Fix [6811a00819]: lsearch performance degradation on Tcl 8.6.11 release (thanks, Sergey!)

13 files changed, 48 insertions, 61 deletions

diff --git a/generic/tclBinary.c b/generic/tclBinary.c
index 916ba4c..9dd9edf 100644
--- a/generic/tclBinary.c
+++ b/generic/tclBinary.c
@@ -512,7 +512,7 @@ MakeByteArray(
 
     for (; src < srcEnd && dst < dstEnd; ) {
 	int ch;
-	int count = Tcl_UtfToUniChar(src, &ch);
+	int count = TclUtfToUniChar(src, &ch);
 
 	if (ch > 255) {
 	    proper = 0;
@@ -2563,7 +2563,7 @@ BinaryDecodeHex(
     if (pure) {
 	ucs4 = c;
     } else {
-	Tcl_UtfToUniChar((const char *)(data - 1), &ucs4);
+	TclUtfToUniChar((const char *)(data - 1), &ucs4);
     }
     TclDecrRefCount(resultObj);
     Tcl_SetObjResult(interp, Tcl_ObjPrintf(
@@ -3028,7 +3028,7 @@ BinaryDecodeUu(
     if (pure) {
 	ucs4 = c;
     } else {
-	Tcl_UtfToUniChar((const char *)(data - 1), &ucs4);
+	TclUtfToUniChar((const char *)(data - 1), &ucs4);
     }
     Tcl_SetObjResult(interp, Tcl_ObjPrintf(
 	    "invalid uuencode character \"%c\" (U+%06X) at position %"
@@ -3203,7 +3203,7 @@ BinaryDecode64(
 	 * of a multi-byte character. */
 
 	/* Safe because we know data is NUL-terminated */
-	Tcl_UtfToUniChar((const char *)(data - 1), &ucs4);
+	TclUtfToUniChar((const char *)(data - 1), &ucs4);
     }
 
     Tcl_SetObjResult(interp, Tcl_ObjPrintf(
diff --git a/generic/tclCmdIL.c b/generic/tclCmdIL.c
index 4474513..c79fd86 100644
--- a/generic/tclCmdIL.c
+++ b/generic/tclCmdIL.c
@@ -5432,8 +5432,8 @@ DictionaryCompare(
 	 */
 
 	if ((*left != '\0') && (*right != '\0')) {
-	    left += Tcl_UtfToUniChar(left, &uniLeft);
-	    right += Tcl_UtfToUniChar(right, &uniRight);
+	    left += TclUtfToUniChar(left, &uniLeft);
+	    right += TclUtfToUniChar(right, &uniRight);
 
 	    /*
 	     * Convert both chars to lower for the comparison, because
diff --git a/generic/tclCmdMZ.c b/generic/tclCmdMZ.c
index 56d4cca..a00762e 100644
--- a/generic/tclCmdMZ.c
+++ b/generic/tclCmdMZ.c
@@ -1220,7 +1220,7 @@ Tcl_SplitObjCmd(
 	Tcl_InitHashTable(&charReuseTable, TCL_ONE_WORD_KEYS);
 
 	for ( ; stringPtr < end; stringPtr += len) {
-	    len = Tcl_UtfToUniChar(stringPtr, &ch);
+	    len = TclUtfToUniChar(stringPtr, &ch);
 	    hPtr = Tcl_CreateHashEntry(&charReuseTable, INT2PTR(ch), &isNew);
 	    if (isNew) {
 		TclNewStringObj(objPtr, stringPtr, len);
@@ -1266,9 +1266,9 @@ Tcl_SplitObjCmd(
 	splitEnd = splitChars + splitCharLen;
 
 	for (element = stringPtr; stringPtr < end; stringPtr += len) {
-	    len = Tcl_UtfToUniChar(stringPtr, &ch);
+	    len = TclUtfToUniChar(stringPtr, &ch);
 	    for (p = splitChars; p < splitEnd; p += splitLen) {
-		splitLen = Tcl_UtfToUniChar(p, &splitChar);
+		splitLen = TclUtfToUniChar(p, &splitChar);
 		if (ch == splitChar) {
 		    TclNewStringObj(objPtr, element, stringPtr - element);
 		    Tcl_ListObjAppendElement(NULL, listPtr, objPtr);
@@ -1891,7 +1891,7 @@ StringIsCmd(
 	for (; string1 < end; string1 += length2, failat++) {
 	    int ucs4;
 
-	    length2 = Tcl_UtfToUniChar(string1, &ucs4);
+	    length2 = TclUtfToUniChar(string1, &ucs4);
 	    if (!chcomp(ucs4)) {
 		result = 0;
 		break;
diff --git a/generic/tclCompExpr.c b/generic/tclCompExpr.c
index 85f475e..e7ea456 100644
--- a/generic/tclCompExpr.c
+++ b/generic/tclCompExpr.c
@@ -2149,13 +2149,13 @@ ParseLexeme(
     if (!TclIsBareword(*start) || *start == '_') {
 	Tcl_Size scanned;
 	if (Tcl_UtfCharComplete(start, numBytes)) {
-	    scanned = Tcl_UtfToUniChar(start, &ch);
+	    scanned = TclUtfToUniChar(start, &ch);
 	} else {
 	    char utfBytes[8];
 
 	    memcpy(utfBytes, start, numBytes);
 	    utfBytes[numBytes] = '\0';
-	    scanned = Tcl_UtfToUniChar(utfBytes, &ch);
+	    scanned = TclUtfToUniChar(utfBytes, &ch);
 	}
 	*lexemePtr = INVALID;
 	Tcl_DecrRefCount(literal);
diff --git a/generic/tclDisassemble.c b/generic/tclDisassemble.c
index 1a41562..68b3b6c 100644
--- a/generic/tclDisassemble.c
+++ b/generic/tclDisassemble.c
@@ -876,7 +876,7 @@ PrintSourceToObj(
     for (;  (*p != '\0') && (i < maxChars);  p+=len) {
 	int ucs4;
 
-	len = Tcl_UtfToUniChar(p, &ucs4);
+	len = TclUtfToUniChar(p, &ucs4);
 	switch (ucs4) {
 	case '"':
 	    Tcl_AppendToObj(appendObj, "\\\"", -1);
diff --git a/generic/tclEncoding.c b/generic/tclEncoding.c
index 074c58e..312ce85 100644
--- a/generic/tclEncoding.c
+++ b/generic/tclEncoding.c
@@ -1581,7 +1581,7 @@ Tcl_UtfToExternalDStringEx(
 		    int ucs4;
 		    char buf[TCL_INTEGER_SPACE];
 
-		    Tcl_UtfToUniChar(&srcStart[nBytesProcessed], &ucs4);
+		    TclUtfToUniChar(&srcStart[nBytesProcessed], &ucs4);
 		    snprintf(buf, sizeof(buf), "%" TCL_SIZE_MODIFIER "d", nBytesProcessed);
 		    Tcl_SetObjResult(interp, Tcl_ObjPrintf(
 			    "unexpected character at index %" TCL_SIZE_MODIFIER
@@ -2551,11 +2551,11 @@ UtfToUtfProc(
 		/* TCL_ENCODING_PROFILE_TCL8 */
 		char chbuf[2];
 		chbuf[0] = UCHAR(*src++); chbuf[1] = 0;
-		Tcl_UtfToUniChar(chbuf, &ch);
+		TclUtfToUniChar(chbuf, &ch);
 	    }
 	    dst += Tcl_UniCharToUtf(ch, dst);
 	} else {
-	    size_t len = Tcl_UtfToUniChar(src, &ch);
+	    size_t len = TclUtfToUniChar(src, &ch);
 	    if (flags & ENCODING_INPUT) {
 		if (((len < 2) && (ch != 0)) || ((ch > 0xFFFF) && !(flags & ENCODING_UTF))) {
 		    if (PROFILE_STRICT(profile)) {
@@ -2799,7 +2799,7 @@ UtfToUtf32Proc(
 	    result = TCL_CONVERT_NOSPACE;
 	    break;
 	}
-	len = Tcl_UtfToUniChar(src, &ch);
+	len = TclUtfToUniChar(src, &ch);
 	if (SURROGATE(ch)) {
 	    if (PROFILE_STRICT(flags)) {
 		result = TCL_CONVERT_UNKNOWN;
@@ -3076,7 +3076,7 @@ UtfToUtf16Proc(
 	    result = TCL_CONVERT_NOSPACE;
 	    break;
 	}
-	len = Tcl_UtfToUniChar(src, &ch);
+	len = TclUtfToUniChar(src, &ch);
 	if (SURROGATE(ch)) {
 	    if (PROFILE_STRICT(flags)) {
 		result = TCL_CONVERT_UNKNOWN;
@@ -3323,7 +3323,7 @@ TableToUtfProc(
 	    } else {
 		char chbuf[2];
 		chbuf[0] = byte; chbuf[1] = 0;
-		Tcl_UtfToUniChar(chbuf, &ch);
+		TclUtfToUniChar(chbuf, &ch);
 	    }
 	}
 
diff --git a/generic/tclInt.h b/generic/tclInt.h
index ada17b3..50c7add 100644
--- a/generic/tclInt.h
+++ b/generic/tclInt.h
@@ -4565,17 +4565,10 @@ MODULE_SCOPE const TclFileAttrProcs	tclpFileAttrProcs[];
  *----------------------------------------------------------------
  */
 
-#if TCL_UTF_MAX > 3
 #define TclUtfToUniChar(str, chPtr) \
 	(((UCHAR(*(str))) < 0x80) ?		\
 	    ((*(chPtr) = UCHAR(*(str))), 1)	\
 	    : Tcl_UtfToUniChar(str, chPtr))
-#else
-#define TclUtfToUniChar(str, chPtr) \
-	(((UCHAR(*(str))) < 0x80) ?		\
-	    ((*(chPtr) = UCHAR(*(str))), 1)	\
-	    : Tcl_UtfToChar16(str, chPtr))
-#endif
 
 /*
  *----------------------------------------------------------------
diff --git a/generic/tclParse.c b/generic/tclParse.c
index 84200ac..92ec1c2 100644
--- a/generic/tclParse.c
+++ b/generic/tclParse.c
@@ -926,13 +926,13 @@ TclParseBackslash(
 	 */
 
 	if (Tcl_UtfCharComplete(p, numBytes - 1)) {
-	    count = Tcl_UtfToUniChar(p, &unichar) + 1;	/* +1 for '\' */
+	    count = TclUtfToUniChar(p, &unichar) + 1;	/* +1 for '\' */
 	} else {
 	    char utfBytes[8];
 
 	    memcpy(utfBytes, p, numBytes - 1);
 	    utfBytes[numBytes - 1] = '\0';
-	    count = Tcl_UtfToUniChar(utfBytes, &unichar) + 1;
+	    count = TclUtfToUniChar(utfBytes, &unichar) + 1;
 	}
 	result = unichar;
 	break;
diff --git a/generic/tclScan.c b/generic/tclScan.c
index 195153b..71af1f2 100644
--- a/generic/tclScan.c
+++ b/generic/tclScan.c
@@ -902,7 +902,7 @@ Tcl_ScanObjCmd(
 	     * Scan a single Unicode character.
 	     */
 
-	    offset = Tcl_UtfToUniChar(string, &i);
+	    offset = TclUtfToUniChar(string, &i);
 	    string += offset;
 	    if (!(flags & SCAN_SUPPRESS)) {
 		TclNewIntObj(objPtr, i);
diff --git a/generic/tclStringObj.c b/generic/tclStringObj.c
index d8b96f7..dba36a6 100644
--- a/generic/tclStringObj.c
+++ b/generic/tclStringObj.c
@@ -618,8 +618,6 @@ TclGetUniChar(
 	return -1;
     }
     const char *begin = TclUtfAtIndex(objPtr->bytes, index);
-#undef Tcl_UtfToUniChar
-    Tcl_UtfToUniChar(begin, &ch);
     return ch;
 }
 
@@ -4133,7 +4131,7 @@ TclStringReverse(
 		 * skip calling Tcl_UtfCharComplete() here.
 		 */
 
-		int bytesInChar = Tcl_UtfToUniChar(from, &chw);
+		int bytesInChar = TclUtfToUniChar(from, &chw);
 
 		ReverseBytes((unsigned char *)to, (unsigned char *)from,
 			bytesInChar);
diff --git a/generic/tclStubInit.c b/generic/tclStubInit.c
index 6b1dea2..e35a401 100644
--- a/generic/tclStubInit.c
+++ b/generic/tclStubInit.c
@@ -64,10 +64,6 @@
 #undef Tcl_ParseArgsObjv
 #undef TclStaticLibrary
 #define TclStaticLibrary Tcl_StaticLibrary
-#undef Tcl_UniCharToUtfDString
-#undef Tcl_UtfToUniCharDString
-#undef Tcl_UtfToUniChar
-#undef Tcl_UniCharLen
 #undef TclObjInterpProc
 #if !defined(_WIN32) && !defined(__CYGWIN__)
 # undef Tcl_WinConvertError
diff --git a/generic/tclUtf.c b/generic/tclUtf.c
index 3dede09..e107081 100644
--- a/generic/tclUtf.c
+++ b/generic/tclUtf.c
@@ -678,11 +678,11 @@ Tcl_UtfToUniCharDString(
     endPtr = src + length;
     optPtr = endPtr - 4;
     while (p <= optPtr) {
-	p += Tcl_UtfToUniChar(p, &ch);
+	p += TclUtfToUniChar(p, &ch);
 	*w++ = ch;
     }
     while ((p < endPtr) && Tcl_UtfCharComplete(p, endPtr-p)) {
-	p += Tcl_UtfToUniChar(p, &ch);
+	p += TclUtfToUniChar(p, &ch);
 	*w++ = ch;
     }
     while (p < endPtr) {
@@ -928,7 +928,7 @@ Tcl_UtfFindFirst(
     int ch)			/* The Unicode character to search for. */
 {
     while (1) {
-	int find, len = Tcl_UtfToUniChar(src, &find);
+	int find, len = TclUtfToUniChar(src, &find);
 
 	if (find == ch) {
 	    return src;
@@ -967,7 +967,7 @@ Tcl_UtfFindLast(
     const char *last = NULL;
 
     while (1) {
-	int find, len = Tcl_UtfToUniChar(src, &find);
+	int find, len = TclUtfToUniChar(src, &find);
 
 	if (find == ch) {
 	    last = src;
@@ -1190,7 +1190,7 @@ Tcl_UniCharAtIndex(
 	i = TclUtfToUniChar(src, &ch);
 	src += i;
     }
-    Tcl_UtfToUniChar(src, &i);
+    TclUtfToUniChar(src, &i);
     return i;
 }
 
@@ -1219,7 +1219,7 @@ Tcl_UtfAtIndex(
     Tcl_UniChar ch = 0;
 
     while (index-- > 0) {
-	src += Tcl_UtfToUniChar(src, &ch);
+	src += TclUtfToUniChar(src, &ch);
     }
     return src;
 }
@@ -1329,7 +1329,7 @@ Tcl_UtfToUpper(
 
     src = dst = str;
     while (*src) {
-	len = Tcl_UtfToUniChar(src, &ch);
+	len = TclUtfToUniChar(src, &ch);
 	upChar = Tcl_UniCharToUpper(ch);
 
 	/*
@@ -1382,7 +1382,7 @@ Tcl_UtfToLower(
 
     src = dst = str;
     while (*src) {
-	len = Tcl_UtfToUniChar(src, &ch);
+	len = TclUtfToUniChar(src, &ch);
 	lowChar = Tcl_UniCharToLower(ch);
 
 	/*
@@ -1438,7 +1438,7 @@ Tcl_UtfToTitle(
     src = dst = str;
 
     if (*src) {
-	len = Tcl_UtfToUniChar(src, &ch);
+	len = TclUtfToUniChar(src, &ch);
 	titleChar = Tcl_UniCharToTitle(ch);
 
 	if (len < TclUtfCount(titleChar)) {
@@ -1450,7 +1450,7 @@ Tcl_UtfToTitle(
 	src += len;
     }
     while (*src) {
-	len = Tcl_UtfToUniChar(src, &ch);
+	len = TclUtfToUniChar(src, &ch);
 	lowChar = ch;
 	/* Special exception for Georgian Asomtavruli chars, no titlecase. */
 	if ((unsigned)(lowChar - 0x1C90) >= 0x30) {
diff --git a/generic/tclUtil.c b/generic/tclUtil.c
index efe7ea9..e15a18f 100644
--- a/generic/tclUtil.c
+++ b/generic/tclUtil.c
@@ -1672,7 +1672,7 @@ TclTrimRight(
 	pp = Tcl_UtfPrev(p, bytes);
 	do {
 	    pp += pInc;
- 	    pInc = Tcl_UtfToUniChar(pp, &ch1);
+ 	    pInc = TclUtfToUniChar(pp, &ch1);
 	} while (pp + pInc < p);
 
 	/*
@@ -1680,7 +1680,7 @@ TclTrimRight(
 	 */
 
 	do {
-	    pInc = Tcl_UtfToUniChar(q, &ch2);
+	    pInc = TclUtfToUniChar(q, &ch2);
 
 	    if (ch1 == ch2) {
 		break;
@@ -1745,7 +1745,7 @@ TclTrimLeft(
      */
 
     do {
-	Tcl_Size pInc = Tcl_UtfToUniChar(p, &ch1);
+	Tcl_Size pInc = TclUtfToUniChar(p, &ch1);
 	const char *q = trim;
 	Tcl_Size bytesLeft = numTrim;
 
@@ -1754,7 +1754,7 @@ TclTrimLeft(
 	 */
 
 	do {
-	    Tcl_Size qInc = Tcl_UtfToUniChar(q, &ch2);
+	    Tcl_Size qInc = TclUtfToUniChar(q, &ch2);
 
 	    if (ch1 == ch2) {
 		break;
@@ -1821,7 +1821,7 @@ TclTrim(
 	if (numBytes > 0) {
 	    int ch;
 	    const char *first = bytes + trimLeft;
-	    bytes += Tcl_UtfToUniChar(first, &ch);
+	    bytes += TclUtfToUniChar(first, &ch);
 	    numBytes -= (bytes - first);
 
 	    if (numBytes > 0) {
@@ -2151,7 +2151,7 @@ Tcl_StringCaseMatch(
 		ch2 = (int)
 			(nocase ? tolower(UCHAR(*pattern)) : UCHAR(*pattern));
 	    } else {
-		Tcl_UtfToUniChar(pattern, &ch2);
+		TclUtfToUniChar(pattern, &ch2);
 		if (nocase) {
 		    ch2 = Tcl_UniCharToLower(ch2);
 		}
@@ -2167,7 +2167,7 @@ Tcl_StringCaseMatch(
 		if ((p != '[') && (p != '?') && (p != '\\')) {
 		    if (nocase) {
 			while (*str) {
-			    charLen = Tcl_UtfToUniChar(str, &ch1);
+			    charLen = TclUtfToUniChar(str, &ch1);
 			    if (ch2==ch1 || ch2==Tcl_UniCharToLower(ch1)) {
 				break;
 			    }
@@ -2181,7 +2181,7 @@ Tcl_StringCaseMatch(
 			 */
 
 			while (*str) {
-			    charLen = Tcl_UtfToUniChar(str, &ch1);
+			    charLen = TclUtfToUniChar(str, &ch1);
 			    if (ch2 == ch1) {
 				break;
 			    }
@@ -2195,7 +2195,7 @@ Tcl_StringCaseMatch(
 		if (*str == '\0') {
 		    return 0;
 		}
-		str += Tcl_UtfToUniChar(str, &ch1);
+		str += TclUtfToUniChar(str, &ch1);
 	    }
 	}
 
@@ -2206,7 +2206,7 @@ Tcl_StringCaseMatch(
 
 	if (p == '?') {
 	    pattern++;
-	    str += Tcl_UtfToUniChar(str, &ch1);
+	    str += TclUtfToUniChar(str, &ch1);
 	    continue;
 	}
 
@@ -2225,7 +2225,7 @@ Tcl_StringCaseMatch(
 			(nocase ? tolower(UCHAR(*str)) : UCHAR(*str));
 		str++;
 	    } else {
-		str += Tcl_UtfToUniChar(str, &ch1);
+		str += TclUtfToUniChar(str, &ch1);
 		if (nocase) {
 		    ch1 = Tcl_UniCharToLower(ch1);
 		}
@@ -2239,7 +2239,7 @@ Tcl_StringCaseMatch(
 			    ? tolower(UCHAR(*pattern)) : UCHAR(*pattern));
 		    pattern++;
 		} else {
-		    pattern += Tcl_UtfToUniChar(pattern, &startChar);
+		    pattern += TclUtfToUniChar(pattern, &startChar);
 		    if (nocase) {
 			startChar = Tcl_UniCharToLower(startChar);
 		    }
@@ -2254,7 +2254,7 @@ Tcl_StringCaseMatch(
 				? tolower(UCHAR(*pattern)) : UCHAR(*pattern));
 			pattern++;
 		    } else {
-			pattern += Tcl_UtfToUniChar(pattern, &endChar);
+			pattern += TclUtfToUniChar(pattern, &endChar);
 			if (nocase) {
 			    endChar = Tcl_UniCharToLower(endChar);
 			}
@@ -2302,8 +2302,8 @@ Tcl_StringCaseMatch(
 	 * each string match.
 	 */
 
-	str += Tcl_UtfToUniChar(str, &ch1);
-	pattern += Tcl_UtfToUniChar(pattern, &ch2);
+	str += TclUtfToUniChar(str, &ch1);
+	pattern += TclUtfToUniChar(pattern, &ch2);
 	if (nocase) {
 	    if (Tcl_UniCharToLower(ch1) != Tcl_UniCharToLower(ch2)) {
 		return 0;