Re: Remaining remaining work on SRFI 178
Wolfgang Corcoran-Mathe 23 Aug 2020 15:40 UTC
On 2020-08-22 20:37 +0200, Marc Nieper-Wißkirchen wrote:
> John Cowan <xxxxxx@ccil.org> schrieb am Sa., 22. Aug. 2020, 19:40:
> > As Marc pointed out, these would have to be called /be and /le or the
> > like. I'm inclined to pull them out altogether, since they are not really
> > related to anything else, and put them somewhere else, perhaps named
> > pack-bits/be, pack-bits/le, unpack-bits/be, and unpack-bits/le, and able to
> > handle integers as well as bitvectors.
> >
>
> Okay. Some bytevector API for these low-level things makes sense.
Does this mean that we want to remove the (BE) bytevector conversions
currently in SRFI 178?
> > Well, not quite for that reason. Using -1 for failure from
> > bitvector-first-bit is compatible with the integer equivalent first-set-bit
> > as defined by SRFIs 33, 60, R6RS, and 151. I think that is determinative.
> >
>
> I didn't review those in detail. I agree that it does not make sense to
> change it to a more sensible value just here.
Thanks. I agree.
Everything seems to be ready, otherwise.
--
Wolfgang Corcoran-Mathe <xxxxxx@sigwinch.xyz>
"... if I found the right chair to work in, all compositional
problems would become nonexistant." --Morton Feldman