Fix for the weird [string first] behaviour.

1 files changed, 23 insertions, 19 deletions

diff --git a/generic/tclStringObj.c b/generic/tclStringObj.c
index 7c1d42b..3a35bcf 100644
--- a/generic/tclStringObj.c
+++ b/generic/tclStringObj.c
@@ -3242,40 +3242,44 @@ TclStringFind(
 	return -1;
     }
 
+    /*
+     * Check if we have two strings of single-byte characters. If we have, we
+     * can use strstr() to do the search. Note that we can sometimes have
+     * multibyte characters when the string could be minimally represented
+     * using single byte characters; we can't assume that a mismatch here
+     * means no match.
+     */
+
     lh = Tcl_GetCharLength(haystack);
-    if (haystack->bytes && (lh == haystack->length)) {
-	/* haystack is all single-byte chars */
+    if (haystack->bytes && (lh == haystack->length) && needle->bytes
+		&& (ln == needle->length)) {
+	/*
+	 * Both haystack and needle are all single-byte chars.
+	 */
 
-	if (needle->bytes && (ln == needle->length)) {
-	    /* needle is also all single-byte chars */
-	    char *found = strstr(haystack->bytes + start, needle->bytes);
+	char *found = strstr(haystack->bytes + start, needle->bytes);
 
-	    if (found) {
-		return (found - haystack->bytes);
-	    } else {
-		return -1;
-	    }
+	if (found) {
+	    return (found - haystack->bytes);
 	} else {
-	    /*
-	     * Cannot find substring with a multi-byte char inside
-	     * a string with no multi-byte chars.
-	     */
 	    return -1;
 	}
     } else {
+	/*
+	 * Do the search on the unicode representation for simplicity.
+	 */
+
 	Tcl_UniChar *try, *end, *uh;
 	Tcl_UniChar *un = Tcl_GetUnicodeFromObj(needle, &ln);
 
 	uh = Tcl_GetUnicodeFromObj(haystack, &lh);
 	end = uh + lh;
 
-	try = uh + start;
-	while (try + ln <= end) {
-	    if ((*try == *un)
-		    && (0 == memcmp(try+1, un+1, (ln-1)*sizeof(Tcl_UniChar)))) {
+	for (try = uh + start; try + ln <= end; try++) {
+	    if ((*try == *un) && (0 ==
+		    memcmp(try + 1, un + 1, (ln-1) * sizeof(Tcl_UniChar)))) {
 		return (try - uh);
 	    }
-	    try++;
 	}
 	return -1;
     }