Rename TkUtfToUniChar32 to TkUtfToUniChar2, and various simplifications needing less #ifdef's

8 files changed, 41 insertions, 173 deletions

diff --git a/generic/tkEntry.c b/generic/tkEntry.c
index 5faf4ef..a66cf18 100644
--- a/generic/tkEntry.c
+++ b/generic/tkEntry.c
@@ -1926,6 +1926,7 @@ EntryComputeGeometry(
      */
 
     if (entryPtr->showChar != NULL) {
+	Tcl_UniChar ch;
 	char buf[4];
 	int size;
 
@@ -1935,15 +1936,8 @@ EntryComputeGeometry(
 	 * characters might end up looking like one valid UTF character in the
 	 * resulting string.
 	 */
-#if TCL_UTF_MAX == 4
-	int ch;
-
-	TkUtfToUniChar32(entryPtr->showChar, &ch);
-#else
-	Tcl_UniChar ch;
 
 	Tcl_UtfToUniChar(entryPtr->showChar, &ch);
-#endif
 	size = Tcl_UniCharToUtf(ch, buf);
 
 	entryPtr->numDisplayBytes = entryPtr->numChars * size;
@@ -3420,11 +3414,7 @@ ExpandPercents(
 				 * list element. */
     int number, length;
     register const char *string;
-#if TCL_UTF_MAX == 4
-    int ch;
-#else
     Tcl_UniChar ch;
-#endif
     char numStorage[2*TCL_INTEGER_SPACE];
 
     while (1) {
@@ -3457,11 +3447,7 @@ ExpandPercents(
 
 	before++; /* skip over % */
 	if (*before != '\0') {
-#if TCL_UTF_MAX == 4
-	    before += TkUtfToUniChar32(before, &ch);
-#else
 	    before += Tcl_UtfToUniChar(before, &ch);
-#endif
 	} else {
 	    ch = '%';
 	}
diff --git a/generic/tkFont.c b/generic/tkFont.c
index 3088959..4a45691 100644
--- a/generic/tkFont.c
+++ b/generic/tkFont.c
@@ -497,11 +497,7 @@ Tk_FontObjCmd(
 	const char *s;
 	Tk_Font tkfont;
 	Tcl_Obj *optPtr, *charPtr, *resultPtr;
-#if TCL_UTF_MAX == 4
 	int uniChar = 0;
-#else
-	Tcl_UniChar uniChar = 0;
-#endif
 	const TkFontAttributes *faPtr;
 	TkFontAttributes fa;
 
@@ -566,32 +562,19 @@ Tk_FontObjCmd(
 	 */
 
 	if (charPtr != NULL) {
-#if TCL_UTF_MAX == 4
-	    Tcl_UniChar *ucPtr;
-#endif
+	    const char *string = Tcl_GetString(charPtr);
+	    int len = TkUtfToUniChar2(string, &uniChar);
 
-	    if (Tcl_GetCharLength(charPtr) != 1) {
+	    if (len != charPtr->length) {
 		resultPtr = Tcl_NewStringObj(
 			"expected a single character but got \"", -1);
-		Tcl_AppendLimitedToObj(resultPtr, Tcl_GetString(charPtr),
+		Tcl_AppendLimitedToObj(resultPtr, string,
 			-1, 40, "...");
 		Tcl_AppendToObj(resultPtr, "\"", -1);
 		Tcl_SetObjResult(interp, resultPtr);
 		Tcl_SetErrorCode(interp, "TK", "VALUE", "FONT_SAMPLE", NULL);
 		return TCL_ERROR;
 	    }
-#if TCL_UTF_MAX == 4
-	    ucPtr = Tcl_GetUnicodeFromObj(charPtr, NULL);
-	    uniChar = *ucPtr;
-	    if (((uniChar & 0xFC00) == 0xD800) && (ucPtr[1] != 0x000)) {
-		if ((ucPtr[1] & 0xFC00) == 0xDC00) {
-		    uniChar = ((uniChar & 0x3FF) << 10) + (ucPtr[1] & 0x3FF);
-		    uniChar += 0x10000;
-		}
-	    }
-#else
-	    uniChar = Tcl_GetUniChar(charPtr, 0);
-#endif
 	}
 
 	/*
@@ -1713,11 +1696,7 @@ Tk_PostscriptFontName(
     } else if (strcasecmp(family, "ZapfDingbats") == 0) {
 	family = "ZapfDingbats";
     } else {
-#if TCL_UTF_MAX == 4
 	int ch;
-#else
-	Tcl_UniChar ch;
-#endif
 
 	/*
 	 * Inline, capitalize the first letter of each word, lowercase the
@@ -1735,11 +1714,7 @@ Tk_PostscriptFontName(
 		src++;
 		upper = 1;
 	    }
-#if TCL_UTF_MAX == 4
-	    src += TkUtfToUniChar32(src, &ch);
-#else
-	    src += Tcl_UtfToUniChar(src, &ch);
-#endif
+	    src += TkUtfToUniChar2(src, &ch);
 	    if (upper) {
 		ch = (Tcl_UniChar) Tcl_UniCharToUpper(ch);
 		upper = 0;
@@ -3276,11 +3251,7 @@ Tk_TextLayoutToPostscript(
     int i, j, len;
     const char *p, *glyphname;
     char uindex[5], c, *ps;
-#if TCL_UTF_MAX == 4
     int ch;
-#else
-    Tcl_UniChar ch;
-#endif
 
     Tcl_AppendToObj(psObj, "[(", -1);
     for (i = 0; i < layoutPtr->numChunks; i++, chunkPtr++) {
@@ -3303,11 +3274,7 @@ Tk_TextLayoutToPostscript(
 	     * international postscript fonts.
 	     */
 
-#if TCL_UTF_MAX == 4
-	    p += TkUtfToUniChar32(p, &ch);
-#else
-	    p += Tcl_UtfToUniChar(p, &ch);
-#endif
+	    p += TkUtfToUniChar2(p, &ch);
 	    if ((ch == '(') || (ch == ')') || (ch == '\\') || (ch < 0x20)) {
 		/*
 		 * Tricky point: the "03" is necessary in the sprintf below,
@@ -3333,11 +3300,9 @@ Tk_TextLayoutToPostscript(
 	     * use the full glyph name.
 	     */
 
-#if TCL_UTF_MAX > 3
 	    if (ch > 0xffff) {
 		goto noMapping;
 	    }
-#endif
 	    sprintf(uindex, "%04X", ch);		/* endianness? */
 	    glyphname = Tcl_GetVar2(interp, "::tk::psglyphs", uindex, 0);
 	    if (glyphname) {
diff --git a/generic/tkInt.h b/generic/tkInt.h
index 367ef3a..6d86e08 100644
--- a/generic/tkInt.h
+++ b/generic/tkInt.h
@@ -1232,8 +1232,10 @@ MODULE_SCOPE Status TkParseColor (Display * display,
 MODULE_SCOPE void	TkUnixSetXftClipRegion(TkRegion clipRegion);
 #endif
 
-#if TCL_UTF_MAX == 4
-MODULE_SCOPE int	TkUtfToUniChar32(const char *src, int *chPtr);
+#if TCL_UTF_MAX > 4
+#   define TkUtfToUniChar2 Tcl_UtfToUniChar
+#else
+    MODULE_SCOPE int TkUtfToUniChar2(const char *src, int *chPtr);
 #endif
 
 /*
diff --git a/generic/tkText.c b/generic/tkText.c
index 834e842..dacadbe 100644
--- a/generic/tkText.c
+++ b/generic/tkText.c
@@ -4508,11 +4508,7 @@ TkTextGetTabs(
     Tcl_Obj **objv;
     TkTextTabArray *tabArrayPtr;
     TkTextTab *tabPtr;
-#if TCL_UTF_MAX == 4
     int ch;
-#else
-    Tcl_UniChar ch;
-#endif
     double prevStop, lastStop;
     /*
      * Map these strings to TkTextTabAlign values.
@@ -4619,11 +4615,7 @@ TkTextGetTabs(
 	 * There may be a more efficient way of getting this.
 	 */
 
-#if TCL_UTF_MAX == 4
-	TkUtfToUniChar32(Tcl_GetString(objv[i+1]), &ch);
-#else
-	Tcl_UtfToUniChar(Tcl_GetString(objv[i+1]), &ch);
-#endif
+	TkUtfToUniChar2(Tcl_GetString(objv[i+1]), &ch);
 	if (!Tcl_UniCharIsAlpha(ch)) {
 	    continue;
 	}
diff --git a/generic/tkTextDisp.c b/generic/tkTextDisp.c
index 5faab36..026023e 100644
--- a/generic/tkTextDisp.c
+++ b/generic/tkTextDisp.c
@@ -7581,8 +7581,8 @@ TkTextCharLayoutProc(
 
     if (bytesThatFit < maxBytes) {
 	if ((bytesThatFit == 0) && noCharsYet) {
-	    Tcl_UniChar ch;
-	    int chLen = Tcl_UtfToUniChar(p, &ch);
+	    int ch;
+	    int chLen = TkUtfToUniChar2(p, &ch);
 
 #if TK_LAYOUT_WITH_BASE_CHUNKS
 	    bytesThatFit = CharChunkMeasureChars(chunkPtr, line,
diff --git a/generic/tkTextIndex.c b/generic/tkTextIndex.c
index f64a6d2..b794cdb 100644
--- a/generic/tkTextIndex.c
+++ b/generic/tkTextIndex.c
@@ -2298,13 +2298,9 @@ StartEnd(
 	    int chSize = 1;
 
 	    if (segPtr->typePtr == &tkTextCharType) {
-#if TCL_UTF_MAX == 4
 		int ch;
-		chSize = TkUtfToUniChar32(segPtr->body.chars + offset, &ch);
-#else
-		Tcl_UniChar ch;
-		chSize = Tcl_UtfToUniChar(segPtr->body.chars + offset, &ch);
-#endif
+
+		chSize = TkUtfToUniChar2(segPtr->body.chars + offset, &ch);
 		if (!Tcl_UniCharIsWordChar(ch)) {
 		    break;
 		}
@@ -2347,13 +2343,9 @@ StartEnd(
 	    int chSize = 1;
 
 	    if (segPtr->typePtr == &tkTextCharType) {
-#if TCL_UTF_MAX == 4
+
 		int ch;
-		TkUtfToUniChar32(segPtr->body.chars + offset, &ch);
-#else
-		Tcl_UniChar ch;
-		Tcl_UtfToUniChar(segPtr->body.chars + offset, &ch);
-#endif
+		TkUtfToUniChar2(segPtr->body.chars + offset, &ch);
 		if (!Tcl_UniCharIsWordChar(ch)) {
 		    break;
 		}
diff --git a/generic/tkUtil.c b/generic/tkUtil.c
index fb796fd..a266cb3 100644
--- a/generic/tkUtil.c
+++ b/generic/tkUtil.c
@@ -1193,24 +1193,15 @@ TkSendVirtualEvent(
     Tk_QueueWindowEvent(&event.general, TCL_QUEUE_TAIL);
 }
 
-#if TCL_UTF_MAX == 4
+#if TCL_UTF_MAX <= 4
 /*
  *---------------------------------------------------------------------------
  *
- * TkUtfToUniChar32 --
+ * TkUtfToUniChar2 --
  *
- *	Copied from Tcl_UtfToUniChar but using int instead of Tcl_UniChar!
- *
- *	Extract the Tcl_UniChar represented by the UTF-8 string. Bad UTF-8
- *	sequences are converted to valid Tcl_UniChars and processing
- *	continues. Equivalent to Plan 9 chartorune().
- *
- *	The caller must ensure that the source buffer is long enough that this
- *	routine does not run off the end and dereference non-existent memory
- *	looking for trail bytes. If the source buffer is known to be '\0'
- *	terminated, this cannot happen. Otherwise, the caller should call
- *	Tcl_UtfCharComplete() before calling this routine to ensure that
- *	enough bytes remain in the string.
+ *	Almost the same as Tcl_UtfToUniChar but using int instead of Tcl_UniChar.
+ *	This function is capable of collapsing a upper/lower pair to a single
+ *	unicode character. So, up to 6 bytes (two UTF-8 characters) might be read.
  *
  * Results:
  *	*chPtr is filled with the Tcl_UniChar, and the return value is the
@@ -1223,75 +1214,29 @@ TkSendVirtualEvent(
  */
 
 int
-TkUtfToUniChar32(
+TkUtfToUniChar2(
     const char *src,	/* The UTF-8 string. */
     int *chPtr)		/* Filled with the Tcl_UniChar represented by
 			 * the UTF-8 string. */
 {
-    int byte;
-
-    /*
-     * Unroll 1 to 3 byte UTF-8 sequences, use loop to handle longer ones.
-     */
-
-    byte = *((unsigned char *) src);
-    if (byte < 0xC0) {
-	/*
-	 * Handles properly formed UTF-8 characters between 0x01 and 0x7F.
-	 * Also treats \0 and naked trail bytes 0x80 to 0xBF as valid
-	 * characters representing themselves.
-	 */
-
-	*chPtr = byte;
-	return 1;
-    } else if (byte < 0xE0) {
-	if ((src[1] & 0xC0) == 0x80) {
-	    /*
-	     * Two-byte-character lead-byte followed by a trail-byte.
-	     */
-
-	    *chPtr = ((byte & 0x1F) << 6) | (src[1] & 0x3F);
-	    return 2;
+    Tcl_UniChar uniChar = 0;
+
+    int len = Tcl_UtfToUniChar(src, &uniChar);
+    if ((uniChar & 0xfc00) == 0xd800) {
+	Tcl_UniChar high = uniChar;
+	/* This can only happen when Tcl is compiled with TCL_UTF_MAX=4,
+	 * or when a high surrogate character is detected */
+	int len2 = Tcl_UtfToUniChar(src+len, &uniChar);
+	if ((uniChar & 0xfc00) == 0xdc00) {
+	    *chPtr = ((high & 0x3ff) << 10) | (uniChar & 0x3ff) | 0x10000;
+	    len += len2;
+	} else {
+	    *chPtr = high;
 	}
-
-	/*
-	 * A two-byte-character lead-byte not followed by trail-byte
-	 * represents itself.
-	 */
-    } else if (byte < 0xF0) {
-	if (((src[1] & 0xC0) == 0x80) && ((src[2] & 0xC0) == 0x80)) {
-	    /*
-	     * Three-byte-character lead byte followed by two trail bytes.
-	     */
-
-	    *chPtr = ((byte & 0x0F) << 12)
-		    | ((src[1] & 0x3F) << 6) | (src[2] & 0x3F);
-	    return 3;
-	}
-
-	/*
-	 * A three-byte-character lead-byte not followed by two trail-bytes
-	 * represents itself.
-	 */
-    } 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 = ((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.
-	 */
+    } else {
+	*chPtr = uniChar;
     }
-
-    *chPtr = byte;
-    return 1;
+    return len;
 }
 #endif
 /*
diff --git a/generic/ttk/ttkEntry.c b/generic/ttk/ttkEntry.c
index 63ebc5f..d80e1fd 100644
--- a/generic/ttk/ttkEntry.c
+++ b/generic/ttk/ttkEntry.c
@@ -282,16 +282,10 @@ static char *EntryDisplayString(const char *showChar, int numChars)
 {
     char *displayString, *p;
     int size;
-    char buf[4];
-#if TCL_UTF_MAX == 4
-    int ch;
-
-    TkUtfToUniChar32(showChar, &ch);
-#else
     Tcl_UniChar ch;
+    char buf[4];
 
     Tcl_UtfToUniChar(showChar, &ch);
-#endif
     size = Tcl_UniCharToUtf(ch, buf);
     p = displayString = ckalloc(numChars * size + 1);
 
@@ -412,11 +406,7 @@ ExpandPercents(
     int number, length;
     const char *string;
     int stringLength;
-#if TCL_UTF_MAX == 4
-    int ch;
-#else
     Tcl_UniChar ch;
-#endif
     char numStorage[2*TCL_INTEGER_SPACE];
 
     while (*template) {
@@ -440,11 +430,7 @@ ExpandPercents(
 	 */
 	++template; /* skip over % */
 	if (*template != '\0') {
-#if TCL_UTF_MAX == 4
-	    template += TkUtfToUniChar32(template, &ch);
-#else
 	    template += Tcl_UtfToUniChar(template, &ch);
-#endif
 	} else {
 	    ch = '%';
 	}