This fixes some potential race condition in the previous implementation
of the GapContentPosition. I used some WeakReference magic there which
was potentially racy (binary searching an ArrayList that has
WeakReferences in it).
I solved this by inserting an indirection layer. Now the
GapContentPosi
Hi.
Cool that this is now fixed.
> However, there is one (minor IMO) problem left. As it is now, the
> positionMarks array is never cleaned up, that means that marks are kept
> there, even if their corresponding GapContentPosition is GC'ed. This
> might get slightly inefficient in large documents
Hi Robert,
> Cool that this is now fixed.
Well, it's not complete yet. This is why I still didn't close the PR.
> > However, there is one (minor IMO) problem left. As it is now, the
> > positionMarks array is never cleaned up, that means that marks are kept
> > there, even if their corresponding