platform/upstream/curl.git
17 years agocheck symbol HAVE_UNISTD_H instead of UNISTD_H to include unistd.h
Yang Tse [Thu, 2 Nov 2006 00:34:21 +0000 (00:34 +0000)]
check symbol HAVE_UNISTD_H instead of UNISTD_H to include unistd.h

17 years agoprototype for gethostname is in unistd.h
Yang Tse [Thu, 2 Nov 2006 00:33:43 +0000 (00:33 +0000)]
prototype for gethostname is in unistd.h

17 years agotest 518 is all about testing libcurl functionality
Yang Tse [Wed, 1 Nov 2006 18:33:50 +0000 (18:33 +0000)]
test 518 is all about testing libcurl functionality
when more than FD_SETSIZE file descriptors are open.
This means that if for any reason we are not able to
open more than FD_SETSIZE file descriptors then test
518 should not be run.

test 537 is all about testing libcurl functionality
when the system has nearly exhausted the number of
free file descriptors. Test 537 will try to run with
very few free file descriptors.

17 years agoUpdated dependency output.
Gisle Vanem [Tue, 31 Oct 2006 20:45:17 +0000 (20:45 +0000)]
Updated dependency output.

17 years agoUpdated dependencies to not include config.h.
Gisle Vanem [Tue, 31 Oct 2006 20:44:36 +0000 (20:44 +0000)]
Updated dependencies to not include config.h.

17 years agoRemoved unneeded stuff.
Gisle Vanem [Tue, 31 Oct 2006 18:01:07 +0000 (18:01 +0000)]
Removed unneeded stuff.

17 years agoAdded Watt-32 section to fix things for Watt32+Win32 targets.
Gisle Vanem [Tue, 31 Oct 2006 17:54:51 +0000 (17:54 +0000)]
Added Watt-32 section to fix things for Watt32+Win32 targets.

17 years agoDon't include "nameser.h" for Watt32. Use the normal BSD-socket headers.
Gisle Vanem [Tue, 31 Oct 2006 17:51:54 +0000 (17:51 +0000)]
Don't include "nameser.h" for Watt32. Use the normal BSD-socket headers.

17 years agoAdded definition of select() for Watt32.
Gisle Vanem [Tue, 31 Oct 2006 17:25:48 +0000 (17:25 +0000)]
Added definition of select() for Watt32.

17 years agoRewritten to use ../packages/DOS/common.dj.
Gisle Vanem [Tue, 31 Oct 2006 17:24:25 +0000 (17:24 +0000)]
Rewritten to use ../packages/DOS/common.dj.

17 years agoChange 'FILETYPE' to ' VFT_APP'.
Gisle Vanem [Tue, 31 Oct 2006 16:25:30 +0000 (16:25 +0000)]
Change 'FILETYPE' to ' VFT_APP'.

17 years agoShow stderr log file for test 518 unconditionally.
Yang Tse [Tue, 31 Oct 2006 01:30:42 +0000 (01:30 +0000)]
Show stderr log file for test 518 unconditionally.

In this way we'll be able to sort out problems that might
arise in the prechek phase of the 518 test.

Once that 518 has been verified this change will be undone.

17 years agoSync comment with code and add three messages more
Yang Tse [Tue, 31 Oct 2006 01:24:03 +0000 (01:24 +0000)]
Sync comment with code and add three messages more

17 years agoAddress some pitfalls in the rlimit() function check that were
Yang Tse [Mon, 30 Oct 2006 17:24:31 +0000 (17:24 +0000)]
Address some pitfalls in the rlimit() function check that were
preventing execution of this test on many platforms

17 years agoAllow 'curl_*printf()' to be used in C++ programs.
Gisle Vanem [Mon, 30 Oct 2006 16:26:24 +0000 (16:26 +0000)]
Allow 'curl_*printf()' to be used in C++ programs.

17 years agoadd contributors from the 7.16.0 release
Daniel Stenberg [Mon, 30 Oct 2006 09:03:34 +0000 (09:03 +0000)]
add contributors from the 7.16.0 release

17 years agostart working on 7.16.1
Daniel Stenberg [Mon, 30 Oct 2006 08:52:12 +0000 (08:52 +0000)]
start working on 7.16.1

17 years ago7.16.0 material
Daniel Stenberg [Sun, 29 Oct 2006 23:03:14 +0000 (23:03 +0000)]
7.16.0 material

17 years agocorrected how tests/memanalyze.pl is used
Daniel Stenberg [Sun, 29 Oct 2006 23:00:52 +0000 (23:00 +0000)]
corrected how tests/memanalyze.pl is used

17 years agoCompiler warning fix
Yang Tse [Sun, 29 Oct 2006 21:19:23 +0000 (21:19 +0000)]
Compiler warning fix

17 years agoMake more human readable and maintainable previous
Yang Tse [Sun, 29 Oct 2006 14:58:59 +0000 (14:58 +0000)]
Make more human readable and maintainable previous
compiler warning fix since it was Ok and actually
avoids the targeted compiler warning.

17 years agoupdated to current status
Daniel Stenberg [Sun, 29 Oct 2006 09:18:32 +0000 (09:18 +0000)]
updated to current status

17 years agoupdated to reflect reality
Daniel Stenberg [Sun, 29 Oct 2006 09:11:44 +0000 (09:11 +0000)]
updated to reflect reality

17 years agoa small unification of the error text on failed server startups
Daniel Stenberg [Fri, 27 Oct 2006 21:07:08 +0000 (21:07 +0000)]
a small unification of the error text on failed server startups

17 years agoSync with lib/setup_once.h
Yang Tse [Fri, 27 Oct 2006 15:37:26 +0000 (15:37 +0000)]
Sync with lib/setup_once.h

17 years agoCompiler warning fix.
Yang Tse [Fri, 27 Oct 2006 15:32:18 +0000 (15:32 +0000)]
Compiler warning fix.

Assigning the const value zero to a pointer to function
results in a null pointer value assignment to the function
pointer.

Assignment of any nonzero value is what should result in a
implementation compiler dependent result.

Since what we want to do here is the first case, this should
not trigger compiler warnings related with conversions from
'pointer to data' to 'pointer to function'.

Our autobuild test suite will judge.

17 years agoFixed 'x_TYPE_ARG2' to match prototypes of recv() and send().
Gisle Vanem [Fri, 27 Oct 2006 14:13:32 +0000 (14:13 +0000)]
Fixed 'x_TYPE_ARG2' to match prototypes of recv() and send().

17 years agoGet rid of the special sread()+swrite() for MSDOS. Use recv()
Gisle Vanem [Fri, 27 Oct 2006 14:07:32 +0000 (14:07 +0000)]
Get rid of the special sread()+swrite() for MSDOS. Use recv()
and send(). Added needed HAVE_x defines.

17 years agoAdded 'RECV_TYPE_ARGx' needed in getinfo.c.
Gisle Vanem [Fri, 27 Oct 2006 13:57:16 +0000 (13:57 +0000)]
Added 'RECV_TYPE_ARGx' needed in getinfo.c.

17 years agoUpdate copyright year, since the file has been modified
Yang Tse [Fri, 27 Oct 2006 03:47:57 +0000 (03:47 +0000)]
Update copyright year, since the file has been modified

17 years agoCompiler warning fix
Yang Tse [Fri, 27 Oct 2006 02:18:29 +0000 (02:18 +0000)]
Compiler warning fix

17 years ago30 seconds isn't long enough for this test on a loaded server.
Yang Tse [Fri, 27 Oct 2006 01:58:59 +0000 (01:58 +0000)]
30 seconds isn't long enough for this test on a loaded server.

17 years agoDo an explicit typecast of data pointers to function pointers
Yang Tse [Fri, 27 Oct 2006 01:04:41 +0000 (01:04 +0000)]
Do an explicit typecast of data pointers to function pointers
to avoid picky compiler warnings, since this is what we want!

17 years agoUse proper 'stat' structure for fstat(). I.e. 'struct _stati64' and '_fstati64()'
Gisle Vanem [Thu, 26 Oct 2006 14:30:11 +0000 (14:30 +0000)]
Use proper 'stat' structure for fstat(). I.e. 'struct _stati64' and '_fstati64()'
on Win32.

17 years agoImproved rlimit logic:
Yang Tse [Thu, 26 Oct 2006 13:55:24 +0000 (13:55 +0000)]
Improved rlimit logic:
- Take in account RLIM_INFINITY.
- Verify that soft limit is actually changed when doing so.
- Show errno in case getrlimit or setrlimit fails.
- Keep file descriptors open only while runing this test.

17 years agoFix Curl_open() not reporting failure when allocation of the
Yang Tse [Thu, 26 Oct 2006 11:15:25 +0000 (11:15 +0000)]
Fix Curl_open() not reporting failure when allocation of the
buffer used to store headers in the SessionHandle failed.

17 years ago30 seconds isn't long enough for this test on a loaded server.
Yang Tse [Thu, 26 Oct 2006 09:50:21 +0000 (09:50 +0000)]
30 seconds isn't long enough for this test on a loaded server.

17 years agoa Smalltalk binding
Daniel Stenberg [Wed, 25 Oct 2006 21:07:26 +0000 (21:07 +0000)]
a Smalltalk binding

17 years agoFixed CURLOPT_FAILONERROR to return CURLE_HTTP_RETURNED_ERROR even for the
Daniel Stenberg [Wed, 25 Oct 2006 20:40:14 +0000 (20:40 +0000)]
Fixed CURLOPT_FAILONERROR to return CURLE_HTTP_RETURNED_ERROR even for the
case when 401 or 407 are returned, *IF* no auth credentials have been given.
The CURLOPT_FAILONERROR option is not possible to make fool-proof for 401
and 407 cases when auth credentials is given, but we've now covered this
somewhat more.

You might get some amounts of headers transferred before this situation is
detected, like for when a "100-continue" is received as a response to a
POST/PUT and a 401 or 407 is received immediately afterwards.

Added test 281 to verify this change.

17 years agoFixed "'x' might be used uninitialized in this function" warning.
Gisle Vanem [Wed, 25 Oct 2006 14:16:01 +0000 (14:16 +0000)]
Fixed "'x' might be used uninitialized in this function" warning.
Removed trailing whitespace.

17 years agoAdded '-d' option for Watt32 targets. Added cvs id.
Gisle Vanem [Wed, 25 Oct 2006 14:13:35 +0000 (14:13 +0000)]
Added '-d' option for Watt32 targets. Added cvs id.

17 years agoCompiler warning fix
Yang Tse [Wed, 25 Oct 2006 10:25:43 +0000 (10:25 +0000)]
Compiler warning fix

17 years agoAdd project notice and file Id
Yang Tse [Wed, 25 Oct 2006 09:20:44 +0000 (09:20 +0000)]
Add project notice and file Id

17 years agoCompiler warning fix
Yang Tse [Wed, 25 Oct 2006 08:52:00 +0000 (08:52 +0000)]
Compiler warning fix

17 years agoupdated copyright year
Daniel Stenberg [Wed, 25 Oct 2006 07:19:45 +0000 (07:19 +0000)]
updated copyright year

17 years agoUse curl_global_init() and curl_global_cleanup().
Yang Tse [Wed, 25 Oct 2006 05:59:46 +0000 (05:59 +0000)]
Use curl_global_init() and curl_global_cleanup().
Improve cleanup in case of initialization failure.

17 years agoother pipelining fixes by Ravi Pratap, that now makes pipelines get used better
Daniel Stenberg [Tue, 24 Oct 2006 21:14:40 +0000 (21:14 +0000)]
other pipelining fixes by Ravi Pratap, that now makes pipelines get used better

17 years agoAbort test if it seems that it would have run forever. This is just to prevent
Yang Tse [Tue, 24 Oct 2006 15:51:42 +0000 (15:51 +0000)]
Abort test if it seems that it would have run forever. This is just to prevent
test hanging and actually is an indication that there's a condition that is
not being properly handled at some point in the library.

Remove a pair of braces and adjust indentation appropriately.

17 years agothe check in ConnectionExists() for not re-using a non-resolved connection now
Daniel Stenberg [Mon, 23 Oct 2006 20:41:50 +0000 (20:41 +0000)]
the check in ConnectionExists() for not re-using a non-resolved connection now
applies for asynch name resolves in general and not only ares

17 years agoRavi Pratap provided a major update with pipelining fixes. We also no longer
Daniel Stenberg [Mon, 23 Oct 2006 20:34:56 +0000 (20:34 +0000)]
Ravi Pratap provided a major update with pipelining fixes. We also no longer
re-use connections (for pipelining) before the name resolving is done.

17 years agoAvoid trying to compare more than strlen bytes.
Yang Tse [Mon, 23 Oct 2006 19:16:19 +0000 (19:16 +0000)]
Avoid trying to compare more than strlen bytes.

17 years ago30 seconds isn't long enough for this test on a loaded server.
Dan Fandrich [Mon, 23 Oct 2006 19:15:14 +0000 (19:15 +0000)]
30 seconds isn't long enough for this test on a loaded server.

17 years agoReplace is*() macros with our own IS*() ones.
Yang Tse [Mon, 23 Oct 2006 19:14:54 +0000 (19:14 +0000)]
Replace is*() macros with our own IS*() ones.

17 years agoNir Soffer fixed a cp line and got rid of an rm
Daniel Stenberg [Sun, 22 Oct 2006 22:18:24 +0000 (22:18 +0000)]
Nir Soffer fixed a cp line and got rid of an rm

17 years agountil we learn how to use Makefile.inc from here, I've added socks.o in here
Daniel Stenberg [Sun, 22 Oct 2006 07:43:34 +0000 (07:43 +0000)]
until we learn how to use Makefile.inc from here, I've added socks.o in here
as well

17 years agoProvide 'datarootdir' parameter to shutup configuration warning,
Yang Tse [Sat, 21 Oct 2006 17:08:48 +0000 (17:08 +0000)]
Provide 'datarootdir' parameter to shutup configuration warning,
'packages/EPM/curl.list.in seems to ignore the --datarootdir setting'

17 years agoFix misplaced runtime library specification for 'release-dll' target
Yang Tse [Sat, 21 Oct 2006 16:25:31 +0000 (16:25 +0000)]
Fix misplaced runtime library specification for 'release-dll' target

17 years agorely on the global LDADD instead of having specific ones for every program
Daniel Stenberg [Sat, 21 Oct 2006 13:00:51 +0000 (13:00 +0000)]
rely on the global LDADD instead of having specific ones for every program

17 years agoNir Soffer for his Makefile.am fix
Daniel Stenberg [Sat, 21 Oct 2006 12:49:45 +0000 (12:49 +0000)]
Nir Soffer for his Makefile.am fix

17 years agoFix copy-paste error
Yang Tse [Sat, 21 Oct 2006 12:36:10 +0000 (12:36 +0000)]
Fix copy-paste error

17 years agoCompiler warning fix
Yang Tse [Sat, 21 Oct 2006 12:35:16 +0000 (12:35 +0000)]
Compiler warning fix

17 years agoNir Soffer made the tests/libtest/Makefile.am use a proper variable for all
Daniel Stenberg [Sat, 21 Oct 2006 11:40:04 +0000 (11:40 +0000)]
Nir Soffer made the tests/libtest/Makefile.am use a proper variable for all
the single test applications' link and dependences, so that you easier can
override those from the command line when using make.

17 years agoArmel Asselin separated CA cert verification problems from problems with
Daniel Stenberg [Sat, 21 Oct 2006 11:32:05 +0000 (11:32 +0000)]
Armel Asselin separated CA cert verification problems from problems with
reading the (local) CA cert file to let users easier pinpoint the actual
problem. CURLE_SSL_CACERT_BADFILE (77) is the new libcurl error code.

17 years agoCompiler warning fix
Yang Tse [Sat, 21 Oct 2006 10:54:41 +0000 (10:54 +0000)]
Compiler warning fix

17 years agomade the arrow for 'Send SSL data' point in the right direction!
Daniel Stenberg [Fri, 20 Oct 2006 21:26:10 +0000 (21:26 +0000)]
made the arrow for 'Send SSL data' point in the right direction!

17 years agoCompiler warning fix
Yang Tse [Fri, 20 Oct 2006 17:54:05 +0000 (17:54 +0000)]
Compiler warning fix

17 years agoSince now src/setup.h includes setup_once.h, src/config-win32.h needs
Yang Tse [Fri, 20 Oct 2006 17:16:06 +0000 (17:16 +0000)]
Since now src/setup.h includes setup_once.h, src/config-win32.h needs
the definitions for the return type and arguments types of functions
recv() and send().

17 years agoOops! Actually set the limit to 30 seconds.
Yang Tse [Fri, 20 Oct 2006 15:45:12 +0000 (15:45 +0000)]
Oops! Actually set the limit to 30 seconds.

17 years agoDecrease the posibility of aborting a test which actually is not
Yang Tse [Fri, 20 Oct 2006 15:39:54 +0000 (15:39 +0000)]
Decrease the posibility of aborting a test which actually is not
stale by replacing loop counters with timeouts. In this way the
main loop of the test will be allowed to run up to 30 seconds on
any platform before aborting it.

17 years agoWhen a resolve is made on a pipelined connection we need to detect it properly
Daniel Stenberg [Fri, 20 Oct 2006 12:25:39 +0000 (12:25 +0000)]
When a resolve is made on a pipelined connection we need to detect it properly
(when the resoling isn't completede yet) and not confuse it with a simple
connection re-use (non-pipelined).

17 years agoSet loop2 counter limit to 60 on this test to avoid a false positive.
Yang Tse [Thu, 19 Oct 2006 23:35:52 +0000 (23:35 +0000)]
Set loop2 counter limit to 60 on this test to avoid a false positive.

17 years agoReplace tabs with spaces and Compiler warning fix.
Yang Tse [Thu, 19 Oct 2006 22:49:33 +0000 (22:49 +0000)]
Replace tabs with spaces and Compiler warning fix.

17 years agoCompiler warning fix
Yang Tse [Thu, 19 Oct 2006 22:48:33 +0000 (22:48 +0000)]
Compiler warning fix

17 years agoWhen aborting, show loop counter values when more than one counter exists.
Yang Tse [Thu, 19 Oct 2006 21:12:27 +0000 (21:12 +0000)]
When aborting, show loop counter values when more than one counter exists.

17 years agoAbort test if it seems that it would have run forever. This is just to prevent
Yang Tse [Thu, 19 Oct 2006 17:29:25 +0000 (17:29 +0000)]
Abort test if it seems that it would have run forever. This is just to prevent
test hanging and actually is an indication that there's a condition that is
not being properly handled at some point in the library.

Loop counter limits might need to be further increased on false positives.

17 years agoHere's an effort to avoid saying 'data not shown' in the debug parts when the
Daniel Stenberg [Thu, 19 Oct 2006 14:28:50 +0000 (14:28 +0000)]
Here's an effort to avoid saying 'data not shown' in the debug parts when the
data is actually shown on screen. Like when you do 'curl -v host' with data
and debug info sent to the same terminal.

17 years agoBuilds using synchronous name resolver dislike marking the connection as async.
Yang Tse [Thu, 19 Oct 2006 02:30:02 +0000 (02:30 +0000)]
Builds using synchronous name resolver dislike marking the connection as async.

17 years agoSync with lib/setup_once.h
Yang Tse [Wed, 18 Oct 2006 21:25:12 +0000 (21:25 +0000)]
Sync with lib/setup_once.h

17 years agoCheck for USE_WINSOCK instead of WIN32 where the check was done
Yang Tse [Wed, 18 Oct 2006 21:05:40 +0000 (21:05 +0000)]
Check for USE_WINSOCK instead of WIN32 where the check was done
to verify winsock API availability.

17 years agoIntroduce symbol USE_WINSOCK which will be defined when
Yang Tse [Wed, 18 Oct 2006 15:57:49 +0000 (15:57 +0000)]
Introduce symbol USE_WINSOCK which will be defined when
using winsock or winsock2 API.

17 years agothe expire timer is a bit too annoying to see all the time ;-)
Daniel Stenberg [Wed, 18 Oct 2006 15:11:24 +0000 (15:11 +0000)]
the expire timer is a bit too annoying to see all the time ;-)

17 years agoWhen a connection is re-used, it can be flagged for re-use before the name
Daniel Stenberg [Wed, 18 Oct 2006 15:10:49 +0000 (15:10 +0000)]
When a connection is re-used, it can be flagged for re-use before the name
resolving is completed so we must make sure to survive it and mark the
connection as async (ie not yet connected completely).

17 years agouse the return code from lseek() to detect problems and bail out if so
Daniel Stenberg [Wed, 18 Oct 2006 14:47:58 +0000 (14:47 +0000)]
use the return code from lseek() to detect problems and bail out if so

17 years agoAdded ISPRINT() required for src/main.c.
Gisle Vanem [Wed, 18 Oct 2006 13:50:23 +0000 (13:50 +0000)]
Added ISPRINT() required for src/main.c.

17 years agoTor's spell fixes
Daniel Stenberg [Wed, 18 Oct 2006 12:59:02 +0000 (12:59 +0000)]
Tor's spell fixes

17 years agochanges done the last few days
Daniel Stenberg [Wed, 18 Oct 2006 11:13:56 +0000 (11:13 +0000)]
changes done the last few days

17 years agocut out matching host names starting with telnet or ftps, since they hardly
Daniel Stenberg [Wed, 18 Oct 2006 11:13:39 +0000 (11:13 +0000)]
cut out matching host names starting with telnet or ftps, since they hardly
ever actually are used

17 years agothe "work in progress" for #25 was ditched a long time ago
Daniel Stenberg [Wed, 18 Oct 2006 07:53:24 +0000 (07:53 +0000)]
the "work in progress" for #25 was ditched a long time ago

17 years agoReplace is*() macros with our own IS*() ones.
Yang Tse [Wed, 18 Oct 2006 03:42:06 +0000 (03:42 +0000)]
Replace is*() macros with our own IS*() ones.
Get rid of non ANSI/ISO isascii().

17 years agoMove definition of IS*() macros to setup_once.h
Yang Tse [Wed, 18 Oct 2006 03:41:19 +0000 (03:41 +0000)]
Move definition of IS*() macros to setup_once.h

17 years agoFixed compile error in HAVE_SIGACTION case.
Dan Fandrich [Tue, 17 Oct 2006 21:45:37 +0000 (21:45 +0000)]
Fixed compile error in HAVE_SIGACTION case.

17 years agoAvoid typecasting a signed char to an int when using is*() functions, as that
Daniel Stenberg [Tue, 17 Oct 2006 21:32:56 +0000 (21:32 +0000)]
Avoid typecasting a signed char to an int when using is*() functions, as that
could very well cause a negate number get passed in and thus cause reading
outside of the array usually used for this purpose.

We avoid this by using the uppercase macro versions introduced just now that
does some extra crazy typecasts to avoid byte codes > 127 to cause negative
int values.

17 years agoclear the struct size not the pointer size, pointed out in bug report
Daniel Stenberg [Tue, 17 Oct 2006 20:34:11 +0000 (20:34 +0000)]
clear the struct size not the pointer size, pointed out in bug report
#1579171

17 years agobuildconf already runs ares/buildconf by itself if there is an ares subdir
Daniel Stenberg [Tue, 17 Oct 2006 11:46:42 +0000 (11:46 +0000)]
buildconf already runs ares/buildconf by itself if there is an ares subdir
present, so there's no use to doing it again in this script!

17 years agoExplicit typecast for Curl_debug() size argument
Yang Tse [Tue, 17 Oct 2006 10:04:13 +0000 (10:04 +0000)]
Explicit typecast for Curl_debug() size argument

17 years agoTypo
Yang Tse [Tue, 17 Oct 2006 09:07:38 +0000 (09:07 +0000)]
Typo

17 years agomake the low_speed check set the expire timer so that it has a chance to work
Daniel Stenberg [Tue, 17 Oct 2006 09:05:44 +0000 (09:05 +0000)]
make the low_speed check set the expire timer so that it has a chance to work
even when using curl_multi_socket() or even using the multi_perform() when
relying on multi_timeout() to be good.

17 years agoPlease welcome our new haxx.se curl mirror, for really fast Swedish access.
Daniel Stenberg [Tue, 17 Oct 2006 08:07:48 +0000 (08:07 +0000)]
Please welcome our new haxx.se curl mirror, for really fast Swedish access.

17 years agoJeff helped me pinpoint that we didn't properly set the expire timer during
Daniel Stenberg [Tue, 17 Oct 2006 08:06:27 +0000 (08:06 +0000)]
Jeff helped me pinpoint that we didn't properly set the expire timer during
c-ares name resolves, but now we do!

17 years agofix the name resolve abort timeout calculation (when signals are used)
Daniel Stenberg [Tue, 17 Oct 2006 08:05:41 +0000 (08:05 +0000)]
fix the name resolve abort timeout calculation (when signals are used)