[LeonarddeR]

Link to issue number:

Closes #20982.
Builds on #20987.

Summary of the issue:

NVDA reads rich edit controls through the Text Object Model exposed using ITextDocumentTextInfo. It reports character formatting and alignment there, but not headings or lists. Users can't tell headings and list items apart from body text.

Description of user facing changes:

In rich edit controls that NVDA reads through the Text Object Model:

• NVDA reports headings and their level when reporting of headings is on.
• NVDA reports each list paragraph as a list item with its level when reporting of lists is on.
• NVDA reports the bullet or number that the control draws before a list item.

Description of developer facing changes:

None.

Description of development approach:

Headings: A rich edit control stores headings as the built-in paragraph styles "heading 1" to "heading 9". ITextDocumentTextInfo maps these styles (tomParaStyleHeading1 to tomParaStyleHeading9) to the heading-level format field.

Lists:

• A paragraph belongs to a list when its list type is not tomListNone. As in Word, each list paragraph is a list item.
• getTextWithFields wraps the format runs of each list paragraph in a list item control field.
• Format runs stop at paragraph ends. So every run that starts before the end of a list paragraph belongs to the same list item.
• The marker goes into the line-prefix of the first format field of the list item.
• Bullets come from the list type. tomListBullet is the default bullet. A list type above 32 is the bullet character itself.
• Numbers come from ITextRange2::GetText2 with tomIncludeNumbering, called on the first character of the paragraph. It returns the number, a tab and that character. GetText2 leaves out bullets unless the document also contains a numbered list, so it is only used for numbers.
• Windows registers no proxy for ITextRange2. Querying for it from NVDA's process fails with E_NOINTERFACE. NVDA therefore calls GetText2 through IDispatch with comtypes' dynamic dispatch, as it does for Word's object model. This uses the Text Object Model 2 symbols from Include Text Object Model 2 in comInterfaces.tom #20987.

Testing strategy:

Manually tested in Paperback with trypsynth/paperback#1117, which stores headings and lists in its reading view. NVDA reported headings and list items after a plain text load, after a rich text load and after the reading view extended its window.

Known issues with pull request:

None.

Code Review Checklist:

• Documentation:
  • Change log entry
  • User Documentation
  • Developer / Technical Documentation
  • Context sensitive help for GUI changes
• Testing:
  • Unit tests
  • System (end to end) tests
  • Manual testing
• UX of all users considered:
  • Speech
  • Braille
  • Low Vision
  • Different web browsers
  • Localization in other languages / culture than English
• API is compatible with existing add-ons.
• Security precautions taken.

[Copilot]

🟡 Changes recommended

List-item fields suppress normal speech and braille reporting, and list levels are off by one.

2 open findings

• Expose bare LISTITEM role and level in normal reading and braille · New
• Avoid adding one to TOM's already one-based list level · New

What changed in this PR

Adds heading and list information to NVDA’s Text Object Model support for rich edit controls.

Changes:

• Maps built-in heading styles to heading levels.
• Wraps list paragraphs in list-item fields with nesting levels.
• Exposes bullets and numbers through line-prefix formatting.

File	Description
source/​NVDAObjects/​window/​edit.py	Adds heading detection, list-item fields, and marker extraction.

🧠 Review effort: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

[LeonarddeR]

This kan now be tested with development builds of Paperback.

[seanbudd]

Is this ready for review or still being worked on?

[LeonarddeR]

@seanbudd yes, the only reason this is a draft is to wait for the 2027.2 dev cycle so I can add the changelog entry.

[CyrilleB79]

@seanbudd you've added 2027.1 milestone. I guess it's a mistake, isn't it?