Seems that the algorithm is not showing the appropriate results when a word has apostrophe. For example typing "etre" does not suggest "être".

A custom third-party spell checker library - NHunspell can be used. 
SDK Example demonstrating this approach can be found in our XAML-SDK portal:
https://github.com/telerik/xaml-sdk/tree/master/RichTextBox/NHunspellSpellChecking

In MS Word, when track changes are enabled, the users can choose to show the revisions in balloons, similar to comments. Add such possibility in RadRichTextBox as well.

The properties applied locally to the content are not preserved and used for the new content when the user changes the paragraph indent.

Workaround: Keep the properties in a variable and apply them after executing the command:

StyleDefinition style;

private void RadRichTextBox_CommandExecuted(object sender, CommandExecutedEventArgs e)
{
    if (e.Command is IncrementParagraphLeftIndentCommand)
    {
        this.radRichTextBox.CurrentEditingStyle.CopyPropertiesFrom(style);
    }
}
        
private void RadRichTextBox_CommandExecuting(object sender, Telerik.Windows.Documents.RichTextBoxCommands.CommandExecutingEventArgs e)
{
    if (e.Command is IncrementParagraphLeftIndentCommand)
    {
        style = new StyleDefinition(this.radRichTextBox.CurrentEditingStyle);
    }
}

When custom annotations are created by inheriting AnnotationRangeStart and AnnotationRangeEnd classes, and the reaction to the delete commands is customized by overriding the DeleteBehavior, BackspaceBehavior, and DeleteSelectedBehavior, the delete command doesn't work as expected in some cases.

The bolder keyword is currently not supported. In the CSS, setting bolder means one font weight heavier than the parent element, so if the parent is with font-weight, font-weight: bolder would mean that the content should be with normal font weight. In MS Word, this value is imported as bold no matter of the parent's value.

The text-decoration-line CSS property is not imported because of wrong name definition in the parsing logic of HtmlFormatProvider.

Add support for positioning floating images relative to section columns. Currently, such images are laid out relative to the first column.

Currently, span elements can only have highlight color and paragraphs can have a background(shading). In MS Word it is possible to apply a background to spans. The import of such a document removes the formatting.

The issue is related to specific steps and is a regression released in R3 2018.

Steps to reproduce:
1. Run RadRichTextBox and insert "test" at the beginning of the document.
2. Move the caret after the 'e'
3. Press Shift+Enter to insert a line break

Observed: NullReferenceException is thrown and the UI becomes unresponsive.
Expected: The break should be successfully inserted in the content

When a DOCX document contains a table with fixed layout, and the table contains <tblGrid> element with improper widths (not coinciding with the ones visualized by MS Word), and such document is imported, the table is laid out with the width of the sum of the column widths imported from the table grid (tblGrid), instead of ones that could be correctly computed by the preferred width of the table cells. 



MS Word, for example, may ignore the widths specified in the tblGrid and compute new ones.



As a side effect, fixed-width tables can be laid out longer than their preferred width.

I want to import a FlowDocument into IDocumentEditor in order to obtain a RadDocument which I can convert to PDF.

There is a NullReferenceException thrown while trying to compare positions for restoring them after pasting the document. The exception is handled but after it is thrown, the layout stops updating.

Using the RadDocumentEditor class, the users can change the margin of the current section. However, there isn't a method allowing them to change the default section margin of the document. 

Add support for the SUBJECT field. The field is evaluated using the document metadata.

Description: https://support.office.com/en-us/article/field-codes-subject-field-f9495276-955a-46a0-b110-0bbc5dd6f0a5?ui=en-US&rs=en-US&ad=US

The character properties applied on the last paragraph symbol for a paragraph in a list (e.g. font family), are exported to HTML for the whole paragraph which is in the list. 

As a notable side effect, when such HTML is imported, the default font size for a paragraph in a default bullet list changes to Symbol.

Regression: The bug is introduced in 2016_2_613.


Workaround: Change the styles to export as inline properties:

but be aware that this may cause issues as described in the related RichTextBox: Character properties of the paragraph symbol are not exported to HTML when StyleExportMode is Inline

Character properties applied on the last paragraph symbol for a paragraph in a list, e.g. font size, are not exported to HTML when StyleExportMode is Inline. 

As a side effect, when the document is imported back, and the text of the bullet is deleted, its properties switch back to the default ones.

Workaround: Change the styles to export as inline properties:
 
Please be aware that this may cause side effects, as described in this issue: RichTextBox: Character properties of the paragraph symbol are exported to HTML for the whole list item when StyleExportMode is Classes

Finding text with regex for the last paragraph is not possible. This is caused by the fact that the search treats the last paragraph symbol as the last paragraph.

If I setup the richtextbox content through the UI as shown in P1.png and try to restore it by pasting the generated html, I get instead P2.png. Expected to be the same. 

I attached the sample project. We're using the recommended settings to use the same html in telerik reports, but can't find the reference on your site, it used on the on this link http://docs.telerik.com/devtools/wpf/controls/radrichtextbox/import-export/features-import-export-settings#htmlformatprovider

 

Typing in Find dialog is not possible when having an ElementHost in a WinForms Windows to host a WPF User Control which contains the RadRichTextBox.

The exception is thrown in RadDocument.IsInNonEditableRange() method, which is called before measuring the document. When using a format provider, the error is thrown but the document can still be opened in RichTextBox. In binding scenario, the document is not imported.

Workaround: Ensure the document is measured before assigning it to RadRichTextBox.