Re: Reading sexp problem: Invalid read syntax: "?"

Re: Reading sexp problem: Invalid read syntax: "?"

[Kim F. Storm]

Christoph Conrad <christoph.conrad@gmx.de> writes:

> Hi,
> 
> i have a strange problem with my CVS Emacs sometimes. Cause i cannot
> reproduce it with "emacs -q --no-site-file" and i suspect that it is a
> configuration problem, not a problem specific to CVS Emacs, i am
> asking here.
> 
> In CVS Gnus contrib/xml.el there ist the following statement in
> method `xml-parse-elem-type':
> 
>       (list '? elem)
> 
> eval-defun reveals
> 
> edebug-read-sexp: Invalid read syntax: "?"
> 
> Changing the line above to
> 
>       (list '?  elem)
> 
> - that is adding one extra space after the question mark - and the
> error is gone.

In CVS Emacs, the Lisp Parser has been modified to require a character
constant to be separated from a following symbol by whitespace.

In the above code, the character constant is a space, but consider
        (list '?melem)

Here, you would probably agree to write (list '?m elem) for clarity.
And so (list '?  elem) isn't any different, conceptually.

However, it does seem plausible (or proven) that existing code does
indeed use the "single space" syntax so maybe we need to relax the
parser so that it only checks the token following a NON-whitespace
character constant.

I've cc:ed to emacs-devel for comments.  So WDYT?


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

Re: Reading sexp problem: Invalid read syntax: "?"

[Christoph Conrad]

Hi Kim,

    > Here, you would probably agree to write (list '?m elem) for
    > clarity. And so (list '? elem) isn't any different,
    > conceptually.

Mhmm, yes, that's plausible.

    > maybe we need to relax the parser so that it only checks the
    > token following a NON-whitespace character constant.

Probably would be helpful, i agree.

Beste Grüße,
cu, -cc-
-- 
Computers are useless.  They can only give you answers.
     Pablo Picasso

Re: Reading sexp problem: Invalid read syntax: "?"

[Richard Stallman]

    However, it does seem plausible (or proven) that existing code does
    indeed use the "single space" syntax so maybe we need to relax the
    parser so that it only checks the token following a NON-whitespace
    character constant.

It makes me sad to agree to this, but looking at the amount of
practical impact that the change has had, I think I have to agree.

Re: Reading sexp problem: Invalid read syntax: "?"

[Kim F. Storm]

Richard Stallman <rms@gnu.org> writes:

>     However, it does seem plausible (or proven) that existing code does
>     indeed use the "single space" syntax so maybe we need to relax the
>     parser so that it only checks the token following a NON-whitespace
>     character constant.
> 
> It makes me sad to agree to this, but looking at the amount of
> practical impact that the change has had, I think I have to agree.

I installed a fix accept the "single space" syntax (for SPC and TAB).

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

Re: Reading sexp problem: Invalid read syntax: "?"

[Kim F. Storm]

"David PONCE" <David.Ponce@wanadoo.fr> writes:

> Would it make sense to issue warnings in such cases, saying that
> that use of "single space" syntax is not correct, and won't be
> supported in future versions of Emacs? 

I think the problem is too specific to be worth the hazzle.

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

Re: Reading sexp problem: Invalid read syntax: "?"

[Richard Stallman]

    Would it make sense to issue warnings in such cases, saying that that use of "single space" syntax is not correct, and won't be supported in future versions of Emacs?

It would make sense to issue such warnings if we can arrange to do it
in a way that is both helpful and not obnoxious.

We could enable these warnings in the compiler only.  Perhaps that
would be sufficiently helpful and nonobnoxious.