Re: Lisp primitives and their calling of the change hooks

[Alan Mackenzie <acm@muc.de>]

Hello, Stefan.

On Wed, Jan 10, 2018 at 15:33:16 -0500, Stefan Monnier wrote:
> > Non-balanced change hooks necessitate special handling.

> I can't imagine why it would be significantly more difficult to handle

>     bcf bcf acf

> than

>     bcf acf acf

> And I can't see why it would be easier to handle

>     bcf bcf acf

> only for the special case where the "missing" acf would have had the
> same length as the previous bcf (i.e. comes from a primitive which
> doesn't change the length of the buffer).

Again, what advantages (other than not changing existing C code) are
there in allowing non-balanced b/a-c-f at all?  The difference in run
time between

    bcf acf acf acf .... acf

(in the cases we currently do this) and

    bcf acf  bcf acf  bcf acf .... bcf acf

is surely negligible, and it is not at all clear that the latter (which
can have simpler hook functions) would be slower than the former (which
may need elaborate special handling).

>         Stefan

-- 
Alan Mackenzie (Nuremberg, Germany).