Issue #207 - deal with 2 cases of an unambiguous ampersand.

html5 allows a naked ampersand unquoted, and now tidy will not issue a
warning. This only deals with a & b, and P&<li>O</li>

More may need to be done for other cases.

Author: geoffmcl

src/lexer.c

@@ -1055,9 +1055,17 @@ static void ParseEntity( TidyDocImpl* doc, GetTokenMode mode )
             if (semicolon)
                 TY_(AddCharToLexer)( lexer, ';' );
         }
-        else /* naked & */
-            TY_(ReportEntityError)( doc, UNESCAPED_AMPERSAND,
+        else
+        {
+            /*\ 
+             *  Issue #207 - A naked & is allowed in HTML5, as an unambiguous ampersand!
+            \*/
+            if (TY_(HTMLVersion)(doc) != HT50) 
+            {
+                TY_(ReportEntityError)( doc, UNESCAPED_AMPERSAND,
                                     lexer->lexbuf+start, ch );
+            }
+        }
     }
     else
     {

src/pprint.c

@@ -982,6 +982,14 @@ static void PPrintText( TidyDocImpl* doc, uint mode, uint indent,
             ixWS = TextStartsWithWhitespace( doc->lexer, node, ix+1, mode );
             ix = IncrWS( ix, end, indent, ixWS );
         }
+        else if (( c == '&' ) && (TY_(HTMLVersion)(doc) == HT50) &&
+            (((ix + 1) == end) || ((ix + 1) < end) && (isspace(doc->lexer->lexbuf[ix+1]))) )
+        {
+            /*\
+             * Issue #207 - This is an unambiguous ampersand need not be 'quoted' in HTML5
+            \*/
+            PPrintChar( doc, c, (mode | CDATA) );
+        }
         else
         {
             PPrintChar( doc, c, mode );