The symbol `@' and sexp scanning

The symbol `@' and sexp scanning

[Michael Heerdegen]

Hi,

the following breaks my "el-search" package when searching in
"skeleton.el".  Am I doing something unkosher, or is this just a bug?

Insert

    (eq element '@)

in an elisp mode buffer.  Put point at the quote or the "@".  Eval

    (goto-char (scan-sexps (point) 1))

and you get an error:


Debugger entered--Lisp error: (scan-error "Containing expression ends prematurely" 15 16)
  scan-sexps(13 1)
  (goto-char (scan-sexps (point) 1))
  eval((goto-char (scan-sexps (point) 1)) nil)
  eval-expression((goto-char (scan-sexps (point) 1)) nil)
  funcall-interactively(eval-expression (goto-char (scan-sexps (point) 1)) nil)
  call-interactively(eval-expression nil nil)
  command-execute(eval-expression)


But if you add one more sexp to the list after "'@", there is no such
error.


TIA,

Michael.

Re: The symbol `@' and sexp scanning

[Andreas Politz]

I believe the answer is in here:

,----[ (info "(elisp) Syntax Flags") ]
|    • ‘p’ identifies an additional prefix character for Lisp syntax.
|      These characters are treated as whitespace when they appear between
|      expressions.  When they appear within an expression, they are
|      handled according to their usual syntax classes.
`----

This flag is set in elisp buffers.  But after evaluating
(modify-syntax-entry ?@ "_") the function performs as expected.

-ap

Re: The symbol `@' and sexp scanning

[Michael Heerdegen]

Hello Andreas,

(we are talking about bug#24542")

> I believe the answer is in here:
>
> ,----[ (info "(elisp) Syntax Flags") ]
> |    • ‘p’ identifies an additional prefix character for Lisp syntax.
> |      These characters are treated as whitespace when they appear
> |      between 
> |      expressions.  When they appear within an expression, they are
> |      handled according to their usual syntax classes.
> `----

FWIW, I can't make much sense out of this description (but the first
sentence).  Nearly anything in an Elisp buffer is inside an expression.
Anything else is either a comment.  Or, the only case that seems to fit
into this description is an isolated "@" at top-level.  According to
this text, it should be treated as whitespace.  Dunno what that exactly
means, but evaluating something like that results in an error saying
that the symbol `@' is unbound, so it doesn't seem to be treated as
whitespace, but as an expression.

> This flag is set in elisp buffers.  But after evaluating
> (modify-syntax-entry ?@ "_") the function performs as expected.

Ok, but I guess it's not syntax classification that is problematic but
how it is treated in the case of the bug.  Also see the seemingly very
related bug#22238, btw.


Regards,

Michael.

Re: The symbol `@' and sexp scanning

[Andreas Politz]

Michael Heerdegen <michael_heerdegen@web.de> writes:

> FWIW, I can't make much sense out of this description (but the first
> sentence).  Nearly anything in an Elisp buffer is inside an
> expression.

I think `expression' can be read as `symbol' in this context.
E.g. calling forward-sexp at the beginning of any of the following
expressions results in the same error.

)
')
'')
'@)
@@)
'@'@')

This is because the sexp scanner treats ' and @ (due to the p flag) in
this context as whitespace, i.e. ignores it (Anyway, that's how I
understand it.).  OTOH the following all `work', because there is
something to scan over, before reaching the closing paren.

x)
' x)
@ x)


> [...] evaluating something like that results in an error saying that
> the symbol `@' is unbound, so it doesn't seem to be treated as
> whitespace, but as an expression.

But the way expressions are read and evaluated has nothing to do with
how the function scan-sexp operates.

> Ok, but I guess it's not syntax classification that is problematic but
> how it is treated in the case of the bug.

AFAIK it operates as intended.  As to why @ is treated as a prefix
character in an Elisp buffer I don't know.

-ap

Re: The symbol `@' and sexp scanning

[Stefan Monnier]

> AFAIK it operates as intended.  As to why @ is treated as a prefix
> character in an Elisp buffer I don't know.

Because of the ,@ "unquote splicing".


        Stefan