Re: New undo element (fun . args)

[Kim F. Storm]

Richard Stallman <address@hidden> writes:

>     100% agreement.  I even suggest we pass the BEG and END args to FUN.
>     I.e. basically keep the current setup, except require the first 2 args to 
> be
>     BEG and END.
>
> That would be a nuisance; you would be unable to use existing functions
> for this purpose.
>
> By the way, note that we would have to change SES to handle the new
> format, if we change the format.

We have two othogonal issues:

a) the format of the undo entry

Here I guess we can agree on (apply DELTA BEG END FUN . ARGS)
And that FUN is called with ARGS only  (if necessary, the
BEG END etc. can be included explicitly in ARGS).

b) the undo-in-region support

I don't know, but in the case of ses, it may not make sense to have
undo in a region, and consequently may be quite hard to put something
sensible into DELTA, BEG and END.

For this purpose, I suppose that we have an alternate, short format
that omits the undo-in-region information:

    (apply FUN . ARGS)

For undo-in-region, this simply says that undo-in-region is not
supported beyond this element in the undo list.

It is trivial to distinguish the two formats - if second element is
an integer, it is the long format, otherwise it is the short format.

That makes it trivial to change ses.el.

-- 
Kim F. Storm <address@hidden> http://www.cua.dk