Beefy Boxes and Bandwidth Generously Provided by pair Networks
good chemistry is complicated,
and a little bit messy -LW

Re: Is reference blessed?

by jeroenes (Priest)
on Nov 23, 2001 at 14:58 UTC ( #127082=note: print w/replies, xml ) Need Help??

in reply to Is reference blessed?

Yes you can find out whether an object is blessed more elegantly (tye and merlyn told me): if UNIVERSAL::isa($ref,'can'){

And as for your type finding, this is from perldoc perlobj:

`isa' returns true if its object is blessed into a subclass of `CLASS' `isa' is also exportable and can be called as a sub with two arguments. This allows the ability to check what a reference points to. Example:
if(UNIVERSAL::isa($ref, 'ARRAY')) {...


"We are not alone"(FZ)

Update: Yeah I screwed up remembering tye's solution. I looked it up (mea culpa):
Is the beast you are looking for.

Update2: D'oh. I took it for granted that you wanted a ref-check anyway, like if( ref $r ){ if ($r->can('isa')){.... Your UNIVERSAL::can($r, 'isa') sounds really neat. I'm curiously awaiting tye's/ merlyn's response(s).

Replies are listed 'Best First'.
Re: Re: Is reference blessed?
by gildir (Pilgrim) on Nov 23, 2001 at 15:29 UTC
    I think it should be
    if (UNIVERSAL::can($ref,'isa')) { ...
    But I see the point now.
    As UNIVERSAL is the parent of every object (blessed reference) and it implements method 'isa', every beast that know how to call isa method must be an object.

    Update: No, $r->can('isa') won't work (as I mentioned in the other reply) because if $r is not blessed, it will die on 'unblessed reference' error. UNIVERSAL::can($r,'isa') is real solution.

    Update2: Even if I do refcheck, situation is the same. I just cannot call any method on non-blessed reference, not even the can() method. Consider $r={}; you cannot do $r->can() because $r is not blessed. And that is what I want to know, if $r is blessed or not. I know that it is a reference. Catching a die exception with eval is TIMTOWDI, but it does not look good to me :-)

      The problem with just UNIVERSAL::can($ref,'isa') is that it can return a true value when $ref isn't even a reference. So you really have to make multiple tests:

      if( ! ref($r) ) { # no reference at all } elsif( ! UNIVERSAL::can($r,'can') ) { # unblessed ref } else { # blessed ref }
      or just:
      if( ref($r) && UNIVERSAL::can($r,'can') ) { # blessed ref }
      if( ref($r) && eval { $r->can('can') } ) { # blessed ref }
      and I can't make a convincing case for one style over the other at the moment.

      And, yes, it would be nice if there were less overloaded versions of these things so that blessed returned the package that a reference was blessed into and ref just always returned the type of thing. That was a certainly a design mistake IMO.

              - tye (but my friends call me "Tye")

Log In?

What's my password?
Create A New User
Node Status?
node history
Node Type: note [id://127082]
[marto]: good morning all, TCIF
[Corion]: Hi marto! Indeed, Crunchy be thanked!
choroba had a Christams party at work yesterday. Slow morning today...
[Corion]: I guess I should not read some discussions on HackerNews - the crowd there is weirdly different - "you have to stay current in tech" (by following all the latest JS trends) is a meme there, but it doesn't really match my experience
[Corion]: At least in my area, experience can beat most new tech anyway because most new tech is just a rehash of things abandoned in the 70s and we still have a lot of that ;)
[Corion]: choroba: Ah, that will be next Friday for $work ;) Drinking enough water is key ;))
[Corion]: But then, maybe that's just my experience with things, and maybe Hacker News is just people much younger who haven't seen more than one tech cycle...
[choroba]: s/water/bear/ and s/enough/too much/
[choroba]: beer

How do I use this? | Other CB clients
Other Users?
Others scrutinizing the Monastery: (6)
As of 2016-12-09 10:17 GMT
Find Nodes?
    Voting Booth?
    On a regular basis, I'm most likely to spy upon:

    Results (150 votes). Check out past polls.