Introduce support for Line Annotation.

They are used by the page and the XObjects drawn on it and deal with the overlaying of the colors. These groups are described on page 556 of PdfReference 1.7. 

The UserDict collection is not parsed and the procedures for the operators are unavailable. The exception is thrown while trying to read the font file of the Univers and Univers Bold fonts.

Internally, an ArgumenNullException is thrown, which results in missing content when the document is shown in RadPdfViewer.

This may result in seeing the pattern as a solid color or even not seeing it at all if this color is white and coincides with the background of the page.

RadPdfViewer does not read image resources defined in Form XObject elements.

Default DctDecode implementation uses System.Windows.Media.Imaging.BitmapImage in order to get the image pixels. However, when the image uses some complex colorspace (for instance Separation based on CMYK) the BitmapImage is initialized with a different format and we use System.Windows.Media.Imaging.FormatConvertedBitmap in order to get the pixels in CMYK colorspace. There seems to be some issue with these calculations as the resulting image has incorrect pixels.

WORKAROUND: You may inherit DctDecode class in order to implement custom decoder and call DecodeWithJpegDecoder method from the base class as mentioned in the note of this documentation article:

http://docs.telerik.com/devtools/wpf/controls/radpdfviewer/customization-and-extensibility/customize-pdf-rendering

A sample implementation of these custom decoder may be seen below:

public class CustomDctDecode : DctDecode
{
    public override byte[] Decode(PdfObject decodedObject, byte[] inputData, DecodeParameters parms)
    {
        return DecodeWithJpegDecoder(inputData);
    }
}

When annotations have annotation flag Print they should be visualized when printing the document. Usually, Widget annotations have such flag and currently RadPdfViewer does not print them at all.

Available in LIB Version 2017.3.918.

NullReferenceException is thrown when importing a document with AcroForm referencing a missing object for fields collection. This is a regression as the same document will be imported without a problem in previous versions of RadPdfViewer as the AcroForm object would be entirely skipped.

The redaction annotation is described in the following PDF reference document:
http://www.adobe.com/content/dam/Adobe/en/devnet/acrobat/pdfs/pdf_reference_addendum_redaction.pdf

You may also see sample document with this annotation type attached.

Since LIB Version 2017.3.925 annotations of type Redact can be visualized.

The current implementation relies on the graphic state color to be always SolidColorBrush. However, it may be other color type and InvalidCastException is thrown. This results in missing content.

Available in LIB version 2017.1.410.

Decode array for image masks with DeviceGray color space is not respected, and always the default array of [0 1] is used for visualization. When the mask is encoded with [1 0], the mask is applied inversed, making the wrong pixels transparent. This may lead to images becoming invisible.

Fix available in LIB Version 2017.3.1120.

Due to excessive memory consumption when decoding images, OutOfMemoryException can be thrown when viewing documents with large images inside them. 

Workarounds: 
- Set ExtensibilityManager.MaxImageSize = null. This will skip one resizing step, and could be beneficial in documents containing few large images.
- Use 64-bit process for the client application. This allows the process to consume much more memory. Note that starting 64-bit process while debugging can be tricky, as described in this blog post: https://weblog.west-wind.com/posts/2016/Dec/19/Visual-Studio-Debugging-and-64-Bit-NET-Applications.

Available in R1 2018 Official release version.

As a result, the glyphs are not measured and positioned properly. The issue applies to the TrueType and Type1 fonts.

The document will be imported with a loss of content. The reason is that ASCII85Decoder will throw an ArgumentException when the stream doesn't end with '~>'. 

Although this is invalid case according to PDF specification, other applications are handling it and the document is successfully visualized in them. We may also consider implementing workaround that handles such invalid streams.

The PdfViewer won't load a document from PdfDocumentSource created from Uri and the ReadOnDemand option.
Here is an example: new PdfDocumentSource(new Uri("webaddress/somefile.pdf"), FormatProviderSettings.ReadOnDemand);

Available in LIB Version 2017.1.130

The exception is thrown while decoding TIFF images. Although the document is properly loaded and the end user is not impacted, the DivideByZeroException is logged and present in the OnException event handler parameters. 

The visualizations of all annotations are defined by their normal appearance streams. This is already implemented in RadPdfViewer for the supported annotation types (Link and Widget annotations). It may be easily adopted for not supported annotation types as well.

Available in 2017 R3 Official Release Version.

Currently, RadPdfViewer uses WPF DrawingContext class for drawing page elements to a Canvas.  When PDF page has many content elements (usually geometries) drawing the page may be time-consuming and freezes the UI. 

Instead, we should consider implementing another approach for rendering the content asynchronously which would improve the user experience. 

When text is exported from PDF document using the TextFormatProvider (it's also used internally for the Copy operation), it is automatically split to lines using the vertical position of the words on the page, and small tolerance. 

Currently the tolerance is hard-coded to 0.1 pixels, which is not suitable for documents which contains scanned and OCR-ed text, and there the text lines could be slightly inclined. The result is that words on one slightly inclined line are recognized as if they are on separate lines.

This allows specifying a soft mask in the external graphics state.