]> oss.titaniummirror.com Git - msp430-gcc.git/blobdiff - libmudflap/testsuite/libmudflap.c/fail33-frag.c
Imported gcc-4.4.3
[msp430-gcc.git] / libmudflap / testsuite / libmudflap.c / fail33-frag.c
diff --git a/libmudflap/testsuite/libmudflap.c/fail33-frag.c b/libmudflap/testsuite/libmudflap.c/fail33-frag.c
new file mode 100644 (file)
index 0000000..ac7c617
--- /dev/null
@@ -0,0 +1,25 @@
+#include <stdlib.h>
+
+#define SIZE 16
+
+char b[SIZE];
+char a[SIZE];
+
+int main ()
+{
+  int i, j=0, k;
+  int a_before_b = (& a[0] < & b[0]);
+  /* Rather than iterating linearly, which would allow loop unrolling
+     and mapping to pointer manipulation, we traverse the "joined"
+     arrays in some random order.  */
+  for (i=0; i<SIZE*2; i++)
+    {
+      k = rand() % (SIZE*2);
+      j += (a_before_b ? a[k] : b[k]);
+    }
+  return j;
+}
+/* { dg-output "mudflap violation 1.*" } */
+/* { dg-output "Nearby object.*" } */
+/* { dg-output "mudflap object.*\[ab\]" } */
+/* { dg-do run { xfail *-*-* } } */