diff --git a/UnicodeCLib/basic.c b/UnicodeCLib/basic.c
index be91265f0..d5189fffb 100644
--- a/UnicodeCLib/basic.c
+++ b/UnicodeCLib/basic.c
@@ -1,9 +1,4 @@
 #include <lean/lean.h>
-#ifdef _WIN32
-#include <winsock2.h>
-#else
-#include <arpa/inet.h>
-#endif
 #include "basic.h"
 
 static inline int unicode_script_is_valid(uint32_t c) {
@@ -19,16 +14,19 @@ static inline int unicode_script_is_valid(uint32_t c) {
 }
 
 uint32_t unicode_script_of_abbrev(b_lean_obj_arg abbr) {
-  lean_string_object * str = lean_to_string(abbr);
-  assert(str->m_size > 4);
-  uint32_t val = *((uint32_t*)(str->m_data));
-  return ntohl(val);
+  assert(lean_string_size(abbr) > 4);
+  const unsigned char * s = (const unsigned char *)lean_string_cstr(abbr);
+  return (uint32_t)s[0] << 24 | (uint32_t)s[1] << 16 | (uint32_t)s[2] << 8 | (uint32_t)s[3];
 }
 
 lean_obj_res unicode_script_to_abbrev(uint32_t scr) {
   assert(unicode_script_is_valid(scr));
   lean_object * abbr = lean_alloc_string(5, 5, 4);
-  lean_to_string(abbr)->m_data[4] = 0;
-  *((uint32_t*)lean_to_string(abbr)->m_data) = htonl(scr);
+  char * s = lean_to_string(abbr)->m_data;
+  s[0] = (char)(scr >> 24 & 0xff);
+  s[1] = (char)(scr >> 16 & 0xff);
+  s[2] = (char)(scr >> 8 & 0xff);
+  s[3] = (char)(scr & 0xff);
+  s[4] = 0;
   return abbr;
 }
diff --git a/lakefile.lean b/lakefile.lean
index f902b6b2e..deb938343 100644
--- a/lakefile.lean
+++ b/lakefile.lean
@@ -19,7 +19,7 @@ target UnicodeCLib pkg : FilePath := do
       let obj := pkg.buildDir / "UnicodeCLib" / ((file.fileName.dropSuffix ".c" |>.copy) ++ ".o")
       let src ← inputTextFile file.path
       let weakArgs := #["-I", (← getLeanIncludeDir).toString, "-I", (pkg.dir / "UnicodeCLib").toString, "-O", "-fPIC"]
-      oFiles := oFiles.push <| ← buildO obj src weakArgs
+      oFiles := oFiles.push <| ← buildO obj src weakArgs (extraDepTrace := getLeanTrace)
   let name := nameToStaticLib "unicodeclib"
   buildStaticLib (pkg.sharedLibDir / name) oFiles