diff --git a/C/alloc.c b/C/alloc.c
index 7dd87c2f0..8e6984a74 100644
--- a/C/alloc.c
+++ b/C/alloc.c
@@ -12,7 +12,7 @@
 * Last rev:								 *
 * mods:									 *
 * comments:	allocating space					 *
-* version:$Id: alloc.c,v 1.76 2005-11-23 03:01:33 vsc Exp $		 *
+* version:$Id: alloc.c,v 1.77 2005-11-23 12:09:50 vsc Exp $		 *
 *************************************************************************/
 #ifdef SCCS
 static char SccsId[] = "%W% %G%";
@@ -81,7 +81,6 @@ minfo(char mtype)
 static inline char *
 call_malloc(unsigned int size)
 {
-  char *tmp;
 #if INSTRUMENT_MALLOC
   if (mallocs % 1024*4 == 0) 
     minfo('A');
diff --git a/C/save.c b/C/save.c
index 870ea57a3..b95993cae 100644
--- a/C/save.c
+++ b/C/save.c
@@ -415,7 +415,6 @@ save_regs(int mode)
 #endif /* SBA || TABLING */
   }
   putout(CurrentModule);
-  putcellptr(AuxSp);
   if (mode == DO_EVERYTHING) {
 #ifdef COROUTINING
     putout(WokenGoals);
diff --git a/changes-5.1.html b/changes-5.1.html
index 22a5376a9..2d30b1959 100644
--- a/changes-5.1.html
+++ b/changes-5.1.html
@@ -16,6 +16,13 @@
 
 <h2>Yap-5.1.0:</h2>
 <ul>
+<li> FIXED: Restore could not recover DL_MALLOC info and was not recovering
+DBTerms correctly. </li>
+<li> FIXED: with locking field, AtomNil and friends might not be
+aligned, and would be confused with int. </li>
+<li> FIXED: variables should always be reset when doing trail
+expansion from the parser, even if there is no shifting (obs Trevor
+Walker). </li>
 <li> FIXED: growglobal with realloc was broken. </li>
 <li> FIXED: with threads, self-deadlocking in new array codes. </li>
 <li> FIXED: in tabling, use malloc when AllocCodeSpace fails. </li>