[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: floating point and other comments



On Saturday 20 December 2003 09:56 am, Alex Shinn wrote:
> One minor question, what case does ~x use for hexstrings?  R5RS
> doesn't specify this for number->string.  An interesting extension
> some schemes use is to use the case of the x (i.e. distinguish between
> ~x and ~X).

I would prefer that the case of ~X matches the case of the hex characters, but 
can't specify more than R5RS here.  Note that the reference implementation 
maes use of NUMBER->STRING.


> Also, the spec says it is an error to pass fewer arguments than
> needed, but is it ok to pass more arguments?

I consider it useful to detect this as an error, but there are other sensible 
ideas.  E.g. one possible return value for FORMAT is the list of unused 
arguments.

I prefer that FORMAT invoke ERROR rather than failing and returning #F, as 
SLIB specifies.

However, I prefer to leave FORMAT's return value unspecified in srfi-48 and 
come to concensus on an advanced format srfi.

Cheers,
-KenD