From jedediah at jdhuntington.com  Thu Mar 20 00:38:23 2008
From: jedediah at jdhuntington.com (JD Huntington)
Date: Wed, 19 Mar 2008 21:38:23 -0700
Subject: [Bus-scheme-discuss] Character Literals
Message-ID: <99F60D00-9175-4448-87BB-28CDECDC81D4@jdhuntington.com>

I did a bit of looking into how character literals are handled in some  
other versions of scheme, and thought it might be useful for  
discussion. As of now, bus-scheme character literals are parsed into  
one-character strings. However, it appears this is not the case in at  
least one other scheme implementation. Here's a snippet from PLT:

 > (char? #\a)
#t
 > (char? "a")
#f

In bus, the second form is true as well:

 > (char? #\w)
true
 > (char? "w")
true

Thoughts?

As an aside, I think I'll make a lookup for the "control" characters.  
I found a list of these for the guile scheme implementation; does this  
look like an adequate list?
http://gnu.cict.fr/software/guile/manual/html_node/Characters.html

-JD Huntington

From phil at hagelb.org  Thu Mar 20 12:13:57 2008
From: phil at hagelb.org (Phil Hagelberg)
Date: Thu, 20 Mar 2008 09:13:57 -0700
Subject: [Bus-scheme-discuss] Fwd:  Character Literals
In-Reply-To: <5F08C562-D802-4C1A-B6CC-F75B73A0853C@jdhuntington.com> (JD
	Huntington's message of "Wed, 19 Mar 2008 22:17:08 -0700")
References: <99F60D00-9175-4448-87BB-28CDECDC81D4@jdhuntington.com>
	<5F08C562-D802-4C1A-B6CC-F75B73A0853C@jdhuntington.com>
Message-ID: <87ve3hxtiy.fsf@hagelb.org>

JD Huntington <jedediah at jdhuntington.com> writes:

>     I did a bit of looking into how character literals are handled in some  
>     other versions of scheme, and thought it might be useful for  
>     discussion. As of now, bus-scheme character literals are parsed into  
>     one-character strings. However, it appears this is not the case in at  
>     least one other scheme implementation. Here's a snippet from PLT:
>
>         (char? #\a)
>    
>     #t
>    
>         (char? "a")
>    
>     #f

Bugger it all. I just checked R5RS, and it says (under section 3.2,
Disjointness of Types) that "No object satisfies more than one of the
following predicates:

boolean?   
pair?
symbol?    
number?
char?      
string?
vector?    
port?
procedure?".

So the standard is quite clear; it's just a question of whether we want
to follow Scheme or Ruby here. I personally prefer Ruby 1.9's approach
(Ruby 1.8's being, of course, totally brain-dead). I think we can leave
it how it is unless there are objections.

-Phil