[ESS-bugs] Re: [ESS] hitting return won't move the cursor down

[Martin Maechler]

>>>>> "StEgl" == Stephen Eglen <S.J.Eglen at damtp.cam.ac.uk>
>>>>>     on Thu, 24 Mar 2005 09:44:02 +0000 writes:

    StEgl> hi Martin, tiago, AFter Tiago sent the report in, I
    StEgl> verified that the error also happened in my xemacs.
    StEgl> I was able to reproduce it in the following fashion.

    StEgl> in Foo.R buffer, try x <- "some very long string that
    StEgl> will line break" (as Kurt suggested before a few
    StEgl> weeks ago; I think you either put auto-fill-mode on
    StEgl> or just try to reformat with M-q) and then instead of
    StEgl> breaking the line as it should, I got the error about
    StEgl> the missing var.

    StEgl> Stephen

Ah, that's very interesting:

I've now just tried to old Xemacsn (21.4.8 and 21.4.9) and a new
Xemacs (21.4.17) on two
different Linux distros, and on  both, I couldn't reproduce the
problem:  the auto filling just worked fine

--> then I checked: 
    
    For both Xemacsen, the  enable-multibyte-characters variable
    was set to 't' for me, from the start, even when I started

 Xemacs -q --no-site-file    

I also notice that both these Xemacsen display "ISO8" in the
very left of the "status bar" (the "bar" above the mini buffer),
which tells me that Xemacs somehow knows that I'm not an
ASCII-only person, even without any initialization files.

Maybe it uses LC_* environment variables or something?
No, that can't be it, since on one platform I haven't set any...

In any case, it seems that Xemacs *does* know about the concept
of and `enable-multibyte-characters', as a matter of fact, in my
Xemacs 21.4.9, doing "C-h v" on this gives

 >> `enable-multibyte-characters' is a built-in boolean variable.
 >> 
 >> Value: t
 >> 
 >> Documentation:
 >> Emacs Compatible; Unimplemented
 >> 
 >> Non-nil means the buffer contents are regarded as multi-byte form
 >> of characters, not a binary code.  This affects the display, file I/O,
 >> and behaviors of various editing commands.
 >> 
 >> Setting this to nil does not do anything.
    ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^

but I still wonder how you guys get a version of Xemacs which
doesn't have this defined..

Note however that my workaround in  .../ess/lisp/ess-emcs.el

;; XEmacs needs this
(if (not (boundp 'enable-multibyte-characters))
    (defvar enable-multibyte-characters nil
      "Non-nil means the buffer contents are regarded as multi-byte characters.
 This concept is handled completely differently on Xemacs."))

will hopefully be sufficient --- waiting for Tiago's feedback on
this.

BTW:  Tomorrow is Good Friday and we have general Easter holiday
      here - holidays from Friday till (including!) Monday.

So don't expect too much from my side.
Martin