[BRLTTY] [HT] The status cell problem...

[Mario Lang]

Nicolas Pitre <nico at cam.org> writes:

> On Thu, 30 Nov 2006, Mario Lang wrote:
>
>> On the Modular Evolution cursor routing keys are currently off by a certain
>> amount (the status cell count).
>> 
>> I propose to "fix" this problem by special-casing
>> models with continuous cells and putting the status cells on the right side
>> for now.  This at least fixes cursor routing (and cut-and-paste and all that).
>
> No.
>
> The proper "fix" is for the driver to report the actual routing key 
> number by substracting the number of status cells as needed.  If the 
> driver knows how many status cells a given display has and tells the 
> core about them, it also must ensure the routing keys reported to the 
> core is coherent with that as well.  Then the routing keys above status 
> cells become extra buttons for other purposes.

Yes, but all the other models in the ht driver (those with status cells)
are not off-by-some-amount.  Only the evolution series is.

This is because previous models never were continuous.  To keep the protocol,
the situation is effectively that cursor routing is not
wrong if we put status cells to the right.  However, previous (non continuous)
models assume that the protocol sends the status cells first (left).
So we need to at least special case and reverse the writing logic.
Since for the previous models, there is no need to calculate anything special.

-- 
CYa,
  Mario