Push the floating-point test into a separate file.

Hopefully, that will make sure we don't optimize away anything that
will, for example, cause us not to do things differently on 32-bit x86
using the x87 instructions.

1 files changed, 3 insertions, 8 deletions

diff --git a/tcpdump.c b/tcpdump.c
index 70cbdedf..107cdbb6 100644
--- a/tcpdump.c
+++ b/tcpdump.c
@@ -150,6 +150,8 @@ The Regents of the University of California.  All rights reserved.\n";
 
 #include "print.h"
 
+#include "fptype.h"
+
 #ifndef PATH_MAX
 #define PATH_MAX 1024
 #endif
@@ -1901,14 +1903,7 @@ main(int argc, char **argv)
 			 * may differ (e.g., it might be 32-bit, 64-bit,
 			 * or 80-bit).
 			 */
-			{
-				union { /* int to float conversion buffer */
-					float f;
-					uint32_t i;
-				} f;
-				f.i = 0x4e93312d;
-				printf("%.3f\n", f.f*8/1000000);
-			}
+			float_type_check(0x4e93312d);
 			return 0;
 
 		default: