merge novem

1 files changed, 28 insertions, 22 deletions

diff --git a/generic/tclStringObj.c b/generic/tclStringObj.c
index d607c1d..336b2be 100644
--- a/generic/tclStringObj.c
+++ b/generic/tclStringObj.c
@@ -3079,6 +3079,7 @@ TclStringCatObjv(
 	    /* Ugly interface! No scheme to init array size. */
 	    objResultPtr = Tcl_NewUnicodeObj(&ch, 0);	/* PANIC? */
 	    if (0 == Tcl_AttemptSetObjLength(objResultPtr, length)) {
+		Tcl_DecrRefCount(objResultPtr);
 		if (interp) {
 		    Tcl_SetObjResult(interp, Tcl_ObjPrintf(
 		    	"concatenation failed: unable to alloc %"
@@ -3126,6 +3127,7 @@ TclStringCatObjv(
 	} else {
 	    objResultPtr = Tcl_NewObj();	/* PANIC? */
 	    if (0 == Tcl_AttemptSetObjLength(objResultPtr, length)) {
+		Tcl_DecrRefCount(objResultPtr);
 		if (interp) {
 		    Tcl_SetObjResult(interp, Tcl_ObjPrintf(
 		    	"concatenation failed: unable to alloc %u bytes",
@@ -3217,40 +3219,44 @@ 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 && ((size_t)lh == haystack->length)) {
-	/* haystack is all single-byte chars */
+    if (haystack->bytes && ((size_t)lh == haystack->length) && needle->bytes
+		&& ((size_t)ln == needle->length)) {
+	/*
+	 * Both haystack and needle are all single-byte chars.
+	 */
 
-	if (needle->bytes && ((size_t)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;
     }
@@ -3433,7 +3439,6 @@ TclStringObjReverse(
 	     * Tcl_SetObjLength into growing the unicode rep buffer.
 	     */
 
-	    ch = 0;
 	    objPtr = Tcl_NewUnicodeObj(&ch, 1);
 	    Tcl_SetObjLength(objPtr, stringPtr->numChars);
 	    to = Tcl_GetUnicode(objPtr);
@@ -3536,7 +3541,7 @@ ExtendUnicodeRepWithString(
 {
     String *stringPtr = GET_STRING(objPtr);
     size_t needed, numOrigChars = 0;
-    Tcl_UniChar *dst;
+    Tcl_UniChar *dst, unichar = 0;
 
     if (stringPtr->hasUnicode) {
 	numOrigChars = stringPtr->numChars;
@@ -3559,7 +3564,8 @@ ExtendUnicodeRepWithString(
 	numAppendChars = 0;
     }
     for (dst=stringPtr->unicode + numOrigChars; numAppendChars-- > 0; dst++) {
-	bytes += TclUtfToUniChar(bytes, dst);
+	bytes += TclUtfToUniChar(bytes, &unichar);
+	*dst = unichar;
     }
     *dst = 0;
 }