Re: end-of-defun is fubsr.

[Alan Mackenzie <acm@muc.de>]

On Sun, Feb 15, 2009 at 02:26:57PM -0500, Stefan Monnier wrote:
> >> I prefer the current text.
> > Was my patch really that bad?

> I'm not saying it's bad, I'm saying that I find it less good than the
> current text.  Most of the differences are nitpicks on which
> we disagree for reasons explained over and over in this thread.

The current text doesn't specify an EOD-F sufficiently to enable a major
mode writer to construct one.  He is forced to read the source in
lisp.el and work it out for himself.  This is a Bad Thing.

> > Look, I'm as sick of all this as you must be.

> AFAIK the only thing still missing is the addition of ARG to
> EOD-function, which is something I prefer to postpone because it's not
> a bug fix and we're in pretest.

There's a lot more "missing" than that.

With all due respect, a change this big should have been discussed on
emacs-devel long before we got to the pretest.  The only discussion I
think there's been so far was in the thread "Useless change in lisp.el?"
started by Andreas Röhler in November 2007.  Your verdict at the time
was, and I quote: "Rather than argue abuot the change itself, just give
us some use case where the new behavior is problematic."

We have now seen such a use case, if not two or three.

Why did this change not get proper discussion on emacs-devel?  Or did I
miss it?

I'd disagree with you about the bug fixness of the thing, but the last
thing we need to do is squabble about what words mean.  The effort
needed to fix it now is of a lesser order of magnitude than what will be
needed later if the current implementation is released in Emacs 23.  In
fact, the effort to fix it now is probably less than what the two of us
have just spent composing emails on the topic.

A suggestion: Let's just revert to the Emacs 22 code and documentation.
That was at least stable and consistent, and reverting won't introduce
new problems.  The way things are at the moment will introduce new
problems for people outside the core Emacs team.  It has already done so
for people inside the team.

If we release with lisp.el the way it currently is in the CVS, the
problems it will cause will be unfixable in the future, or at least
fixable only with an inordinate degree of uncreative effort.  I suspect
a lot of major mode maintainers just won't bother.

Then we can discuss things properly and fix them properly for Emacs 24.

>         Stefan

-- 
Alan Mackenzie (Nuremberg, Germany).