Fixed crashes when Fl_Text_* detects illegal UTF 8 sequences. Widgets will not do any further processing but just jump over the character. Screen representation depends largely on whatever the underlying OS does with those sequences, but I feel that this is out of the scope of this library. (STR 2348)

git-svn-id: file:///fltk/svn/fltk/branches/branch-1.3@7965 ea41ed52-d2ee-0310-a9c1-e6b18d33e121

3 files changed, 40 insertions, 19 deletions

diff --git a/src/Fl_Text_Buffer.cxx b/src/Fl_Text_Buffer.cxx
index c70ae1692..d1aea36f1 100644
--- a/src/Fl_Text_Buffer.cxx
+++ b/src/Fl_Text_Buffer.cxx
@@ -1025,7 +1025,7 @@ int Fl_Text_Buffer::search_forward(int startPos, const char *searchString,
           *foundPos = startPos;
           return 1;
         }
-        int l = fl_utf8len(c);
+        int l = fl_utf8len1(c);
         if (memcmp(sp, address(bp), l))
           break;
         sp += l; bp += l;
@@ -1077,7 +1077,7 @@ int Fl_Text_Buffer::search_backward(int startPos, const char *searchString,
           *foundPos = startPos;
           return 1;
         }
-        int l = fl_utf8len(c);
+        int l = fl_utf8len1(c);
         if (memcmp(sp, address(bp), l))
           break;
         sp += l; bp += l;
@@ -1602,7 +1602,7 @@ int Fl_Text_Buffer::prev_char(int pos) const
 int Fl_Text_Buffer::next_char(int pos) const
 {
   IS_UTF8_ALIGNED2(this, (pos))  
-  int n = fl_utf8len(byte_at(pos));
+  int n = fl_utf8len1(byte_at(pos));
   pos += n;
   if (pos>=mLength)
     return mLength;
diff --git a/src/Fl_Text_Display.cxx b/src/Fl_Text_Display.cxx
index d54bbcca9..c55a0d17c 100644
--- a/src/Fl_Text_Display.cxx
+++ b/src/Fl_Text_Display.cxx
@@ -753,7 +753,7 @@ void Fl_Text_Display::overstrike(const char* text) {
   /* determine how many displayed character positions are covered */
   startIndent = mBuffer->count_displayed_characters( lineStart, startPos );
   indent = startIndent;
-  for ( c = text; *c != '\0'; c += fl_utf8len(*c) )
+  for ( c = text; *c != '\0'; c += fl_utf8len1(*c) )
     indent++;
   endIndent = indent;
   
@@ -1735,7 +1735,7 @@ int Fl_Text_Display::handle_vline(
   style = position_style(lineStartPos, lineLen, 0);
   for (i=0; i<lineLen; ) {
     currChar = lineStr[i]; // one byte is enough to handele tabs and other cases
-    int len = fl_utf8len(currChar);
+    int len = fl_utf8len1(currChar);
     if (len<=0) len = 1; // OUCH!
     charStyle = position_style(lineStartPos, lineLen, i);
     if (charStyle!=style || currChar=='\t' || prevChar=='\t') {
@@ -1829,7 +1829,7 @@ int Fl_Text_Display::find_x(const char *s, int len, int style, int x) const {
   // TODO: use binary search which may be quicker.
   int i = 0;
   while (i<len) {
-    int cl = fl_utf8len(s[i]);
+    int cl = fl_utf8len1(s[i]);
     int w = int( string_width(s, i+cl, style) );
     if (w>x) 
       return i;
@@ -3204,7 +3204,7 @@ double Fl_Text_Display::measure_proportional_character(const char *s, int xPix,
     return (((xPix/tab)+1)*tab) - xPix;
   }
   
-  int charLen = fl_utf8len(*s), style = 0;
+  int charLen = fl_utf8len1(*s), style = 0;
   if (mStyleBuffer) {
     style = mStyleBuffer->byte_at(pos);
   }
@@ -3284,7 +3284,7 @@ int Fl_Text_Display::wrap_uses_character(int lineEndPos) const {
   
   c = buffer()->char_at(lineEndPos);
   return c == '\n' || ((c == '\t' || c == ' ') &&
-                       lineEndPos + fl_utf8len(c) < buffer()->length());
+                       lineEndPos + fl_utf8len1(c) < buffer()->length());
 }
 
 
diff --git a/src/fl_utf8.cxx b/src/fl_utf8.cxx
index 94aff0fb8..ccbe98e95 100644
--- a/src/fl_utf8.cxx
+++ b/src/fl_utf8.cxx
@@ -112,9 +112,11 @@ Toupper(
 }
 
 /**
-  return the byte length of the UTF-8 sequence with first byte \p c,
-  or -1 if \p c is not valid.
-  */
+ return the byte length of the UTF-8 sequence with first byte \p c,
+ or -1 if \p c is not valid.
+ This function is helpful for finding faulty UTF8 sequences.
+ \see fl_utf8len1
+ */
 int fl_utf8len(char c)
 {
   if (!(c & 0x80)) return 1;
@@ -137,15 +139,34 @@ int fl_utf8len(char c)
 } // fl_utf8len
 
 
-#if 0
-int fl_utflen(
-        const unsigned char     *buf,
-        int                     len)
+/**
+ Return the byte length of the UTF-8 sequence with first byte \p c,
+ or 1 if \p c is not valid. 
+ This function can be used to scan faulty UTF8 sequence, albeit ignoring invalid
+ codes.
+ \see fl_utf8len
+ */
+int fl_utf8len1(char c)
 {
-	unsigned int ucs;
-	return fl_utf2ucs(buf, len, &ucs);
-}
-#endif
+  if (!(c & 0x80)) return 1;
+  if (c & 0x40) {
+    if (c & 0x20) {
+      if (c & 0x10) {
+        if (c & 0x08) {
+          if (c & 0x04) {
+            return 6;
+          }
+          return 5;
+        }
+        return 4;
+      }
+      return 3;
+    }
+    return 2;
+  }
+  return 1;
+} // fl_utf8len1
+
 
 /**
   returns the number of Unicode chars in the UTF-8 string