emacs-pretest-bug
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: scheme-mode+auto-fill-mode minor bug


From: Dave Love
Subject: Re: scheme-mode+auto-fill-mode minor bug
Date: Sun, 11 May 2003 14:03:14 +0100
User-agent: Gnus/5.1002 (Gnus v5.10.2) Emacs/21.2 (gnu/linux)

"Stefan Monnier" <monnier+gnu/emacs/address@hidden> writes:

> This is due to Dave's change:
>
> revision 1.57
> date: 2003/02/13 15:54:19;  author: fx;  state: Exp;  lines: +4 -0
> (comment-indent): Ensure space before added comment.
> ----------------------------

Sorry, I thought it was trivial...

> I've just installed the patch below which should fix it.
> Dave can you confirm that it still does what you wanted it to ?

I haven't tried it, but it was to fix the case in sh-mode that a line
ended at comment-column and you did M-;.  Then # was appended to the
last word on the line and caused a syntax error.

> PS: I'm less and less happy with the `whitespace' syntax-class because
>     it's really not clear what it's supposed to stand for.

I'm not sure that's the problem.

>     The absence of \n from that class in language where \n is a
>     comment-ender is a common problem,

Yes, I think it should effectively be in two classes.  I assume that
could be dealt with by flags somehow (like the complex comment stuff),
but I haven't looked into it.

>     and the presence of things like ' in that class
>     (in Scheme mode until recently)

Gosh.  I wonder why I didn't fix that.

>     or ~ (in TeX mode) bothers me.

I think that's wrong for the same reason NO-BREAK SPACE doesn't have
whitespace syntax.

>     I think we need to clarify what we mean by `whitespace'.

We agreed we need to do something better with syntax, particularly in
relation to what Unicode says (see a todo item, I think).  [The
Unicode syntax classes (or whatever they're called are listed in one
of the Unicode data files and you can look those up in the
emacs-unicode codebase with C-u C-x = if you have the data file.]

By the way, you can potentially use matches of char categories in
difficult situations where the syntax isn't good enough.  I tried that
once with Fortran mode and abandonned it for reasons I forget, but I
don't think it has a fundamental problem.




reply via email to

[Prev in Thread] Current Thread [Next in Thread]