Steel Bank Common Lisp

On Wed, Jan 28, 2009 at 10:35 AM, Leslie P. Polzer
<sky@...:

>
> I have two CLOS objects that seem to be pretty identical
> -- same print output and same address.
>
> Yet EQ returns NIL when they are compared.
>
> How is this possible?


Is it a quiz?

> Is it a quiz?

No, I'm asking for internal criteria that could make
two CLOS objects that are apparently identical not EQ.

-- 
LinkedIn Profile: http://www.linkedin.com/in/polzer
Xing Profile: https://www.xing.com/profile/LeslieP_Polzer
Blog: http://blog.viridian-project.de/

On Wed, 28 Jan 2009 10:42:24 +0100, Pierpaolo Bernardi wrote:

> Is it a quiz?<div class="gmail_quote">On Wed, Jan 28, 2009 at 10:35 AM,
> Leslie P. Polzer <span dir="ltr">&lt;<a

Would you mind sending e-mails to this list in plaintext?

Tamas

http://www.lispworks.com/documentation/HyperSpec/Body/f_eq.htm

"Pointer equality" as it were.

On Wed, Jan 28, 2009 at 9:48 AM, Leslie P. Polzer
<sky@...:

>
> > Is it a quiz?
>
> No, I'm asking for internal criteria that could make
> two CLOS objects that are apparently identical not EQ.
>
> --
> LinkedIn Profile: http://www.linkedin.com/in/polzer
> Xing Profile: https://www.xing.com/profile/LeslieP_Polzer
> Blog: http://blog.viridian-project.de/
>
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> SourcForge Community
> SourceForge wants to tell your story.
> http://p.sf.net/sfu/sf-spreadtheword
> _______________________________________________
> Sbcl-help mailing list
> Sbcl-help@...
> https://lists.sourceforge.net/lists/listinfo/sbcl-help
>

2009/1/28 Leslie P. Polzer <sky@...>:
>
> I have two CLOS objects that seem to be pretty identical
> -- same print output and same address.
>
> Yet EQ returns NIL when they are compared.
>
> How is this possible?

It shouldn't be. Do you have a reproducible example?

You may also try using the following sanity-checking function:

  (defun sanity-check (obj1 obj2)
    (declare (optimize (speed 1) (safety 1) (debug 1)))
    (let ((addr1 (sb-kernel:get-lisp-obj-address obj1))
	  (addr2 (sb-kernel:get-lisp-obj-address obj2)))
      (format t "~&obj1 is at ~X~%obj2 is at ~X~%(eq obj1 obj2) => ~S~%"
	      addr1 addr2 (eq obj1 obj2))))

> 	  (addr2 (sb-kernel:get-lisp-obj-address obj2)))

That sheds a bit of light:

EQM(19): one

#<OBJ {AAFD571}>>
EQM(20): two

#<OBJ {AAFD571}>>

EQM(24): (sb-kernel:get-lisp-obj-address one)

199984081
EQM(25): (sb-kernel:get-lisp-obj-address two)

179295617

Both ONE's and TWO's hex display corresponds to the second value.

  Leslie

-- 
LinkedIn Profile: http://www.linkedin.com/in/polzer
Xing Profile: https://www.xing.com/profile/LeslieP_Polzer
Blog: http://blog.viridian-project.de/

> Does destructive modification of one visibly alter the other?

No.

-- 
LinkedIn Profile: http://www.linkedin.com/in/polzer
Xing Profile: https://www.xing.com/profile/LeslieP_Polzer
Blog: http://blog.viridian-project.de/

On Jan 28, 2009, at 6:10, Leslie P. Polzer wrote:

>
>> 	  (addr2 (sb-kernel:get-lisp-obj-address obj2)))
>
> That sheds a bit of light:
>
> EQM(19): one
>
> #<OBJ {AAFD571}>>
> EQM(20): two
>
> #<OBJ {AAFD571}>>
>
> EQM(24): (sb-kernel:get-lisp-obj-address one)
>
> 199984081
> EQM(25): (sb-kernel:get-lisp-obj-address two)

Is there any chance that 'one' has a PRINT-OBJECT method which would  
be printing 'two'? The extra #\> is particularly suspicious.

It might also be useful to consult TYPE-OF, CLASS-OF, and INSPECT for  
the two objects.

-- 
Kevin Reid                            <http://homepage.mac.com/kpreid/>

"Leslie P. Polzer" <sky@...> writes:

>> 	  (addr2 (sb-kernel:get-lisp-obj-address obj2)))
>
> That sheds a bit of light:
>
> EQM(19): one
>
> #<OBJ {AAFD571}>>

This isn't the default print-object method: look at that extra ">".
How certain are you that your print-object method is actually printing
the address of the object?

Cheers,

Christophe

> This isn't the default print-object method: look at that extra ">".
> How certain are you that your print-object method is actually printing
> the address of the object?

I use this printer:

(defmethod print-object ((object obj) stream)
  (print-unreadable-object (object stream :type t)
    (ignore-errors (format stream ...))))

-- 
LinkedIn Profile: http://www.linkedin.com/in/polzer
Xing Profile: https://www.xing.com/profile/LeslieP_Polzer
Blog: http://blog.viridian-project.de/

> What's in the ...? If you're not using :identity t in the
> print-unreadable-object options, the hex value being printed is coming
> from your own format string.

Ah, I see. That explains it, thanks everyone!

  Leslie

-- 
LinkedIn Profile: http://www.linkedin.com/in/polzer
Xing Profile: https://www.xing.com/profile/LeslieP_Polzer
Blog: http://blog.viridian-project.de/

Leslie P. Polzer wrote:
> I use this printer:
> 
> (defmethod print-object ((object obj) stream)
>   (print-unreadable-object (object stream :type t)
>     (ignore-errors (format stream ...))))

What's in the ...? If you're not using :identity t in the 
print-unreadable-object options, the hex value being printed is coming 
from your own format string.
-- 
Brian Mastenbrook
brian@...
http://brian.mastenbrook.net/

Does destructive modification of one visibly alter the other?

On Wed, Jan 28, 2009 at 11:10 AM, Leslie P. Polzer
<sky@...> wrote:
>
> >         (addr2 (sb-kernel:get-lisp-obj-address obj2)))
>
> That sheds a bit of light:
>
> EQM(19): one
>
> #<OBJ {AAFD571}>>
> EQM(20): two
>
> #<OBJ {AAFD571}>>
>
> EQM(24): (sb-kernel:get-lisp-obj-address one)
>
> 199984081
> EQM(25): (sb-kernel:get-lisp-obj-address two)
>
> 179295617
>
> Both ONE's and TWO's hex display corresponds to the second value.
>
>  Leslie
>
> --
> LinkedIn Profile: http://www.linkedin.com/in/polzer
> Xing Profile: https://www.xing.com/profile/LeslieP_Polzer
> Blog: http://blog.viridian-project.de/
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> SourcForge Community
> SourceForge wants to tell your story.
> http://p.sf.net/sfu/sf-spreadtheword
> _______________________________________________
> Sbcl-help mailing list
> Sbcl-help@...
> https://lists.sourceforge.net/lists/listinfo/sbcl-help

Just because I feel the need to say something... <smile>

I think your question is not about what #'eq means (pointer equality)  
but why your two objects which appear (from the print-object method)  
to have the same address are not eq. Is that right?

Since it's unlikely #'eq is wrong (I think that the universe would  
grind to a halt if it was), it's likely that something is awry with  
your print-object method. Others have already provided means to  
determine more about what SBCL is thinking and that's the way forward,  
I think.

OTOH (or on the additional hand), #'describe might also be your friend.

HTH


On Jan 28, 2009, at 4:35 AM, Leslie P. Polzer wrote:

>
> I have two CLOS objects that seem to be pretty identical
> -- same print output and same address.
>
> Yet EQ returns NIL when they are compared.
>
> How is this possible?
>
>  Thanks,
>
>    Leslie
>
> -- 
> LinkedIn Profile: http://www.linkedin.com/in/polzer
> Xing Profile: https://www.xing.com/profile/LeslieP_Polzer
> Blog: http://blog.viridian-project.de/
>
>
> ------------------------------------------------------------------------------
> This SF.net email is sponsored by:
> SourcForge Community
> SourceForge wants to tell your story.
> http://p.sf.net/sfu/sf-spreadtheword
> _______________________________________________
> Sbcl-help mailing list
> Sbcl-help@...
> https://lists.sourceforge.net/lists/listinfo/sbcl-help

--
Gary Warren King, metabang.com
Cell: (413) 559 8738
Fax: (206) 338-4052
gwkkwg on Skype * garethsan on AIM * gwking on twitter