Dave Rolsky wrote:
>
> On Thu, 4 Mar 2004 [EMAIL PROTECTED] wrote:
>
> > Accepting 'undef' in a map operation is ok, but a Set cannot contain
> > 'undef' !
>
> The problem is that if the callback is called in a scalar context, and it
> does a bare return, like this:
>
> $set->iterate( sub { return } )
>
> then that "return" returns an undef. In fact, I think you don't even need
> a bare return. Simply following off the end of the subroutine would do
> the trick:
>
> sub foo { if ( $_[0] > $x ) { ... } }
>
> so if \&foo were the callback, and the argument wasn't greater than $x,
> then it would return undef!
>
> > So we only have two options if we get an undef: to die, or to discard
> > the value.
>
> I think we need to die and provide a different way to remove elements,
> probably a separate method.
>
> > > just to avoid the effect. I think if you want to be able to modify
> > > the set, you should provide different methods for that. ->map and
> > > ->grep come to mind.
> >
> > People are already confused with lists vs. sets.
> >
> > I think using map and grep would lead them to think sets _are_ lists.
>
> True, but the map/grep analogue would make the functionality obvious. How
> about set_grep() and set_map()?
Looks good. How about:
* set_map
$set2 = $set->set_map( sub { return } );
# same as $set->clone
$set2 = $set->set_map( sub { $_[0]->add( hours => 1 ) } );
# returns a new set, which is offset by 1 hour; $set is not modified.
$set2 = $set->set_map( sub {
$_[0]->add( hours => 1 );
return;
}
);
# same as before.
* set_grep
$set2 = $set->set_grep( sub { 1 } );
# same as $set->clone
$set2 = $set->set_grep( sub { return } );
# make an empty set
$set2 = $set->set_grep( sub { $_[0]->day > 15 } );
# makes a new set, all set elements have day > 15
In DT::Set, $_[0] is a DateTime.
In DT::SpanSet, $_[0] is a DateTime::Span.
- Flavio S. Glock
