Re: 5.7.X : problem with attributed variables and memberchk ?

On Sat, Oct 04, 2008 at 10:08:59PM +0200, Bart Demoen wrote:
> 
> > ?- A in 0..5, B in 6..7, memberchk(A, [B, C]).
> > false.
> > 
> > 
> > Shouldn't the last one return :
> > A = C,
> > C in 0..5,
> > B in 6..7.
> > 
> > ?
> 
> First a look at:
> 
>        ?- help(memberchk).
>        memberchk(?Elem, +List)
>            Equivalent to member/2, but leaves no choice point.
> 
> This is enough to answer your question: yes, you are correct about
> your expectation.
> 
> In fact, the implementation of pl_memberchk in pl-list.c cuts the
> choicepoint before executing woken goals. It is a bit tricky to get
> this correct (from within C), but currently it doesn't deliver what
> the manual promises I think. If memchk/2 is written in Prolog, you get
> what you expect.
> 
> How did you find this bug ?

It is complicated to explain, in simple :
I have a big program using memberchk and it was not doing what it should 
be doing :>

I tried for a long time understanding what happened, at first I
thought it was the constraints on the variables that were not
compatibles so the results would have been normal, but after a lot
of test (now I know how to make attributes(portray) for
write_term/2 works :) I reduced it to this little example.

The program is a general argumentation framework that combines
preference reasoning, abduction and constraint programming.
I am working on the inclusion of constraint programming to it, you
can take a look at if you want, there is a hg branch for it there :
https://dev.crazydwarves.org/trac/Gorgias

Regards,

Victor