Since prosemirror doc has its own mechanism to compute text position, I can not use a common string search algorithm to index the target substring.
For example, we have a paragraph, there’re two text nodes inside.
p text "cae" [mark: bold] text "eddc"
I want to search “caeed”
But it seems like very difficult to compute the range(start, end) of the found substring
@marjin Thanks for the great work of ProseMirror! Any idea about this question?