[R6RS] draft Unicode SRFI

[Marc Feeley]

On 1-Jul-05, at 9:18 AM, Mathyew Platt wrote:

> <r6rs-srfi-unicode.html>

I just noticed an inconsistency with the string escapes.  Here's a  
table comparing some of the C/Java string escapes and the proposed  
string escapes:

   1)  \'        proposed in the SRFI and exists in C/Java
   2)  \e        proposed in the SRFI and does not exist in C/Java
   3)  \v        proposed in the SRFI and exists in C but not in Java
   4)  \a        proposed in the SRFI and exists in C but not in Java
   5)  \?    not proposed in the SRFI and exists in C but not in Java

\' exists in C/Java for a good reason, but it is not required in  
Scheme, so why not apply the same reasoning to \?.  Note also that  
C99 specifies a 4 hex digit \u, and an 8 hex digit \U.

I think we should iron out the differences, or give a better  
rationale for departing from the C syntax.  I propose:

   1) drop \e (and to be consistent, #\esc)
   2) keep \' \v \a
   3) add \?
   4) require 8 hex digits after \U  (I think \U will occur very  
infrequently
      in code, so we might as well make it compatible with C99,  
Python and other
      languages).  The value still has to fall in 0..#x10ffff.

Marc