mirror/curl
2
0
Fork 0
mirror of https://github.com/curl/curl.git synced 2024-12-09 06:30:06 +08:00
Code Issues Packages Projects Releases Wiki Activity
20,596 Commits 18 Branches 221 Tags 185 MiB
dc68f2dab9
Commit Graph

272 Commits

Author SHA1 Message Date
Yang Tse
305e03905f Cleanup. Warnings related with FD_SET, FD_ISSET, and FD_ZERO macros 
are not icc 9.0 specific.
 2007-04-03 02:45:04 +00:00
Dan Fandrich
59c620bfa5 Daniel Johnson's fix for shared object extension detection on Mac OS X. 2007-03-27 05:10:20 +00:00
Yang Tse
2288094b26 fix wrong macro name introduced in las commit 2007-03-22 18:59:14 +00:00
Yang Tse
273b2b230b Add check for compiler variadic macro support in configuration script 2007-03-22 18:25:38 +00:00
Yang Tse
8920606b8b attempt to keep message length below 80 chars 2007-03-22 14:41:10 +00:00
Yang Tse
d381dd68cf icc 9.0 when compiling its generated code for its own FD_SET, 
FD_ISSET, and FD_ZERO macros emits warnings #1469 and #593.
So for icc 9.0 we also ignore warnings #1469 and #593.
* 593 warns on "variable __d0 was set but never used"
* 1469 warns on "cc clobber ignored"
 2007-02-15 14:02:32 +00:00
Dan Fandrich
23d1041bd8 Added last-resort dynamic library names. 2007-02-13 19:59:58 +00:00
Dan Fandrich
5565f45f5e Properly use libtool macros to fix OpenLDAP library name detection on Darwin. 2007-01-31 23:15:47 +00:00
Yang Tse
73226415fc Added a check in configure that verifies if <signal.h> is available, 
defining HAVE_SIGNAL_H if the header is available.

Added a check in configure that tests if the sig_atomic_t type is
available, defining HAVE_SIG_ATOMIC_T if it is available. Providing
a suitable default in setup_once.h if not available.

Added a check in configure that tests if the sig_atomic_t type is
already defined as volatile, defining HAVE_SIG_ATOMIC_T_VOLATILE
if it is available and already defined as volatile.
 2006-11-22 18:41:34 +00:00
Daniel Stenberg
67e8d22958 Added a check in configure that simply tries to run a program (not when 
cross-compiling) in order to detect problems with run-time libraries that
otherwise would occur when the sizeof tests for curl_off_t would run and
thus be much more confusing to users. The check of course should run after
all lib-checks are done and before any other test is used that would run an
executable built for testing-purposes.
 2006-10-16 08:30:54 +00:00
Yang Tse
32ac4edeed Check for struct timeval at configuration time 2006-10-13 01:35:14 +00:00
Yang Tse
7240acdebc Being unable to link or find out recv() or send() args types is a fatal error. 2006-08-04 00:39:34 +00:00
Yang Tse
ae8a01ead6 Avoid the risk of a false positive detection of MSG_NOSIGNAL when cross compiling a Windows target. 2006-07-31 18:41:29 +00:00
Yang Tse
77b3bc239d First step trying to avoid the multiple header inclusion and recursion nightmare. 
Reintroduce checking for HAVE_MSG_NOSIGNAL in configure script, so that we don't depend on header inclusion order for a valid check.
 2006-07-28 14:19:02 +00:00
Yang Tse
13616f8f96 Simplify check for NEED_MALLOC_H, and make more explicit that NEED_MALLOC_H shall be defined if <malloc.h> header file must be included even when including <stdlib.h>. 2006-07-25 13:49:49 +00:00
Yang Tse
4d95d23d99 Define NEED_MALLOC_H if including <stdlib.h> is not enough for proper compilation and <malloc.h> must also be included. 2006-07-11 20:40:38 +00:00
Yang Tse
8ef454dcbe Finally get rid of CURL_CHECK_HEADERS_ONCE since it adds very little value and has portability issues. 
Change some shell if...then...fi tests into case...esac tests which demand less resources.
 2006-07-07 17:34:40 +00:00
Yang Tse
551a041283 Substitution of the literal '-' is only done if it's the first or last character. 2006-07-07 14:03:17 +00:00
Yang Tse
03288943af Using backslashes and slashes in the strings of the sed 'y' command shall be avoided since its interpretation is not the same across platforms. 
Now we use the sed 's' command with a bracket expression.
 2006-07-07 12:59:45 +00:00
Yang Tse
05edd48ad0 Fix excessive escaping. 2006-07-07 05:39:13 +00:00
Yang Tse
266ab95557 Fix CURL_CHECK_HEADERS_ONCE 2006-07-07 04:42:11 +00:00
Yang Tse
0a4bba565c Fix MinGW/MSYS support in CURL_CHECK_FUNC_RECV and CURL_CHECK_FUNC_SEND. 2006-07-06 15:51:03 +00:00
Yang Tse
bc2f0c7dcb Prevent definition of HAVE_WINxxx_H symbols and avoid inclusion of Windows headers when compiled with Cygwin in POSIX emulation mode. 2006-07-05 23:10:37 +00:00
Yang Tse
55329b56cb Use a more descriptive var name. 2006-07-04 17:19:15 +00:00
Yang Tse
7e43d06b60 Get qualifier of arg 2 for send() apart into SEND_QUAL_ARG2. 2006-07-04 16:54:10 +00:00
Yang Tse
01fa02d0b5 Find out return types and argument types for functions recv() and send() at configuration stage. 2006-07-04 13:03:32 +00:00
Yang Tse
bec1977137 Use CURL_CHECK_FUNC_GETNAMEINFO results in CURL_CHECK_NI_WITHSCOPEID 2006-07-03 15:32:12 +00:00
Yang Tse
4c08eb4b11 Make CURL_CHECK_NI_WITHSCOPEID actually try to compile NI_WITHSCOPEID when cross-compiling. 2006-07-02 23:09:46 +00:00
Yang Tse
0163730437 Fix shell globbing in CURL_CHECK_FUNC_GETNAMEINFO 2006-07-02 01:21:54 +00:00
Yang Tse
39745ac38e Fix shell globbing in CURL_CHECK_FUNC_GETNAMEINFO 2006-07-02 01:17:37 +00:00
Yang Tse
a1c6d5861a Get some debug info 2006-07-01 17:07:12 +00:00
Yang Tse
589c4596d8 Get some debug info 2006-07-01 15:01:48 +00:00
Yang Tse
7a6d7fca42 Avoid shell globbing 2006-07-01 12:53:24 +00:00
Yang Tse
5305c9f1e8 Get qualifier of arg 1 for getnameinfo apart. Take 3. 2006-07-01 11:21:38 +00:00
Yang Tse
b1022ea4c1 Get qualifier of arg 1 for getnameinfo apart. 2006-07-01 03:07:07 +00:00
Yang Tse
2e0ad842d0 Remove experimental notice from CURL_CHECK_FUNC_GETNAMEINFO 2006-06-30 00:22:38 +00:00
Daniel Stenberg
2df622fd14 detect ICC and pass on "-we 147" so that the configure checks for function 
arguments work properly - and the option is not harmful for the rest of the
curl build either!
 2006-04-19 11:11:10 +00:00
Daniel Stenberg
65afc576ea configure no longer warns on "missing" if the current path contains a space 2006-01-18 10:00:36 +00:00
Yang Tse
e4388643f1 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
 2005-12-21 20:44:54 +00:00
Yang Tse
bc4208201c Undefine HAVE_GETNAMEINFO if unable to find 
proper types to use for getnameinfo args
 2005-12-21 17:51:08 +00:00
Yang Tse
a0d69d52a1 Undefine HAVE_GETNAMEINFO if unable to find 
proper types to use for getnameinfo args
 2005-12-21 17:20:22 +00:00
Yang Tse
e3657644d6 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.
 2005-12-20 23:49:14 +00:00
Yang Tse
df2b1251a0 fix ioctlsocket detection 2005-12-20 20:29:53 +00:00
Daniel Stenberg
d551412a32 added our standard source header 2005-12-20 09:02:36 +00:00
Daniel Stenberg
6de67a134e fix closing parentheses 2005-12-20 08:51:48 +00:00
Yang Tse
fd0d560b47 Fix quoting 2005-12-20 03:23:49 +00:00
Yang Tse
7b488a3bef Use native type SOCKET instead of int when testing functionality of ioctlsocket on Windows 2005-12-20 00:27:59 +00:00
Yang Tse
f448168501 Ooops 2005-12-19 23:32:52 +00:00
Yang Tse
898bb397b1 Add checking for type DWORD as argument 4 and 6 of getnameinfo 2005-12-19 21:45:43 +00:00
Yang Tse
74ed5b5ebd Adjust more windows header includes 2005-12-19 21:38:10 +00:00
Yang Tse
d6eb1a7b98 Fix guard detection of _WIN32_WINNT for MingW in CURL_FUNC_GETNAMEINFO_ARGTYPES 2005-12-19 05:57:17 +00:00
Yang Tse
4ff56b15e9 Add check for 'unsigned int' as type of arguments 4 and 6 of getnameinfo 2005-12-19 05:32:33 +00:00
Yang Tse
5fd8dd2dce When checking the type of the first argument of getnameinfo do it in the 
following order: 'struct sockaddr *' 'const struct sockaddr *' 'void *'.
 2005-12-18 20:24:35 +00:00
Yang Tse
8a3280a2de MingW guards getnameinfo, getaddrinfo and freeaddrinfo with _WIN32_WINNT >= 0x0501 2005-12-18 06:07:10 +00:00
Yang Tse
939d368d5f Check first arg of getnameinfo with and without const qualifier. 2005-12-17 17:33:14 +00:00
Yang Tse
47c06fa308 Avoid breaking configure due to CURL_FUNC_GETNAMEINFO_ARGTYPES failure, since at this point nothing depends on it. 2005-12-17 02:41:33 +00:00
Yang Tse
b7f447f8d8 const qualifier in getnameinfo check 2005-12-17 02:32:55 +00:00
Yang Tse
03af76b631 Since there is no proof of the existence of a platform which would 
justify checking for socklen_t in more than one function, the code
used to find a valid socklen_t replacement is simplified back. The
only function that will be used to find a socklen_t replacement is
getpeername, as it has been since revision 1.4 of curl/acinclude.m4
 2005-12-16 23:15:33 +00:00
Yang Tse
d6ffb4c177 TYPE_SOCKLEN_T completely replaced by CURL_CHECK_TYPE_SOCKLEN_T. 
CURL_FUNC_GETNAMEINFO_ARGTYPES now also checks first argument.
All related changes taken to cares configuration scripts.
 2005-12-16 18:18:00 +00:00
Yang Tse
fea5ddf585 More quotes 2005-12-16 07:28:56 +00:00
Yang Tse
b9f39c2711 Fix copy paste bug 2005-12-16 06:50:03 +00:00
Yang Tse
5acac0309a Test CURL_CHECK_TYPE_SOCKLEN_T 2005-12-16 05:05:55 +00:00
Yang Tse
78febad718 Oops. Wrong double quotes 2005-12-16 04:54:56 +00:00
Yang Tse
a6a5bba0a9 Experimental check for socklen_t CURL_CHECK_TYPE_SOCKLEN_T 2005-12-16 04:18:44 +00:00
Yang Tse
51581c034d Tests to check the availability of compilable and 
valid windows.h winsock.h winsock2.h and ws2tcpip.h
header files:
CURL_CHECK_HEADER_WINDOWS
CURL_CHECK_HEADER_WINSOCK
CURL_CHECK_HEADER_WINSOCK2
CURL_CHECK_HEADER_WS2TCPIP
 2005-12-15 19:39:44 +00:00
Yang Tse
7f5d092223 Some preprocessors have problems if the # character isn't at position 1. 2005-12-14 21:09:30 +00:00
Yang Tse
2645782f89 Determine the correct type to be passed to four of the getnameinfo' function's arguments, and define those types in GETNAMEINFO_TYPE_ARG2', GETNAMEINFO_TYPE_ARG46', and GETNAMEINFO_TYPE_ARG7'. 2005-12-14 19:00:44 +00:00
Daniel Stenberg
be524fed38 on windows (mingw32) the sockaddr_storage struct is in winsock2.h 2005-11-12 19:11:02 +00:00
Daniel Stenberg
c7b7bf32f8 removed the unreachable code warning from gcc debug builds, even the most 
recent gcc versions give far too many false positives for this to be valuable
 2005-08-19 15:07:50 +00:00
Daniel Stenberg
ef35519696 removed some inaccurate comments about the TYPE_IN_ADDR_T check 2005-08-19 07:02:24 +00:00
Dan Fandrich
e81d4ac8c8 Add -Wdeclaration-after-statement to gcc to detect accidental C99-style 
variable declarations.
 2005-07-20 21:58:23 +00:00
Dan Fandrich
b37a0b0512 Detect (or at least infer) glibc-style strerror_r even when cross-compiling. 2005-06-30 04:53:05 +00:00
Daniel Stenberg
b9e5302f27 adjusted the strerror_r test more, use _REENTRANT instead of _THREAD_SAFE 
when looking for the prototype
 2005-05-18 15:15:00 +00:00
Daniel Stenberg
dd7b6aef39 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).
 2005-05-18 13:24:23 +00:00
Daniel Stenberg
07e58aaa79 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.
 2005-05-02 14:06:27 +00:00
Daniel Stenberg
a31ddd363b Now configure checks for struct sockaddr_storage and the ftp code tries 
to survive without it if not found. AIX 4.3 targetted adjustment.
 2005-05-02 11:56:15 +00:00
Daniel Stenberg
4e555c7b32 When cross-compiling, we do some better checking for the NI_WITHSCOPEID option instead of just assuming it is present. 2005-03-22 18:02:06 +00:00
Dan Fandrich
1f68fa19c7 Finally fixed the LDAP library searching bug on libtool ver. 1.5 2005-03-15 04:47:07 +00:00
Dan Fandrich
228e627475 Use the libtool variables better to make LDAP library search work on 
more platforms.
 2005-03-14 19:37:07 +00:00
Dan Fandrich
b5112096d7 Work around a bug in libtool ver. 1.5 during LDAP library detection. 2005-03-11 23:07:40 +00:00
Dan Fandrich
205f8b266c Fixed LDAP library file name bug (KNOWN_BUGS #1). configure now auto-detects 
the correct dynamic library names by default, and provides override switches
--with-ldap-lib, --with-lber-lib and --without-lber-lib.  Added
CURL_DISABLE_LDAP to platform-specific config files to disable LDAP
support on those platforms that probably don't have dynamic OpenLDAP
libraries available to avoid compile errors.
 2005-03-11 05:28:07 +00:00
Daniel Stenberg
58f4af7973 Marcin Konicki provided two configure fixes and a source fix to make curl 
build out-of-the-box on BeOS.
 2004-12-22 22:28:10 +00:00
Dan Fandrich
0e59182945 gcc 2.7 can't handle a few warning options that gcc 2.95 can. 2004-12-20 18:20:17 +00:00
Dan Fandrich
38181fbc74 Fix the --enable-debug compiler warning options for older versions of gcc. 2004-12-15 21:09:53 +00:00
Daniel Stenberg
18dd8154e2 Reverted the previous change and redid it differently as it seemed to not 
work. This is supposed to detect cross-compiling and alert the user, and not
do the POSIX-check for strerror_r() if it already detected a glibc-compatible
strerror_r().
 2004-06-07 08:30:36 +00:00
Daniel Stenberg
e0b436f8e1 1 - do better when cross-compiling when checking for strerror_r() - alert 
the user.
2 - don't check for POSIX-style if glibc-style is found first
 2004-06-04 19:01:58 +00:00
Daniel Stenberg
55c015c136 typo AGAIN 2004-05-06 11:02:50 +00:00
Daniel Stenberg
bde6c6685e When using the icc compiler, we also ignore remark #1418 "external definition 
with no prior declaration" since this is a habit we have in the code.
 2004-05-06 10:58:57 +00:00
Daniel Stenberg
6f74820cfc if no strerror_r prototype is found, we provide our own to prevent picky 
compilers to warn
 2004-05-06 10:49:40 +00:00
Daniel Stenberg
aae521d086 gcc 3.4 now uses the -Wunreachable-code option, I believe we can make older 
ones use this too...
 2004-05-05 06:59:27 +00:00
Daniel Stenberg
1fc7ff878e Applied Joe Halpin's bugfixes to the NI_WITHSCOPEID test program. 2004-04-01 06:53:11 +00:00
Daniel Stenberg
5d27f50f2f HAVE_NI_WITHSCOPEID spelled right! 2004-03-31 20:13:53 +00:00
Daniel Stenberg
40d9855df2 Andrs Garca fixed a warning in the ioctlsocket() usage. 2004-03-31 10:34:53 +00:00
Daniel Stenberg
ecf7adba15 modified the NI_WITHSCOPEID to use an AF_INET6 socket immediately and 
added some more debug output to make it easier to detect failure reasons
in the autobuild logs
 2004-03-31 10:31:08 +00:00
Daniel Stenberg
a2ea0abf7f Added CURL_CHECK_NI_WITHSCOPEID that checks if NI_WITHSCOPEID exists and 
works. No code actually uses the HAVE_NI_WITHSCOPEID (that a positive test
results in), but this is still only for testing purposes.
 2004-03-30 10:35:54 +00:00
Daniel Stenberg
7461592a16 strerror_r() detection changes: 
1. Try with _THREAD_SAFE instead of _REENTRANT, as AIX seems to require it
   and if _REENTRANT is required we should already have it set since one of
   the previous tests.
2. Added API-detection for what kind of strerror_r() that is provided. The
   POSIX style or the glibc style.

Tor Arntsen provided the necessary feedback these changes are based upon.
 2004-03-25 12:15:00 +00:00
Daniel Stenberg
242be55771 added check for strerror_r() 2004-03-24 08:45:58 +00:00
Daniel Stenberg
78e0ba82b0 ignore the icc warning 981 "operands are evaluated in unspecified order" 2004-03-10 09:52:04 +00:00
Daniel Stenberg
7e6043f04f we ignore the ICC warning 1419 as well 2004-03-08 11:37:12 +00:00
Daniel Stenberg
6789dfc5f5 Perhaps -Wundef is better on gcc versions after 2.95, since the autobuild 
on FreeBSD gives us lots of warnings in system headers and I suspect this
option is what causes them!
 2004-03-01 13:10:35 +00:00
Daniel Stenberg
193a652e6a When setting aggressive pedantic compiler options, display what options 
that were set. For easier debugging/changing of this.
 2004-03-01 13:02:34 +00:00
Daniel Stenberg
85af357d81 Make icc ignore "invalid format string conversion" warnings as well. They 
appear because of our home-grown option '%Od' for the curl_off_t output.
 2004-02-26 12:40:09 +00:00
Daniel Stenberg
0943f33438 Use __INTEL_COMPILER instead of __ICC to the cpp to detect the Intel icc 
compiler!
 2004-02-26 11:37:14 +00:00
Daniel Stenberg
6bd2de0c42 set debug options when using the icc compiler 2004-02-25 14:14:30 +00:00
Daniel Stenberg
4fa58560bf Moved most of the set-debug-options-depending-on-compiler logic to the new 
CURL_CC_DEBUG_OPTS function in acinclude.m4
 2004-02-25 10:19:02 +00:00
Daniel Stenberg
4c70c3e820 Added escaping of the function name in the AC_DEFUN() macros, so that 
automake 1.8+ stops complaining!
 2004-01-15 12:48:13 +00:00
Daniel Stenberg
5f4fb857f0 Added proper #ifdef's for the #include files when checking for in_addr_t, 
which made the test start working nicely on BeOS. Shard verified it.
 2003-10-30 07:08:45 +00:00
Daniel Stenberg
bf20ed5984 O_NONBLOCK does not work on BeOS according to Shard 2003-10-06 22:03:11 +00:00
Daniel Stenberg
0bbe184f1f improved the non-block check a lot: 
* several tests ALWAYS failed due to bad test programs
* the IoctlSocket() test now is linked on test as it could otherwise compile
  fine on many systems that doesn't support it
* added a test for the BeOS way of doing non-blocking sockets (based on a
  patch from Shard)
 2003-10-05 08:35:55 +00:00
Daniel Stenberg
de9b76cef0 change the order of the in_addr_t tests, so that 'unsigned long' is tested 
for first, as it seems to be what many systems use
 2003-05-15 21:13:36 +00:00
Daniel Stenberg
b9c60df04b added description in all AC_DEFINE() calls 2003-01-20 15:16:56 +00:00
Daniel Stenberg
325391aef9 Albert Chin: 
Forgot one case. On HP-UX 11.00, gethostbyname_r() is properly defined
if -D_REENTRANT is used. Without it, the compiler still accepts the
function prototype but gives a warning about hostent_data going out of
scope. This is because struct hostent_data is not declared. So, we
force an error by trying to set a variable to the struct.
 2002-01-31 07:53:20 +00:00
Daniel Stenberg
d220389647 Stoned Elipot's patch for the in_addr_t test 2002-01-14 07:53:09 +00:00
Daniel Stenberg
820de919b6 now sets a type for in_addr_t even if it isn't found in the #include files 
like on my linux box
 2001-11-28 23:14:20 +00:00
Daniel Stenberg
d09b436937 Added an in_addr_t check 2001-11-28 13:16:56 +00:00
Daniel Stenberg
6ebac3dc76 now we make sure that NULL is defined in the gethostbyname_r() compiles 
as it turned out they aren't everywhere, and that causes compiles to fail
and then we don't find the proper function call!
 2001-11-08 14:48:50 +00:00
Daniel Stenberg
c3363f833c Albert Chin's improved gethostbyname_r() tests 2001-10-04 14:05:12 +00:00
Daniel Stenberg
6ca45beaed Added non-blocking sockets test 2001-10-01 08:58:03 +00:00
Daniel Stenberg
a1cec0e49a Albert Chin's neat configure/package fixes 2001-08-20 07:10:28 +00:00
Daniel Stenberg
5c6859e537 not used in autoconf version 2.50 2001-05-29 19:28:50 +00:00
Daniel Stenberg
8755c44d40 Albert Chin's patch posted to the mailing list 19 Apr 2001 2001-04-19 06:01:48 +00:00
Daniel Stenberg
76229d3da1 largefile.m4 was turned into acinclude.m4 2000-08-24 19:16:26 +00:00