Steel Bank Common Lisp

Koenraad De Smedt <desmedt@...> writes:

> I am using SBCL on OSX under Apple's X11.

> Problem: upper-ascii characters are not being read; the reader just stops.

> Examples:  "=C3=A6=C3=B8=C3=A5"  '=C3=A6=C3=B8=C3=A5

> This occurs when simply running SBCL in an xterm or via Emacs, using
> iso-latin-1 characters and corresponding Emacs encoding.

> This might be an X11 implementation problem; however, other process
> communication from Emacs does handle iso-latin-1 characters.  In
> Emacs, I can for example start a shell and grep such characters
> without a problem.

Oh, my word.  PAIN AND SUFFERING.

OS X's terminal defaults to utf-8 input/output, so that's what SBCL
does.  How are you setting up your X11 to deal with latin-1
input/output?  (OS X doesn't provide the posix-defined nl_langinfo()
method for querying the system ourselves for natural language
environmental questions).

Cheers,

Christophe

On Jan 17, 2005, at 3:25 PM, Christophe Rhodes wrote:

> Oh, my word.  PAIN AND SUFFERING.
>
> OS X's terminal defaults to utf-8 input/output, so that's what SBCL
> does.  How are you setting up your X11 to deal with latin-1
> input/output?  (OS X doesn't provide the posix-defined nl_langinfo()
> method for querying the system ourselves for natural language
> environmental questions).

It might be slightly more painful than you think, Christophe. OS X 10.3 
provides nl_langinfo, but OS X 10.2 does not. However, 10.3's version 
lies: even when the terminal is set to UTF-8, it returns the equivalent 
of iso latin 1. Since I couldn't figure out a good solution to this 
problem, I decided that always returning UTF-8 was a good enough 
solution, as most other programs on OS X expect and know how to deal 
with UTF-8.

The solution to this problem might just be to use a unicode xterm.
--
Brian Mastenbrook
bmastenb@...
http://cs.indiana.edu/~bmastenb/

Brian Mastenbrook wrote:

> The solution to this problem might just be to use a unicode xterm.

% xterm -u8
open ttydev: Permission denied.
Any hints on where to get a unicode xterm?

On the other hand, I got Emacs to use the proper coding by the following:

(defvar sbcl-buffer-process-coding-system 'utf-8 "UTF8 I/O for SBCL.")
(modify-coding-system-alist 'process "sbcl"
			    sbcl-buffer-process-coding-system)

I suspect that advanced lisp interaction packages like Slime do something similar but I didn't bother to get those.

I found surprisingly little information about character coding in the SBCL manual.

Thanks,

K

Koenraad De Smedt writes:
 > I am using SBCL on OSX under Apple's X11.
 >=20
 > Problem: upper-ascii characters are not being read; the reader just =
stops.
 >=20
 > Examples:  "=E6=F8=E5"  '=E6=F8=E5
 >=20
 > This occurs when simply running SBCL in an xterm or via Emacs,
 > using iso-latin-1 characters and corresponding Emacs encoding.
 >=20
 > This might be an X11 implementation problem; however, other process
 > communication from Emacs does handle iso-latin-1 characters.  In
 > Emacs, I can for example start a shell and grep such characters
 > without a problem.

You can get SBCL to use Latin-1 encoding instead of the default UTF-8,
by running it in an environment where either LC_CTYPE or LANG are set
to "C".

The depressing thing is that even if they fix nl_langinfo for 10.4,
we're pretty much stuck with this mess until 10.5 comes out and it's
reasonable to stop supporting 10.3.  Yuck.  Maybe we can steal the
langinfo code from one of the BSDs :-)