Re: Terminal matrices

[martin rudalics]

> I've committed something to master which prevents an optimization that
> cannot work on tty root frames when they have a visible child. For me,
> that had the side effect that your example with moving point under a
> child frame to the right until the cursor re-appears started working.
> The cursor was no longer moved to the end of the line. Reason:
> set_cursor_from_row not being called with the current matrix row in this
> case.
>
> If that works for you too now, that would be a good thing because the
> terminal matrix stuff makes problems. Among them something in
> build_frame_matrix_from_leaf_window which is so hard to describe that
> I'll spare me the details for now. Actually 2 problems there, and both
> might apply to master as well, and they have nothing to do with the
> terminal matrices as such.
>
> Or IOW, this all is not fun :-(. If the above cursor movement case now
> works for you on master, I'll put the se-matrix branch on hold.

Doesn't work for me.  All I can say so far is that set_cursor_from_row
is now no more called with delta > 0 regardless of whether there exists a
child frame or not.

martin