Better implementation of [a179564826], not depending on surrogate handling by Tcl any more

6 files changed, 14 insertions, 16 deletions

diff --git a/generic/tkEntry.c b/generic/tkEntry.c
index 00c6fc4..9c53769 100644
--- a/generic/tkEntry.c
+++ b/generic/tkEntry.c
@@ -1923,7 +1923,7 @@ EntryComputeGeometry(
 
     if (entryPtr->showChar != NULL) {
 	int ch;
-	char buf[4];
+	char buf[6];
 	int size;
 
 	/*
diff --git a/generic/tkUtil.c b/generic/tkUtil.c
index cac3dc8..7a15410 100644
--- a/generic/tkUtil.c
+++ b/generic/tkUtil.c
@@ -1240,8 +1240,9 @@ TkUtfToUniChar(
  *
  * TkUniCharToUtf --
  *
- *	Almost the same as Tcl_UniCharToUtf but producing 4-byte UTF-8
- *	sequences even when TCL_UTF_MAX==3. So, up to 4 bytes might be produced.
+ *	Almost the same as Tcl_UniCharToUtf but producing 2 x 3-byte UTF-8
+ *	sequences for out-of-bmp characters when TCL_UTF_MAX==3.
+ *	So, up to 6 bytes might be produced.
  *
  * Results:
  *	*buf is filled with the UTF-8 string, and the return value is the
@@ -1255,16 +1256,13 @@ TkUtfToUniChar(
 
 int TkUniCharToUtf(int ch, char *buf)
 {
-    if (((unsigned)(ch - 0x10000) <= 0xFFFFF)) {
-	/* Spit out a 4-byte UTF-8 character */
-	*buf++ = (char) ((ch >> 18) | 0xF0);
-	*buf++ = (char) (((ch >> 12) | 0x80) & 0xBF);
-	*buf++ = (char) (((ch >> 6) | 0x80) & 0xBF);
-	*buf = (char) ((ch | 0x80) & 0xBF);
-	return 4;
-    } else {
-	return Tcl_UniCharToUtf(ch, buf);
+    if ((sizeof(Tcl_UniChar) == 2) && (((unsigned)(ch - 0x10000) <= 0xFFFFF))) {
+	/* Spit out a 4-byte UTF-8 character or 2 x 3-byte UTF-8 characters, depending on Tcl
+	 * version and/or TCL_UTF_MAX build value */
+	int len = Tcl_UniCharToUtf(0xD800 | ((ch - 0x10000) >> 10), buf);
+	return len + Tcl_UniCharToUtf(0xDC00 | (ch & 0x7FF), buf + len);
     }
+    return Tcl_UniCharToUtf(ch, buf);
 }
 
 
diff --git a/generic/ttk/ttkEntry.c b/generic/ttk/ttkEntry.c
index 44e45de..96f3cf2 100644
--- a/generic/ttk/ttkEntry.c
+++ b/generic/ttk/ttkEntry.c
@@ -279,7 +279,7 @@ static char *EntryDisplayString(const char *showChar, int numChars)
     char *displayString, *p;
     int size;
     int ch;
-    char buf[4];
+    char buf[6];
 
     TkUtfToUniChar(showChar, &ch);
     size = TkUniCharToUtf(ch, buf);
diff --git a/macosx/tkMacOSXFont.c b/macosx/tkMacOSXFont.c
index c8fce8b..c3b0b66 100644
--- a/macosx/tkMacOSXFont.c
+++ b/macosx/tkMacOSXFont.c
@@ -176,7 +176,7 @@ TkUtfAtIndex(
 	*code = CFStringGetLongCharacterForSurrogatePair(
 	    uniChar, lowChar);
 	ptr += Tcl_UniCharToUtf(uniChar, ptr);
-        ptr += Tcl_UniCharToUtf(lowChar, ptr);
+	ptr += Tcl_UniCharToUtf(lowChar, ptr);
 	return ptr - uni;
     } else {
 	*code = (int) uniChar;
diff --git a/win/tkWinFont.c b/win/tkWinFont.c
index c430d48..cef505a 100644
--- a/win/tkWinFont.c
+++ b/win/tkWinFont.c
@@ -2182,7 +2182,7 @@ FontMapLoadPage(
     Tcl_Encoding encoding;
     int i, j, bitOffset, end, segCount;
     USHORT *startCount, *endCount;
-    char buf[16], src[4];
+    char buf[16], src[6];
 
     subFontPtr->fontMap[row] = (char *)ckalloc(FONTMAP_BITSPERPAGE / 8);
     memset(subFontPtr->fontMap[row], 0, FONTMAP_BITSPERPAGE / 8);
diff --git a/win/tkWinKey.c b/win/tkWinKey.c
index 29f2ff0..c7fa20c 100644
--- a/win/tkWinKey.c
+++ b/win/tkWinKey.c
@@ -98,7 +98,7 @@ TkpGetString(
 {
     XKeyEvent *keyEv = &eventPtr->xkey;
     int len;
-    char buf[4];
+    char buf[6];
 
     Tcl_DStringInit(dsPtr);
     if (keyEv->send_event == -1) {