bug#5082: 23.1.50; print-circle and make-hash-table

bug#5082: 23.1.50; print-circle and make-hash-table

[川幡太一]

Dear Emacs Developers,

I've found an Emacs behavior that seems a bug concerning to `print-circle'
and make-hash-table, so I would like to report it.

By setting `print-circle' to `t', printing Lisp Object with recursive
or repetitive
list should be displayed as #N= and #N# syntax.

However, it seems this is not true when printing lisp objects within
hash tables
(which become possible in Emacs 23.1.50.)  Following is an example.

Is it possible to display the repetitive lisp object within printed hash tables
by #N= and #N# syntax, when `print-circle' is set to `t'?

I notify this behavior as a bug, but it may not be so.  I apologize in that
case.

(let ((print-circle t)
      (x (list 10 20))
      (y (make-hash-table)))
  (insert (format "%S\n" (list x x)))
  (puthash 1 x y)
  (puthash 2 x y)
  (insert (format "%S\n" y))
  (setcar x 3)
  (insert (format "%S\n" y)))

Result:
> (#1=(10 20) #1#)
> #s(hash-table size 65 test eql rehash-size 1.5 rehash-threshold 0.8
> data (1 (10 20) 2 (10 20)))
> #s(hash-table size 65 test eql rehash-size 1.5 rehash-threshold 0.8
> data (1 (3 20) 2 (3 20)))
nil

Cordially Yours,

-- 
---------------------------------------------------------------------
　KAWABATA, Taichi　　　E-mail: kawabata.taichi@gmail.com

bug#5082: marked as done (23.1.50; print-circle and make-hash-table)

[Emacs bug Tracking System]

[-- Attachment #1: Type: text/plain, Size: 897 bytes --]

Your message dated Mon, 30 Nov 2009 16:08:38 -0500
with message-id <jwv8wdnencg.fsf-monnier+emacsbugreports@gnu.org>
and subject line Re: bug#5082: 23.1.50; print-circle and make-hash-table
has caused the Emacs bug report #5082,
regarding 23.1.50; print-circle and make-hash-table
to be marked as done.

This means that you claim that the problem has been dealt with.
If this is not the case it is now your responsibility to reopen the
bug report if necessary, and/or fix the problem forthwith.

(NB: If you are a system administrator and have no idea what this
message is talking about, this may indicate a serious mail system
misconfiguration somewhere. Please contact owner@emacsbugs.donarmstrong.com
immediately.)


-- 
5082: http://emacsbugs.donarmstrong.com/cgi-bin/bugreport.cgi?bug=5082
Emacs Bug Tracking System
Contact owner@emacsbugs.donarmstrong.com with problems

[-- Attachment #2: Type: message/rfc822, Size: 4125 bytes --]

From: 川幡太一 <kawabata.taichi@gmail.com>
To: emacs-pretest-bug@gnu.org
Subject: 23.1.50; print-circle and make-hash-table
Date: Tue, 1 Dec 2009 00:09:30 +0900
Message-ID: <270879b0911300709o29a12798m2ae6c6194c25e53@mail.gmail.com>

Dear Emacs Developers,

I've found an Emacs behavior that seems a bug concerning to `print-circle'
and make-hash-table, so I would like to report it.

By setting `print-circle' to `t', printing Lisp Object with recursive
or repetitive
list should be displayed as #N= and #N# syntax.

However, it seems this is not true when printing lisp objects within
hash tables
(which become possible in Emacs 23.1.50.)  Following is an example.

Is it possible to display the repetitive lisp object within printed hash tables
by #N= and #N# syntax, when `print-circle' is set to `t'?

I notify this behavior as a bug, but it may not be so.  I apologize in that
case.

(let ((print-circle t)
      (x (list 10 20))
      (y (make-hash-table)))
  (insert (format "%S\n" (list x x)))
  (puthash 1 x y)
  (puthash 2 x y)
  (insert (format "%S\n" y))
  (setcar x 3)
  (insert (format "%S\n" y)))

Result:
> (#1=(10 20) #1#)
> #s(hash-table size 65 test eql rehash-size 1.5 rehash-threshold 0.8
> data (1 (10 20) 2 (10 20)))
> #s(hash-table size 65 test eql rehash-size 1.5 rehash-threshold 0.8
> data (1 (3 20) 2 (3 20)))
nil

Cordially Yours,

-- 
---------------------------------------------------------------------
　KAWABATA, Taichi　　　E-mail: kawabata.taichi@gmail.com


[-- Attachment #3: Type: message/rfc822, Size: 2520 bytes --]

From: Stefan Monnier <monnier@IRO.UMontreal.CA>
To: 川幡太一 <kawabata.taichi@gmail.com>
Subject: Re: bug#5082: 23.1.50; print-circle and make-hash-table
Date: Mon, 30 Nov 2009 16:08:38 -0500
Message-ID: <jwv8wdnencg.fsf-monnier+emacsbugreports@gnu.org>

> I've found an Emacs behavior that seems a bug concerning to `print-circle'
> and make-hash-table, so I would like to report it.

Thanks for catching this bug and for the clear testcase.  I've installed
the patch below which should fix it.


        Stefan


=== modified file 'src/print.c'
--- src/print.c	2009-11-24 15:30:54 +0000
+++ src/print.c	2009-11-30 21:03:36 +0000
@@ -1420,6 +1420,13 @@
 	    size &= PSEUDOVECTOR_SIZE_MASK;
 	  for (i = 0; i < size; i++)
 	    print_preprocess (XVECTOR (obj)->contents[i]);
+	  if (HASH_TABLE_P (obj))
+	    { /* For hash tables, the key_and_value slot is past
+	        `size' because it needs to be marked specially in case
+	        the table is weak.  */
+	      struct Lisp_Hash_Table *h = XHASH_TABLE (obj);
+	      print_preprocess (h->key_and_value);
+	    }
 	  break;
 
 	default:

Re: bug#5082: 23.1.50; print-circle and make-hash-table

[Ted Zlatanov]

On Tue, 1 Dec 2009 00:09:30 +0900 川幡太一 <kawabata.taichi@gmail.com> wrote: 

> Dear Emacs Developers,
> I've found an Emacs behavior that seems a bug concerning to `print-circle'
> and make-hash-table, so I would like to report it.

> By setting `print-circle' to `t', printing Lisp Object with recursive
> or repetitive
> list should be displayed as #N= and #N# syntax.

> However, it seems this is not true when printing lisp objects within
> hash tables
> (which become possible in Emacs 23.1.50.)  Following is an example.

> Is it possible to display the repetitive lisp object within printed hash tables
> by #N= and #N# syntax, when `print-circle' is set to `t'?

> I notify this behavior as a bug, but it may not be so.  I apologize in that
> case.

> (let ((print-circle t)
>       (x (list 10 20))
>       (y (make-hash-table)))
>   (insert (format "%S\n" (list x x)))
>   (puthash 1 x y)
>   (puthash 2 x y)
>   (insert (format "%S\n" y))
>   (setcar x 3)
>   (insert (format "%S\n" y)))

> Result:
 (#1=(10 20) #1#)
 #s(hash-table size 65 test eql rehash-size 1.5 rehash-threshold 0.8
 data (1 (10 20) 2 (10 20)))
 #s(hash-table size 65 test eql rehash-size 1.5 rehash-threshold 0.8
 data (1 (3 20) 2 (3 20)))
> nil

I looked at the source in src/print.c.  At the print_object() level the
behavior is correct for lists but fiendishly complicated so I couldn't
figure out the parallel structure for hashtables quickly.  I'll put it
on my TODO list if no one else is interested.

Ted