#<Object:0x2c1c8f0> <--- why is the displayed number == object_id*2

[Meinrad Recheis]

irb(main):001:0> VERSION
=> "1.8.5"
irb(main):002:0> o=Object.new
=> #<Object:0x2c1c8f0>
irb(main):003:0> ObjectSpace._id2ref 0x2c1c8f0
RangeError: 0x2c1c8f0 is not id value
from (irb):3:in `_id2ref'
from (irb):3
irb(main):004:0> ObjectSpace._id2ref 0x2c1c8f0/2
=> #<Object:0x2c1c8f0>
irb(main):005:0>

I think that is pretty confusing. I am sure there is a good reason for
this. But, don't you think it'd be better to display the real
object_id instead of the doubled value?

-- henon

 

[Robert Klemme]

irb(main):001:0> VERSION
=> "1.8.5"
irb(main):002:0> o=Object.new
=> #<Object:0x2c1c8f0>
irb(main):003:0> ObjectSpace._id2ref 0x2c1c8f0
RangeError: 0x2c1c8f0 is not id value
from (irb):3:in `_id2ref'
from (irb):3
irb(main):004:0> ObjectSpace._id2ref 0x2c1c8f0/2
=> #<Object:0x2c1c8f0>
irb(main):005:0>

I think that is pretty confusing. I am sure there is a good reason for
this. But, don't you think it'd be better to display the real
object_id instead of the doubled value?

The output of #inspect is not meant to give you any means to get at the
original object although it may seem so at times. Having said that,
#object_id and #inspect are completely unrelated and it's just an
implementation artifact that some classes actually return something that
resembles the object id.

As far as I remember this list you are the first one that mentions this.
So it's probably not that big an issue (at least not as confusing as
"singleton class" )

Kind regards

robert

 

[Meinrad Recheis]

The output of #inspect is not meant to give you any means to get at the
original object although it may seem so at times. Having said that,
#object_id and #inspect are completely unrelated and it's just an
implementation artifact that some classes actually return something that
resembles the object id.

As far as I remember this list you are the first one that mentions this.
So it's probably not that big an issue (at least not as confusing as
"singleton class" )

Kind regards

robert

Robert,
Thanks for your answer. I came across it while debugging a running
application via an eval console. It's not really a problem because I
know that I need to divide the displayed id by 2 to get the object. I
posted it just because ruby didn't work as I expected it to ... which
is completely against the principles of the language.

cheers,
-- henon

 

[Robert Klemme]

Thanks for your answer. I came across it while debugging a running
application via an eval console. It's not really a problem because I
know that I need to divide the displayed id by 2 to get the object.

Well, but that won't work for all cases.

I
posted it just because ruby didn't work as I expected it to ... which
is completely against the principles of the language.

That's only one way to look at it. The other way is of course to assume
that your expectations were "wrong" (whatever that means).

Kind regards

robert

 

[Rick DeNatale]

Well, but that won't work for all cases.

Nor for all versions of ruby. The relationship between object_id and
the VALUE representation for various classes has been changing even
within the 1.8.x stream. $Diety only knows what implementations like
JRuby, rubinius, IronRuby etc would use for inspect which is the
method which produces what you see after the ==> in irb, or call
Kernel#p

Inspect is intended to be used for the 'programmer' representation of
an object, to help activities like debugging. So in the case of
classes like Object that representation will likely have some info
useful to the current internal implementation, and unless it's
documented somewhare, there's no guarantee that it won't change.

 

[james.d.masters]

(The reason, presumably, is that the ID is left-shifted by one, because
objects with a 1 in the low-order bit are fixnums.)

When I came across this behavior, this was also my best guess as the
reason.