Re^3: goto superclass method

Well, calling can directly can break if you have something in @ISA that isn't a class (such as a coderef: Class::Dynamic). Out of habit I try to call UNIVERSAL::can unless the $object->can('can');

Perhaps I'm being overly paranoid.

Cheers,
Ovid

New address of my CGI Course.

Comment on Re^3: goto superclass method

Replies are listed 'Best First'.
Re^4: goto superclass method by chromatic (Archbishop) on Dec 22, 2004 at 21:53 UTC
How do you know in general whether the object or class has its own `can` though? (Easy answer, but not one you'll like.) I prefer to use Scalar::Util's `blessed` to see if I can invoke operations on an object, though wrapping the call in an `eval` block is safer if you expect class names as well.	[reply] [d/l] [select]
Re^5: goto superclass method by Ovid (Cardinal) on Dec 22, 2004 at 22:19 UTC
I guess in this case it would be preferable to call call can as a method and let the code die and then worry about what to do if it fails. That's probably better than my silently failing or doing the wrong thing because of an overridden can. Cheers, Ovid New address of my CGI Course.	[reply]
Re^6: goto superclass method by Aristotle (Chancellor) on Dec 22, 2004 at 22:37 UTC
My thinking is that if a superclass has overriden `can()`, then they have good reason to, so in a situation where you know you have a blessed ref or a package name, you should prefer calling it as a method. I only use `UNIVERSAL::can` directly when I'm unsure about that. Makeshifts last the longest.	[reply]
Re^4: goto superclass method by Joost (Canon) on Dec 22, 2004 at 21:52 UTC
This is the first time i've ever seen an excuse for using UNIVERSAL::can directly, but I still don't like it :-) Doesn't this just mean that putting coderefs in your @ISA is just breaking things anyway? "What should it profit a man, if he should win a flame war, yet lose his cool?"	[reply]


"be consistent"
	PerlMonks