[Patch] multibyte encodings in strings

[m4tze at exitno dot de]

Dear binutils team,

I am sending you a patch fixing an issue in the binary 'strings'.
The issue concerned finding multibyte encoded strings at odd offsets.

To test this issue try the following line

    echo "aa53007400720069006e0067003100aa53007400720069006e0067003200bb" | xxd -p -r | strings -el

Only the second string will be found.

The attached patch fixes this problem by doing only a single byte step if an
invalid character has been found.

Do not hesitate to contact me in case of questions or feedback.

Happy to contribute,

Mathias

diff --git a/binutils/strings.c b/binutils/strings.c
index 74545dbbdc..5bd31213c0 100644
--- a/binutils/strings.c
+++ b/binutils/strings.c
@@ -540,9 +540,41 @@ print_strings (const char *filename, FILE *stream, file_ptr address,
 	      free (buf);
 	      return;
 	    }
-	  if (! STRING_ISGRAPHIC (c))
-	    /* Found a non-graphic.  Try again starting with next char.  */
+	  if (! STRING_ISGRAPHIC (c)){
+	    /* Found a non-graphic. Try again starting with next char. */
+        if (encoding_bytes > 1){
+          /* In case of multibyte encodings rewind using magic buffer. */
+          if (magiccount == 0){
+            /* If no magic buffer exists: use memory of c */
+            switch (encoding)
+            {
+              default:
+                break;
+              case 'b':
+                c = c & 0xff;
+                magiccount += 1;
+                break;
+              case 'l':
+              case 'L':
+                c = c >> 8;
+                magiccount += (encoding_bytes -1);
+                break;
+              case 'B':
+                c = (( c & 0xff0000) >> 16) | ( c & 0xff00)
+                    | (( c & 0xff) << 16);
+                magiccount += 3;
+                break;
+            }
+            magic = (char *)&c;
+          }
+          else {
+             /* If magic buffer exists: rewind. */
+             magic = magic - (encoding_bytes -1);
+          }
+          address = address - (encoding_bytes -1);
+        }
 	    goto tryline;
+      }
 	  buf[i] = c;
 	}
 
@@ -621,8 +653,40 @@ print_strings (const char *filename, FILE *stream, file_ptr address,
 	  c = get_char (stream, &address, &magiccount, &magic);
 	  if (c == EOF)
 	    break;
-	  if (! STRING_ISGRAPHIC (c))
+	  if (! STRING_ISGRAPHIC (c)){
+        if (encoding_bytes > 1){
+        /*In case of multibyte encodings rewind using magic buffer. */
+          if (magiccount == 0){
+            /* If no magic buffer exists: use memory of c */
+            switch (encoding)
+            {
+              default:
+                break;
+              case 'b':
+                c = c & 0xff;
+                magiccount += 1;
+                break;
+              case 'l':
+              case 'L':
+                c = c >> 8;
+                magiccount += (encoding_bytes -1);
+                break;
+              case 'B':
+                c = (( c & 0xff0000) >> 16) | ( c & 0xff00)
+                    | (( c & 0xff) << 16);
+                magiccount += 3;
+                break;
+            }
+            magic = (char *)&c;
+          }
+          else {
+             /* If magic buffer exists: rewind. */
+             magic = magic - (encoding_bytes -1);
+          }
+          address = address - (encoding_bytes -1);
+        }
 	    break;
+      }
 	  putchar (c);
 	}