[Patch] multibyte encodings in strings

[m***@exitno.de]

bump

Should there be a problem with the patch format or the way of submission please let me know.

Mathias

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)
+ {
+ break;
+ c = c & 0xff;
+ magiccount += 1;
+ break;
+ c = c >> 8;
+ magiccount += (encoding_bytes -1);
+ break;
+ 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)
+ {
+ break;
+ c = c & 0xff;
+ magiccount += 1;
+ break;
+ c = c >> 8;
+ magiccount += (encoding_bytes -1);
+ break;
+ 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);
}