On Sat, 17 Dec 2011, Dimitry Andric wrote:

Log:
 In usr.bin/csup/proto.c, use the correct printf length modifier to print
 an off_t.
...
Modified: head/usr.bin/csup/proto.c
==============================================================================
--- head/usr.bin/csup/proto.c   Sat Dec 17 13:14:44 2011        (r228625)
+++ head/usr.bin/csup/proto.c   Sat Dec 17 13:52:53 2011        (r228626)
...
@@ -751,7 +752,7 @@ proto_printf(struct stream *wr, const ch
                        break;
                case 'O':
                        off = va_arg(ap, off_t);
-                       rv = stream_printf(wr, "%llu", off);
+                       rv = stream_printf(wr, "%" PRId64, off);
                        break;
                case 'S':
                        s = va_arg(ap, char *);

PRId64 is another incorrect printf format.

off_t is typedefed so that it can be changed as neccessary.  Using
PRId64 hard-codes the assumption that it is precisely a 64 bit signed
integer.  It is indeed a signed integer (POSIX 2001 standard).  In
1990 POSIX, it was only required to be a signed arithmetic type, so
portable code had to handle the possibility that it was floating point,
and on systems with C90 compilers and 32-bit longs, it needed to be
floating point for it represent values a bit larger than 2**31-1.

In FreeBSD-1, it was 32 bits, so neither of the above would compile.
In FreeBSD[2-10], it is 64 bits integral.  FreeBSD depended on using
a non-C90 compiler even to declare it, and never needed floating point
for it, except for strict C90 support it would have needed a compat
layer with the int64_t kernel off_t trranslated to a long double
userland off_t.

Bruce
_______________________________________________
svn-src-all@freebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/svn-src-all
To unsubscribe, send any mail to "svn-src-all-unsubscr...@freebsd.org"

Reply via email to