-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 According to Bruno Haible on 8/13/2009 3:41 AM: > Hi Eric, > > Wow! What a large change.
Largely repetitive, but it adds up fast! > You handled locale.h, stddef.h, stdio.h, stdlib.h, string.h, unistd.h, > wchar.h, > but time.h also should define NULL, according to POSIX. Oh, good point. I've included <time.h> in the patch refresh. > I got some "malformed patch" errors while applying the patch. Due to wrapped > lines I hate gmane's patch mangling, with no option for patch attachments. Updated patches are available here, with your changes squashed in: $ git pull git://repo.or.cz/gnulib/ericb.git master I still need time to test on my NetBSD access point. > The vasnprintf code is also shared with libintl, which does not use any > replacements of <locale.h>, <stdlib.h>, nor <stddef.h>. Therefore defining > wchar_t for all platforms, while possible, would not simplify vasnprintf. I agree that because of libintl, it would not reduce the amount of #ifdef, but it could at least let us support wchar_t in the -posix variants. But that's for another day. And my question still remains - are there any modern platforms that flat out lack the wchar_t type in <stddef.h>, regardless of whether <wchar.h> is supported? > > The replacement value that you are using for NULL is not right for C++. In > ISO C++ <http://www.open-std.org/jtc1/sc22/wg21/docs/papers/2005/n1905.pdf> > sections 18.1.(3) and footnote 188 say > "The macro NULL is an implementation-defined C++ null pointer constant in > this International Standard (4.10). > Possible definitions include 0 and 0L, but not (void*)0." IIUC, the draft version of C++0x is introducing __nullptr as a version of NULL that can't be confused with an integer context. So using 0 is correct for the current C++ standard, but not future-proof. > // g++ 3.1...4.3.3 all call A::foo (int). > // The openSUSE g++ 4.3.1 calls A::foo (long). Once the new C++ standard is out with its improved rules for NULL, a conformant compiler should call A::foo (void *) (or maybe it will fail with a complaint that the overload resolution between void* and char* is ambiguous; I'd have to reread the proposed spec to be sure). But I'm less worried about fixing that now (as you mentioned, the patch is already big), so I just went with your patch as-is. Besides, NetBSD's bug with under-parenthesized NULL appears to be C only; their C++ definition looks correct - from the /usr/include/sys/null.h of the NetBSD 1.6 machine that I have access to, I see: #ifndef NULL #if !defined(__GNUG__) || __GNUG__ < 2 || (__GNUG__ == 2 && __GNUC_MINOR__ < 90) #if defined(__AUDIT__) && !defined(__cplusplus) /* * This will cause GCC to emit a warning if you attempt to use NULL * with some non-pointer object, return value, etc. */ #define NULL (void *)0 #else #define NULL 0 #endif /* __AUDIT__ && !__cplusplus */ #else #define NULL __null #endif - -- Don't work too hard, make some time for fun as well! Eric Blake e...@byu.net -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (Cygwin) Comment: Public key at home.comcast.net/~ericblake/eblake.gpg Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAkqED2QACgkQ84KuGfSFAYASsACeKKs334+OAg9ph1QFLHgSUWz9 pXwAoLH5UiXh7eyJ96aTbfNJVZEDcOVa =Nxr5 -----END PGP SIGNATURE-----
