[BRLTTY] Cursor routing appears to fail with 4.3 in Lynx?

[Shérab]

Hi,

> The problem it resolves is the cursor being temporarily moved to an unexpected 
> location, so that must be happening.
> 
> To do this, brltty has to wait a little longer to ensure that the cursor has 
> stopped moving around. I'm very interested, therefore, in as many users as 
> possible testing cursor routing in the development stream to see if, perhaps, 
> it has become too slow now.

What I've noticed e.g. when running brltty in the Iris is that one can
see the cursor moving on the display to its final position and the whole
process seems to me a bit slower than it was before.

Would it make sense tomodify the algorithm in the following way: for the
first few vertical/horizontal motions, wait a bit longer in case the
cursor moves temporaily. If that does not occur, then do the following
motions faster. Is it possible for brltty to detect these temporary
motions at all ?

Sherab.