"Randal L. Schwartz" wrote: > >>>>> "R" == R Joseph Newton <[EMAIL PROTECTED]> writes: > > R> "Randal L. Schwartz" wrote: > >> >>>>> "Dan" == Dan Anderson <[EMAIL PROTECTED]> writes: > >> > Dan> my $class = ref($proto) || $proto; > >> > >> Don't do this! > >> > >> <http://www.perlmonks.org/index.pl?node_id=52089>. > > R> Hi Randal, > > R> Read the article, and it percursor thread. I'm still a little mystified as to > R> what you find offensive there. Perhaps one should be aware of the contents, if > R> any, of @ISA, before writing the new method. It doesn't always work that way, > R> though, and having a flexible [and standardized] constructor does not seem to do > R> any harm. Is there some potential harm you see? > > Sorry you missed the point then. > > The point is that > > $instance->new > > could mean either "clone", or "make one of the same class as". You > don't need it for "make one of the same class as", because you've got: > > (ref $instance)->new > > to do that explicitly. And if you really wanted that to do clone, > CALL IT CLONE, don't call it ->new. > > It obscures more than it clarifies, and hence is a *bad* name > for an instance method. > > And because it's a bad name for an instance method, it should > not respond as an instance method. That simplifies the "standard > constructor immensely", to something like: > > sub new { > my $class = shift; > my $self = { ... }; > other fixups > return bless $self, $class; # return is optional here > } > > This makes your code both clearer, and faster.
Got it. Thanks. You know, I never would have imagined anybody trying to use a constuctor to clone, except perhaps by passing an object reference in as an argument. I had asumed that it had something to do with importing parent classes. Thanks for clearing that up. I guess I'll do some inheritance experiments to see how the constructor works without that construct. Joseph -- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
