commit 3b2623b6c00a91f83b51e6ef8ddf0398eb8cdf5c
Author: Sam Trenholme <git@samiam.org>
Date:   Mon Oct 5 10:47:39 2026 -0600

    Sometimes, clang does the wrong thing when rotating by word size
    
    If you have a say 8-bit number, “foo”, and you rotate it eight bits
    to the left (or right), GCC does the right thing: The number becomes 0.
    clang, on the other hand, can and does sometimes make it a value
    besides 0, and, yes, the C99 spec allows that, from page 84 the
    the C99 spec:
    
    >If the value of the right operand is negative or is greater than or equal
    >to the width of the promoted left operand, the behavior is undefined.
    
    (As I recall, C11 defines negative shifts)

diff --git a/deadwood-github/src/DwSys.c b/deadwood-github/src/DwSys.c
index 1ca3d7e..b6dc660 100644
--- a/deadwood-github/src/DwSys.c
+++ b/deadwood-github/src/DwSys.c
@@ -1043,7 +1043,12 @@ void make_netmask(int num, uint8_t *str, int len) {
 
         /* The last byte in the string is determined by the modulo */
         last = 0xff;
-        last <<= (8 - rem);
+	/* clang crazyness workaround */
+        if(rem == 0) {
+                last = 0;
+        } else {
+                last <<= (8 - rem);
+        }
 
         /* This kind of coding is dangerous.  I have triple-checked
          * the following code and don't see any possible overflows */