Not so much FILL! [was Re: Also SUBSEQ and FILL!] Donovan Kolbly 14 Mar 2005 14:22 UTC

On Mon, 14 Mar 2005, Michael Sperber wrote:

> >>>>> "Donovan" == Donovan Kolbly <xxxxxx@rscheme.org> writes:
>
> Donovan> I would like to additionally suggest:
>
> Donovan>   (BYTE-VECTOR-FILL! byte-vector start-offset end-offset(or length?) value)
>
> Donovan> for the many times I find myself needing to fill just a portion of a byte
> Donovan> vector.
>
> For analogy with R5RS, this would make sense, and I was thinking about
> including it when I wrote the draft.  However, by leaving it out, I
> acknowledged the fact, that I never ever used BYTE-VECTOR-FILL! (or
> some equivalent) in my own code---and I did want to keep the SRFI
> minimal.
>
> Your experience seems to differ.  Could you provide a couple of
> examples to help convince me?

Ha on me!  You are exactly right.  I thought I did it all the time, but I
grepped around my code base and found that every occurrence would be
served just as well with an initializing fill value (which RScheme lacks,
hence my mistaken impression).

--
-- Donovan Kolbly                    (  xxxxxx@rscheme.org
				     (  http://www.rscheme.org/~donovan/