When exporting, an option of how to create the PDF bookmarks should be provided:
- Using the RadFlowDocument bookmarks
- Using document Headings

The non-breaking hyphen element is currently not supported in the model and is stripped when importing the document.

The calculations are wrong, leading to single lines on a page. As a result, the content of the PDF document is laid out on a bigger number of pages.

Workaround: Change the line spacing and its type before exporting to PDF:

foreach (var paragraph in this.document.EnumerateChildrenOfType<Paragraph>())
{
	HeightType? heightType = paragraph.Properties.LineSpacingType.GetActualValue();
	if (heightType == HeightType.Exact || heightType == HeightType.AtLeast) 
	{
		paragraph.Properties.LineSpacingType.LocalValue = Telerik.Windows.Documents.Flow.Model.Styles.HeightType.Auto;
		paragraph.Properties.LineSpacing.LocalValue = 2;
	}
}

When inserting content in an empty paragraph, the styles applied to it are the default document styles. However, if the properties are present in the last paragraph symbol, the content should inherit them.

Workaround: Copy the properties of the marker after inserting the content:

run.Properties.CopyPropertiesFrom(paragraph.Properties.ParagraphMarkerProperties);

This element specifies the settings for the document grid, which enables precise layout of full-width East Asian language characters within a document by specifying the desired number of characters per line and lines per page for all East Asian text content in this section.

If Latin text is interspersed on this line, then it is placed across the number of grid units needed to fit the
content, but all other grid positions are unaffected.

In a DOCX document, the users can define an image which is bigger than the size of the page. When such an image is at the beginning of the document and is converted with PdfFormatProvider, an additional page is added in the result PDF.

foreach (var image in document.EnumerateChildrenOfType<ImageInline>())
{
    var section = image.Paragraph.Parent as Section;
    if (image.Image.Height > section.PageSize.Height - section.PageMargins.Top - section.PageMargins.Bottom)
    {
        var height = section.PageSize.Height - section.PageMargins.Top - section.PageMargins.Bottom - (image.Paragraph.Spacing.SpacingAfter * (image.Paragraph.Spacing.LineSpacing - 0.97));
        image.Image.Size = new System.Windows.Size(image.Image.Width, height);
    }
}

When a HTML has a wrong encoding set (content="text/html; charset=utf-16") we use the UTF-16 encoding when importing this HTML. This leads to wrong import. 
Microsoft Word detects that the set encoding is wrong and uses UTF-8 instead so that the imported result is correct. 

The HTML5 specification forbids the use of the meta element to declare UTF-16, because the values must be ASCII-compatible: https://www.w3.org/International/questions/qa-html-encoding-declarations#utf16 

When the value in the document variable collection and the argument of the DOCVARIABLE fields are with different casing, the field is not updated.

Available in LIB Version 2018.3.1029.