[climacs-devel] keyboard macros, display protocol

Christophe Rhodes csr21 at cam.ac.uk
Sun May 7 17:53:20 UTC 2006


Nikodemus Siivola <nikodemus at random-state.net> writes:

> Robert Strandh <strandh at labri.fr> writes:
>
>> I would prefer that this case be handled by having the commands that
>> require the syntax to be up to date explicitly call whatever function
>> updates the syntax.
>
> I'm not following Climacs too closely, so sorry for butting in, but...
>
> Why on earth? Isn't that the classic recipe for commands that only work
> under magic assumptions, and duplicating work?
>
> I'm obviously missing something here.

Whether or not you're missing something, there are multiple clients
for an up to date syntax, and they have different needs; we start from
the observation that doing a complete parse of the entire buffer at
every keystroke (or even more often: again, consider keyboard macros)
is not acceptable in terms of performance for arbitrary grammar.
Therefore, there cannot be a general automagic system that ensures
that the syntax is sufficiently up to date for a given command to
operate correctly.  (Stop me there if there's something you want to
disagree with).

Since only each individual client of a syntax knows how much of the
syntax needs to be up to date (and there we will need to extend the
current support, so that it is reasonable to say "please update the
syntax until you find the end of this top-level form" or similar), the
logical place to put the calls to extend the up-to-date syntax region
is where they're needed.

Cheers,

Christophe



More information about the climacs-devel mailing list