Am Freitag, den 16.01.2009, 16:31 -0500 schrieb Raymond Toy:
> Dieter Kaiser wrote:
> > Today, I have checked out a check to prevent an assignment to a symbol which is
> > declared to be a SYSCONST. This is related to the bug report SF[783847]
> > "assignments to %i, inf, ...".
> >
> > All works well, but I have seen to late that we have got a problem with the test
> > file rtest_intensor.mac from the share_testsuite.
> >
> >
> Also, what about %o1 and friends? Should I be able to assign to them?
> It works now:
>
> (%i1) sin(1.0);
> (%o1) .8414709848078965
> (%i2) %o1:42;
> (%o2) 42
> (%i3) asin(%o1);
> (%o3) asin(42)
>
> Likewise for %i.
Thank you very much for your fast response.
Perhaps, you have recognized that I try to work on the "old" bugs. I
think there too much of them.
The change to prevent an assignment to the symbols %i, inf, minf,
infinity, ind and und was already proposed in 2003.
I had a further look at the complete code. I have found an assignment to
IND only in the file share/algebra/charset/charsets.mac (95 matches) and
share/algegra/charset/test1.mac (2 matches).
I am not really sure what to do. It seems to me natural to prevent the
assignment to a system constant. Should we allow the assignment to
system constants and close the bug report anyway?
Another possible would be to prevent the assignment only for %i.
Dieter Kaiser