platform/upstream/curl.git
19 years agoupdate
Daniel Stenberg [Wed, 18 May 2005 20:00:50 +0000 (20:00 +0000)]
update

19 years agoadjusted the strerror_r test more, use _REENTRANT instead of _THREAD_SAFE
Daniel Stenberg [Wed, 18 May 2005 15:15:00 +0000 (15:15 +0000)]
adjusted the strerror_r test more, use _REENTRANT instead of _THREAD_SAFE
when looking for the prototype

19 years agoThe configure check for strerror_r() failed to detect the proper API at
Daniel Stenberg [Wed, 18 May 2005 13:24:23 +0000 (13:24 +0000)]
The configure check for strerror_r() failed to detect the proper API at
times, like on my HP-UX 10.20 tests. And then lib/strerror.c badly assumed
the glibc version if the posix define wasn't set (since it _had_ found a
strerror_r).

19 years ago#15 is now fixed
Daniel Stenberg [Wed, 18 May 2005 10:38:21 +0000 (10:38 +0000)]
#15 is now fixed

19 years agoclarified for GnuTLS
Daniel Stenberg [Wed, 18 May 2005 10:14:25 +0000 (10:14 +0000)]
clarified for GnuTLS

19 years agoseveral updates
Daniel Stenberg [Wed, 18 May 2005 10:12:08 +0000 (10:12 +0000)]
several updates

19 years agoignore resolve too
Daniel Stenberg [Wed, 18 May 2005 10:05:38 +0000 (10:05 +0000)]
ignore resolve too

19 years agouse less code and prevent compiler warning
Daniel Stenberg [Wed, 18 May 2005 10:01:46 +0000 (10:01 +0000)]
use less code and prevent compiler warning

19 years agoremoved the separate table with download links, and extended the curl site
Daniel Stenberg [Wed, 18 May 2005 09:26:57 +0000 (09:26 +0000)]
removed the separate table with download links, and extended the curl site
list with all current mirrors

19 years agoscan for gmake and make to prefer gmake on systems that have it
Daniel Stenberg [Tue, 17 May 2005 12:07:08 +0000 (12:07 +0000)]
scan for gmake and make to prefer gmake on systems that have it

19 years agoMade test case 241 precheck that the given name resolves to an ipv6 address,
Daniel Stenberg [Tue, 17 May 2005 10:27:11 +0000 (10:27 +0000)]
Made test case 241 precheck that the given name resolves to an ipv6 address,
or the test is skipped. Ideally, we should let this test case go over a few
frequently used IPv6 localhost aliases...

19 years agoMoved more generic functions to util.[ch]
Daniel Stenberg [Tue, 17 May 2005 10:22:22 +0000 (10:22 +0000)]
Moved more generic functions to util.[ch]
Added resolve.c to simply resolve a given host name

19 years agocheck if getsockname() returns failure before using the address it provides
Daniel Stenberg [Tue, 17 May 2005 09:18:17 +0000 (09:18 +0000)]
check if getsockname() returns failure before using the address it provides

19 years agoreduced typecasts, from two to one
Daniel Stenberg [Tue, 17 May 2005 09:15:21 +0000 (09:15 +0000)]
reduced typecasts, from two to one

19 years agoMore of the same
Dominick Meglio [Tue, 17 May 2005 04:20:58 +0000 (04:20 +0000)]
More of the same

19 years agoMore compiler warning cleanups
Dominick Meglio [Tue, 17 May 2005 04:18:41 +0000 (04:18 +0000)]
More compiler warning cleanups

19 years agobad formatting
Daniel Stenberg [Mon, 16 May 2005 22:30:00 +0000 (22:30 +0000)]
bad formatting

19 years agoMade ares_free_hostent man page refer to ares_parse_aaaa_reply
Dominick Meglio [Mon, 16 May 2005 19:23:57 +0000 (19:23 +0000)]
Made ares_free_hostent man page refer to ares_parse_aaaa_reply

19 years agoCleaned up some compile warnings
Dominick Meglio [Mon, 16 May 2005 19:14:25 +0000 (19:14 +0000)]
Cleaned up some compile warnings

19 years agoAdded ares_getnameinfo which mimics the getnameinfo API
Dominick Meglio [Mon, 16 May 2005 18:06:54 +0000 (18:06 +0000)]
Added ares_getnameinfo which mimics the getnameinfo API

19 years agoModified the gmtime_r check to not check for it until the "check for a working
Daniel Stenberg [Mon, 16 May 2005 15:09:26 +0000 (15:09 +0000)]
Modified the gmtime_r check to not check for it until the "check for a working
one" is made, and only if that test runs ok we define it as present. Unless
crosscompiling, since then we use the former AC_CHECK_FUNCS method.

19 years agodefine GMTIME_R to 0 if not working
Daniel Stenberg [Mon, 16 May 2005 14:53:48 +0000 (14:53 +0000)]
define GMTIME_R to 0 if not working

19 years agoattempt to detect a bad (as in HPUX 10.20 bad) gmtime_r function
Daniel Stenberg [Mon, 16 May 2005 14:40:10 +0000 (14:40 +0000)]
attempt to detect a bad (as in HPUX 10.20 bad) gmtime_r function

19 years agostart working on 7.14.1
Daniel Stenberg [Mon, 16 May 2005 13:27:41 +0000 (13:27 +0000)]
start working on 7.14.1

19 years agoVersion 7.14.0
Daniel Stenberg [Mon, 16 May 2005 12:58:41 +0000 (12:58 +0000)]
Version 7.14.0

19 years agoreturn, not exit, on several places
Daniel Stenberg [Mon, 16 May 2005 07:07:00 +0000 (07:07 +0000)]
return, not exit, on several places

19 years agoConverted some macros to use NS_* so they work on non-IPv6 systems
Dominick Meglio [Sun, 15 May 2005 16:31:01 +0000 (16:31 +0000)]
Converted some macros to use NS_* so they work on non-IPv6 systems

19 years agoForgot to ares_-ize inet_ntop
Dominick Meglio [Sun, 15 May 2005 04:38:53 +0000 (04:38 +0000)]
Forgot to ares_-ize inet_ntop

19 years agofix warning about redefined symbol
Daniel Stenberg [Sat, 14 May 2005 21:15:36 +0000 (21:15 +0000)]
fix warning about redefined symbol

19 years agoreplaced the CRLF newlines with plain LF ones
Daniel Stenberg [Sat, 14 May 2005 20:45:49 +0000 (20:45 +0000)]
replaced the CRLF newlines with plain LF ones

19 years agoAdded an inet_ntop function from BIND for systems that do not have it
Dominick Meglio [Sat, 14 May 2005 18:35:20 +0000 (18:35 +0000)]
Added an inet_ntop function from BIND for systems that do not have it

19 years agoUpdated generated dependencies.
Gisle Vanem [Sat, 14 May 2005 06:04:21 +0000 (06:04 +0000)]
Updated generated dependencies.

19 years agoSome patches for (a stricter/smarter) gcc 4.0 and
Gisle Vanem [Sat, 14 May 2005 06:00:40 +0000 (06:00 +0000)]
Some patches for (a stricter/smarter) gcc 4.0 and
warnings like:
  'x' may be used uninitialized in this function.

19 years ago'ssize_t' seems to be a gcc 4.x built-in.
Gisle Vanem [Sat, 14 May 2005 05:59:26 +0000 (05:59 +0000)]
'ssize_t' seems to be a gcc 4.x built-in.

19 years agoChange for systems with >1 ways of setting (non-)blocking
Gisle Vanem [Sat, 14 May 2005 05:58:52 +0000 (05:58 +0000)]
Change for systems with >1 ways of setting (non-)blocking
mode. (djgpp/Watt-32 has 3 ways). Should rewrite this using
"#elif ..", but maybe there is still broken cpp around?

19 years agoupdated
Daniel Stenberg [Fri, 13 May 2005 23:00:06 +0000 (23:00 +0000)]
updated

19 years agouses select() instead of poll() even on Mac OS X 10.4
Daniel Stenberg [Fri, 13 May 2005 22:24:50 +0000 (22:24 +0000)]
uses select() instead of poll() even on Mac OS X 10.4

19 years agoadjusted the configure to always skip the fine-poll() test on Mac OS X (darwin)
Daniel Stenberg [Fri, 13 May 2005 21:19:21 +0000 (21:19 +0000)]
adjusted the configure to always skip the fine-poll() test on Mac OS X (darwin)

19 years agoremove blank lines
Daniel Stenberg [Thu, 12 May 2005 21:56:26 +0000 (21:56 +0000)]
remove blank lines

19 years agoCURLOPT_SSLVERSION clarified
Daniel Stenberg [Thu, 12 May 2005 21:49:42 +0000 (21:49 +0000)]
CURLOPT_SSLVERSION clarified

19 years ago-z bad use warning and NTLM proxy auth in reconnect fix
Daniel Stenberg [Thu, 12 May 2005 14:00:27 +0000 (14:00 +0000)]
-z bad use warning and NTLM proxy auth in reconnect fix

19 years agooops, found by bug reported in bug report #1200661
Daniel Stenberg [Thu, 12 May 2005 13:44:25 +0000 (13:44 +0000)]
oops, found by bug reported in bug report #1200661

19 years agospell
Daniel Stenberg [Thu, 12 May 2005 12:53:02 +0000 (12:53 +0000)]
spell

19 years agotypecast to fix warning on 64bit systems
Daniel Stenberg [Thu, 12 May 2005 08:51:30 +0000 (08:51 +0000)]
typecast to fix warning on 64bit systems

19 years agowarn about bad -z syntax
Daniel Stenberg [Thu, 12 May 2005 07:28:03 +0000 (07:28 +0000)]
warn about bad -z syntax

19 years agoMatrixSSL and yaSSL are two free libs we _could_ support
Daniel Stenberg [Wed, 11 May 2005 11:56:04 +0000 (11:56 +0000)]
MatrixSSL and yaSSL are two free libs we _could_ support

19 years agomention the select() error fix as well
Daniel Stenberg [Wed, 11 May 2005 10:23:59 +0000 (10:23 +0000)]
mention the select() error fix as well

19 years agothe new HTTP headers
Daniel Stenberg [Wed, 11 May 2005 09:56:53 +0000 (09:56 +0000)]
the new HTTP headers

19 years agoremoved duplicate
Daniel Stenberg [Wed, 11 May 2005 09:56:10 +0000 (09:56 +0000)]
removed duplicate

19 years agoModified the default HTTP headers used by libcurl:
Daniel Stenberg [Wed, 11 May 2005 09:52:59 +0000 (09:52 +0000)]
Modified the default HTTP headers used by libcurl:

A) Normal non-proxy HTTP:

 - no more "Pragma: no-cache" (this only makes sense to proxies)

B) Non-CONNECT HTTP request over proxy:

 - "Pragma: no-cache" is used (like before)
 - "Proxy-Connection: Keep-alive" (for older style 1.0-proxies)

C) CONNECT HTTP request over proxy:

 - "Host: [name]:[port]"
 - "Proxy-Connection: Keep-alive"

19 years agoprevent NS_IN6ADDRSZ from getting set to zero if the struct doesn't exist
Daniel Stenberg [Wed, 11 May 2005 06:47:09 +0000 (06:47 +0000)]
prevent NS_IN6ADDRSZ from getting set to zero if the struct doesn't exist

19 years agoHm, this doesn't feel right. The error bits returned from Curl_select() can
Daniel Stenberg [Tue, 10 May 2005 23:02:37 +0000 (23:02 +0000)]
Hm, this doesn't feel right. The error bits returned from Curl_select() can
be returned at times when we want to ignore them. Test case 160 fails on Linux,
so I modify the comparison to check for _only_ the error bit set...

19 years agome stupid, errno is not set for mere select()-exceptions
Daniel Stenberg [Tue, 10 May 2005 22:48:26 +0000 (22:48 +0000)]
me stupid, errno is not set for mere select()-exceptions

19 years agoinclude protos to fix warnings
Daniel Stenberg [Tue, 10 May 2005 22:46:24 +0000 (22:46 +0000)]
include protos to fix warnings

19 years agoIf Curl_select() returns with the error bit set, bail out.
Daniel Stenberg [Tue, 10 May 2005 22:44:08 +0000 (22:44 +0000)]
If Curl_select() returns with the error bit set, bail out.

19 years agoprevent 64bit warnings
Daniel Stenberg [Tue, 10 May 2005 11:21:57 +0000 (11:21 +0000)]
prevent 64bit warnings

19 years agoallow the ares/config.h display to fail
Daniel Stenberg [Tue, 10 May 2005 11:19:12 +0000 (11:19 +0000)]
allow the ares/config.h display to fail

19 years agoJeremy Brown's OpenSSL thread-locking example
Daniel Stenberg [Mon, 9 May 2005 21:12:03 +0000 (21:12 +0000)]
Jeremy Brown's OpenSSL thread-locking example

19 years agonew counter
Daniel Stenberg [Mon, 9 May 2005 13:57:58 +0000 (13:57 +0000)]
new counter

19 years agoJamie Lokier added. And I now recounted the amount better: 437 named as of now.
Daniel Stenberg [Mon, 9 May 2005 13:53:42 +0000 (13:53 +0000)]
Jamie Lokier added. And I now recounted the amount better: 437 named as of now.

19 years agoupdate the "PORTS" section a little
Daniel Stenberg [Mon, 9 May 2005 13:26:41 +0000 (13:26 +0000)]
update the "PORTS" section a little

19 years agoadd multi-thread details for GnuTLS
Daniel Stenberg [Mon, 9 May 2005 13:13:38 +0000 (13:13 +0000)]
add multi-thread details for GnuTLS

19 years agonew mirror, added amount of contributors
Daniel Stenberg [Mon, 9 May 2005 12:34:55 +0000 (12:34 +0000)]
new mirror, added amount of contributors

19 years agoJeff is short for Jeffrey
Daniel Stenberg [Mon, 9 May 2005 11:43:23 +0000 (11:43 +0000)]
Jeff is short for Jeffrey

19 years agoupdated with the current RELEASE-NOTES names
Daniel Stenberg [Mon, 9 May 2005 11:39:20 +0000 (11:39 +0000)]
updated with the current RELEASE-NOTES names

19 years agoI decided to make this list more complete. I took the 5-year anniversary list
Daniel Stenberg [Mon, 9 May 2005 09:11:39 +0000 (09:11 +0000)]
I decided to make this list more complete. I took the 5-year anniversary list
from 2003 and added all names from all release notes in the CVS (there is a
slight gap though). I removed names with only first names (Like "Chris" and
"Ralph") , as that won't make anyone happy and we might list their full names
as well anyway.

This list is now intended to include _all_ people that contribute: big or
small. 389 names at the time of this commit.

19 years agono need to display src/config.h anymore since it is a duplicate of lib/config.h
Daniel Stenberg [Mon, 9 May 2005 07:45:30 +0000 (07:45 +0000)]
no need to display src/config.h anymore since it is a duplicate of lib/config.h
but we could use having a look at ares/config.h when that is used

19 years agosilence compiler warnings
Daniel Stenberg [Sun, 8 May 2005 22:45:01 +0000 (22:45 +0000)]
silence compiler warnings

19 years agofix warnings about unused variables for non-debug builds
Daniel Stenberg [Sat, 7 May 2005 20:41:16 +0000 (20:41 +0000)]
fix warnings about unused variables for non-debug builds

19 years agofix
Daniel Stenberg [Sat, 7 May 2005 20:28:39 +0000 (20:28 +0000)]
fix

19 years agoJanuary 2003. Started working on the distributed curl tests. The autobuilds.
Daniel Stenberg [Sat, 7 May 2005 14:23:32 +0000 (14:23 +0000)]
January 2003. Started working on the distributed curl tests. The autobuilds.

19 years agoDEBUGF() is a new conveniant macro to add infof() calls (or similar) for
Daniel Stenberg [Sat, 7 May 2005 13:57:07 +0000 (13:57 +0000)]
DEBUGF() is a new conveniant macro to add infof() calls (or similar) for
debug builds only. Made the ftp code use it on several places.

19 years agoAdded an active disconnected state, to make the code clearer.
Daniel Stenberg [Sat, 7 May 2005 13:52:07 +0000 (13:52 +0000)]
Added an active disconnected state, to make the code clearer.

19 years agoremoved unnecessary logging to ease REAL debuggin
Daniel Stenberg [Sat, 7 May 2005 08:55:37 +0000 (08:55 +0000)]
removed unnecessary logging to ease REAL debuggin

19 years agoone more command line option, fixed the AIX 4.3 enabled IPv6 build (it now
Daniel Stenberg [Fri, 6 May 2005 23:46:35 +0000 (23:46 +0000)]
one more command line option, fixed the AIX 4.3 enabled IPv6 build (it now
detects a bad Ipv6 situation and disables it automatically)

19 years agoAdded two test cases for multipart formpost over a proxy with --anyauth. Our
Daniel Stenberg [Fri, 6 May 2005 23:22:33 +0000 (23:22 +0000)]
Added two test cases for multipart formpost over a proxy with --anyauth. Our
HTTP test server is a bit limited though, as it never responds to the POST
request until all data has been sent (and received)...

19 years agoWhen a server is clearly running, curl is now invoked to verify that it can
Daniel Stenberg [Fri, 6 May 2005 23:21:34 +0000 (23:21 +0000)]
When a server is clearly running, curl is now invoked to verify that it can
download a file from the server before the server is considered fine to use
for the given test case. This should fix the cases where the server can run
but curl cannot work with it.

19 years agouse calloc instead of malloc to save a call to memset()
Daniel Stenberg [Thu, 5 May 2005 06:04:00 +0000 (06:04 +0000)]
use calloc instead of malloc to save a call to memset()

19 years agonow add --trace-time by default for curl tests
Daniel Stenberg [Wed, 4 May 2005 21:58:47 +0000 (21:58 +0000)]
now add --trace-time by default for curl tests

19 years agoremoved lots of (now) redundant logging
Daniel Stenberg [Wed, 4 May 2005 21:57:07 +0000 (21:57 +0000)]
removed lots of (now) redundant logging

19 years agomodify a value we are allowed to
Daniel Stenberg [Wed, 4 May 2005 21:51:09 +0000 (21:51 +0000)]
modify a value we are allowed to

19 years agoimproved logging (all FTP protocol data, both ways) to possibly help us realize
Daniel Stenberg [Wed, 4 May 2005 21:49:30 +0000 (21:49 +0000)]
improved logging (all FTP protocol data, both ways) to possibly help us realize
why sometimes the control connection dies after a RETR has been sent

19 years agotowards 7.14.0 - really
Daniel Stenberg [Wed, 4 May 2005 15:11:09 +0000 (15:11 +0000)]
towards 7.14.0 - really

19 years agoprevent memory leak when built SSL disabled
Daniel Stenberg [Wed, 4 May 2005 14:52:51 +0000 (14:52 +0000)]
prevent memory leak when built SSL disabled

19 years ago*MAN* was this hard to track down. Had I just read the docs properly from the
Daniel Stenberg [Tue, 3 May 2005 23:14:43 +0000 (23:14 +0000)]
*MAN* was this hard to track down. Had I just read the docs properly from the
start... Anyway, fork() + exec() makes _two_ pids (in perl) that we need to
track and kill after use. Thankyouverymuch.

19 years agoadd more info to the log to ease debugging
Daniel Stenberg [Tue, 3 May 2005 23:13:24 +0000 (23:13 +0000)]
add more info to the log to ease debugging

19 years agoimproved failf() error messages
Daniel Stenberg [Mon, 2 May 2005 22:53:55 +0000 (22:53 +0000)]
improved failf() error messages

19 years agothe new functions and the upcoming ipv6 calls for the next version to become
Daniel Stenberg [Mon, 2 May 2005 22:33:57 +0000 (22:33 +0000)]
the new functions and the upcoming ipv6 calls for the next version to become
1.3.0

19 years agocorrected copyright years
Daniel Stenberg [Mon, 2 May 2005 14:33:58 +0000 (14:33 +0000)]
corrected copyright years

19 years agocorrected copyright year
Daniel Stenberg [Mon, 2 May 2005 14:33:07 +0000 (14:33 +0000)]
corrected copyright year

19 years agoSort of "fixed" KNOWN_BUGS #4: curl now builds IPv6 enabled on AIX 4.3. At
Daniel Stenberg [Mon, 2 May 2005 14:06:27 +0000 (14:06 +0000)]
Sort of "fixed" KNOWN_BUGS #4: curl now builds IPv6 enabled on AIX 4.3. At
least it should no longer cause a compiler error. However, it does not have
AI_NUMERICHOST so we cannot getaddrinfo() any numerical addresses with it (we
use that for FTP PORT/EPRT)! So, I modified the configure check that checks if
the getaddrinfo() is working, to use AI_NUMERICHOST since then it'll fail on
AIX 4.3 and it will automatically build with IPv6 support disabled.

19 years agoNow configure checks for struct sockaddr_storage and the ftp code tries
Daniel Stenberg [Mon, 2 May 2005 11:56:15 +0000 (11:56 +0000)]
Now configure checks for struct sockaddr_storage and the ftp code tries
to survive without it if not found. AIX 4.3 targetted adjustment.

19 years agoanother <case> converted to sysread
Daniel Stenberg [Mon, 2 May 2005 11:55:17 +0000 (11:55 +0000)]
another <case> converted to sysread

19 years agoread from the open2 filehandle with sysread, not <handle>
Daniel Stenberg [Mon, 2 May 2005 11:31:15 +0000 (11:31 +0000)]
read from the open2 filehandle with sysread, not <handle>

19 years agoFixed the FTP server read stuff when waiting for a connect after a PASV/EPSV.
Daniel Stenberg [Mon, 2 May 2005 10:22:09 +0000 (10:22 +0000)]
Fixed the FTP server read stuff when waiting for a connect after a PASV/EPSV.

Made the ftp server use the passed in pidfile name, and made runtests.pl
pass it in properly.

19 years agofix the server for the slow response case
Daniel Stenberg [Mon, 2 May 2005 10:03:12 +0000 (10:03 +0000)]
fix the server for the slow response case

19 years agoAdded --trace-time that when used adds a time stamp to each trace line that
Daniel Stenberg [Mon, 2 May 2005 09:38:19 +0000 (09:38 +0000)]
Added --trace-time that when used adds a time stamp to each trace line that
--trace, --trace-ascii and --verbose output. I also made the '>' display
separate each line on the linefeed so that HTTP requests etc look nicer in the
-v output.

19 years agoWhen starting the ftp server, wait a few seconds to make really sure that
Daniel Stenberg [Mon, 2 May 2005 09:08:44 +0000 (09:08 +0000)]
When starting the ftp server, wait a few seconds to make really sure that
a pidfile for the server appears as otherwise it failed.

19 years agoMake sure there's no pidfile if we cannot start the initial sockfilt tool -
Daniel Stenberg [Mon, 2 May 2005 09:08:02 +0000 (09:08 +0000)]
Make sure there's no pidfile if we cannot start the initial sockfilt tool -
this happens for some ipv6-enabled hosts on which sockfilt cannot listen
on ipv6.

19 years agotwo bugfixes, one change and one test script modification
Daniel Stenberg [Mon, 2 May 2005 08:40:20 +0000 (08:40 +0000)]
two bugfixes, one change and one test script modification