[R6RS] draft Unicode SRFI

[Matthew Flatt]

[Sorry - I missed this one before.]

At Fri, 8 Jul 2005 05:58:07 +0200, "Manuel Serrano" wrote:
> I'm not sure to understand the issue about CR-LF vs Newline. I don't
> understand the problem and the arguments. I have read about "opening
> a file in binary or text mode". I have also read about "Unix files".
> This looks like to me OS idiosyncrasy appearing in the proposal. Do
> we really have to discuss these points in R6?

I think it's an important practical issue, but in the end, the SRFI
draft says nothing about it, and just assumes that newlines appear in
the right way in the input stream.

> I'm really not in favor of here-strings. 
> [...]
> - May be its because they are imposing constraints on the source code itself?
>   What is the encoding used for expressing here-strings. Is is ascii,
>   iso-latin, ut8, ucsXXX? How do we specify that?

At the parsing level, everything is characters, so decoding has already
taken place. The program gets a string, and if it needs a byte
sequence, it will have to encode it somehow.

> - I don't think that the syntax scales up very well. Let's me illustrate this 
>   with two personal examples:
>   In Scheme programs of my own, at two different places, I'm using "mixed"
>   code (i.e., code that mixes two different syntaxes). First, in Skribe, as 
>   I have said earlier with the [...] syntax. Second in a Web server where
>   I use a {...} syntax for inserting JavaScript codes inside Scheme programs. 
>   May be I'm just unlucky but for these two applications the here-string syntax
>   is of no help for me! 

I added here strings to MzScheme after several requests from users, who
now seem to be happy. I haven't found them to be particularly useful
myself.

Matthew