[PATCH] RTF: underline and other stuff

[Nicolas Goutte]

The patch is for the directory filters/kword/rtf/import

I have tried to fix the loading of AbiWord's test file rtftest.rtf.

Changes:
- support advanced underline modes (thick, dot, dash dot, dash dot dot.)
- support conditional hyphen
- do not generate control characters under 32 (apart TAB, LF, CR.)

While working on the underline modes, I have found a bug in KWord, which does 
not respect the dot and dash-dot modes. (Bug reported, number pending.)

I have found another bug, this time more in the filter and in the file that I 
have tested. The problem is the sequence (in a \uc1 context )
\u8164\'20
The filter does not recognised the Unicode character value and tries to 
generated a control character outside XML's valid characters. The sequence 20 
is already considered as normal text.

Have a nice day/evening/night!
-------------- next part --------------
Index: rtfimport.cpp
===================================================================
RCS file: /home/kde/koffice/filters/kword/rtf/import/rtfimport.cpp,v
retrieving revision 1.24
diff -u -r1.24 rtfimport.cpp
--- rtfimport.cpp	2002/07/23 09:11:00	1.24
+++ rtfimport.cpp	2002/07/24 19:48:54
@@ -72,6 +72,7 @@
 	PROP(	0L,		"|",		insertSymbol,		0L, 0x00b7 ),
 	PROP(	0L,		"}",		insertSymbol,		0L, '}' ),
 	PROP(	0L,		"~",		insertSymbol,		0L, 0x00a0 ),
+	PROP(	0L,		"-",		insertSymbol,		0L, 0x00ad ),
 	PROP(	0L,		"ansicpg",	setCodepage,		0L, 0 ),
 	MEMBER(	0L,		"b",		setToggleProperty,	state.format.bold, 0 ),
 	MEMBER(	"@colortbl",	"blue",		setNumericProperty,	blue, 0 ),
@@ -206,9 +207,13 @@
 	MEMBER(	0L,		"u",		insertUnicodeSymbol,	state.format.uc, 0 ),
 	MEMBER(	0L,		"uc",		setNumericProperty,	state.format.uc, 0 ),
 	MEMBER(	0L,		"ul",		setToggleProperty,	state.format.underline, 0 ),
-	MEMBER(	0L,		"uld",		setFlagProperty,	state.format.underline, true ),
+	MEMBER(	0L,		"uld",		setFlagProperty,	state.format.underlineDot, true ),
+	MEMBER(	0L,		"uldash",	setFlagProperty,	state.format.underlineDash, true ),
+	MEMBER(	0L,		"uldashd",	setFlagProperty,	state.format.underlineDashDot, true ),
+	MEMBER(	0L,		"uldashdd",	setFlagProperty,	state.format.underlineDashDotDot, true ),
 	MEMBER(	0L,		"uldb",		setFlagProperty,	state.format.underlined, true ),
 	MEMBER(	0L,		"ulnone",	setFlagProperty,	state.format.underline, false ),
+	MEMBER(	0L,		"ulth",		setFlagProperty,	state.format.underlineThick, true ),
 	MEMBER(	0L,		"ulw",		setFlagProperty,	state.format.underline, true ),
 	MEMBER(	0L,		"up",		setUpProperty,		state.format.baseline, 6 ),
 	MEMBER(	0L,		"v",		setToggleProperty,	state.format.hidden, 0 ),
@@ -716,6 +721,11 @@
     format.vertAlign	= RTFFormat::Normal;
     format.underline	= false;
     format.underlined   = false;
+    format.underlineDash       = false;
+    format.underlineThick      = false;
+    format.underlineDot        = false;
+    format.underlineDashDot    = false;
+    format.underlineDashDotDot = false;
     format.bold		= false;
     format.italic	= false;
     format.strike	= false;
@@ -948,7 +958,7 @@
     token.type = RTFTokenizer::PlainText;
     token.text = buf;
 
-    if (ch > 0)
+    if (ch > 31)
     {
 	if (ch > 0x007f)
 	{
@@ -962,6 +972,14 @@
 	}
 	*text++ = ch;
     }
+    else
+    {
+        // We have a control character, so we must check if it is XML-compactible
+        if ( ch == 9 || ch == 10 || ch == 13 )
+            *text++ = ch ;
+        else
+            kdWarning() << "RTFImport::insertUTF8: tried to insert control character " << ch << endl;
+    }
     *text++ = 0;
 
     (this->*destination.destproc)();
@@ -1787,14 +1805,48 @@
 	    node.setAttribute( "value", format.fmt.italic );
 	    node.closeNode( "ITALIC" );
 	}
-	if (!baseFormat || format.fmt.underline != baseFormat->underline || format.fmt.underlined != baseFormat->underlined)
+        if (!baseFormat || format.fmt.underline != baseFormat->underline
+            || format.fmt.underlined != baseFormat->underlined
+            || format.fmt.underlineDash != baseFormat->underlineDash
+            || format.fmt.underlineThick != baseFormat->underlineThick
+            || format.fmt.underlineDot != baseFormat->underlineDot
+            || format.fmt.underlineDashDot != baseFormat->underlineDashDot
+            || format.fmt.underlineDashDotDot != baseFormat->underlineDashDotDot )
 	{
 	    node.addNode( "UNDERLINE" );
-            QCString st;
+            QCString st,styleline;
             st.setNum(format.fmt.underline);
             if ( format.fmt.underlined )
                 st="double";
-	    node.setAttribute( "value", st );
+            else if ( format.fmt.underlineDash )
+            {
+                st="1";
+                styleline="dash";
+            }
+            else if (format.fmt.underlineThick )
+            {
+                st="single-bold";
+                styleline="solid";
+            }
+            else if (format.fmt.underlineDot )
+            {
+                st="1";
+                styleline="dot"; // ### FIXME: KWord has bug and considers it as DashDot
+            }
+            else if (format.fmt.underlineDashDot )
+            {
+                st="1";
+                styleline="dashdot"; // ### FIXME: KWord has bug mixing Dot and DashDot
+            }
+            else if (format.fmt.underlineDashDotDot )
+            {
+                st="1";
+                styleline="dashdotdot"; // ### FIXME: KWord has bug mixing Dot and DashDot
+            }
+            node.setAttribute( "value", st );
+            if ( !styleline.isEmpty() )
+                node.setAttribute( "styleline", styleline );
+
 	    node.closeNode( "UNDERLINE" );
 	}
 	if (!baseFormat || format.fmt.strike != baseFormat->strike || format.fmt.striked != baseFormat->striked)
Index: rtfimport.h
===================================================================
RCS file: /home/kde/koffice/filters/kword/rtf/import/rtfimport.h,v
retrieving revision 1.11
diff -u -r1.11 rtfimport.h
--- rtfimport.h	2002/07/23 09:11:00	1.11
+++ rtfimport.h	2002/07/24 19:48:54
@@ -133,6 +133,7 @@
     int color, bgcolor;
     int uc;
     bool underline, underlined, bold, italic, strike, striked, hidden, caps;
+    bool underlineDash, underlineThick, underlineDot, underlineDashDot, underlineDashDotDot;
 };
 
 // Style sheet entry