Re: XSRETURN_UNDEF, and the remainder of the stack

On 1999-10-14 19:38, Stephen Zander wrote:
> Jan Dubois:

>> I would argue that functions that *always* return a single
>> value in the non-error mode should return C<undef>  and *not*
>> the empty list in case of an error.
>
> Whereas I would argue that should return *nothing* in case of an
> error.  It's a philisophical question so we could go back and forth
> all day. :)


   my %h = ( a => func('a'), b => func('b' );

Now suppose func('a') sometimes returns the empty list,
so all code then becomes like this:

   my %h = ( a => ( func('a') // undef ), b => func('b' );

Etc. etc.

-- 
Ruud (is obviously with Jan)