How do I set the caret position for text runs?

[phatchman]

I'm currently implementing the text_run role for DVUI.

Most of the reading part is working, i.e. setting the word-lengths, character positions etc and communicating text selection using anchor and focus. But I can't find any way to set the caret position.

Is it possible to set this with AccessKit? (using the C API).

Edit: I understand from the code that caret position should be set to "focus"? But I do not see any value for CaretPosition in Accessibility Insights. (Using v 0.17.1 of accesskit-c).

Thanks.

[phatchman]

I think I understand this a bit better now. There is no need to update the caret position to the screen reader as the reader keeps a separate, internal start and end selection position.

When a text_selection action is received it will the reader's internal cursor positions are sent and the widget should update it's selection accordingly.

If the user moves the cursor/caret or selects text outside the screen reader, the new selected text will be updated to the reader, and (depending on the reader) updates the screen reader's internal start/end positions.

[DataTriny]

Hi @phatchman,

Yes, the caret position or text selection must be updated everytime, no matter if it results from a click, a key press, an action from an assistive technology or a programmatic update from the app or UI toolkit.

AccessKit represents caret position as a degenerate text selection, which means that both anchor and focus fields of the accesskit_text_selection struct must have the same accesskit_text_position value.

With that being said, I don't think Accessibility Insights shows the caret position in the list of properties for the Text pattern. I can help you check with a real screen reader if you need me to, just send me the name of your git branch.

[phatchman]

I've also been testing against NVDA and the word / character navigation and selection look to be working, though it is lightly tested. But, yes appreciate if you could check as well.

It should have the logic you describe, where it sets anchor and focus to start/end selection, and sets both anchor and focus to caret position if there is no selection.

david-vanderson/dvui#661 or https://github.com/phatchman/dvui-dev/tree/accesskit_text_run

Thanks and happy new year!