core: Fix double caret rendering in justified text

[kjarosh]

Progresses #1891.

Before	After

[torokati44]

My finger is hovering over the merge button, but the fact that I don't understand the change (or even the code it changes) completely, keeps me from pressing it.
Would you mind explaining in a sentence or two how the logic produced two (and not more) carets to be rendered, and how the change results in at most one being rendered, @kjarosh? 🤔

[kjarosh]

Added a description to the commit message which describes the fix.

The condition

visible_selection.start() >= *start && visible_selection.end() <= *end

was inaccurate, because the end of the selection is exclusive. That caused the condition to be true for two adjacent boxes. For instance:

  box 1: from 0 to 6,  "hello "
  box 2: from 6 to 11, "world"

The caret was rendered for both boxes when it was at position 6.

When applying a correct condition (i.e. treating the end as exclusive) there is a problem with rendering the caret at the very end of the text, because the condition will not be triggered for any box (position 11 in the example above).

That is why a condition specific to this case is added, i.e.

*end == text_len

When the box is the last box in the text, we are forcing the caret to be rendered.

[torokati44]

Oookay, it all makes sense now, thanks!