -P fix
Daniel Stenberg [Fri, 20 Jan 2006 17:50:12 +0000 (17:50 +0000)]
the second -P fix

18 years agoDuane Cathey was one of our friends who reported that curl -P [IP]
Daniel Stenberg [Thu, 19 Jan 2006 23:52:03 +0000 (23:52 +0000)]
Duane Cathey was one of our friends who reported that curl -P [IP]
(CURLOPT_FTPPORT) didn't work for ipv6-enabed curls if the IP wasn't a
"native" IP while it works fine for ipv6-disabled builds!

In the process of fixing this, I removed the support for LPRT since I can't
think of many reasons to keep doing it and asking on the mailing list didn't
reveal anyone else that could either. The code that sends EPRT and PORT is
now also a lot simpler than before (IMHO).

18 years agoJon Turner pointed out that doing -P [hostname] with curl (built ipv4-only)
Daniel Stenberg [Thu, 19 Jan 2006 22:02:46 +0000 (22:02 +0000)]
Jon Turner pointed out that doing -P [hostname] with curl (built ipv4-only)
didn't work.

18 years agoclarify what "-P -" does
Daniel Stenberg [Thu, 19 Jan 2006 20:40:08 +0000 (20:40 +0000)]
clarify what "-P -" does

18 years agocorrected factual mistake about BSD license in the krb4.c code
Daniel Stenberg [Thu, 19 Jan 2006 09:53:33 +0000 (09:53 +0000)]
corrected factual mistake about BSD license in the krb4.c code

18 years agoreality sync
Daniel Stenberg [Wed, 18 Jan 2006 12:17:20 +0000 (12:17 +0000)]
reality sync

18 years agoconfigure no longer warns on "missing" if the current path contains a space
Daniel Stenberg [Wed, 18 Jan 2006 10:00:36 +0000 (10:00 +0000)]
configure no longer warns on "missing" if the current path contains a space

18 years agoDavid Shaw: Here is the latest libcurl.m4 autoconf tests. It is updated with
Daniel Stenberg [Tue, 17 Jan 2006 17:39:39 +0000 (17:39 +0000)]
David Shaw: Here is the latest libcurl.m4 autoconf tests.  It is updated with
the latest features and protocols that libcurl supports and has a minor fix to
better deal with the obscure case where someone has more than one libcurl
installed at the same time.

18 years agohappy new year!
Daniel Stenberg [Tue, 17 Jan 2006 07:53:29 +0000 (07:53 +0000)]
happy new year!

18 years agoDavid Shaw finally removed all traces of Gopher and we are now officially
Daniel Stenberg [Mon, 16 Jan 2006 22:14:37 +0000 (22:14 +0000)]
David Shaw finally removed all traces of Gopher and we are now officially
not supporting it. It hasn't been functioning for years anyway, so this is
just finally stating what already was true. And a cleanup at the same time.

18 years agoimproved the description of the -L/--location option
Daniel Stenberg [Mon, 16 Jan 2006 00:00:28 +0000 (00:00 +0000)]
improved the description of the -L/--location option

18 years agoBryan Henderson turned the 'initialized' variable for curl_global_init()
Daniel Stenberg [Sun, 15 Jan 2006 23:55:53 +0000 (23:55 +0000)]
Bryan Henderson turned the 'initialized' variable for curl_global_init()
into a counter, and thus you can now do multiple curl_global_init() and you
are then supposed to do the same amount of calls to curl_global_cleanup().
Bryan also updated the docs accordingly.

18 years agoadjusted to use curl_multi_setopt() to set the callback
Daniel Stenberg [Sun, 15 Jan 2006 23:17:46 +0000 (23:17 +0000)]
adjusted to use curl_multi_setopt() to set the callback

18 years agoadjusted to the new concept of the callback
Daniel Stenberg [Sun, 15 Jan 2006 23:15:24 +0000 (23:15 +0000)]
adjusted to the new concept of the callback

18 years agoAndrew Benham fixed a race condition in the test suite that could cause the
Daniel Stenberg [Fri, 13 Jan 2006 12:16:16 +0000 (12:16 +0000)]
Andrew Benham fixed a race condition in the test suite that could cause the
test script to kill all processes in the current process group!

18 years agoFixed FTP_SKIP_PASV_IP and FTP_USE_EPSV to "do right" when used on FTP thru
Daniel Stenberg [Thu, 12 Jan 2006 22:18:38 +0000 (22:18 +0000)]
Fixed FTP_SKIP_PASV_IP and FTP_USE_EPSV to "do right" when used on FTP thru
HTTP proxy.

18 years agoMichael Jahn fixed ftp over CONNECT
Daniel Stenberg [Thu, 12 Jan 2006 12:40:04 +0000 (12:40 +0000)]
Michael Jahn fixed ftp over CONNECT

18 years agomention the "secret" option as I've got no feedback and it is actually present
Daniel Stenberg [Tue, 10 Jan 2006 23:08:38 +0000 (23:08 +0000)]
mention the "secret" option as I've got no feedback and it is actually present
in 7.15.1

18 years agoWhen using a bad path over FTP, as in when libcurl couldn't CWD into all
Daniel Stenberg [Tue, 10 Jan 2006 23:03:22 +0000 (23:03 +0000)]
When using a bad path over FTP, as in when libcurl couldn't CWD into all
given subdirs, libcurl would still "remember" the full path as if it is the
current directory libcurl is in so that the next curl_easy_perform() would
get really confused if it tried the same path again - as it would not issue
any CWD commands at all, assuming it is already in the "proper" dir.

Starting now, a failed CWD command sets a flag that prevents the path to be
"remembered" after returning.

18 years agoMade the copyright year match the latest modification's year.
Daniel Stenberg [Mon, 9 Jan 2006 13:17:14 +0000 (13:17 +0000)]
Made the copyright year match the latest modification's year.

18 years agoAlexander Lazic improved the getservbyport_r() configure check.
Daniel Stenberg [Mon, 9 Jan 2006 08:31:48 +0000 (08:31 +0000)]
Alexander Lazic improved the getservbyport_r() configure check.

18 years agoone more mirror, now fortunately in Japan
Daniel Stenberg [Sun, 8 Jan 2006 23:28:33 +0000 (23:28 +0000)]
one more mirror, now fortunately in Japan

18 years agouse the proper dash
Daniel Stenberg [Sun, 8 Jan 2006 22:55:13 +0000 (22:55 +0000)]
use the proper dash

18 years agoMike Jean fixed so that the second CONNECT when doing FTP over a HTTP proxy
Daniel Stenberg [Sat, 7 Jan 2006 22:24:16 +0000 (22:24 +0000)]
Mike Jean fixed so that the second CONNECT when doing FTP over a HTTP proxy
actually used a new connection and not sent the second request on the first
socket!

18 years agobuildconf fixes
Daniel Stenberg [Fri, 6 Jan 2006 22:59:34 +0000 (22:59 +0000)]
buildconf fixes

18 years agoAs Alexander Lazic pointed out, run the buildconf from the ares dir if that
Daniel Stenberg [Fri, 6 Jan 2006 22:08:39 +0000 (22:08 +0000)]
As Alexander Lazic pointed out, run the buildconf from the ares dir if that
is present instead of trying to duplicate that stuff in this script.

18 years agoUse $ACLOCAL_FLAGS too, pointed out by Alexander Lazic
Daniel Stenberg [Fri, 6 Jan 2006 22:07:51 +0000 (22:07 +0000)]
Use $ACLOCAL_FLAGS too, pointed out by Alexander Lazic

18 years agosummary of what we have
Daniel Stenberg [Thu, 5 Jan 2006 14:58:34 +0000 (14:58 +0000)]
summary of what we have

18 years agoJames Bursa fixes: find the hosts file on RISC OS, and made it build with
Daniel Stenberg [Thu, 5 Jan 2006 07:57:32 +0000 (07:57 +0000)]
James Bursa fixes: find the hosts file on RISC OS, and made it build with
newer gcc versions that no longer defines "riscos".

18 years agoYang Tse has been helping out
Daniel Stenberg [Thu, 5 Jan 2006 07:56:43 +0000 (07:56 +0000)]
Yang Tse has been helping out

18 years agomodified output to prevent the autobuild system to trap on the 'FAILED' output
Daniel Stenberg [Wed, 4 Jan 2006 23:02:40 +0000 (23:02 +0000)]
modified output to prevent the autobuild system to trap on the 'FAILED' output
mistaking it for an actual failed test case

18 years agoAdded ares_getsock.obj.
Gisle Vanem [Wed, 4 Jan 2006 14:21:57 +0000 (14:21 +0000)]
Added ares_getsock.obj.

18 years agoupdated test programs to use the API as it currently works
Daniel Stenberg [Wed, 4 Jan 2006 14:11:35 +0000 (14:11 +0000)]
updated test programs to use the API as it currently works

18 years agoremoved easy handle argument from proto
Daniel Stenberg [Wed, 4 Jan 2006 14:09:57 +0000 (14:09 +0000)]
removed easy handle argument from proto

18 years agoupcoming new error code
Daniel Stenberg [Wed, 4 Jan 2006 14:09:42 +0000 (14:09 +0000)]
upcoming new error code

18 years agoAndres Garcia made the TFTP test server build with mingw.
Daniel Stenberg [Wed, 4 Jan 2006 10:07:36 +0000 (10:07 +0000)]
Andres Garcia made the TFTP test server build with mingw.

18 years agonot much recent stuff, but still I had this modified locally
Daniel Stenberg [Wed, 4 Jan 2006 10:04:23 +0000 (10:04 +0000)]
not much recent stuff, but still I had this modified locally

18 years agoAdded remake of the arpa/tftp.h file to make the TFTP server build on systems
Daniel Stenberg [Tue, 3 Jan 2006 22:47:07 +0000 (22:47 +0000)]
Added remake of the arpa/tftp.h file to make the TFTP server build on systems
without the real header file.

18 years agokilled trailing whitespace
Daniel Stenberg [Tue, 3 Jan 2006 22:44:58 +0000 (22:44 +0000)]
killed trailing whitespace

18 years agoAndres Garcia made the TFTP test server build with mingw ("I also had to copy
Daniel Stenberg [Tue, 3 Jan 2006 22:19:15 +0000 (22:19 +0000)]
Andres Garcia made the TFTP test server build with mingw ("I also had to copy
the 'tftp.h' file from a linux box, since it doesn't come with mingw.")

18 years agoCURLOPT_PROGRESSFUNCTION is really not a good idea when using the multi
Daniel Stenberg [Tue, 3 Jan 2006 15:53:29 +0000 (15:53 +0000)]
CURLOPT_PROGRESSFUNCTION is really not a good idea when using the multi
interface

18 years agofixed the prototype
Daniel Stenberg [Tue, 3 Jan 2006 15:52:59 +0000 (15:52 +0000)]
fixed the prototype

18 years agomodified to hush compiler warnings
Daniel Stenberg [Tue, 3 Jan 2006 12:18:22 +0000 (12:18 +0000)]
modified to hush compiler warnings

18 years agoRemoved inaccurate comment for upcoming curl_multi_socket() and family.
Daniel Stenberg [Mon, 2 Jan 2006 23:37:48 +0000 (23:37 +0000)]
Removed inaccurate comment for upcoming curl_multi_socket() and family.
Modified the callback proto used for it.

18 years agominor edits
Daniel Stenberg [Mon, 2 Jan 2006 23:32:36 +0000 (23:32 +0000)]
minor edits

18 years agoInitial description of the upcoming curl_multi_timeout() function
Daniel Stenberg [Mon, 2 Jan 2006 23:00:04 +0000 (23:00 +0000)]
Initial description of the upcoming curl_multi_timeout() function

18 years agoI removed the timeout argument from the socket callback and did some other
Daniel Stenberg [Mon, 2 Jan 2006 22:58:56 +0000 (22:58 +0000)]
I removed the timeout argument from the socket callback and did some other
cleanups of this man page. The lengthy description has now also been removed
from curl/multi.h since it immediately got tedious to maintain the info on
two places when I did major updates...

18 years agoInclude <sys/ioctl.h> before redefining ioctl().
Gisle Vanem [Mon, 2 Jan 2006 18:35:58 +0000 (18:35 +0000)]
Include <sys/ioctl.h> before redefining ioctl().

18 years ago1. sws now supports two new "commands" and 2. if built with
Daniel Stenberg [Mon, 2 Jan 2006 12:19:12 +0000 (12:19 +0000)]
1. sws now supports two new "commands" and 2. if built with
CURL_SWS_FORK_ENABLED defined it forks for each new connection and thus can
support any amount of connection clients (used for hiper tests and not for the
standard plain curl test suite)

18 years agowe're working on 1.3.1 (or more)
Daniel Stenberg [Mon, 2 Jan 2006 09:13:56 +0000 (09:13 +0000)]
we're working on 1.3.1 (or more)

18 years agofix questionable compare
Marty Kuhrt [Fri, 30 Dec 2005 00:35:21 +0000 (00:35 +0000)]
fix questionable compare

18 years agofix questionable compare compiler error (unsigned can't be < 0)
Marty Kuhrt [Fri, 30 Dec 2005 00:20:46 +0000 (00:20 +0000)]
fix questionable compare compiler error (unsigned can't be < 0)

18 years agoremoved defunct email address
Marty Kuhrt [Fri, 30 Dec 2005 00:07:25 +0000 (00:07 +0000)]
removed defunct email address

18 years agoremoved .h and .sdl
Marty Kuhrt [Fri, 30 Dec 2005 00:07:25 +0000 (00:07 +0000)]
removed .h and .sdl

18 years agoputting back into dist
Marty Kuhrt [Fri, 30 Dec 2005 00:07:25 +0000 (00:07 +0000)]
putting back into dist

18 years agoput back into dist to lessen build confusion for some
Marty Kuhrt [Fri, 30 Dec 2005 00:07:25 +0000 (00:07 +0000)]
put back into dist to lessen build confusion for some

18 years agoput back into dist since most people didn't want to use SDL
Marty Kuhrt [Fri, 30 Dec 2005 00:07:25 +0000 (00:07 +0000)]
put back into dist since most people didn't want to use SDL

18 years agochanged HAVE_STRTOK to follow CRTL version
Marty Kuhrt [Fri, 30 Dec 2005 00:07:25 +0000 (00:07 +0000)]
changed HAVE_STRTOK to follow CRTL version

18 years agoadded TFTP errors to match curl.h
Marty Kuhrt [Fri, 30 Dec 2005 00:07:25 +0000 (00:07 +0000)]
added TFTP errors to match curl.h

18 years agoKirill Vasiliev fixed the 'release-ssl-dll' target to properly build a
Daniel Stenberg [Fri, 23 Dec 2005 23:40:40 +0000 (23:40 +0000)]
Kirill Vasiliev fixed the 'release-ssl-dll' target to properly build a
static libcurl using openssl as dll.

18 years agoclarified that curl_global_init() isn't thread-safe and that it might affect
Daniel Stenberg [Fri, 23 Dec 2005 23:22:23 +0000 (23:22 +0000)]
clarified that curl_global_init() isn't thread-safe and that it might affect
curl_easy_init() if you don't call curl_global_init() explicitly in your app

18 years agoMention that PKG_CONFIG_PATH is preferred to --with-ssl
Dan Fandrich [Fri, 23 Dec 2005 22:33:06 +0000 (22:33 +0000)]
Mention that PKG_CONFIG_PATH is preferred to --with-ssl

18 years agoThis function was added in c-ares [version]
Daniel Stenberg [Thu, 22 Dec 2005 15:31:32 +0000 (15:31 +0000)]
This function was added in c-ares [version]

18 years agoadded ares_getsock()
Daniel Stenberg [Thu, 22 Dec 2005 15:29:12 +0000 (15:29 +0000)]
added ares_getsock()

18 years agoAdded ares_getsock() to extract sockets to wait for action on, without being
Daniel Stenberg [Thu, 22 Dec 2005 15:27:41 +0000 (15:27 +0000)]
Added ares_getsock() to extract sockets to wait for action on, without being
limited to select().

18 years agoThe inital early embryos to describe the curl_multi_socket() API. Committed
Daniel Stenberg [Thu, 22 Dec 2005 15:11:11 +0000 (15:11 +0000)]
The inital early embryos to describe the curl_multi_socket() API. Committed
now to enable them to get added as web pages easier, they are not ready
for anything "real" just yet.

18 years agothe curl_multi_socket() test application (still using select())
Daniel Stenberg [Thu, 22 Dec 2005 14:14:14 +0000 (14:14 +0000)]
the curl_multi_socket() test application (still using select())

18 years ago#31 curl-config --libs" will include details set in LDFLAGS when configure is
Daniel Stenberg [Thu, 22 Dec 2005 08:33:02 +0000 (08:33 +0000)]
#31 curl-config --libs" will include details set in LDFLAGS when configure is
run that might be needed only for building libcurl.

18 years agoChecking for function getnameinfo and its arguments is finally
Yang Tse [Wed, 21 Dec 2005 20:44:54 +0000 (20:44 +0000)]
Checking for function getnameinfo and its arguments is finally
done in one single function CURL_CHECK_FUNC_GETNAMEINFO which
will only define HAVE_GETNAMEINFO if the function has been found
AND the type of its arguments has been properly been detected

18 years agoUndefine HAVE_GETNAMEINFO if unable to find
Yang Tse [Wed, 21 Dec 2005 17:51:08 +0000 (17:51 +0000)]
Undefine HAVE_GETNAMEINFO if unable to find
proper types to use for getnameinfo args

18 years agoUndefine HAVE_GETNAMEINFO if unable to find
Yang Tse [Wed, 21 Dec 2005 17:20:22 +0000 (17:20 +0000)]
Undefine HAVE_GETNAMEINFO if unable to find
proper types to use for getnameinfo args

18 years agoMake sure we're using 'c-ares' sources and not 'ares' ones.
Yang Tse [Wed, 21 Dec 2005 16:08:35 +0000 (16:08 +0000)]
Make sure we're using 'c-ares' sources and not 'ares' ones.

18 years agoadded our regular source header
Daniel Stenberg [Wed, 21 Dec 2005 09:15:54 +0000 (09:15 +0000)]
added our regular source header

18 years agoallow more evironment variables to control what tools to check for and use
Daniel Stenberg [Wed, 21 Dec 2005 08:09:12 +0000 (08:09 +0000)]
allow more evironment variables to control what tools to check for and use

18 years agouse ACLOCAL even when using 'find' to find the aclocal tool
Daniel Stenberg [Wed, 21 Dec 2005 07:59:03 +0000 (07:59 +0000)]
use ACLOCAL even when using 'find' to find the aclocal tool

18 years agoIn CURL_FUNC_GETNAMEINFO_ARGTYPES, when cross-compiling a windows
Yang Tse [Tue, 20 Dec 2005 23:49:14 +0000 (23:49 +0000)]
In CURL_FUNC_GETNAMEINFO_ARGTYPES, when cross-compiling a windows
target use calling convention WSAAPI for getnameinfo() prototype.
Checking type DWORD as argument 4 and 6 of getnameinfo not needed.

18 years agoexplain tld_check_name()
Daniel Stenberg [Tue, 20 Dec 2005 22:46:12 +0000 (22:46 +0000)]
explain tld_check_name()

18 years agoChanges for PellesC compiler under Win32.
Gisle Vanem [Tue, 20 Dec 2005 22:20:42 +0000 (22:20 +0000)]
Changes for PellesC compiler under Win32.

18 years agoChanges for PellesC compiler under Win32. A bit limited, but
Gisle Vanem [Tue, 20 Dec 2005 22:20:04 +0000 (22:20 +0000)]
Changes for PellesC compiler under Win32. A bit limited, but
we just love swedish products...

18 years agoFix PellesC warning.
Gisle Vanem [Tue, 20 Dec 2005 20:58:51 +0000 (20:58 +0000)]
Fix PellesC warning.

18 years agoChanges for PellesC for Win32. It needs <unistd.h> for 'ssize_t'. Hence the
Gisle Vanem [Tue, 20 Dec 2005 20:48:38 +0000 (20:48 +0000)]
Changes for PellesC for Win32. It needs <unistd.h> for 'ssize_t'. Hence the
rearrangement in ares_process.c.

18 years agofix ioctlsocket detection
Yang Tse [Tue, 20 Dec 2005 20:29:53 +0000 (20:29 +0000)]
fix ioctlsocket detection

18 years agoFix, header checks must be done before using its results.
Yang Tse [Tue, 20 Dec 2005 18:50:37 +0000 (18:50 +0000)]
Fix, header checks must be done before using its results.

18 years agoshiper is the new test tool for the new API
Daniel Stenberg [Tue, 20 Dec 2005 09:19:30 +0000 (09:19 +0000)]
shiper is the new test tool for the new API

18 years agoshow dl speed
Daniel Stenberg [Tue, 20 Dec 2005 09:19:20 +0000 (09:19 +0000)]
show dl speed

18 years agoadded our standard source header
Daniel Stenberg [Tue, 20 Dec 2005 09:02:36 +0000 (09:02 +0000)]
added our standard source header

18 years agofix closing parentheses
Daniel Stenberg [Tue, 20 Dec 2005 08:51:48 +0000 (08:51 +0000)]
fix closing parentheses

18 years agouse AC_PROG_LIBTOOL after AC_DISABLE_SHARED
Daniel Stenberg [Tue, 20 Dec 2005 08:51:34 +0000 (08:51 +0000)]
use AC_PROG_LIBTOOL after AC_DISABLE_SHARED

18 years agoFix quoting
Yang Tse [Tue, 20 Dec 2005 03:23:49 +0000 (03:23 +0000)]
Fix quoting

18 years agoGive third argument to AC_DEFINE_UNQUOTED
Yang Tse [Tue, 20 Dec 2005 02:48:48 +0000 (02:48 +0000)]
Give third argument to AC_DEFINE_UNQUOTED