Don't ever allow UTF-8 sequences of more than 4 characters to be generated or parsed, even when TCL_UTF_MAX>4: According to current Unicode standard, a byte string of >4 characters can never form a single UTF-8 character.

And a few minor micro-optimizations related to UTF-8 handling.

6 files changed, 33 insertions, 53 deletions

diff --git a/generic/tclExecute.c b/generic/tclExecute.c
index a2a465a..704e494 100644
--- a/generic/tclExecute.c
+++ b/generic/tclExecute.c
@@ -3173,7 +3173,7 @@ TEBCresume(
 	    Tcl_Obj *copyPtr = Tcl_NewListObj(objc - opnd + 1, NULL);
 
 	    Tcl_ListObjAppendElement(NULL, copyPtr, objPtr);
-	    Tcl_ListObjReplace(NULL, copyPtr, LIST_MAX, 0, 
+	    Tcl_ListObjReplace(NULL, copyPtr, LIST_MAX, 0,
 		    objc - opnd, objv + opnd);
 	    Tcl_DecrRefCount(objPtr);
 	    objPtr = copyPtr;
diff --git a/generic/tclUtf.c b/generic/tclUtf.c
index 6c4cb7f..b33bf6a 100644
--- a/generic/tclUtf.c
+++ b/generic/tclUtf.c
@@ -73,16 +73,7 @@ static const unsigned char totalBytes[256] = {
 #else
     1,1,1,1,1,1,1,1,
 #endif
-#if TCL_UTF_MAX > 4
-    5,5,5,5,
-#else
-    1,1,1,1,
-#endif
-#if TCL_UTF_MAX > 5
-    6,6,6,6
-#else
-    1,1,1,1
-#endif
+    1,1,1,1,1,1,1,1
 };
 
 /*
@@ -105,14 +96,14 @@ int
 TclUtfCount(
     int ch)			/* The Tcl_UniChar whose size is returned. */
 {
-    if ((ch > 0) && (ch < UNICODE_SELF)) {
+    if ((unsigned)(ch - 1) < (UNICODE_SELF - 1)) {
 	return 1;
     }
     if (ch <= 0x7FF) {
 	return 2;
     }
 #if TCL_UTF_MAX > 3
-    if ((ch > 0xFFFF) && (ch <= 0x10FFFF)) {
+    if (((unsigned)(ch - 0x10000) <= 0xfffff)) {
 	return 4;
     }
 #endif
@@ -146,7 +137,7 @@ Tcl_UniCharToUtf(
 				 * large enough to hold the UTF-8 character
 				 * (at most TCL_UTF_MAX bytes). */
 {
-    if ((ch > 0) && (ch < UNICODE_SELF)) {
+    if ((unsigned)(ch - 1) < (UNICODE_SELF - 1)) {
 	buf[0] = (char) ch;
 	return 1;
     }
@@ -174,11 +165,7 @@ Tcl_UniCharToUtf(
 		}
 	    }
 #endif
-	three:
-	    buf[2] = (char) ((ch | 0x80) & 0xBF);
-	    buf[1] = (char) (((ch >> 6) | 0x80) & 0xBF);
-	    buf[0] = (char) ((ch >> 12) | 0xE0);
-	    return 3;
+	    goto three;
 	}
 
 #if TCL_UTF_MAX > 3
@@ -193,7 +180,11 @@ Tcl_UniCharToUtf(
     }
 
     ch = 0xFFFD;
-    goto three;
+three:
+    buf[2] = (char) ((ch | 0x80) & 0xBF);
+    buf[1] = (char) (((ch >> 6) | 0x80) & 0xBF);
+    buf[0] = (char) ((ch >> 12) | 0xE0);
+    return 3;
 }
 
 /*
@@ -308,9 +299,6 @@ Tcl_UtfToUniChar(
 	 * A two-byte-character lead-byte not followed by trail-byte
 	 * represents itself.
 	 */
-
-	*chPtr = (Tcl_UniChar) byte;
-	return 1;
     } else if (byte < 0xF0) {
 	if (((src[1] & 0xC0) == 0x80) && ((src[2] & 0xC0) == 0x80)) {
 	    /*
@@ -326,31 +314,23 @@ Tcl_UtfToUniChar(
 	 * A three-byte-character lead-byte not followed by two trail-bytes
 	 * represents itself.
 	 */
-
-	*chPtr = (Tcl_UniChar) byte;
-	return 1;
     }
 #if TCL_UTF_MAX > 3
-    {
-	int ch, total, trail;
-
-	total = totalBytes[byte];
-	trail = total - 1;
-	if (trail > 0) {
-	    ch = byte & (0x3F >> trail);
-	    do {
-		src++;
-		if ((*src & 0xC0) != 0x80) {
-		    *chPtr = byte;
-		    return 1;
-		}
-		ch <<= 6;
-		ch |= (*src & 0x3F);
-		trail--;
-	    } while (trail > 0);
-	    *chPtr = ch;
-	    return total;
+    else if (byte < 0xF8) {
+	if (((src[1] & 0xC0) == 0x80) && ((src[2] & 0xC0) == 0x80) && ((src[3] & 0xC0) == 0x80)) {
+	    /*
+	     * Four-byte-character lead byte followed by three trail bytes.
+	     */
+
+	    *chPtr = (Tcl_UniChar) (((byte & 0x0E) << 18) | ((src[1] & 0x3F) << 12)
+		    | ((src[2] & 0x3F) << 6) | (src[3] & 0x3F));
+	    return 4;
 	}
+
+	/*
+	 * A three-byte-character lead-byte not followed by two trail-bytes
+	 * represents itself.
+	 */
     }
 #endif
 
diff --git a/generic/tclVar.c b/generic/tclVar.c
index 55eb436..20bc208 100644
--- a/generic/tclVar.c
+++ b/generic/tclVar.c
@@ -2983,7 +2983,7 @@ ArrayAnyMoreCmd(
     if (varPtr == NULL) {
 	return TCL_ERROR;
     }
-    
+
     /*
      * Get the search.
      */
@@ -3056,7 +3056,7 @@ ArrayNextElementCmd(
     varPtr = VerifyArray(interp, varNameObj);
     if (varPtr == NULL) {
 	return TCL_ERROR;
-    } 
+    }
 
     /*
      * Get the search.
diff --git a/win/tclWinFile.c b/win/tclWinFile.c
index 3a856a1..9458933 100755
--- a/win/tclWinFile.c
+++ b/win/tclWinFile.c
@@ -3166,8 +3166,8 @@ TclWinFileOwned(
            case we are in all likelihood not the owner */
         return 0;
     }
-        
-    /* 
+
+    /*
      * Getting the current process SID is a multi-step process.
      * We make the assumption that if a call fails, this process is
      * so underprivileged it could not possibly own anything. Normally
@@ -3191,10 +3191,10 @@ TclWinFileOwned(
         LocalFree(secd);            /* Also frees ownerSid */
     if (buf)
         ckfree(buf);
-    
+
     return (owned != 0);        /* Convert non-0 to 1 */
 }
-    
+
 /*
  * Local Variables:
  * mode: c
diff --git a/win/tclWinPipe.c b/win/tclWinPipe.c
index 382addd..4666deb 100644
--- a/win/tclWinPipe.c
+++ b/win/tclWinPipe.c
@@ -1337,7 +1337,7 @@ ApplicationType(
 	Tcl_DStringFree(&ds);
 
 	ext = strrchr(fullName, '.');
-	if ((ext != NULL) && 
+	if ((ext != NULL) &&
             (strcasecmp(ext, ".cmd") == 0 || strcasecmp(ext, ".bat") == 0)) {
 	    applType = APPL_DOS;
 	    break;
diff --git a/win/tclWinPort.h b/win/tclWinPort.h
index b486466..159a708 100644
--- a/win/tclWinPort.h
+++ b/win/tclWinPort.h
@@ -360,7 +360,7 @@ typedef DWORD_PTR * PDWORD_PTR;
 #   define S_IFLNK        0120000  /* Symbolic Link */
 #endif
 
-/* 
+/*
  * Windows compilers do not define S_IFBLK. However, Tcl uses it in
  * GetTypeFromMode to identify blockSpecial devices based on the
  * value in the statsbuf st_mode field. We have no other way to pass this