When the 'font-family' CSS attribute's value contains Chinese, Japanese or Korean symbols, the HTML parser could not parse them correctly, and the applied font families are skipped. As a result, the text is imported with Arial font.

Steps to reproduce:
- Register the following font "HGS明朝B":
FontFamily myCustomFont = new FontFamily("HGS明朝B");
FontManager.RegisterFont(myCustomFont);
- Write some text and apply the font over the text.
- Export to HTML.
- Import the HTML.

Expected: the text should be with the HGS明朝B font family.
Actual: the text is with Arial font family

When a document with table of contents is exported to HTML, the fore color, alignment of the page numbers are not respected; the tab leader is missing and the entries are underlined as hyperlinks.

Most of the CJK characters are not defined or don't look good in fonts not specifically designed for CJK.

Because of this it would be convenient the font to be automatically changed to MS Mincho in case the user starts writing CJK characters. 

Note that if the current font supports CJK (for example Arial Unicode MS), the font is not changed to MS Mincho.

The problem occurs in specific situations when a list depends on a style that hasn't been added to the fragment in the clipboard yet.

Workaround: Replace the fragment in the clipboard with one that is already "fixed", as in the attached MainWindow.cs.txt

If there is a paragraph with set background color, the styles for paragraphs are not reset after the colored one and this results in additional background color when importing RTF documents.

In MS Word the user could add different effects to the selected text. This could be done through Font -> Text Effects... button.

In HTML the width of the columns in a table could be set as proportions of the remaining free space using width="3*"

Example:
<TABLE>
<COLGROUP>
   <COL width="30">
<COLGROUP>
   <COL width="30">
   <COL width="0*">
   <COL width="2*">
<COLGROUP>
   <COL width="1*">
   <COL width="3*">
<THEAD>
<TR><TD> ...
...rows...
</TABLE>



First the user agent will allot 30 pixels to columns one and two. Then, the minimal space required for the third column will be reserved. The remaining horizontal space will be divided into six equal portions (since 2* + 1* + 3* = 6 portions). Column four (2*) will receive two of these portions, column five (1*) will receive one, and column six (3*) will receive three.

Note: This is not supported even by MS Word. 

When a document containing fields and with a custom theme is exported to .docx and opened in MS Word, the theme is different.

Available in R3 2018 Official Release.

If a document with default tab stops width 0 or imported or the value is set from the UI, measuring the document when it contains a tab causes an InvalidOperationException with message "MeasureOverride returned Nan".

Workaround: If the case is importing a document, change the default tab stop width to value which is not 0.

            if (document.DefaultTabWidth == 0)
            {
                  // set this to large number if there are performance problems, e.g. 0.5. 
                  document.DefaultTabWidth = 0.01;
            }

Steps to reproduce:
1. Load RadRichTextBox and add text in it.
2. Change default tab stops value from the Paragraph Properties Dialog -> Tabs Dialog to 0''
3. Add a tab to the paragraph.
Observed: InvalidOperationException with message "MeasureOverride returned NaN" is thrown.

Fix available in LIB Version 2017.3.1120.

Selection geometry is not rectangular when the selection contains paragraphs with specific fonts and indents, e.g. two empty paragraphs with with font family Georgia and the second one having First Line Indent = 6 px.

Steps to reproduce:
- Add 3 empty paragraphs
- Select the first 2 and apply font family Georgia
- Apply First line indent of 6 px to the second paragraph
- Select all
Expected: The selection geometry should be rectangular.
Actual: The selection geometry is not a rectangle.

When a table with PreferredWidth=Auto and fixed columns width is inserted in another table's cell with the same settings, the inserted table changes the width of the other table. 

InvalidOperationException: 'ColumnSpan must be at least 1.' is thrown when importing HTML document with table column containing colspan="0".

The colspan="0" has special meaning according to the HTML specification http://www.w3.org/TR/html401/struct/tables.html:
-------------------
colspan = number [CN]

This attribute specifies the number of columns spanned by the current cell. The default value of this attribute is one ("1"). The value zero ("0") means that the cell spans all columns from the current column to the last column of the column group (COLGROUP) in which the cell is defined.
-------------------
Workaround: Preprocess the HTML, and delete all occurrences of colspan="0".

Add support for importing of imaged defined with VML's v:imagedata inside v:image or v:shape. Currently such images are just skipped.

Documents that are protected using WordsProcessing could be protected using SHA256, SHA512 and SHA384 algorithms. However these algorithms are currently not supported by RichTextBox. Thus, a document protected via WordsProcessing could not be unprotected in RadRichTextBox and NotSupportedException is thrown. Add support for these three algorithms.
Documents protected in MS Word 2013 could not be unprotected in RRTB as well. Exception thrown: Unsupported hashing algorithm: SHA512.

Add support for = (Formula) fields, including the scenario with referencing other table cells.

More information: https://support.office.com/en-us/article/Field-codes-Formula-field-32d5c9de-3516-4ec3-80ed-d1fc2b5bc21d?ui=en-US&rs=en-US&ad=US

When a table is auto fitted its cells are exported with incorrect width value. An additional requirement is the cells to not have any preferred width (NULL).
This problem will occur only if the document is converted directly from one format to another.
Possible workaround is to apply auto preferred width to all cells in that table.

Documents containing <pre> tag and line breaks (CR LF) cause unexpected visual behavior (loss of content) and invalid layout. Sporadically, document positions seem to also get invalid which leads to exceptions.

The width of the image is inherited from the table element it is in when it doesn't have a specific width set to it. Sample HTML reproducing the problem is:

<html>
	<body>
		<table style="width: 700px;">
			<tbody>
				<tr>
					<td>
						<img src="" />
					</td>
				</tr>
			</tbody>
		</table>
	</body>
</html>

Instead, the real size of the image should be used.

Workaround: Subscribe to LoadImageFromUri and set size to the image.

            settings.LoadImageFromUrl += (s, arg) =>
                {
                    if (arg.Url != null)
                    {
                        string extension = Path.GetExtension(arg.Url);
                        Uri uri = new Uri(arg.Url, UriKind.RelativeOrAbsolute);
                        System.Net.WebClient oWebClient = new System.Net.WebClient();

                        using (MemoryStream stream = new MemoryStream())
                        {
                            oWebClient.OpenRead(uri).CopyTo(stream);
                            arg.ImageElement.Init(stream, extension);
                            arg.ImageElement.Size = new Size(164, 80);

                            e.Handled = true;
                        }
                    }
                };

The table width value is exported with different decimal separator depending on the used culture. One the problem is fixed the width value should be exported with invariant culture.