Re: JACAL, scm

[Marius Vollmer]

Dirk Herrmann <address@hidden> writes:

> On 18 Oct 2001, Neil Jerram wrote:
> 
> > >>>>> "Radey" == Radey Shouman <address@hidden> writes:
> > 
> >     Radey> Please see the "Programs and Definitions" section of the
> >     Radey> formal syntax, where we see
> > 
> >     Radey>    <definition> --> (begin <definition>*)
> > 
> >     Radey> Which indicates that (BEGIN) should be allowable anywhere
> >     Radey> definitions may appear.
> > 
> > This seems convincing enough.  Do you (everyone) think it's acceptable
> > to fix this such that (begin) is allowed anywhere, or do you think we
> > need to fix it so that it's allowed for definitions but disallowed for
> > non-definition expressions?
> 
> I don't mind either way.  However, as long as the issue is not decided, a
> test should be added to the test suite that marks the appearances of
> (begin) in non-definition expressions as an expected failure - a comment
> should be added to the test case to describe the issue.

I think we should allow `(begin)' everywhere, and make `(begin)' in a
non-definition context return #<unspecified>.

I will do that in `unstable'.