Re: bitstreams

["Pat Rogers" <progers@NOclasswideSPAM.com>]

<mark_lundquist@my-deja.com> wrote in message
news:9420qv$2d$1@nnrp1.deja.com...
> In article <9406cs$g4i$1@nnrp1.deja.com>,
>   Robert Dewar <robert_dewar@my-deja.com> wrote:
> > In article <93vik8$u08$1@nnrp1.deja.com>,
> >   mark_lundquist@my-deja.com wrote:
> > > Note that shift and rotate functions are primitive operations
> > > of the Signed_<n> and Unsigned_<n> families of types (RM B.2)
> >          ^^^^^^^^^^
> >
> > A surprising mistake :-)
>
> You got that right... it sure suprised me!
>
> Sloppy, sloppy.  Yesterday was NOT a good news-posting day for me...
>
> :-) :-)
>
> > It is of course the case that these
> > functions do NOT apply to signed numbers (where there meaning
> > would be obscure or at least implementation defined, because
> > of the sign bit issue).
> >
> > One interesting bit of non-portability comes from whether it
> > is possible to define these operators for *USER* defined
> > unsigned types. The RM does not require this capability, but
> > in GNAT you can define your own intrinsic shift functions for
> > your own modular types, and GNAT will permit this. I do not
> > know if other Ada 95 compilers provide this capability.
>
> That is interesting...

When I want one of my own, I always derive from those in Interfaces to
inherit the primitives.