Bug report: inn-2.4.3 on 64 bit machines

[Ernst Boetsch]

Hello *,


in  "inn-2.4.3/tests/lib/snprintf-t.c"  there is a bug which is
only triggered on 64 bit machines (like AMD 64):

   int count;
   long lcount;

   [ ... ]

   test_format(23, true, "abcdefghijklmnopqrstuvwxyz01234", 36,
			"%ln%s%n", &count, string, &lcount);

The length modifier  "l"  is at the wrong place (patch as unified diff
at the end of the e-mail).  Therefore, test  "lib/snprintf 25"  fails.

The bug is only triggered on 64 bit machines because on 32 bit machines
there is no difference between  "int"  and  "long".


Kind regards,
    Ernst Boetsch

-- 
******************************************************************
Ernst N. Boetsch                     |  Leibniz Computer Center
                                     |     of the Bavarian Academy
E-Mail:  boetsch at lrz.de              |     of Sciences
Phone:   +49 89 289-28831 (/-28784)  |  Barer Strasse 21
Fax:     +49 89 2809460              |  D-80333 Muenchen, Germany

########################################################################

--- ./tests/lib/snprintf-t.c.orig	2006-03-20 05:14:57.000000000 +0100
+++ ./tests/lib/snprintf-t.c	2006-04-02 17:07:52.000000000 +0200
@@ -125,7 +125,7 @@
     test_format(21, true, "abcdefghij4444", 14, "%.10s%n%d", string, &count,
                 4444);
     ok(22, count == 10);
-    test_format(23, true, "abcdefghijklmnopqrstuvwxyz01234", 36, "%ln%s%n",
+    test_format(23, true, "abcdefghijklmnopqrstuvwxyz01234", 36, "%n%s%ln",
                 &count, string, &lcount);
     ok(24, count == 0);
     ok(25, lcount == 31);