(Fsubst_char_in_region): Use replace_range in the case

where we may need to combine bytes.
(Ftranslate_region): Likewise.

diff --git a/src/editfns.c b/src/editfns.c
index aca1a3455e1864b66328ad42a63c64ee984f2bfc..c3cdad96aa0dcaf14afa8161e4c09d833e66d2a0 100644 (file)
--- a/src/editfns.c
+++ b/src/editfns.c
@@ -1938,9 +1938,44 @@ Both characters must have the same length of multi-byte form.")
              changed = 1;
            }
 
-         if (NILP (noundo))
-           record_change (pos, 1);
-         for (i = 0; i < len; i++) *p++ = tostr[i];
+         /* Take care of the case where the new character
+            combines with neighboring bytes.  */ 
+         if (len == 1
+             && ((! CHAR_HEAD_P (tostr[0])
+                  && pos_byte > BEGV_BYTE
+                  && ! ASCII_BYTE_P (FETCH_BYTE (pos_byte - 1)))
+                 ||
+                 (! ASCII_BYTE_P (tostr[0])
+                  && pos_byte + 1 < ZV_BYTE
+                  && ! CHAR_HEAD_P (FETCH_BYTE (pos_byte + 1)))))
+           {
+             Lisp_Object tem, string;
+
+             struct gcpro gcpro1;
+
+             tem = current_buffer->undo_list;
+             GCPRO1 (tem);
+
+             /* Make a multibyte string containing this
+                single-byte character.  */
+             string = Fmake_string (make_number (1),
+                                    make_number (tochar));
+             SET_STRING_BYTES (XSTRING (string), 1);
+             /* replace_range is less efficient, because it moves the gap,
+                but it handles combining correctly.  */
+             replace_range (pos, pos + 1, string,
+                            0, 0, 0);
+             if (! NILP (noundo))
+               current_buffer->undo_list = tem;
+
+             UNGCPRO;
+           }
+         else
+           {
+             if (NILP (noundo))
+               record_change (pos, 1);
+             for (i = 0; i < len; i++) *p++ = tostr[i];
+           }
        }
       INC_BOTH (pos, pos_byte);
     }
@@ -1995,9 +2030,33 @@ It returns the number of characters changed.")
          nc = tt[oc];
          if (nc != oc)
            {
-             record_change (pos, 1);
-             *p = nc;
-             signal_after_change (pos, 1, 1);
+             /* Take care of the case where the new character
+                combines with neighboring bytes.  */ 
+             if ((! CHAR_HEAD_P (nc)
+                  && pos_byte > BEGV_BYTE
+                  && ! ASCII_BYTE_P (FETCH_BYTE (pos_byte - 1)))
+                 ||
+                 (! ASCII_BYTE_P (nc)
+                  && pos_byte + 1 < ZV_BYTE
+                  && ! CHAR_HEAD_P (FETCH_BYTE (pos_byte + 1))))
+               {
+                 Lisp_Object string;
+
+                 string = Fmake_string (make_number (1),
+                                        make_number (nc));
+                 SET_STRING_BYTES (XSTRING (string), 1);
+
+                 /* This is less efficient, because it moves the gap,
+                    but it handles combining correctly.  */
+                 replace_range (pos, pos + 1, string,
+                                1, 0, 0);
+               }
+             else
+               {
+                 record_change (pos, 1);
+                 *p = nc;
+                 signal_after_change (pos, 1, 1);
+               }
              ++cnt;
            }
        }