Re: Newbee question

[Ole-Hjalmar Kristensen <oleh@vlinux.voxelvision.no>]

"Marin David Condic" <mcondic.auntie.spam@acm.org> writes:

<snip>

> 
> Well, my observations have come about from trying to interface along
> communications paths with other things that are not necessarily Ada.
> Sometimes things overlap byte, word or longword boundaries. Having a
> compiler tell me I can't do that instantly makes the rep-clause useless to
> me because the only way I can guarantee that I can handle all the cases that
> come up - or ever may come up - is to go to a byte array and start picking
> out the bits myself, ala C. Maybe my problems aren't the most common ones,
> but it happens enough to make one wish that the rep clause could be used to
> arbitrarily overlay a contiguous stream of bits without restrictions - or
> you start asking yourself "Why bother having a rep clause?" The C answer is
> "We don't have a rep clause - go pick the bits apart yourself." The Ada
> answer seems to have been "Here. Let me help you with all that bit
> twiddling. But only if you agree not to twiddle all the bits you might want
> to twiddle..."
> 
> I'll give you the compiler efficiency issues. But that just suggests that
> perhaps some other construct is necessary for the language to be able to
> express representation and then deal with packing/unpacking. I'd be happy if
> I could somehow express the record representation in some manner that let me
> put the record on the wire or take the record off the wire and then do some
> kind of "Pack/Unpack" operations if I wanted to do anything computational. I
> don't have a prepackaged answer, but it is something to think about...
> 
> MDC
> --
> ======================================================================
> Marin David Condic
> I work for: http://www.belcan.com/
> My project is: http://www.jast.mil/
> 
> Send Replies To: m c o n d i c @ a c m . o r g
> 
>     "I'd trade it all for just a little more"
>         --  Charles Montgomery Burns, [4F10]
> ======================================================================
> 
> 

I picked up this on the net by chance, perhaps Lutz could comment if
this could be an answer to such problems? :

        16.00 Workshop 1

        Generic Bidirectional Mapper
        Vorgestellt wird ein Yacc-Nachfolger, der ausreichen komplex ist, 
        um die praktischen externen Datenstrukturen in einem Ritt zu lesen 
        und zu schreiben.
        Lutz Donnerhacke


-- 
Ole-Hj. Kristensen

******************************************************************************
* You cannot consistently believe this sentence.
******************************************************************************