10:57 a.m.
Hi,
IMHO a routine detecting nested forms is needed - for
all kind of things which have a start and a end.
Implemented a kind of this in sh-beg-end.el.
Such a form should take start and end as arguments,
then report nesting levels and positions of beginnings
and ends.
Maybe it exists already? Then sh-beg-end should use it,
otherwise I'm going to write it.
Cheers
Andreas
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
1:23 p.m.
Andreas Roehler writes:
IMHO a routine detecting nested forms is needed - for
all kind of things which have a start and a end.
What is it you need that a sexp (s-expression, balanced expression)
doesn't give you?
Such a form should take start and end as arguments,
then report nesting levels and positions of beginnings
and ends.
Well, really it ought to take a list of start-end pairs, no?
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
3:41 p.m.
Stephen J. Turnbull wrote:
Andreas Roehler writes:
> IMHO a routine detecting nested forms is needed - for
> all kind of things which have a start and a end.
What is it you need that a sexp (s-expression, balanced expression)
doesn't give you?
Best to see in a shell-script: how to jump at the beginning or end
of a nested "if/fi"?
IMO it was not possible in a simple way.
sh-beg-end.el implements that.
> Such a form should take start and end as arguments,
> then report nesting levels and positions of beginnings
> and ends.
Well, really it ought to take a list of start-end pairs, no?
Precisely. sh-beg-end does it, but is hard-coded for its purpose.
Just to make it up into the general.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
4:11 a.m.
Andreas Roehler writes:
Precisely. sh-beg-end does it, but is hard-coded for its purpose.
Just to make it up into the general.
OIC. Why not just have a general LALR parser, then? semantic
(regular XEmacs package) or synfl (check out
XEmacs/packages/unsupported/stephen/synful from cvs.xemacs.org).
Or create a loadable module to embed bison in XEmacs .... (That's
hard, since it needs to work on Lisp buffers rather than raw text.)
Didier Verna <didier(a)xemacs.org> suggested embedding a scannerless
parser in XEmacs. You could ask him if he ever did anything with
that.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
8:30 a.m.
Stephen J. Turnbull wrote:
Didier Verna <didier(a)xemacs.org> suggested embedding a
scannerless
parser in XEmacs. You could ask him if he ever did anything with
that.
Nope, like 99% of my XEmacs-related projects in the last 6 years :-(
Or create a loadable module to embed bison in XEmacs ....
(That's
hard, since it needs to work on Lisp buffers rather than raw text.)
That was more the idea, yeah.
My original idea (IIRC) was that the notion of major mode as it is right
now is shaky/obsolete. The way to go is do natively (I insist on
"natively") something in the lines of mmm. For instance, in a code
buffer, you want text-mode in comments and <code>-mode in the code
areas. There would be some kind of "meta-mode" to coordinate all
involved (sub-)major-modes, but no <prog-lang>-mode should have to deal
with the contents of comments in itself[1]. For that (and also to ease
fontification) we would really need a proper syntactic analysis in
buffers.
Footnotes:
[1] Of course, things can be more complex than that, for instance in
strings where you would like text-oriented editing, but the language
also has something to say about string syntax.
--
European Lisp Symposium, May 2009: http://www.european-lisp-symposium.org
European Lisp Workshop, July 2009: http://elw.bknr.net/2009
Scientific site: http://www.lrde.epita.fr/~didier
Music (Jazz) site: http://www.didierverna.com
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
12:59 p.m.
Didier Verna writes:
My original idea (IIRC) was that the notion of major mode as it is
right now is shaky/obsolete. The way to go is do natively (I
insist on "natively") something in the lines of mmm.
Yup. I have some (not even close to compiling) code somewhere which
adds a "mode" locale to specifiers, and the next step after that would
have been a primitive "mode" property for extents.
For that (and also to ease fontification) we would really need a
proper syntactic analysis in buffers.
Yup. That's precisely what synfl was aimed at.
_______________________________________________
XEmacs-Beta mailing list
XEmacs-Beta(a)xemacs.org
http://calypso.tux.org/cgi-bin/mailman/listinfo/xemacs-beta
5988
days inactive
5989
days old
5 comments
3 participants
participants (3)
-
Andreas Röhler -
Didier Verna -
Stephen J. Turnbull