platform/upstream/curl.git
16 years agologmsg already appends '\n'
Yang Tse [Tue, 9 Oct 2007 23:21:29 +0000 (23:21 +0000)]
logmsg already appends '\n'

16 years agofix socklen_t for MSVC6 & 7.
Gunter Knauf [Tue, 9 Oct 2007 22:10:17 +0000 (22:10 +0000)]
fix socklen_t for MSVC6 & 7.

16 years agoadded two more module dependencies for LDAPS.
Gunter Knauf [Tue, 9 Oct 2007 20:15:27 +0000 (20:15 +0000)]
added two more module dependencies for LDAPS.

16 years agoDocumented error codes 77-80, and fixed the one for 60.
Dan Fandrich [Tue, 9 Oct 2007 16:49:41 +0000 (16:49 +0000)]
Documented error codes 77-80, and fixed the one for 60.

16 years agoAdd a paragraph about CURLOPT_CUSTOMREQUEST not actually changing libcurl's
Daniel Stenberg [Tue, 9 Oct 2007 14:53:50 +0000 (14:53 +0000)]
Add a paragraph about CURLOPT_CUSTOMREQUEST not actually changing libcurl's
behavior, it only changes the actual request method keyword and this is not
always what the user/app wants.

16 years agoMichal Marek removed the no longer existing return codes from the curl.1
Daniel Stenberg [Tue, 9 Oct 2007 08:42:50 +0000 (08:42 +0000)]
Michal Marek removed the no longer existing return codes from the curl.1
man page.

16 years agoAdded needed 'HAVE_*' defines.
Gisle Vanem [Mon, 8 Oct 2007 14:39:52 +0000 (14:39 +0000)]
Added needed 'HAVE_*' defines.

16 years ago'FD_CLOXEC' is meaningless on MSDOS/Watt-32.
Gisle Vanem [Mon, 8 Oct 2007 14:38:51 +0000 (14:38 +0000)]
'FD_CLOXEC' is meaningless on MSDOS/Watt-32.

16 years agoKnown bug #47, which confused libcurl if doing NTLM auth over a proxy with
Daniel Stenberg [Sun, 7 Oct 2007 08:28:03 +0000 (08:28 +0000)]
Known bug #47, which confused libcurl if doing NTLM auth over a proxy with
a response that was larger than 16KB is now improved slightly so that now
the restriction at 16KB is for the headers only and it should be a rare
situation where the response-headers exceed 16KB. Thus, I consider #47 fixed
and the header limitation is now known as known bug #48.

16 years ago<ws2tcpip.h> needed for 'socklen_t' typedef.
Gisle Vanem [Sat, 6 Oct 2007 17:20:06 +0000 (17:20 +0000)]
<ws2tcpip.h> needed for 'socklen_t' typedef.

16 years agoadd url to the wikipedia article for a longer description
Daniel Stenberg [Fri, 5 Oct 2007 15:18:34 +0000 (15:18 +0000)]
add url to the wikipedia article for a longer description

16 years agoAlexey Pesternikov documented CURLOPT_OPENSOCKETDATA and
Daniel Stenberg [Fri, 5 Oct 2007 15:16:18 +0000 (15:16 +0000)]
Alexey Pesternikov documented CURLOPT_OPENSOCKETDATA and
CURLOPT_OPENSOCKETFUNCTION

16 years agoMichael Wallner made the CULROPT_COOKIELIST option support a new magic
Daniel Stenberg [Fri, 5 Oct 2007 14:37:33 +0000 (14:37 +0000)]
Michael Wallner made the CULROPT_COOKIELIST option support a new magic
string: "FLUSH". Using that will cause libcurl to flush its cookies to the
CURLOPT_COOKIEJAR file.

16 years agoThe new file docs/libcurl/ABI describes how we view ABI breakages, soname
Daniel Stenberg [Thu, 4 Oct 2007 22:05:25 +0000 (22:05 +0000)]
The new file docs/libcurl/ABI describes how we view ABI breakages, soname
bumps and what the version number's significance to all that is.

16 years agoI enabled test 1009 and made the --local-port use a wide range to reduce the
Daniel Stenberg [Thu, 4 Oct 2007 21:26:26 +0000 (21:26 +0000)]
I enabled test 1009 and made the --local-port use a wide range to reduce the
risk of failures.

16 years agoKim Rinnewitz reported that --local-port didn't work with TFTP transfers.
Daniel Stenberg [Thu, 4 Oct 2007 10:01:41 +0000 (10:01 +0000)]
Kim Rinnewitz reported that --local-port didn't work with TFTP transfers.
This happened because the tftp code always uncondionally did a bind()
without caring if one already had been done and then it failed. I wrote a
test case (1009) to verify this, but it is a bit error-prone since it will
have to pick a fixed local port number and since the tests are run on so
many different hosts in different situations I add it in disabled state.

16 years agoRemoved a piece of redundant code (process_answer already takes care of it).
Steinar H. Gunderson [Thu, 4 Oct 2007 08:12:12 +0000 (08:12 +0000)]
Removed a piece of redundant code (process_answer already takes care of it).

16 years agoAnother timeout fix in ares_getnameinfo().
Steinar H. Gunderson [Thu, 4 Oct 2007 08:09:52 +0000 (08:09 +0000)]
Another timeout fix in ares_getnameinfo().

16 years agoSend the timeout count in ares_getnameinfo().
Steinar H. Gunderson [Thu, 4 Oct 2007 08:09:04 +0000 (08:09 +0000)]
Send the timeout count in ares_getnameinfo().

16 years agoMoved the NULL check for channel upwards in ares_destroy().
Steinar H. Gunderson [Thu, 4 Oct 2007 08:07:47 +0000 (08:07 +0000)]
Moved the NULL check for channel upwards in ares_destroy().

16 years agoClarified the comment over ares_cancel.
Steinar H. Gunderson [Thu, 4 Oct 2007 08:06:25 +0000 (08:06 +0000)]
Clarified the comment over ares_cancel.

16 years agoOn error, close "log/server.response"
Yang Tse [Thu, 4 Oct 2007 02:09:33 +0000 (02:09 +0000)]
On error, close "log/server.response"

16 years agoIf TCP_NODELAY is not defined we can't disable the Nagle algorithm
Yang Tse [Wed, 3 Oct 2007 23:38:07 +0000 (23:38 +0000)]
If TCP_NODELAY is not defined we can't disable the Nagle algorithm

16 years agoCleanup no longer used macros
Yang Tse [Wed, 3 Oct 2007 16:58:10 +0000 (16:58 +0000)]
Cleanup no longer used macros

16 years agoFix compiler warning: local variable may be used without having been initialized
Yang Tse [Wed, 3 Oct 2007 16:26:56 +0000 (16:26 +0000)]
Fix compiler warning: local variable may be used without having been initialized

16 years agoUpgrade OS400 wrappers and RPG copy file according to latest code updates
Patrick Monnerat [Wed, 3 Oct 2007 15:09:21 +0000 (15:09 +0000)]
Upgrade OS400 wrappers and RPG copy file according to latest code updates

16 years agoFix issue related with the use of ares_timeout() result.
Yang Tse [Wed, 3 Oct 2007 13:19:34 +0000 (13:19 +0000)]
Fix issue related with the use of ares_timeout() result.

16 years agoexported symbols must use lowercase "curl_", and I also fixed two compiler
Daniel Stenberg [Wed, 3 Oct 2007 08:58:40 +0000 (08:58 +0000)]
exported symbols must use lowercase "curl_", and I also fixed two compiler
warnings, one C99 thing and the bad pointer sent to the callback

16 years agorecount contributors after the 7.17.0 release
Daniel Stenberg [Wed, 3 Oct 2007 08:54:35 +0000 (08:54 +0000)]
recount contributors after the 7.17.0 release

16 years agopeople from the 7.17.0 announcement
Daniel Stenberg [Wed, 3 Oct 2007 08:51:36 +0000 (08:51 +0000)]
people from the 7.17.0 announcement

16 years ago"97 - check ip callback", check
Daniel Stenberg [Wed, 3 Oct 2007 08:46:45 +0000 (08:46 +0000)]
"97 - check ip callback", check

16 years agoAlexey Pesternikov introduced CURLOPT_OPENSOCKETFUNCTION and
Daniel Stenberg [Wed, 3 Oct 2007 08:45:00 +0000 (08:45 +0000)]
Alexey Pesternikov introduced CURLOPT_OPENSOCKETFUNCTION and
CURLOPT_OPENSOCKETDATA to set a callback that allows an application to replace
the socket() call used by libcurl. It basically allows the app to change
address, protocol or whatever of the socket. (I also did some whitespace
indent/cleanups in lib/url.c which kind of hides some of these changes, sorry
for mixing those in.)

16 years agoI renamed the CURLE_SSL_PEER_CERTIFICATE error code to
Daniel Stenberg [Wed, 3 Oct 2007 08:07:50 +0000 (08:07 +0000)]
I renamed the CURLE_SSL_PEER_CERTIFICATE error code to
CURLE_PEER_FAILED_VERIFICATION (standard CURL_NO_OLDIES style), and made this
return code get used by the previous SSH MD5 fingerprint check in case it
fails.

16 years agoBased on a patch brought by Johnny Luong, libcurl now offers
Daniel Stenberg [Wed, 3 Oct 2007 08:00:42 +0000 (08:00 +0000)]
Based on a patch brought by Johnny Luong, libcurl now offers
CURLOPT_SSH_HOST_PUBLIC_KEY_MD5 and the curl tool --hostpubmd5. They both make
the SCP or SFTP connection verify the remote host's md5 checksum of the public
key before doing a connect, to reduce the risk of a man-in-the-middle attack.

16 years ago"99 - curl_easy_close()" seems to have gone uninteresting
Daniel Stenberg [Tue, 2 Oct 2007 22:00:03 +0000 (22:00 +0000)]
"99 - curl_easy_close()" seems to have gone uninteresting

16 years ago102, check
Daniel Stenberg [Tue, 2 Oct 2007 21:59:06 +0000 (21:59 +0000)]
102, check

16 years agoFix memory leak under low memory conditions.
Yang Tse [Tue, 2 Oct 2007 19:19:47 +0000 (19:19 +0000)]
Fix memory leak under low memory conditions.

16 years agoAvoid a segfault when generating a DNS "Transaction ID" in internal
Yang Tse [Tue, 2 Oct 2007 18:26:48 +0000 (18:26 +0000)]
Avoid a segfault when generating a DNS "Transaction ID" in internal
function init_id_key() under low memory conditions.

16 years agoRenamed a couple of global variables to avoid shadowing warnings
Yang Tse [Tue, 2 Oct 2007 16:05:28 +0000 (16:05 +0000)]
Renamed a couple of global variables to avoid shadowing warnings

16 years agoFix compiler warning
Yang Tse [Tue, 2 Oct 2007 15:26:30 +0000 (15:26 +0000)]
Fix compiler warning

16 years ago<winber.h> needed for Windows LDAP client 32 API support
Yang Tse [Tue, 2 Oct 2007 14:48:38 +0000 (14:48 +0000)]
<winber.h> needed for Windows LDAP client 32 API support

16 years agoLinking with wldap32.lib needed for Windows LDAP client 32 API support
Yang Tse [Tue, 2 Oct 2007 14:26:04 +0000 (14:26 +0000)]
Linking with wldap32.lib needed for Windows LDAP client 32 API support

16 years agoAdd ares_llist.c and ares_llist.h to MSCV project file.
Yang Tse [Tue, 2 Oct 2007 11:13:58 +0000 (11:13 +0000)]
Add ares_llist.c and ares_llist.h to MSCV project file.

16 years agoknown bug #46: chunked-encoded CONNECT responses from a http proxy now works.
Daniel Stenberg [Tue, 2 Oct 2007 10:21:36 +0000 (10:21 +0000)]
known bug #46: chunked-encoded CONNECT responses from a http proxy now works.
Added test case 1008 to verify. Note that #47 is still there.

16 years agoDisable the Nagle algorithm and send back responses in small chunks in an
Daniel Stenberg [Tue, 2 Oct 2007 10:13:37 +0000 (10:13 +0000)]
Disable the Nagle algorithm and send back responses in small chunks in an
attempt to force smaller bits to get read by clients.

16 years agodocument --post301, based on the phrasing in curl_easy_setopt.3 for
Daniel Stenberg [Tue, 2 Oct 2007 09:57:48 +0000 (09:57 +0000)]
document --post301, based on the phrasing in curl_easy_setopt.3 for
CURLOPT_POST301 written by Philip Langdale

16 years agoCURLOPT_POST301 section, added by Philip Langdale
Daniel Stenberg [Tue, 2 Oct 2007 09:56:52 +0000 (09:56 +0000)]
CURLOPT_POST301 section, added by Philip Langdale

16 years agoFixed the problem where next_lookup would use 'status' uninitialized. Now
Daniel Stenberg [Tue, 2 Oct 2007 08:12:30 +0000 (08:12 +0000)]
Fixed the problem where next_lookup would use 'status' uninitialized. Now
it gets passed the initial value as an argument.

16 years agoAvoid inline C99ism, and move c-ares routines for managing doubly-linked lists.
Yang Tse [Tue, 2 Oct 2007 02:18:01 +0000 (02:18 +0000)]
Avoid inline C99ism, and move c-ares routines for managing doubly-linked lists.

16 years agoares_strerror() segfaulted if the input error number was out of the currently
Daniel Stenberg [Mon, 1 Oct 2007 22:52:31 +0000 (22:52 +0000)]
ares_strerror() segfaulted if the input error number was out of the currently
supported range.

16 years agoPrevent ares_strerror() from segfaulting if an invalid error code is passed
Daniel Stenberg [Mon, 1 Oct 2007 22:51:38 +0000 (22:51 +0000)]
Prevent ares_strerror() from segfaulting if an invalid error code is passed
in as argument!

16 years agoAdded test536 that was accidentally missing. I also wrote up a new makefile
Daniel Stenberg [Sun, 30 Sep 2007 22:58:24 +0000 (22:58 +0000)]
Added test536 that was accidentally missing. I also wrote up a new makefile
target called 'filecheck' so that if you run 'make filecheck' in this directory
it'll check if the local files are also mentioned in the Makefile.am so that
they are properly included in release archives!

16 years agoAlex Fishman reported a curl_easy_escape() problem that was made the
Daniel Stenberg [Sun, 30 Sep 2007 22:40:24 +0000 (22:40 +0000)]
Alex Fishman reported a curl_easy_escape() problem that was made the
function do wrong on all input bytes that are >= 0x80 (decimal 128) due to a
signed / unsigned mistake in the code. I fixed it and added test case 543 to
verify.

16 years agoFix compiler warning
Yang Tse [Sun, 30 Sep 2007 19:43:23 +0000 (19:43 +0000)]
Fix compiler warning

16 years agocheck availability of <netinet/tcp.h>
Yang Tse [Sun, 30 Sep 2007 02:12:11 +0000 (02:12 +0000)]
check availability of <netinet/tcp.h>

16 years agoFix missing right parenthesis
Yang Tse [Sun, 30 Sep 2007 01:27:39 +0000 (01:27 +0000)]
Fix missing right parenthesis

16 years agoFix comparison between signed and unsigned
Yang Tse [Sun, 30 Sep 2007 01:01:43 +0000 (01:01 +0000)]
Fix comparison between signed and unsigned

16 years agoimprove portability, defining MAXDNAME and MAXCDNAME
Yang Tse [Sun, 30 Sep 2007 00:37:47 +0000 (00:37 +0000)]
improve portability, defining MAXDNAME and MAXCDNAME

16 years agoFix a memory leak that I recently inadvertedly introduced.
Steinar H. Gunderson [Sun, 30 Sep 2007 00:08:01 +0000 (00:08 +0000)]
Fix a memory leak that I recently inadvertedly introduced.

16 years agoUse ISDIGIT instead of isdigit; fixes a gcc warning.
Steinar H. Gunderson [Sat, 29 Sep 2007 21:57:05 +0000 (21:57 +0000)]
Use ISDIGIT instead of isdigit; fixes a gcc warning.

16 years agoImmanuel Gregoire fixed a problem with persistent transfers over SFTP - the
Daniel Stenberg [Sat, 29 Sep 2007 21:34:34 +0000 (21:34 +0000)]
Immanuel Gregoire fixed a problem with persistent transfers over SFTP - the
previous proto struct was kept.

16 years agoPort the TCP socket fix made in ares_fds() to ares_getsock() as well.
Steinar H. Gunderson [Sat, 29 Sep 2007 19:26:59 +0000 (19:26 +0000)]
Port the TCP socket fix made in ares_fds() to ares_getsock() as well.

16 years agoPreviously, processing a large batch of timeouts was O(n^2) in the number of
Steinar H. Gunderson [Sat, 29 Sep 2007 18:18:47 +0000 (18:18 +0000)]
Previously, processing a large batch of timeouts was O(n^2) in the number of
outstanding queries, and processing a DNS response packet was O(n) in the
number of outstanding queries. To speed things up in Google, we added a few circular,
doubly-linked lists of queries that are hash-bucketed based on
the attributes we care about, so most important operations are now O(1).

It might be that the number of buckets are higher than most people would need,
but on a quick calculation it should only be 100kB or so even on a 64-bit
system, so I've let it stay as-is.

16 years agoWe should standarise on C comments.
Gisle Vanem [Sat, 29 Sep 2007 14:37:47 +0000 (14:37 +0000)]
We should standarise on C comments.

16 years agoFix compiler warning in setsockopt().
Gisle Vanem [Sat, 29 Sep 2007 14:34:59 +0000 (14:34 +0000)]
Fix compiler warning in setsockopt().

16 years agoTCP queries can time out too, not just UDP queries. (Patch from the Google tree.)
Steinar H. Gunderson [Sat, 29 Sep 2007 14:25:14 +0000 (14:25 +0000)]
TCP queries can time out too, not just UDP queries. (Patch from the Google tree.)

16 years agoRead and process as many packets as possible in read_udp_packets, to avoid having...
Steinar H. Gunderson [Sat, 29 Sep 2007 14:21:47 +0000 (14:21 +0000)]
Read and process as many packets as possible in read_udp_packets, to avoid having to run the entire event loop once per packet. (Patch from the Google tree.)

16 years agoThere are two different places in write_tcp_data() that advance the send_queue; howev...
Steinar H. Gunderson [Sat, 29 Sep 2007 14:09:14 +0000 (14:09 +0000)]
There are two different places in write_tcp_data() that advance the send_queue; however, they are slightly different and only the first one properly uses a while loop. Consolidate both into a single function that DTTR. (Patch from the Google tree.)

16 years agoReject names that are longer than 255 characters, to avoid problems with strict or...
Steinar H. Gunderson [Sat, 29 Sep 2007 13:58:23 +0000 (13:58 +0000)]
Reject names that are longer than 255 characters, to avoid problems with strict or buggy DNS server implementations. (Patch from the Google tree)

16 years agoIn ares_mkquery, make sure we set buflen and buf to reasonable values if there's...
Steinar H. Gunderson [Sat, 29 Sep 2007 13:56:36 +0000 (13:56 +0000)]
In ares_mkquery, make sure we set buflen and buf to reasonable values if there's an error. (Patch from the Google tree)

16 years agoBe stricter about what's a valid IP address in fake_hostent. (Patch from the Google...
Steinar H. Gunderson [Sat, 29 Sep 2007 13:52:14 +0000 (13:52 +0000)]
Be stricter about what's a valid IP address in fake_hostent. (Patch from the Google tree.)

16 years agoHandle the root of the DNS tree correctly in ares_expand_name.
Steinar H. Gunderson [Sat, 29 Sep 2007 13:38:17 +0000 (13:38 +0000)]
Handle the root of the DNS tree correctly in ares_expand_name.

16 years agoAdapted the c-ares code to the API change c-ares 1.5.0 brings in the
Daniel Stenberg [Fri, 28 Sep 2007 21:48:28 +0000 (21:48 +0000)]
Adapted the c-ares code to the API change c-ares 1.5.0 brings in the
notifier callback(s).

16 years agorename variable to prevent shadow warning
Daniel Stenberg [Fri, 28 Sep 2007 21:45:05 +0000 (21:45 +0000)]
rename variable to prevent shadow warning

16 years agotoday's modifications by Steinar and me
Daniel Stenberg [Fri, 28 Sep 2007 20:28:20 +0000 (20:28 +0000)]
today's modifications by Steinar and me

16 years agoBumped version to 1.5.0 for next release and soname bumped to 2 due to ABI
Daniel Stenberg [Fri, 28 Sep 2007 20:28:06 +0000 (20:28 +0000)]
Bumped version to 1.5.0 for next release and soname bumped to 2 due to ABI
and API changes in the progress callback (and possibly more coming up from
Steinar)

16 years agoRenamed a variable to avoid shadowing a global declarations.
Dan Fandrich [Fri, 28 Sep 2007 18:47:59 +0000 (18:47 +0000)]
Renamed a variable to avoid shadowing a global declarations.

16 years agoUnrevert previous 'missing' hunks. They were missing since the patch is still in...
Steinar H. Gunderson [Fri, 28 Sep 2007 15:56:28 +0000 (15:56 +0000)]
Unrevert previous 'missing' hunks. They were missing since the patch is still in for review :-)

16 years agoYet more missing hunks... Nggh.
Steinar H. Gunderson [Fri, 28 Sep 2007 15:55:11 +0000 (15:55 +0000)]
Yet more missing hunks... Nggh.

16 years agoAlways register for TCP events even if there are no outstanding queries, as the other...
Steinar H. Gunderson [Fri, 28 Sep 2007 15:53:10 +0000 (15:53 +0000)]
Always register for TCP events even if there are no outstanding queries, as the other side could always close the connection, which is a valid event which should be responded to.

16 years agoForgot to include a few hunks from ares_process.c earlier. Fixing now.
Steinar H. Gunderson [Fri, 28 Sep 2007 15:51:00 +0000 (15:51 +0000)]
Forgot to include a few hunks from ares_process.c earlier. Fixing now.

16 years agoSupport a few more socket options, and refactor the option setting a bit. (Patch...
Steinar H. Gunderson [Fri, 28 Sep 2007 15:15:39 +0000 (15:15 +0000)]
Support a few more socket options, and refactor the option setting a bit. (Patch from the Google tree.)

16 years agoMake the query callbacks return the number of timeouts that happened during the execu...
Steinar H. Gunderson [Fri, 28 Sep 2007 14:46:51 +0000 (14:46 +0000)]
Make the query callbacks return the number of timeouts that happened during the execution of a query, and update documentation accordingly. (Patch from the Google tree.)

16 years agoThree fixes in one commit (sorry): a) Take care of the tcpbuf if it ends while queued...
Steinar H. Gunderson [Fri, 28 Sep 2007 14:28:14 +0000 (14:28 +0000)]
Three fixes in one commit (sorry): a) Take care of the tcpbuf if it ends while queued for transmission, note broken servers and close them in the main loop, and store TCP socket generation number in order not to send the same query twice over the same socket.

16 years agoDon't skip a server if it's the only one. (Bugfix from the Google tree.)
Steinar H. Gunderson [Fri, 28 Sep 2007 14:26:11 +0000 (14:26 +0000)]
Don't skip a server if it's the only one. (Bugfix from the Google tree.)

16 years agoDon't strdup an empty string
Dan Fandrich [Thu, 27 Sep 2007 18:39:10 +0000 (18:39 +0000)]
Don't strdup an empty string

16 years agoRenamed a few variables to avoid shadowing global declarations.
Dan Fandrich [Thu, 27 Sep 2007 18:12:03 +0000 (18:12 +0000)]
Renamed a few variables to avoid shadowing global declarations.

16 years agoRemoved cut-and-paste cruft leading to fclose() of an unopened file
Dan Fandrich [Thu, 27 Sep 2007 17:22:10 +0000 (17:22 +0000)]
Removed cut-and-paste cruft leading to fclose() of an unopened file

16 years agoa name resolve that times out is still a failed name resolve
Daniel Stenberg [Thu, 27 Sep 2007 12:05:39 +0000 (12:05 +0000)]
a name resolve that times out is still a failed name resolve

16 years agowrong, revert the previous "fix" and instead check that the fd_set pointer
Daniel Stenberg [Thu, 27 Sep 2007 12:04:56 +0000 (12:04 +0000)]
wrong, revert the previous "fix" and instead check that the fd_set pointer
is non-NULL before we FD_CLR

16 years agoeek, fix the conditions to return on either problem instead of requiring
Daniel Stenberg [Thu, 27 Sep 2007 12:02:15 +0000 (12:02 +0000)]
eek, fix the conditions to return on either problem instead of requiring
both to occur

16 years agoRenamed a few variables to avoid shadowing global declarations.
Dan Fandrich [Thu, 27 Sep 2007 02:45:58 +0000 (02:45 +0000)]
Renamed a few variables to avoid shadowing global declarations.

16 years agoEnabled a few more gcc warnings with --enable-debug. Renamed a few
Dan Fandrich [Thu, 27 Sep 2007 01:45:22 +0000 (01:45 +0000)]
Enabled a few more gcc warnings with --enable-debug.  Renamed a few
variables to avoid shadowing global declarations.

16 years agoFix compiler warning: the address of 'env' will always evaluate as 'true'
Yang Tse [Thu, 27 Sep 2007 00:58:41 +0000 (00:58 +0000)]
Fix compiler warning: the address of 'env' will always evaluate as 'true'

16 years agowe added a curl_easy_setopt() option too
Daniel Stenberg [Wed, 26 Sep 2007 12:46:03 +0000 (12:46 +0000)]
we added a curl_easy_setopt() option too

16 years agoPhilip Langdale provided the new CURLOPT_POST301 option for
Daniel Stenberg [Wed, 26 Sep 2007 12:44:59 +0000 (12:44 +0000)]
Philip Langdale provided the new CURLOPT_POST301 option for
curl_easy_setopt() that alters how libcurl functions when following
redirects. It makes libcurl obey the RFC2616 when a 301 response is received
after a non-GET request is made. Default libcurl behaviour is to change
method to GET in the subsequent request (like it does for response code 302
- because that's what many/most browsers do), but with this CURLOPT_POST301
option enabled it will do what the spec says and do the next request using
the same method again. I.e keep POST after 301.

The curl tool got this option as --post301

Test case 1011 and 1012 were added to verify.

16 years agoMax Katsev reported that when doing a libcurl FTP request with
Daniel Stenberg [Wed, 26 Sep 2007 12:00:01 +0000 (12:00 +0000)]
Max Katsev reported that when doing a libcurl FTP request with
CURLOPT_NOBODY enabled but not CURLOPT_HEADER, libcurl wouldn't do TYPE
before it does SIZE which makes it less useful. I walked over the code and
made it do this properly, and added test case 542 to verify it.

16 years agoMake glibc define the prototype for strcasestr
Dan Fandrich [Tue, 25 Sep 2007 17:33:56 +0000 (17:33 +0000)]
Make glibc define the prototype for strcasestr

16 years agoa new Lua binding and I shortened the wording on several bindings by cutting
Daniel Stenberg [Tue, 25 Sep 2007 08:46:49 +0000 (08:46 +0000)]
a new Lua binding and I shortened the wording on several bindings by cutting
out "written"

16 years ago#ifdef out a few more functions when SSL is disabled.
Dan Fandrich [Tue, 25 Sep 2007 06:45:05 +0000 (06:45 +0000)]
#ifdef out a few more functions when SSL is disabled.