Avoid wrong-order initialization on big endian

llarp/net/uint128.h

@@ -23,18 +23,24 @@ namespace llarp
 #endif
 
     // Initializes with 0s
-    constexpr uint128_t() : lower{0}, upper{0}
+    constexpr uint128_t() : uint128_t{0, 0}
     {
     }
 
     // Initializes with least-significant value
-    constexpr uint128_t(uint64_t lower) : lower{lower}, upper{0}
+    constexpr uint128_t(uint64_t lower) : uint128_t{0, lower}
     {
     }
 
     // Initializes with upper and lower values
     constexpr uint128_t(uint64_t upper, uint64_t lower)
+// clang-format off
+#ifdef __BIG_ENDIAN__
+        : upper{upper}, lower{lower}
+#else
         : lower{lower}, upper{upper}
+#endif
+    // clang-format on
     {
     }